''i'ii^^iiit!;;; ■^Kuiiv-;;;;'!-;!!);::;?:;^^ mmi-tti'.r :i:';:H;.,:.x: •«s^;rjJ3] ;iHa»ai^ wi? ^^m '%^' •^v" .'f-^ ^ -^^ i^^^ -^. .A^ ,0 o ■='/ " ■>:-y^: \^°. a- CO .^■^" "^.- S^' .-^.^ ■^ o,"i' ,v5 ' -J^. c'b' "^^ v*^ .,0 •^. e,-^'- .^"^ •^*. / Ube. Science Series EDITED BY professor 5. mcTkccn CattcII, /ID.H., pb.S). AND 3f. TE, 3Bc^^ar^, '^.H., df.tR.S. RIVERS OF NORTH AMERICA RIVERS OF NORTH AMERICA A READING LESSON FOR STUDENTS OF GEOGRAPHY AND GEOLOGY BY ISRAEL C. RUSSELL PROFESSOR OF GEOLOGY IN THE UNIVERSITY OF MICHIGAN AUTHOR OF "lakes OF NORTH AMERICA," "GLACIERS OF NORTH AMERICA," "volcanoes OF NORTH AMERICA," ETC. NEV/ YORK G. P. PUTNAM'S SONS LONDON JOHN MURRAY 1898 \^\^ e) 18419 Copyright, i3q8 BY G. P. PUTNAM'S SONS 2^0 ow. . •-^ "cocivtU. 2ncf COPY, 1G33. Zbc f^nicfccrbockcr prcas, lAcw )i?ocfe "Every river ap])ears to consist of a main trunk, fed from a variety of branches, each running in a valley proportioned to its size, and all of them together forming a system of valleys, communicating with one another, and having such a nice adjustment of their declivities, that none of them join the principal valley either on too high or too low a level ; a circumstance which would be infinitely improbable if each of these valleys were not the work of the streams that flow through them." — Illusti-ations of the Hiittoniaii Theory of the Earth : by John Flay fair. Edinburgh, 1S02, p. 102. TO THE READER EVERY person is familiar with the beating of the rain upon the surface of the land, and the gathering of the waters that fall into rills, rivulets, and brooks which fre- quently unite to form larger rivers. Everyone is aware, also, that streams are turbid after heavy rains. But, al- though these facts have been known to us from childhood, yet comparatively few people have thought out the chain of events of which they form a part, or recognised the results toward which they lead. Standing by the side of a river, we see its waters flowing continually in one direction and in many instances bearing along a load of sediment. We know that the mud which discolours the waters was derived from the lands bordering the stream and is journeying to the sea. So far as we can ordinarily discern, there is no compensation for this re- moval. Evidently, if the process goes on without being •counteracted by other agencies, all of the material forming land areas will in time be removed, and the hills and even the grandest mountains will be degraded to the level of the sea. We know of no reason why this process of soil removal may not have been in operation since rain first fell on land, or why it may not continue so long as continents and islands exist. As the land has not been reduced to sea-level, one VI TO THE READER of two conclusions seems evident : either the time that has elapsed since the process began has not been sufficiently long to bring about the final result, or else there is some compensating process by which land areas are renewed. A person who wanders along a river bank with these and kindred thoughts in mind naturally seeks for evidences of the work the rivers have accomplished and endeavours to learn how it has been carried on. One observes the river flowing through a narrow, steep-sided trench, or perhaps meandering in broad, graceful curves over the bottom of a wide, fruitful valley. In all directions the view is limited by hills or steeply ascending slopes. On gaining a command- ing station on a hilltop, the broader prospect, including both hills and valleys, very likely will reveal the fact that the hills, ridges, and more or less isolated peaks rise to the same general height, and appear as a level plain w^hen the eye, nearly on a level with their summit, ranges over them; and sunken in this plain is the river valley. If the river has been engaged for ages in carrying away the material of the land in the way it is now doing, the valley must represent a part or the whole of the work done. The thought that the river is older than the valley, and that the valley has been excavated by the river, comes to one like a revelation. In fancy we see the valley filled with rocks like those in the bordering hills and the plain restored. Evidently the sur- face of the region, less diversified than now, must have been a plain or plateau before the stream excavated its valley. When this idea has taken root in the mind, our powers of observation are stimulated and our faith strengthened in what has been termed the scientific use of the imagination. TO THE READER VI i As our vision ranges over valleys and hills, the fact is recog- nised that the neighbouring mountains are but uplands of larger size, separated one from another by gorges and val- leys, in each of which a stream is flowing, and we are startled by the vividness of the pictures that crowd them- selves on our fancy, each portraying a difl'erent stage in the development of a landscape which had previously charmed us simply by its assemblage of attractive forms and its harmonious blendings of colour. The mental pictures of the loiterer by a stream-side or the dreamer on a hilltop are not confined to the region im- mediately before him. What is true of the stream at his feet must also pertain in general to other streams in what- ever clime. Again, a curtain is lifted and he sees that every stream on the earth's surface is engaged in changing the as- pect of the land. Valleys have been excavated on every continent and island. Every mountain has been sculptured. Changes due to running water are everywhere in progress. If these conclusions are well founded, it is evident that valleys and mountains are but transient forms in a long process of topographical development, and the history of past changes should find expression in the relief of the land. The leading idea which absorbs the attention when a living interest is once awakened in the meaning of the many and diversified features of the earth's surface, is that they are not fixed and changeless forms, but have undergone many orderly modifications in the past and will continue to change under the action of definite laws in the future. It is not the shape of the earth as it exists to-day, the present VI li TO THE READER distribution of land and water on its surface, or the relief of the land, or of the floor of the sea, but the changes that each of these conditions has passed through in order to reach its present state, and the modifications still in progress, which claim the greatest share of the geographer's attention. Evolution is the leading theme of the inanimate as well as of the animate world. The features of the earth's surface, from the continents and oceans to the smallest islands and tiniest rills, all have what may be termed their life histories. It is the recognition of this fact that has given new interest and imparted a fresh impetus to geographical study. When once the idea is grasped that each and every one of the elements in a landscape has a history w^hich can be read, and that the end is not yet, but still other transformations are to come, an insatiable desire is awakened for more knowledge concerning especially the work of the streams to which so many of the changes that have been made on the earth's surface are due. What laws do the streams obey ? What conditions modify their normal behaviour ? What are the various stages in the transformations they are making everywhere about us ? Are there other forces in action, tending to counteract their destructive work ? Are the conditions to which man has become adjusted to pass away ? What changes are to come ? Is the re-modelling of the land to be continued for ever, or will there be a final condi- tion beyond which the expression of the face of Nature will become sphinx-like and unchanging ? These and many more queries crowd themselves on the mind when once an interest in the daily scenes about us is awakened. Many, but by no means all, of the questions which we wish to ask TO THE READER IX of the mountains and streams can be satisfactorily answered at the present day. It is with the hope of assisting the reader both in ques- tioning the streams and in understanding their answers, and at the same time creating a desire for more light on other and related chapters of the earth's history, that the book before you was written. The study of the earth's surface should be of especial in- terest to American students not only because of the mag- nificent and varied scenery of our native land, but for the reason that new life and vividness have been given the sub- ject by the labours of men who are still among us. The marked advances made during the present century both in the study of the ancient life on the earth and of the surface changes still in progress, show the influence of en- vironment. The Geological Survey of the State of New York gave a marked impetus to the study of the invertebrate life of distant ages, largely for the reason that the rocks of the New York series are rich in such relics. When geolo- gists visited the central and western portions of the United States, the sediments of ancient lakes, rich in vertebrate fossils, were discovered. The veritable menagerie of mar- vellous birds, reptiles, and mammals that has been made to appear from these cemeteries is more varied than the most fantastic dreams of fable. We feel that the uncouth pro- cession has only begun to pass in review, but are at a loss to imagine what as yet unknown commingling of fish, rep- tile, bird, and mammal in one and the same individual can possibly be found. When investigators of surface geology and geography made their bold explorations into the vast X TO THE READER arid region of the south-west, they discovered a land of wonders, where the mask of vegetation which conceals so many countries is absent, and the features of the naked land are fully revealed beneath a cloudless sky. The facts in the earth's history which there impress themselves most forcibly on the beholder are such as have resulted from the action of streams and of atmospheric agencies. It was in this arid region of strong relief that a revival of interest in the surface forms of the earth was engendered. The seeds of what is practically a new science, — physiography, — gathered in this desert land by J. S. Newberry, J. W. Powell, G. K. Gilbert, C. E. Button, and others, when carried to other regions bore abundant fruit. It was found that the surface of the land, when once suggestion had en- abled men to see topographic forms and interpret their meaning, is a manuscript on which wonderful events are recorded. A younger generation of active workers has extended the study of the earth's surface, so greatly stimulated by the pioneer explorers just named, and has read for us the his- tories of valleys, plains, hills, and mountains throughout the length and breadth of the land. Of these younger investi- gators we are indebted to none so much as to W. M. Davis, Professor of Physical Geography at Harvard, who has made New England, New Jersey, and Pennsylvania classic ground to all future students of physiography. Gilbert has ex- tended his studies to the basins of the Laurentian lakes and other regions. The writings of J. C. Branner, IM. R. Camp- bell, T. C. Chamberlin, N. H. Darton, J. S. Diller, C. W. Hayes, Arthur Keith, W J McGce, R. D. Salisbury, R. TO THE READER XI S. Tarr, Bailey Willis, and others, have greatly enlarged our knowledge of the laws governing streams, and of the origin of topographic forms. The publications of the United States Geological Survey and of the earlier national surveys of which it is a continua- tion, the Geological and Natural History Survey of Canada, various State surveys, the Geological Society of America, and the National Geographic Society, together with the Journal of Geology^ the American Journal of Science j etc., are the sources of published information drawn on most largely in the preparation of this book. In the following chapters many references are given to the writings of the distinguished investigators named above, but all of the help derived from them while writing this book can scarcely be acknowledged. Much assistance has also been derived from conversation and correspondence with my colleagues, and, although fully appreciated, the portion derived from each one is scarcely known even to myself. My part in presenting this book is largely that of a guide who points out the routes others have traversed. My reward will be ample if even a few students follow the paths indicated and are led to explore their many as yet unknown branches. Israel C. Russell. University of Michigan, December lo, 1897. CONTENTS PAGE To THE Reader v CHAPTER I The Disintegration and Decay of Rocks Mechanical Disintegration — Chemical Disintegration or Rock Decay — Removal and Renewal of Surface Debris. CHAPTER n Laws Governing the Streams ........ 12 How Streams Obtain their Loads — Transportation — Debris Carried by Ice — Corrasion — Pot- Holes — Lateral Corrasion — Meandering Streams — Other Curves — Deflection of Streams Owing to the Earth's Rotation — Questioning the Rivers — Erosion — Baselevel of Erosion — Peneplains — Influence of Vegetation on Erosion. CHAPTER III Influence of Inequalities in the Hardness of Rocks on River- side Scenery 52 Waterfalls — The Migration of Waterfalls — Bluffs Bordering Aged Streams. CHAPTER IV Material Carried by Streams in Suspension and in Solution 67 The Visible Loads of Streams — Bottom Load — Measures of Material in Suspension — The Invisible Loads of Streams — Rate of Land Degradation — Mechanical Degradation — Chemical Degradation — Rate of Both Mechanical and Chemical Degradation — Underground Streams. XIV CONTENTS CHAPTER V FACE Stream Deposits 97 Alluvial Cones — Talus Slopes — Flood-Plains — Natural Levees — Deltas : Deltas of High-Grade Streams — Deltas of Low-Grade Streams — Effects of Changes in the Elevation of the Land on the Growth of Deltas — Variations in Normal Stream Deposition — Influence of Elevation and Depression of the Land on Stream Deposition — Influence of Variations in Load on Stream Deposition — Influence of Changes of Climate on Stream Deposition — The General Process of Stream Corrasion and Deposition — Profiles of Streams — The Longitudinal Profile — Cross-Profiles. CHAPTER VI Stream Terraces Origin of Terraces during the Process of Normal Stream Develop- ment — Terraces Due to Climatic Changes — Terraces Due to Eleva- tion of the Land — Bottom Terraces — Delta Terraces and Current Terraces — Glacial Terraces — Relative Age of Terraces — Other Ter- races — General Distribution of Stream Terraces. CHAPTER VII Stream Development 184 Consequent Streams — Subsequent Streams — Ideal Illustration of Stream Adjustment and Development — Examples of Stream Develop- ment and Adjustment in the Appalachian Mountains — Influence of Folds in the Rocks on Stream Adjustment — Water-Gaps and Wind- Gaps — Stream Conquest — Ancient Peneplains — Synclinal Mountains and Anticlinal Valleys — Effects of Elevation and Subsidence on Stream Development — Some of the Effects of Elevation — Some of the Effects of Subsidence — Some of the Influences of Volcanic Agencies on Stream Development— Some of the Modifications in Stream Development Due to Climatic Changes — Variations in Pre- cipitation — Variations in Temperature — Fluctuations of Streams — Some of the Influences of Glaciers on Stream Development — Some of the Influences of Vegetation on Stream Development — Driftwood — Superimposed Streams — Migration of Divides. CONTENTS XV CHAPTER VIII PAGE Some OF THE Characteris lies OF American Rivers . . . 254 Drainage Slopes : Atlantic, St. Lawrence, Hudson Bay, Arctic, Bering, Pacific, Great Basin, Gulf, and Caribbean — Leading Features of the Several Drainage Slopes — New England Rivers — A Drowned River — Appalachian Rivers — Rivers of Glaciated Lands — Southern Rivers — Alluvial Rivers — The Mississippi — Canyon Rivers — Sierra Nevada Rivers — "Where Rolls the Oregon" — Rivers of the Far North-West — Glacier-Born Rivers — Arctic Rivers — Rivers of the *' Great Lone Land" — Rivers Flowing to Fresh- Water Seas — Niagara — Retrospect. CHAPTER IX The Life History of a River ........ 301 Index 321 ILLUSTRATIONS IN THE TEXT FIGURE 1. A POT-HOLE BEING SCOURED OUT BY A STREAM 2. A Young Stream near Ithaca, New York 3. Profile of Niagara Falls 4. Cross-Profile of a Floodplain .... 5. Map of the Lower Mississippi showing Crevasses 6. Radial Section of a Delta .... 7. Longitudinal Profile of a Young Stream 8. Successive Changes in the Profile of a Divide 9. Ideal Profile of a Divide 10. Cross-Profile of a Terraced Valley 11. Alluvial Terraces 12. Alluvial Terraces 13. Cross-Section of a Valley with Terraces in Solid 14. Cross-Section of a Current-built Terrace 15. Section of tilted Peneplain .... 16. Sketch-Map, showing Young Streams 17. Sketch-Map, illustrating Stream Development 18. Sketch-Map, showing Mature Streams . ig. Anticlinal and Synclinal . 20. Map illustrating River Piracy 21. Section through Lookout Mountain 22. Map of Chesapeake Bay 23. Cross-Profile of Colorado Canyon Table A. Analysis of American River-Waters etc., Alabama Rock facing PAGE 33 55 60 118 119 126 147 148 149 152 156 157 165 168 186 187 189 190 198 200 212 219 272 78 FULL-PAGE ILLUSTRATIONS PLATE FACING PAGE I. a. Marion River, New York. b. Ingall's Ceeek, Wash- ington 12 II. Views on the Yukon, Alaska 24 III. a. Ray Brook, Adirondacks, New York. b. Moccasin Bend, Tennessee River 38 IV. a. Fall on Black Creek, near Gadsden, Alabama, b. Echo River, Mammoth Cave, Kenjucky ... 60 V. Map of the Delta of the Mississippi 98 VI. a. Sketch of Alluvial Cones, b. Indian Creek, Cali- fornia 102 VII. a. Big Goose River, Wyoming, b. New River, Tennessee, 108 VIII. a. Terraces on Fraskr River, British Columbia, b. Ter- races in Connecticut Valley . . . . . 154 IX. Map of the Northern Appalachians 196 X. Map of Western Portion of the Anthracite Basin, Pennsylvania 204 XI. Map illustrating Stream Adjustment . . . .210 XII. a. Beaver Dam, Wyoming, b. Dam of Drift-wood, West Fork of Teanaway River, Washington ... 238 XIII. Map of a Portion of the Catskill Moun-^ains, New York, 250 XIV. Map of North America showing Drainage Slopes . . 256 XV. a. Columbia River, b. Hudson River .... 262 XVI. a. An aggraded Valley near Fort Wingate, New Mexico, b. Shenandoah Peneplain, near Harper's Ferry, West Virginia 26S XVII. Canyon of the Colorado 274 RIVERS OF NORTH AMERICA CHAPTER 1 THE DISINTEGRA TION AND DEC A Y OF ROCKS THE study of rivers, from the point of view of the geo- grapher, necessitates the consideration of the nature and origin of many topographic forms; the reason being that streams are among the most important agencies which give form and expression to the surface of the land. The study of streams, therefore, involves, to a great extent, the consideration of the origin of hills and mountains, plains and valleys, and the changes they pass through. One of the principal tasks performed by streams is the moving of rock fragments and their transportation to the sea. Another function of streams is the deepening and widening of their channels and valleys. These propositions will be demonstrated later. It will be shown, also, that clear water has but little power to wear away the rocks over which it flows. In order to do this, the flowing water must be charged with hard particles or rock fragments of greater or less size. That is, the streams must be supplied with tools with which to excavate. It is now well under- 2 RIVERS OF NORTH AMERICA stood that the tools used by streams in abrading the rocks are mainly silt, sand, gravel, and stones, which are carried in suspension or rolled and pushed along the bottom. One of the primary questions, therefore, in order to understand how streams are enabled to remove material from the land, and, in so doing, to deepen and broaden their valleys, is: How are the rocks broken or otherwise prepared for stream transportation ? The soil which nearly everywhere forms the surface of the land is composed mainly of disintegrated rock. This loose surface layer of more or less comminuted and decayed material, much of it, however, far too coarse to be termed soil, is the storehouse from which the streams derive the principal part of their loads. The study of the agencies at work in breaking and other- wise disintegrating the earth's crust has shown that they may be classified in two groups: 1st, those acting mechani- cally; and 2d, those whose influence is principally chemical. Although the various agencies in these two groups co- operate and are frequently in action at the same time, it is convenient to consider them separately. Mechanical Disintegration. — Changes of temperature, as between day and night, or from season to season, cause un- equal expansion and contraction of the minerals and grains of which rocks are composed. Various and complex stresses are thus produced which cause even the most compact granite to crumble. The freezing of water contained in crevasses in rocks or in the interspaces between grains or crystals, is accompanied by expansion, which exerts a powerful force tending to fracture and disintegrate them. DISINTEGRATION AND DECAY OF ROCKS 3 The roots of trees enter crevices in the rocks, and as they enlarge, force off fragments frequently of large size. The undercutting of cliffs and banks by streams and by the waves and currents of lakes and of the ocean, causes the dis- lodgment of vast quantities of earth and stone. The fall of rocky material produced by these and still other causes, leads to still further breakage. Rock masses are also loos- ened or caused to fall by earthquake shocks. Volcanoes discharge a great volume of fragmental material into the air, and the cooling of lavas causes them to become frac- tured and jointed. When molten lava enters surface-water bodies, steam and gas explosions occur, and the rock is perhaps blown to dust. Rain-drops, snow crystals, and hail by beating on the rocks exert a force tending to break off fragments loosened by other and principally chemical agencies, and to wear, and frequently to polish, exposed surfaces. Sand and dust, blown by the wind, on coming in contact with rock exposures, wear away the softer parts and loosen the harder grains and crystals. Glaciers as they flow down mountain valleys or move over the surface of more level land, tear away projecting ledges, and when charged with sand and stones abrade and grind away the rock over which they move. Avalanches and landslides rush down declivities, carrying destruction in their paths, and sweep along loosened rock fragments which are broken still finer and in part ground to powder. The streams themselves, under certain conditions, roll along stones and even large boulders, which become rounded and broken, at the same time abrading the rocks over which they are carried, and thus aid in the general process of rock disintegration which 4 RIVERS OF NORTH AMERICA prepares the material composing the land for stream trans- portation. All of the agencies just enumerated are mechanical in their action, although accompanied by chemical changes, and are confined to the surface, or, at most, to an extremely superficial portion, of the earth's crust. There are also im- portant mechanical agencies which act deep below the sur- face and lead to the fracturing of the rocks in such manner as greatly to facilitate the agencies producing disintegration in operation at the surface. And, besides, on account of the continual lowering of the surface in many regions owing to the removal of material, rock fragments originating at a greater or less depth become mingled with those produced at the surface in the several ways just enumerated, and thus become of interest in the study of the manner in which rocks are reduced to fragments of such size that they can be moved by streams. Of the mechanical agencies leading to the fracturing of rocks below the reach of frost and of normal changes in temperature, the most important are movements in the earth's crust, the nature of which it is impracticable to dis- cuss at this time, which cause even the most massive layers to become folded and broken. These movements are fre- quently accompanied by the crushing of the rocks in zones of various widths, as when a fracture is formed and its walls ground against each other. Such breaks, accompanied or followed by differential movements of their wails, are termed faults, and the rock fragments produced are desig- nated fault breccias. The world over and to a great but indefinite depth, the DISINTEGRATION AND DECAY OF ROCKS 5 rocks are divided by what are known as joints, the origin of which is obscure. These dividing planes are similar, we may fancy, to gashes made by a sharp blade without appre- ciable thickness, drawn through the rocks. There are fre- quently two series of joints nearly at right angles to each other, and more or less nearly vertical ; these are intersected many times by approximately horizontal cuts of the same character, and frequently also by planes of bedding. The rocks are divided in this manner into masses that are some- times nearly true cubes. In many instances the joints cross each other irregularly and divide the rocks into blocks of many shapes. Joint blocks of whatever form vary in size, from a small fraction of a cubic inch to many cubic feet. In some instances these are contributed directly to streams, as when they fall from the face of a precipice, but more com- monly they are broken and variously modified by atmos- pheric agencies before being fed to the flowing waters. The joints in rocks, although of inappreciable width deep below the surface, are planes of weakness along which chemical and mechanical agencies find favourable lines of attack. They open when the rocks are exposed to the weather, and greatly favour the further disintegration re- sulting from changes of temperature, the freezing of water, etc. The jointing of rocks is one of the primary and most important methods by which they become divided into blocks, thus exposing greater surfaces to the attack of chemical agencies, and in many regions, particularly in rugged mountains and in canyon walls, exerts a direct and pronounced influence on topographic forms. Among the agencies that lead to the fracturing and me- 6 RIVERS OF NORTH AMERICA chanical disintegration of rocks deep below the surface, should be noted, also, injections of molten rocks forced up- ward into the earth's crust, earthquake shocks, the friction of debris carried by subterranean streams, and the falling of cavern roofs. Still another agency, as has been pointed out by G. P. Merrill, in part chemical and in part mechanical in its action, results from the combination of water with certain mineral substances, producing what is termed hydration. This is accompanied by an increase in the bulk of the min- erals affected, and the consequent production of stresses in the rocks containing them. In some instances, apparently unaltered rock, when removed from mines and tunnels, rapidly crumbles from this cause when exposed to the air. In nature, the lowering of the surface by erosion and the exposure of previously deeply buried rocks would bring about similar changes. There are yet other alterations in progress in the rocks due to chemical action, that promote mechanical disintegration, which cannot be noted at this time. By the several processes just enumerated, the rocks are broken into blocks of all shapes and dimensions, some of which are of the size of gravel, sand, and dust grains, and are thus rendered. suitable for stream transportation, and to act as tools by means of which flowing water promotes the process of rock breakage. Chemical Disintegration or Rock-Decay, — Water is a solv- ent for probably all substances that occur in the earth's crust, although in many instances acting with extreme slow- ness. The readiness with which most substances are taken into solution by water is enhanced by an increase of tem- perature, and in nature is also greatly assisted by various DISINTEGRA TION AND DEC A V OF ROCKS 7 substances, especially organic acids, with which it becomes charged. Even rain-water is never pure, but contains various salts and gases derived from the air. Principal among these is carbon dioxide, or carbonic acid. Rain-water on reaching the earth flows over the surface, or percolates for a time through the soil and rocks, and thus comes into intimate re- lations with the great store of organic acids supplied by the waste and decay of animal and vegetable life. The chemi- cal energy of the water is thus greatly enhanced, and it becomes an active solvent for most mineral substances. Some of the minerals composing rocks are more soluble than others, and, being removed, allow those that remain to crumble and fall apart. The mineral substances taken in solution are, for the most part, contributed to streams and by them carried to the sea as an invisible load ; but a por- tion is taken below the surface by downward-percolating waters and undergoes many changes in composition and at the same time produces various alterations in the rocks through which it passes. The chemical changes produced in the rocks by percolat- ing waters, while most active near the surface, occur also at considerable depths, and are there augmented by the inter- nal heat of the earth. There is a lower limit to this process, however, due to the increasing density of the rocks with pressure and to the rise of temperature with increase in depth. There are good reasons for concluding that surface waters cannot descend more than twenty thousand or thirty thousand feet below the surface. The chemical changes due to percolating water are influ- 8 RIVERS OF NORTH AMERICA enced in a variety of ways by temperature. The rocks are dissolved most readily in warm, moist regions. It is in such regions also that vegetation is most luxuriant and animal life most abundant, and hence the waters are most highly charged with organic acids. Chemical action, in most in- stances, is retarded by cold; vegetation is less abundant and decay less rapid in cold than in warm climates; it is, therefore, in cold regions that the decay of the rocks is at a minimum. In warm, humid countries, deep rock-decay has usually taken place, but a thick surface sheet of decomposed ma- terial is not necessarily found, as the loosened debris may be carried away as fast as it is produced. In the southern Appalachians, and in many other warm temperate or equa- torial regions, the rocks are so broken and decayed, even at a depth of one hundred and fifty or two hundred feet from the surface, that they may be crumbled between the fingers or moulded like clay. In such instances the soil usually shows various tints of red and yellow, the colours being due to the oxidation and hydration of iron present in them. In warm, dry countries chemical changes in the surface material are retarded, although the rocks may be greatly shattered by changes of temperature. In such regions the soils are seldom red. Chemical changes produced by percolating water below the superficial portion of the earth — that is, below, per- haps, one hundred feet — increase with depth, on account of progressively increasing temperature, but these changes are beyond the immediate subject under discussion. An important agency in rock disintegration and decay having DISINTEGRA TION AND DEC A V OF ROCKS g its source deep within the earth, however, is manifest espe- cially in volcanic regions where steam charged with various acids rises through fissures and other openings. During volcanic eruptions, but more particularly after a volcano has passed to the condition of a fumarole or a solfatara, heated vapour and gases charged with sulphuric, hydrochloric, car- bonic, and other acids escape in large volumes, sometimes continuously for centuries, and produce conspicuous changes in the rocks through which they rise. Similar but usually less copious exhalations occur from lava streams, and pro- duce alterations in the lava which influence the character of the soil resulting from them. The chemical alterations produced by percolating water, and less commonly by volcanic gases, in rocks near the surface, are in part by solution, and in part oxidation, hydration, precipitation, etc. These changes, except the last mentioned, may be grouped, at least in a general way, under the term rock-decay. In decaying, the rocks are more or less disintegrated, however, since the more soluble minerals are removed, thus allowing the less soluble rock constituents to crumble and fall apart. The processes of rock-disintegration and rock-decay mutually assist each other, and progress at the same time. The result is that the surface layer of the earth's crust is profoundly altered, and a sheet of modified material is produced, which is designated in part as soil and in part as rock detritus.* Tlie Removal and Renewal of Surface Debris. — The sur- ' The name Regolith, meaning blanket-stone, has recently been proposed for the superficial material covering the earth, by G. P. Merrill, A Treatise oh Rocks, Rock- Weathering and Soils^ 1897, p. 299. lO RIVERS OF NORTH AMERICA face changes just considered have been in progress since the first appearance of land, and will continue as long as conti- nents and islands exist. In former geological periods the agencies enumerated, particularly those of a chemical nature, were more active than now, and have varied from time to time, in probably all portions of the earth's surface, with climatic and other changes. Throughout all geological ages, the streams have been actively engaged in removing the dis- integrated and more or less chemically altered surface por- tions of the earth's crust. In places and at certain times, the debris has been removed as fast as formed, and bare, hard rock surfaces have been exposed ; at other times, the supply has been in excess of the demand, and deep accumu- lations have resulted. The surface sheet of debris has been continually wasting and continually renewed. Throughout the history of the earth, topographic changes have been in progress. Mountain ranges and systems have been upraised, the rocks composing them fractured and chemically altered, and borne away by streams. Where once a magnificent mountain range reared its battlements among the clouds, there is now a plain but little elevated above the sea. Not only one, but several such geographical cycles have run their courses in many lands. During the cycles still in progress human agencies have been added to those previously in action. This new element in the earth's history has become more and more important as man has advanced in civilisation. In part, human indus- tries have retarded the work of physical and chemical agen- cies, but in the main man has been a destroyer. The removal of the portion of the earth's crust rising DISINTEGRA TION AND DECA V OF ROCKS 1 1 above the sea, during each cycle, has been done almost wholly by streams. The manner in which the rocks are prepared for transportation, however, is quite as important to the geographer as the methods employed for their re- moval, but the brief review given above of this division of the general process must suffice for the present. The student who may wish to continue the studies out- lined in this chapter will find assistance in the books men- tioned below. ^ ^ George P. Merrill. A Treatise on Rocks, Rock- Weathering and Soils, pp. 172-398. The Macmillan Co., 1897. Israel C. Russell. The Decay of Rocks and the Origin of the Red Colour cf Certain Formations. U. S. Geological Survey, Bulletin No. 52, 1889. Israel C. Russell. "A Reconnoissance in South-Eastern Washington." U. S. Geological Survey, Water-Supply and Irrigation Papers, No. 4, pp. 57- 69, 1897. George P. Marsh. The Earth as Modified by Human Action. Charles Scribner's Sons, 1885. John C. Branner. '* Decomposition of Rocks in Brazil." Bulletin of the Geological Society of America, vol. vii., pp. 255-314, 1896. Alexis A. Julien. "On the Geological Action of the Humus Acids," in American Association for the Advancement of Science, Proceedings, vol. xxviii., pp. 311-410, 1879. Walter Maxwell. Lavas and Soils of the Hawaiian Islands. Honolulu, i8q8. CHAPTER II ZAIVS GOVERNING THE STREAMS THE water which flows off from the land, as is well known, is supplied by the condensation of vapour in the air. A part of the water reaching the earth flows over the surface and gathers into rills which unite to form larger streams, and a part sinks below the surface and, after follow- ing an underground course, usually by percolation through porous soil or rocks, emerges in springs, many of which join the surface flow. It is also well known that streams ranging in size from the smallest rills to the mightiest rivers are engaged either oc- casionally, as during floods, or continually, in carrying away material that was previously a portion of the land. The manner in which this material is acquired by the streams, the way it is transported, the effects it has on the flow of the streams, and on their bottoms and sides, the modifica- tions in the configuration of the surface of the land due to the removal and re-deposition of debris, etc., are all phe- nomena that obey definite laws and are variously modified by conditions. If one can ascertain the laws governing the behaviour of a single stream, they should also apply not only to the streams of North America, but to those of all land areas. 12 i Plate I. Fig. a. Marion River, Adirondacks, New York. Summer stage, showing stones which are moved during high-water. (Photograph by S. R. Stoddard.) Fig. B. Ingall's Creek, Washington. Showing boulders too large for the stream to move even during high-water stages. LAWS GOVERNING THE STREAMS 1 3 No one stream, perhaps, in the limited time that an in- dividual student is enabled to examine it, will furnish illus- trations of all of the modifying conditions influencing the life of a great river. By selecting typical examples, how- ever, affected by different modifying conditions, we may sketch a composite picture which will represent the various phases in the life history of a single river that has carried on its work for tens of thousands of years. How Streams Obtain their Loads, — Rain-drops strike the earth with a certain force, dependent on their size, the dis- tance they descend, the direction and force of the wind, etc. If rain-drops fall on the surface of a still pool we may see them rebound. If we face a rain-storm, the sting of the beating drops again assures us that they exert a consider- able force on the objects against which they strike. When the drops fall on a solid rock surface, they gather in rills of clear water, but if they fall on loose soil, as a newly ploughed field, for example, the finer particles of earth are disturbed, and as the waters gather into rills and flow away in obedience to gravity, they are turbid with earth particles held in suspension. The turbid rills unite in brooks, and these again combine to form larger streams. The fine silt disturbed by the impact of the rain-drops is carried by the rills to the brooks and thence onward, perhaps with many halts, to the sea. After heavy rains even large rivers be- come muddy. The lakes and large areas in the sea near the mouths of rivers are then discoloured. This happens during every storm the world over, and evidently, if suffi- cient time be allowed, must lead to changes of great mag- nitude both in the topography of the land from which 14 RIVERS OF NORTH AMERICA material is removed and in the shape of the basins where it is deposited. When the surface of the land is dry, and especially when bare of vegetation, earth particles are moved by the wind in much the same manner that streams take up and transport the flakes and grains of rock which they are competent to transport. The dust and sand carried by the wind and falling in streams is another source from which they obtain material suitable for removal. The fine rock powder, or glacial meal as it is termed, produced by the grinding of stones held in the ice against each other and on the rocks over which the glaciers flow, is contributed directly to the waters formed by the melting of the ice. For this reason nearly every glacier-born stream is turbid and heavy with silt. Volcanoes during times of violent eruption discharge vast quantities of fine dust, and in many instances equally abund- ant rock fragments of the size of sand and gravel. When material extruded in these forms falls in streams, or is car- ried in by tributary rills and by the wind, another source is fur- nished from which streams receive their initial loads. There are yet other methods by which streams are supplied with material in a suitable condition to be transported. Among these may be noted the fall of cosmic dust, disturbances produced by avalanches and landslides, the uprooting of trees, the impact of driftwood and floating ice on the bot- toms and sides of stream channels, the movement of roots and overhanging branches by the wind and by the currents of the streams themselves, disturbances of the material forming the bottoms and sides of stream channels by ani- LAWS GOVERNING THE STREAMS I 5 mals, — as by beavers, for example, — contributions of shells and siliceous cases from organisms like mollusks and diatoms living in the waters, etc. Man promotes the transfer of solid matter to the streams in various ways, more especially by ploughing and otherwise disturbing the soil, and by the removal of forests. All of the methods mentioned in this paragraph, however, are of secondary importance in com- parison with the influence of rain, wind, and glaciers. The particles carried in suspension by streams tend to fall to the bottom, being continually pulled down by grav- ity, but in flowing water there are various currents, some of which tend upward and exert an influence on the falling particles in opposition to gravity. The currents in the water move the suspended particles in various directions and retard their fall to the bottom, but the resultant move- ment is in the direction of the flow of the stream. The particles carried by streams fall to the bottom many times during their journeys, and rest there for a period per- haps brief but possibly long, and are again lifted by upward currents and brought within the influence of the onward flow. Material thus transported by a stream may for con- venience be termed its load. Streams not only receive their initial loads in the various ways just stated, but there are other methods by which the same result is reached. As will be considered later, flowing water exerts a pressure on objects against which it strikes. If this force be great enough to move the objects in the path of a stream, they will be pushed along, rolled over, or, with the assistance of upward currents, taken in suspension. The strength of the water current determines the size of the particles it can 1 6 RIVERS OF X OR Til AMERICA carry, so that a stream of a given velocity, but without an initial load, would be able to remove from its channel all of the loose particles it is competent to carry and thereafter would run clear unless its velocity were increased. The principal methods by which streams receive their initial loads insure a waste of the land between the drain- age lines, and consequently this land changes in topographic form. The deepening and broadening of the stream chan- nels is accomplished principally by the friction of the debris carried through them, aided also by solution. The laws governing this complex process will be considered later. Transportation, — The debris acquired by streams in the several ways considered above is carried along by them. A convenient term for this process is transportation. Light objects like leaves, wood, pumice, etc., are floated on the surface, but meet with various delays, and undergo more or less chemical and mechanical changes during their journeys, and sooner or later sink to the bottom. Material like fine sand and silt remains in suspension to a great extent and is carried bodily onward. Heavier objects, like pebbles and boulders, are either rolled or pushed along the bottom, or remain at rest until they are reduced in size by the friction and solution, or shattered by the impact, of material swept against them. The size, weight (specific gravity), and form of the loose material within the influence of a stream deter- mine whether or not it will be moved by a current of a given velocity. The smaller the divisions into which a mass of rock is broken, the larger the ratio of surface to weight. The force which a current of a given velocity exerts against objects in its path varies as the area of the opposing sur- LAJVS GOVERNING THE STREAMS 1 7 face. The smaller the parts into which a rock mass be- comes divided, therefore, the greater the tendency of the current to move them. The ability of the stream to carry debris in suspension, however, depends not only on its velocity and the degree of comminution of the material within its influence, but also, as previously stated, on the presence of secondary and es- pecially of upward currents which tend to lift the particles i brought within their influence. While a particle is in sus- pension the onward currents bear it along, but gravity is all the while acting, and, unless counteracted, finally pulls it to the bottom. The journey of a rock fragment from the mountains to the sea consists of a great number of upward and onward excursions, with rests of greater or less length between. More definitely, the ability of a stream to hold debris in suspension is due to the fact that different layers of water are actuated by different velocities, and these exert different pressures upon the different sides of the suspended particles. Hence, the greater the differences in the veloci- ties of consecutive layers, the greater will be the tendency to hold material in suspension. It is stated by Humphreys/ and Abbot, from whose report on the Mississippi much ofl this discussion of the mechanics of stream flow is taken, that the change in the velocity of the waters of streams in horizontal planes is greatest near the shore and least near the thread of maximum current; and in vertical planes, is greatest near the bottom and surface and least at about one- third of the depth of the stream — that is, where the abso- lute velocity is greatest. If, then, the water be either charged to its maximum capacity or overcharged with sedi- 1 8 RIVERS OF NORTH AMERICA Iment, the highest percentage of material in suspension will be found near the banks and near the surface and bottom, and the least amount near the thread of the maximum cur- rent and at a depth of about one-third of that of the stream. If, however, the water is undercharged with material in suspension, the distribution will not follow any law, the amount at any locality being determined by what may be .considered as accidental swirls, boils, etc. As most streams are undercharged, it follows that samples of water from several points in a cross-section should be examined in order to ascertain. approximately the amount of material that is being carried. Rock fragments too heavy to be lifted may be rolled or pushed along the bottom, or perhaps turned over from time to time by the resultant onward cur- rent. There is thus an adjustment between the strength of the current and the specific gravity of the material trans- ported. It is well known that the power flowing water has to trans- port rock debris increases with increase of velocity. Experi- ments have shown that if water is made to flow through an even channel and the rate of flow is gradually increased by increasing the inclination of the channel, it will move ma- terial added to it approximately as follows ^ : VELOCITY OF CURRENT. SIZE OF MATERIAL MOVED. 3 inches per second. Fine clay and silt. 6 Fine sand. 12 '* " " Pebbles \ inch in diameter. 2 feet ii I *' *' ^ David Stevenson, Canal and River Engineering, p. 315 ; A. J. Jukes- Browne, Physical Geology, 1892, p. 130 ; Archibald Geikie, Text-Book of Geology, 2d edition, 1885, p. 354 ; Joseph Le Conte, Elements of Geology y. 4th edition, 1896, pp. 18-20. See also other elementary works on geology. LAWS GOVERNING THE STREAMS 19 VELOCITY OF CURRENT. SIZE OF MATERIAL MOVED. 2.82 feet per second. Pebbles 2 inches in diameter. 3.46 3 " 4 4 - 4.47 5 " 4.90 6 " 5.29 7 " 5.65 8 " 6 9 " It must be understood that the currents referred to in this table are bottom currents, and in general may be taken at about one-half the central surface current. An important fact shown by these and other sirnilar ex- periments is that the transporting power of running water increases in a greater ratio than the increase in velocity. It has been demonstrated that if the surface of an object opposed to a current of water, as the pier of a bridge, for example, remains constant, the force of current striking it varies as the square of its velocity. Also, that the trans- porting ^^ow^x of a current, or the weight of the largest frag- 1 ment it can carry, varies as the sixth power of the velocity,^ Under this law it will be seen that doubling the velocity of a current increases its transporting power sixty-four times. If a stream flowing with a given velocity is able to move stones weighing one pound, by doubling the velocity boulders weighing sixty-four pounds can be carried ; and if the velocity were increased ten times, rocks weighing one mil- lion pounds could be moved. This enables us to see how streams are capable of producing such striking results during floods, when their velocities are increased on account of an increase in volume. The gradient of a river, or its average ^ A demonstration of this proposition may be found in Joseph Le Conte's Elements of Geology, 4th edition, pp. 19, 20. Appleton & Co., 1896. 20 RIVERS OF NORTH AMERICA fall in a given distance, as a rule, progressively decreases from near its source to its mouth. With this general de- crease in gradient there is a decrease in velocity, and con- sequently a loss in transporting power and a diminution or total check of friction on the stream's bottom. As a result of these conditions, we usually find that the streams are actively engaged in deepening their channels in their upper courses, and are consequently able to extend their branches farther and farther, thus acquiring new territory, and at the same time to deposit material in their lower and less steep courses nearer their mouths. Variations in this process oc- cur not only from season to season but from day to day, on account principally of variation in velocity due to changes in volume. The carrying of debris consumes some of the energy of flowing water. As an extreme example, it is readily seen that an excessive quantity of fine mud contributed to ? stream will entirely check its flow. If but little mud is added, however, it is carried forward without sensibly diminishing the strength of the current. Without attempt- ing to present a complete analysis of the laws governing stream transportation, it will be sufficient at this time to note that streams exert a selective power, taking up and carrying forward the finer and lighter material within their reach, and, if this be sufficient to consume their available energy, leaving the larger and heavier masses, although they may not be too heavy to be removed if the energy of the stream is not otherwise taxed. The principal laws governing stream transportation may be briefly formulated as follows: LAWS GOVERNING THE STREAMS 2 1 1. The greater the slope of a stream channel, the greater the amount of material in suspension the stream can carry; the reason being that the greater the slope the swifter is the flow of the water descending it, other conditions remaining unchanged. The increase in transporting power with in- crease of slope is greater than a single ratio. That is, if the declivity of a stream is double, its transporting power is more than double. 2. An increase in the volume of a stream increases its ability to transport. The greater the volume of a stream, the greater will be its velocity, and the less its loss of power due to friction in proportion to its energy. Here, again, the increase in transporting power is greater than a simple ratio. 3. The capacity of a stream to transport is greater for fine debris than for coarse; for the reason that to move fine material requires less power for the same weight than for coarser material, and, also, when the material is fine a greater portion of the stream's energy can be utilised than when the load is coarse. One of the most important principles connected with stream transportation is that flowing water assorts the debris delivered to it. Fine particles are more easily carried than coarser ones of the same specific gravity, and are first removed. This is true both of particles in suspension and of material rolled along the bottom. If the fine ma- terial is sufficient to consume the available energy of a stream, all coarser debris is left until its energy is increased, as during storms, or until the fragments too large to be re- moved are reduced in size. There is a delicate adjustment between the velocity of a stream and the size of the debris 22 RIVERS OF NORTH AMERICA it can carry, which may be termed the selective power of currents. The influence of this selective power is seen not only in the character of the material moved by streams, and in the debris left on their bottoms, but in the deposits which they make, whether on their border, or in the lakes and sea to which they contribute their loads. As most sedimentary rocks are formed of stream-born debris, this assorting pro- cess must evidently be of vast geological importance. Debris Carried by Ice, — An interesting and at times an important factor in stream transportation is the assistance furnished by ice, which frequently enables streams to move objects that would otherwise exceed their power. During winter the water of streams frequently freezes to the bottom, more especially along their margins, and stones, gravel, sand, etc., forming the beds of their channels, be- come firmly attached to the ice. In spring, when the streams are swollen, the ice, on account of its buoyancy, breaks away from the bottom, but frequently retains large quantities of debris, which is carried with it down-stream, and may make a long journey before being dropped or de- posited by the stranding of the ice.^ Stones carried in this manner are frequently of large size. A rudely spherical boulder measured by me on the bank of the Yukon, which had certainly travelled scores of miles from its parent ledge, was a little over six feet in diameter. Many others very nearly as large were seen which had recently been forced several yards up the banks of the river. All of these were ^ An account of the method of transportation here discussed may be found in Lyell's Principles of Geology (nth edition, vol. i., pp. 359-363, Appleton & Co., 1873), accompanied by an illustration of large boulders along the shores of the St. Lawrence, which had been moved through the agency of ice. LAWS GOVERNING THE STREAMS 23 far beyond the reach of former glaciers, and, without ques- tion, had been deposited in their present positions at a very recent date; some of them, in fact, during the floods of the preceding spring. Another method by which the ice of a large river some- times becomes freighted with debris may be observed where high-grade tributaries occur. In such an instance, when spring approaches, the small streams may first become freed of ice and be able to sweep down debris upon the still frozen surface of the river. Again, when a river is bordered by steep bluffs, material loosened from the faces of the cliffs falls upon the ice and is ready for removal when freshets occur. Avalanches may also bring debris to a frozen river in the same way that they do to glaciers. The assistance in transportation rendered to streams by ice is, of course, greatest in high latitudes, but is not inconsider- able as far south as Virginia. Along the Potomac there are frequently boulders much too large for the unaided waters to move, and which it is presumed have been buoyed up by ice during a part at least of their journeys, since they are well beyond where the river loses velocity on passing from its high-grade upper course to the plain near the sea. In the method of transportation here considered it is not neces- sary that a river should freeze from side to side. The ice that forms about a partially submerged boulder near the shore of a stream tends to buoy it up. When the water surface is raised, if sufificient ice has formed about the stone, it will be floated away.* In the terraces of the Potomac ^ The weight of a cubic foot of water at 32° F. is 62.417 pounds ; a cubic foot of ice weighs 57.2 pounds. 24 RIVERS OF NORTH AMERICA about Washington, there are boulders two to three feet or more in diameter, resting on fine sand and clay, which it is thought were attached to ice-cakes at the time of their re- moval to their present sites. The locality referred to, it will be remembered, is south of the southern limit of former glaciers. The stone-charged ice carried each spring by the rivers in high latitudes acts much like a glacier in grinding the bot- tom and sides of the channel down which it moves. In the case of a large river this action is most pronounced near its borders, where the water is shallow, and probably does not occur at all in the deeper portions. On the border of Por- cupine River, Alaska, I have seen large areas exposed during low water where the bottom consisted of stones em- bedded in tenacious clay so as to form a veritable pavement. The ice charged with debris had previously moved over this pavement, and not only pressed down the stones so as to produce a generally even surface, but ground their exposed portions so as to make facets which were polished and striated. The pebbles and in some cases flat stones a foot or more in diameter, bearing these markings, have a remark- able resemblance to glaciated boulders. In other instances along the Yukon I found the solid rock, on prominent points, to a height of twenty feet or more above the summer level of the river, smoothed and striated by the action of river ice in much the same manner that is familiar in formerly glaciated valleys.* During spring floods in northern rivers, ice-blocks are fre- ^ I. C. Russell, *' Notes on the Surface Geology of Alaska," in Btilletin of the Geological Society of America, vol. i., pp. 116-122, 1890. Plate Views on the Yukon, Alaska. A.— Looking from the river across a portion of its delta. B. — River-bank of perennially frozen gravel. C. and D. — Stones left by floating ice during spring floods. E, — Bluff of hard rock on the border of a deeply cut valley. F. — Small cut-terraces in sand deposited during high-water. LAIVS GOVERNING THE STREAMS 2$ quently stranded and even forced far up the banks. Where flat cakes of ice accumulate in this manner, they sometimes have gravel and sand washed over them ; this material lodges in the cracks and openings between the cakes, and is left in curious heaps and ridges when the. ice is melted. Sometimes these deposits surround small areas so as to make shallow basins. There is yet another method by which ice assists streams in moving debris down their channels, and one which oper- ates in midstream, where the current is swift. I refer to the formation, during excessively cold weather, of what is termed anchor ice, or grouitd icCy at the bottom of streams, especially where the waters plunge over small obstructions and comparatively quiet bottom-eddies are produced. This bottom ice forms about stones, and by its buoyancy tends to lift them from the bottom, and thus to assist the currents in carrying them away. An instructive account of the formation of anchor ice in one of the rivers of New Brunswick, during the winter of 1869-70, has been recorded by W. G. Thompson,' one of the engineers of the Intercolonial Railway. In this ac- count, quoted below, it is stated that not only were small stones lifted from the bottom and floated down-stream, but, what is of still greater interest, the ice increased in thickness in some instances until it formed dams, and the stream was turned from its course. Mr. Thompson's ac- count is as follows: " The Matapediac, which is fed by large fresh-water springs, runs over a rocky bottom covered with loose stones, ranging in '^Nature, vol. i., p. 555, 1870. 26 RIVERS OF NORTH AMERICA size from coarse gravel to boulders as large as a hogshead, and the average current is about four miles an hour. " Early in November last the temperature went down in one night to 12° F., and on going out of camp the following morning I noticed large quantities of what appeared to be snow saturated with water floating down the stream, but not a particle of snow had fallen near us for many miles round, as far as I could see by the mountain-tops, nor had any ice formed on the surface of the river. ** The water opposite where I stood was about six feet deep, and perfectly clear, so that I could see every stone on the bottom, and, with the exception of the floating slush, the river was as it had been the previous day when the temperature was about 50° F. I got into a canoe and paddled with the current for half a mile or so, and in shooting some small rapids, where the water in places was not more than two or three feet deep, I noticed on the bottom masses of the slush clustered round and between the boulders, and a slight touch with the paddle was sufficient to free these clusters, when they rose to the surface, and were carried away by the current. I continued down the stream for three or four miles, and noticed the same thing in every rapid, where the water was shallow and ruflied by stones at the bottom. " The buoyancy of this slush was such that when detached from the bottom it rose so rapidly as to force itself well out of the water, and then floated off about half submerged. " I watched this forming of slush for many days, and in several cases found small stones embedded in the floating slush, having been torn from the bottom when the buoyancy of the slush, aided by the running water, caused it to rise. '* The temperature continued getting lower daily, and the slush in the rapids formed more rapidly than it was carried away, so much so that a bar or dam was formed across the river at each rapid, backing up the water in some cases five or six feet, when it generally found an outlet over the adjoining land, and into its natural bed again, or the head of water became sufficient to tear away the obstruction, which by this time had become a solid frozen mass. LAWS GOVERNING THE STREAMS 2J ** All this time, no properly crystallised ice had formed on the surface of the river, the current being too rapid, but the slush of ' anchor ice,' as the trappers call it, was forming in deeper water than it had formed in before, indeed all over the river bottom, and was rising and floating away as I have already de- scribed. Eventually the temperature got down to two and three degrees below zero, when the river surface began to freeze in the eddies and along the edges, and the open-water space became narrower every day, and was filled with floating ' anchor ice ' and detached masses of solid ice, which here and there became jammed and frozen together, so as to form ice-bridges on which we could cross. " These ice-bridges served as booms to stop much of the float- ing ice, which froze solid the moment it came to rest; and in this manner the river at last became completely frozen over for about forty miles of its length, but not until after we had experienced five weeks of steady cold, with the thermometer never above 12" F., and frequently down to — 16° F." When we recall the fact that the conditions of tem- perature described above recur every winter throughout nearly one-half of North America, it becomes evident that anchor ice must play an important part in stream trans- portation. But little attention has been directed to this matter, however, and it is highly desirable that someone favourably located for such studies should make a careful record, especially as to the number and size of the stones picked up from the bottom of stream channels, owing to the buoyancy of the ice formed about them. It will be noted that this process of transportation is brought into operation simply by a lowering of temperature, and does not require a rise of the water in order to float the debris attached to the ice, as is the case when surface ice becomes fastened to the bottom. Anchor ice operates in the way 28 RIVERS OF NORTH AMERICA described in midstream, where the water is not only swift but may be comparatively deep ; while the similar work of surface ice is practically confined to the shallow water on the margins of streams. Corrasion, — Clear streams, as we ordinarily see them, are such as have removed from their channels all of the particles within their reach which they are competent to transport, although they may still roll and push coarse fragments along their bottoms. It is to be noted, however, that clear streams, when of a given velocity, may become muddy if the velocity is increased. The friction of clear running water is but slight, hence streams not charged with material in suspension wear their channels very slowly. In such in- stances chemical solution of the rocks over which the waters flow may be in excess of mechanical abrasion. When a stream receives an initial load of silt and sand from rain-wash, the action of the wind, glacial abrasion, etc., an important change in its behaviour occurs. The transported fragments on being brought in contact with the bottom and sides of the channel of the stream produce abrasion. The flowing waters charged with silt and sand act not unlike a strip of sandpaper that is drawn over an object continually in one direction. The transported frag- ments abrade the rocks with which they are brought in contact, and are themselves worn and broken. Gravel and larger rock fragments too heavy to be carried in suspension, except during floods, are worn and broken by smaller frag- ments coming in contact with them, and when moved during high-water stages, assist in a marked way in promoting the process of channel enlargement. The friction and impact LAWS GOVERNING THE STREAMS 29 of the particles that are carried forward tend to loosen and dislodge other fragments, and thus increase the amount of material available for transportation. It is convenient to consider the process of stream abrasion due to the friction of transported material and to chemical solution, as a part of the general process of land degradation, and to give it a separate name. To meet this want, the term corrasion has been proposed.^ The deepening and widening of a stream channel — that is, corrasion — is carried on mainly by mechanical wear due to the friction of silt, sand, gravel, boulders, etc., carried through it by the flowing waters, but is assisted, many times in an important manner, by solution. There are conditions that limit or modify the competency of a stream to transport debris, as has already been briefly considered. In a similar way there are conditions which modify and limit the rate of stream corrasion. Streams, as is well known, vary in rate of flow, in volume, in declivity, in the degree to which they are loaded, in chemical com- position, etc. Changes in any one of these conditions will manifestly exert an influence on the rate at which a stream is enabled to deepen and widen its channel. The nature of the load carried by a stream also varies in different instances, and even from month to month and ^ As the nomenclature of dynamical geology and physical geography is not yet definitely fixed, it may be suggested that corrasion furnishes a convenient generic term, and may be made to include the processes of abrasion by stream- like movements of other substances than water, when charged with rock fragments. The grinding of rocks by glaciers may be designated as glacial corrasion ; the process of wearing of rocks by dust and sand transported by air- currents becomes aolian corrasion ; and lake and ocean shores furnish examples of wave and current corrasion. 30 J^IVERS OF NORTH AMERICA from day to day in the same stream. The particles or frag- ments carried are fine or coarse, hard or soft, rounded or angular; all of these conditions have an influence on the amount of friction exerted on the stream bed, and hence modify the rate of corrasion. Again, the rate at which a stream channel is enlarged under the supposition that the velocity of the stream, the character of its load, etc., remain constant, will vary with the nature of the rocks over which it flows. Hard rocks are worn more slowly than soft rocks, easily soluble rocks are more rapidly removed than those of difficult solubility. There are still other conditions pertain- ing to the beds of streams which influence corrasion. Mas- sive rocks yield less readily than those perhaps of equal hardness and equally soluble, but which are much jointed, or occur in thin layers. Rock texture thus exerts an import- ant influence on corrasion, as does also the inclination of the rocks or their dip. When hard and soft beds alternate, other conditions being the same, corrasion is more rapid when they are inclined than when they are horizontal. It is unnecessary to trace the effects of variations in these several conditions, as the student may do this for himself, and thus have the pleasure of making independent dis- coveries. In the case of a stream flowing under stated con- ditions, let the student postulate an increase in volume, in declivity, in character of load, in hardness of the rocks forming its channel, etc., other conditions remaining the same in each case, except so far as the reaction on them of the postulated change is concerned, and trace the effects on the rate of corrasion. The principal laws governing corrasion are, briefly: LAWS GOVERNING THE STREAMS 3 I 1. The rate at which a stream corrades its channel, other conditions remaining the same, increases with increase in load to a certain point which varies with the character of the load. If the load continues to increase, friction on the bottom decreases, and ceases when the entire energy of the stream is consumed in transportation. 2. Other conditions remaining the same, corrasion in- creases with declivity and with volume of water, since each of the changes increases velocity. For a more detailed discussion of the laws governing stream transportation and corrasion than it is practicable to present at this time, the reader is referred to Gilbert's * ad- mirable analysis of land sculpture, already cited many times in the present treatise. The conditions controlling the amount of detritus a stream can transport are mainly velocity and volume. Velocity is increased by an increase in volume and also by increased declivity. In nature we find that streams ordinarily vary in volume with seasonal changes and also from day to day, and hence their ability to transport, and consequently to corrade, undergoes many fluctuations. . The gradients of stream channels vary from place to place along their courses, and hence their ability to deepen their channels is not the same in all parts. The load that a stream carries in one portion of its course may be too great a burden in another portion, and some of it, always the coarser portion, will be dropped. The journeys of stream-borne debris are thus far from being continuous. The transported material ' G. K. Gilbert, Report on the Geology of the Henry Mountains, 4to, pp. gg- 150. Department of the Interior, U. S. Geographical and Geological Survey of the Rocky Mountain Region, J. W. Powell in charge. Washington, 1877. 32 RIVERS OF NORTH AMERICA is laid aside from time to time in bars and flood-plains. It may require tens of thousands of years for a given rock fragment loosened on a mountain-side to reach its final resting-place in the sea. In nature we find, as a rule, that the gradients of streams decrease from their sources to their mouths, but it must be remembered that this is the result of the action of the streams themselves and follows a long period of development and adjustment. As increased declivity favours corrasion, it is to be expected that the mountain tracts of streams will be deepened at a greater rate than their valley tracts, and that they will be enabled to extend their branches farther and farther, and thus acquire new territory. This, in fact, is the case, as we know, for the great majority of rivers are corrading their channels in the highlands and de- positing in the lowlands. Rivers are ordinarily supplied by many branches, however, which means that the volume of water in the branches is less than in the trunk stream, and accompanying decreased volume, other conditions remain- ing the same, is a decreased corrasion. The behaviour of a stream in reference to corrasion and deposition is thus a resultant of many and frequently opposing conditions. As will be shown later, the ability of a stream to corrade or de- posit in a given portion of its course varies ordinarily with its age, or, more accurately, with its stage of development. The portion of a stream channel where corrasion is in active progress during its youth, may become a region of deposi- tion at a more advanced stage in its history under the pro- cess of normal development which streams experience even if no changes occur in land elevation. LAWS GOVERNING THE STREAMS 33 Pot-Holcs. — One of the minor phases of stream corrasion is illustrated by the cylindrical holes frequently worn in the beds of streams by stones swept about by strong currents. These holes are sometimes saucer-shaped, but more fre- quently have steep sides and rounded bottoms, and resemble the insides of the familiar cast-iron kettles used for culinary purposes ; this similarity has suggested the name pot-hole, by which they are commonly designated. Their walls are usually smooth, and sometimes exhibit grooves and ridges Fig. I. A Pot-Hole being Scoured out by Stream Action. (After R. S. Tarr.) in horizontal planes or arranged more or less spirally. In these grooved holes one sometimes finds well-worn pebbles, - or even large boulders, and discovers a relation between the size of the grooves and of the stones that made them. In many instances, also, these mills are still in working order, and a stream of water is plunging into them, as is shown in the accompanying photograph. 3 34 RIVERS OF NORTH AMERICA Pot-holes are of all sizes, from shallow depressions a few- inches in depth to vertical borings five or six feet across and fifteen or twenty feet or more deep. From these well de- fined examples there is a gradation up to the basins pro- duced by the stones swirled about at the bases of waterfalls, as the pool into which Niagara plunges, for example, which might perhaps be termed compound pot-holes. In the making of these characteristic depressions the mill- stones may be worn out, but new ones are supplied from time to time, and the process goes on. Similar excavations are made also beneath glaciers, where streams flowing on the surface of the ice plunge into crevasses, or deep well- like openings termed moulinSy and reach the rocks below. In fact, any strong current by being deflected may cause loose stones to be whirled about so as to grind the rocks on which they rest, and produce depressions of the nature here considered. Favourable conditions result when pebbles and boulders of hard material occur in a stream where the bottom is of soft rock, and where also the current is swift, and eddies, swirls, whirlpools, etc., are produced. Lateral Corrasion, — Streams in moving material along their channels not only wear away the rocks over which they flow, but abrade the sides of their channels as well. It is difficult to formulate the laws governing lateral cor- rasion, but the general manner in which it is accomplished may be readily understood. In considering the process of vertical corrasion the influ- ence of upward currents in flowing water was recognised. But besides the upward currents there are also lateral cur- rents. In fact, the flow of water, even through smooth, LAWS GOVERNING THE STREAMS 35 ft straight troughs, is complex, and many secondary currents are generated. This complexity is vastly increased when the channel is rough and irregular. If we watch a swift- flowing stream, we shall be enabled to see that there are many swirls and eddies due to, or accompanied by, currents moving in all directions. Along the sides of a stream channel the secondary currents strike the rocks and dash against them whatever material the waters may hold in sus- pension. Friction results from the impact of the floating particles, and tends to wear away the sides of the channel. The larger fragments rolled along the bottom of the stream can take but little part, directly, in this process of channel- widening. It is the finer material — the silt and sand held in suspension — which does most of the work. Moreover, the courses of stream channels are seldom, if ever, straight for any considerable distance, but are a succession of con- cave and convex curves. The water is alternately thrown against one bank, and the direction of the current being de- flected, impinges on the opposite bank lower down-stream. At the locality where the thread of swiftest flow nears the bank, the rocks are worn away, and the irregularity of the stream's course increased. The material removed is carried down-stream, and in part deposited in the slack water on the concave side of the next bend. Swift streams are not so easily turned aside as those which flow less impetuously. Hence the former maintain straighter courses than the latter. Lateral corrasion may go on when vertical corrasion is checked by decrease in declivity, deposition, or for other reasons. After a stream has cut down its channel at its 36 RIVERS OF NORTH AMERICA mouth to the level of the still water into which it discharges and established a low gradient, lateral corrasion may con- tinue. It is under these conditions that the widening of river valleys is mainly accomplished. In general, vertical is so far in excess of lateral corrasion in the case of high-grade streams, that the valleys produced are narrow, and V-shaped in cross-section, while under similar conditions in respect to climate, rock texture, etc., in a low-grade stream, although its actual rate of lateral corrasion may be less than in the first instance cited, the ratio of lateral to vertical corrasion is greater, and a flat-bottomed valley results. The cross-pro- file of a valley widened by lateral corrasion is U-shaped, and if the process is long continued becomes broad-bottomed. The statement frequently made that a stream-cut valley is V-shaped in cross-profile, in distinction from the U-shape of valleys formerly occupied by glaciers, is not strictly true, as it considers only young stream-valleys. Meanderifig Streams, — The serpentine courses followed especially by sluggish streams, just referred to, is a matter of more than passing interest to the student of geography. Much of the charmingly picturesque which enters into and many times forms the leading feature of stream-side scenery, as well as the secret of the process by w^hich valleys are broadened, and the adjacent uplands removed, results from the meandering and lateral migration of streams. By this same general process, too, as will be considered later, the flood-plains of rivers are spread out. Illustrations of the curves characteristic of many streams, are given on Plate 3. The causes leading to the meandering of streams have been studied by various observers. As stated by Fergus- LAPVS GOVERNING THE STREAMS 37 son/ a river is a body of water in unstable equilibrium, whose normal condition is that of motion down an inclined plane, and if all inequalities in the material forming the bottom and sides of its channel could be removed, it would flow continuously in a straight line. (It is to be noted, however, that the influence of the earth^s rotation is not considered in this discussion.) Any obstruction or inequal- ity, however, necessarily induces an oscillation, and, the action being continuous, the effects are cumulative, and the oscillation goes on increasing till it reaches a mean be- tween the force of gravity tending to direct the current in a straight line, and the force due to the obstruction tending to give a direction more or less at right angles to the former. In nature not one but many disturbing conditions occur, and the streams flow in a series of curves, each of which bears a definite relation to their volumes and the gradients of their channels. The stage in the lives of rivers when they meander in broad curves through rich bottom-lands, is usually, or most commonly, reached late in their lives, when the task of re- ducing their channels to the level of the still water into which they discharge has been nearly completed. They then flow sluggishly, and may be said to be enjoying the rest to which a long life of activity entitles them. A slack current and a tortuous course are not infallible indications of old age, however. Young streams flowing across an abandoned lake bed, for example, or over lands recently raised from the sea, may have these characteristics. A 'James Fergusson, *' On Recent Changes in the Delta of the Ganges," in The Quarterly Journal of the Geological Society of Londo7i, vol. xix., p. 323, 1S63. 38 RIVERS OF NORTH AMERICA winding course may be retained by a river which has been given renewed energy by a re-elevation of the land, even after it has cut a deep trench and is a hurrying torrent. The tendency to meander is strongest in streams that are heavily loaded and are depositing a portion of their burdens in flood-plains. Although the tendency to meander characterises all streams, for the reason that their channels are never straight or composed of homogeneous material for any con- siderable distance, yet the process carries with it certain limitations, as will be shown in discussing the origin and nature of flood-plains and terraces in a subsequent chapter. Other Curves, — Not only do streams bend to the right and left of their general courses, as where a river meanders through a broad, partially alluvial-filled valley, but, as will be described later, form curves in a vertical plane as well. Where corrasion is in progress, the longitudinal profile of the channel produced is concave to the sky, and where de- position occurs, curves convex upward result. More or less complete spiral curves about vertical or inclined axes, like the twists of a corkscrew, may be seen when a high-grade stream is excavating soft clays, and where a brook on the surface of a glacier plunges into a well-like opening in the ice. The influence of the graceful sweep of stream-curves, on the beauty of many landscapes, is due to their infinite variety; no two in the course of even a great river being identical. This marvellous diversity, produced by simple means, becomes still more impressive when it is remembered that no one of these many curves remains the same for any considerable period of time. Plate III. ^N^.^'.\. uJJCkL ' . ..'i^^'H^.k Fig. a. Ray Brook, Adirondacks, New York. (Photograph by S. R. Stoddard.) Fig. B. Moccasin Bend, Tennessee River, from Lookout Mountain ; Chattanooga at the Right. LAM^S GOVERNING THE STREAMS 39 The several classes of curves in the channels of streams are supplemented in an interesting manner by other curves in the surfaces of the streams themselves. Not only are graceful curves produced by the flow of water in eddies and swirls, or when they arch over or circle about obstacles, and are thrown into waves by the wind and other causes, but the surface of a stream in cross-section is not a straight line, although this condition is very nearly reached when the current is gentle, and the waters deep. If there is a strong central current, however, the surface there forms a convex curve, which rises well above the more gently flowing waters on either side. In swift rivers this difference in level frequently amounts to five or six feet, and in certain instances, as at the whirlpool below Niagara, is reported to be two or three times these measures. In such examples the surface line of a cross-profile would show a pronounced upward curve in the central part, bordered on each side by a downward curve. The bordering downward curves are gentle, but may be recognised, although they probably de- part but little from a straight line. Driftwood carried by a stream with a swift central current, as will be described more fully in advance, tends to leave the elevated central part and collect along the banks. This tendency may be seen especially when swift streams are rising; when the waters fall, however, driftwood leaves the slack water adjacent to the shore, and tends to concentrate in mid-stream. At such times the stream in cross-profile probably presents a concave surface-line. Deflection of Streams Owing to the Rotation of the Earth, — The earth, as is well known, makes one rotation from west 40 RIVERS OF NORTH AMERICA to east about an axis passing through the poles, in 23 hours 56 minutes and 4 seconds. The circumference of the earth being about 24,000 miles, any point on the equator must, therefore, travel over 1000 miles an hour. North and south of the equator this motion gradually decreases, and becomes zero at the poles. This motion has an influence on the flow of streams, and tends to cause them to follow curved instead of straight courses. This may be most readily understood in the case of streams having either north or south direc- tions, but affects all streams on the earth's surface, unless they follow strictly the path marked out by the equator. Water flowing northward from the equator would start with the motion from west to east, which pertains to that location, but as it advanced would cross regions which have progressively less and less motion from west ta east. The current due to gravity, we will assume, tends due north, but the waters have also a motion from west to east, due to the earth's rotation, which is in excess of the similar motion of the region necessarily invaded. The re- sultant of these two forces will carry the stream to the east of the meridian on which it started, and the stream will curve to the right of its initial course. In a similar way, a stream in the northern hemisphere, flowing toward the equator, would be continually invading territory having greater and greater motion from west to east and would curve to the west of the path it would follow if influenced only by gravity. Thus, in the northern hemisphere, the tendency of the earth's rotation is to cause the streams, no matter what their direction of flow, to corrade their right more than their left banks. In the southern hemisphere the direction LAWS GOVERNING THE STREAMS 4 1 of curvature due to the earth's rotation is reversed, and the streams, no matter what their direction, tend to corrade their left more than their right banks. This is an applica- tion to streams of Ferrel's law, namely: '' If a body moves in any direction on the eartlis surface, tJiere is a deflecting force arising front the earth' s rotation whicli deflects it to the right in the nortliern hemisphere, but to the left in the southern hemisphere,'' * The tendency of a stream to maintain a straight course as it invades territory having a progressively changing rate of motion is greater the greater the velocity of the stream ; the same is true of the parts of a stream. The thread of maxi- mum current in a stream following an approximately straight course, is in the centre near the surface. The change in direction owing to rotation is, therefore, less quickly mani- fest in the thread of maximum current than in the more sluggish waters on either side, and the current undergoes greater deflection. Where streams follow winding courses this tendency leads to an increase in their meanderings to the right, in the northern hemisphere, more than to the left of their general direction. There is thus a tendency, due to the earth's rotation, for them to excavate their right more than their left banks, and to migrate to the right of their initial courses. This tendency is slight, but all the time operative. Owing, however, to inequalities in hardness of the banks of streams, and other disturbing conditions, it is difficult to discover examples where the earth's rotation has ^ William Ferrel, " The Motions of Fluids and Solids Relative to the Earth's Surface," in The Matheviatical Monthly, vol. i., p. 307, 1859. ^ he influence of the earth's rotation on air-currents is clearly explained in W. M. Davis's Elementary Meteorology^ pp. loi-iii. Ginn & Co., Boston, 1894. 42 RIVERS OF NORTH AMERICA plainly controlled their migrations. An illustration of such a result is thought to be furnished, however, by the streams on the south side of Long Island, where there is a plain with a remarkably even descent and gentle slope. This plain is crossed by a number of small streams which have excavated shallow valleys in essentially homogeneous gravel. Each of these little valleys is bordered on the west, or right side, by a bluff from ten to twenty feet high, while its gentle slope on the left side merges imperceptibly with the general plain. The stream in each case follows closely the bluff at the right. As stated by Elias Lewis, and affirmed by Gilbert,^ there seems to be no room for reasonable doubt that these peculiar features result from the influence of terrestrial rotation. It is to be remembered that the force of rotation, like gravity, is all the time operative, but its influence is greatest on streams flowing north or south, is greater in high than in low latitudes, and increases with the rapidity with which the waters are transferred from an area having a certain motion to another area having a different motion. The results of this influence, although not conspicuous, are nev- ertheless important. There is a slight tendency through- out the length of every stream in America and at all times, to erode its right more rapidly than its left bank. In the case of the Mississippi, shown by Gilbert in the article ' G. K. Gilbert, " The Sufficiency of Terrestrial Rotation for the Deflection of Streams," in American Journal of Science, vol. xxvii., pp. 427-432, 3d series, 1884, An abstract of this paper, accompanied by an extension of the discussion, may be found in Science, vol. iv., pp. 28, 29, 1884. See, also, \V. M. Davis, " An Early Statement of the Deflective Effect of the Earth's Rota- tion," in Science, vol. i., p. 98, 1S83. ZAPVS GOVERNING THE STREAMS 43 cited above, the selective tendency thus determined toward the right bank is nearly nine per cent, greater than toward the left bank. Questioning the Rivers, — To illustrate the laws govern- ing the behaviour of streams, let us see how some of the leading features of the rivers of North America can be accounted for. Why, for example, are the waters of the St. Lawrence clear and those of the Missouri usually muddy ? The former is obviously a clear stream for the reason that the Great Lakes it drains act as settling basins and retain the sediment brought down by countless tributaries. Many small streams join the St. Lawrence below Lake Ontario, but most of these also have lakes in their courses, and the amount of the sediment reaching the main river from rills and brooks is not sufficient to materially change its charac- ter. The' clear waters of the St. Lawrence have but little power to corrade. The current is swift in places, but all of the fragments in its bed which the current is competent to move have long since been carried away. Corrasion has gone on with extreme slowness throughout the present geographical cycle, and the river has not yet entrenched itself, but is practically a surface stream. The reader will, no doubt, at once remark that the Mis- souri and the Platte are also surface streams, although heav- ily loaded with silt and sand. These rivers, however, rise in high mountains and flow across a broad plateau. Their many branches in the mountains are swift and bear along heavy loads of detritus. On leaving the mountains and entering their plain tracts, velocity is checked, and the less 44 RIVERS OF NORTH AMERICA swift waters are no longer able to carry the loads they pre- viously transported with ease, and deposition occurs. Then, too, the surfaces of the Great Plateaus are composed of easily eroded rocks. During every rain quantities of soil, etc., are washed into the rivers, and during the long, dry summers, the winds are busy in performing a similar task. At present not only is corrasion nil throughout the plain tracts of the Missouri and the Platte, but sedimentation is in progress. These rivers are aggrading previously eroded valleys. The question of how deep and how wide a river valley shall be, depends not alone on the elevation of the land above sea-level, but also on the ratio of stream corrasion to the general waste or erosion of the bordering lands. When the rate of stream corrasion is in marked excess of the rate at which the general surface of the land is being eroded, deep, narrow stream channels result. But if the erosion progresses at nearly the same rate as corrasion, the relief of the region will be mild. Between these two ex- tremes there are many intermediate stages. Overloaded streams, by dropping a portion of their burdens, may not only spread out broad flood-plains, but also elevate their channels so as to flow at a higher level than the sur- rounding land. The Platte and the Missouri are now de- positing material throughout their courses across the Great Plateaus, and, as just stated, are aggrading previously formed broad-bottomed valleys. Many conditions besides those just noticed exert an influ- ence on the character and expression of stream-cut valleys. When the rocks are hard, they tend to form precipitous LAWS GOVERNING THE STREAMS 45 bluffs; when soft, they crumble, and the valleys have flaring sides. When the climate is arid, the wasting away of cliffs is long delayed, and if corrasion is actively progressing, steep-sided gorges, or canyons, result. Vegetation retards surface erosion, although favouring rock decay, and has a varied influence on the lives and character of the streams. These and still other influences modify the results of stream corrasion, which find expression in the scenery of the land. It may be asked, why is it that the Colorado has carved the most magnificent canyon in North America, while the Platte, rising in the same mountains, is bordered throughout much of its course by perhaps the least picturesque scenery of any large river on the continent ? Much of the answer has already been given. The Platte, as we have seen, is aggrad- ing its channel to the east of the Rocky Mountains. The Colorado is still corrading from its source, with the excep- tion of certain slack-water reaches in soft rocks, all the way to its place of discharge. The rocks through which the Platte flows are soft ; while those cut by the Colorado are hard. The region of the High Plateaus crossed by the Colorado has experienced a somewhat recent uplift of several thousand feet ; while the country traversed by the Platte to the east of the Rocky Mountains has, so far as is known, undergone but slight changes in elevation during the same period. The climate of the Colorado region is arid, and surface waste from rains and the action of the wind probably less than in the region of the Platte. It is in these and per- haps still other contrasts of conditions that the striking differences in the scenery along the border of these two rivers may be accounted for. 46 RIVERS OF NORTH AMERICA By comparing the scenery of various other regions and seeking for the underlying causes to which their differences are due, the student may arrive at a juster appreciation of the characteristics of river work than a formal statement of the subject will furnish. Erosion, — Weathering, transportation, and corrasion ^re three agencies which by their combined action lead to the removal of land upraised above the sea, and the production of a vast array of ever-changing topographic forms. To this far-reaching and highly complex process, the name erosion has been given. Briefly stated, the removal of material from land areas is accomplished by: ist. The disintegration of the rocks by both mechanical and chemical means, through the action of the varied and complex process termed weathering, 2d. The removal especially of the finer products produced by weathering, by the wind, general rain-wash, and rills, and of both fine and coarse debris by brooks and rivers, by a pro- cess termed trajisportation. 3d. The friction and impact of the* material transported, accompanied by solution, lead to the deepening and broadening of stream channels, or c or r a si 071, A necessary accompaniment of erosion is the deposition of the material removed, or sedimentation. A temporary phase of sedimentation is the laying aside of stream-carried detritus in flood-plains and stream channels, but its final resting-place, so far as a single geographical cycle is in- volved, is on the floor of the sea. Baselevel of Erosion. — The depth to which a stream, flow- ing into a lake or the sea, can lower its channel by mechani- LAWS GOVERNING THE STREAMS 4/ cal corrasion is limited by the surface level of the receiving water-body. As mechanical corrasion decreases in more than a simple ratio with decrease in declivity, the final stages in the lowering of a stream channel to the level of the still water into which it flows must be extremely slow. Corrasion is not limited to mechanical processes, however, but includes solution as well. The final reduction of a stream channel to sea-level must, therefore, be by solution. Every stream, when its entire history is reviewed, will be found to be engaged in deepening its channel to the horizon referred to, or else has accomplished a part or the whole of the task. The datum-plane limiting downward corrasion is reached first at the mouth of a stream and is then continued progressively towards its source. When many streams are considered, various stages in this process may be recognised. The depths to which streams may excavate their channels is evidently a matter of great importance in their develop- ment and in the history of the topographic changes of the land. This fundamental principle was first clearly defined by Powell,^ who employed the term baselevel to designate the lower limit of stream action. It is the baselevel of cor- rasion. A lake determines the baselevel for the streams flowing into it, but as lakes are short-lived, the real base- level toward which all streams are working is the surface level of the sea. To use Powell's own words in this connection: " We may consider the level of the sea to be a grand base- level, below which the dry land cannot be eroded; but we may ^ J. W. Powell, Exploration of the Colorado River of the West and its Tributaries, p. 203, 4to. Washington, D. C, 1875. 48 RIVERS OF NORTH AMERICA also have, for local and temporary purposes, other baselevels of erosion, which are the levels of the beds of the principal streams which carry away the products of erosion. I take some liberty in using the term level in this connection, as the action of a run- ning stream in wearing its channel ceases, for all practical pur- poses, before its bed has quite reached the level of the lower end of the stream. What I have called the baselevel would, in fact, be an imaginary surface, inclining slightly in all its parts toward the lower end of the principal stream draining the area through which the level is supposed to extend, or having the inclina- tion of its path raised in direction as determined by tributary streams." When a stream has lowered its channel nearly to base- level, downward corrasion is retarded, but lateral corrasion continues. Low-grade streams, as we have seen, are the ones most inclined to meander, and to broaden their valleys. If this process is continued for a sufficient time in any region, it will lead to the removal of all land within reach of the streams, down to their own level. Baselevel of corrasion thus becomes practically the baselevel of erosion. The ulti- mate result of erosion is to reduce a land area to a plain at sea-level. Such perfect plains, however, are exceedingly rare, but approximations to the ultimate result are common, and plains in this penultimate stage have been named pejie- plains by Davis. A peneplain is the normal result of the erosion of the land, provided elevation or depression do not occur to check the process. If a portion of the earth crust remains essentially stable for a sufficient length of time to allow the streams flowing from it to broaden their channels, a more or less extensive peneplain is produced. Such periods of % LAWS GOVERNING THE STREAMS 49 stability, when movements in the earth's crust are not suffi- cient to check the normal process of baselevelling, are termed geograpliical cycles, A peneplain may be defined as an ap- proximately perfect plain produced by erosion during a geographical cycle. If, after a portion of a land area has been reduced to a peneplain, elevation occurs, a new geo- graphical cycle will be initiated, and the process of base- levelling again begun. During one geographical cycle all of the land may not be reduced to a plain, but isolated uplands between broad valleys remain. These remnants are left as an inheritance to the next succeeding geographical cycle. Mount Monad- nock, in southern New Hampshire, is an example of such a remnant, and forms a prominent feature on the surface of an elevated peneplain. The study of the relief of the land in various regions has shown that there are many such rem- nants, left by incomplete planation. A name is needed for such topographic features, and to meet this, want Davis has proposed that they be termed monadnocks, after the typical example just referred to. A monadnock, then, is a hill or mountain left standing on a peneplain, owing to incomplete planation. The laws, just stated, governing the reduction of land areas to baselevel, although wide-reaching and fundamental to the student of geography, are not strictly true, or, rather, exceptions to them may be found. The downward limit of mechanical corrasion is not in all cases the sea-level. Gla- ciers may enter the sea and continue their destructive work at a depth of a few hundred feet below its surface. Ice- bergs may also disturb the bottom of the sea at considerable 50 RIVERS OF NORTH AMERICA depths. Currents in the sea sometimes corrade the bot- tom ; the downward limit to which this process may be carried is unknown, but may be hundreds of fathoms. The removal of rocks in solution may be carried on deep below sea-level; the lower limit has not been determined, but is certainly many thousands of feet. Change in the position of rock material through the agency of plants and animals is not limited downward, by the surface level of the sea, but goes on below that horizon. All of these processes, how- ever, are of minor importance, and need not be considered as sensibly modifying the conclusion that the downward limit to which land areas may be reduced is the horizon of the surface of the sea. Strictly speaking, baselevel is the lower limit of the mechanical corrasion of streams, but prac- tically, as we say, it is also the downward limit of erosion. Influence of Vegetation on Erosion, — The influence of vegetation on the general process of denudation is varied, and both retards and accelerates the process. Vegetation breaks the force with which rain-drops strike the earth, and besides, when the ground is covered with leaves, or with grasses, moss, or other plants of low growth, the surface w^aters are filtered of such debris as they may have taken in suspension. Vegetation thus retards transportation and decreases mechanical corrasion. On the other hand, vegeta- tion furnishes the percolating water with organic acids, principally humus acids, which greatly enhance their solvent power. Hence vegetation favours chemical corrasion in a high degree. The roots of plants bind the soil together, and thus assist it in resisting mechanical agencies tending to remove it. LAWS GOVERNING THE STREAMS 5 1 But roots furnish organic acids as they decay, and besides open passageways for descending water, thus facilitating chemical changes. The student may readily observe other modifications in the lives of streams due to vegetation and to climate. Interesting results would no doubt be obtained from the study of the vegetation which grows in the streams themselves, such as the algae and certain higher forms of plant life. The direct influence of driftwood and of fallen timber is considered in a subsequent chapter. Although the summary of the laws governing the be- haviour and work of streams just presented is confessedly incomplete, yet it is the wTiter's hope that it will serve to interest the reader in the processes of land sculpture nearly everywhere in progress where the earth's surface rises above the sea, and suggest questions to which more technical treatises, or, better still, the rills and rivers themselves will furnish answers. Note. — Since this chapter was written, a highly instructive paper by Hunt- ington Hooker has appeared, on " The Suspension of Solids in Flowing Water," Transactions of ihe Americait Society of Civil Engineers, vol. xxxvi., 1897, pp. 239-340, which the reader is recommended to study. CHAPTER III INFLUENCE OF INEQUALITIES IN THE HARD- NESS OF ROCKS ON RIVER-SIDE SCENERY IF we watch a hillside rill which is born during a heavy shower and runs dry when the sun again shines, it will be noted that in the steeper portion of its descent it cuts a narrow trench and deposits much of the material removed farther down its course. It soon becomes apparent that the little stream is corrading where its descent is steep, and raising its bed by depositing the material brought from above where the grade becomes gentle. The processes of corrasion, transportation, and deposition may all be seen in operation in a stream only a few rods in length. The topo- graphic forms resulting are, on a minute scale, the same as those which give grandeur to many far-reaching views of river-side scenery. The process of excavation and deposition carried on by a rill leads to the making of a uniform grade down which the waters continue to carry debris. This gradient, as will be shown later in discussing the profiles of streams, is not an inclined plane, but in longitudinal profile is a curve, steepest near the source of the stream, and flattening out and approaching nearer and nearer a straight line the nearer 52 INEQUALITIES IN THE HARDNESS OF ROCKS 53 the mouth of the rill is approached. The immediate task of the rill may be said to be the making of a certain gradient, which is best suited to its volume and other conditions. In the portion of the rill channel where corrasion is in progress, the waters are broken by little cascades and mini- ature rapids, and it is evident that a uniform gradient in that portion of its bed is far from perfect. The reason is that the clay, earth, or other material in which the rill is working is of varying degrees of hardness. When a boulder is crossed, a cascade results. When the material is soft the channel is quickly deepened. Each passage from a hard to a soft layer is marked by a cascade. Evidently the charac- ter of the material which the rill is removing exerts a con- trolling influence on the miniature scenery along all of its upper course. If we enlarge our field of observation, similar characteris- tics will be found in many brooks, creeks, and rivers. The study of many streams will soon show, however, that some of them are broken by cascades and rapids, while others, except on their extreme headwaters, have even descents and flow with a generally uniform current. Those which are broken by cascades, we soon learn, for the most part oc- cupy narrow, steep-sided trenches, and are apt to have lakes along their courses ; while those with a uniform descent from near their sources to their mouths, are not associated with lakes, except perhaps such as are formed by the streams themselves in alluvial-filled valleys. A comparison of many streams will show that the differ- ences just referred to, depend on their age, or, more accur- ately, on the stage of development they have reached. 54 RIVERS OF NORTH AMERICA Youne streams, or such as have not cut down their channels so as to produce a uniform gradient, are the ones suppHed in part by lakes, and are broken by places of rapid descent ; while older streams have removed the inequalities from their channels, and the lakes that may formerly have existed along their courses have been drained. It thus becomes evident that the degree to which variations in the hardness of the rocks influence the scenery of streams is greatest in youth and gradually beomes less and less, but seldom, if ever, entirely vanishes. All stages in development, from extreme youth to slug- gish old age, may be recognised in the streams of North America, and among the most marked characteristics of this slow change is the presence of cataracts and rapids in the courses of the streams which have not yet made a decided advance in their appointed tasks. Waterfalls, — Young streams are obliged to accept in- herited conditions of slope, and may discover that their courses are broken by places of deep descent, and rapids and cascades result. If, for example, a stream originates on a tableland, with an irregular surface, or is perhaps bounded by an escarpment, it will have an uneven channel, at least for a time, and be broken by places of steep descent. Again, streams coming into existence on the withdrawal of an ice-sheet, or the draining of a lake, will usually find in- equalities in their channels which will cause waterfalls. In such instances the conditions producing the falls are an in- heritance from pre-existing topographic conditions. As stream development progresses, however, the 'cascades re- sulting from inherited conditions disappear; but others due INEQUALITIES IN THE HARDNESS OF ROCKS 55 to inequalities in rock texture, to the more rapid rates at which a trunk stream may deepen its channel than its branches, to the loads sometimes deposited in sluggish streams by swifter tributaries, and still other causes, make their appearance. In illustration of the processes by which cascades origi- nate through the ac- tion of the streams themselves, it is evi- dent that when a stream flows across alternating hard and soft beds, the soft beds will be removed more easily than those of greater resist- ance, and when the streams leave a hard bed, a rapid, cascade, or waterfall may be Fig. 2. A Young Valley being Cut in Shale produced. The refer- ^^ck, Central New York. (After R. S. Tarr.) ences just made to a trunk stream cutting more rapidly than its tributaries, and the deposition of debris in a sluggish stream at the mouth of a high-grade tributary, need no special explanation. All of the conditions just referred to, including inherited 56 RIVERS OF NORTH AMERICA inequalities of channel and the production of places of steep descent during stream development, pertain prnicipally to young streams. As a stream advances in its task of cutting down to baselevel and acquires the gradient best adapted to its work, inequalities in the slope of its channel disap- pear. Cataracts, of whatever character, are usually an index of immature stream development. The development of a stream progresses, however, from its mouth towards its source, and the feeding brooks of even well-developed river systems are young, and may have cascades, while the lower portions of the same drainage system may have reached perfect adjustment. The streams draining the southern Appalachians have been allowed to progress with the execu- tion of their tasks without serious interruption for a long period of time, all lakes and waterfalls resulting from in- herited conditions, and practically all cascades produced by irregularities in rock texture, have long since disappeared throughout their lower courses, but their head branches are still young and are yet being extended. On these young twigs of the drainage system cascades are common. An illustration of a cascade of this nature is presented in Plate IV. An exception to the rule that cascades are not developed in the courses of mature streams, is sometimes found in lime- stone regions where surface drainage enters underground channels. The breaking of the roof of a cavern may lead to the production of a cascade at any stage in the life of a stream, but the chances of such an accident, as it may be termed, become less and less as baselevel conditions are ap- proached. INEQUALITIES IN THE HARDNESS OF ROCKS 57 In addition to the causes just considered, there are changes produced by movements in the earth's crust, as when the rocks are folded or faulted ; the birth and growth of glaciers; volcanic eruptions; the deposition of rock ma- terial from solution, as when a spring precipitates traver- tine, siliceous sinter, etc., in the course of a stream; the work of animals, as when beavers build their dams ; the stranding of driftwood so as to block drainage, etc., which may interrupt the even flow of water, and give origin to rapids, cascades, and waterfalls. The tens of thousands of waterfalls in North America may be arranged principally in two classes: those resulting from the excavation of alternating hard and soft layers, and occurring mostly on the head branches of well-developed streams, as in the southern Appalachians; and those due to previous glacial conditions. Of these two classes the second is by far the more numerous, and furnishes the most magnifi- cent examples of waterfall of all grades. As already stated, the northern half of North America was formerly covered by ice-sheets. Local centres of snow accumulation and of glaciers also existed on the Rocky and Cascade Mountains, and the Sierra Nevada, far to the south of the southern limit of the former continental glaciers. Throughout nearly all of the vast regions which were form- erly ice-covered, the melting of the glaciers left the- land encumbered with debris. The surface inherited by the post- glacial streams was essentially a new-land area, and the streams have not progressed far enough in their develop- ment to have removed the inequalities in their channels, and waterfalls are common. 58 RIVERS OF NORTH AMERICA Illustrations of this class of cascades are furnished by those in the picturesque streams of the Catskills, Trenton Falls, the many beautiful cataracts near Ithaca, the well- known instances in Watkins Glen, and numberless others of the same general character in New York. Still more numerous instances might be cited in Canada, as, for example, the leap made by the water at the Falls of Mont- morenci, near Quebec, the Great Falls in Labrador. In fact, scarcely a stream can be ascended in all of the eastern and northern portion of the formerly glacier-covered region, without discovering that it has recently been turned from its former channel, or exhibits the characteristics of youth from mouth to source. On the headwater of the Mississippi, and about the upper Great Lakes, the drift sheet is thicker than farther eastward, and the streams have less frequently cut down to the hard rocks beneath, so as to develop cascades. It is only the stronger rivers in this more thoroughly drift-covered country that have progressed far enough with their recently added task, to lay bare the solid rock beneath the superficial covering. A far-reaching result of the disturbance produced in stream development by the glacial epoch is seen in the dis- tribution of water power produced by it. In New England, manufacturing industries were soon established after the coming of Europeans, and a decided impression made on the character of the people by this circumstance. South of the glacial boundary, water power was far less abundant, and mostly within the more inaccessible portions of the mountains; the development of manufacturing industries INEQUALITIES IN THE HARDNESS OF ROCKS 59 was hence delayed, and attention given more largely to agriculture, for which climatic and other conditions were more favourable. When steam was introduced as a motive power, the waterfalls at the north declined in importance as sources of energy, but in this budding age of electricity they are again coming into demand. In all of the various phases of stream development and of the diversity in the relief of the land produced by stream erosion, thus far considered, a marked feature has been that changes are continually in progress. To this rule, the water- falls furnish no exception. They have their periods of growth and decline, and in many instances shift their positions, or migrate. In the case of waterfalls resulting from inherited topo- graphic conditions, they may spring into existence all at once, and at the very start be grander than ever after. Niagara, when it first leaped from the summit of the escarp- ment near the present site of Lewiston, was higher than at any subsequent period of its history.' When falls result from the wearing away of soft rocks so as to make adjacent hard beds prominent, there is usually at first a small difference in relief, producing a rapid, and then greater changes resulting in a cascade the height of which is increased by reason of the increased energy of the waters as they plunge into the pool below, but there comes a time when the stream channel below the fall can be low- ered no farther. The cascade, or waterfall, as we may choose to call it, then reaches its greatest development, or ^ G. K. Gilbert, *' Niagara Falls and their History," in National Geographic Monographs, vol. i., pp. 203-236. American Book Co., 1895. 6o RIVERS OF NORTH AMERICA its majority. But the lowering of the stream channel above the fall continues, and its height gradually decreases. Such a sequence in the life of a waterfall originating from unequal stream corrasion in soft and hard rocks, w^ould result when the fall did not migrate up stream but remained in one place. But most waterfalls are subject to a process of migration. TJie Migration of Waterfalls, — When a hard layer causing a waterfall is horizontal, or but slightly inclined, the escarp- ment over which the waters plunge recedes, owing, in most instances, to the removal of softer rocks beneath, by the friction of stones washed about by the swirling waters, and the fall migrates up stream, leaving a more or less canyon- like valley to mark the path along which it travelled.' Thus, below Ni- agara Falls there is a canyon about seven miles long, and approximately two hundred feet deep, which has been left by the migration of the cataract. If the hard bed cut through by a stream so as to produce a cascade has a sharp downward slope in the direc- tion opposite to the flow of the stream, the fall will become lower and lower as corrasion progresses, and when the hard layer is passed, the life of the cataract will come to an end. If, however, as may happen, the hard layer dips downstream, the rapid or fall produced will manifestly increase in magni- FiG. 3. Profile and Sec- tion at Middle of Horse- shoe Fall, Niagara, Showing Hard Lime- stone above Soft Shale, and Probable Depth of the Pool into which the Waters Plunge. Scale : I inch = 384 feet. (After G. K. Gilbert.) Plate IV. Fig. a. Fall on Black Creek near Gadsden, iVlabama. A young branch of Coosa River, at the south end of Lookout Mountain ; hard sandstone above shale. Fig. B. Echo River in Mammoth Cave, Kentucky. (Copyrighted photograph by H. C. Ganter.) INEQUALITIES IN THE HARDNESS OF ROCKS 6 1 tude until the locality where the outcropping edge of the hard bed comes to the surface is cut through, and then a comparatively sudden lowering and adjustment of grade will follow. When the hard layer which a stream has to cut through is horizontal instead of being inclined either with or against the current, it presents a greater task to a stream cutting through it than in any other position, because the mass of rocks necessary for the stream to remove in order to reduce the grade of its channel is greater than if the bed is inclined. When the rocks are horizontal, the life of a cataract may be immensely prolonged. The only cases in which waterfalls produced by hard ad- jacent to soft rocks do not migrate, are when the hard layer is vertical. In such a case the change in the position is limited to the thickness of the resistant bed. In the Cas- cade Mountains there are numerous rapids and cascades due to vertical dikes of basalt. These dikes are harder than the adjacent rock, and cause inequalities in the beds of the streams crossing them, but the positions of the falls produced remain essentially the same throughout their lives. In the streams flowing eastward from the Appalachians, falls and rapids occur where they leave the hard crystalline rocks forming the Piedmont Plateau and enter the soft rocks of the Coastal Plain. As these falls recede up stream, they leave canyons to mark the paths they follow. Many of the falls in the drift-covered region of North America are due to the turning of streams from pre-glacial valleys in such a way as to cause them to flow over what were formerly divides or rocky spurs between adjacent streams and plunge into valleys. In some instances, also. 62 RIVERS OF NORTH AMERICA they have cut through a covering of drift and been lowered upon harder rocks beneath. In either case falls may result. Cascades are also produced where these streams leave a region of hard rock and enter areas of drift which is more easily removed. Niagara Falls came into existence when a large lake, which formerly flooded both the Ontario and Erie basins, was lowered so as to be divided into two water bodies by a ridge trending east and west, formed by the summit of the Lewis- ton escarpment. The Falls of St. Anthony are due to the Mississippi having been turned from its pre-glacial course by deposits of drift, and made to flow over the surface of a compara- tively thin horizontal sheet of limestone resting on soft sandstone. A cataract about one hundred feet high was produced where the river left the edge of the limestone layer and plunged into a pre-glacial valley. From this escarpment the fall has receded about eight miles, leaving a steep-sided canyon, as in the case of Niagara. Shoshone Falls, Idaho, were produced by a hard sheet of trachyte which Snake River discovered as it sank its chan- nel in nearly horizontal layers of basalt. The falls have migrated up stream, leaving a narrow canyon as a record of the work already performed. Many beautiful cascades in the Rocky and Cascade Mountains, and others no less picturesque in the Sierra Nevada, are due to changes produced by a former period of glaciation. In some instances in the Sierra Nevada, large Alpine glaciers flowed down the main valleys, and blocked the streams in lateral gorges so as to cause them to cease ' INEQUALITIES IN THE HARDNESS OE ROCKS 63 corrading. The glaciers deepened the main valleys, how- ever, during the time the development of their tributary streams was arrested, and when the ice melted and water drainage was once more established, the branches of the main streams were compelled to descend steep precipices and in many instances form fine cascades, in order to reach the bottoms of the glacier-deepened main valleys. The ancient glaciers, to which so many references have been made, brought destruction in their paths as they ad- vanced and left fields of desolation as they retreated, but in many ways the beauty of the region they occupied was en- hanced by the changes they made. Our greatest debt to the vanished glaciers, so far as the revolutions they wrought appeal to our artistic sense, is for the tens of thousands of placid lakes they left strewn over the land, and the tens of thousands of leaping waterfalls which sprang into exist- ence on their retreat. The former are emblems of rest, the latter of ceaseless activity. Bluffs Bordering Aged Streams, — As previously stated, the influence of the unequal yielding of hard and soft rocks is most marked in the case of streams that are still young, and decreases as they advance in development, but seldom entirely disappears. Topography being largely the result of the action of streams, it follows that the various features in the relief of the land must be due, to a marked extent, to the un- equal waste of hard and soft rocks. The hard rocks stand as bluffs, ridges, and peaks, while the soft rocks are worn away more rapidly, and dells and valleys appear. The various stages from youth to old age, so characteristic of 64' RIVERS OF NORTH AMERICA streams, find a counterpart in the general changes in form and expression experienced by the surfaces of land areas. A new land area may have a generally even surface, but as time passes, and topographic maturity is reached, it becomes roughened, and if the elevation has been great, and marked inequalities in the hardness of the rocks occur, exceedingly rugged topographic forms will be developed. When a land area has been long exposed, the inequalities of surface due to differential weathering gradually decrease, but except in the rare instances of nearly complete baselevelling, do not disappear. Throughout the courses of streams that have passed their periods of maturity, and even after having developed a gentle gradient characteristic of old age, their valley walls frequently retain evidences of the great topographic diver- sity that characterised them during youth. Rivers flowing through lands having in general all the characteristics of topographic old-age may yet be bordered in places by steep bluffs and overshadowed by towering precipices. The Highlands of the Hudson, where the river valley is narrow and bordered on each side by rugged mountains, in contrast with the wider portions above and below, where the bordering uplands are less precipitous, reveal the influence of hard rocks on the scenery of an ancient river. The picturesque ** coves** in the Southern Appalachians, as along the upper course of the Hiawassee, have been hol- lowed out in soft beds, and are surrounded by precipitous mountains of hard rock. Many bold headlands in the upper Mississippi valley are remnants of ancient eminences, rounded and worn by long INEQUALITIES IN THE HARDNESS OF ROCKS 65 exposure, which in several instances rise directly from the border of the river that sweeps about them and has long since passed its period of youth. Much of the wonderfully impressive scenery of the Columbia is due to great bluffs of basalt which rise di- rectly from the river's brink, and on account of their hardness, in contrast with softer beds adjacent, remained prominent even after the river had cut down its channel to an even grade and become navigable. The same sequence of events was noted by the writer in many instances while ascending the Yukon. That noble river, although well adjusted to the various rock conditions it discovered as it deepened its channel, and flowing with such an even grade that it can be ascended by steamboats for over fifteen hundred miles, is bordered in places, as shown in Plate II., by magnificent bluffs of hard rock which intervene between long reaches where the valley is several miles broad, and has been excavated in softer beds. While the persistence of the topographic forms on the borders of river valleys is conspicuous, and accounts for many of the more prominent features adjacent to aged rivers, yet the lives of many streams have been so greatly prolonged that movements in the earth's crust have produced changes simulating those just considered. The rocks crossed by a great river may be upraised so as to form ridges, or even mountain ranges, athwart its course, and dam its waters or turn them aside. When such changes occur, however, with sufficient slowness to allow the river to deepen its channel as fast as the rocks rise, it will maintain its right of way, and excavate a gorge or canyon through the obstruction. 5 66 RIVERS OF NORTH AMERICA In such instances a portion of the stream will have the charac- teristics of youth, while adjacent portions above and below, whose development was unchecked, present all the features of old age. Rivers which maintain their right of way in the manner just cited, and carve gorges and canyons through newly elevated lands, have been termed antecedent rivers by Powell, in recognition of the fact that they are antecedent to the movement which causes the rocks to be elevated. CHAPTER IV MATERIAL CARRIED BY STREAMS IN SUSPEN- SION AND IN SOLUTION THE waters flowing from the land back to the sea, whence they came as vapour, carry material with them, as is well known, in two distinct ways: namely, in suspension and in solution. The debris carried in suspen- sion and rolled along the bottom, or the visible load, as it may be termed, sooner or later finds its way to the sea and forms stratified deposits. The material dissolved by the waters during their excursion through the air and over the land, or their invisible load, goes to increase the salinity of the sea and to supply marine plants and animals with sub- stances necessary for their growth. THE VISIBLE LOADS OF STREAMS The material transported mechanically by streams may be divided into two classes: 1st, the portion rolled and pushed along the bottom; and, 2d, the portion lifted well above the bottom and carried forward in suspension. The divid- ing plane between these two classes is indefinite, as much of the material moved along the bottom makes short up- ward excursions, and the fine particles normally carried forward in suspension from time to time rest on the bottom. 67 68 RIVERS OF NORTH AMERICA Bottom Load, — Concerning the manner in which the bot- tom load, as it may be termed, is moved, and the amount of such transportation in a given stream, but httle information is available. If we watch a clear stream supplied with sand or gravel of such size that the current has power to move it, it will be seen that the debris does not advance as a continuous sheet, but rather as a succession of wave-like forms. The action of the water-current in this respect is similar to the be- haviour of air-currents when moving over dry sand. The ripple-like ridges on the bottom of streams are frequently and probably always broad in reference to their height. The up-stream slope of each ridge is gentle and its down- stream border short and precipitous. Grains of sand are moved over the broad gently ascending surface and rolled down its steep down-stream margin. At the base of the steep border of each ripple-like sheet, there are secondary currents caused by the plunging of the water, and the particles forming the bottom are there disturbed and carried onward and the process repeated. When the material at the bottom is in excess of the transporting power of the stream, sedimentation takes place, and cross-stratified or current-bedded accumulations result ; but if the bottom cur- rent is under-loaded, the material is carried forward by being removed from the up-stream margin of a broad ripple-like sheet, and re-deposited on its steep down-stream margin. The process just described goes on at the bottom of clear streams, and illustrates the fact that such streams, contrary to what is sometimes stated, have power to corrade. It is only when their bottoms are swept clean of all grains of such MATERIAL CARRIED BY STREAMS 69 sfze as are within the capacity of the stream to sweep away, that mechanical corrasion ceases. These statements concerning the bottom loads of streams may be said to be qualitative, inasmuch as measures of the amount of material thus transported are lacking. It is diffi- cult and at present seemingly impossible to ascertain how much material a large river is moving in the manner just considered. Bottom transportation will evidently vary with changes in conditions, being favoured by swiftness of cur- rent and the character of the debris available for transporta- tion. Variations probably also occur in reference to the amount of material a stream carries in suspension. If a stream is heavily charged with silt, the friction of flow will be increased, and as this friction is greatest in proportion to rate of flow, near the bottom, it is to be expected that bot- tom transportation will be checked while transportation in suspension is still actively progressing. It would seem, therefore, as if bottom transportation is favoured by de- crease of material in suspension ; or, other conditions being the sam.e, clear streams have a greater power to move bot- tom loads than muddy streams. I must confess, however, that this is theory rather than a deduction from observations and experiments, and the reader is invited to test the con- clusion for himself. It is evident that the principal conditions favouring bot- tom transportation are velocity and volume of water. As velocity increases with declivity, we should expect that high-grade streams would move proportionately heavier loads along their bottoms than low-grade streams, other conditions being the same. The ratio of bottom load to 70 RIVERS OF NORTH AMERICA total transportation should be greater during floods than during low-water stages. Observation seems to confirm these conclusions. The proportion of bottom load to the amount of material carried by a stream in suspension is dependent largely on the character of the debris within the reach of the stream — that is, whether it is fine or coarse ; but in general it seems true, as just stated, that the bottom load in swift streams is greater in proportion to the amount of material in suspen- sion, than is the case in slower streams under similar second- ary conditions. Although the manner in which bottom loads are carried is not thoroughly understood, and the amount of such trans- portation difficult to determine, the fact remains that much of the energy of streams is consumed in rolling debris along their bottoms. In many measures of the rate at which streams are removing rock debris from their drainage basins, the quantity moved along their bottoms is not considered. For this reason most estimates of the rate at which land areas are being lowered by denudation require important modifications. Measures of Material in Suspension, — The methods em- ployed for ascertaining the amount of sediment carried by a stream in suspension are illustrated by the careful work done in this connection by Professor Forshey, during the survey of the Mississippi by the United States Topographic Engineer Corps.* Stations were selected near Carrollton, a short distance above New Orleans, one about three hundred ' Humphreys and Abbot, Report upon the Physics and Hydraulics of the Mississippi River ^ p. 137, 186 1. MATERIAL CARRIED BY STREAMS 7 1 feet from the east bank of the river, the next in mid-stream, and a third about four hundred feet from the west bank. The high-water depths at these stations were lOO, lOO, and 40 feet respectively. Samples of water were collected daily at surface, mid-depth, and bottom at the first two stations; and at surface and bottom at the third station for a period of one year. During the succeeding year, the ratio between the sediment contained in the water at any one station and that contained in the entire cross-section of the river having been ascertained, one sample was taken each day from the surface at the station near the east bank. The samples from below the surface were secured by means of a small keg heavily weighted at the bottom and provided at each of its ends with a large valve opening up- ward. These valves allowed a free passage to the water while the keg was sinking to the required depth, but pre- vented its escape while being drawn up. When the keg reached the surface, the water contained in it was thor- oughly stirred and a bottle filled from it. The sediment contained in the water samples was subsequently filtered out and weighed after drying. From the tabulated results of the first year's work re- ferred to, it was found that the greatest amount of sediment was carried in June, during the annual high-water stage of the river, the weight of sediment then being -g-J^ of the weight of the river water containing it; the minimum was obtained during the low-water stage late in October, the ratio of weight of sediment to weight of water then being as I to 6383. The mean for the year was y^Vs"^ ^^ ^^^ ^^^ ^f sediment to 1808 tons of water. 72 RIVERS OF NORTH AMERICA A discussion of a still larger number of observations, made under the direction of Humphreys and Abbot, gave the ratio of i of sediment to 1500 of water by weight, and of about I to 2900 by volume. The variation in the amount of sediment with positions in the stream is indicated in the following table of the weekly means for the two months of highest and lowest water respectively : SEDIMENT IN THE MISSISSIPPI AT CARROLLTON FIRST POSITION. SECOND POSITION. THIRD POSITION. NUMBER OF WEEK. 1 3 C/3 6 PQ S oi 3 CO First in June, 1851 Second *' " " Third '' " " Fourth " " " . 0.345 0.456 0.917 0.498 0.407 0.507 0.960 0.570 0.187 0.510 0.940 0.557 0.365 0.477 0.731 0.528 0.415 0.515 0.981 0.597 0.410 0.517 1. 105 0.601 0.285 0.365 0.666 0.427 0.3900.365 0.4570.442 1.0460.447 0.5360.452 Mean for June 0.559 O.61I 0.548 0.525 0.627 0.658 0.436 O.6O7JO.426 First in October, 1851.. Second '' '' " . . Third '' Fourth" 0.137 0.120 O.IOO 0.068 0.187 0.169 0.132 0.096 0.220 0.170 0.136 0.106 0.125 0.109 0.097 0.059 0.215 0.193 0.146 O.I 15 0.235 0.220 0.159 0.116 0.096:0.265 0.170 0.107 0.235 0.092 0.0890.195 0.071 0.061 0.136 o.oSi Mean for October 0.106 0.146 0.158 0.096 0.172 0.182 1 1 0.0880.2080.104 The figures denote the number of grammes of dry sediment contained in 600 grammes of river water. Knowing the amount of water discharged annually by the Mississippi and the proportion of sediment contained in it, the amount of material carried by the stream each year in MATERIAL CARRIED BY STREAMS 73 suspension may be readily computed. The mean annual discharge, as determined by the survey in charge of Hum- phreys and Abbot,* is 19,500,000,000,000 cubic feet, and the amount of soHd matter carried in suspension 812,500,000,- , 000 pounds.^ The average specific gravity of this material is about 1.9; with this density, the sediment carried annually would occupy 6,718,694, 400 cubic feet, or sufficient to cover one square mile to the depth of 241 feet. In addition to the silt carried in suspension, it has been estimated by the engineers cited above, that the amount of sand and gravel rolled along the bottom and contributed each year to the filling of the Gulf of Mexico is about 750,000,000 cubic feet ; making the total visible load carried by the river each year about 7,468,694,400 cubic feet, or sufficient to cover one square mile to a depth of 268 feet. The Mississippi has been more carefully studied than any other river in North America, but it is well known that other streams are doing a similar work. In many streams the proportion of material in suspension to the amount of water is greater than in the lower Mississippi ; while rivers might be selected, as, for example, the St. Lawrence, in which the percentage of sediment is much less. An inspec- ' Report on the Mississippi River, p. 149. If I understand this portion of Humphreys and Abbot's report correctly, the above measures do not include the three outlet bayous, which leave the main river above New Orleans. It is stated on page 93 of the report, that, including these bayous, the annual discharge is 21,300,000,000,000 cubic feet of water. -Taking the specific gravity of water as i, the relative weight of coarse river-sand is 1.88; fine sand, 1.52 ; clay, 1.90; alluvial matter, from 1.92 to 2.72. A cubic foot of water weighs 62.5 lbs. ; of coarse sand, 11 7. 5 lbs. ; fine sand, 95 lbs. ; clay, 118.75 lbs. ; alluvial matter, 120 to 170 lbs. ; silt, 103 lbs. W. H. Wheeler, Tidal Rivers ^ p. 62. Published by Longmans, Green, «& Co., 1893. 74 RIVERS OF xXORTH AMERICA tion of the following table, compiled by C. C. Babb,' in which data concerning the amount of material that is being carried by several large rivers are presented, shows that the Mississippi, although commonly recognised as a muddy stream, holds a smaller percentage of silt in suspension than several other rivers with which it may be compared. DISCHARGE AND SEDIMENT OF LARGE RIVERS z < Ji Q MEAN ANNUAL DIS- CHARGE IN CUBIC FEET PER SECOND. SEDIMENT. RIVER. 'c3 Ratio of sedi- ment to water by weight. Height of col- umn, one sq. mile base. Feet. Potomac Mississippi Rio Grande Uruguay Rhone 11,043 1,244,000 30,000 150,000 34,800 27,100 320,300 1,100,000 125,000 20,160 610,000 1,700 150,000 65,850 62,200 315,200 113,000 475,000 5,557,250 406,250,000 3,830,000 14,782,500 36,000,000 67,000,000 108,000,000 54,000,000 291,430,000 3,575 1,500 291 10,000 1,775 900 2,8So 2,050 1,610 4.0 241.4 2.8 10.6 31. 1 59.0 93.2 38.8 209.0 .00433 .00223 .00116 .00085 .01075 .01139 .00354 .00042 .02005 Po Danube Nile Irrawaddy Mean 334,693 201,468 109,649,972 i: 2,731 76.65 .00614 Estimates similar to those given in the above table have been published by several geologists. / One series of these, probably as reliable as any, by Archibald Geikie,'* is here copied for the purpose in part of showing that the observa- tions now available concerning the work of streams are de- fective. Even the areas of hydrographic basins are stated differently by different writers, and with perhaps a few ' Science, vol. xxi., p. 343, June, 1S93. ^ Text-Book of Geoiogy, 2d edition, p. 428. Macmillan & Co., 1885. MATERIAL CARRIED BY STREAMS 75 exceptions the measures given of the annual discharge of large rivers, the amount of sediment they carry, etc., should be considered as subject to corrections. SEDIMENT OF RIVERS RIVER. AREA OF BASIN IN SQUARE MILES. ANNUAL DISCHARGE OF SEDIMENT IN CUBIC FEET. FRACTION OF FEET OF ROCK BY WHICH THE AREA DRAINED IS LOWERED IN ONE YEAR. Mississippi 1,147,000 143,000 700,000 25,000 234,000 30,000 7,459,267,200 6,368,077,440 i7,52o,ooo,ooo(?) 600,381,800 1,253,738,600 1,510,137,000 1 Ganges (UpDer) "5" "011X7 Hoang Ho Rhone Danube Po ^ST^ A brief discussion of the rate at which land areas are being lowered by the removal of material by streams will be given after the measures of mineral matter in solution have been considered. THE INVISIBLE LOADS OF STREAMS Water as it reaches the land as rain, snow, dew, etc., is never chemically pure, but contains both organic and in- organic matter in solution and dust particles in suspension. The substances most commonly occurring in solution in rain-water are shown by the following analysis of a sample collected near London, England ^ : .99 part in 1,000,000 of water. .22 .50 ** .07 Organic carbon Organic nitrogen Ammonia Nitrogen as nitrates and nitrites. Chlorine 6. 30 parts in Total solids 39-50 " * Quoted by W. P. Mason, Water Supply, p. 204. John Wiley & Sons, 1896. 76 RIVERS OF NORTH AMERICA Examinations for chlorine in water samples representing the average condition of the rain-water at Troy, New York, for one year, gave a mean of 1.64 parts in a million. That is, each million pounds of rain-water contained 1.64 pounds of chlorine in solution.' The impurities in rain-water vary in character and amount in different localities. In general they are greatest near cities, and least in the open country at a distance from vol- canoes, gas springs, etc. They also vary with climatic con- ditions, being greatest in arid and least in humid regions, and greater in dry than in wet seasons. The amount of common salt is large near the sea, and normally decreases inland, but probably reaches a maximum in the neighbour- hood of saline lakes and over salt deserts. Rain-water, then, comes to the earth with its solvent power increased by the presence of various substances washed out of the air, but its ability to take up mineral matter in solution is greatly increased as it flows over the land. The soil usually contains organic matter which is easily dissolved. The most common substances thus added to the water, which enhance its chemical activity, are carbonic acid or carbon dioxide (COg), and a large group of organic acids, known as the humus acids; these, however, are unstable, and soon change to carbon dioxide. The organic acids are derived mainly from the decay of vegetation, but in part are of animal origin.'^ The percentage of the organic acids taken in solution by ^ W. p. Mason, Water Supply, p. 205. * A. A. Julien, " On the Geological Action of the Humus Acids." in Amer- ican Association for the Advancement of Science, Proceedings, vol. xxviii., pp. 311-410, 1879. MATERIAL CARRIED BY STREAMS J 7 a given quantity of water percolating through the soil varies with different localities, being greatest when decaying vegetation is most abundant and where the temperature is high. In all portions of the earth's surface, however, the water, on coming in contact with the soil or with solid rocks, has the power to dissolve portions of them. The water which runs over the surface and is gathered quickly into streams has less opportunity to take up mineral matter in solution than that which percolates through the soil and in many instances descends into the hard rocks beneath and comes to the surface again as springs. The water flowing quickly over the surface and that following more or less ex- tensive underground courses are commingled in the streams, and send their combined tribute of dissolved matter to the sea. Chemical denudation thus assists the mechanical ac- tion of flowing water in lowering the land, and is an import- ant factor in the process. The rate at which rocks are dissolved varies not only with the rain-fall, with the amount of organic acids in surface and subterranean water, and with temperature, but is influenced especially by the nature of the rocks in various regions. The solution of mineral matter in general is greater, other con- ditions remaining the same, the higher the temperature, although this does not apply to limestone, and is greatest where the rocks are composed of easily soluble minerals. For these reasons the chemical composition of river-water varies, but the departure from a mean, as shown by a large number of analyses, is less than might at first be expected. By the time the surface waters have united to form rills they contain sufificient mineral and organic matter to give 78 RIVERS OF NORTH AMERICA them a complex chemical composition. Throughout their journeys to the ocean, as they form brooks, creeks and rivers, and especially when travelling underground, they be- come more and more highly charged with dissolved mineral matter. The longer the waters are in contact with soil and rocks, and with the finely divided material held by them in suspension, temperature conditions, etc., remaining the same, the more highly charged they become with substances in solution. Evaporation also tends to concentration, but this process, particularly in humid regions, is more or less completely counteracted by direct precipitation. River-waters, filtered of all material in suspension, and evaporated to dryness, leave a solid residue, which is the principal portion (the more volatile substances escaping) of the foreign matter previously held in solution. These waters are fresh in the every-day use of the term, but in fact owe their agreeable taste and, to a certain extent, their health-giving qualities, to the mineral salts and gases con- tained in them. Irx Table A, analyses are given of the waters of a number of American rivers, which show that the principal substances in solution are calcium and carbonic acid, probably combined as calcium bicarbonate. In some instances, however, as in the case of Jordan River, Utah, calcium sulphate is in excess of all other salts. From a large number of analyses of water samples ob- tained from the rivers of Canada and the United States, it has been found that the average amount of total solids in solution is o. 1 5044 part in a thousand by weight ; of this material, 0.056416 part in a thousand is calcium carbonate. In a table of forty-eight analyses of European river-waters MATERIAL CARRIED BY STREAMS 79 given by Bischof,' the average of total solids in solution is 0.2127, and the average of calcium carbonate 0. 1139 part per thousand. From the analyses of thirty-six European river-waters, published by Roth,'^ including some of those tabulated by Bischof, the average of total solids is 0.2033, and of calcium carbonate 0.09598 part per thousand. In both American and European river-waters, so far as can be determined from the data in hand, the average of total solids is 0.1888, and of calcium carbonate 0.088765 part per thousand. These figures may be assumed to repre- sent the average of the solids in solution in the waters of normal rivers. It will be noticed that the average for cal- cium carbonate is nearly one-half the average for total solids. Knowing the annual discharge of a river and the percent- age of mineral matter carried in solution, we can ascertain the amount of dissolved matter that the river contributes annually to the ocean, or enclosed lake into which it flows. To one unfamiliar with studies of this nature, the amount of rock-forming material thus annually transported by a large river in an invisible state is astonishing. The follow- ing table, showing the total amounts of solids in solution carried by certain rivers, has been compiled from various sources : Rhine. 5,816,805 tons per year. 8,290,464 Rhone Danube 22,521,434 Thames 613,930 Nile 16,950,000 Croton 66,795 Hudson 438,000 Mississippi 112,832.171 Chemical Geology, vol. i., pp. 76, 77. English edition, London, 1854. '■ Allgemein und chemische Geologic, vol. i., pp. 456, 457. Berlin, 1879. 8o RIVERS OF NORTH AMERICA The fact that streams transport great quantities of dissolved mineral matter, derived from the rocks in the basins they drain, may be shown by computing the numbers of tons of material in solution in a cubic mile of river-water. This has been done by John Murray,' and the result, based on the average composition of the waters of nineteen of the principal rivers of the world, is given below: MATERIAL IN SOLUTION IN ONE CUBIC MILE OF AVERAGE RIVER-WATER ^ CONSTITUENTS. TONS IN CUBIC MILE. Calcium carbonate (CaCOg) 326,710 Magnesium carbonate (MgCOg) 112,870 Calcium phosphate (CagPgOg) 2,913 Calcium sulphate (CaSO^) 34,30i Sodium sulphate (NagSO^) 31,805 Potassium sulphate (K2SO4) 20,358 Sodium nitrate (NaNOg) 26,800 Sodium chloride (NaCl) 16,657 Lithium chloride (LiCl) 2,462 Ammonium chloride (NH4CI) 1,030 Silica (Si02) 74,577 Ferric oxide (FegOg) , 13,006 Alumina (AlgOg) I4,3I5 , Manganese oxide (MngOg) 5,703 I Organic matter 79,020 I Total dissolved matter 762,587 It has also been computed by Murray, and published in the article just cited, that the volume of water flowing to the sea in one year, including all the land areas of the earth, is about 6524 cubic miles. From the average chemical com- position of river-water, it follows that about 4,975,117,588 tons of mineral matter in solution are being removed annu- ^ Scottish Geographical Magazine^ vol. iii., p. 76, 1887. ' Acids and bases combined according to the principles indicated by Bunsen. MATERIAL CARRIED BY STREAMS 8 1 ally from the land area of the earth. This process of removing material of the land in solution has, very properly, been termed clicmical demidation. It is instructive to follow the history of the material carried in solution by rivers, and to see what changes occur, especially in inland seas where ordinary river-waters are concentrated by evaporation, and in many instances the salts they contain precipitated in a crystalline form, and to ex- tend such studies to the ocean. Another fruitful line of investigation in this connection is the manner in which mineral matter in solution is eliminated. This occurs in part, as just mentioned, by chemical precipitation, but is effected to an equally great extent, and from dilute solu- tions, through the action of plant and animal life. These interesting studies, however, lie beyond the scope of our present thesis. RATE OF LAND DEGRADATION Measures of the amount of material carried by streams both mechanically and in solution furnish a means of ap- proximately determining the rate at which the surface of the land is being degraded. Mechanical Degradation. — As shown in the table on page 74, the amount of silt carried annually by the Mississippi, if taken uniformly from the area it drains, would lower it j^tt of a foot. That is, considering only the material carried in suspension, the basin is now being lowered at the rate of one foot in 5376 years. If we take into account also the material rolled along the bottom, computed to be 750,000,- 82 RIVERS OF NORTH AMERICA ooo cubic feet per year/ we find that the basin is being^ lowered at the rate of one foot in 4638 years. Studies of the Potomac River, conducted by the United States Geological Survey, have determined the fact con- cerning that river presented in the table on page 74. As- suming that one cubic foot of the silt carried by the Potomac weighs 100 pounds, the average annual amount transported would cover one square mile to a depth of 3.98 feet. If this amount should be taken uniformly from all parts of the area drained, it would be lowered 0.0043 of an inch, or ^7^5- of a foot. In other words, the Potomac is lowering its hydrographic basin at the rate of one foot in 2772 years. Other similar estimates are given in the table on page 75, which, if approximately correct, might be taken as indicat- ing the work that the rivers of the world are doing. It is probable, however, that except in the case of the Mississippi, in the table compiled by Geikie, the bottom loads of the rivers are not included. I am also inclined to doubt the ac- curacy of some of the other measures referred to. The average for the nine rivers tabulated is one foot of denuda- tion in about 9000 years. Chemical Degradation. — The importance of the slowly acting and invisible process by which the surface of the land is being lowered by solution, has only recently been recog- nised. The earliest definite discussion of the rate of chemi- cal degradation now in progress, so far as I am aware, is in a series of three papers by T. Mellard Reade.' In these • A comparatively slight discrepancy comes in here, since the specific gravity of the bottom load and of the silt in suspension is not the same. ^ Republished with the title, Chemical Denudation in Relation to Geological Time. Daniel Dogue, London, 1879. MATERIAL CARRIED BY STREAMS 83 instructive essays it is estimated that the amount of material removed in a century by the streams of England and Wales in solution, if spread evenly over the land from which it is derived, would have a thickness of .0077 of a foot. That is, it will take 12,987 years to denude the surface of England and Wales of one foot of solid matter by the process here considered, under the supposition that the material is taken evenly from all parts of the surface. The Mississippi, as previously stated, carries annually about 112,832,171 tons of mineral matter in solution. This amount of material may be considered as about equivalent to 1,350,000,000 cubic feet of limestone, and if spread evenly over the Mississippi basin would cover it to the depth of about ^y-J-o-o ^^ ^^^ iooX. ; or, in other words, chemical degradation is lowering that area at the rate of one foot in 25,000 years.' Rate of Both Mechanical and Chemical Degi'adation. — The best and in fact the only approximately reliable measures we have of the rate of the combined mechanical and chemi- cal degradation by the rivers of North America, is in the case of the Mississippi. On account of the large size of the drainage area of that river and the variety of rocks forming its surface, as well as the diversity of climate included within its border, it may be reasonably assumed to represent about the average rate of degradation which is being performed by rivers in general. ^ The material in solution is taken in part from the surface and in part from below the surface. While an estimate of the average lowering of the surface by degradation during a single year need not perhaps include the material removed in solution from below the surface, yet this should certainly be taken into account in estimates of average degradation. a given region is accompanied by a decrease in the volume of the draining streams, and consequently a loss in their trans- porting power. The behaviour of streams under such con- ditions is materially influenced by the rate at which they are supplied with debris. During heavy rains a stream may be overloaded and caused to deposit, in spite of its increased 142 RIVERS OF NORTH AMERICA velocity due to greater volume, and the amount of work done in a given time is far in excess of that accomplished during an equal time when the precipitation is less. The influence of variations in precipitation is illustrated by the annual change in streams during rainy and dry sea- sons. During rainy seasons, more especially in spring in temperate latitudes, when the rain causes the melting of previously accumulated snow, the streams are swollen and heavily charged with debris. They overspread their banks and deposit material on their flood-plains. It is during the time the streams overflow their banks that the greater amount of material is deposited. Much debris is also laid down at such times, however, in the stream beds, even when the current is swift, and in some instances the less heavily loaded water, when much decreased in volume, corrades channels through deposits made during high-water stage. This may be seen in many wayside rills, which spread out in sheets, heavily charged with debris, during storms, and make deposits through which the shrunken and less heavily charged rills at a later stage excavate channels. THE GENERAL PROCESS OF STREAM CORRASION AND DEPOSITION The action of streams in corrading, transporting, and de- positing .debris is such a complex process that it is con- venient to consider the different phases of their work separately. For this reason, an effort has been made in this chapter to direct special attention to the manner in which streams lay aside their loads during the process of STREAM DEPOSITS 1 43 development that they pass through. The behaviour of streams is much like the action of a complex piece of machinery, as a watch, for example; changes cannot be made in one portion of the mechanism without affecting the action of the whole, and necessitating adjustments through- out. Considering deposition alone, we find that streams in general, in passing from a high to a low grade, make de- posits, as where a river leaves its mountain tract and enters a valley tract, or passes from a swift to a more quiet reach. Streams subject to floods make deposits over the lands they inundate during high-water stages, and spread out flood- plains. At such times, also, the heaviest deposit is in the immediate border of the low-water channel, and natural levees are built. A high-grade stream, tributary to a low- grade and consequently less swift stream, unless the differ- ence in grade is more than counterbalanced by the volume of the receiving stream, makes deposits and the waters of the main stream are more or .less completely ponded. Local overloading from the action of the wind or of glaciers produces similar results. The debris carried in suspension by streams or rolled along their beds is also deposited in lakes as deltas, or dis- tributed over their bottom. As lakes in many instances are of the nature of expansions of streams, the filling of their basins may be considered as a part of the general process of stream deposition by which stream channels are aggraded. In discussing corrasion it was shown that a stream, at least in humid climates, cuts down its channel to baselevel most quickly at its mouth, and that the process of deepening pro- gresses up stream. The head-waters of a well-developed 144 RIVERS OF NORTH AMERICA stream are steeper than the lower portion of its trunk. A general view of stream deposition shows that a similar order is followed in the process of stream deposition. When the seaward portion of the trunk of a stream has been lowered to baselevel, the stream continues to corrade laterally, and thus makes it possible for flood-plains to form. As a stream continues to widen its channel farther and farther from its mouth, the flood-plain follows. If a stream is making a delta, its length of flow is increased, and its flood-plains and channel are raised by deposition in order to furnish the necessary slope. When a stream reaches maturity, its plains tract and valley tract are greatly length- ened at the expense of the high-grade portions of its course in the uplands. The high-grade branches, then, bring more material than the trunk stream can bear away, and the flood-plains along its sides are raised by the de- position of material laid aside. During the upbuilding of the flood - plains the stream channel is also raised. Hence, for a long time after a normal river has cut down its channel in its lower course practically to baselevel, building is in progress and the valley becomes filled in with abandoned debris. There comes a time, however, when the highlands from which the river flows have been lowered so that the branches of the main stream are not as swift as previously, and the stream is enabled to devote a portion of its energy not consumed in the friction of flow to the re- excavation of its channel farther seaward. As this process is continued, the highest flood-plain is abandoned and new ones formed at lower levels, thus giving origin to terraces, as will be shown in the next chapter. During this stage of STREAM DEPOSITS I45 a stream's development, as in the preceding stages, changes occur, also, in the longitudinal profile of the stream throughout its length. The manner in which flood-plains are formed and advance up stream as the down cutting of the upper portion of a stream channel progresses, shows that only an approximation to baselevelling is reached during the earlier stages of a stream's development. It is after a stream has lowered its head-water channels so as to permit of the removal of the flood-plains built lower downstream, that what may be termed a second approximation to baselevel is normally reached. PROFILES OF STREAMS In a discussion of the succession of changes experienced by a stream during its life, consideration should be given to the orderly variations in shape that occur in the valley it excavates. The shape of a valley may be illustrated by two classes of profiles; one longitudinal and the other transverse. A gen- eralised longitudinal profile of a stream would be what is understood as a projection on a vertical plane ; that is, it is approximately the profile which the stream would have if it flowed in a perfectly straight course from source to mouth. Such a profile, together with a sufficient number of cross- profiles, would enable one to construct a model of a valley showing the actual relations and proportions of its several parts. The Long it itdhial Profile, — A young stream flowing down the surface of a tilted plain, we will assume, will necessarily 146 RIVERS OF NORTH AMERICA have the same gradient as the land which gave direction to its current. As such a stream entrenches itself by corrading the bottom of its channel, and during the process of cutting down to baselevel spreads out flood-plains, which are sub- sequently dissected, it will develop a series of profiles to suit the various stages of its development. An ideal example of the succession of longitudinal profiles which a stream makes, may be had by assuming that it works in homogeneous rocks throughout its course and is not disturbed by changes of climate, the formation of glaciers, or other modifying conditions. When the typical profile of a young stream and the changes it passes through as the stream advances in its appointed task is understood, the modifications due to climatic and other disturbances, or accidents, as they may be termed, can be readily recognised. As has been shown with exceptional clearness by Hicks, * corrading streams have curved profiles, the curvature being concave upward, while deposits laid down by currents, such as alluvial cones, have a reverse curvature, that is, they are convex to the sky. The longitudinal profile of a stream which is corrading in its mountain tract and spreading out a flood-plain farther down its course, must therefore have a double curvature — that is, it will be concave in its upper course but convex in its lower course. The concave portion of the curve is much more conspicuous than the more gentle curve due to deposition, and it is frequently stated that the profile of a stream is a concave curve throughout its length. This, however, can only be strictly true when a * L. E. Hicks, " Some Elements of Land Sculpture," in Bulletin of the Geo- logical Society of America^ vol. iv., pp. 133-146, 1893. STREAM DEPOSITS I47 Stream is engaged in corrading its channel from source to mouth. A generalised profile of a stream which is corrading its channel throughout is shown approximately in the following diagram. It will be noticed that the curvature is compara- FlG. 7. Longitudinal Profile of a Young Stream. tively great near the source of the stream, but decreases and becomes nearly horizontal on approaching its mouth. There is a suggestive resemblance between such a profile and cycloid curves. As is well known, a cycloid curve is the curve of quickest descent for a body moving from a given point to a lower one not in the same vertical line. Should accurate survey show that streams corrading homogeneous rocks actually produce cycloid curves, or the curves of quickest descent for their debris-charged waters, it will fur- nish another illustration of the wonderful harmony that pre- vails in nature. A stream in cutting down its channel to baselevel must evidently reach that limit first at its mouth, and will then continue to deepen its bed progressively up stream. If this operation should be allowed to go on with- out deposition and the formation of flood-plains, the result would evidently be the flattening of the curved profile from the mouth of the stream to its source at the same time that the elevation of the stream channel above sea-level was lowered progressively and in an increasing ratio from mouth to source. Corrasion, however, is accompanied by sedi- 148 RIVERS OF NORTH AMERICA mentation. In young streams, corrasion may occur through- out their courses, but as soon as their mouths are lowered to baselevel, deposition begins and progressively advances up stream. The longitudinal profiles of most streams result from both corrasion and sedimentation, and have a double curvature. Corrasion is more active in the mountain tract than in the valley and plains tracts, and until these divisions are obliterated by advancing age, the profile of a stream is, in part, due to corrasion and in part to sedimentation. With advancing age the portion of the curve due to deposi- tion advances up stream at the expense of the steeper portions of the profile where corrasion is still in progress. There comes a time in the development of a stream, how- ever, when this advance is checked, and when the flood-plain deposits begin to be dissected ; the swing is then the other way, and the portion of the profile due to corrasion is lengthened and progresses toward the mouth of the stream. In old age the profiles of streams become flattened and ap- proach more and more nearly a straight line, but probably never reach that condition. Fig. 8. Successive Changes in the Profile of a Divide Owing to Corrasion and Weathering : Vertical Scale Exaggerated. The heavy broken line indicates the profile of an uplift as it might appear had there been no erosion ; the smaller broken lines show weather-curves ; the dotted lines, successive cor- rasion curves ; and the solid curved line below, the profile of the resulting old-land surface. In the above diagram an attempt is made to show quali- tatively the successive changes that the profiles of streams pass through from youth to old age. In the case assumed, STREAM DEPOSITS 1 49 two streams flow in opposite directions from a common divide, and are so nicely balanced against each other that the divide has been lowered in a single vertical plane. The concave curvature of the profiles in their upper courses in- creases during early youth, reaches its maximum when the streams are mature, and then decreases with advancing age. On account of the exceedingly gentle concave curves due to deposition, it is impossible to represent them on the scale here used. When the profiles of oppositely flowing streams meet at the crest of a mountain range, there should be, if no modify- ing conditions intervene, a sharp divide, as is indicated in Fig. 9. On some mountain crests this condition is very nearly reached. As one follows up a stream and approaches Fig. 9. Ideal Profile of a Divide between the Head-W^aters of Two Opposite- Flowing Streams : Vertical Scale Exaggerated. its ultimate source, the rate of corrasion progressively di- minishes, for the reason that the water supply becomes smaller and smaller. The rocks, however, are everywhere exposed to the denuding agencies of the air, namely, rain, wind, frost, etc., and at the heads of drainage lines the action of these agencies is in excess of stream corrasion, and convex curves due to weathering modify or replace the con- cave curves due to stream action. Unless the rocks on a divide between two drainage systems which head against each other are unusually resistant, and maintain angular forms as they weather, the concave profile leading up to the 150 RIVERS OF NORTH AMERICA divide from either side changes to convex curves before uniting.* The usual profile in such instances is shown in the lower curves in Fig. 8. Cross-Profiles. — The cross-profiles of stream-cut valleys change in the same locality with the age of the stream, and are modified by the weathering of the valley sides, the texture of the rocks, etc. If, as above, we conceive of a valley being cut out of homogeneous rocks and ascertain what changes in its cross-profile at a given locality will result from stream action and weathering, the modifications due to other causes may be more easity recognised. A young and rapidly corrading stream working in moder- ately hard rocks produces a gorge or canyon with steep sides. The cross-profile of such a gorge is markedly V- shaped, except that the bottom of the V is slightly rounded. The width of the concave bottom of the trench varies with the size of the stream. If the stream is work- ing in hard rocks the sides of the trench cut by it rnay be vertical. As a stream advances with its task of cutting down its channel to baselevel, its energy available for corrasion is more largely exerted in the direction of broadening its val- ley. The cross-profiles of the valleys of old streams be- come broadly U-shaped. The valleys of streams where an approximation to baselevel has been reached, or when flood- plains are being formed, generally have flat bottoms with more or less flaring sides. The cross-profiles then resemble more or less closely the figure which would be obtained by ' L. E. Hicks, ** Some Elements of Land Sculpture," in Bulletin of the Geo- logical Society of A vi erica, vol. iv., pp. 133-146, 1893. STREAM DEPOSITS I5I breaking a plate straight across. That is, the bottom is a horizontal line bordered by ascending lines. The graceful double curves in the profile on each side of an aggrading stream have already been referred to. As a stream advances in age, the cross-profile at a given locality gradually changes from a V-shape to a U-shape, and then to a -^ — /-shape. In extreme old age the bottom becomes greatly broadened with reference to the height of the sides. The slope of the sides of a valley, whatever its age, de- pends on the texture of the rocks and on weathering. In hard rocks the slopes are steeper than in soft rocks. If the rate at which a stream deepens or widens its valley is rapid in reference to the rate of weathering, the sides will be steep, but if the reverse is the case the slopes will be gentle. It is thus evident that the character of the cross-profile of a stream-cut valley depends largely on climate, on rock text- ure and rock structure, on relative rate of corrasion and weathering, and on the stage in development that the stream has reached. CHAPTER VI STREAM TERRACES A TERRACE may be defined as a step-like area with a nearly even and approximately level surface, bounded on one margin by an ascending and on the other by a descending slope. A stairway may be considered as an example of a series of terraces bounded by vertical escarp- ments. In nature there are many departures from the reg- ularity in form implied in the above statements, due in part to the conditions under which they were made, but more commonly to subsequent changes. The surface of a terrace is frequently uneven, and cut across by rill-channels and gullies, or talus slopes and landslides may encumber it. The bounding slopes may be steep, or depart but slightly from the horizontal. A cross-profile of a river valley with terraces on each side is shown in the following diagram. Fig. io. Ideal Cross-Profile of a Terraced Valley. This figure is intended simply to illustrate the general char- acteristics of stream terraces, and not to indicate the precise conditions which the student may expect to find when he supplements his reading by cross-country tramps. 152 STREAM TERRACES I 53 Terraces of this general character, from a few feet to several rods broad, may frequently be traced for many miles on each border of a river valley. In numerous instances several terraces one above another with various intervals be- tween may be recognised on the same slope. They follow all of the windings of the valleys, sweeping about prominent bluffs and into adjacent embayments in broad, beautiful curves. Much of the charm alike of sheltered dells and of broad valleys is frequently due to the symmetrically curving lines formed by the terraces on the bordering slopes of the adjacent uplands. This is true more especially of the valleys of the Northern Appalachians and New England and thence westward through the vast areas formerly occupied by glaciers. Many of the valleys in the mountains of the Cordilleran region are also terraced in a remarkable manner. River valleys, as we know, have been excavated by the streams flowing through them, and it is at once evident that the terraces beautifying their sloping sides must, in most instances, be due to the same agency. Another obser- vation confirming this conclusion, is that the terraces are not horizontal when followed in the direction of their lengths, but have a gradient similar to that of the stream flowing through the bottom of the valley in which they occur, but not precisely coinciding with it. The fact that stream terraces are not horizontal in the direction of their lengths serves to distinguish them from similar topographic forms made by the waves and currents of lakes or of the ocean. The surfaces of standing water- bodies are horizontal in the every-day sense of the term, 154 RIVERS OF NORTH AMERICA and the terraces made by such water bodies on the land confining them are also horizontal. The presence of terraces on the borders of stream-cut valleys suggests that they owe their origin to the processes of corrasion or of deposition which characterise the work of streams. The study of the topographic forms under con- sideration has shown that they may be due to either of these processes, or to their combined action. Certain stream ter- races have been formed by excavation, others are the result of deposition, while still others owe their existence to a combination of the two processes. We might classify them as cut terraces, built terraces, and cut-and-built, or com- pound, terraces. Such a classification has but little signifi- cance, however, unless the relation of the terraces to the life histories of the streams which gave them origin is under- stood. When the life histories of streams are reviewed, and the modifications in their normal development due to climatic changes and to secular movements in the earth's crust are considered, it will be found that there are three principal causes which lead to the origin of terraces. These are : ist. the normal changes in a stream valley due to the successive processes of corrasion, flood-plain building, and re-excava- tion ; 2d. climatic changes which cause variations in the volumes of streams or lead to excessive deposition for a time, and the re-excavation of the partially filled valleys; and 3d. oscillations in the land which vary the rate of cor- rasion and of deposition. Let us consider these three methods in the order named. Origin of Terraces during the Process of Normal Stream Plate VIM. Fig. a. Terraces on Fraser River, British Columbia. Showing post-glacial re-excavation, (Photograph by Geological Survey of Canada,) Pig. B. Terraces in Connecticut Valley, near Bellows Falls, Vermont. (Photograph by C. H, Hitchcock.; STREAM TERRACES I 55 Development, — In discussing the combined process of stream corrasion and deposition, when not seriously modified by climatic changes or movements in the crust of the earth, it was found that a river in flowing from highlands to the sea first cuts down its channel to baselevel at its mouth and then lowers it progressively up stream, but during its early life makes only an approximate adjustment to the level of the receiving water-body. Succeeding the first stage of excavation and following it progressively up stream, the valley is aggraded. This combined pro- cess is checked when the head branches of the river no longer supply more debris than the trunk stream can carry away, or, less commonly, when the course of the river is lengthened by the formation of a delta. The stream then begins to excavate a channel through the flood-plain pre- viously formed. When this process of re-excavation begins the stream is usually meandering in broad curves over a flood-plain. As the stream deepens its channel and sinks below the level of the flood-plain, it retains its windings; although the accelerated velocity of the stream may tend appreciably to straighten its course. When the stream lowers its channel, portions of the original flood-plain are left as terraces on the sides of the valley. At the time a stream begins to deepen its channel, it may, in one portion of its course, be in the centre of its flood-plain, and will then leave a terrace on each side, or may flow on one side or the other of its valley, and therefore leave a terrace only on one border of its course. The stream may then broaden its channel, and spread out a second flood-plain in the valley excavated through the previously formed deposit. 156 RIVERS OF NORTH AMERICA A stream in flowing down a flood-plain, it will be remem- bered, makes not only short bends, but broad sweeps which carry it from one side of its valley to the other. The short bends are made during periods of time measured by tens or hundreds of years, while the great migrations from side to side of a broad valley require thousands of years to com- plete a single swing. The short bends which combine to make much greater curves have been referred to in the case of the Mississippi, and may be readily recognised on any good map of that river. While a stream is deepening its channel in a broad alluvial plain and building a second flood-plain at a lower level, the down-cutting, between the time it leaves one border of its valley, migrates to the other side, and returns, may be so great that on its return it will be flowing at a sufficiently lower level to prevent its re- flooding its previously formed flood-plain. When this hap- pens, and the stream in its migrations does not swing back to its previous position, a portion of the flood-plain is left and forms a terrace. Successive terraces may be left at lower and lower levels by a continuance of this process. A cross-section of a valley terraced in the manner just de- scribed would present the features shown in the following diagram. Each terrace is a portion of a flood-plain deposit^ Fig. II. Ideal Cross-Section of a Partially Filled Valley with Terraces Left during Re-Excavation. and the highest in the series is the oldest. The material forming the superficial portion of the second terrace from STREAM TERRACES 1 57 the top has been removed by the stream, and re-deposited as a portion of the second-formed flood-plain, and this process has been repeated also in the case of the third terrace. In the normal development of a stream after the stage in a certain portion of its course, indicated in Fig. 11, is reached, the stream will continue to deepen its channel, and may cut into the rock below the flood-plain deposits. This stage in the process is illustrated by the cross-section shown below. Should subaerial erosion remove the alluvial ma- terial indicated by dots in the diagram, a rock terrace would be left. If stream development progresses and a second Fig. 12. Ideal Cross-Profile of a Partially Alluvial- Filled Valley Re-Excavated to below its Original Depth. approximation to baselevel is made, all of the alluvial ma- terial and a portion of the rocky floor on which it rests may be removed. Other ways in which normal alluvial terraces might be formed have been cited by Dodge.* Suppose that a stream whose load is slightly in excess of its carrying power ac- quires by capture the head-waters of another stream, as will be considered later. In the district thus acquired there might be an excess of carrying power over load ; if such was the case, the capturing stream would have its carrying power increased without a corresponding increase in load, and ' R. E. Dodge, *' The Geographical Development of Alluvial Terraces," in Boston Society of Natural History, Proceedings, vol. xxvi., p. 263, 1S94. 158 RIVERS OF NORTH AMERICA therefore be able to deepen its channel in previously de- posited alluvium, and terrace it. Again, as cited by Dodge, a stream which had been work- ing in soft rocks might cut down into hard rocks underneath the soft ones. The effect of such a change on the head- waters of a stream would be to decrease its load and enable it to corrade in its alluvial tract. Hence, without varying in volume a stream might be able to terrace an alluvial plain formed while it was previously removing soft rock. Thus in several ways, or as a result of the combined in- fluence of two or three normal variations in streams, alluvial terraces might result. These processes of terrace-making, however, are slow, and the topographic forms resulting may be greatly modified or even obliterated by subaerial denuda- tion as fast as they appear. These processes, also, are a part of a larger process, i. e., cutting to baselevel, which in- sures the ultimate destruction of the topographic features referred to. For these reasons the methods of terrace- making just considered have received but little attention, and their results are difficult to recognise. And, besides, other methods of terrace-forming are apt to produce such conspicuous results that the terraces due to what has been termed the normal stream development are usually masked or obliterated. Terraces Due to Climatic Changes. — In considering the various influences of changes of climate on stream deposi- tion, it was shown that heavy rains may cause the tributaries of a stream to bring to the main channel more debris than can be removed, and deposition takes place. In a similar way a secular change of climate producing an increase in STREAM TERRACES I 59 precipitation, might lead to the filling, especially of low- grade river-valleys, and the raising of the flood-plains throughout all of their lower courses. A climatic change which would admit of the birth and growth of glaciers on the higher portions of a mountain range, previously deeply stream-sculptured, would lead to the overloading of the streams below the glaciers and the thickening and broadening of the flood-plains throughout their lower courses. Climatic conditions favourable for the birth and growth of glaciers are usually, and probably always, accompanied by increased precipitation and decreased evaporation. Thus for several reasons the occurrence of a glacial epoch like that in late geological time, when one-half of North America was occupied by ice-sheets, would favour the filling of pre- existing valleys with debris. When the climate experienced a reverse change and the glaciers melted, the draining streams would for a time be still more deeply flooded, and additional quantities of debris carried from high to low regions. If a warmer and drier climate should succeed a glacial epoch, the streams, no longer heavily loaded, would begin the task of removing the debris deposited in their valleys during the preceding time of overloading. As the streams deepened their channels in the alluvium previously deposited, portions of the flood-plains left intact would ap- pear as terraces, and the elevation of their surfaces would record the depth to which the valleys had been filled with debris. This process of removing the accumulations of debris clogging a valley might be accompanied by the formation of l6o RIVERS OF NORTH AMERICA terraces at lower levels, according to the laws, cited above, governing the normal development of streams. As will be shown later, however, a still more potent agency in the formation of the lower terraces would be climatic changes and periodic elevation of the land. Should several glacial stages occur with intervals of milder and less humid climatic conditions intervening, it is evident that the terraces resulting from the trenching of the first- formed flood-plains might be obliterated by subsequent de- position, and the surface of the debris in the valleys be carried higher than during the first ice invasion ; or the valleys, cut in the first-formed flood-plain, might be only partially re- filled, and when excavation was renewed, lower terraces would be formed. The conclusion that glacial conditions would lead to the filling of pre-existing valleys downstream from alpine glaciers or about the margins of piedmonts and continental ice-sheets, and portions of these deposits be left as terraces when corrasion was resumed, is sustained by an abundance of examples throughout the northern portion of the United States and Canada. In the valleys in this region terraces excavated in m^aterial deposits by glacial streams are mag- nificently displayed. On the head-waters of Columbia River in Washington and Idaho, terraces of the nature here con- sidered are perhaps as well developed, and their history as easily read, as in any other portion of the continent. The great canyon of Snake River, the principal tributary of the Columbia, was excavated to its present depth, — four thou- sand feet throughout a considerable portion of its course, — previous to the Glacial epoch. During that epoch, glaciers STREAM TERRACES l6l existed in the more elevated valleys and about the sum- mits of the mountains of Idaho. The branches of the Snake were flooded, and brought such quantities of debris to the canyon of the main stream that throughout hun- dreds of miles of its course it became filled to a depth of three hundred and sixty feet. When the glaciers passed away and the streams were no longer supplied with debris by them, and still more effectually when a mild and but moderately humid climate prevailed, the streams were -enabled to attack their flood-plains and cut valleys and canyons through them. Snake River has now removed by far the greater portion of the coarse gravel and boulders that formerly occupied its canyon, and has resumed the task of deepening its channel in the hard rock beneath. Episodes similar to that just referred to in the history of Snake River, but with various minor modifications, occurred in the lives of tens of thousands of streams not only in the northern part of North America, but as far south as the Gulf of Mexico and also in the Rocky Mountains and Sierra Nevada, as a result of the climatic change to which the Glacial epoch was due. In studying the effects of changes in climate on the be- haviour of streams, the fact should be borne in mind that such changes, although by reason of the comparatively brief time during which man has taken account of secular vari- ations in atmospheric phenomena they are commonly considered as exceedingly slow in their occurrence and em- bracing but a moderate range, appear relatively rapid and of well-marked amplitude when such periods of time as are involved in geographic cycles are studied. Many of our 1 62 RIVERS OF NORTH AMERICA rivers, as, for example, the Susquehanna, Mississippi, and Columbia, were far advanced in their development before the beginning of the Glacial epoch. The time that has elapsed since the melting of the continental glaciers on the head-waters of these rivers is but a small fraction of the cur- rent geographic cycles. Many annual climatic changes, as is well known to everyone, occur while even a meadow brook undergoes but slight modifications ; in a similar way^ as is well known to geographers, many secular changes in cli- matic conditions may take place during the life history of a great river. Terraces Due to Elevation of the Land. — The manner in which a stream carries on its work, it will be remembered^ is controlled in an important way by declivity. Conse- quently, changes in the elevation of the land must have a direct bearing on the history of the streams draining an area thus affected. The movements in the earth's crust referred to are known to have modified the surface slopes throughout large areas, and frequently to be of the nature of a tilting of the land. Other movements occur, but at present let us consider simply the effects of the tilting of a region drained by a large river on the problem of terrace- making. A tilting of the rocks which decreases the gradient, and consequently the velocity, of a stream, other conditions re- maining the same, will favour deposition, and may lead to the partial or complete filling of its previously formed valley. The flood-plain deposits would then increase in thickness and become broader at the surface. In other words, a de- crease in velocity favours the process of aggrading. STREAM TERRACES 163 If the region drained by the Connecticut, for example, be considered as a plane gently inclined southward, and to be affected by a movement in the earth's crust which de- creases the gradient of the river, the depression of the land being least at the south and progressively increasing north- ward, — that is, the hinge-line, so to speak, on which the tilted block of the earth's crust moves, being situated near its southern margin, — the main trunk of the river would have its current slackened, and its transporting power diminished, while the gradients of the branches of the river coming in from the east or west would be but slightly affected. The direct result of such a change would be to favour deposition in the main valley and to a less extent in its branches. If, after such a change of grade as has been postulated, when the valley of the Connecticut has become deeply filled and a broad flood-plain spread out, we imagine the land to remain stationary, the branches of the main river would cut down their channels, thus decreasing their velocities and diminishing the amount of debris carried annually to the main valley. When this stage had been reached, the Con- necticut would begin to cut a channel through its previously formed flood-plain, as in the case of normal stream develop- ment already considered. In case the inclined plane drained by the Connecti- cut should experience a reverse movement after its valley had become deeply filled — that is, if elevation should occur, the hinge-line retaining its former position, — the gradient of the main stream and of all its branches flowing southward would be increased, while the lateral branches would be but little affected. The increased gradient of the 164 RIVERS OF NORTH AMERICA main stream would give its waters greater velocity, thus favouring corrasion at the expense of deposition, and a channel would be cut through the previously formed flood- plain. Portions of the flood-plain not removed would re- main as terraces. Imagine the re-elevation at the northern border of the tilted area to be one hundred feet, and to de- crease to zero at the hinge-line at the south. The result would be acceleration of velocity in the extreme head branches flowing southward ; this might cause them to bring more debris to the main stream than it could transport, but the branches from the east and west being but slightly affected, the more probable result would be the deepening of the bed of the main stream throughout. The river would excavate a channel through its previously formed flood-plain, leaving portions of it on either side of the valley as terraces. When the river, after adjusting itself to the new conditions, began to broaden its channel and spread out a second flood-plain, it would be flowing a hundred feet below its former bed- in the upper portion of its course, but this difference would gradually decrease downstream and become zero where the hinge-line was crossed. The south- ward or down-stream slope of the surface of the old flood- plain, portions of which remain for a time as terraces, would therefore be greater than the slope or gradient of the readjusted stream. This postulated case thus furnishes an explanation of the fact that when a number of stream ter- races occur on the border of a valley, they are not only not horizontal, but have different gradients. The gradient of every stream terrace is determined by the gradient of the parent stream at the time it was formed. STREAM TERRACES 165 The terraces originating in the several ways thus far con- sidered consist of alluvium, which was deposited in a pre- viously formed river valley, and the surface of each terrace is a portion of a flood-plain. In cross-section, such terraces would have the characteristics shown by the diagram on page 156, introduced in connection with the discussion of what are termed normal terraces, and would be cut through or finally removed during subsequent stream development in the manner already described, unless subsidence carried them below baselevel. In the discussion just presented, we have assumed a river valley to have been deeply filled with alluvium previous to the elevation of the land which enabled the stream to deepen its channel. This assumption is not Fig. 13. Ideal Cross-Section of a Valley with Terraces Cut in Solid Rock and Covered with Alluvium. necessary, however, and numerous instances might be cited where terraces in solid rock have resulted from accelerated corrasion due to periodic uplifts. Imagine a stream like the Connecticut to have broadened its valley and spread out a flood-plain, and then an elevation to take place as before. Accelerated velocity may enable the stream to lower its bed so as to cut through the flood-plain deposits and into the rocks beneath. A broadening of the new channel may then occur, and renewed elevation allow the process to be re- peated. With each upheaval the stream cuts deeper into the rocks, leaving each time a terrace of solid rock with a sheet of alluvium on its surface. The characteristic features 1 66 RIVERS OF NORTH AMERICA of a cross-section of such a terraced valley are shown in the ideal diagram, Fig. 13. As the excavation of solid rock is normally a slow process, the sheet of alluvium covering the terraces would be apt to be removed by rain, rills, etc., and rock terraces but scantily covered or without debris be exposed. The formation of terraces during what has been termed the normal development of a stream — that is, when changes of level have not occurred, and climatic variations, etc., have not materially affected its volume, velocity, or load — is an extremely slow process, and, as previously stated, it is prob- able that atmospheric agencies under most climatic condi- tions would destroy the terraces as fast as formed. For this and other reasons it is believed that most of the terraces on the borders of stream-cut valleys are records of climatic changes which caused excessive deposition in low-grade valleys followed by a period of erosion ; or are due to land oscillation. Bottom Terraces, — Still another variety of terraces is formed by streams by deposition when their bottom loads are small. When their bottom currents are underloaded, as we may term the condition here referred to, the material is carried forward like a wave, in the manner in which a ripple in sand is produced under the influence of a wind- or water- current, and deposited with a steep escarpment facing down- stream. These bottom terraces have broad, gently ascending surfaces in the direction of the flow of the current, and steep escarpments, facing downstream, and trend in general at right angles to the flow of the water, but are usually lobed on their lower margins. Such terraces or broad STREAM TERRACES 1 67 ripples may be seen in process of growth in many clear streams which have moderate bottom-loads of coarse sand and gravel. They are frequently several feet or even yards broad, with escarpments from a few inches to a few feet high. Although of minor importance when considered in connection with associated stream-made topographic forms, yet under special conditions, as when a broad stream is mov- ing debris over a gentle slope, they might become relatively conspicuous if the stream should be diverted. There is a gradation between bottom terraces of the nature just con- sidered and delta terraces which would repay investigation. Delta Terraces and Current Terraces, — In the discussion of deltas in a previous chapter, it was shown that they are formed where streams deposit their loads on entering still water. Now, streams sometimes expand and have sluggish currents so as to simulate lakes. When this happens a tributary stream freighted with debris may drop a por- tion of its load and build up a delta-like deposit at its mouth. The most favourable conditions for this process are when low-grade, sluggish rivers extend into embayments on their borders, as the mouths of tributary valleys, and a stream from the tributary valley brings in sediment. A lowering of the main stream after such a delta has been formed would leave it as a terrace. Such structures have been termed delta terraces by Edward Hitchcock.^ A section of a delta terrace would reveal a series of inclined beds, as shown in the diagram on page 126, and possibly the upper ^ " Illustrations of the Earth's Surface," Smithsonian Contributions to Knoiv- ledge, vol. ix., pp. 32-34, 1857. 1 68 RIVERS OF NORTH AMERICA and lower members of a typical delta built by a high-grade stream as well. The surface of such a delta terrace would have a slope corresponding with the grade of the supplying; stream. The current of a river washes its banks in much the same way as the currents in lakes wash their shores. The study of the action of lake currents has shown that they bear along debris, and drop it in part so as to form what are known as built terraces. The current, especially of a broad river, behaves in much the same manner. Debris brought by tributary streams, or derived from localities where the river is corrading its banks, is carried down stream and may be deposited adjacent to the shore, so as to form a built terrace. A subsidence of the waters would leave the terrace exposed. Its surface would slope gently toward the stream, and, as in the case of all river terraces, would have a gradient, when followed along the valley on the side of which it was formed, corresponding with the surface gradient of the building stream. In cross-section such a terrace would reveal the structure characteristic of built lake- terraces, the general features of which are shown in the following ideal diagram. The slope rising above such a Fig. 14. Ideal Cross-Section of a Current-Built Terrace. terrace may be the valley side, produced by stream corrasion and weathering, or be a steeper slope due to lateral corra- STREAM TERRACES 1 69 sion of the current and correspond more nearly with the '* sea-cHff " above a lake terrace. The waves and currents of a broad river may lead to cor- rasion along its shores in the same manner as in a lake, and cut terraces result. A miniature example of this is shown in Fig. F, Plate II. So far as the present knowledge of stream terraces allows one to judge, it does not appear that those built after the manner of lake terraces, as just described, are common. In fact, delta terraces and current terraces, as they may be termed, depend for their origin on a delicate balancing of conditions which apparently is seldom reached. Delta terraces and current terraces formed on the sides of streams are of interest, as they constitute a group, although small and of minor importance, which may be designated as built terraces in distinction from other stream terraces which are due to both deposition and excavation, or to excavation alone. The downward slope bordering the nearly flat sur- face of a built terrace is due to deposition ; in the other varieties, this slope is produced by excavation. Glacial Terraces, — The terraces built by streams con- jointly with glaciers need not claim much attention at this time, since they derive their greatest interest from their connection with the ice-bodies about which they are formed. When a glacier, however, or perhaps more frequently a stag- nant ice-mass, occupies a valley, streams sometimes bring gravel, sand, etc., and deposit it along the margin of the ice so as to give a level floor to the space intervening between the ice and the valley border. After this space has been filled to a greater or less depth and the ice melts, the deposit re- IJO RIVERS OF NORTH AMERICA mains as a terrace.' Such terraces have approximately level surfaces, are composed of current-bedded gravel and sand, and perhaps certain occasional boulders or angular rock- masses, but do not exhibit the arrangement of coarse and fine material characteristic of flood-plains; and, besides, their down-stream gradients are markedly different from those of true stream-terraces. Relative Age of Terraces, — When stream terraces occur one above another on the side of a valley, the highest in the series is usually the oldest. But exceptions to this rule may occur, as when changes of level lead to the build- ing of delta or current terraces on the surface of previously formed terraces. Again, a valley in which a terrace has been cut in solid rock might become filled with alluvium so as to bury the terrace, and re-excavation again bring it to light, and form another terrace at a higher level; the lower terrace would then be older than the one above it. In a series of alluvial terraces in which the highest is the oldest, each one or each pair, if fragments of the same flood- plain are left on each side of the valley, is a remnant of a flood-plain, and the material in the highest terrace is younger than the main portion of each lower terrace ; but the surface portion of each terrace was worked over and re- distributed at the time the flood-plain of which it is a part was formed, and hence may be said to be younger than the material in each higher terrace. ' The terraces here referred to have been termed " kame terraces" by R. D. Salisbury, Geological Survey of New Jersey, Annual Report for i8gj, pp. I55» 156. Similar topographic forms were previously termed "moraine ter- races" by G. K. Gilbert, U. S. Geological Survey, MonograpJis, vol. i., p. 8l, 1890. STREAM TERRACES I7I Other Terraces, — Terraces similar to those formed by streams originate in other ways, and it is important that the student of geography and geology should be able to dis- tinguish those which owe their origin to one series of agencies from those belonging to other categories. Cut terraces in rock or in loose material are a characteris- tic feature of lake and ocean shores, as are also delta and current terraces. In nearly all of their main features, these terraces are similar to stream terraces except that they are essentially horizontal when traced in the direction of their length. Movements in the earth's crust, however, may tilt a previously horizontal terrace so as to give it a gradient closely approximating to the normal slope of a stream terrace. A similar tilting of the land might affect a river terrace so as to alter its gradient and perhaps make it hori- zontal. In cases of this sort associated topographic features would usually furnish the best clue to the true history. River terraces are formed in comparatively narrow valleys, while lakes may occupy valleys of any shape. Lake ter- races are usually accompanied by escarpments which rise above them, termed *' sea-cliffs " ; these may or may not be characteristically different from the corresponding slopes above river terraces. The normal lake-terrace is either cut- and-built — that is, it is a shelf made by excavation with a current-built covering on its surface, — or may be entirely a deposit formed by construction. Stream terraces do not usually have this structure, but yet may have it. As may be judged, the tests just suggested might not lead to definite conclusions. In fact, in the case of abandoned stream- and lake-terraces, that is, when a former lake basin has been 1/2 RIVERS OF NORTH AMERICA emptied and perhaps has a stream flowing through it, or when a former stream valley is no longer a line of drainage, it is frequently difficult to satisfactorily determine their origin. In such an instance, if the former topography is not greatly altered, it may be possible to work out the history of the changes that have occurred and to construct a map of the country as it existed when the terraces were formed, and thus be able to decide whether flowing water or bodies of still water were responsible for the terraces. River terraces frequently make a direct connection with lake terraces so that the place of junction may be difficult to determine. The main difference to be looked for in such an instance would be a change from horizontality to an in- clination in the surfaces of the terraces where it passed from the lake valley into the stream valley. In a tilted or other- wise disturbed region, where both lake and stream terraces occur, a difference in their gradients may still be recognis- able, and assist in their discrimination. Instances occur, however, in disturbed and eroded re- gions, where only fragments of terraces remain, when it is practically impossible to tell whether they record lacustral or stream conditions. River terraces also pass into terraces made about the borders of estuaries and on ocean shores. Here, again, when disturbances occur and the estuaries are emptied of their water and the streams have been diverted, difficulties in interpreting the records might arise. The sedimentary deposits made on the floor of the estuaries and the evidences of life buried in them, as well as the fossils in the terraces themselves, might here furnish assistance. The shells in river terraces will be fresh-water or land species; STREAM TERRACES 1 73 while those in the estuary sediment and terraces will be, in part at least, such as inhabit brackish or saline water. Terraces also result from the weathering of the outcrops of alternating hard and soft strata. When the strata are horizontal or but slightly inclined, the hard beds may stand out as shelves or terraces, as, for example, in the sides of a valley cut in stratified rocks. The downward slope of such a terrace is the exposed edges of hard layers, and may be steep or gentle according to climatic and other conditions; the slope rising above the terrace is formed of the edge of the weak strata above the terrace-making layer, and is usually a gentle slope unless the layer is thin and another hard terrace-making layer occurs just above. The arrange- ment of resistant and weak strata in the case of these terraces of differential erosion usually makes it easy to dis- tinguish them from river terraces. There is an absence, also, on the terraces of this nature, of stream deposits, and this negative evidence might assist in the diagnosis. Weathered debris falling on the surface of a terrace of differential erosion may simulate stream deposits, however^ and lead to erroneous conclusions. Fractures in the rocks along which differential movement of the sides has taken place, producing what are known as faults, may also give origin to topographic forms of a terrace- like character, but these are usually irregular, with reference to both horizontal and vertical planes, and in most instances are easily distinguishable from stream or other terraces. Landslides also produce terrace-like forms, but these are seldom continuous for considerable distances, and are usually so irregular and bear such relations to the slopes from which 1/4 RIVERS OF NORTH AMERICA the fallen blocks descended, that their origin may usually be readily determined. When a landslide occurs it fre- quently happens that the displaced material acquires a sur- face slope toward the place from which it came. This backward slope frequently produces basins in which lakes and swamps occur, thus furnishing additional evidences bearing on the origin of the terrace-like forms produced.* Terraces due to still other causes might be enumerated and the means for their discrimination indicated, but I be- lieve those most nearly simulating stream terraces have been referred to. The reader who may desire to follow this sub- ject farther will find assistance in the treatises named below.* General Distribution of Stream Terraces, — In North America stream terraces may be said to occur on the borders of nearly every river valley north of the central part of the United States, but are less conspicuous in the more south-eastern States and about the Gulf of Mexico: the reason being that the northern half of the continent was occupied by glaciers in late geological time, and has also undergone movements of the nature of elevation and depression throughout broad areas; while in the southern half of the continent there are no records of glaciation ex- cept on high mountains in the south-west, and evidences of recent changes of level are seldom pronounced. Stream terraces extend far southward from the formerly glaciated region, for the reason that the glaciers drained by ^ I. C. Russell, " Topographic Changes Due to Landslides," Popular Science Monthly, vol. liii., i8g8, in press. ^ G. K. Gilbert, U. S. Geoh^ical Survey, Monographs, vol. i., pp. 78-86, i8qo. W J McGee, CJ. S. Geological Survey, nth Annual Report, part i., pp. 256- 273, 18S9-90. STREAM TERRACES 1 75 southward-flowing streams furnished more debris than the streams could remove, and they became overloaded and con- sequently filled in their previously excavated valleys. When the glaciers disappeared and the streams were no longer over- loaded, they cut channels through their previously formed flood-plains, and left portions of them as terraces on their sides. This process was aided also by a general depression at the north, due to some extent, it is believed, to the weight of the ice, and in part to the effect of the lowering of the temperature to a considerable depth in the earth's crust and a consequent contraction and depression of the surface, and a partial re-elevation when the glaciers vanished. Many of the beautiful river-valleys of New England owe much of their attractiveness to the gracefully bending curves traced on their borders. Numerous towns and villages in that region are indebted for their sightly locations to the terraces on which they are built. Present flood-plains and abandoned portions of former flood-plains afford rich agricul- tural lands. These were among the first areas to be cleared and cultivated after European immigration began. A direct relation between the effects of distant and far-reach- ing changes in geography and the advance and growth of civilisation is here abundantly illustrated. What has been said of the terraced valleys of New Eng- land is true in varying degrees of a great area to the north embraced in the south-eastern provinces of Canada, and of a still broader region to the west including New York, Penn- sylvania, Ohio, and thence north-westward to the Pacific. The valleys of the Ohio and of many of its tributaries are noted for their terraces. In this region and also in the val- 1/6 RIVERS OF NORTH AMERICA leys tributary to the upper Mississippi, the numerous terraces are due principally to the re-excavation of pre-glacial val- leys, in which overloaded glacial streams dropped the burdens too heavy for them to carry. The effects of the changes in drainage accompanying the great ice invasion at the north may be traced throughout the length of the Mis- sissippi, but become less and less conspicuous towards its mouth. The borders of the streams flowing to the Gulf of Mexico (other than the Mississippi), including the larger rivers of the Texas region and those draining the eastern slope of the Appalachians south of Maryland, are mostly without con- spicuous terraces. The streams in this region did not feel the direct influence of the glaciers at the north and have passed their period of youth, except on their extreme head- waters. The terraces they may have formed as a part of their normal aggrading and re-excavation have been re- moved principally by weathering, and the comparatively gentle slopes of their valleys show vertical scorings due to the action of rills and not the nearly horizontal lines which record higher water stages. Such terraces as do occur along the sides of southern rivers are in part remnants of ancient baselevel plains, or, in some instances, the result of local changes in elevation. These statements are inten- tionally made general, as it is only the more marked differ- ences between the characteristics of southern and northern valleys to which attention is here sought to be directed. The rivers flowing eastward from the Rocky Mountains, like the Platte, Missouri, Arkansas, etc., are bordered by terraced slopes in a portion of their valley tracts, and for STREAM TERRACES I 77 many miles in the plains tracts after leaving the mountains, but well out on the Great Plains they are in part, and per- haps mostly, engaged at the present time in aggrading previously formed valleys, and terraces are not conspicuous. One reason for the presence of conspicuous terraces adja- cent to the mountains and in the valleys eroded in them previous to the great extension of the glaciers, is that the streams dropped the coarser and heavier portions of their loads at those localities, and bore on only fine material to their low-grade plains tracts. Subsequently, when the flood- plains thus formed were terraced, the escarpments in coarse and in part cemented gravel and boulders retained their slopes for a longer period than the similar escarpments in fine material a hundred miles or more downstream. In addition to this, it is to be noted that the gradients of the highest terraces are greater than the gradients of the streams as they flow at the present day. For this reason the verti- cal interval between the terraces and the present streams which they follow — that is, the height of the downward- sloping escarpments bordering them on their valley margins — progressively decreases from the gateways in the mountain to the sea. The lower terraces composed of soft material far out on the plains have melted down and been removed to a great extent, under the action of the atmosphere, while the higher terraces in firmer material have retained their characteristic topographic forms. In the terraces of the Rocky Mountains and adjacent portions of the Great Plains, the wide-reaching influence of the Glacial epoch is again recorded. The Rocky Mountains had their local glaciers at that time, and the streams were 178 RIVERS OF NORTH AMERICA flooded especially during the final nielting of the ice, when^ it is inferred, the rain-fall was also abundant, but the swol- len streams were overloaded and their channels became deeply filled. The effects of more or less periodic changes in the elevation of broad areas above the sea may perhaps be made out in this region from the terrace records, but as yet too little attention has been given to this subject, and in fact to the general surface features of the Rocky Mountain region, to warrant one in offering anything more than pro- visional answers to the questions here suggested. In the Great Basin region variations of climate are again recorded by the work of the streams. Changes in the rate at which streams deposit, it will be remembered, depend on changes in velocity and in load. Velocity depends in part on volume. A change in climatic conditions from arid to humid means an increase in the volume of the draining streams. Whether this increment in velocity will be ac- companied by deeper corrasion or by aggrading, is deter- mined by the accompanying change in the rate at which the streams are loaded. An increased rain-fall would be accompanied by a greater removal of loose material from the highlands, and consequently a greater contribution of debris to the stream. There would, therefore, be a bal- ancing of conditions in any one part of a stream's course. If the supply of debris was not in excess of the transporting power of the draining streams, they would corrade, but if the loads delivered to them were too great, deposition would result. When the whole extent of a river is con- sidered, a change from arid to humid conditions must increase both corrasion and deposition. More active cor- STREAM TERRACES 1 79 rasion in the upper portion of a drainage system usually necessitates a greater rate of deposition lower down. The streams of the Great Basin felt the effects of the change of climate which caused, or accompanied, the Gla- cial epoch, and, as the results indicate, were overloaded. Many of them are bordered by terraces in such a manner as to show that they were formerly of greater volume than at present, and subsequently decreased in volume, but were able for a time to cut channels and broaden their valleys in previously formed flood-plains. In many instances the streams have diminished to such an extent, however, that they are now aggrading. In their enfeebled condition they are unable to carry even the small burdens that are imposed upon them. In the region of the high plateaus drained by the Colo- rado and its branches, there are many terraces. The majority of those which attract the eye, however, are due to the weathering of the outcrop of alternating hard and soft strata, but stream terraces also occur. Many broad valleys are deeply filled and without terraces, as is illustrated by Fig. A, Plate XVI., owing to the fact that aggrading has been in progress throughout a large portion of the present arid period. The Colorado River, as is well known, flows through a canyon within a canyon. The outer canyon is in places some fifteen miles broad and comparatively level-floored. Sunken in this floor is the deeper, inner canyon, which in places is from one to two miles broad. The floor of the outer canyon is thus a great terrace, as may be seen on inspect- ing Plate XVII. The general history to be read in these l8o RIVERS OF NORTH AMERICA features is that the region was at one time lower than now by an amount about equal to the depth of the inner canyon. The Colorado cut down its channel to baselevel and then broadened it into a wide valley. Subsequent elevation renewed the energy of the stream and enabled it to cut down nearly to baselevel once more, leaving large portions of the bottom of its older valley as a terrace on each side of its later gorge. The terrace thus formed extends into many of the tributaries of the main river, and in fact the change to which it was due affected a large part if not the entire drainage system. Our knowledge of the terraces in the valleys of North America is too immature to permit us to state with confi- dence their distribution throughout the continent, and one more illustration must suffice for this brief review. Columbia River, in the portion of its course known as the Big Bend, flows through a narrow, canyon-like valley on the side of which there are numerous terraces. A short account of these taken from a report by the present writer,* will indicate not only the nature of the stream records so abundant on the sides of many of the canyons and valleys of the far Northwest, but also show how stream terraces are frequently associated with similar topographic forms originating in other ways: " On descending the side of the canyon [of the Columbia, op- posite the entrance of Chelan valley] by means of a road following a deep, high-grade gorge, we notice that there are many terraces on each side of the river. The most remarkable of these, and ' T. C. Russell, "A Geological Reconnoissance in Central Washington," U. S. Geological Survey, Bulletin No. io8, pp. 78, 79, 1893. STREAM TERRACES l8l one of the finest examples of terrace structure that can be found anywhere, is a level-topped shelf formed of gravel and water- worn boulders, the surface of which is seven hundred feet above the Columbia. This truly remarkable terrace is best developed about two miles below where we descended into the canyon. It is there several hundred feet broad and runs back into lateral gorges, showing that the sides of the main canyon were deeply scored by lateral drainage before the gravel forming the terrace was deposited. On the west side of the valley there are other fragments of the same deposits, forming a less conspicuous shelf, which has been built against the steep slope, and has the same level as the great terrace on the east side of the river. The valley was excavated lower than the bottom over which the Columbia now flows, and then filled in from side to side with stream-borne stones, gravel, and sand before the present channel was excavated. In the re-excavation, fragments of the deposits filling the canyon have been left clinging to its slopes. Streams flowing down lateral gorges have cut channels across the terrace, thus revealing the structure even more plainly than the steep slope leading to the river. " Above the valley opening in the west wall of the canyon and leading to Lake Chelan, there are other large remnants of the same great terrace, this time on the west side of the river. In the broad plain formed by the surface of the terrace, there stands a lofty pyramid of solid rock completely surrounded by the gravel deposit and rising like an island from its level surface. " The terrace gravels extend into the valley of Lake Chelan and form conspicuous terraces about its lower end. For many miles both up and down the Columbia, other fragments of the same level-topped deposit occur, always forming striking features in the landscape, owing to the marked contrast of their smooth horizontal lines with the vertical line due to the erosion of rills and creeks. " Beside the great terrace described above there are many other but less conspicuous horizontal lines on the sides of the Columbia canyon. Some of these below the horizon of the main 1 82 RIVERS OF NORTH AMERICA terrace are stream terraces, made by the river in lowering its bed. A more numerous but much less regular class are due to land- slides, of which there have been many. Other horizontal lines are due to the unequal weathering of the strata of basalt and of interstratified sedimentary beds. " Still another class of terraces, both numerous and conspicu- ous, has been formed as moraines on the sides of the glacier that once filled the canyon up to an elevation of 1200 feet above the river as it flows to-day. The moraine terraces are of older date than the great terrace described above, and about the en- trance of Chelan valley have been partially buried by it. '* In the canyon of the Columbia for several miles above Lake Chelan its rugged sides are strewn with thousands of perched boulders, left by the retreat of the ice. These have a definite upper limit, but mingled with them are masses of basalt that have fallen quite recently from the cliffs above. " In embayments along the sides of the main canyon and back of the ridges of stone and boulders left by the ancient glaciers, there are flat areas which have been filled in with fine material, washed from higher levels. These plains have in some instances been cut by small streams flowing across them, thus adding other horizontal lines to the complex topography of the canyon walls. '* It is not practicable to describe these terraces in detail, but those who visit Lake Chelan will have an opportunity to read for themselves the remarkable history which they record. In study- ing them, however, the traveller must bear in mind that the canyon, after being cut through various rocks to a depth greater than it now has, was occupied by a large glacier, and then by an arm of a large lake, and that river, glacial, and lacustral records are inscribed on the same slope. In addition, there have been many landslides, producing deceptive, terrace-like forms, and terraces due to the unequal weathering of hard and soft beds." A continuation of this review of the distribution of ter- races would lead us northward to the Mackenzie and STREAM TERRACES 1 83 Yukon/ where many records similar to those just con- sidered are known to exist, but our present knowledge of the changes these streams have experienced is even more scanty than for the central and southern portions of the continent. The most important result of this hasty excursion through terraced river-valleys is perhaps the recognition of the fact that terraces exist along the sides of stream-cut valleys throughout the length and breadth of North America. Volumes of history are recorded in those graceful curves which give beauty to the varied scenery of valley borders from the tropical forests of Central America and Mexico to beyond the Arctic circle. The interpretation of these records has only recently been undertaken, and much that is new unquestionably awaits the patient explorer. The general principles to be used in this study have been pre- sented in the present chapter, but as investigation pro- gresses, much that is novel in details, and probably also the discovery of as yet unknown laws or modifications of those now recognised, will reward the student. * I. C. Russell, •' Notes on the Surface Geology of Alaska," in Bulletin of the Geological Society of America, vol. i., pp. 144-146, 1890. CHAPTER VII STREAM DEVELOPMENT Consequent Streams, — In case a portion of the sea floor should be upraised so as to make what geographers term a new land-area, the streams flowing from it would take the easiest courses, as determmed by the slope of the surface,, regardless of the structure of the rocks beneath. If a dome-shaped uplift should occur in a broad plain underlain by previously horizontal layers of rock, the rain falling on its surface would form rills, and these uniting would give origin to creeks and perhaps to rivers, which would flow away in all directions from the summit portion of the uplift. In each of these hypothetical cases the streams would evi- dently have their directions determined by the pre-existing topography, and hence may be termed conseqiieiit streams. Subsequent Streams, — As consequent streams deepened their channels they might discover differences in the rate at which they can remove the rocks, and hence have their directions variously modified by differences in hardness or by the greater or less solubility of the various beds that they encountered. New branches or tributaries to the main streams would be developed, the positions of which would be determined by the down-cutting of the channels of the 184 STREAM DEVELOPMENT 1 8 5 master streams, and by the relative ease with which the various rocks coming to the surface could be removed. That is, as a drainage system develops, streams originate^ the directions of which are regulated by the hardness and solubility of the rocks. Such streams appear subsequently to the main topographic features in their environment, and are termed subsequent streams. Ideal Illustration of Stream Adjustment and Development. — Perhaps the best way in which to obtain a graphic idea of the changes passed through by a river system in the course of its development and adjustment to the geological conditions it discovers, when unaffected by marked climatic changes and not seriously disturbed by movements in the earth's crust, is to form a mental picture of a gently sloping plateau with an essentially even surface on which rain falls and gives origin to rills and brooks which unite to form larger consequent streams, and picture to ourselves the changes that would result under the influence of a moder- ately humid climate. Imagine such a plateau, we will say, one hundred miles long from right to left and fifty miles broad, sloping gently toward an assumed point of view, and follow in fancy the changes in the streams, and the result- ing modifications in the relief of the surface as normal stream development progresses. To complete our concep- tion, we may assume that, beyond the sky-line bounding the far side of the landscape, the surface gently declines in the opposite direction. That is, the plateau before us is the side of a long ridge, the central axis of which is raised, we will say, one thousand feet above sea-level. The rocks beneath the surface of the lilted plane, we will I 86 RIVERS OF NORTH AMERICA assume, are in inclined layers, which slope toward our point of view, at a greater angle than the surface of the plane, and besides are composed of hard and soft beds. A section through the tilted plane at right angles to the axis of the uplift has the structure indicated in the following diagram. Fig. 15. Section at Right Angles to the Lines A A and B B in Fig. 16. The harder rocks are shaded. The lines in which the inclined beds join the surface of the plane run in the direction of its length. The conditions here assumed are such as might result if a region underlain by inclined stratified rocks had been planed away nearly to baselevel and then upraised so as to produce a gently sloping peneplain. The rain-water, falling on the surface of the plane before us, gathers in part in depressions on its surface and forms ponds and lakes. These are but temporary, however, as the shallow basins are soon filled with sediment, or have their outlets cut down by the overflowing water, and are drained. The rills supplied directly from the rain and those starting from the lakelets unite to form larger streams, which flow down the inclined plane to the sea in obedience to gravity. The directions of these initial streams are deter- mined by the slope of the surface. They are, therefore, consequent streams. Their number will be determined by the inequalities of the surface which cause the rills and rivulets to unite. Some will be longer than others. Some will have a greater volume than others. A common feature, shared at first by all, is that they have but few tributaries. STREAM DEVELOPMENT 187 A diagram showing these initial consequent streams, in their infancy, is presented in the following ideal map/ Fig. 16. Ideal Sketch-Map Showing Young Consequent Streams. The consequent streams, a to i, follow courses determined by the slope of the surface and approximately straight. The hardness or softness of the underlying rock does not affect them at first, for the reason that they are surface streams. They differ in volume, as is indicated to some extent by the length of the lines representing them in the diagram. All ^ Figures 16, 17, and 18, together with almost all of the account of the development of streams here presented, have been taken from a highly instruc- tive article by W. M. Davis, on " The Development of Certain English Rivers," in The Geographical Journal {oi the Royal Geographical Society), vol. v., pp. 127-146. London, 1895. I 88 RIVERS OF NORTH AMERICA of them begin at once the task of deepening their channels to a certain grade determined by their volumes and their loads. This work would not progress at the same rate in all the streams because they are of unequal length, of differ- ent volumes, and are variously loaded. The longer streams would, under most conditions, be the larger, and would corrade their channels most rapidly. The drainage from the inter-stream spaces flows to the master streams and de- velops feeding branches. The gradients of these branches depend on the rate at which the master streams deepen their channels. The branches of the more rapidly corrading master streams have their velocities, and consequently their corrading power, increased more rapidly than the similar branches of their weaker neighbours. Hence the branches of the stronger consequent streams cut back by head-water corrasion and increase in length more rapidly than the branches of the weaker consequent streams. More and more of the water falling on the territory between the main streams is thus carried to the more favoured conse- quent streams, and increases still more their advantage over their weaker neighbours. The initial streams, it will be remembered, flowed down the original slope, as shown in Fig. 15, at right angles to the strike, that is, across the edges of the strata composing the tilted block of the earth crust we have in mind ; the branches of the streams, however, flow parallel with the edges of the strata where they come to the surface, and hence find hard and soft bands parallel with their courses. As erosion progresses, the edges of the resistant beds are left in relief, forming ridges, while the less resistant beds are STREAM DEVELOPMENT 189 more rapidly removed and determine the courses of the subsequent branches. As this process goes on, our sloping plane loses its smoothness, channels are cut by the conse- quent streams, and depressions are made by their branches trending in general at right angles to their courses. Be- tween the lateral valleys ridges appear, marking the posi- FiG. 17. Ideal Sketch-Map Illustrating Stream Development. tions of the edges of the resistant beds. These ridges are at first straight, and mark the intersection of the hard beds with the original, gently sloping surface. As the strata are inclined, however, one side of a ridge formed by the out- cropping edge of a hard bed will have a more gentle slope than the opposite side. The sides of the ridges slope gently IQO RIVERS OF NORTH AMERICA toward our assumed point of view and present steep escarp- ments in the opposite direction. The condition of the sloping surface before us at this stage is shown in Fig. 17. The hard layers, indicated by broken bands, stand up as ridges, and the branches of the original consequent stream have begun to develop valleys in the soft rocks. Fig. 18. Ideal Sketch-Map Showing an Advanced Stage in Stream Develop- ment. Former Shore Shown by Broken Line The consequent stream c, being stronger and corrading more rapidly than a, deepens its channel through the hard ridge A A more rapidly than its weaker neighbour. The branch m' of the strong consequent stream c, having its place of discharge lowered by the corrasion of the stream STREAM DEVELOPMENT I9I to which it is tributary, is able to remove the soft rocks forming its bed and to grow in length by head-water corra- sion more rapidly than the corresponding subsequent branch of a. As in' increases in length, it captures more and more of the drainage of a, thus weakening that stream, and at length draws off all of its water above 0. The original stream a is thus broken in two, or is beheadedy to use a term proposed by Davis. The notch that a has cut in the ridge of hard rock A A is left as a wind-gap. The bottom of this gap is a divide from which the waters flow each way. The beheaded stream a' holds its former course below the divide, but is weakened by the loss of its head-waters; the portion of the stream a, from the divide on the hard bed to where the subsequent stream m' intersected it, is reversed. For such reversed streams Davis has proposed the name obseqitent. As time goes on, changes similar to those accompanying the backward cutting of in' take place also in the other streams, as may be seen from Fig. 18, in which a more advanced stage in stream development is indicated. Another feature of the changes in progress is shown by the fact that the ridges of hard rock, indicated by the bands A A and B B, are no longer straight. When the stronger streams cut through them, forming water-gaps, the cliffs recede by weathering and by the sapping of their bases by the streams. This wasting of the hard ridges goes on most rapidly on the side toward which they slope most steeply; that is, on the farther side from our assumed point of view. The cliffs, formed by the steeper slope of the ridges of hard rock, thus gradually migrate in the direction of the dip of the hard beds, that is, toward our point of 192 RIVERS OF NORTH AMERICA view, under the influence of general erosion and sapping throughout their entire length. But this recession is most rapid where the stronger consequent streams cross them, and they become lobed or scalloped, as shown in Fig. 17. With the process of stream development, the alignment of the cliffs becomes more and more modified, until the reces- sion in the neighbourhood of the master stream is checked by the streams having cut down nearly to baselevel. At this stage, the cliffs at the ends of the V-shaped gorges, at the apex of which the master streams cross the hard beds, will remain stationary and crumble away, while those por- tions of the cliffs between the master streams, which before receded more slowly than the portions near the stream, will retreat more rapidly than the portions of the escarpment near where the master streams have cut to baselevel. In an advanced stage of stream adjustment and of topographic development, the lines of cliffs, at first straight and then deeply lobed, will again approach an even alignment, but the position of the ridges will change with this development, and move in the direction of the dip of the hard beds. The -cliffs in an early stage of development were of faint relief; when cut into deep lobes, they stand up prominently, but as their alignment is again established, they become sub- dued, and when the process is far advanced nearly or quite disappear. To return to the development of drainage. The manner in which the subsequent stream m\ Fig. 17, captured and diverted the head-water of a and divided that stream into a beheaded portion, a\ a reversed portion, 0, and a diverted portion, a" ^ will serve to illustrate the similar process fol- STREAM DEVELOPMENT I93 lowed by other streams. In each instance, the subsequent branches of the stronger primary stream cut back until they divert a portion of the drainage of their weaker neighbours. The result of this process can be easily predicted from a study of the map. At a certain stage in the process, the stronger streams c and h, will have captured the head-water of all of their rivals above the ridge A A, and a competition between the two conquering streams will ensue. An ad- vanced stage in this struggle is indicated in Fig. 18, where the subsequent branches m!' and n" of the stronger mas- ter stream c have captured and diverted the head-waters of >^^ In the map forming Fig. 18, it will be noted that the ridges of hard rock are again nearly straight, and also that the lower courses of c and h have become tortuous. The reason for the curves in the lower courses of the stronger stream is that after cutting down their channels nearly to baselevel and becoming sluggish, they continue to corrade laterally and form flood-plains on which they meander from side to side, at the same time broadening their valleys. As has been shown on a previous page, a sluggish stream, having little power to overcome obstacles, is more easily deflected than a swifter stream, and besides, the lower por- tions of the valley and plains tracts of a stream during ad- vanced stages in development are regions of deposition. The migrations of the streams a and c have brought them together, as shown in Fig. 18, thus illustrating another process of capture. The examples of stream development we have been fol- lowing are ideal, but, I believe, true to nature. The reason 194 RIVERS OF NORTH AMERICA for sketching an ideal illustration is that in nature various disturbing conditions usually modify the process and in- crease the difificulty of separating what is normal from that which may be termed accidental. Stream development is a slow process even when not disturbed by marked climatic changes or by movements in the earth's crust. The life of a man, or even of a nation, is too short to embrace the time necessary for the development of a river system. It is only by studying many streams in various stages of their develop- ment that geographers are able to sketch generalised pictures of the normal changes a great river passes through in its life- span of millions of years. When one attempts to apply the elementary conceptions of stream development outlined above to actual streams, it is found that many modifying conditions have to be taken into account. Broad surfaces with even initial slopes are rare; the rocks forming the earth's crust are frequently folded and faulted, especially in uplifted regions, and con- sequently the development of subsequent streams is fre- quently greatly modified; more puzzling complications arise, however, from the fact that the land is not stable, but is subject to up-and-dov/n movements, which disturb or entirely arrest the slow process of stream development be- fore it can run its normal course. These and still other modifying conditions have to be considered in studying the history of the streams which drain the land and throughout their history are continually modifying the relief of the surface. With the introduction to the principle of stream develop- ment in mind, let us turn to a portion of our own land STREAM DEVELOPMENT 1 95 where the processes just outlined have been long in action and see if we can read a portion of the history recorded by the valley and intervening hills. EXAMPLES OF STREAM DEVELOPMENT AND ADJUSTMENT IN THE APPALACHIAN MOUNTAINS The leading geographical feature of the Appalachian Mountains, more especially of their northern half, is the large number of curving but generally parallel, level-topped ridges with valleys between, which compose the uplifted region. Crossing the ridges and valleys approximately at right angles is a series of rivers, such as the Delaware, Sus- quehanna, Potomac, and James, which have their sources to the west of the mountains, and flow eastward to the sea. These master streams receive many branches from the valleys crossed by them. The general features referred to are shown on the map forming Plate IX. The ridges in the northern Appalachians are known to be due to folds in the rocks, which have been truncated or planed off to a certain general level, and the surface thus formed upraised and eroded so as to leave the edges of hard beds in bold relief. The first question suggested by an in- spection of the accompanying map is : How has it come about that the main rivers flow through the ridges of hard rock by means of gaps cut in them, instead of being turned aside and pursuing what would seem to be much easier courses to the sea? The ideal case of river development we now have in mind will assist in solving this problem. The study of the northern Appalachians, conducted by a large number of geologists, and especially by Davis and 196 RIVERS OF NORTH AMERICA Willis, has shown that after the rocks were folded the region existed as a land area, probably more elevated than now, and was worn down nearly to sea-level, or, in other words, was reduced to the condition of a peneplain. This pene- plain was subsequently elevated and tilted so as to slope toward the south-east. The conditions were then essentially the same as in the ideal case already discussed, except that the rocks beneath the tilted plane had a complex structure. The rocks were also of many degrees of hardness and solu- bility. The south-eastern margin of this tilted peneplain was at sea-level, while its north-western border, in the region now embraced in the central and western portion of New York, Western Pennsylvania, and West Virginia, was ele- vated, not all at once, but slowly, to a height of probably two or three thousand feet. We have designated the sloping surface referred to as a tilted plane, more accurately it should be considered as the side of an elongated dome-like uplift. The pre-existing streams flowing with slack currents in their old age, and young consequent streams originating on the tilted peneplain, took the direction of easiest descent, and flowed south-eastward to the sea. The courses of these streams were determined by the slope of the surface irre- spective of the position or character of the rocks beneath, and hence are consequent streams. As they deepened their channels, the edges of hard and soft beds were cut through. With this process of deepening the channels of the conse- quent streams, many subsequent branches originated which also entrenched themselves, but their directions were con- trolled by the hardness or solubility of the rocks. Those ., I *v LW"V> 1-: f^--^^^:#\ The Nortliern Apj^alachiaii Apf^roxiinate scale: one in After liailey Willis.) >venty-seven miles. STREAM DEVELOPMENT I97 originating on soft beds maintained their positions, while those which flowed at first along the outcrops of hard beds were soon shifted to the softer beds adjacent. The hard beds were thus left as ridges between the valleys excavated by the subsequent streams along the outcrops of soft beds. As the master streams flowing south-eastward lowered their channels their subsequent branches were given greater velocity and deepened their channels also. This sinking of the rivers and of all their branches produced a roughening of the topography and the once nearly smooth plain became a rugged mountainous region. In this general down-cutting the large consequent streams first reached baselevel at their mouths, and then a low gradient, or an approximation to baselevel, was produced progressively up stream. The tend- ency of all the streams, or their chief aim, as we may say, was to reduce the land to a second baselevel. This would be accomplished by the downward corrasion of the streams in their upper courses, and a broadening of their valleys in their lower courses ; the broadening process progressing up stream as fast as the valleys were deepened at a certain progressively decreasing rate. During the process outlined above, each of the subsequent branches of the main streams entered into competition with its neighbours for the possession of the territory between them, as in the ideal illustration of stream development previously considered. The branches of the larger master- streams, by having their places of discharge lowered more rapidly than adjacent subsequent streams flowing to weaker consequent streams, were able to extend their head branches and capture new territory in the manner already discussed. 198 RIVERS OF NORTH AMERICA This process was modified in many ways, however, owing to complex folding in the rocks exposed by erosion. Influence of Folds i?i the Rocks on Stream Adjiistvient, — A fold in stratified rocks of various degrees of resistance, when the axis is horizontal, will produce parallel ridges and valleys with tapering ends. If the axis of the fold is not horizon- tal but inclined so as to pass below a horizontal plane in one direction and rise above it in the opposite direction, it is evident that if the region where the fold occurs is carved away to a horizontal plane, and then etched so as to leave the edges of the hard layers in relief, the resulting ridge will not be parallel throughout, but form more or less elliptical curves/ Synclinal Fold, with Central Canoe-Shaped Valley. Anticlinal Fold, with Hemi- Cigar-Shaped Mountain. Fig. 19. Topographic Forms Resulting from the Erosion of Folded Rocks. (After Bailey Willis). The topographic changes resulting from the weathering and erosion of rocks of various degrees of resistance, when ' Folds in the rocks, if traced to where they die out, will be found either to flatten and spread so as to merge with undisturbed areas or become narrow and more or less sharp-pointed. Individual folds are more or less conical and when cut by planes of erosion give figures which are conic sections. STREAM DEVELOPMENT 1 99 folded, are shown in Fig. 19. In one instance the fold is downward, so that the strata on the borders dip toward the longer axis, and is termed a synclinal ; and in the other instances the strata forming the arch dip away from the longer axis, making an anticlinaL Water-Gaps and Wind-Gaps, — The Appalachian Mount- ains are due to the upraising of a great belt of country in which the rocks have been folded into anticlinal and syn- clinal, as in the illustration just given. The longer axes of the folds trend N. E. and S. W., and are either horizontal or pitch at various angles. The western side of each fold is usually steeper than the eastern side. The ends of the folds frequently overlap, one dying out and another begin- ning and continuing sometimes for scores of miles, before it in turn disappears and is replaced by another similar fold. It is the ridges in these variously truncated folds, due to the weathering out of resistant la3^ers, that give to the Appalachians their highly characteristic topography. The softer beds have been eroded away by the subsequent streams. The ridges of resistant rock form the divides be- tween the branches of the large rivers. The crests of the ridges are nearly level for the reason that the region was worn down to a peneplain before the etching process which gave them prominence was initiated. These level crest- lines are broken, however, by deep notches where the master streams pass through them, and are also indented by less deep notches where streams which have been beheaded formerly crossed them. The process of river conquest, as it has been termed, by which notches have been left in the crest-lines of the ridges. 200 RIVERS OF NORTH AMERICA is illustrated by the following typical example in Virginia, borrowed from an admirable essay on the northern Appa- lachians, by Willis.' The Potomac near Harper's Ferry flows through two deep THE ^ KITTATINNV y^ /^ PLAIM yfC .y -.-'^■^*^^^»x-,^^4oV /> ( 5V <.' Fig. B. Water-Gaps Cut by the Potomac through 1 wo Kidges ot Hard Rock, near Harper's Ferry, W. Va. The point of view is on the Shenandoah peneplain ; the Potomac flows through a steep-sided trench about 225 feet deep, sunken in this peneplain. Loudoun Heights on the left and Maryland Heights on the right in the background. SOME CHARACTERISTICS OF AMERICAN RIVERS 269 system of branching valleys formed by the sinking into the rocks of the preglacial Mississippi drainage, but its upper portion was deeply buried by ice during the height of the Glacial epoch. When the glaciers finally melted, the re- born streams found their channels blocked, and in many instances were turned from their former courses in the same manner as in the case of the streams of Wisconsin and Minnesota. When the Laurentian glacier retreated to the northward of the height-of-land now dividing the streams which feed the Mississippi from those flowing to Hudson Bay and the Great Lakes, several lakes came into existence, which were retained on their northern margins by the face of the re- treating glacier, and supplied southward-flowing streams. One of these glacier-dammed lakes, named in honour of Louis Agassiz, occupied what is now the valley of Red River and the Winnipeg basin, and supplied River Warren which flowed to the Mississippi. Another similar lake was found in the western part of the present drainage basin of Lake Superior, and had its outlet a few miles west of the site of the city of Duluth. Other lakes in this same category oc- cupied the southern part of the basin of Lake Michigan, and the western part of the Erie basin ; the former discharged in- to the Mississippi through the channel now being converted into a canal, just west of Chicago, and the latter flowed through the valley now occupied by the Wabash to the west of Fort Wayne, Indiana. During the time the Winnipeg and Laurentian basins were sending their surplus waters southward to the Mis- sissippi, not only was the run-off from the land probably 270 RIVERS OF NORTH AMERICA greater than now on account of heavier rain-fall, but the vast snow- and ice-sheet which covered Canada was melting, and all the stream channels leading away from it were flooded. The volume of water contributed in these several ways and flowing through the Mississippi valley to the Gulf of Mexico must at all seasons have been far in excess of the greatest of the modern floods. It has been estimated by James E. Todd ^ that the Mississippi, during the geological springtime following the great winter known as the Glacial epoch, carried annually from eleven to twenty times the volume of water now reaching the Gulf of Mexico through the same channel in a single year. Whether the current of the rivefduring this great flood stage was vastly increased or not, depends on the former elevation of the land. These are reasons for believing that the region occupied by ice was depressed several hundred feet below its present posi- tion, and that the gradient of the Mississippi was much less than at present. The expanded rivers then resembled a great sea in which the loess and other similar deposits now occupying the greater Mississippi valley were spread out. The long preglacial history of the Mississippi, the many changes impressed by the glaciers directly on its tributaries from the Appalachians to the crest of the Rocky Mount- ains, and indirectly, owing to the vastly increased water- supply, on the character of the river all the way to the Gulf, make it a most instructive subject for study. An ad- ditional chapter in the life of the river is supplied by a modern submergence which allowed the sea to reach to the mouth of the Ohio, and of still later re-elevation. The ac- ' Geological Survey of Missouri, i8g6, vol. x., p. 203. SOME CHARACTERISTICS OF AMERICAN RIVERS 27 1 / cidents, as they have been termed, in the normal develop- ment of streams, due to climatic changes and to movements in the earth's crust, thus find numerous and graphic illustration in the Mississippi Valley. Until the entire basin, however, has been examined as a unit, disregarding political boundaries, even a satisfactory outline of its entire geographical history cannot be written. Another phase of the wonderful story of the Mississippi deals with its influence on the early explorations of the in- trepid emissaries of Spain and France, and the final con- quests of its basin by the English, the vast agricultural importance of its rich lands, and its value as an avenue of commerce; but the influence of geographical history on human events pertains more properly to the domain of the historian, and cannot be treated even briefly at this time. Canyon Rivers. — The Colorado River, rising in the mount- ains of Colorado, Wyoming, and Utah, and flowing through a high and for the most part arid table-land, has carved in the solid rocks the most magnificent canyon that has yet been studied. The river, with its load of sand and mud, has been able to deepen its channel more rapidly than its bounding walls have been lowered by rain, rills, and other destructive agencies. The result is a steep-walled trench of such stupendous proportions that when its sides are seen from below they appear to be towering mountain ranges. The tributaries of the Colorado have also deepened their channels at approximately the same rate as the main stream has excavated its canyon. A great river with many branches has thus been sunken into the rocks, to the depth, over a vast area, of about one m.ile. Between the larger 2/2 RIVERS OF NORTH AMERICA streams there are flat-topped table-lands, remnants of the great plateau across which the Colorado flowed in its infancy. The plateau has been slowly elevated, while the sand- charged streams, acting like saws, have dissected it. Throughout a region tens of thousands of square miles in area, every stream is in the bottom of a profound gorge of its own making. The remnants of the plateau between the canyons are waterless and desert. The Colorado throughout a large part of its course flows through a canyon that is from four to six thousand feet deep. The canyon walls are, for the most part, of horizon- tally bedded rocks of many tones and tints, and various degrees of hardness. Weathering has increased the variety of colours, and rendered them more brilliant than they are in the unchanged rocks. The rain and wind have sculptured the cliffs so as to give them the greatest imaginable variety of forms. The most vivid dream-pictures of gorgeous Oriental architecture fail to rival the temple- and cathedral- like forms, incrusted with harmoniously tinted decorations, which overshadow the Colorado for hundreds of miles. Fig. 23. Cross-Profile of the Canyon of the Colorado. (After W. H. Holmes.) Vertical and horizontal scale the same : one inch = 6375 feet. There is nothing of the same class in the whole world, so far as is known, to compare either in extent and height, in richness of colour, or in variety and intricacy of detail with SOME CHARACTERISTICS OF AMERICAN RIVERS 273 the canyon walls in the southern portion of the Pacific drainage slope. The canyon of the Colorado is not an even- sided canal, but a great valley some fifteen or more miles across. In the bottom of this greater canyon, as may be seen from the accompanying illustration, Plate XVII., re- produced from a drawing by W. H. Holmes, one of the few artists who are true to nature, is sunken a much narrower and deeper inner canyon. The reader will be able to read in this picture some of the leading events in the geographi- cal history of the region of the Great Plateau. This outer canyon is clearly the record of a time when the land was some four thousand feet lower than now, and remained at that horizon for tens of thousands of years, while the river cut down its channel to baselevel and by lateral corrasion broadened its valley. The climate, at least near the close of this long period of uninterrupted work, was arid, as is shown by the precipitous character of the cliffs bordering the valley that was excavated. The river meandered in broad curves over the nearly level bottom of its valley, and when the land was again raised maintained its winding course, and owing to renewed energy, due to greater ve- locity on account of an increase in gradient, again began the task of corrading to baselevel. This new task imposed upon the river is not yet completed. The waters still flow swiftly, and vertical corrasion is still in excess of lateral wear and of weathering. The precipitous character of the cliffs border- ing the inner gorge, and the details in their sculpture, indi- cate that the climate has had its present characteristics throughout the greater part, and probably the whole, of the time since the energy of the river was renewed. 274 RIVERS OF NORTH AMERICA The walls of the canyon of the Colorado are not even- surfaced precipices, but on either side of the river are but- tressed by outstanding ridges and retaining walls, with many lateral branches. Everywhere there are towers and pin- nacles, as well as innumerable alcoves and recesses. The main buttresses extend out for miles from the brink of the gorge and partially fill the profound chasm into which they descend. From within the purple depth of the canyon rise wondrous temple-like forms as gorgeous in colour and as rich in fretwork and arabesque as a Moorish palace. These shrines for Nature-worship, although minor features in the sublime panorama, tower as far above the shining stream flowing past their bases as the summit of Mount Washing- ton rises above the sea. As an illustration of the endless variety, both in form and colour, of architectural forms that Nature can sculpture from an upraised block of the earth's crust under certain condi- tions of climate and rock texture, the Colorado region is unrivalled. The student of earth-forms there finds many illustrations of the various phases that an upraised region passes through, in what may be termed its youth. In a re- view of the life histories of rivers, the Colorado furnishes an example of a stream yet young, so far as its advance in its appointed task is concerned, but one which, owing to un- usual opportunities, has surpassed many older but less favoured streams in the magnificence of the results accom- plished. The Colorado is not only a young stream, but has been termed a precocious youth. Its success, however, in producing wonderful scenery of a novel type, lies not so much in the amount of work performed, as in the fact that SOME CHARACTERISTICS OF AMERICAN RIVERS 275 destructive agencies have spared the canyon walls as the stream entrenched itself. The climate is arid, and the wasting of the cliffs consequently retarded. Stupendous as are the results achieved by the Colorado, and wonderfully impressive as is the scenery along its course, the amount of work it has done — that is, the num- ber of cubic miles of rock removed — is small in compari- son with what has been accomplished in many regions of mild relief, where rivers in their old age flow sluggishly over a plain from which they have removed nearly every vestige of a former mountain range. The region of great plateaus drained by the Colorado will, under the action of the agencies now in operation, be reduced to such a plain, unless future upheaval again renews the youth of the river. Sierra Nevada Rivers. — The numerous bright, leaping rivers of the Sierra Nevadas, flowing through valleys three to four thousand feet deep and overshadowed by pine- clothed mountains, suggest many questions in reference especially to the influence of rock texture, changes in eleva- tion and glaciation on stream erosion, and on the origin and development of topographic forms. The valleys are nar- row, with usually little if any bottom-land. The rivers are swift and strong and carry along with ease all of the debris delivered to them by the bordering slopes and tributaries. Not only do they bear away all of the fine material that reaches them, but in times of high water roll along large boulders, and yet their capacity to transport is not satisfied, and they are clear, limpid, joyous streams during a large part of the year. The conditions are there the reverse of what is so manifest in the rivers of the Gulf States, where previously 276 RIVERS OF NORTH AMERICA eroded valleys are being filled and broad areas of new land have been formed. In the Sierra Nevadas the streams are all at work at the task of deepening their channels; the stage in their lives when they will broaden their val- leys more rapidly than they deepen them has not been reached. The Tuolumne, King, Truckee, and many other rivers are not only young, but are still broken by cataracts and rapids, and in many instances are supplied in part by the overflow of lakes. These are plain evidences of youth. A little study shows one, however, that the tireless activity of the streams is largely due to a recent uplifting of the mountains, which has given them steeper slopes, and that the presence of waterfalls and lakes along their courses is in many instances due to the former existence of great snov/- fields and magnificent glaciers in all of the higher valleys. The energy with which the streams are working is thus seen to be due to a revival of activity, or a rejuvenation, rather than to actual youthfulness. The westward-flowing streams from the Sierra Nevadas experience a sudden change on leaving the mountains and entering the flat-bottomed valleys where they unite to form the San Joaquin and Sacramento. With loss of grade the waters flow less rapidly, and their burdens are dropped. Deposition and aggrading are then the rule instead of abra- sion and valley-deepening. Borings made in the bottom of the broad, nearly level-floored valley of California, show that a great depression between the Sierra Nevada and Coast mountains has been filled to a depth of many hun- dreds of feet. Much of this filling is due to the deposition of material swept out of the bordering mountains in order SOME CHARACTERISTICS OF AMERICAN RIVERS 277 to form the gorges and canyons which give them much of their diversity and beauty. Conditions similar to those already noted on the Atlantic coast, where a subsidence of the land has transformed the river valleys into estuaries, are again manifest on the western border of the continent. The story of stream development and of changes in the re- lief of the land on the Pacific coast, due to the upheaval of the land, is supplemented by the effects of subsidence on the geography where land and ocean meet. The bay of San Francisco and its outlet through the Golden Gate show that valleys have been drowned, owing to a downward movement of the land. Surveys of the sea bottom adjacent to the present coast-line reveal the fact that former river courses may, in some instances, be traced over the conti- nental border now depressed beneath the Pacific, in the same manner that soundings have demonstrated a former seaward extension of the Hudson, St. Lawrence, and other streams of the Atlantic slope. It seems scarcely necessary to mention, so obvious is it, the intimate relation between geographical history and human activities, illustrated by the origin and marvellous growth of the metropolis of the Pacific coast on the border of a partially submerged river valley. The magnificent bay of San Francisco, one of the very finest harbours in the world, is a direct result of a long series of geographical changes. The subsidence which converted a portion of the valley of the Sacramento into an arm of the sea has had a direct and far-reaching influence not only on the lives of millions of people, but on the building of a nation. The future greatness of San Francisco, assumed by her com- 278 RIVERS OF NORTH AMERICA manding geographical position, will make her an important factor in the spread of civilisation, not in America alone, but in the countries bordering the distant shores of the Pacific. *' Where Rolls the Oregon,'' — The Columbia and its main tributary, the Snake, rise in the Rocky Mountains, and flow across a region of small rain-fall, thus simulating some of the main conditions which have influenced the history of Colo- rado River. Snake River crosses a basaltic plateau and has excavated a magnificent canyon. Although inferior in the richness of its colouring and the profusion of details in its sculptured cliffs, it is comparable in many ways with the Grand Canyon of the Colorado. The walls of Snake River canyon are composed mainly of black basalt in horizontal layers, which assumes a great variety of cathedral-like and monumental forms on weathering. The architecture is locally varied where granite and schist are exposed in the lower portions of the profound gulf, but throughout hun- dreds of miles of great escarpments the dark basalt gives a sombre and even an oppressive gloom to the strange scenery. In its deepest portion, on the east flank of the Blue Mount- ains, the canyon is about four thousand feet deep and fifteen miles broad. As in the vast canyon carved by the Colorado, ridges and abutments from the main walls extend from either side far into the profound depths, and fill the depression so as to make it appear much narrower and deeper than it is in reality. The Columbia also flows in a canyon-like valley for much of its course after leaving the Rocky Mountains. In its wild passage through the Cascade Mountains, it is bordered by some of the most rugged river scenery to be found on SOME CHARACTERISTICS OF AMERICAN RIVERS 279 the Pacific coast, but nowhere has it formed a canyon com- parable with that traversed by Snake River. The main subjects of interest to admirers of bold scenery as well as to the student of topographic forms and of stream development, presented by the vast region drained by the Columbia, centre in the relation of the drainage lines to the disturbances which have affected the rocks. In the Appa- lachians, as we have seen, many of the rivers flow across folded rocks and have cut water-gaps through the ridges; in the region drained by the Columbia the streams fre- quently cross rocky ridges formed by the upraised edges of tilted blocks of the earth's crust, and also give origin to water-gaps. In several instances sharp-crested walls of rock from a few hundred to two or three thousand feet high, have been upraised directly athwart the course of the Columbia or of its branches, but the rivers have not been turned aside. As the blocks were tilted and their edges slowly elevated, the rivers deepened their channels as fast as the land rose and thus maintained their right of way. In other instances, the waters were held in check for a time by the rising land, and caused lakes to form, but the barriers were slowly cut across by the out-flowing streams, and again, what may be termed gateways were opened through the ridges. The thickness of ancient lake sediments over broad areas in the region under discussion shows that earth- movements, similar to those which influenced the character of the present topography, have been long in progress and have produced profound geological as well as geographical changes. The tilting of blocks of the earth's crust in the region drained by the Columbia has not only produced ridges of 28o RIVERS OF NORTH AMERICA the nature just referred to, but in certain instances the surface has been depressed, thus lessening the grade of the streams and causing them to deposit their loads and ag- grade their valleys. Broad areas have for this reason been transformed into nearly level alluvial plains. In the instructive Columbian region, and especially in that portion of central Washington known as the Big Bend country, where the climate is now arid and the rate of gen- eral waste from the surface due to atmospheric agencies small, lines of fracture and of moderate faulting have deter- mined the direction of the stream courses. The streams flow along lines of fracture and in some instances have ex- cavated canyons with one wall higher than the other. This is the only region in North America, so far as has been recognised, where the relation of streams to fractures in the earth's crust favours the once prevalent idea that valleys are due to breaks in the rocks, instead of resulting from the wearing action of streams. Even these minor examples, however, of the influence of fractures on drainage fail ta support the hypothesis referred to, since the breaks simply gave direction to the streams which subsequently excavated the valleys, instead of directly producing the depressions. Another feature of especial interest in the land of the ** Oregon," illustrating the influence of climate on the lives of streams, is furnished by the Grand Coulee, a deep, steep- sided canyon which cuts across the plateau partially enclosed by the Big Bend of the Columbia. The Columbia once flowed through this great trench, having been turned from its present channel by the advance of a glacier from the mountains to the north. This dam of ice held the river in SOME CHARACTERISTICS OF AMERICAN RIVERS 28 1 check and caused it to rise and form a long, narrow lake, the outlet of which was through the previously eroded canyon now known as the Grand Coulee. The Columbia flows directly through the Cascade Mount- ains nearly at right angles to their trend, in a wild and ex- ceedingly picturesque water-gap, the greatest of its class on the continent. The complete history of this most impres- sive topographical feature has not been made out, but the facts in hand suggest that the mountains, like the narrow, sharp-crested ridges to the eastward, are due to the upraising of a block or a series of blocks of the earth's crust, along a line or belt of faulting, and that the river deepened its channel as fast as the rocks rose. Possibly the elevation of the land was not uniform, but progressed by stages, and that when most rapid, the river, unable to maintain its grade, was ponded and lakes forrned. This explanation of the origin of the Dalles of the Columbia, and other gorges both above and below, must not be accepted too hastily, however, as the possibility of cross-fractures having given direction to the river and assisted it in its task has not been fully considered. Where the Columbia nears the ocean its waters lose their energy and expand into an estuary, in which the tides rise and fall for a distance of about one hundred and forty miles from the ocean.' From what has been said concerning the ' Rev. Earl M. vVilbur of Portland, Oregon, has informed the writer by let- ter, on the authority of government engineers, " that the tide is felt in the Columbia as far as the Lower Cascades, which is, I believe, about 140 miles from the mouth of the river ; the extreme range there being al)out six inches. ** In the Willamette, the Columbia's largest affluent, the tide is felt as far as Oregon City, where there are falls ; about 115 miles from the ocean. " The extreme range noted at Portland, about 100 miles from the ocean, is 3.2 feet." 282 RIVERS OF NORTH AMERICA drowning of stream-cut valleys both on the Atlantic and Pacific coasts, it will be readily seen that a modern subsi- dence has recently affected a great extent of the coastal region of the North-west, including south-eastern Alaska, and has allowed the sea to encroach on the land and trans- form the lower courses of many valleys into tideways. The extremely interesting problems presented by the region drained by the Columbia, and the magnificence and novelty of the scenery existing there, tempt me to detain the reader and consider more fully the origin of the mas- sive cliffs and of the terraces and landslides on their faces. There are yet other connections between the elements of scenery and the work of streams, and a wonderful story of the time when the land was again and again inundated by floods of molten lava, but as the object of this fireside recon- noissance is simply to indicate some of the more instructive features of the land which the study of streams assists in interpreting, we must hasten on. Our journey is northward. Rivers of the Far North- West. — Fraser River, fed by tens of thousands of twig-like branches on the western slope of the Cordilleran Mountain system, furnishes much informa- tion in reference to the manner in which a broad, high region becomes dissected by the streams flowing from it. Many of the branches of this splendid river have their sources in fine glaciers, high up among the glorious peaks of the Selkirks and neighbouring ranges, flow through wild, steep-sided gorges and valleys and unite to form a trunk stream which has sunken three or four thousand feet into the rocks. In following the steep bank of the Fraser, while making the transcontinental journey over the Canadian SOME CHARACTERISTICS OF AMERICAN RIVERS 283 Pacific, one sees on every hand evidences of the work of streams. The topography is yet young, although deeply and boldly cut, but the valleys are narrow, barely wide enough to give the rushing, foaming waters a passageway. Throughout much of the trunk portion of the Fraser drainage-tree, the grade is sufficiently steep to insure a rapid current. The debris brought from glaciers, and fed by tributary rills and creeks, supplies the swiftly running waters with an abundance of tools with which to deepen their channels. Many conditions favour rapid work, and it is not surprising that the swift, debris-charged river has literally sawed a great mountain system into blocks, and is progressing rapidly with the task of removing the masses still remaining between its branches. The walls of its main canyon, although less precipitous than the bordering cliffs of the Colorado or the Snake, are wonderfully varied and picturesque. When seen from below they appear like deeply sculptured, forest-clothed mountain ranges. The river is yet young, but has accomplished a herculean task, and is still working with the energy of youth. As in the case of Snake River, the Fraser was interrupted in its work of cor- rasion during the Glacial epoch and its canyon deeply filled ; more recent corrasion has removed much of the alluvium, however, leaving well-marked terraces, as is illustrated on Plate VIII. Its canyon, although three or four thousand feet deep, has not yet reached the limit to which down- cutting is possible. Vertical corrasion is still in excess of lateral wear and weathering, and the great trench is V- shaped in cross-section instead of being broadly U-shaped, as will be the case in its mature life. Like the streams of 284 RIVERS OF NORTH AMERICA the Sierra Nevadas, its energy is not all consumed in trans- porting the debris delivered to it, and for a large part of the year it rushes along as a foaming, roaring torrent carry- ing its load easily until it enters the coastal region, where a recent depression of the land causes a decrease in grade and a consequent loss of velocity. The river is shorter than formerly, for the reason that its trunk near the sea has been transformed into an estuary. The drainage-tree has been betrunked by subsidence and drowning. Glacier-Born Rivers, — North of the Fraser, and similar to it in the chief points of their histories, are the Stickine, Taku, Alsec, and other rivers which have their sources to the east and north of the mountains near the coast and flow through rugged and as yet but little-known lands to the sea. Probably all of the region drained by these rivers was ice- covered at a comparatively recent date, and thousands of glaciers still remain. Many are the lessons illustrated by the rugged landscapes of British Columbia and Alaska of the manner in which streams and glaciers modify topography, and the way that a subsidence of a deeply dissected land leads to the production of a ragged coast-line fringed with islands. This region includes the highest mountain and the largest glaciers in North America. The chief lesson that invites the geographer amid the ice-covered mountains near the coast is the influence of climate and of topography on the birth, growth, and decline of glaciers. This theme has been considered in a preceding volume.* The streams, many of them veritable rivers, flowing be- neath the glaciers, make highly interesting deposits of ^ I. C. Russell, Glaciers of North America. Ginn & Co., Boston, 1897. SOME CHARACTERISTICS OE AMERICAN RIVERS 285 gravel in the tunnels they occupy, and form alluvial cones and broad sand-plains after escaping from the ice. The study of these peculiar accumulations, still in process of formation, furnishes an explanation of many riddles in formerly glaciated lands. To the north of the narrow coastal portion of southern Alaska, where the surface waters are discharged directly to the Pacific, lies the great region drained principally by the Yukon, which forms the Bering drainage slope. The hydro- graphic basin of the Yukon embraces about 440,000 square miles, and the volume of the river, although as yet un- measured, is comparable with that of the Mississippi. The Yukon presents many of the characteristics of the rivers of more southern latitudes, and also possesses certain features peculiar to the streams of northern countries. Flowing, as it does in its upper course, from south to north, the wave of sunshine and warmth that sweeps from the equatorial to polar regions each recurring springtime reaches the lands drained by its head-waters and loosens the icy grasp of winter, while its lower portion is still ice- bound. The melting of the snow and ice and the spring rains at the south cause the streams to rise in floods, which advance laden with floating ice upon the still frozen country to the northward. Ice-dams are formed, and the streams expand and inundate the forest-covered valley bottoms. The rising waters finally break the ice-dams and rush on down the valleys carrying destruction in their paths. Trees are uprooted, or cut off by the floating ice as with a scythe. Vast quantities of earth and stones, enclosed in the ice that formed in shallow water, are borne along and 286 RIVERS OF NORTH AMERICA deposited in part over the flood-plains of the streams when the ice melts. The energy with which the Yukon modifies its banks, on account especially of the ice-laden floods, is unrivalled by any more southern river. Another important variation in what may be considered as the normal action of streams arises, in the Alaskan region, from the constantly frozen condition of the soil. Throughout nearly the whole of the area drained by the Yukon the soil in the low lands is continually frozen. The winters are long and severe, the summers short and hot. The soil at a depth of a few inches beneath the usual cover- ing of moss, shrubs, and trees is perennially frozen. The thickness of the frozen subsoil is not known, but excavations twenty-five feet in depth have failed to penetrate it. Ice- clifl"s along the Kowak River, in North-western Alaska, re- veal a thickness of fully two hundred feet of dirt-stained ice beneath a thin layer of black mucky soil on which grasses and other vegetation thrive. From these and other observ- ations, and especially the records of certain borings made in a similar region in Siberia, it is safe to assume that the average thickness of the frozen layer in Alaska is probably in excess of one hundred feet and possibly two or three hundred feet or more in depth. These conditions have an important bearing on the work of streams. Frost renders otherwise loose and inadherent material as firm as solid rock. The action of flowing waters on the land is thus checked, and stream development as well as rock disintegra- tion and decay and general surface erosion greatly retarded. The climatic conditions in the region under discussion are such that the ground almost everywhere is covered with a SOME CHARACTERISTICS OE AMERICAN RIVERS 287 dense growth of mosses and lichens, which make a living mat through which the surface waters percolate as through a layer of sponges, and are filtered of all matter in suspen- sion. Thus, again, the work of the streams is delayed, for the reason that sand and silt, which ordinarily constitute the principal tools with which flowing waters abrade the rocks, are removed. This process of filtering the water is illus- trated by the contrasts in the character of the tributaries of the Yukon which come to it from the north and from the south. Every stream, so far as is known, which joins the great river along its right bank is clear, although usually amber-coloured on account of the organic material contained in solution ; while the tributaries entering from the left, or, in general, the southern bank, are mostly turbid and heavily loaded with sediment, for the reason that they have their sources in glaciers. White River, one of the principal tributaries of the Yukon from the south, is charged with material in suspension not only because it is fed by melting glaciers, but for the reason that it flows through a region that is covered with fine volcanic dust, some of which is washed into the stream by every rain. The accidents to streams, as they have been termed, due to glacial and to volcanic agencies here find abundant illustration. The valley of the Yukon and of several of its important tributaries, particularly to the east of the Alaskan boundary, are marked by conspicuous terraces. A part of these are lake terraces, formed at a time when the waters were held in check by a lava dam, but other and equally conspicuous terraces were formed by the streams, and record changes in the altitude of the land, or the overloading of the waters 288 RIVERS OF NORTH AMERICA with detritus during a time when glaciers near their sources were much more abundant and of far larger size than now. The head-branches of the Yukon drainage-tree rise in a country which was formerly covered with a continuous ice- sheet, but in its middle and lower courses evidence of for- mer ice occupation is wanting. Marked differences in the scenery beheld in journeying from one of these regions to the other have been noted by several travellers. Bering Sea, into which the Yukon empties, is shallow, at least in the portions bordering Alaska, and is without strong currents or high tides. The great river on entering the sea drops its heavy burden of silt, and has built up a delta comparable in extent with that of the Mississippi. The river divides into many branches, or sends off several distributaries in the delta portion of its course. The first of these diverging channels leaves the main river about a hundred miles from its mouth. The low, swampy area built by the stream is treeless, but clothed in summer with a dense growth of mosses, lichens, grasses, rushes, and a great variety of less conspicuous flowering plants. This luxuriant garden of brilliant flowers and luscious green fronds and leaves is but a veneer of verdure concealing a frozen morass. This is a portion of the vast treeless tract of perennially frozen ground known as the tundray which fringes the shores of Bering Sea and the Arctic Ocean. The several distributaries of the river flow through this new- made land in meandering courses, and enter the sea at various localities, over a breadth of seventy miles of coast. With the exception of the delta portion of the Yukon, its banks are fringed with spruce trees, cottonwoods, and SOME CHARACTERISTICS OF AMERICAN RIVERS 289 willows. The annual floods and ice-gorges cause large numbers of trees to be swept into the river, and the swift ■current at numerous localities cuts away the banks in such a way as to undermine the trees growing on them and cause them to fall into the waters with roots and branches at- tached. Great quantities of drift-wood are thus contributed to the river, as has already been described. The conserva- tive influence both of growing trees and of stranded drift- wood on the banks of the river are well marked and important, as are also the destructive tendencies of the same agencies. The trees on being uprooted tear away the banks, and on being stranded frequently deflect the current so as to cause it to cut away neighbouring shores and in- crease the number and extent of the windings of the river. Arctic Rivers. — Of the streams flowing down the Arctic drainage slope but little can be said, for the reason that no traveller especially interested in the study of modern geography has visited that region. The Mackenzie prob- ably illustrates the characteristics of a northward-flowing Arctic river even better than the Yukon. Much of the region it traverses is forested, and vast floods occur each spring when the thick ice of winter breaks up and is swept northward. The sudden changes that occur at the turn of the annual tide of temperature must be even grander than along the Yukon, but in this connection but little informa- tion is available. On entering the Arctic Ocean when the tides are low and currents produced by the winds mostly lacking, owing to the fact that the sea is covered with ice- floes throughout the year, the river deposits its sediment and is engaged in building a large delta. The three great 19 290 RIVERS OF NORTH AMERICA delta-making rivers of North America are the Mississippi^ Yukon, and Mackenzie. Rivers of the '' Great Lone La7td/* — On the Hudson Bay drainage slope there are tens of thousands of lakes, which, for the most part, occupy basins due in one way or another to the former occupation of the land by glacial ice. There are also many rivers, but the way in which they illustrate the principles of stream development has received but slight attention. The reports of explorers, especially those connected with the Geological and Natural History Survey of Canada, show that the drainage is immature. The streams have not cut down their channels so as to furnish direct and ready avenues of discharge for the surface waters. This is demonstrated especially by the countless lakes. The streams are not only young, having come into existence or having been rejuvenated since the last retreat of the glaciers, but have developed slowly on account of adverse circum- stances. Among the conditions that have retarded stream development may be noted the general low altitude of the land, and, consequently, gentle gradients of the stream channels and lack of energy in the flowing waters. The winter climate is severe, and the streams either ice-covered or frozen to their bottoms for several months each year. Snow protects the ground in winter. The subsoil, as in the Yukon basin, remains solidly frozen in many places even during the warm season. Erosion and the transportation of debris by the streams is thus limited to one half, or even less, of the year. Forests with undergrowths of mosses, lichens, and other plants shield the soil in summer from the beating of rain, and filter the percolating surface waters, SOME CHARACTERISTICS OF AMERICAN RIVERS 29I thus robbing them of the means of abrading the rocks over which they flow. There are no glaciers to supply the streams with sediment. The vegetation retards the gather- ing of the waters into rills, and equalises the flow of the streams in such a manner that the floods caused by melting snow have their energy diminished. The river banks are clothed with trees and shrubs, especially willows and alders, and their roots bind the soil and increase its ability to resist the attacks of the flowing waters. Drift timber lodged against the sides of the streams, especially fallen trees which still retain a hold on the land, also protect the banks. For these and still other reasons, the work of the streams has progressed slowly. They illustrate retarded stream develop- ment, or a long-continued youthful stage. In this respect they afford a marked contrast to the Colorado, where the opportunities for development have been unusually great. There is still another reason for the slow development of the streams flowing to Hudson Bay, although its full signifi- cance has not been determined. That is, the region toward which they flow is believed to be a rising area. The upward movement of the land is slow, although the rate is not known. An elevation of a very few inches a century would have a decided effect on the flow of the streams in a region of such mild relief. A glance at a map of North America suggests that a large number of islands into which the land is broken on the north-eastern border of the continent is due to a recent sub- sidence. There is geological evidence that over a vast area at the north, the land was depressed during the Glacial epoch and has since been slowly rising, but has not regained 292 RIVERS OF NORTH AMERICA the position it held previous to the birth of the ice-sheets which once covered it. This re-elevation is thought to be still in progress, and should this conclusion be maintained it will furnish an additional reason for the present immature condition of the northward-flowing streams just referred to. Rivers Flowing to Fresh-Water Seas. — The St. Lawrence drainage slope, with its great lakes and magnificent rivers, affords numerous features of interest to the geographer be- sides its beautiful scenery. Soundings made in the Gulf of St. Lawrence, and even well to the eastward of the most easterly cape of Nova Scotia, have revealed the fact that the submerged channel of the St. Lawrence River may be traced on the floor of the ocean as far as the submarine es- carpment marking the true continental border. From the eastern extremity of this submerged channel through the Gulf of St. Lawrence and up the narrowing estuary to near Montreal, where the river at present meets tide-water, is more than a thousand miles. The Saguenay River, bordered by towering walls, occupies a canyon excavated by a branch of the Greater St. Lawrence. The same conditions are recorded in a less marked way by the Ottawa and other branches of the present river. We have here the most remarkable example of a drowned river-system that is known. The marginal portion of the continent with broad valley near the sea, leading inland to deep canyons, has been depressed in recent times so as to allow the sea to en- croach on the land. The valleys have become gulfs, bays, and estuaries, and the canyons narrow tideways ; highlands, that separated the former river valleys, when not completely submerged have been transformed into capes and head- SOME CHARACTERISTICS OF AMERICAN RIVERS 293 lands, and in part surrounded by the sea so as to form islands. The influence of the geographical history of the St. Law- rence on the course of human events is even more strongly marked than in the case of similar changes along the Atlantic border to the southward. The estuary of the St. Lawrence furnished an easy passageway, reaching far inland, for early explorers, and its connection, by means of the unsubmerged portion of the ancient river, with the Great Lakes tempted the Jesuit missionaries to make bold canoe journeys into the very heart of the continent. This same route led to the establishment of missions and the planting of white settle- ments and trading stations on the shore of the Great Lakes and in the Mississippi valley before the passes in the Ap- palachians to the southward became known. In later years, a series of canals to facilitate navigation between the St. Lawrence estuary and the Great Lakes stimulated industry by bringing tens of thousands of square miles of forest and of rich agricultural land into communication with the markets of Europe. Far-reaching plans for establishing deep waterways along this general course of early canoe navigation are now being matured, and the influence of geographical conditions favourable to commerce will be felt still more potently in the future than they have been in the past.* To the south of the St. Lawrence estuary lies the charm- ing valley of Lake Champlain, which was excavated by a stream tributary to the Greater St. Lawrence when the land 'I. C. Russell, "Geography of the Laurentian Basin," in Bulletin of the American Geographical Society ^ vol. xxx., pp. 226-254, 1898. 294 RIVERS OF NORTH AMERICA stood higher than now. After acquiring about its present form, the Champlain valley was depressed and became an arm of the sea, which was inhabited by marine mollusks and frequented by whales. A tideway reaching southward connected with the submerged Hudson River valley, making New England an island. A partial re-elevation of the land caused the former gulf to be separated from the ocean, so as to form a saline lake. The rains and feeding streams furnished a supply of fresh water in excess of the amount lost by evaporation, and the salt waters were flooded out and the present stage in the history of the valley initiated. This marvellous transformation of a broad and well-de- veloped river valley to an arm of the sea, to a saline lake, and then to a fresh lake, in which the blue Adirondack Hills and the equally picturesque mountains of Vermont are re- flected, is one of the most instructive pages in the later geographical history of America. The story of the St. Lawrence valley and its tributary branches is supplemented by the no less instructive history of the basins of the Great Lakes, some account of which has been given in a companion to the present volume.* The student of river development and of the changes made by streams in the topography of the land, as he sails the Great Lakes and visits the thriving cities on their shores, sees records of a time when rivers flowed through the now water-filled basins, and for ages worked slowly at their ap- pointed task of deepening and widening their valleys. This great task, when far advanced, was more than once inter- rupted by the invasion of the entire St. Lawrence region by ' I. C. Russell, Lakes of North Auicrica, Ginn & Co., Boston, 1895. SOME CHARACTERISTICS OF AMERICAN RIVERS 295 glaciers from the north. Conditions now characteristic of central Greenland then prevailed where millions of homes are now situated, and where fruitful farms have replaced the desolation of ice-fields. When the great geological winter had passed, the former stream channels were clogged with debris, so as to retard the waters and cause them to choose new courses. Elevation and depression of the land over tens of thousands of square miles still further complicated the difficulties that the re-born streams had to contend with. The surface waters were held in check, and formed vast lakes in the partially obstructed and warped and deformed preglacial river valleys. When one marshals in fancy the changes that the St. Lawrence drainage slope has passed through from the time when it was more elevated than now and supplied a well- developed river-system, — that is, a river with many branches, which had cut down its channel nearly to sea-level, so as to have a low gradient for two thousand miles or more, and had broadened its valley so as to form a wide, open plain which extended far into the many tributary valleys, — through the marvellous changes incident to the glacial invasions, and the partial submergence beneath the sea, and the partial re-elevation of the drowned portion, the damming of the stream by glacial debris, and the changes due to warping of the earth's crust, the brief time that civilised man has been acquainted with the region becomes insignificant. In this hasty outline of a million or more years of geographical history, although seemingly crowded with important events, all of the changes experienced by the St. Lawrence drainage slope have not been included. There is evidence that the 296 RIVERS OF NORTH AMERICA St. Lawrence basin has been in communication with the Mississippi River drainage. In the development of these two great river-systems, there has been a struggle for the possession of the land where they approach each other (analogous in some ways to the wars of the French and English for the possession of the same territory), which will be of interest to the reader who has followed the discussion of the backward cutting of drainage lines in a preceding- chapter. More than this, there are suggestions that the excavation of the basins of the Great Lakes was due in part to streams flowing southward instead of eastward, and that a change in direction was caused by movements in the earth's crust which are still in progress. The student of geography thus finds two chief lines of interest in the region under consideration, one dealing with the origin and history of the land forms, and the other with their bearing and in^ fluence on the current of human events. Niagara. — The lakes that first came into existence in the Laurentian basin during the final retreat of the glaciers were small and numerous. Many of them were short-lived, and were drained as the ice-dam retaining them withdrew north-eastward ; but some of them expanded with the retreat of the ice, and became vast inland seas, larger than any of their present representatives. At a late stage in the melt- ing of the glaciers, the basins now occupied by Lakes Erie and Ontario were occupied by a single great water-body. When the ice withdrew still more and the Mohawk valley was uncovered, a lower outlet became available and the waters escaped so as to lower the lake and cause it to be divided. The lake in the Erie basin overflowed across the dividing SOME CHARACTERISTICS OF AMERICAN RIVERS 297 land to the Ontario basin, and Niagara River was born. An example of the manner in which a river may originate, not previously considered in this book, is thus furnished. The many happy tourists who have listened to the thunder of mighty Niagara and wandered along the brink of the gorge occupied by the waters in their mad course below the cataract, have many illustrations thrust upon their atten- tion of the manner in which streams modify the land. The cataract was once at the margins of the bold escarpment near Lewiston, and has slowly receded, leaving a great gorge as a record of its work. Unlike the magnificent canyon of the Colorado, or the almost equally remarkable example through which Snake River flows, the steep-walled gorge of the Niagara has not been worn out by the flow of silt- and sand-laden waters, but as already described in discussing the migration of waterfalls, illustrates another process by which the land may be deeply trenched. The waters of Niagara come directly from a great lake in which they have left all of the sediment they may once have held in suspension, and are clear. The deep tourmaline-green of the plunging cata- ract is never clouded. Clear streams, as we are aware, have but little power to deepen their channels, as all the debris available for transportation is soon removed, and the chem- ical action of the waters in dissolving the rock over which they flow is so slow that only in exceptional instances are they able to deepen their channels more rapidly than the adjacent surface is lowered by weathering. How, then, has the canyon below the Falls of Niagara been excavated ? The energy of Niagara River available for canyon cutting is concentrated at the base of the cataract. The river came 298 RIVERS OF NORTH AMERICA into existence with a cataract, which was even grander when it first leaped from the crest of the escarpment at Lewiston than it has been since the first white man kneeled in its awful presence. Owing to the southward dip of the rocks, the height of the fall has been continually decreasing and will continue to decrease until it has receded to the lake from which the river flows, or the westward tilting of the land, known to be in progress, diverts its waters. The sur- face rock along Niagara River is a hard limestone about eighty feet thick; beneath this there are soft shales, much broken by joints and easy of removal. The dash of the spray, the grinding of ice-blocks, and, to some extent, -the freezing of absorbed water, leads to the removal of the shale so as to leave the limestone above projecting. From tim^e to time masses of the limestone break off and fall into the pool below. When the plunging waters have sufficient power to move these blocks, they are dashed against the cliff and act as millstones in deepening and widening the basin below. When the descending waters do not have sufficient power to sweep about the blocks of stone, as in the case of the American Fall, they accumulate and form a talus slope which protects the cliffs, and retard their seces- sion. This explanation, first offered by Gilbert, furnishes a reason for the marked differences between the American and Canadian portions of the cataract. Many other inter- esting and instructive features of Niagara are described and explained in the monograph just referred to. Retrospect. — We have made this rapid review of the prin- cipal drainage slopes of North America for the purpose of refreshing our memories concerning the more pronounced SOME CHARACTERISTICS OF AMERICAN RIVERS 299 geographical features of the continent due to erosion. Another aim has been to suggest questions which the student of geography will find pleasure in answering. I fear, how- ever, our hasty journey has in some respects left an errone- ous impression on the reader's mind, for the reason that in considering each drainage slope only its more pronounced features have claimed attention. If each river appears to be principally and essentially different from all other streams, and the several drainage systems present an infinite variety of disconnected facts, modifications and corrections of such ideas are necessary. A more detailed study of the behaviour of streams will show that law and order prevail. From the purling rill to the majestic river, where at first, perhaps, endless variety appears, the flow of water and the changes it produces in the relief of the land are governed by inflexible laws. The streams one and all are engaged in a definite and well-circumscribed task, which leads to an orderly succession of topographic forms. When all of the modifying conditions are taken into account, the successive changes experienced by a given land area, from the time of its upraising above the sea, to the time when it is worn down nearly to sea-level once more, are seen to be as much in obedience to law as the seasonal changes in a landscape or the development of an individual man from childhood to old age. Although the origin of topographic forms and the many metamorphoses they undergo, claim the special attention of students of geography, the fact should be borne in mind that the knowledge thus gained is but the basis of a more profound study, — the relation of man to nature. The in- 300 RIVERS OF NORTH AMERICA fluence of the earth's history on human history, although in many instances not fully realised, is an underlying and ever- present source of interest and enjoyment to the geologist and geographer. The brief review of some of the characteristics of rivers given in this chapter, it is hoped, will stimulate a desire, es- pecially in American students, to know more of the many and varied charms of their native land. CHAPTER IX THE LIFE HISTORY OF A RIVER AN application of the laws governing the behaviour of streams in interpreting the origin and history of topo- graphic forms can be made in almost any land area on the earth. In order to group in a single panorama, however, all of the various phases which a river passes through from its birth and youth to its old age and death, the conditions presented by many streams in various stages of growth and decline have to be combined, for the reason that the life of a man is too brief to enable him to observe more than a few minor changes in the history of a single river. But know- ing the laws which govern stream development, one can easily picture in his mind the leading events in the life of a majestic river whose murmurs we may be pardoned for fancying make audible the memoir of a million years. In order to sketch in outline the life history of an ideal river, let the reader imagine that the floor of the sea in temperate latitudes over an area of a hundred square miles has been upraised so as to form an island ; and trace the changes which will follow as the rain-water falls on its sur- face, and gathers into rills which unite one with another until 301 302 RIVERS OF NORTH AMERICA a series of rivers conducts the contributions from the clouds down their shining courses to the sea. The surface of our imaginary island is mildly irregular. In the central portion it has an elevation of a thousand feet, and slopes gradually but somewhat irregularly in all direc- tions to the sea. In places the waters gather in hollows and form lakes. These consequent lakes are soon filled, or their overflowing waters cut notches in the rims of their basins, and they are drained. The first streams that are born of the showers, like the children of men, have their courses marked out for them in early life, or, in more prosaic language, are consequent streams. Later in life they carve out their own fortunes and influence their surroundings. In these fireside fancies we assume the point of \'iew granted the novelist, to whom time and distance offer no limitation. A Scott or a Hawthorne tells us with confidence the most secret thoughts of a prisoner in his cell a century before they themselves were born. We accept the illusion so long as the laws governing human nature are not violated. Why should a similar privilege be denied the geographer ? Let us, then, trace the changes that our island will undergo in obedience to the laws of the inanimate world,' accepting the remark of Lamarck applied to the development of species, that ** time is nothing." The only supernatural condition which I will ask the reader to accept, is that the promontory on which we keep our vigil remains unchanged. Looking across the shimmering sea of fancy, v/e see the new-born consequent streams appearing like shining threads of silver when the skies are clear, but when the rain descends in torrents and the soil is loosened and disturbed they be- THE LIFE HISTORY OF A RIVER 303 come yellow with sediment. Already changes are in pro- gress. The streams charged with silt and sand are corrading their channels. The lines thus produced are delicate at first, but soon become more and more deeply engraved. These infant streams have their sources not at the summits of the island, but in general midway down its sides. The deep- ening of the channels leading from the higher portions of the island to the sea makes the waters flowing down them master streams. As they sink deeper and deeper into the rocks, lateral streams are developed on the original inter- stream areas. These branches become swifter as the main streams deepen their channels, and in turn develop branches to which they themselves are masters. The secondary and tertiary branches cannot excavate below the level of the stream to which they contribute their waters, but the down- cutting at their mouths may keep pace with the lowering of the receiving channel. This process of throwing out new branches and the growth of each branch by terminal bud- ding, as it were, soon lead to the complete drainage of the land. Water falling on any portion of the island finds a system of channels, delicately adjusted in size in accord with the part they have to play, v/hich lead it back to the sea. Our island, we will assume for simplicity, is composed of nearly horizontally bedded rock of various degrees of hard- ness. The influence of the dip of the beds beneath the original surface, discussed in a previous page in connection with the adjustment of subsequent streams and the develop- ment of drainage systems under the conditions there de- scribed, need not be repeated. The headward growth of the feeding rills and brooks of 304 RIVERS OF NORTH AMERICA the main consequent streams brings them into rivalry with each other. The boundaries between opposite-flowing streams in the central portion of the island become more and more sharply defined, and the positions of the divides can be easily traced. From these divides the descent into the valleys on either side is steep. Hard layers in the nearly horizontal beds cause many cascades. The young, joyous streams fill the air with laughter. The cascades came into existence low down the course of the streams and gradually retreated toward the centre of the uplift, leaving shadowy gorges as records of their migrations. It is only when the streams in their lower courses have deepened their channels nearly to sea-level that they cease to be w^hitened by cataracts and rapids. As we watch the growth of the streams we note that they deepen their channels most rapidly not at their mouths, nor at their sources, but at some locality between, which differs in its relative position, in various instances, with the size of the stream. The rate at which the streams we are observ- ing entrench themselves depends, as we know, on their volume, the declivity of their channels, and on the amount and character of the loads they carry. It is the resultant of these main conditions, rock texture being essentially the same throughout their courses, which determines at what locality conspicuous changes will first appear. Near their sources the grade is steep, but the waters are divided, flow- ing in numerous channels, and the work they are enabled to accomplish is not so great as farther down the slopes where many tributaries have united their energies. We need not again consider the various elements of a stream energy, THE LIFE HISTORY OF A RIVER 305 but from the fact that the channels through which they flow become conspicuously deeper midway up the slopes of the island, it is evident the most rapid corrasion is there taking place. Below the locality of most rapid corrasion the slope is less precipitous, and although the volume of water is greater, the rate at which the streams corrade decreases all the way to the sea. The amount of rock that has to be removed in order to admit of the sinking of the channels to base- level, however, is less and less the nearer they approach the coast-line. The streams near their mouths are thus enabled to reach the downward limit of their task sooner than at any locality higher up their courses, in spite of the fact that they there work more slowly than elsewhere, except perhaps at their extreme head-waters. Whatever the conditions, it is evident that any portion of a stream at a distance from its mouth cannot be lowered to baselevel more quickly than the portion nearer the sea, unless possibly by solution, as the material removed would in such a case have to be carried up instead of down a gradient. As we watch the changes in progress, we note that after the first adjustment to inherited conditions is made, all of the material removed where the stream beds are steep, is not carried directly to the sea. At first, perhaps, the slopes were such that the debris contributed to the master streams could be carried all the way to their mouths, but such an adjustment of gradient to load at the start would be of the nature of chance. The probabilities of a stream's inheriting a gradient perfectly adapted to its needs are almost in- finitely small. Throughout the life of a stream, even 306 RIVERS OF NORTH AMERICA though external conditions remain unchanged, there is a con- stant process of adjustment of gradient to suit the changing conditions due to corrasion and sedimentation in various por- tions of its channel, and also to variations in volume and load. This process of adjusting the gradient of a stream channel in its several parts to particular conditions of volume and load, is so delicate that no two of the streams we are watching will carry on their work in precisely the same way. In most instances, the debris removed midway- down the course of a stream, where corrasion is most active, will in part be deposited lower down and aggrading begin. Other streams will deepen their channels at their mouths to baselevel, and then begin to broaden their valleys and spread out flood- plains. As the streams grow older, the portions of their courses where corrasion is in progress will slowly recede up stream and be followed, at least for a time, by an extension in the same direction of the increasing flood-plains. Clouds gather about our island from time to time, and it experiences all the vicissitudes of climate entailed by the position it occupies on the earth's surface. Vegetation springs into existence, and the land is clothed with grasses and flowers, or deeply shadowed by forests. The length of our vigil" is so great that possibly the character of the flora undergoes many variations owing to climatic changes. Al- though these modifications in conditions vary the lives of the streams, they do not stop their w^ork. When the streams have deepened their channel where they approach the sea nearly or quite to baselevel, vertical corrasion ceases and is followed by aggrading, while lateral corrasion continues. THE LIFE HISTORY OF A RIVER 307 If we select one of the several larger consequent streams for special study, we find during the earlier stages of its life, that debris is continually being supplied by its swift upper branches in excess of the amount the sluggish current in its main trunk can carry away. In consequence, the flood- plain downstream, from the localities where corrasion is in progress, is built higher and higher. During high-water stages accompanying heavy rains, the stream meanders at will over the flat bottom it has given to its valley, and divides into many branches. The position of the stream is unstable, for the reason that abundant deposition of debris raises its bottom and borders, thus elevating it above the adjacent areas. When floods occur, the stream breaks through its levees and chooses a new channel, which is built upon as before, and the process repeated. At this stage in its history the stream to which we have directed special attention will have many high-grade branches, in which corrasion is in active progress, and a low-grade trunk portion where debris is being deposited. When the stream is corrading, the valleys or gorges are steep-sided and present V-shaped cross-profiles, but below the region of waste where a flood-plain is being formed, the valley is wide, essentially flat-bottomed, and has flaring sides. It is to be noted also that in the alluvial-filled valley there are no terraces. During this still youthful stage the trunk stream has many curves, and divides into several branches during floods, so as to enclose low% sandy islands. These changes in the position of the main channel are ifregular, and frequently rapid. The grading up of the main valley in the manner just 308 RIVERS OF NORTH AMERICA noted, was necessitated by the high grade of its tribu- taries. The aim, we may say, was to make an approxima- tion to the easiest attainable pathway for the debris on its journey to the sea. Theoretically, such a pathway would be what is known as the '' curve of quickest descent. '' But the down-cutting of the stream channels in their upper courses continually changes the conditions, and a continual process of adjustment in the lower and flatter portions of the curve is thus necessitated. As the grade of the tribu- tary streams is thus reduced, the region of previous aggrad- ing must also be modified. Hence there comes a time when the stream in its trunk portion begins to excavate a channel through its previously formed flood-plain. In this stage of adjustment, centuries being numbered as hours, we see the river writhing in its course through its more level tract; now cutting away the rocks on one side of its valley and then swinging bodily across its flood-plain and attacking the opposite bluff. Each of these migrations is accompanied by a multitude of minor contortions. Its course is always serpentine. On each of the minor bends we note that the immediate river bank is steepest on the concave side of the curve made by the stream, while on the opposite or convex side the bank slopes gently upward to the level of the flood-plain. The stream is plainly at work in removing material from the concave, and making additions to the convex, side of each curve. It soon becomes appar- ent that the stream is working over the material forming at least the surface portion of its flood-plain. With each dis- turbance of the detritus previously deposited, it is carried farther on its way to the sea, but each journey is short for THE LIFE HISTORY OF A RIVER 309 all but the very finest material, which is taken in suspen- sion and may be borne to the sea with but short rests on the bed of the stream. This process, as we know, leads to an assorting of the debris of which the flood-plain is formed : the coarsest portions are dropped first, and on them finer and finer sediment is laid down, the last addition to the plain being the finest of all. With each period of rest in the flood-plain, the debris undergoes chemical changes, and is more or less affected by frost and variations of tem- perature, which softens and weakens it so as to favour more rapid wear when next it is removed. The increase in the curvature of the minor bends of the meandering stream leads from time to time to the cutting through of the neck of land between two adjacent curves, and the straightening of the contorted channel. A shorter course is thus made for the waters, which is followed by re- adjustment of grade both up and down stream, and the former abrupt curve is left as a bayou, the entrance and exit to which soon become closed and an '* ox-bow lake ** is formed. A single migration of the river across its flood-plain re- quires thousands of years. But during this time its channel sinks deeper and deeper into the previously deposited debris, and an entire migration from one side of the valley to the other and back again is not always completed before a second swing is begun. The portion of the flood-plain not worked over during one of these incomplete migrations re- mains as a terrace. With each migration a flood-plain is spread out, and wherever a flood-plain formed during the preceding migration is not completely worked over, a terrace 'is left as a record of the unfinished task. 3IO RIVERS OF NORTH AMERICA The branches of the river have now cut down their chan- nels so as to have a comparatively low grade, except at their extreme head-waters, and the trunk of the drainage-tree is contorted and bordered by alluvial terraces. While the changes outlined above have been in progress, the fine debris carried by the river to its mouth has been deposited so as to form a low-grade delta, which makes an addition to the land, thereby increasing the distance to which the river has to carry its load before it can finally lay it aside. Even the depositing of debris in a delta, however, can scarcely terminate the influence of the river upon it, as the surface of the delta is but a continuation of the flood- plain, and future adjustments to ever-changing conditions may necessitate its removal and re-deposition in a later seaward extension of the land. / The growth of the delta, by increasing the length of the river, necessitates that its gradient throughout the portion where flood-plains occur should be raised in order to facili- tate the transportation of fresh debris over it. A continual check is thus placed on the process of down-cutting in the alluvial filling of the valley, necessitated by the constantly decreasing grade of the corrading tributaries. This con- tinual re-adjustment of the gradients throughout a drainage system, be it a meadow brook or the Mississippi, with ever- changing conditions of corrasion and sedimentation, is one of thousands of illustrations of the harmony of Nature. It was the result of this process of continual adjustment which forcibly impressed Hutton and Play fair, nearly a century ago, as is shown by the passage quoted on a prefatory page of this book. THE LIFE HISTORY OF A RIVER 31 1 During the centuries that have passed while we have been considering the work of the streams flowing from the island before us, wonderful changes have taken place in the topo- graphy of its surface. The sinking of the stream channels, but partially counteracted by aggrading, has left ridges be- tween the drainage lines. The land has been roughened by the cutting of valleys and canyons. The degree of this roughening depends mainly on the altitude of the land, the stage of development reached by the streams in various por- tions of their courses, and the amount the land has been lowered by general erosion. The streams are yet young, but, as already noted, have advanced farthest with their task of removing the rocks down to sea-level in their sea- ward portions. In the regions of low relief, near the sea, the streams have already passed the youthful stage, and are subduing the landscape not only by lateral corrasion but by •deposition. In this portion of the island, also, vertical cor- rasion having ceased, the tendency of weathering to reduce the interstream areas to the level of the adjacent valleys is no longer counteracted. In the higher portions of the island the divides between the main streams and between neigh- bouring branches of the same trunk drainage-line, are sharp- crested ridges. There is a wonderful and beautiful system displayed by these ridges. They are not level-topped, but marked by peaks and downward-curving saddles. On each ridge, whether a main divide or the crest of a branching spur, where two streams head against each other, there is a sag or saddle in its crest-line, and, where lateral spurs or sec- ondary or tertiary ridges join the main divides or a second- ary ridge, there are peaks or rounded knobs. The upland 312 RIVERS OF NORTH AMERICA with its multitude of crest-lines and many peaks resembles a vast tent supported by many poles. The place of each sup- port is marked by a peak, rounded and given a convex curvature by weathering, and between the peaks the tent- cloth descends in gracefully curving folds. The branches of the stream are most numerous where the original slopes were steep ; the interstream areas are there narrow and sharp crested ; farther toward the sea, where the slopes are more gentle, the interstream areas are broader and flatter. There are two principal reasons for these dif- ferences. On the steeper slopes the run-off is greater in proportion to the total rain-fall than on lower slopes, because the less the slope the longer the waters are retained on the surface and the greater the loss from evaporation and per- colation. A more important result, perhaps, is that on steep slopes corrasion is rapid in proportion to weathering and general degradation, while with progressively decreas- ing declivity weathering more and more nearly keeps pace with corrasion. The ridges between adjacent streams not only lose their even crests as determined by the original slope of the land, with the progress of stream sculpturing, but if we look down on them from a point vertically above, it is apparent that they are sinuous lines. The eating back into the up- lands of opposite-flowing streams has not been uniform, but the divides have been pushed one way or another according to the rate at which the rival streams have been enabled to progress with their work. The divides migrate toward the weaker streams. The ridges forming both the main and secondary divides are sharp and steep-sided where opposing THE LIFE HISTORY OF A RIVER 3 13 streams have eaten back the farthest, and are broader and have more gentle slopes where diverging ridges meet. The hills as well as the valleys are ever changing and record the workings of laws which produce infinite variety with the constant preservation of harmony and beauty. At a more advanced stage in the history of our island, the peaks standing at the junction of lateral ridges become more and more prominent as the saddles between them are deepened. Those well down the general slope in time be- come so far isolated that they stand as individual eminences, and their connection with the central system is at the first glance not apparent. The island has now reached its greatest topographic diver- sity, and, unless the orderly progression is disturbed, as by renewed elevation, for example, future changes will be in the direction of subduing its relief and smoothing out its contours. The process of broadening the valleys in their lower courses is extended farther and farther toward their sources. Broad flood-plains in time reach well into the central group of hills. In the uplands, valley-deepening continues, but the ratio of corrasion to general erosion becomes less and less with de- crease in the gradients of the streams. With decrease in eleva- tion, general erosion or degradation also diminishes, and both corrasion and degradation cease when baselevel is reached. As the gradient of the streams in their upper courses diminishes, the loads they carry to lower tracts also become less, and the streams are enabled to cut still deeper into their previously formed flood-plains. With the advance of old age the gradients of the streams become less and less 314 RIVERS OF NORTH AMERICA throughout their lengths, but are always steeper near their sources than at any locaHty farther downstream. Our island now consists of two portions in reference to topographic development : a central region of prominent peaks and ridges, and a surrounding baselevel plain, covered with a sheet of stream-deposited debris. If at this stage a comparatively sudden elevation of the whole island occurs, which carries it up, we will assume, one hundred feet, conspicuous changes will follow. The gradi- ents of the streams at their mouths will be increased. Some of them may plunge into the sea over escarpments, thus forming cascades, which will recede up stream, leaving sharply cut ravines. The gradients of the streams will be re-adjusted to meet the requirements of the changed condi- tions. Their revived energy will enable them to corrade throughout their length and to quickly deepen their chan- nels, especially in the previously alluvial-filled valleys. This rapid deepening will lead to the abandoning of portions of previously occupied flood-plains, and terraces will appear. The streams flowing through their new steep-sided channels will at first retain the positions they chanced to have at the time of the uplift, and will be sinuous, but their increased energy will tend to straighten their courses. The terraces left by the sinking of the streams will be broad in the coastal plain and become narrower and narrower farther inland. At the same time the vertical elevation of each terrace above the adjacent portion of the new stream channel will decrease from the sea margin inland to where the stream it borders ceases to be a depositing stream and the tract where corra- sion is in progress is reached. THE LIFE HISTORY OF A RIVER 315 At this stage in the history of the island, its marginal tract is an upraised peneplain, surrounding a group of hills, or monadnocks. The streams again broaden their valleys in their lower courses, the broadening terraces are removed, and this change sweeps inland until perhaps all records of the once conspicuous peneplain disappear. While we are pondering on the changes produced by slowly acting forces when the time limit is ample, our island, in obedience to unseen causes deep within the earth, is depressed two hundred feet. The sea encroaches on the land, and, extending far up the valleys, converts them into estuaries. The highlands between the valleys become capes, and possibly some of the outer members of the central group of hills are entirely surrounded by water and are transformed into islands. The coast-line, which previous to the subsidence, was conspicuously regular and formed long, sweeping curves, is now markedly irregular. The sea-water, extending far up the deeper and more thoroughly developed valley, forms long, narrow estuaries or downward valleys^ of the type of the Hudson estuary. In some instances the drowning of the valleys has extended above where the lower branches of the former stream joined the main trunk, and the lower courses of the tributary valleys are now bays on the side of the main estuary. These are estuaries of the Chesapeake type. In such instances the trunks of the former river systems have disappeared, and only their dismembered branches remain. The drainage-trees have been betrunked by subsidence. At this stage of greatest geographical diversity, so far as the relations of sea 3l6 RIVERS OF NORTH AMERICA and land are involved, the shores of the island are indented by numerous bays, many of them having flat alluvial lands at their heads. In fancy we have clothed our island with a varied flora. The picture presents a pleasing grouping of swelling hills with rounded summits and gradually sweeping sides, wide valleys with gently sloping borders, separated in part by broad but yet well-drained plains. These various forms are modulated and their details concealed beneath a living man- tle of vegetation. Seasonal changes, recurring like the fig- ures in a dance of merry children, come and go with the ebb and flow of the annual tide of temperature. Each springtime the willow-fringed brooksides blush w^ith the pulsations of renewed youth. Flowery banks and shadowy vistas in the forests reveal cool retreats in summer, when in the stillness of the evening we hear the distant mellow song of the wood-thrush. The deep, strong, harlequin colours of autumn make the island a garden of gorgeous flowers edged about by the silvery surf. In winter the babble of the brooks is hushed beneath icy coverings, and the bare trees are etch- ings on the white pages of the snow. These minor har- monies are interwoven all through the melody of the ages. Like the white fretwork on the waves of the sea, they ac- company the greater changes wrought by unseen agencies. We are overpowered by the multitude of questions suggested by this great drama of Nature. We long to know what the end may be. Why all this beauty and variety in form and colour ? Why the scented breeze, the hum of insects, the songs of birds, the music of the brooks, the coming and going of hills and valleys, and the thousands of other evi- THE LIFE HISTORY OF A RIVER 317 dences of harmoniously working laws ? Was the elevation of the land in a far distant time, the crumbling and decay of the rocks, the long journeys of the finer fragments down the streams, and their deposition in alluvial lands, but to form a soil in which violets and lilies might take root and furnish nectar for the bees ? We trace the origin and development of material things to intangible laws. These at first seem but the children of our own brains. We soon learn, how- ever, that they are not only external to ourselves, but sway and guide us. Man, too, gathers honey from the flowers. Such a mighty vision rises before the mind as we watch our island passing through its orderly transformations or glance upward at the changing constellations above it, that we pause, fearing to go farther, lest our fancy lead us astray. Our studies have brought us to the threshold of a vast temple: to explore it we must grope our way at first and laboriously gather facts to guide us in the same manner as in attempting to trace the life history of a river. We are recalled from dreamland by a new element in the scene before us. A canoe, buoyant and graceful, rounds a distant headland, traverses the belt of dark water just out- side the beating surf, enters one of the sheltered bays, and touches the shore. Dark men clad in skins step upon the beach. The light canoe is drawn part way out of the water. Soon a column of blue smoke rises above the tree-tops, spreads inland, and vanishes in the steady blow of the breeze from the sea. As time passes, other savages come to the island. Villages are built. Fires sweep through the forests leaving black ruin in their wakes. The soil is stripped of its natural covering, and for a time erosion is 3l8 RIVERS OF NORTH AMERICA accelerated. Large quantities of soil and other rock debris are washed down from the hills and encumber the more level lands below, destroying for a time their fertility. Generations of savages come and go, until a change of as great moment to them as was their coming to the animals and plants of the island takes place. For the first time a sail breaks the even sky-line of the sea. A Half-Moon borne proudly on by gentle breezes nears the island and enters one of its forest-fringed harbours. The changes which follow the coming of civilised man need not be dwelt upon. Chief among the events due to the greater w^ants of civilised than of savage men, is the removal of the forests. The land is cleared of its trees and shrubs. Other plants which grew beneath their shelter are extermin- ated. Ploughing greatly facilitates the work of the rills and rivulets. The precious layer of soil, thin at best, is more rapidly removed than formerly, and the sources of its re- newal to a great extent destroyed. In time, fields become too impoverished to repay cultivation while virgin lands can still be had on neighbouring shores for the taking, and are abandoned. Destruction follows apace. Gullies and deep canyon-like trenches are cut in the sides of the hills, and even greater desolation results in the plains below than fol- lowed the wild-fire started by the Indians. The island loses its archaic loveliness. Its flora is largely laid waste, or supplanted by the growth of seeds brought intentionally or by accident from other lands. The native birds and animals disappear, or, like the plants, are displaced by others and in part alien, species. The changes are so profound that they are felt not only throughout the fauna and flora, but THE LIFE HISTORY OF A RIVER 319 impress themselves on the topography of the land. No longer a source of immediate gain, the island is neglected and abandoned. The rivers with their increased freight due to the debris v^ashed from abandoned fields, progress more rapidly with their appointed tasks than before the forests were removed. Deltas are formed at the mouth of the streams, the estuaries are filled, and in time the waters of the sea are displaced, and broad grassy plains due to construction make their ap- pearance. The coast-line again becomes a series of sweep- ing curves. Tens of thousands of years elapse before the last of the conspicuous results due to elevation and subsi- dence become obliterated, and during this interval marked changes have taken place in the group of monadnocks forming the central highland of the island. Some of the original consequent streams were larger or had shorter courses to the sea than their competitors, and were enabled to develop more rapidly. The divides at the heads of the more energetic streams receded and new territory is added to their hydrographic basins. This process of cap- ture and diversion leads to still greater diversity in the topography. The struggle between streams for the posses- sion of territory in progress along every divide is not unlike the struggle for existence and the survival of the fittest in the animal world. The streams develop in accordance with their environment. Those most favoured capture the waters of their less favoured neighbours and wax stronger at the expense of the weak. In its old age our island loses the roughness of surface produced by stream corrasion. The ridges and peaks be- 320 RIVERS OF NORTH AMERICA come subdued and their outlines more flowing, and the valleys broader. In time low mounds alone remain to mark the site of once picturesque peaks, and in the broad valleys sluggish streams meandering in sweeping curves carry off the decreased water supply. Even the hills, after a pro- longed old age, disappear, and an undulating plain but slightly elevated above the encircling sea remains. A geographical cycle has run its course. The resulting peneplain has even less diversity than the surface of the new-born island. The streams, now flowing sluggishly on account of the lowering of their gradients, are too feeble to carry burdens, and run clear, but their chemical activity is undiminished, and they still bear invisible loads in solution. Chemical degradation, previously of minor importance in reference to the work of mechanical agencies, now becomes the more potent, and the final reduction of the land to sea- level is secured by the removal of its material in solution. The waves and currents of the sea have been active throughout this long history in producing changes which, however, are beyond the limits of the present discussion. After the streams ceased to bring debris to the shore, which, we may presume, either in part or wholly counter- acted the attacks of the sea on the land, the low coastal plains are washed away, and finally the waves roll over the site of the vanished island. INDEX yEolian corrasion, mention of, 2g Ages of terraces, relative, 170, 171 Aggrading, explanation of the term, 98 Alaska, characteristics of the rivers of, 284-289 Alluvial cones, description of, 101- 109 — fans, reference to, loi — rivers, characteristics of, 264, 265 Alsec River, Alaska, mention of, 284 Analyses of river-water, average, 80 — table of, 78 Analysis of rain-water, 75 Anchor ice, influence of, on stream transportation, 25-28 Anticlinals, influence of, on topo- graphy, 198 Appalachian Mountains, stream ad- justment in, 195-203 — rivers, brief account of, 260, 261 Arctic drainage slope briefly defined, 256, 257 Atlantic drainage slope briefly de- fined, 256 Babb, C. C, observations by, 74 Baselevel, definition of, 47 Baselevelling, discussion of, 46-50 Bear River, Wyoming, analysis of the water of, 78 Beheaded streams, explanation of the term, 191 Bering drainage slope briefly defined, 257 Big Bend of Columbia River, Wash- ington, mention of, 280 Big Wills Creek, Alabama, adjustment of, 208-213 Bischof, G., cited on water analyses, 79 Blatchley, W. S., reference to writ- ings of, 96 Bonneville, Lake, reference to deltas of, 125 Bottom loads of streams, 68-70 — terraces, origin and nature of, 166, 167 Branner, J. C, reference to work of, X., II Breccia, due to faulting, mention of, 4 Calcium carbonate, solubility of, 92 Call, R. E., reference to writings of, 94 Campbell, M. R., reference to work of, X. Canada, characteristics of the rivers of, 290-292 Canadian Geological Survey, reference to, xi. Canyon of Snake River, Washington, reference to, 160 — rivers, characteristics of, 271-275 Caribbean drainage slope briefly de- fined, 257 Carrollton, Mississippi, sediment in the Mississippi at, 71, 72 Cascade Mountains, reference to waterfalls of, 57, 61 — streams of, 280 Catskill Mountains, migration of di- vides on, 251-253 Cephalonia, Greece, reference to "sea-mills" of, 95 321 3^ INDEX Chamber! in, T. C, reference to the work of, X. Chandler, C. F., water analyses by, Characteristics of American rivers, 254-299 Chattanooga, Tenn., geography near, 208-214 Chelan, Lake, Washington, terraces near, 182 Chemical degradation, 82-84 — denudation, discussion of, 80, 81 — disintegration, discussion of, 6-1 1 Chesapeake Bay, map of, 219 Chipaway River, reference to, 138 Clarke, F. W., water analyses by, 78 Climate, influence of, on streams, 140- 142 Climatic changes, influence of, on streams, 223-233 — on terrace-making, 158-160 Colorado River, characteristics of, 271-275 — reference to, 133 — still currading, 45 Columbia River, characteristics of, 278-282 - - — terraces of, 180-182 Corrasion, discussion of, 28-36 — lateral, discussion of, 34-36 — stream, genera] process of, 142- 145 ^ Corthell, E. L., reference to writings of, 132 Crevasses, origin and nature of, 120 Crosby, F. W., reference to writings of, 95 — W. O., reference to writings of, 95 Croton River, New York, analysis of the water of, 78 — material carried in suspension by, Cumberland River, Tennessee, analy- sis of the water of, 78 Current terraces, origin and nature of, 167-169 Curves made by streams, 36-39 Danube River, data concerning, 74, 75 — material carried in suspension by, 79 Darton, N. H., cited on drainage of Catskill Mountains, 251 — reference to the work of x. Davis, W. M., cited on lakes, 122 — cited on peneplains, 48 — cited on stream development, 187 — references to the writings of, x., 41, 42, 214 — and J. VV. Wood, cited on super- imposed drainage, 243 Decay of rocks, i-ii Deflection of streams owing to the earth's rotation, 39-43 Degradation of the land, general rate of, 81-84 Delaware River, analysis of the water of, 78 Delta of the Yukon River, brief ac- count of, 288 — terraces, origin and nature of, 167-169 Deltas, origin and structure of, 123- 142 Deposition, stream, general process of, 142-145 Deposits made by streams, variations in the, 136-142 Development of streams, 63-66 Diller, J. S., reference to the work of, X. Discharge of the Mississippi, 267 Disintegration of rocks, i-ii Distributaries, explanation of the term, 103 Diverted streams, explanation of the term, 192 Divides, migration of, 247-253 Dodge, R. E., cited on terraces, 157, 158 Drainage slopes of North America briefly defined, 256, 257 Drew, F., reference to the writings of, lOI Drift-wood, influence of, on stream development, 240-244 Drowned rivers, examples of, 260 Dunes, influence of, on streams, 139 Dutton, C. E., reference to explora- tions by, X. Earth's rotation, influence of, on streams, 39-43 INDEX 323 Elevation, effects of, on stream devel- opment, 215-217 England, rate of land degradation in, Erosion, baselevel of, discussed, 46- 50 — general discussion of, 46-51 "Fall line" of the Atlantic coast, brief account of, 261 Fault breccia, mention of, 4 Fergusson, J., reference to the writ- ings of, 37 Ferrel, \V., cited on the rotation of the earth, 41 Flood-plains, origin and nature of, 1 10-116 Floods in rivers, brief account of, 229-233 Fluctuations of streams, discussion of, 229-233 Forshey, Professor, observations by, 70 Eraser River, British Columbia, char- acteristics of, 282-284 Ganges River, data concerning, 75 Geikie, A., references to the writings of, 18, 74 Genesee River, New York, analysis of the water of, 78 Geographical cycles, definition of, 49 Gilbert, G. K., cited on Niagara Falls, 59, 60 — explorations by, x. — reference to writings of, 31, 42, 125 Glacial corrasiou, mention of, 29 — meal, contnbution of, to streams, — terraces, origin and nature of, 169, 170 Glaciated lands, rivers of, 262, 263 Glaciers, influence of, on stream de- velopment, 234-236 Grand Coulee, Washington, mention of, 280 Great Basin, climatic condition of, 17S, 179 — drainage, brief account of, 257 Great Falls, Canada, mention of, 58 Great Lakes, rivers flowing to, 292- 3QO Great Plateaus, references to, 44, 45 Green River, Kentucky, references to, 88-90, 94 Ground ice, influence of, on stream transportation, 25-28 Gulf drainage slope briefly defined, 257 Hayes, C. W., cited on geography of Southern Appalachians, 208-214 — references to the work of, x., 207, 208 Hicks, L. E., cited on flood-plains, 118 — cited on profiles of streams, 146, 150 High Plateaus, reference to, 45 Hitchcock, E., cited on delta terraces, 167 Hoang Ho River, data concerning, 75 Holmes, W. H., cited on Colorado River, 273 Hosford, E. N., water analvsis by, 78 Hovey, H. C, references to the writ- ings of, 94, 96 Pludson Bay drainage slope briefly ) defined, 256 j Hudson River, New York, analysis of ' the water of, 78 — brief account of, 260 — cited as an example of a drowned river, 218 — — material carried in suspension by, 79 <-~ Humboldt River, Nevada, analysis of the water of, 78 Humphreys and Abbot, cited in ref- erence to the Mississippi, 70, 73, 253, 267 ; cited on drift-wood, .243 — cited on Mississippi delta, 132 Hunt, T. S., water analysis by, 78 Hydration, influence of, on rock dis- integration, 6 Ice, influence of, on stream transport- ation, 22-28 — weight of, 23 Indian Creek, California, reference to, I II Invisible load of streams, 75-81 324 INDEX Irrawaddy River, data concerning, 74 James River, Virginia, analysis of the water of, 78 Jones, W. J., water analysis by, 78 Jordan River, Utah, analysis of the water of, 78 Jukes-Browne, A. J., reference to the writings of, 18 Julian, A. A., references to the writ- ings of, II, 76 Keyes, C. R., cited on the meander- ings of streams, 113 Kittatinny peneplain, brief account of, 200, 206 Kowak River, Alaska, mention of, 286 Lake Pepin, Wisconsin- Minnesota, reference to, 138 — St. Clair, delta in, 133 — Tahoe, California-Nevada, analy- sis of the water of, 78 Lakes, climatic changes indicated by, 220 Laurentian Basin, rivers of, 292-300 Le Conte, J., references to the writ- ings of, 18, 19 Levees, natural, origin and nature of, I 16-123 Life history of a river, 301-320 Limestone, solubility of, 92 Loads of streams, how obtained, 13-16 Loew, O., water analysis by, 78 London, England, analysis of rain- water at, 75 Lookout Mountain, Tennessee-Ala- bama, drainage of, 208-214 Los Angeles River, California, analy- sis of the water of, 78 Lost rivers, reference to, 226 Lupton, N. T., water analysis by, 78 Luray Cavern, Virginia, references to, 93, 94 Lyell, C, reference to the writings of, 22 Mackenzie River, delta of, 133 Maine, coast topography of, 218 Mammoth Cave, Kentucky, brief ac- count of, 88, 89, 91, 94 Marsh, G. P., reference to the writings of, II Mason, W. P., references to the writ- ings of, 75, 76 Matapediac River, New Brunswick, anchor ice in, 25-27 Material in suspension, measures of, 70-75 Maumee River, Ohio, analysis of the water of, 78 Maxwell, W., reference to the writ- ings of, II McGee, W. J., reference to the work of, X. Meandering streams, discussion of, 36-3S Mechanical disintegration of rocks, 2-6 Merrill, G. P., cited on hydration, 6 — references to the writings of, 9, 11 Migration of divides, discussion of, 247-253 — of waterfalls, discussion of, 60-63 Mississippi River, analysis of the water of, 78 — characteristics of, 265-271 — Commission, reference to map by, 122 — data concerning, 74, 75 — delta of, 131 — influence of earth's rotation on, 42 — inundation of, 119-121 — material carried in suspension by, — rate of degradation in basin of, 83, 84 — sediment in waters of, 70-74 Missouri River, an aggrading stream, 43, 44 Mohawk River, New York, analysis of the water of, 78 Monadnock, definition of, 49 — Mount, New Hampshire, refer- ence to, 49 Montmorenci Falls, Canada, reference to, 58 Morrill, P., cited on the Mississippi, 267 — reference to the writings of, 121, 253 Moses Lake, Washington, reference to, 139 Moulins, mention of, 34 INDEX 325 Murray, J., cited on water analyses, So Natural Bridge, Virginia, reference to, 94 — levees, origin and nature of, 116- 123 Newberry, J. S., explorations by, x. New England rivers, characteristics of, 259, 260 New Orleans, Louisiana, depth of delta deposits at, 132 Niagara Falls, profile and section at, 60 — reference to, 62 Niagara River, characteristics of, 296- '298 Nile River, data concerning, 74, 79 Nita crevasse, Louisiana, an account of the, 121, 122 Ottawa River, (Canada, analysis of the water of, 78 — mention of, 292 Pacific drainage slope, brief account of, 257 Passaic River, New Jersey, analysis of the water of, 78 Peneplain, definition of, 48 Peneplains, ancient, in the Appalach- ians, 205-207 Platte River, Nebraska, an aggrading stream, 43, 44 Po River, Italy, data concerning, 74, Porcupine River, Alaska, ice-work on banks of, 24 Pot-holes, origin and nature of, 33, 34 Potomac River, Virginia, data con- cerning, 74 — rate of degradation by, 82 Powell, J. W., cited on baselevel, 47 — cited on moisture necessary for vegetation, 237 — cited on rapids in Colorado River, 138 — reference to explorations by, x. Precipitation, influence of variations in, on streams, 224-228 Profiles of streams, 145-151 Rain-water, impurities in, 75, 76 Reade, T. M., cited on chemical de- gradation, 82 Red River, Louisiana, lakes on the sides of, 122 Regolith, meaning of the term, 9 Reversed streams, explanation of, 192 Rhine River, material carried in sus- pension by, 79 Rhone River, data concerning, 74, 75, 79 Rio Grande, data concerning, 74 Rio Grande del Norte, analysis of the water of, 78 River piracy, discussion of, 203-205 Rocky Mountains, reference to water- falls of, 57 Rotation of the earth, influence of, on streams, 39-43 Roth, Professor, cited in water analy- sis, 79 Russell, L C., cited on glaciers, 236 — cited on Lauren tian basin, 293 — cited on Laurentian lakes, 294 — cited on terraces of the Columbia, 180-182 — references to the writings of, 11, 24, 123, 124, 134, 137, 174 — T., reference to the writings of, of, 253 Sacramento River, California, analysis of the water of, 78 St. Anthony P'alls, Minnesota, refer- ence to, 62 S't. Clair Lake, delta in, 133 St. Lawrence drainage slope briefly described, 256 St. Lawrence River, analysis of the water of, 78 — character of, 43 — submerged portion of, 218 Salisbury, R. D., reference to the writings of, x., 170 Schooley peneplain, reference to, 205 Screes, explanation of the term, 109 Shaler, N. S., cited on caverns, go Shenandoah peneplain, brief account of, 200 Shoshone Falls, Idaho, reference to, 62 Sierra Nevada, reference to water- falls of, 57 326 INDEX Sierra Nevada rivers, characteristics of, 275-278 Sink-holes, reference to, 93 Snake River, Idaho-Washington, char- acteristics of, 27g, 280 Snickers Gap, Virginia, character and origin of, 200, 201 Southern rivers, characteristics of, 263, 264 Stevenson, D., reference to the writ- ings of, 18 Stickine River, Alaska-Canada, men- tion of, 284 Stream conquest, discussion of, 203- 205 — deposition, discussion of, 97-151 — development, discussion of, 184- 195 Subimposed drainage, term suggested, 246 Subsequent streams, origin and nature of, 184, 185 Subsidence, effects of, on stream de- velopment, 217-221 Superimposed drainage, explanation of, 244-246 Synclinal mountains and anticlinal valleys, 207-214 Synclinals, influence of, on topo- graphy, 198 Tahoe Lake, California-Nevada, an- alysis of the water of, 78 Taku River, Alaska, mention of, 284 Talus slopes, origin and nature of, 109, no Tarr, R. S., cited on drowned rivers, 219 — cited on young valleys, 55 — reference to the work of, xi. Teanaway River, Washington, dam of drift-wood on, 243 Temperature, influence of variation in, on streams, 228, 229 Terraces, origin and nature of, 152- 183 Thames River, England, material carried in suspension by, 79 Thompson, W. C, cited on anchor ice, 25-27 Tides in Columbia River, 280 Todd, J. E., cited on the Mississippi, 270 Transportation by streams, discussion of, 14-28 Trenton Falls, New York, reference to, 58 Troy, New York, impurities in rain- water at, 76 Truckee River, Nevada, analysis of the water of, 78 Tundra of Arctic shores, brief account of, 133, 288 Underground streams, 84-96 United States Geological Survey, re- ference to work of, xi. Uruguay River^ data concerning, 74 Vegetation, influence of, on stream development, 236-244 Visible loads of streams, 67-75 Volcanic agencies, influence of, on stream development, 231-233 Volcanic dust, contributed to streams, 14 Von Hosen, J., table of analyses com- piled by, 78 Wales, rate of land degradation in, 83 Walker River, Nevada, analysis of the water of, 78 Walla Walla River, Washington, re- ference to, 137 Waller, E., water analysis by, 78 Water, weight of, 23 — analyses, table of, 78 Waterfalls, nature and history of, 54-62 Water-gaps, origin of, 199-205 Watkins Glen, New York, reference to, 58 Wheeler, W. H., reference to the writings of, 73 White River, Washingson, mention of, 287 Wilbur, E. M., cited on tides in Co- lumbia River, 2S0 Willis, B., cited on stream adjust- ment, 198 — reference to the work of, xi. — reference to the writings of, 200 Wills Creek, Alabama, adjustment of, 208-213 INDEX 327 Wind-gaps, explanation of the term, 199-205 Wurtz, H., water analysis by, 78 Wyandotte Cavern, Indiana, refer- ence to, 93 Yazoo River, Louisiana, reference to, 123 Young valleys, illustrations of, 55 Yukon River, ice-work on the banks of, 24 — drift-wood on, 242 The Science Series Edited by Professor J. McKeen Cattell, Columbia Uni- versity, with the cooperation of Frank Evers Beddard, F.R.S., in Great Britain. Each volume of the series will treat some department of science with reference to the most recent advances, and will be contributed by an author of acknowledged authority. Every effort will be made to maintain the standard set by the first volumes, until the series shall represent the more im- portant aspects of contemporary science. The advance of science has been so rapid, and its place in modern life has become so dominant, that it is needful to revise continually the statement of its results, and to put these in a form that is intelligible and attractive. The man of science can himself be a specialist in one department only, yet it is necessary for him to keep abreast of scientific progress in many directions. The results of modern science are of use in nearly every pro- fession and calling, and are an essential part of modern edacation and culture. A series of scientific books, such as has been planned, should be assured of a wide circulation, and should contribute greatly to the advance and diffusion of scientific knowledge. The volumes will be in octavo form, and will be fully illus- trated in so far as the subject-matter calls for illustrations. G. P. PUTNAM'S SONS, New York & London THE SCIENCE SERIES (Volumes ready, in press, and in preparation.) The Study of Man. By Professor A. C. Haddon, M.A., D.Sc, Royal College of Science, Dublin. Illustrated. The Groundwork of Science. A Study of Epistemology. By St. George Mivart, F.R.S. Rivers of North America. A Reading Lesson for Students of Geography and Geology. By Israel C. Russell, LL.D., Professor of Geology in the University of Michigan. Illustrated. Earth Sculpture. By Professor James Geikie, F.R.S., University of Edinburgh. Illustrated, The Stars. By Professor Simon Newcomb, U.S.N., Nautical Almanac Office, and Johns Hopkins University. Meteors and Comets. By Professor C. A. Young, Princeton University. The Measurement of the Earth. By Professor T. C. Mendenhall, Worcester Polytechnic Institute, formerly Superintendent of the U. S. Coast and Geodetic Survey. Volcanoes. By T. G. Bonne y, F.R.S., University College, London. Earthquakes. By Major C. E. Button, U.S.A. Physiography; The Forms of the Land. By Professor W. M. Davis, Harvard University. The History of Science. By C. S. Peirce. General Ethnography. By Professor Daniel G. Brinton, University of Pennsylvania. Recent Theories of Evolution. By J. Mark Baldwin, Princeton University. Whales. By F. E. Beddard, F.R.S., Zoological Society, London. The Reproduction of Living Beings. By Professor Marcus Hartog, Queen's College, Cork. Man and the Higher Apes. By Dr. A. Keith, F.R.C.S. Heredity. By J. Arthur Thompson, School of Medicine, Edinburgh. Life Areas of North America : A Study in the Distribution of Animals and Plants. By Dr. C. Hart Merriam, Chief of the Bio- logical Survey, U. S. Department of Agriculture. Age, Growth, Sex, and Death. By Professor Charles S. Minot, Harvard Medical School. Bacteria. Dr. J. H. Gladstone. History of Botany. Professor A. H. Green. Planetary Motion. G. W. Hill. Infection and Immunity. Geo. M. Sternberg, Surgeon-General U.S.A. G. P. PUTNAM'S SONS, New York & London * V * o . ^ V/. 0- X • *^ o^ ^. v\ t/' ^•^ '/• ^" "^^ V aV^' ">. <" ^/. \ \ ; o ^ . r .^^^^ y. -.^'^ ^ ^ % ..x^'' ■/ ^ '" " / .4* xXV-^. ■.,,.* .A i\ ■ .-.'■■■♦ • o .-X' ,\ 0' >" \"' -^^ V/ i: '■Jo « ;^'-' r.--' -• - '^ -/. ,,N .-^'^ \ ,\ - . . ,, ,/:. ^-"0^ '"'^V'^^-/^ >^ -^^fr?^ o. / , >^ xN^ -'^:r. ^^■ s- ^. A^ ^^A^ '■> ". '^'% >" .v'«. ^/^ ^'^ ^ ^ 0.\^ .^^\^.^ .^^ ,0 -/ .^*\ ■\^^ .^ .x-' ■■..,-. ■• -(-.. ,-ji j-- <^ V 'N^ '^^- >^^ .-,-^ •'■r: %.<^ ,\' ^> o "♦ •/ nX^^' ''^.. ■^•';^' •^^^^ . ^<> ,0' ■<>;^' ^ ' '^t' .-A v^'' ■•'■ O^ ^. S •'', ^•'•0^ -^^ A- •'^^<=.'^ ■\<^'' V V^ r:)-^vi^^^WKfrny)jm.yiisf^^^\K"' • ■-- ^. OF CONGRESS 708 3247 l!«Wt« '?v>''>''i^'§;-:fS!iii:; mmm'.. wm. irm'E'ji'-it -?i;!iiSI'''