Charmouth the river Char, have both been driven a considerable distance westward by the drifting shingle.

The outfall of the river Sid has been completely blocked by a bank of shingle, which has drifted across its mouth.

At Looe, in Mount's Bay, Cornwall, the drifting shingle forms a complete bank across the estuary of the river Caher and two smaller streams, and prevents the outflow of the water, creating a large fresh-water lake 7 miles in circumference, called Looe Pool. This bank is from 300 to 400 yards wide, and its top is 34 feet above L.W.S.T. In summer the lake maintains an even level, the water coming down the river passing away by evaporation or percolating through the shingle; but in wet weather the water rises 10 feet above its normal level, and a channel is then cut through the shingle, and the water allowed to run off. After the flood water subsides the channel fills up again.

Dieppe offers an illustration of the effect of shingle on ports. In remote times vessels went 13 miles up the valley to Bouteilles, where the port at that time existed. The mouth of the valley formed a deep roadstead, to which there were two means of access, one at the foot of the cliff to the east by Eaulue and Bethune, and the other at the foot of the rocks to the west by the river Arques. The west entrance became choked with shingle, and the town now stands on this bed of shingle.

In tidal rivers like the Thames, the Severn, the Humber, and the Seine, although their estuaries may be encumbered with large beds of sand, the ebb and flow of the large volume of tidal water is sufficient to maintain a deep-water channel, across which the beach material cannot drift in sufficient quantity to form a bar, but in smaller rivers, especially when situated in depressions in the coast-line, the balance of power of the tidal current is not sufficiently dominant to prevent the beach material, whether sand or shingle, drifting across the mouth of the channel in the form of a ridge, the continuance of which is aided by the rotary motion of the water.

The most remarkable instance of this kind was the bar of the Mersey, which, extending across the channel from a mass of sandbanks on either side, consisted of a ridge of sand, having a depth of 50 feet of water on one side and 40 feet on the other, with only 9 feet on the crest. The bar at the mouth of the Boston Deeps, at the entrance of the northern side of the Wash, consists of three separate ridges of sand extending over a mile, with 12 feet at low

The

water on the crests, and from 4 to 5 fathoms in the channel. entrance to the south side of the Wash, which has a depth of from 14 to 15 fathoms, has no bar.

The rivers Deben and Blyth, on the East Coast, have to fight their way to sea through large banks of shingle which have drifted across their outfalls, and the Adur at Shoreham is only kept open by constant dredging.

On other parts of the coast the shingle has drifted across indents or bays and completely enclosed them from the sea.

At Slapton a deep indent, into which two small rivers discharge, has been enclosed by a bank of shingle which has drifted across it, and two fresh-water lakes formed about 2 miles in length and the sixth of a mile wide.

At Northam, in Bideford Bay, the pebble ridge has drifted across a tract of low land, 900 acres in extent, and formed a natural embankment, enclosing it entirely from the sea.

The harbour at Pagham, near Selsea Bill, is now completely shut off from the sea by a large bank of shingle, and the channel blocked up. The western harbour of Rye and the outfall of the river Brede, after repeated attempts to keep it open, had to be abandoned owing to the drift of shingle from the west.

At Lancing, on the South Coast, the shingle has drifted across an indent in the coast, forming a bank 150 yards wide, which now protects the low land inside from high tides.

It is probable that the enclosure of the large tract known as Romney Marsh was first effected by a bank of shingle drifting across the wide estuary of the river Rother, the bank being enlarged and strengthened by the Romans.

Transporting Agency.-Attention was first directed to the movement of shingle in a paper presented by Mr. Palmer in 1834 to the Royal Society,1 in which he divided the action into three classes (1) Heaping up or accumulating; (2) disturbance or breaking down; and (3) progressive movement; and contended that all of these were due to the effect of wind, and that tidal action alone was not capable of effecting the result produced.

This view was generally accepted at the discussions which took place at the Institution of Civil Engineers on the Chesil Bank.2

1 "Observations on the Motion of Shingle Beaches," H. R. Palmer, Phil. Trans. of the Royal Society, April, 1834.

"On the Chesil Bank," J. Coode, Min. Proc. Inst. C.E., vol. xii. ; Ibid., J. Prestwich, Min. Proc. Inst. C.E., vol. xi.

The observations of the author have led him to a different conclusion as regards the first. With regard to the second conclusion, he is in accord. With regard to the third, although it is admitted that wind is one of the principal factors in the movement of shingle, the facts hereafter given show that tidal action is capable of producing, and does effect, the regular and continuous drift of shingle that takes place along the coast.

The effect of waves due to gales is principally destructive, pulling down and dispersing the banks accumulated by tidal action in calm weather.

The action of waves due to gales of wind is intermittent, variable in direction, and irregular.

On-shore gales invariably cut out and lower shingle beaches, and cause the pebbles to be spread over the beach at a lower level. As soon as the gale ceases and the wind becomes calm or offshore, the shingle begins to move upwards again, and in course of time is built up into a steep bank.

Winds blowing off-shore have little or no effect on the movement of shingle.

The effect of an on-shore gale is to reduce the slope of a shingle bank from 35 or 45 degrees, or say 1 in 13 to 1 in 1, to a slope of 5 degrees, or 1 in 10. Under this process, however, although the bank is cut out and partially destroyed, the top may be raised to a considerable height above the normal height of high water, and remain so permanently. This height and the size of the stones thrown up depend on the force of the gale, the length of the fetch to which the bank is exposed, and the height to which the tide rises.

If the wind is not blowing directly on-shore, but obliquely, the waves will also break in an oblique direction on the bank, but the return wave will flow down the bank along the shortest course or directly down the slope; and in addition to the results above described, the shingle will be moved along the shore in the same direction as the wind is blowing, the coast being thus denuded at the windward end, and the material transported to the leeward.

If the material be drifted by the wind-waves in the direction from which the normal travel takes place, the shingle, after the gale ceases, will be drifted back in due course. If, however, it is drifted to leeward, it will not return. It is generally accepted as an axiom that shingle once drifted to

leeward is lost for ever to the part of the beach from which it was moved.

With regard to the building of banks of shingle and its movement along the coast being due to the action of the "prevailing winds," the following facts deserve consideration.

The prevailing winds in this country are those from the southwest, which blow from that quarter on a greater number of days than from any other. The term "prevailing wind" is, however, in this case sometimes taken to be that wind which blows in the same direction as the flood current and which causes the highest tides, and the one most destructive to the coast, these gales being distinguished as the "predominant winds." There is no doubt that gales blowing with the set of the flood tide do the greatest damage to the coast, and so provide the greatest amount of débris from which the shingle is supplied. If, however, the regular and continuous direction of the drift of the shingle be taken all round the coast of this country or on the opposite shores, or, in fact, on any tidal coast where ordinary conditions prevail, it will be found that it is not governed either by the prevailing or predominant winds, but invariably follows the set of the floodtide.

The flood-tide coming from the Atlantic sets northwards along the west coast of this country, and the general drift of the shingle is in the same direction. The drift, however, follows the flood tide up the Bristol Channel, working round the various headlands and bays. Thus, in the Bay of Barnstaple the direction of the set of the drift varies no less than six times.

In the Irish Sea the flood tide sets into it both from the north and from the south, the two waves meeting in Morecambe Bay, and there is a corresponding variation in the set of the drift. On the north side of the Ribble estuary the shingle sets eastward up the estuary to St. Anne's and Lytham, where at the former place it has formed one of the shingle-banks known as "The Double Stanner," extending out from the shore, and which has recently been made use of to enclose a marine lake, and another similar bank enclosing some marsh land is to be found at Lytham. From St. Anne's and Blackpool to Rossal the drift is northwards, bending round this coast with the set of the tide eastward to Fleetwood. From Whitehaven the drift is southward, following the tide into the estuary of the Duddon up to Millom. Southward of the estuary along Walney Island the set is again southward

to the end of the island, when it follows the tide up the back of the island. Thus on the north and south sides of Morecambe Bay the drift is in directly contrary directions.

On the East Coast the general set of the drift follows the set of the flood tide from north to south. There are, however, many deviations in bays and estuaries. Thus in the Wash the set is eastwards up to Wolferton, from Weybourne to Blakeney it is westerly, and along the coast of Cromer eastwards; while beyond this at Aldborough and Lowestoft its general direction is south. North of Harwich the shingle travels south-easterly; and until preventive measures were taken, part of it was extending in a long spit nearly across the harbour mouth, the remainder working round Landguard Point with the tides and travelling along the beach of the harbour in a north-easterly direction.

In the English Channel the general direction of drift is easterly with the flood tide, but, moving round several of the indents in Lyme Bay, the direction varies from north-east to south-east. In both Weymouth, Bournemouth, and Chichester bays there are counter-tides working into the bays round the eastern headlands, and setting in opposite directions, and the drift of the shingle follows the same course.

In Newhaven Bay the drift was originally eastwards. The projection of the harbour wall has caused a counter-set of the tide, which now strikes Seaford Point and curls round into the bay westwards. Since this has occurred the direction of the drift of the shingle has altered, and the beach at Seaford has become denuded, causing the destruction of a sea-bank and of a concrete wall which had been erected for the protection of the shore, and the material has accumulated at the west end of the harbour, at the back of the pier.

From Bognor to Beachy Head the general direction of the drift is almost due east; from Beachy Head to Dover it is northeast; from Dover to the North Foreland, northerly; and from the North Foreland to the mouth of the Thames, westerly; and north of the Thames, south-westerly. In all these cases the shingle moves in the same direction as the set of flood tide.

On the north coast of France the prevailing winds are as on the other side of the Channel, from the south-west, and the main set of the flood tide from west to east. The drift of material from the waste of the chalk cliffs between Barfleur and Isigny is from north to south, the flood tide setting along the shores of the bay