IT is difficult, or almost impossible, even approximately to settle precisely what are the geological dates of the valleys through which many rivers run; or, in other words, when they first began to be scooped out, and through what various periods their excavation was intermittently or continuously carried on. No one has yet thoroughly analysed this subject, and only of late years have 1 begun clearly to see my way into it. Nevertheless a good deal has been done even now, and a great deal more will be accomplished when, with sufficient data, the whole subject may come to be investigated. In Wales, for example, there are vast numbers of rivers and brooks, small and large, and when we examine the relation of these streams to the present surface of the country, we often find it very remarkable. Fig. 97 is a diagram representing no particular section, but simply the general nature of the sections across the Lower Silurian strata of Cardiganshire, as shown by myself in a paper given to the British

[Plains of Marine Denudation. 497]

Association at Oxford in 1847. The dark-coloured part represents the form of the country given in the original sections on a scale of six inches to a mile horizontally and vertically. The strata of this area, and, indeed, of much of South Wales, are exceedingly contorted. The level of the sea is represented by the lower line; and if we take a straight-edge, and place it on the topmost part of the highest hill, and incline it gently seaward, it touches the top of each hill in succession, in the manner shown by the line b b. This line is as near as can be straight, and, on the average, has an inclination of from one to one and a half degrees; and it is a curious circumstance that in the original line of sections there were no peaks rising above that line—they barely touched it and no more. It occurred to me when 1 first observed this circumstance that, at a period of geological history of unknown date, perhaps older than the beginning of the deposition of the Permian and New Red Sandstones, this inclined line that touches the hill-tops must have represented a great plain of marine denudation.

FIG. 97

Atmospheric degradation, aided by sea waves on the cliffs by the shore, are the only  powers I know that can denude a country so as to shave it across, and make a plane surface either horizontal or slightly inclined. If a country be sinking very gradually, and the rate of waste by all causes be proportionate to the rate of sinking, this will greatly assist in the production of the phenomena we are now considering: and a little reflection will show,

[498 Plains of Marine Denudation.]

that the result would be an inclined plane like that of the straight line b b in the diagram. Let South Wales be such a country: then when that country was again raised out of the water, the streams made by its drainage immediately began to scoop out valleys; and though some inequalities of contour forming mere bays may have been begun by marine denudation during emergence, yet in the main I believe that the inequalities below the line b b have been made by the influence of rain and running water. Hence the number of deep valleys, many of them steepsided, that diversify Wales, all the way from the Towey in Caermarthenshire to the slaty hills near the southern flanks of Cader Idris and the Arans.

On ascending to the upper heights, indeed, anywhere between the Vale of Towey and Cardigan Bay, it is impossible not to be struck with the average uniformity of elevation of the flat-topped hills that form a principal feature of the country. The country already described as seen from Ramsey Island is part of this plain,
1 and much further north let anyone ascend Aran Mowddy or Cader Idris in Merionethshire, and look south and south-east. From thence he will behold, as far as the eye can reach, a wide extent of flat-topped hills, which form the relics of a vast tableland, now intersected by numerous rivers, which, in the long lapse of untold ages, have scooped out unnumbered labyrinthine valleys eastward into Montgomeryshire, and far south into Cardiganshire. Between the rivers Towey and Teifi, and in other areas, these hills, in fact, form the relics of a great plain or tableland in which the valleys have been scooped out; and in the case of the country represented in fig. 97, 'the higher land, as it now exists, is

1 See p. 487.

[Marine Denudation. Escarpments. 499]

only the relic of an average general gentle slope, represented by the straight line (b b) drawn from the inland heights towards the sea.
1 Mr. Jukes applied and extended the scope of the same kind of reasoning to the south of Ireland, with great success. In various parts of Europe, notably in those regions that have been longest above the water-on the banks of the Moselle and of the Rhine, and in the great coalfield west of the Appalachian chain in North America-we find unnumbered valleys intersecting tablelands, of a form that leads us to believe that they also have been made by the long-continued action of atmospheric waste and running waters; and I believe that the valleys of South Wales have been formed in the same way, and in their origin are even often of latest palæozoic dates.

Nothing is more remarkable in the history of rivers than the circumstance that very frequently they run straight through bold escarpments, which at first sight we might suppose ought to have barred the course of the streams.2 The Wye in South Wales, for example, runs through a bold escarpment of Old Red Sandstone hills; and the same is the case with the Usk.

For long it was customary to attribute such breaches in escarpments and indeed valleys in general, to disturbances and fractures of the strata, producing a wide separation, and actually making hills. But when we realise that thousands of feet of strata have often been removed by denudation since the great disturbances of the Welsh strata took place, it becomes clear that the present valleys are in no way immediately connected with them; for even if there be dislocations or faults

1 Reports, British Association, p. 66, 1847.
2 This has already been alluded to in the case of the rivers of the Wealden, pp. 108-119.

[500 Valleys and Dislocations.]

in some of the valleys, these faults when formed were, as far as regards the present surface, thousands of feet deep in the earth. All they could do might have been

FIG. 98.

a Present surface of the ground.
The dotted lines show the continuation of an anticlinal curve broken by a fault f.
The dotted lines above the surface a a represent a certain amount of strata removed by denudation.
to establish lines of weakness along which subsequent denudation may have excavated valleys.

The real explanation of such cases as those of the Wye and the Usk is this. At some period, now uncertain, the beds of the Old Red Sandstone, well seen in the escarpment of the Beacons of Brecon, a, and the Caermarthen Fans, once spread much farther westward, forming a great plain, b b (fig. 99), the result of earlier denudations. This plain sloped gently eastward, and the dotted line shows the general state of old outcrops of the strata. The river then ran over ground perhaps even higher than the tops of the hills of the present escarpment, and by degrees it cut itself a

[The Wye and the Usk. 501]

FIG. 99.
Diagram showing the mode of formation of the escarpment of
the Beacons of Brecon and
the Caermarthen Fans, South Wales.

bb, Original plain sloping gently eastward.
1, 2, 3, 4, 5 successive escarpments.

502 Escarpments and Rivers.]

channel approximately in its present course, but varied and widened by subsequent river action; and, as it cut out that valley, the escarpment, by the influence of rain and other atmospheric causes, gradually receded to the points marked 1, 2, 3, 4, 5, and a, the last being the present escarpment. For all observation tells us that escarpments of a certain kind work back in this way, that is to say, in the direction of the dip of the strata.

One reason of this is, that escarpments often partly consist of hard beds lying on softer strata. The softer strata are first more easily worn away along the line of strike, and thus an escarpment begins to be formed. Once established, the weather acting on the joints and other fissures in the rocks, takes more effect on the steep slope of the scarp than on the gentle slope that is inclined away from the scarp. The loosened detritus on the steeper slope slips readily downward, and is easily removed by floods of rain; and thus the escarpment constantly recedes in a given direction, while on the opposite gentle slope, the loosened detritus, smaller in amount, travels so slowly that it rather tends to block the way against further waste. In this way we can explain how the Wye and the Usk break through the Old Red Sandstone and find their way to the estuary of the Severn; why the Severn itself breaks through the Upper Silurian escarpment of Wenlock Edge; why certain other rivers—such as the Dee in Wales, and the Derwent in Cumberland—cut through escarpments of Carboniferous Limestone; and how, indeed, the same kind of phenomena are everywhere prevalent under similar circumstances. Of this I shall say more when I come to treat of the Oolitic and Cretaceous escarpments.

But when we have to consider the origin of some of

[The Severn. 503]

the larger river valleys, there is a great deal that is difficult to account for. One thing is certain, that before the Glacial epoch the greater contours of the country were much the same as they are now. The mountains of Scotland, Wales, and of Cumberland, and the great Pennine chain, existed then, somewhat different in outline, and yet the same essentially; the central plains of England were plains then, and the escarpments of the Chalk and Oolites existed before the Glacial period. All that the ice did was to modify the surface by degradation, to smooth its asperities by rounding and polishing them, to deepen valleys where glaciers flowed, and to scatter quantities of moraine-detritus, partly in the shape of boulder-clay and of marine boulder beds, and sands and gravels, over the plains that form the east of England, and the Lias and New Red Sandstone in the middle.

If we examine the valley of the Severn from Bristol northwards through Coalbrook Dale, we find that for a large part of its course the river runs down a broad valley, between the old Palaeozoic hills and the escarpment formed by the tableland of the Cotswold Hills which are highest in the neighbourhood of Cheltenham. That valley certainly existed before the Glacial epoch, because we find boulders and boulder-drift far down towards Tewkesbury; and therefore, I believe that before the Glacial epoch this part of the Severn ran very much in the same course that it does at present. During part of the Glacial epoch the country sank beneath the sea, and Plinlimmon itself, where the river rises, was perhaps buried in part beneath the waters. When the country again emerged, the old system of river-drainage in that area was resumed; and the Severn, following in the main its old course, cut a

[504 Early Physical Changes.]

channel for itself through the boulder-clay that partially blocked up the original valley in which it ran. When that original valley was formed through which the older Severn ran is the point that I shall now attempt to discover. This subject is intimately connected with the origin and geological dates of the channels of many of the other large rivers of England, most of which, unlike the Severn, flow eastward to the English Channel and the German Ocean.

I must begin the subject by a rapid summary of certain physical changes that affected the English Secondary and Eocene strata long before he Severn, after leaving the mountains of Wales, took its present southern and south-western course along the eastern side of the Pa1aozoic rocks that border that old land.

About the close of the Oolitic epoch the Oolitic formations were raised above the sea, and remained a long time out of water; and, during that period, those atmospheric influences that produced the sediment of the great Purbeck and Wealden delta were slowly wearing away and lowering the land, and reducing it to the state of a broad undulating plain. At this time the Oolitic strata still abutted on the mountain country now forming Wales and parts of the adjacent counties. They also completely covered the Mendip Hills, and passed westward as far as the mountains of Devon passing out between Wales and Devonshire through what is now the Bristol Channel. The whole of the middle of England was likewise covered by the same deposits, overlying the rocks that now form the plains of Shropshire, Cheshire, Lancashire, and the adjoining areas, so that the Lias and Oolites passed out to the area now occupied by the Irish Sea, over and beyond the present estuaries of the Dee and the Mersey,

FIG. 100.

General Arrangement of the Lower Secondary Formations
before the Deposition of the Upper Secondary Strata.

1. Disturbed and hilly Silurian Old Red and Carboniferous strata.
2. New Red beds.
3. Lower Lias.
4. Marlstone and Upper Lias.
5. Lower Oolites.
6. Middle Oolites
7. Upper Oolites and Purbeck beds.
[506 Cretaceous Overlap.]

which lie between North Wales and the hilly ground of Lancashire, formed of previously disturbed Carboniferous rocks. In brief, most of the present mountainous and hilly lands of the mainland of Britain were mountainous and hilly then, and must have been much higher than now, considering how much they have since suffered by denudation.

At this period, south of the Derbyshire hills, and through Shropshire and Cheshire, the Secondary rocks lay somewhat flatly; while in the more southern and eastern areas they were tilted up to the west, so as to give them a low eastern dip. The general arrangement of the strata would then be somewhat as in fig. 100.

The submersion of this low lying area brought the deposition of the Wealden strata to a close, and the Cretaceous formations were deposited above the Wealden and Oolitic strata, so that a great unconformable overlap of Cretaceous strata took place across the successive outcrops of the Oolitic and older Secondary formations. (See fig. 101.)

The same kind of overlapping of the Cretaceous on the Oolitic formations, took place at the same time in the country north and south of the present estuary of the Humber, the proof of which is well seen in the unconformity of the Cretaceous rocks on part of the Oolites and Lias of Lincolnshire and Yorkshire.

At this time, the mountains of Wales, and other hilly regions made of Pa1aozoic rocks, must have been lower than they were during the Oolitic epochs; partly by the effect of long-continued waste due to atmospheric causes, but much more because of gradual and greatly increased submergence during the time that the Chalk was being deposited. It is even possible that

FIG. 101.

Early Overlap of the Cretaceous Strata on the Lower Secondary Formations.

1. Disturbed hilly land, as in fig. 34.
2. New Red Liassic and Oolitic strata, &c.
3. Cretaceous strata lying unconformably on and overlapping the denuded edges of No. 2, and at * abutting on the half-submerged mountains of Wales.
[508 Miocene Continent.]

during the Upper Cretaceous period Wales sunk almost entirely beneath the sea.

I omit any detailed mention of the phenomena connected with the depositions of the freshwater and marine Eocene strata because at present this subject is not essential to my argument.

The Miocene period of old Europe was essentially a continental one. Important disturbances of strata brought it to a close, at all events physically, in what is now the centre of Europe; and the formations partly formed in the great fresh-water lakes that lay at the bases of the older Alps were, after consolidation, heaved up to form new mountains along the flanks of the ancient range; and all the length of the Jura, and far beyond to the north-east, was elevated by disturbance of the Jurassic, Cretaceous, and Miocene strata. The broad valley of the lowlands of Switzerland began then to be established, long afterwards to be overspread by the huge glaciers that abutted on the Jura, deepened the valleys, and scooped out all the rock-bound lakes.

One marked effect of this extremely important elevation, after Miocene times, of so much of the centre of Europe was, that the flat, or nearly flat-lying Secondary formations that now form great part of France and England (then united), were so far affected by the renewed upheaval of the Alps and Jura that they were to a great extent tilted, at low angles, to the north-west. That circumstance gave the initial northwesterly direction to the flow of 80 many of the existing rivers of France, and led them to excavate the valleys in which they run, including the upper tributaries of the Loire and Seine, the Seine itself, the Marne, the Oise, and many more of smaller size; and

[Origin of the Severn Valley. 509]
FIG. 102.

Post-Miocene Westerly Dip of the Cretaceous Strata of England.

1. Old Wales.
2. Oolitic strata, &c.
3. Chalk, &c. The surface of the Chalk slopes gently from east to west.

510 Severn and Avon, Tewkesbury.]

my surmise is, that this same westerly and north-westerly tilting of the Chalk of England formed a gentle slope towards the mountains of Wales, as shown in fig. 102, and the rivers of the period of the middle and south of England at that time flowed westerly. This first induced the Severn to take a southern course between the hilly land of Wales and Herefordshire and the long slope of Chalk then rising to the east. Aided by the tributary streams of Herefordshire, it began to cut a valley towards what afterwards became the Bristol Channel, and established the beginning of the escarpment of the Chalk, e, fig. 102, which has since gradually receded, chiefly by atmospheric waste, so far to the east. If this be so, then the origin of the valley of the Severn between e and 1 is of immediate postMiocene date, and is one of the oldest in the lowlands of England.

The course of the Avon, which is a tributary of the Severn, and joins it at Tewkesbury, is, I believe, of later date than the latter river. It now rises at the base of the escarpment of the Oolitic rocks east of Rugby, and gradually established and increased the length of its channel in the low grounds now formed of Lower Lias and New Red Marl as that escarpment retired eastward by virtue of that law of waste which causes all inland escarpments to retire away from the steep slope and in the direction of the dip of the strata.

If the general slope of the surface of the Chalk of this part of England had been easterly instead of westerly at the post-Miocene date alluded to, then the initial course of the Severn would also have been easterly, like

1 Many of the valleys of Wales must be very much older.

[The Avon, Bristol. 511]

that of the Thames and the rivers that flow into the Wash and the Humber.

One of the best known rivers that enters the estuary of the Severn is the other Avon, which flows through Bath and Bristol. Its physical history, on a small scale almost precisely resembles that of the Rhine between Basle and Bonn.

West of Bristol there is a high plateau of Carboniferous Limestone, the flat top of which attains a height of nearly 400 feet above the level of the sea. Through a deep narrow gorge in this limestone (fig. 103) the river flows, between Clifton and Durdham Downs on the east and Leigh Wood on the west, north-west of which it enters the low grounds and finds its way to the estuary of the Severn at King's Road. Above Bristol, north and south of the river, the country consists of a number of isolated flat-topped hills, of which Dundry Hill and the Mendips form conspicuous members, while in the neighbourhood of Bath, Lansdown, Charmy Down, Odd Down all the minor Oolitic plateaux now form portions of what was once a continuous broad tableland with minor undulations. In these regions the Avon takes its rise, swelled by many north-flowing tributaries, one of which, the Chew, rises on the north flank of the Mendip Hills. North of Bath, several minor streams flow into the Avon through beautiful valleys which have been scooped out of the Oolitic plateau, while the Boyd, the Siston, and the Frome pass through the soft undulating grounds of Lias, New Red Marl, and Coal-measures that lie west of the bold Oolitic escarpment between Bath and Wotton-under-edge. Some of these streams rise at

1 'On the Physical History of the Rhine,' 'Journ. Geol. Society,' 1874, A. C. Ramsay.

[512 The Avon.]

heights approximately as high as the summit level of the limestone gorge through which the Avon flows below Bristol.

The vulgar notion respecting the Avon and its gorge is, that before that ravine was formed all the low ground through which the river and its tributaries flow was a large lake, that 'a convulsion of nature' suddenly

FIG. 103.
Gorge of the Avon
Gorge of the Avon at Clifton, Bristol, looking down the river.

rent the rocks asunder and formed the gorge through which the river afterwards flowed, and so drained the hypothetical lake. It is scarcely necessary to add, that had there been a large lake in that area, we might expect to find lacustrine deposits and organisms in some parts of these valleys, but none exist.

The true explanation is, that in some late tertiary period of geological history, the surface of the country on either side of the river above the gorge formed a

[The Thames. 513]

great plain, somewhat higher than the summit level of the Carboniferous Limestone plateau. This plain being slightly inclined to the west at the time the Severn was scooping out its valley, as I have already explained at p. 508, the ancient Avon flowed over the top of the plateau of Clifton and Durdham Downs, through a minor inequality of the surface, and, as rivers do, it steadily worked at the deepening of its own channel. As it did this, so in like proportion the river and its tributaries in the upper part of their courses gradually wasted and lowered the hill-sides and valleys through which they flowed, being aided by rains and snows and all the ordinary agents of atmospheric denudation; and thus it happens, that what was once a high slightlyinclined tableland, has been converted partly into flattopped fragments of a high plain, and partly into undulating hills and vales; while in the great Oolitic plateau, that stretches eastward as far as the Chalk escarpment, we have still remaining a large tract of the ancient plain, with this difference, that the average gentle slope of its surface is now east instead of west.

This naturally leads to the question, Why is it that the Thames, and some other rivers that flow through the Oolites and Chalk, run eastward? The answer seems to be, that after the original valley of the Severn was well established by its river, a new disturbance of the whole country took place, by which the Cretaceous and other strata were tilted eastward, not suddenly, but by degrees, and thus a second slope was given to the Chalk and Eocene strata, in a direction opposite to the dip, that originally led to the scooping out of the present valley of the Severn. This dip lay east of the comparatively newly-formed escarpment of the Chalk indicated by the dark line in fig. 102 marked e. The

[514 The Thames.]

present Chalk escarpment, in its beginning, is thus of older date than the Oolitic escarpment (fig. 57, p. 304), but it would be hard to prove this, except on the hypothesis I have stated.

When this slope of the Chalk and the overlying Eocene strata was established, the water that fell on the long inclined plain east of the escarpment of the Chalk necessarily flowed eastward, and the Thames, in its beginning, flowed from end to end entirely over Chalk and Eocene strata.

The river was larger then than now, for I am inclined to believe, that in these early times of its history, the south of England was joined to France, the Straits of Dover had no existence, and the eastern part of the Thames as a river, not as a mere estuary, ran far across land now destroyed, perhaps directly to join an exterision of the north flowing river which we now call the Rhine. At its upper end, west of its present sources, the Thames was longer by about as much probably as the distance between the well-known escarpment of the Cotswold Hills and the course of the Severn as it now runs, for the original escarpment of the Chalk must have directly overlooked the early valley of the Severn, which was then much narrower than now (see p. 509). But by processes of waste identical with those that formed the escarpment of the Wealden (figs. 71, 72, 73, pp. 337-343), the Chalk escarpment gradually receded eastward, and as it did this the valley of the Severn widened, and the area of the drainage of the Thames was contracted.

By-and-by the outcropping edges of the Oolitic strata becoming exposed, a second and later escarpment began to be formed, while the valley of the Severn gradually deepened; but the escarpment of the Chalk being more

[The Thames. 515]

easily wasted than that of the Oolite, its recession eastward was more rapid, and this process having gone on from that day to this, the two escarpments in the region across which the Thames runs are far distant from each other.

All this time the Thames was cutting a valley for itself in the Chalk, and by-and-by, when the escarpment had receded to a certain point, its base became in part lower than the edge of the Oolitic escarpment that then, as now, overlooked the valley of the Severn, only at that time the valley was narrower. While this point was being gradually reached, the Thames by degrees was joined by the growing tributary waters that drained part of the surface of the eastward slope of ihe Oolitic strata, the western escarpment of which was still receding; and thus was brought about, what at first sight seems the unnatural breaking of the river through the high escarpment of Chalk between Wallingford and Reading.

From the foregoing remarks it will be understood why the sources of the Thames, the Seven Springs and others, rise so close to the great escarpment of the Inferior Oolite, east of Gloucester and Cheltenham. But just as in times long gone, the sources of the Thames once rose westward of the Seven Springs, so well known on the Cotswolds, so the sources of the river now, are not more stationary than those that preceded. The escarpments, both of Chalk and Oolite, are still slowly changing and receding eastward; and as that of the Oolite recedes the area of drainage will diminish and the Thames decrease in volume. This is a geological fact, however distant it may appear to persons unaccustomed to deal with geological time

[516 The Frome.]

A change in the story of an old river, even more striking than that of the Thames, has taken place in the history of what was once an important stream further south. Before the formation of the Straits of Dover, the solid land of England, formed of Cretaceous and Eocene strata, extended far south into what is now the English Channel. The Isle of Wight still exists as an outlying fragment of that land. At that time the Nine Barrow Chalk Downs, north of Weymouth Bay and Purbeck, were directly joined as a continuous ridge with the Downs that cross the Isle of Wight from the Needles to Culver Cliff. Old Harry and his Wife, off the end of Nine Barrow Downs, and the Needles, off the Isle of Wight, are small outlying relics, left by the denudation of the long range of Downs that once joined the Isle of Wight to the so-called Isle of Purbeck, and of the land that lay still farther south of Portland Bill the Isle of Wight and Beachy Head.

North of this old land, the Frome, which rises in the Cretaceous hills east of Beaminster, still runs, and, much diminished, discharges its waters into Poole Harbour. But in older times the Solent formed part of its valley, where, swollen by its affluents, the Stour, the Avon, the Test, and the Itchin, it must have formed a large river, which, by great subsequent denudations and changes in the level of the land, has resulted in the synclinal hollow through which the semi.estuarine waters of the Solent now flow.

The same kind of argument that has been applied

1 See Mr. T. Codrington 'On the Superficial Deposits of the South of Hampshire and the Isle of Wight.' Quart. Jour. Geol. Soc. 1870, vol. xxvi., p. 528, and Mr. John Evans, 'Stone Implements,' Chap. XXV.

[The Trent. 517]

to the Thames is equally applicable to the Ouse, the Nen, the Welland, the Glen, and the Witham, rivers flowing into the Wash, all of which rise either on or close to the escarpment of the Oolites, between the country near Buckingham and that east of Grantham, which rocks were once covered by the Chalk.

With minor differences, the same general theory equally applies to all the rivers that run into the Humber. I believe the early course of the Trent was established at a time when, to say the least, the Lias and Oolites overspread all the undulating plains of New Red Marl and Sandstone of the centre of England, spreading west to what is now the sea, beyond the estuaries of the Mersey and the Dee. A high-lying antielinal line threw off these strata, with low dips, to the east and west; and, after much denudation, the large outlier of Lias between Market Drayton and Whitchurch in Shropshire, is one of the western results. Down the eastern slopes the Trent began to run across an inclined plain of Oolitic strata. Through long ages of waste and decay the Lias and Oolites have been washed away from these midland districts, and the long escarpments formed of these strata lie well to the east, overlooking the broad valley of New Red Marl through which the Trent flows.

The most important affluent of the Trent is the Derwent, a tributary of which is the Wye of Derbyshire. The geological history of the Wye is very instructive. It runs right across part of the central watershed of England, formed by the great boss of the Carboniferous Limestone of Derbyshire. This course, at first sight seems so unnatural, that the late Mr. Hopkins of Cambridge stated that it was caused by two fractures in the strata, running parallel to the winding course of

[518 The Wye, Derbyshire.]

the river.
1 There are no fractures there of any importance. The true explanation is as follows:

At an old period of the physical history of the country, the valley north and west of Buxton had no existence, and the land there actually stood higher than the tops of the limestone hills to the east. An inclined 'plain of marine denudation,' stretched eastwards, and gave an initial direction to the drainage of the country. The river began to cut a channel through the limestone rocks; and as it deepened and formed a gorge, the soft Carboniferous shales in which the river rose, were also worn away by atmospheric action, and streams from the north and west began to run into the Wye. By the power of running water, those valleys were deepened simultaneously and proportionately to their distance from the sources of the river; and the farther the Wye flowed, the more was its volume increased by the aid of tributary streams and springs Thus it happens that the Wye seems to the uninitiated unnaturally to break across a boss of hills, which, however, were once a mere slightly undulating unbroken plain of limestone. There is no breakage of the rocks, and nothing violent in the matter. It was and is, a simple case of the wearing action of running water cutting a channel for itself from higher to lower levels, till, where Rowsley now stands, it joined the Derwent, which flows in a long north and south valley scooped

1 'On the Stratification of the Limestone District of Derbyshire,' by W. Hopkins, M.A., &c. For private circulation. 1834. In p. 7 he says, 'When two longitudinal faults, ranging parallel, are not, very distant from each other, they sometimes form a longitudinal valley, of which the valley of the Wye is a splendid instance. In such cases, however, it is curious that the faults do not generally coincide with the steep sides of the valley, but are distant from them by perhaps from 50 to 200 or 300 yards.'

[The Humber. 519]

by itself, chiefly in comparatively soft Yoredale shales between the high-terraced hard moorland scarps of Millstone Grit, and the still harder grassy slopes of the Carboniferous Limestone.

When we come to the other rivers that enter the Humber north and west of the Trent, the case is more puzzling. The Oolites in that region were extensively denuded before the deposition of the Chalk; so that between Market Weighton and Kirkby-under-dale in Yorkshire, the Chalk is seen to overlap unconformably the Oolitic strata, and to rest directly on the Lower Lias, which there, as far as it is exposed, is very thin. The Chalk, therefore, overspread all these strata to the west, and lay directly on the New Red beds of the Vale of York, till, overlapping these, it probably intruded on the Carboniferous strata of the Yorkshire hills farther west. At this time the Oolites of the northern moorlands of Yorkshire seem also to have spread westward till they also encroached on the Carboniferous slopes, the denuded remains of which now rise above the beautiful valleys of Yoredale and Swaledale, the whole, both Carboniferous and Secondary, strata having gentle eastern and south-eastern dips. These dips gave the rivers their initial tendency to flow south-east and east; and thus it was that the Wharfe, the Ouse, and the Swale, cutting their own channels, formed a way to what is now the estuary of the Humber, while the escarpments of the Chalk and Oolite were gradually receding eastward to their present temporary positions.

That the Oolitic strata spread northward beyond their present scarped edges is quite certain; but whether or not they extended far enough north to cover the whole of the Durham and Northumberland coalfield I am unable to say. Whether they did so or not

[520 The Tees, Wear, Tyne, &c.]

does not materially affect the next question to be considered; for if they did spread over part of these Carboniferous strata, they must have thinned away to a feather edge in times long before the Oolitic escarpment began to be formed.

Taken as a whole, from the great escarpment of Carboniferous Limestone that overlooks the Vale of Eden on the east, all the Carboniferous strata from thence to the German Ocean have a gentle eastern dip; so gentle, indeed, that, on Mallerstang and other high hills over]ooking the Vale of Eden, outlying patches of Millstone Grit, still remain to tell that once the whole of the Coalmeasures spread across the country as far as the edge of the Vale, and even far beyond in pre-Permian times, for the Carboniferous Limestone on both sides of the Vale of Eden, now broken by a fault, was once continuous, and the Whitehaven coalfield was then united to that of Northumberland. These gentle eastern and south-eastern dips, caused by upheaval of the strata on the west and north-west, gave the initial tendency of all the rivers of the region to flow east and south-east. Thus it happens that the Tees, the Wear, the Derwent, the Tyne, the Blyth, the Coquet, and the Alne, have found their way to the German Ocean, cutting and deepening their valleys as they ran, the sides of which, widened by time and subaerial degradation, now often rise high above the rivers in the regions west of the Coal-measures, in a succession of terraces of limestone bands, tier above tier, as it were in Titanic steps, till on the tops of the hills we reach the Millstone Grit itself.

I now turn to the western-flowing rivers, about which there is far less to be said.

First, the Eden :—This river flows along the whole

[The Eden 521]
FIG. 104.

Section from the Cumbrian Mountains across the
Vale of Eden to the Northumberland Coalfield.

1. Silurian flocks. 2. Carboniferous Limestone. 3. Yoredale rocks and Millstone Grit. 4. Coalmeasures. 5. Permian Conglomerates, Sandstones, and Marie.

[522 Rivers of Wales.]

length of that beautiful valley, through various Permian rocks, for nearly forty miles. At the mouth of the valley, at and near Carlisle, a patch of New Red Marl lies on the Permian sandstones, and on the Marl rests the Lias. Whether the whole length of the Permian strata of the Vale of Eden was once covered by these rocks it is impossible to determine, but I believe that it must have been so to a great extent, and also that the Lias may have been covered by Oolitic strata. A great fault east of the Eden has thrown these formations down on the west, so that the faulted edge of the Permian beds now abuts on the high Carboniferous hills that form the eastern side of the valley. As these Permian and Secondary were denuded away by time, the present river Eden began to establish itself, and now runs through rocks in a faulted hollow, in the manner shown in fig. 104. What is the precise geological date of the origin of this great valley and its river course in their present form, I am unable to say; but I believe that it may approximately be of the same age as the valleys last described: that is to say, of later date than the Oolites, and probably it is later than the Cretaceous and Eocene, or even than the Miocene epoch. And so with the other rivers of the west of England—the Lune, the Ribble, the Mersey, and the Weaver.

In Wales, the Dyfi partly runs in a valley formed by denudation along an old line of fault; and the Teifi in Cardiganshire, and the Towey in Caermarthenshire, in parts of their courses along lines running in the direction of the strike of soft Llandeilo flags, sometimes slaty and easily worn down by water, their valleys being bounded on either side by hills to a great extent formed of harder Silurian grits.

To sum up the subject: It seems to me that all

[Rivers of Wales. 523]

the rivers of Wales, whether flowing through Silurian, Old Red, or Carboniferous rocks, have been busy scooping out their valleys ever since the close of that great continental epoch that ended with the influx of the Rhatic and Liassic sea across the Triassic salt lakes, and though these valleys were modified by ice, and partially filled with detritus, during a short episode of submergence in glacial times, the rivers re-asserted their rights to their old channels when emergence took place. All the important rivers, therefore, that flow east and west and north and south through the Silurian rocks of Wales, are in their origin approximately of the same age, and from Cader Idris to Pembrokeshire they have all cut their way through a tableland with minor undulations, while here and there remains a higher hill, the rocks of which were unusually hard. This old upland was indeed of great extent, and its relics stretch far and wide into the northern part of Denbighshire, and into Montgomeryshire and South Wales. As already stated, standing on the summit of Cader Idris or of Aran Mowddwy, 2,960 feet high, and looking east and south, the eye, as far as it can reach, ranges across a vast extent of old tableland, the plane surface of which near the Arans is about 1,900 feet above the level of the sea, or more than 1,000 feet below the summits of the neighbouring mountains. All intersected by unnumbered valleys, to the ordinary observer it is merely a hilly country, while an eye versed in physical geology at once recognises that all the diversities of feature are due to fluviatile erosions that have scooped out the valleys.

For this reason it also happens that the Dee now cuts right across the Carboniferous escarpment west of Erbistock and the lower area of the Permian strata;

[524 The Dee.]

for when the Dee began to run, that escarpment had no existence, and the strata of these formations stretched further to the west, ending along some line now unknown in a sort of feather edge, and forming part of the great inclined plane over which the Dee ran at a level hundreds of feet above the bottom of its present valley. By-and-by, as the river channel deepened, the escarpment began to be formed, its face sloping in a direction at right angles to the general dip of the strata, after the habit of all such escarpments. The whole was strictly analogous to the manner in which the rivers of the Weald acted at a later date, and also for the same reason that the Thames now cuts across the escarpment of the Chalk. Escaped into the low country of the New Red series, the history of the Dee becomes simple, and requires no special illustration.

But this process of ordinary fluviatile erosion is not the only agent that has been at work in Wales, for in later geological times the Glacial epoch supervened, and the moving ice of thick glaciers exercised a strong abrading power. Then it was that in the mountain-region of the west, ice-smoothing, mammillations, and striations were so strongly impressed on the sides of so many valleys, and so many lake-basins were scooped out, and among others the rock-bound basin of Bala Lake; and though the face of the country is always being slowly changed, the time that has elapsed since the close of the Glacial epoch is comparatively so short, that the large essential rocky features of the regions traversed by the rivers have since that time undergone no important alteration.

In the 'Journal of the Geological Society' for 1876, I published the Physical History of the Dee. It is too long, and the necessary diagrams are too large

[The Dee. 525]

for publication in such a book as this, but the leading features of the story are, that before entering the plains of Cheshire, the river, passing through Bala Lake, runs through the beautiful Vale of Llangollen, which as far as the behaviour of the river is concerned, may on a small scale be compared to that of the Moselle (see p. 534, Chap. XXXI.). At the mouth of its valley the river passes through a bold escarpment of Carboniferous Limestone and Millstone Grit, whence suddenly bending to the north it passes through flats of New Red Sandstone to its long shallow estuary beyond Chester.

The greater part of the Silurian region on either side of Bala Lake, and of the Dee, stood high above the level of the sea, from remote geological times, and formed a wide tableland, extending far to the south, and also to the east and north-east, and on its edges rose the more mountainous land, formed by the Lower Silurian volcanic rocks, splendid relics of which still remain in the peaks of Cadir Idris, the Arans, and Arenigs.

When, by the drainage of this old land, the Dee, induced by minor undulations of the ground, began to flow in its earliest channel, it is clear that its present source, Bala Lake, had no existence; for whereas the river at that time must have flowed on a surface of land. not less high than that on either side of the present valley near Corwen and Llangollen (now, in places, from 1,600 to 1,800 feet high), the surface of Bala Lake is only 600 feet above the level of the sea, while the neighbouring watershed between the lake arid Dolgelli is only 200 feet higher. As the river could not flow. up hill, it is clear that in that early stage of its history, the valley of the Dee about Bala, must have been at least from 1,300 to 1,400 feet higher than it is now, and

[526 The Dee,]

consisted of a mass of Silurian rocks, great part of which has since been removed by denudation.

In my opinion this region of North Wales has never been depressed beneath the sea since the beginning of the Permian epoch, excepting in part during a short episode in Glacial times (see p. 413). During that long lapse of geological ages, there was therefore ample time for the action of all the ordinary processes of subaerial denudation, the most powerful of which is the action of rain, rivers, and glaciers, and thus it happened that the Dee, a river of very ancient date, wandering hither and thither, by degrees deepened its channel in the same manner that the Rhine and the toituous Moselle have cut out theirs, as described in my memoir 'On the Physical History of the Valley of the Rhine.' While this process was going on, minor tributary valleys were cut by rain and rivers in the tableland to right and left of the great main channel, and other smaller rivers in adjacent regions playing the same general part, this wide tableland of marine denudation was gradually turned by the scooping out of unnumbered valleys, into a region of hill and dale.

The Vale of Clwyd is of extreme antiquity, for it was a valley before the deposition of the New Red Sandstone, and it may be that the Clwyd has flowed ever since the end of the Triassic epoch, and the Conwy like the Dee is at least as old.

I cannot pretend to give a detailed account of the river systems of Scotland. My personal knowledge of the subject is less minute, and however minute it might be, the subject is difficult.
1 Something of the subject

1 Professor Geikie, who fully realises the difficulty of the subject, nevertheless enters into it and explains it, as far as his present. knowledge will allow, in his work, the Scenery and Geology of Scotland.'

[Rivers of Scotland. 527]

I know myself, but for fuller details the reader must refer to Professor Geikie's work, from which part of what I have to say is drawn.

By referring to any good geological map of Scotland and the north of England, it will be seen that the country is intersected by two great valleys, running from north-east to south-west, viz., the valley of Loch Ness running from Moray Firth to Loch Linnhe, and also the valleys of the Forth and Clyde. If we go farther south another valley traverses England from Tynemouth to the Solway Firth. The general strike of all the older formations of Scotland is more or less from south-west to north-east, and starting from the watershed of the north-west of Scotland between Loch Linnhe and Cape Wrath, it will be seen that almost all the larger rivers flow to the east and south-east, transverse to the strike of the strata. In fact, like the Thames, they may be said to start from a great scarped watershed facing the Atlantic, and run from thence more or less in accordance with the general dip of the strata, or rather in conjunction with that, down a sloping plain of marine denudation, till they find their way into the sea or into the great valley of Loch Ness. Thus, in some degree, they follow the same general law that guided the eastflowing rivers of England, though traversing much more mountainous ground, they have cut their valleys in hard, greatly disturbed, and metamorphic Lower Silurian strata.

South of the Great Valley, the rivers follow a northeast course, in Strath Dearn and Strath Spey, approximately parallel to the trend of the Great Valley, running in valleys probably excavated in lines of strike occupied by strata, less hard than the general mass of the country. The Tay does the same in the upper part.

[528 Rivers of Scotland.]

of its course. South of Strath Spey, the rivers find their way east and south-east to the German Ocean; the Tay and the Forth from a high watershed that crosses Scotland from the neighbourhood of Fraserburgh on the east to Crinan on the west coast. To a great extent it is formed of hard granitic rocks and associated gneiss, and on this account it is high because of its power to resist denudation.

Like so many other rivers, the Tay has cut its way in old times over, and now through, a high belt of ground, that of the Sidlaw Hills just above the estuary; and the Forth, the Teith, and the Allan have in like manner breached that long range of Trappean Hills, known as the Ochils and the hills of Campsie.

The whole of the estuary of the Forth and the greater part of the valley of the Clyde lie in an exceedingly ancient area of depression. That country is also covered more or less with Boulderclay, and with later stratified detritus of sand and gravel which were formed in part by the remodelling of the Glacial drifts. These rivers ran in that area before the commencement of these deposits, and indeed for unknown ages before that period. But we have no distinct traces of those earlier epochs when we try to trace them as regards the history of the rivers of Scotland; and we know little besides this, that the Forth and the Clyde ran in their valleys long before the deposition of the Boulder-clay, and with other rivers resumed to some extent their old courses after the emergence of the country.

As of the rivers already mentioned, this may also be said of the Tweed, that we know nothing for certain of its history, except that its valley is of later age than the Old Red Sandstone and Carboniferous rocks.

[Rivers of Scotland. 529]

My own opinion is, that all the valleys of the South of Scotland may be said to have been formed generally contemporaneously with the valleys of the adjoining region of the north of England already described.

Of this we are certain, that some very ancient valleys in Scotland are older than the Old Red Sandstone, the deposition of which has more or less filled them with detritus, and they are now being re-excavated by running water. Taken as a whole, most of them may be said to be as old as the river-made valleys of Wales and Cumberland, for the disturbances which affected the Silurian and other palæozoic formations of Scotland were coeval with those that first raised the mountains of Cumberland, Wales, and Ireland high above the level of the sea.

1 A model of the Thames Valley, by Mr. J. B. Jordan, coloured geologically, may be seen at the Geological Museum, Jermyn Street. It clearly explains the relation of the river to the Oolitic and Cretaceous escarpments, pp. 513-15.