Chapter XXV: Antiquity of the River-Drift (1)

In order to discuss this subject, it will be necessary to enter into some geological details; as it is evident that the least antiquity that can be assigned to the implements is that of the beds of gravel, sand, and clay in which they occur, and of which, in fact, they may be regarded as constituent portions. Whether they may not in some instances have been derived from beds of even greater antiquity than those in which they are found, is another question, which will subsequently be dealt with; but any one examining the condition of the beds in which the implements occur, will have no difficulty in seeing that they have not been disturbed since their deposit; while in most cases, the colouring of the worked and of the unworked flints they contain is similar, and affords proof of their having long lain together under the same conditions.

That the containing beds have, at all events in most cases, been deposited by fresh water, and not by the sea, is proved by the occasional abundance in them of land and freshwater shells, and the absence of those of marine origin; while their general analogy with the flood deposits of existing rivers, and their almost universal contiguity to them, raises the strongest possible presumption of their existence being due to river action. At the risk of being thought to have prejudged the question, I have, therefore, made no scruple in treating them hitherto as being River-drift. To show that for the most part they are so in reality, and to enable the reader to form some opinion of the manner in which deposits originally formed in and about the beds of streams or lakes, now in some cases occupy the tops of hills, and cover the slopes of valleys, far above the level of any existing neighbouring river, or even at a considerable distance from any stream, it will, I think, be well to state a hypothetical case; and then to compare the actual phenomena with it, and see how far they correspond. |663|

Should it appear that with a certain given configuration of the land surface, a certain character of rock, a certain climate, and a certain number of years, certain effects must, judging from all analogy, have been produced; and should we in the case of these ancient Drifts find some of the conditions to have existed, and all the phenomena to be in accordance with the hypothesis, we may with some confidence assume that the other original conditions existed also; and build up a connected theory which will account for the whole of the observed results, and will also throw light on their causes, as well as on the duration of time necessary for their operation to have produced such effects. In stating the case, I lay no claim to originality, and do little more than follow in the steps of Sir Charles Lyell, Sir Joseph Prestwich, and others who have made a study of the character and effects of fluviatile action.

As it is in the gravels of Chalk districts that Palæolithic implements have been chiefly, though by no means exclusively, found, let us base the hypothesis on the assumption that an extensive and almost horizontal area of Upper Chalk, covered for the most part with beds of marine clay and shingle, gradually rose from beneath the sea, to an elevation of 200 feet above its level. Let us also assume that the land was elevated at a rate far in excess of that at which any subaërial action, such as rain, frost, or snow, would enable a river flowing over it to excavate its valley to the depth of 200 feet in the space of time required for its elevation to that height. Let us further assume, that the winter climate was somewhat more rigorous than that which at present prevails in this country, and that there was a considerably greater annual rainfall. We may also, for the purposes of the argument, take the position of the coast-line as permanent, instead of its constantly receding in consequence of the eroding power of the sea upon the cliffs.

Let us now see what would theoretically be the effect produced by subaërial causes on the river-valleys in this area during an indefinite number of centuries.

Under ordinary circumstances, and with our present amount of rainfall, there is no geological formation less liable to floods than the Chalk, or at all events, its upper portion. It is of so absorbent a nature that it is only in the extraordinary event of the ground being hard frozen at the time of a heavy fall of rain, or of a rapid thaw of snow; or of some inches of rain falling in the course of a few hours, that the soil is unable to absorb the water as fast as it is delivered upon it. The moisture when once in the soil is |664| either carried off again by evaporation and vegetation, or descends to a point at which the chalk is saturated with water, which is, however, constantly being drained off by springs along the valleys. This body of water has been termed “the subterranean reservoir” in the Chalk. The consequence of this absorbent power of the soil is that the streams and rivers in a Chalk country are not liable to floods, and moreover that their flow is but little affected at the time by rain; they being almost entirely dependent on perennial springs, which, during the driest of summers, still continue to deliver the water that in the course of the preceding winter, or even previously, has accumulated in the body of the Chalk.

The surface of the “subterranean reservoir” in the Chalk is by no means level, but always presents a gradient towards the point at which the springs are delivering its contents, so that within a chalk-hill forming a watershed between two streams there is what may be termed a hill of subterranean water, the summit of which need not, and often does not, correspond with the apparent watershed on the surface. The angle of the water-surface gradient depends principally on two factors, the degree of friction in passing through the chalk, and the amount of rain that finds its way down from the surface.

The height of saturation varies much in different seasons, as is evinced by the intermitting streams, often known as bournes,[2758] which perhaps only flow for a few months once in every six or seven years. Near the Chalk escarpment in Hertfordshire, at a spot several miles distant from any stream, I have known this height of saturation, as shown by the level of water in a deep well, to vary as much as 70 feet in the course of a single year. But with a greater rainfall than at present, the Chalk might at all times be in a state of saturation up to within a few feet of the surface; and this would be materially assisted, were there no deep valleys in existence into which the subterranean water could be delivered; as, of course, if the outfall were raised, the level of permanent saturation would be raised also. Were the Chalk in a less porous condition than at present, of course also its absorbent powers would not be so great. Under the circumstances, therefore, which have been supposed, the river-and spring-water from a Chalk district would be delivered in a manner very materially differing from that which at present prevails. The delivery of water by springs would be but small in shallow valleys; and, indeed, the only |665| important springs would be those along the sea-shore; while irrespective of this, the greater rainfall would keep the soil so saturated, that floods would be as readily produced by heavy storms of rain as if the soil were the most unabsorbent of rocks. If after some lapse of time the rainfall diminished, and the valleys were deepened, so that the outlets for the springs were at a considerably lower level than that of the principal area of the country, the case would be altered, and the tendency to floods would be immediately reduced.

At the commencement of the state of things supposed in our hypothesis, these outlets, with the exception of those on the sea-shore, would be but little lower than the general surface of the country, which, however, would not be perfectly plane. For it seems probable that the waters of the retreating sea would, during the elevation of the tract of land, form shallow channels, cutting down some little distance into the clay or chalk; and thus, as it were, mark out a course along which streams or rivers would flow, after the land was completely free from the sea. In some places, perhaps, shallow lakes might be left, but these also would have channels draining off their waters when they rose above a certain elevation.

With a bare surface, such as a newly-elevated tract would expose, there can be no doubt that the eroding power of heavy rains would be highly effective; as may be seen at the present day in the far greater effects of heavy showers on bare soil than on that which is protected by turf and vegetation. At the same time, with a rigorous climate, such as that supposed, the winter accumulation of snow and ice would be great, and its thawing during the summer months would add enormously and rapidly to the streams draining the area, which would in consequence have great power to deepen and widen their channels. The outflows from the lakes, if any such existed, would also be enlarged, while their upper portions would be filled with material brought down by the streams, and eventually they would be drained, with the exception of some channels in their beds through which the streams would pass.

We may therefore readily suppose that in the course of no very great interval of time, geologically speaking, a river-system for carrying off the waters falling from the heavens, analogous in character to those of the present day, but with shallower valleys, would be formed on the surface of the elevated tract. Let us |666| suppose that while this, as it may be termed, preliminary configuration of the surface has been taking place, the land has become tenanted by various trees, shrubs, and plants affording means of subsistence to different forms of animal life; while the streams also have been occupied by colonies of freshwater _testacea_; and let us now trace what would be the action of the rivers. To use the words of Sir Charles Lyell,[2759] “when we are speculating on the excavating force which a river may have exerted in any particular valley, the most important question is, not the volume of the existing stream nor the present levels of its channel, nor even the nature of the rocks; but the probability of a succession of floods at some period since the time when the valley may have been first elevated above the sea.”

Now in the first place, all rivers whose banks are not artificially protected, and whose channels are not kept clear, are of necessity more liable to floods than those in civilized countries, which bear much the same relation to rivers flowing through uncultivated lands, as domesticated animals do to wild. We have, moreover, _ex hypothesi_, a fruitful source of floods in a greater rainfall and in a more rigorous winter climate. The marvellous effects of such floods in excavating channels, and in transporting materials, can only be estimated by those who have seen their results, or have studied the accounts given of them. When we read of a small rivulet on the Cheviots,[2760] swollen by heavy rain, having transported several thousand tons of gravel and sand into the neighbouring plain, and having carried blocks of stone, weighing upwards of half a ton, two miles down its course, while another block weighing nearly two tons was transported the distance of a quarter of a mile, we may form some conception of the effects of even a flooded brook. The blocking of a stream by ice or fallen trees, so as to keep back its waters, and thus form a lake, which is suddenly drained by the breaking of the barrier; a heavy fall of rain; or a rapid fall of snow on ground hard frozen, and therefore impervious, are common causes of floods; and such as we may presume to have prevailed in our hypothetical case. What, therefore, would be the effect of such floods?

The first effect would no doubt be to cause the streams to overflow their banks, and spread over the bottom of the valleys in which they usually flowed. The shallower the valley the greater |667| probably would be the sinuosities of the stream, and the wider would its waters spread. The greater also would be the probability of the stream, on the cessation of the flood, not returning to its original channel, which might have become obliterated or filled up, but of its flowing along some new course, it may be miles away from its former channel. Even when not flooded so as to overflow their banks, rivers along which a larger body of water flowed than there does at present, would, so long as they were not confined within deep valleys, have a tendency to wander over a much wider tract of country than that now occupied by their valleys. The tendency of all rivers to produce sinuosities in their course is well known; but Mr. Fergusson, in his excellent paper on recent changes in the Delta of the Ganges,[2761] has called attention to the fact that all rivers oscillate in curves, the extent of which is directly proportionate to the quantity of water flowing through them.

But rivers in a state of flood, or passing even at a moderate speed over soft or incoherent soil, are always turbid, owing to the presence in their waters of earthy matter which they are transporting towards the sea. The character of the solid matter thus transported by water in motion is entirely dependent on its velocity. A velocity of 300 yards per hour is sufficient to tear up fine clay; of 600 yards, fine sand; of 1,200 yards, fine gravel; and of a little over two miles per hour, to transport shivery angular stones of the size of an egg.[2762] Considering the small velocity requisite to remove the finer particles of the soil, and to retain them in suspension, a river such as has been supposed, must have been excessively turbid, so long as any fine earthy particles were accessible to its waters, or to those of the streamlets delivering into it.

The amount of solid matter suspended in turbid water is greater than might be imagined. Mr. A. Tylor has calculated that the detritus carried down by the Ganges is equivalent to what would result from the removal of soil a foot in depth over the whole of the area which it drains in 1,791 years,[2763] and that brought down by the Mississippi to one foot in 9,000 years. Other estimates fix this at one foot in 6,000 years, while the sediment contained in its stream has been estimated at from 1∕1245 to 1∕1500 of the weight of the water.[2764] Taking this latter proportion, an inch of rain |668| falling on a square mile of ground, and flowing off it in a turbid state, would carry with it at least forty-three tons of sediment; and were we to assume an annual rainfall of fifty-four inches—which, though exceptional, is by no means unknown even in the British Isles—about 2,300 tons of fine earthy matter would be removed from a square mile of country in a single year. Taking a cubic yard of solid ground as equal to a ton in weight, this would involve the removal of one foot in depth from the surface in about 450 years. If, however, a portion of the rainfall were delivered by springs, or fell on hard or rocky ground, so as not to be rendered turbid, of course the effect would be proportionally diminished. Sir Archibald Geikie[2765] has estimated that practically, at the present day, the Thames (apart from about 450,000 tons of chalk and other matter carried away annually in solution), lowers its basin at the rate of one foot in 11,740 years; the Boyne, one foot in 6,700 years; the Forth, one foot in 3,111 years; and the Tay, one foot in 1,482 years. It is, however, with water moving with far greater velocity than that merely sufficient to keep fine sediment in suspension, that we have to deal in this hypothetical case; and we may readily suppose the streams, at more or less regular intervals, liable to violent floods, eroding the chalk and the superimposed clays and gravels, and carrying with them not only the finer particles and sand, but the pebbles, large and small, of the gravel, and the flints washed out of the chalk.

Let us now consider what would be the condition of the surface of a broad shallow valley, on the cessation of a flood such as that which has been supposed. In certain parts removed from the main current, and where the water had been nearly stationary, we should find deposits of fine mud or clay; in others, where the water had still moved with sufficient velocity to retain the clay and fine silt in suspension, the heavier particles of sand would have accumulated; in others, again, the smaller stones and pebbles; while near the main current, especially on the inner side of any curves which it had made, and where of course its velocity had been diminished, we should find the larger flints and pebbles, probably to some extent intermixed with part of the finer materials. In the beds of mud and sand, we should probably find the shells of some of the molluscs inhabiting the waters, and also those of terrestrial species, washed in from the inundated land surface, or brought down from the banks of the tributary rivulets; while |669| mixed among the larger pebbles we might expect to find any animal bones that had been lying on the land contiguous to the stream, or any of the larger and heavier objects of human workmanship, that would have been carried off by such an inundation, had mankind been living on the banks of the river.

Were men, or any of the larger animals overwhelmed and drowned by the flood, it seems probable that, owing to the slight difference between their specific gravity and that of water, they would eventually have been carried down to the sea, unless by some means accidentally arrested in their course, or carried into the more stagnant waters. In either case, they would, on the waters subsiding, probably be exposed on or near the surface, and not be imbedded in any of the deposits of the stream. Assuming the existence at that time of a respect for the dead, such as may be regarded as almost instinctive in man, any human remains would be buried or otherwise disposed of, while the bones of the other carcases would be left within reach of the waters, should another flood occur.

At the mouth of the river, where it joined the sea, its excavating power would be considerably greater than farther inland; for at first, on account of the land having—as was presumed, in this hypothetical case—risen faster than the river could excavate its valley, the stream must have fallen as a cascade into the sea. This, by the cutting back of the lip in such a soft rock as the Chalk, would soon be converted into a rapid, where the greater velocity of the water would much add to its erosive power; and, ere long, a mouth to the river would be formed, which would soon become tidal. Before tracing the results that would be due to this greater declivity of the river-bed in the immediate neighbourhood of the sea, it will be well to consider what would be the results of successively recurring floods, in the less inclined broad shallow valley, on which we have been speculating.

There can be no doubt that with each succeeding flood the valley would be deepened; and the fact of its being thus deepened would tend to make it narrower, by restricting the windings of the river. We can, however, hardly imagine that in this deepening process the whole of the deposits spread by the former floods over the bottom and slopes of the valley would be removed, but must acknowledge the extreme probability of some portions of them having remained intact, especially those which were left at the greatest distance from the course eventually taken by the river |670| during its period of flood. When once they had been thus left, the chances of their being again assailed by the stream would become more and more remote with each successive flood; and though the waters might reach some deposit of the larger pebbles formerly carried down by the main stream, but now at a distance from it, yet they would only belong to the more sluggish portions of the flood, and at first might envelope them in beds of sand; and subsequently, when they were only accessible to the more stagnant turbid waters, leave layer upon layer of muddy silt or clay upon them. In forming the more loess-like beds the action of the wind in transporting sand and dust might also assist. In some cases, and especially at the extremity of curves, and at the end of the tongue between two streams, the accumulation of one period, though at a lower level than that of earlier date, might abut upon it, or even become mingled with it, so that an almost continuous coating of Drift-deposits might extend from the highest level to the lowest.

The bulk, however, of the deposits of one inundation would be moved by the next, or by one of those which subsequently recurred; and stones, and pebbles, and other objects might thus be transported down stream, from place to place, an indefinite number of times, and form constituent parts of an indefinite number of gravelly beds along the bottom of the flooded stream. They might, under some circumstances, lie for a long period of years in some particular bed, in which they would become stained by salts of iron or otherwise, and subsequently be transported and re-deposited among unstained, or differently stained pebbles. The angles of any flints thus transported from place to place would also become rolled, as would, in like manner, those of bones or teeth. In the same way, assuming, as we have done, that the surface of the Chalk in the district was in part, or wholly, covered with beds of marine clay and shingle, it is evident that in the earlier deposits, when the river flowed at the higher level, and was, as it were, commencing to excavate its valley, the proportion of the pebbles derived from these beds to the flints washed out from the Chalk, would be much greater than at a later period. For in the course of time the river would have worked its way below the level of these upper beds, and many of the pebbles at first deposited in its gravels would have been disturbed, again and again, in their beds; on each disturbance carried farther down the stream, and eventually so far as the sea or the tidal portion of the river. At the same time the |671| river itself would be principally excavating the Chalk which had been freed from the marine shingle, and would therefore be forming the gravel in its bed, for the most part, from flints derived from the Chalk.

In the same manner, pebbles brought from a distant part of the country, and higher up the river, would eventually become more abundant in the deposits near its mouth, than they were at the first. Still no amount of transport of this kind could bring any pebbles into the bed of the river, which did not, in some form or other, exist within its drainage area.

Besides the transporting power of water, which by itself is, under favourable circumstances, capable of producing considerable excavations in a comparatively short period, there is another force at work, where, as has been supposed in this case, the climate is severe, which not only aids in the transport of pebbles and blocks of stone from one part of the bed of a river to another, but is a fertile source of floods. This is the formation of ground-ice. Sir Joseph Prestwich,[2766] in his second “Memoir on the Flint Implement-bearing Beds,” has given numerous instances of the transporting power of this agent, and shown the method of its occurrence in running streams, when the cold suffices to reduce the temperature of the water, and of the bed of the river itself, to the freezing point. Under such circumstances a gravelly river bed—and on mud alone, ice rarely forms—may become coated with ice, which being lighter than water will, on acquiring certain dimensions, overcome the forces which keep it at the bottom, and rise to the surface, carrying with it all the loose materials to which it adhered.

M. Engelhardt,[2767] director of the forges at Niederbronn, in the Vosges, has, perhaps, more minutely than any one else investigated the causes of the formation of ground-ice; and to prevent its effects in causing floods, actually removed each year from the bed of the stream supplying the motive power to his works, the stones and other extraneous bodies round which it was likely to form. His account of the effects of ground-ice in causing floods in the upper part of the Rhine and the Danube is worth transcribing. These two rivers having “a rapid current, do not freeze, like the Seine, by being covered with a plane and uniform stratum; they bear along large blocks of ice, which cross and impinge upon one another, and becoming thus heaped together, finally barricade the river. It is a grand spectacle, when the Rhine is thus charged, |672| to see these countless drifts adjust themselves in their relative position, where they unite by congelation, and convey the idea of the fall of some mountain which has covered the plain with rocks of every dimension. But it is not this accumulation of ice-drifts in the Rhine which is of itself the cause of danger; it is, on the contrary, the _débâcle_, or breaking-up, which is often productive of calamitous consequences. When this _débâcle_ commences in the upper part of the river, above the point where the latter is completely frozen, the masses of ice, drifting with the current and unable to pass, are hurled upon those already soldered together; thus an enormous barrier is formed, which the water, arrested in its course, cannot pass over, and hence overflows to the right and left, breaking the dykes, inundating the plains, and spreading devastation and suffering, far and near. The disasters caused by the _débâcles_ of the Rhine have taught the riparian inhabitants to observe attentively the facts which may serve them as a prognostic, and put them on their guard against the irruption of the ice. It is thus that they have been led to observe the _grund-eis_—that is to say, the ice formed at the bottom of the rivers—for it is this ice which, in becoming detached from the bottom and rising towards the surface, unites itself to the under surface of the masses already in place, and by further embarrassing the discharge, exposes the country to inundation.”

Another most effective agent in transporting the pebbles and larger blocks of stone along the course of rivers is shore-ice. During a severe winter masses of thick ice are formed which enclose the larger stones on the bottom of the river towards its edge; these masses are dislodged and carried away by subsequent floods, whether arising from rapid thaws or from rain higher up the river, or from accumulations of ice, such as those described, having formed a temporary barrier across the stream through which the pent-up water eventually burst and carried all before it. The lateral pressure of such dams of ice, with a large body of water behind, must be enormous; and we can readily conceive their crumbling-up any beds of gravel on the banks of the rivers against which they might happen to abut.

But there is still another way in which a severe climate, such as has been supposed, would act upon the rocks, namely, by their being rent and disintegrated by frost. This has been well pointed out by Sir Joseph Prestwich,[2768] who has cited numerous instances |673| of its effects, and mentions having seen a low cliff of chalk, 15 feet high, form a talus or heap of fragments at its foot, 6 feet broad and 4 feet high, in the course of an ordinary winter.

As I am by no means attempting an exhaustive geological essay on this subject, which is indeed hardly needed, I think that enough has been said to show that under conditions such as have been supposed in this hypothetical case, the great subaërial agents—rain and snow, ice and frost—would, in the course of time, enable rivers to excavate their valleys to an almost indefinite extent. Indeed, one can conceive the process being carried on, until what had been rivers became estuaries or arms of the sea; or, until a large island once traversed by rivers became converted into several smaller islands, by the cutting back, and subsequent junction, of its various river-valleys.

Without, however, carrying the excavatory process to such an extreme, let us now consider what would be the condition of our hypothetical river-valley when excavated to a depth of say 100 feet, at a point about midway between its source and the sea. We have already seen that at an earlier period—when the river ran at a higher level by 100 feet than that it is now supposed to occupy—its valley must have been broader, and its bottom strewn with detritus of various kinds, in the shape of gravel, sand, and clay, and, it may be, some larger blocks of stone. In the further process of excavating by agents such as have been described, it has also been seen, that it is in the highest degree improbable that the succeeding floods and other transporting agents should have entirely removed and obliterated the deposits left by those of earlier date. We should, therefore, expect to find, at various heights on the slope of the valley, remains of such beds of detritus, and especially at points such as the junctions of affluents with the river, and the inner side of the bends it makes in its course, which would naturally be the least exposed to the violent invasion of the stream. In these beds we might reasonably search for the remains of the surface and freshwater life of the period; and had there been any amelioration of climate during the process of excavation, a larger proportion of silt and clay, and less of coarse gravel, in the lower and more recent deposits, would testify to the fact. Looking also at the power possessed by rivers of levelling the bottoms of their valleys, during their successive changes of course, we might expect to find in places, tracts of these old valley-bottoms left as terraces on the slopes of the more deeply excavated valleys. The |674| upper surface of any such relics of a former condition of things would, of course, be covered with _débris_ and rain-washed clay, brought down from a higher level on the slopes, but on digging into them their true nature might be recognized.

Nearer the sea, and farther up the valleys, the state of things would be somewhat different. At the mouth of the river, as has already been pointed out, the declivity of the stream would have been greater, and its excavating power therefore increased. If, as originally assumed, the bed of the river, when the land was first elevated, was, at a mile distant from the sea, 200 feet above its level, the declivity would be 200 feet to the mile; when the 200 feet level was 4 miles from the sea, the slope would still be 50 feet to the mile; at 10 miles distance it would still be 20 feet, and it would not be until the 200 feet level was 15 miles from the sea that the ordinary slope of the bottom of the Chalk valleys of Hertfordshire, which is about 13 feet 6 inches to the mile, would be attained. In the meantime, however, if the sea were encroaching on the shore, or were, owing to the nature of the rocks, widening and extending that portion of the river subject to tidal influences, the actual point of contact with the sea would be carried far inland, and—assuming the rock traversed to be of one uniform nature and hardness—it would be long before the river towards its mouth ceased to have a greater declivity than nearer its source. We see, then, that the amount of excavation effected by the river, during the time necessary for the deepening of the valley by 100 feet, at a point midway in its course, would, near the sea, have been twice as great, or 200 feet. We should, therefore, expect to find beds of the same age as those which, at the middle of its course, were 100 feet above the river, at relatively twice that elevation near the mouth; and any intermediate beds would also be proportionally higher above the then existing stream, than contemporary beds farther up the valley.

At the heads of the valleys, the excavation would, on the contrary, have been less than towards the middle of the course of the river; partly owing to there always being less water present, partly to the reduced liability to floods, and partly to other causes. The heads of the valleys would, however, be constantly receding in all cases, and their retrogression would in most instances be aided by springs issuing from them. In cases where, from some geological cause, the heads of two valleys running in opposite directions receded in the same line, we can readily imagine their |675| meeting eventually at the watershed, and cutting through it so as to form apparently but a single valley, though on either side of the highest portion of its bottom, the waters flowed in opposite directions.

The mention of springs recalls another denuding agent, which has been already discussed in connection with caverns, and seems to have assisted in moulding the surface of the country and in excavating the valleys. It is well known that the water flowing in the streams of a chalk-country contains, in solution, a considerable amount of chalk, or rather, of bi-carbonate of lime; the water on entering the ground deriving a certain amount of carbonic acid from the decaying vegetable matter contained in the soil, and when thus charged, becoming capable of dissolving a corresponding quantity of the chalk. The amount is usually 17 or 18 grains in the gallon; and even in the Thames at London, not a purely chalk-stream, there are about 14 grains. Taking the proportion of 17 grains to the gallon, it will be found by calculation that every inch of rain which falls over a square mile of chalk-country, and passes off by springs, carries with it, in solution, and without in the slightest degree interfering with its brightness, no less than from 15 to 16 tons of solid chalk. The quantity of rain which thus finds its way to the springs has, as already stated, been ascertained by experiment to be as much as 9 inches per annum in average seasons, giving an amount of about 140 tons of chalk thus annually carried away from each square mile of country at the present day; so that the loss is still going on at the rate of 140,000 tons of dry chalk to each square mile in every ten centuries.

The lowering of level from this cause is probably not uniform over the whole surface. For the acidulated water sinking into the chalk on the top of a hill, and descending one or two hundred feet before reaching the surface of “the subterranean reservoir,”[2769] might, in its almost vertical passage, become saturated with carbonate of lime, and only render the chalk through which it passed somewhat more porous, without materially affecting the level of its surface. On the other hand, that absorbed in a valley would probably, to some extent, acquire the chalk which it eventually held in solution during its almost horizontal passage to the point of its delivery by springs; and as this would be at no great depth, the abstraction of solid matter would become more perceptible on the surface, so that the level of the valley would be lowered more |676| rapidly than that of the hill. With an increased rainfall, such as we have supposed, this removal of solid matter by solution must have been considerable; but still nothing in comparison with that effected by the other denuding agencies which have been mentioned. It is, moreover, to be borne in mind that, as will shortly be seen, until the valleys had been excavated to a considerable depth, the amount of water delivered by the springs would, with the same rainfall, have been far less than at present. The springs would also, to some extent, have been affected by the chalk being in a less porous condition than it now is, owing to its not having lost so much of its substance by the chemical action which has just been described.

Before comparing the actual phenomena with the results of the conditions which have been assumed, it will be well to say a few words as to the probable effects of an amelioration of climate, and a diminution in the rainfall, upon a valley already excavated to an average depth of 100 feet, such as has already been described. It is evident that any transport of materials due to the action of ice, by floating loose stones and pebbles from one part of the bed of the stream to another, would be materially diminished; as would also the number of floods resulting from the thawing of the winter accumulation of ice and snow, and from rain falling on frozen ground. The only remaining principal cause for floods would be the heavy fall of rain during storms or wet seasons; but here, a comparatively slight alteration in the conditions will have made a vast difference in the results. When the valleys were once excavated to a certain depth, the level of the springs or outfalls carrying off the accumulation of water in the absorbent soil, would be proportionally reduced, as would also be the line of permanent saturation in the chalk. The effect of this would be that during any dry interval, the water contained in the upper part of the chalk would gravitate downwards, until it reached the subterranean reservoir of water saturating the chalk; and thus leave the surface soil in the same absorbent condition as it is at present, and capable of receiving a much greater amount of rain than formerly, before any would flow from off its surface.

Even with a constant and excessive rainfall, the result of the continued deepening of the valleys would be to cause more and more to flow off by the springs, and less from the surface; but with the valleys once deepened, a small diminution in the rainfall, or its more even distribution over the whole year, might cause the |677| flow from the surface almost entirely to cease, and allow the whole to be carried off by the springs. Whenever this was the case, any great and rapid excavation of the valleys from rain alone would be rendered almost impossible; and with no extreme reduction in the total amount of annual flow of the rivers, yet by their originating in perennial springs subject to but slight variations, and from their being no longer to any extent immediately connected with the surface drainage, there would cease to be that immense difference between their maximum and minimum volume, which must have formerly existed. The result of this comparatively uniform flow would be a great diminution in the tendency of any river to change its bed, and even if it occasionally received a great accession of water, it would find relief by overflowing into the wide valley due to its former more violent action. In the less inclined portions of its valley, the parts now almost deserted by the stream would be favourable for vegetation, such as would result in the formation of peat, and any occasional overflowing of the banks might, owing to the less torrential character of the inundations, have a tendency to fill up and level these marginal spaces rather than to excavate them deeper. The deposits of gravel, sand, and clay at the low levels would also be more continuous than those at the higher.

In tracing the effects of subaërial action in forming valleys, I have assumed the subsoil or rock in which they were formed to have been chalk, as it is principally in valleys in the Chalk that the gravels containing Palæolithic implements are known to occur. This is probably on account of the greater natural abundance of flints in such valleys, which of course led to implements being there chipped out in greater numbers, as well as to their being less cared for, from their being more easily replaced than they would be where flint was scarce. The effects on other soft and absorbent soils would not materially differ from those on chalk. On clay, the general amount of denudation would perhaps be greater, but the valleys broader, and with less inclined slopes on their sides. In a clay country we might, I think, expect to find the old river-gravels not unfrequently at greater distances from the existing streams than in a chalk-district.

It must, however, be borne in mind that in such a country the materials from which river-gravels can be formed are usually absent, and can only have been derived from older superficial beds, or brought from Chalk higher up the valley. In some |678| valleys, partly or almost entirely excavated in Pre-Glacial times, gravels belonging to the Glacial Period exist, and tend to complicate the question of the more recent River-drifts.

Any theory of the valleys having been excavated at some remote period in some unknown manner, and then having been filled with gravels derived from an unknown source, and again re-excavated, presents such difficulties that, to my mind, it cannot well be entertained. If, however, such a view be accepted, it seems to add to the time necessary for the excavation of the valleys; as much of the rainfall might find a subterranean vent at a low level through the gravel lining the bottom of the filled-up valleys, and thus keep the upper soil in a more absorbent condition and therefore less liable to erosion.

I must not, however, dwell too long upon this hypothetical case, which perhaps is such as may not have found an absolutely exact analogue in nature, but which may yet, I think, be accepted as a fair typical example of the results which, under the supposed conditions, must, judging from what we know of the action of subaërial causes, in all probability have ensued.

Let us now compare the phenomena as we find them in the gravel-beds of our present river-valleys, with those of the hypothetical case, and we shall, I think, find them coincide in a remarkable manner.

In the first place, the constituent parts of the gravels of the beds of Drift containing Palæolithic implements are always, petrologically, such as are to be found in the existing river-basins, as they must also of necessity have been in the hypothetical case. This fact, which holds good both in France and England, has been insisted on by Sir Joseph Prestwich, and such insistency cannot be too often reiterated. Where old superficial marine deposits of the Glacial or any other period, consisting of pebbles of various ages and origins, exist within a river-basin, there also will such pebbles be found in its gravels, but the originally derivative character of the pebbles prevents any strong argument being founded upon their presence. Where, however, no such beds exist, the case can clearly be made out. Unless a river traverses a granite or slate country, no granite or slate is found in the Quaternary gravels of its valley: unless it passes over Oolite, Purbeck, or Greensand, no blocks or pebbles of these rocks occur. This fact suffices to prove that the gravels are due to some local cause, such as river-action, and not to any general submergence or supposed |679| “wave of translation,” which would of necessity bring in materials not to be found in the existing basins.

That the various deposits resulting from a flooded river, should contain some of the land and freshwater shells, and animal bones of the period, is, as has been shown, most natural. Such shells and remains are of constant occurrence in the Quaternary gravels. If they prove nothing else, their evidence as to the freshwater origin of the beds must be accepted as conclusive. It is true that in all cases such land and freshwater remains have not as yet been found; but if in a dozen instances we find beds of a certain character containing these remains, and also flint instruments wrought by the hand of man; and in a dozen other instances, similar beds in analogous positions, also containing implements of the same kind, but, so far as is known, no such organic remains; we are justified in regarding both sets of beds as due to the same original cause, and in believing that the organic remains, if actually absent, are so from some accidental circumstance. We may indeed accept the implements as being truly characteristic fossils of a certain class of deposits. The character of the beds, consisting as they do, of gravel, sand, and fine silt, brick-earth or loess, and their manner of deposition, are also absolutely in accordance with the river-hypothesis.

On the higher levels above but near the valleys, we frequently find these beds at a considerable distance from the existing stream; we find them at all levels on the flanks of the valleys, and occasionally almost at their bottom, or even below it. In these lower beds, the implements, if of the same form and character as those in the upper beds nearer the source, are, in accordance with what would be the case under the hypothesis, very frequently much rolled and water-worn. The beds at the low level are also usually, so far as the gravel is concerned, of a finer character than those at the high level, and present a greater abundance of sand and brick-earth. They seem, in fact, indicative of some such amelioration of climate as that supposed.

Looking again at the position of the deposits with regard to the neighbouring rivers, we find them, as a rule, exactly in such positions as might have been expected, had their presence been due to the action of a stream in the process of excavating its valley, in such a manner as that described. So constantly is this the case, that a practised geologist, from a mere inspection of the Ordnance map, could with almost certainty predict where deposits |680| of River-drift would occur, of such an age and character as to be likely to contain Palæolithic implements. In more than one instance, indeed, as has already been mentioned, the probability of certain gravels containing these relics of human art, was pointed out before their actual discovery.

These are some, but by no means all, of the points in which the actual phenomena agree with those which must have resulted from river-action such as suggested in the hypothesis, and they are alone sufficient to raise the strongest presumption that the phenomena are due to such action, and that the theory that would account for them in this manner, cannot be far from the truth.

I will, however, now pass in review some of the principal localities where Palæolithic implements have been found in Drift-deposits, and see what other points of accordance, and what difficulties, if any, they present.

Taking first the basin of the Ouse and its tributaries, we find at Biddenham, near Bedford, one of the principal localities for Drift-implements, the gravel on the inner side of a bold sweep made by the river, and from forty to fifty feet above it. Its constituent stones are all derived either from the rocks in the neighbourhood, or from the Glacial beds which cap them, and which have evidently been cut through by the river. Throughout the beds are seams containing numerous freshwater shells, mixed with some derived from the land and from marshy places; numerous bones of terrestrial mammals also occur. In the valley of the Lark remains of such shells occur at Bury St. Edmunds, in the same beds as the implements. Farther down, at Icklingham, the beds at Rampart Field cap a rounded knoll on the inner side of a curve of the river, which appears, however, to have somewhat straightened its course since they were deposited. Below Icklingham, the whole surface of the country, and its drainage, have been so much modified by the invasion of the sea, which produced the wide level of the Fens, that we should expect to find any deposits of an ancient river, which existed before that great planing down of the adjacent country, in somewhat anomalous positions.

I need not here enter into the history of the origin of the Fens; it is enough to say that the subsoil of almost the whole district consists of clays, belonging either to the Oolitic or Cretaceous series, and unprotected by any rocks of a more durable nature towards the sea, which has thus been enabled to invade it. The presence of the sea is attested in various localities by marine |681| remains. _Buccinum_, _Trophon_, _Littorina_, _Cardium_, and _Ostrea_ are abundant in the gravel at March.[2770] In the valley of the Nene, near Peterborough, oysters and other marine shells occur, mixed with those of land and freshwater origin. In Whittlesea Mere, remains of walrus and seal, and sea shells are found; while so far south as Waterbeach, less than ten miles from Cambridge, remains of whale have been discovered.

The old land-surface having been thus destroyed, we cannot with certainty trace the course of the ancient representative of the river Lark, below Mildenhall; it seems, however, to have proceeded northwards by Eriswell and Lakenheath, to join the Little Ouse. At Eriswell, a gravel of the same character as that near Mildenhall, occurs on the slope of the hill towards the Fen; but in it, as yet, few implements are recorded to have been found. At Lakenheath, however, they occur in the gravel now capping the hill overlooking the Fen, as well as on the slope.

Owing to the distance of these beds from any existing rivers, the late Mr. Flower[2771] found great difficulty in reconciling them with any theory which would account for their presence by the action of rivers. If, however, we regard the great denudation of the Fen country as subsequent in date to the deposit of the gravels, it appears to me that any difficulty on this point vanishes. That this denudation was in fact, at all events in part, subsequent to the deposit of the gravels, is proved by the position of the beds at Shrub Hill, which there cap a small area of Gault, and which, being above the general level of the Fens, can hardly have been deposited in the position they now occupy, when the configuration of the country was at all like what it now is. Such beds must, on the contrary, have been deposited in the bottom of a valley; and it appears as if in this case, by their superior hardness to the clay around them, or from some other accidental cause, they had protected this small spot from tidal action, which in the adjacent river, previously to the construction of Denver Sluice, extended nearly as far as Brandon.

The rolled condition of so many of the implements found at Shrub Hill, proves that they must have been transported some distance by water, from beds of a higher level.