torrents are numerous, a succession of such cones or fans, nearly or quite touching each other, spreads over the floor of a valley. From this cause, so large an amount of detritus has within historic times been swept down into some of the valleys of the Tyrol that churches and other buildings are now half-buried in the accumulation.

Looking more closely at the materials brought down by the torrents, we find them arranged in rude irregular layers, sloping downwards into the plain, the coarsest and most angular detritus lying nearest to the mountains, while more rounded and water-worn shingle or sand extends to the outer margin of the cone. This grouping of irregular layers of angular and half-rounded detritus is characteristic of the action of torrents. Hence, where it occurs, even though no water may run there at the present day, it may be regarded as indicating that at some former time a torrent swept down detritus over that site.

Quitting the more abrupt declivities, and augmented by numerous tributaries from either side, the stream whose course we are tracing loses the character of a torrent and assumes that of a river. It still flows with velocity enough to carry along not only mud and sand, but even somewhat coarse gravel. The large angular blocks of the torrential part of its course, however, are no longer to be seen, and all the detritus becomes more and more rounded and smoothed as we follow it towards the plains. At many places, deposits of gravel or sand take place, more especially at the inner side of the curves which the stream makes as it winds down the valley. Sweeping with a more rapid flow round the outer side of each curve, the current lingers in eddies on the inner side and drops there a quantity of sediment. When the water is low, strips of bare sand and shingle on the concave side of each bend of the stream form a distinctive feature in river scenery. It is interesting to walk along one of these strips and to mark how the current has left its record there. The stones are well smoothed and rounded, showing that they have been rolled far enough along the bottom of the channel to lose their original sharp edges, and to pass from the state of rough angular detritus into that of thoroughly water-worn gravel. Further, they will be found not to lie entirely at random, as might at first sight be imagined. A little examination will show that, where the stones are oblong, they are generally placed with their longer axis pointing across the stream. This would naturally be the position which they would assume where the current kept

rolling them forward along the channel. Those which are flat in shape will be observed usually to slope up stream. That the sloping face must look in the direction from which the current moves will be evident from Fig. 11, where a current, moving in the direction of the arrow and gradually diminishing in force, would no longer be able to overturn the stones which it had so placed as to offer the current flowed from the opposite quarter, it would have found the upturned edges of the stones exposed to it, and would have readily overturned them until they found a position in which they again presented least resistance to the water. In a section of gravel, it is thus often quite possible to tell from what quarter the current flowed that deposited the pebbles.

FIG. 11.-Flat stones in a bank of river-shingle, showing the direction of the current (indicated by the arrow) that transported and left them.

least obstacle to its passage.

Had the

Yet another feature in the arrangement of the materials is well seen where a digging has been made in one of the alluvial banks, but better still in a section of one of the terraces to be immediately referred to. The layers of gravel or sand in some bands may be observed to be inclined at a

In

steeper angle than in others,
as shown in the accompany-
ing figure (Fig. 12).
such cases, it will be noticed
that the slope of the more
inclined layers is down the
stream, and hence that their
direction gives a clue to that
of the current which ar-
ranged them. We may watch
similar layers in the act of
deposition among shallow
pools into which currents

are discharging sediment. The gravel or sand may be observed moving along the bottom, and then falling over the edge of a bank into the bottom of the pool. As the sediment advances by successive additions to its steep slope in front, it gradually fills the pool up. Its progress may be compared to that of a railway embankment formed by the discharge of waggon-loads of rubbish down its end. A section through such an embankment would

reveal a series of bands of variously coloured materials inclined steeply towards the direction in which the waggon-loads were thrown down. Yet the top of the embankment may be kept quite level for the permanent way. The nearly level bands (b, c) in Fig. 12 represent the general bottom on which the sediment accumulated, while the steeper lines in the lower gravel (a) point to the existence and direction of the currents by which sediment was pushed forward along that bottom. (Compare pp. 174, 175.)

As the river flows onward through a gradually expanding valley, another characteristic feature becomes prominent. Flanking each side of the flat land through which the stream pursues

its winding course, there runs a steep slope or bank a few feet or yards in height, terminating above in a second or higher plain, which again may be bordered with another similar bank, above which there may lie a third plain. These slopes and plains form a group of terraces, rising step by step above and away from the river, sometimes to a height of several hundred feet, and occasionally to the number of 6 or 8 or even more (Fig. 13). Here and there, by the narrowing of the intervening strip of plain, two terraces merge into one, and at some places the river in winding down the valley has cut away great slices from the terraces, perhaps even entirely removing them and eating back into the rock out of which the valley has been excavated. Sections are thus exposed showing a succession of gravels, sands, and loams like those of the present river. From the line of the uppermost terrace down to the spits of shingle now forming in the channel,

we have evidently a chronologically arranged series of riverdeposits, the oldest being at the top and the youngest at the bottom. But how could the river have flowed at the level of these high gravels, so far above its present limits? An examination of the behaviour of the stream during floods will help towards an answer to this question.

When from heavy rains or melted snows the river overflows its banks, it spreads out over the level ground on either side. The tract liable to be thus submerged during inundations is called the flood-plain. As the river rises in flood, it becomes more and more turbid from the quantity of mud and silt poured into it by its tributaries on either side. Its increase in volume likewise augments its velocity, and consequently its power of scouring its bed and of transporting the coarser detritus resting there. Large quantities of shingle may thus be swept out of the ordinary channel and be strewn across the nearer parts of the flood-plain. As the current spreads over this plain, its velocity and transporting capacity diminish, and consequently sediment begins to be thrown down. Grass, bushes, and trees, standing on the flood-plain, filter some of the sediment out of the water. Fine mud and sand, for instance, adhere to the leaves and stems, whence they are eventually washed off by rain into the soil underneath. In this way, the flood-plain is gradually heightened by the river itself. At the same time, the bed of the river is deepened by the scour of the current, until, in the end, even the highest floods are no longer able to inundate the flood-plain. The difference of level between that plain and the surface of the river gradually increases; by degrees the river begins to cut away the edges of the terrace which it cannot now overflow, and to form a new flood-plain at a lower level. In this manner, it slowly lowers its bed, and leaves on either side a set of alluvial terraces to mark successive stages in the process of excavation. If during this process the level of the land should be raised, the slope of the rivers, and consequently their scour, would be augmented, and they would thereby acquire greater capacity for the formation of terraces. There is reason to believe that this has taken place both in Europe and North America.

While it is obvious that the highest terraces must be the oldest, and that the series is progressively younger down to the terrace that is being formed at the present time, nevertheless, in the materials comprising any one terrace, those lying at the top must be the youngest.

This apparent contradiction arises from

the double action of the river in eroding its bed and depositing its sediment. If there were no lowering of the channel, then the deposits would follow the usual order of sequence, the oldest being below and the youngest above. This order is maintained in the constituents of each single terrace, for the lowermost layers of gravel must evidently have been accumulated before the deposit of those that overlie them. But when the level of the water is lowered, the next set of deposits must, though younger, lie at a lower level than those that preceded them. In no case, however, will the older beds, though higher in position, be found really to overlie the younger. They have been formed at different levels.

The gravel, sand, and loam laid down by a river are marked, as we have seen, by an arrangement in layers, beds, or strata lying one upon another. This stratified disposition indeed is characteristic of all sedimentary accumulations, and is best developed where currents have been most active in transporting and assorting the materials (p. 172). It is the feature that first catches the eye in any river-bank, where a section of the older deposits or "alluvium" is exposed. Beds of coarser and finer detritus alternate with each other, but the coarsest are generally to be observed below and the finest above. The "deltas " ассиmulated by rivers in lakes and in the sea will be noticed in Chapters IV and VII.

But besides the inorganic detritus carried forward by a river, we have also to consider the fate of the remains of plants and the carcases of animals that are swept down, especially during floods. Swollen by sudden and heavy rains, a river will rise above its ordinary level and uproot trees and shrubs. On such occasions, too, moles and rabbits are drowned and buried in their burrows on the alluvial flood-plain. Birds, insects, and even some of the larger mammals are from time to time drowned, swept away by floods and buried in the sediment, and their remains, where of a durable kind or where sufficiently covered over, may be preserved for an indefinite period. The shells and fishes living in the river itself may also be killed during the flood, and may be entombed with the other organisms in the sediment.

Summary.—The material produced by the universal decay of the surface of the land is washed off by rain and swept seawards by brooks and rivers. The rate at which the general level of the land is being lowered by the operation of running water may be approximately ascertained by measuring or estimating the amount of mineral matter carried seaward every year from a definite region,