RIVER ENGINEERING. The improvement of rivers ma,y be considered under two aspects for rivers form the natural channels for conveying the surp'lus rainfall from the districts through which they pass to the sea, and they can also be utilized for the purposes of inland navigation. If a river, owing to the small section of its channel, or the slight inclination of its bed, is incapable of discharging the whole volume of water which drains into it in rainy seasons, the lands along its banks become flooded, frequently to the great detriment of the crops, and sometimes with disa,strous results to life and property. If, on the other hand, a river is impeded by rapids, by shoals, or by a bar at its mouth, it is prevented from serving as a. natural highway for the traffic of the district through which .it flows. Accordingly the mitigation of floods and the regulation of rivers are the problems which have 'to be grappled with in the engineering of rivers. The first aims at remedying an existing evil, and the second deals with the development. of the resources and trade of a country by the improve-ment of its water communications.
Floods. - Floods are of twO kinds, according to the nature of the country traversed by the rivers producing them. Tor-rential rivers, flowing over impermeable strata and having a rapid fall, rise rapidly after a heavy rainfall, and produce a high flood which quickly subsides. Gently flowing rivers on the contrary rise slowly, and do not attain the same height as torrential rivers ; but their floods subside slowly, and consequently, though less high, remain longer on the land than torrential floods. The valleys, moreover, of torrential rivers are steeper and less fertile than the alluvial plains of gently flowing rivers ; and, consequently, the high short floods of the former are less injurious than the long continuing lower floods of the latter. The long duration of a flood is also the more prejudicial, as some-times a flood remains long enough on the land for a second flood to come down before the first has subsided, thereby producing an increased rise.
Floods are generally largest in the winter months, for, owing to the absence of evaporation, a much larger pro-portion of rainfall finds its way into the rivers at that period; and the greatest floods occur when rain falls on melting snow. High floods, however, sometimes occur in the summer after an exceptionally heavy rainfall ; and they are necessarily far more injurious at that time of the year devastating the crops on the land which they inundate.
Floods are due to the inadequacy of the river channel . to carry off the water poured into it within a given period. The bed of a river, being formed and maintained by its stream, is merely adequate to carry the ordinary dis-charge. Large floods occur at too distant intervals to scour a sufficient channel for their passage, and conse-quently they overflow the banks and inundate the adjacent districts. Other causes, moreover, tend to aggravate this evil.
A river carries down a large amount of solid matter which has been either ground from the mountain rocks by glaciers, washed from the land by the inflowing streams, or thrown into it as refuse from towns and manufactories. This material tends to settle in the channel wherever the current is checked, and consequently raises its bed and impedes the flow of the stream. Moreover, a river flow-ing through a plain gradually increases its serpentine course, thereby dirninishing its fall and reducing its velo-city. Accordingly the tendency of rivers is to deteriorate when left to themselves ; and the discharging capacity of their channels becomes less, whilst the extension of subsoil drainage causes the rain to flow more rapidly and com-pletely into the river upon whose basin it falls.
The available fall of the river is frequently diminished by the erection of fixed weirs across the channel, at various places, for the purpose of forming a head of water for mills, or of providing still-water navigation. These weirs are generally constructed with high sills, and of inadequate width ; and where flood openings closed by draw-doors are adopted the doors are frequently not fully raised till a flood has actually arrived. These weirs consequently not merely reduce the discharging capacity of the channel by diminishing the available fall of the water surface, but also actually restrict the section of the channel. The result is that floods occur more frequently, rise higher, and remain on the land for a longer period.
Prevention of Floods. - The entire prevention of floods would entail a larger expenditure than the results would justify. In most cases, the prevention of summer floods and the mitigation of winter floods would suffice ; for, whilst summer floods are always very injurious winter floods prove sometimes beneficial in depositing ;he mud which they bring down, provided they do not remain very long upon the land. Occasionally, however, where large tracts of low-lying country are exposed to inundation, and especially where portions of towns are below the flood level, it is necessary to extend the protection so as to ensure entire immunity from floods.
There are three methods by which floods may be pre-vented or mitigated, namely, - (1) improvement of channel ; (2) embankment of channel ; (3) pumping.
(1) Improvement of Channel. - The discharging capacity of a river may be increased by enlarging the section of its channel ; by the formation of straight cuts, which reduce its length, and consequently increase its fall; by dredging away shoals, and thus rendering the fall of its bed more uniform ; and by removing obstacles to its flow at weirs.
The channels of the English Fen rivers have been enlarged and straightened, and additional straight drains have been excavated for the more effectual drainage of the lo»--lying Fen country. Catchwater drains have also been formed to collect the rainfall of the higher lands and convey it into the river lower down, thus gaining a better fall than could be obtained if the upland waters were allowed to flow down to the low lands, besides relieving the low lands of this additional discharge.
Straight cuts are very useful when the fall is slight and the velocity of flow is consequently small ; but they are not so suitable for more rapid streams, and, besides modifying the flow, and thus tending to produce shoals below, their straight course is liable to be altered by the irregularities of the current, especially if joining a sharp bend above. The banks in such cases need protection against erosion, which adds considerably to the cost of the works. The improvement of the upper Mississippi by cut-offs has not proved satisfactory ; and it has been found preferable to train the river by brushwood mattresses and dykes.
Solid weirs across a river form serious impediments to its flow ; and draw-door or movable weirs should be pro-vided with adequate waterways, and sills level with the bed of the river. Old bridges, also, with wide piers and narrow arches retard the discharge of a river, and their rebuilding with wide openings would afford considerable relief. All obstacles in the river bed, such as weeds, fallen trees, and refuse, should be periodically removed; and fish-traps at weirs should be discarded, as they collect floating leaves and rubbish in their meshes and thus soon become entirely blocked up.
(2) Embankment of Channel. - When it is essential that the lands bordering a river should be absolutely protected from inundation, the enlargement of a river bed to an adequate extent for discharging the greatest floods would be too costly, especially when the fall is small; and it becomes necessary to resort to the expedient of increasing the channel, above the surface of the ground, by forming embankments along each side. By making the banks with material excavated from the channel, the earthwork serves the double purpose of enlarging the river bed and forming a bank. A flood channel of considerable dimensions can be readily obtained by placing the embankments some distance back from the margin of the river, thus greatly enlarging the section when the waters rise above the level of the land, whilst leaving the natural river bed unaltered for the ordinary flow. In some cases merely low embank-ments are constructed, which retain small floods but are submerged when large floods come down. Embankments, however, formed to secure the surrounding country from inundation must be high enough to exceed the highest flood level of the river, strong enough to resist the pres-sure of the water at that level, and perfectly watertight. If water can percolate through the bank, a breach is readily formed ; and if a high bank is overtopped by the river, the rush of the stream over it soon destroys a portion of the embankment and produces a disastrous inundation from the large volume of water suddenly liberated. The Fens of Lincolnshire, a large portion of Holland, the valley of the Po, and large tracts of low-lying land bordering the Mississippi are protected by embankments.
The defects in the system of embanking rivers are - that weak points in the banks are liable to be breached ; that the banks are liable to be overtopped by unusually_ high floods; and that the muddy waters of the river in flood deposit their sediment in the bed of the river, instead of spreading it over the adjacent land, and thus gradually raise the river bed and consequently the height of the floods. The remedies for these defects are strong high banks made of the best materials, and a periodical cleans-ing of the river bed. Neglect of these precautions has led to serious disasters. Large tracts in the Fens have been occasionally flooded by the bursting or overtopping of badly constructed banks. Numerous breaches have occurred in the embankments of the Po, resulting in the devastation of its valley; and the flood level of the Po has been so much raised that it has been decided not to heighten the embankments for fear of occasioning still greater disasters. The gradual silting up of the river Theiss, near Szegedin, produced a rise in its flood level which led to the overtopping of the protecting embank-ment in 1879, and the formation of a breach ; and the water thus set free destroyed a portion of Szegedin, and inundated a tract of 200 square miles. The rise of the bed of some rivers in Japan, from the deposit of silt, has been followed up by the gradual raising of the embank-ments ; and this system has been carried out to such a degree, and the accumulated deposit is so great, that some of these embanked rivers have their beds as much as forty feet above the level of the plains over which they flow. These high embankments nece,ssarily require constant attention ; and any failure is attended with serious in-undations. They serve as a warning against the extensive raising of embankments to counteract the silting up of a river.
(3) Pumping. - When lands are very flat and low, lying sometimes actually below the general drainage level of the district and the waterlevel of the streams, it is impossible for rivers to perform their ordinary function of draining the land by gravitation. It is necessary in such cases to create an artificial fall by pumping the drainage waters up so as to be discharged into the adjacent streams. This method has the advantage of ensuring the effectual drain-age of the lands, provided adequate pumping power is supplied; but it forms an additional tax on the land, as steam has to be applied to do what is under ordinary con-ditions effected by nature, and the land has also to be surrounded by banks.
This system has been adopted for the drainage of the Haarlem Meer reclamation, and also for the lands re-claimed from Lake Y in the construction of the Amster-dam ship canal. The drainage of the Fens is, in several instances, supplemented by pumping ; and a portion of the Witham basin has been secured against floods by this means.
The formation of large reservoirs in river valleys has been proposed for storing the surplus waters till a flood has subsided. A reservoir has indeed been formed, by constructing a high masonry dam across a narrow gorge of the Furens valley, which both supplies the town of St Etienne with water and preserves it from inundation. It is also proposed to prevent the floods of the river Chagres from interfering with the Panama Canal, by impounding its flood waters in an extensive reservoir to be formed by building a high dam across a suitable point of its valley. In these cases, however, the deep valleys with their narrow gorges are peculiarly well adapted for the forma-tion of reservoirs having a considerable capacity, whereas most river valleys are unsuitable for reservoirs, and the construction of the long lengths of the banks that would be.required would entail a very large expenditure, so that this system could only have a very restricted application.
Improvements of the Upper Portion,s of Rivers. - Most rivers are not suitable by nature for navigation in their non-tidal portion, or, in the case of tideless rivers, at a considerable distance from their mouths, as the fall of their bed increases towards their source, and they gene-rally present irregularities in depth and flow, with occa-sional sharp bends. Even where the depth is adequate, an irregular or rapid flow offers a great obstacle to up-stream traffic. Moreover, the fall of the waterlevel in dry seasons would often make a river too shallow for navigation. Accordingly, whilst improving the worst bends and remov-ing shoals, it is frequently necessary to retain the water, when the flow is sinall, so as to maintain a sufficient depth. This is accomplished by dividing the river into a series of sections or reaches, and pounding up the water at the end of each reach by means of a dam or weir.
Formerly rivers used to be penned in by a series of stanches near shoal places, which held up the water, and, when several boats were collected in the pool above a stanch, it was suddenly opened, and the sudden rush of water floated the boats over the shallows below. This primitive method of navigation, termed flashing, was formerly practised on the Thames and the Severn, and also on some of the rapidly flowing rivers in France, such as the Yonne. The stanches on the Severn were removed in 1842 ; but a few still exist across the Thames above Oxford, where the barge navigation follows a lateral canal. These stanches, consisting of be-ams swung across the river and supporting a series of spars and paddles, were easily removed in flood time or for the passage of boats. The stanches on the Yonne, which were more recently erected, were of a more elaborate description, known as needle weirs, and are still retained as weirs for holding up the water, though the process of flashing has been discon-tinued.
As the demands of navigation increased, these primitive methods proved inadequate, and, moreover, they were quite unsuited for up-stream traffic. Accordingly weirs were substituted for stanches, to hold up the water in each reach ; and locks were constructed, in suitable side channels, for enabling vessels to be passed from one reach to the next with little loss of water, and with equal facility either up or down. Rivers have been thus converted into still-water navigations, with level reaches forming a series of steps, having a fall at each lock, in place of the natural inclination of the river bed ; so that the up-river traffic is in a great measure relieved from the serious hindrance of an opposing current. In order that the water held back by the weir may be retained within the channel, it is necessary to raise the banks on each side for some distance above the weir ; and, as the gradual rise in the river bed towards the upper end of the reach reduces the depth, it is necessary to deepen the river along the upper portion of the reach to secure a uniform draught of water. A river is thus practically converted into a canal, with this sole difference that it has still to discharge the drainage waters of its basin. This primary object of rivers, to which indeed they owe their existence, was in many instances somewhat overlooked when rivers were utilized for naviga-tion ; and weirs appear to have been often regarded merely as dams for retaining the water, rather than as regulators of its flow.
Weirs. - Locks have been already considered in the article OR CANALS (q.v.); so that it will suffice here to describe briefly the different forms of weirs, which are essentially river works.
Weirs have been divided into three classes, namely, overfall weirs, draw-door weirs, and movable weirs.1 Ovelfall Weirs. - An overfall weir is a solid barrier placed across a stream for the purpose of raising the waterlevel (Plate V. fig. 12), and only affords an outlet for the discharge of the river when tho water rises above its crest. The waterlevel of the river is thus per-manently raised, not merely in dry seasons, but also in flood time, and its fall is correspondingly reduced. Accordingly, the dis-charging capacity of a river is materially diminished by the erection of overfall wehs ; and this is only partially remedied by the deepening of the channel for navigation, especially as deposit more readily accumulates in the lower part of a reach, owing to the reduction in velocity of tho current by the conversion of the river into level reaches. Endeavours have been made to alleviate this defect by placing the weir in a wide place on the river, and at an angle to the cross section of the channel, thereby increasing the length of its sill, and consequently the discharge over it for any definite height of the river (fig. 2). The gain in length of an oblique weir is somewhat neutralized by the weir not being at right angles to the direction of the current ; and, even if the cross section of the channel above the sill of the weir is as large as the avemge section of the river bed, tbe change in shap ,e and the small hydraulic radius of the section over the weir, check the discharge of the stream.
Draw-door Weirs. - In order to afford a freer flow than is attain-able with the best-designed overfall weir, draw-door weirs are sometimes adopted, which serve equally well to retain the water above during dry weather, and provide a large opening for the discharge of the stream in floods. Draw-door weirs consist of a row of doors, or sluice gates, sliding vertically in grooves formed at the sides of frames, piles, or piers, which are shut down when the flow is small, but are raised to admit the passage of flood waters. These weirs generally serve to supplement an ordinary overfall weir ; and, whilst the overfall weir regulates the flow in dry weather, the draw-door weir provides for its more rapid discharge in flood time. The relief, however, afforded by draw-door weirs depends entirely on the opening they furnish ; they are rarely, if ever, made equal in section to the raver channel, and their sills are usually raised some feet above the bed of the river ; but, nevertheless, they are much superior to overfall weirs in respect of drainage, especially when the river has a small fall and low banks.
A large oblique overfall and draw-door weir has been erected across the Thames, by Mr Leach, at the limit of its tidal flow at Teddington, having a total length of 480 feet. This weir is divided into four bays, the two side bays being overfalls, whilst the two central bays, 172i feet and 69i feet wide respectively, are closed by large iron draw-doors sliding in grooves at the sides of strong iron frames supporting a foot bridge from which the doors are raised. The frames rest upon piles, and on the top of a rubble mound raised about a foot above low-tido level.] The friction of large dmw-doors against the grooves, in being lifted, is considerable when there is a head of water on one side ; but this has been much reduced by Mr F. Stoney, in a large weir in Brazil, by making the doors, 20 feet in width, bear on each side against a row of free rollers suspended in the grooves.2 Movable Weirs. - Although draw-door weirs afford a much freer discharge for a river than overfall weirs, the vertical frames or piers in which the doors slide offer more or less impediment to the flow. This defect is avoided by movable weirs, which, whilst equally efficient in retaining the water when raised, can be entirely lowered or removed so as to leave the channel quite open in flood-time.
There are three types of movable weirs which have been regularly adopted abroad, whilst other forms have been occasionally tried. The two types most extensively used are the frame or needle weir and the shutter weir, whereas the drum weir has been only erected on the Marne.
The frame weir consists of a series of movable iron frames, placed at intervals across the channel of a river end on to the current, calving a foot bridge at the top, and supporting a wooden water-tight barrier which forms the actual weir. °Till recently the barrier was always composed of a series of long square wooden spars, or needles, placed close together and nearly vertical, resting against a sill at the bottom and against a horizontal bar con-necting the frames near the top (fig. 3). This type of weir has, accordingly, been very commonly called a needle weir (barrage a aiguilles); but this term would not now include every form of the fmme weir. The first needle weir was erected across the Yonne in 1834 ; and till 1881 all the weirs on the Seine below Paris were of this form. The needle weir is opened by lifting each needle successively from the foot bridge • one of the end frames is then disconnected from the rest by un'fastening and withdrawing the connecting bars, the corresponding portion of the foot bridge is taken up, and the frame, which is hinged at the bottom, is lowered by chains on to the apron of the weir, and the whole of the frames are sindlarly lowered in succession, leaving the passage quite clear. The weir is closed. again by a precisely reverse senes of operations.
As the weight of the needle, which should be readily lifted by one man, imposes a limit to the height of the weir, the large frame weir recently erected at Port-Villez near Vernon, having a height of 18 feet, has been-closed by a sort of wooden binged shutter which spans the interval of 3i feet between each frame, and can be rolled up from the bottom and removed when the weir is to be opened.
A series of similar hinged shutters are designed to close double intervals between the frames, about 7 feet wide, of the frame weir in progress at Poses, the next weir above Martot weir which is at the boundary of the tidal Seine above Rouen. Poses weir has a form quite distinct from all frame weirs hitherto constructed, for its vertical frames are suspended from an overhead girder, and rest against a sill at the bottom when down, but can be raised entirely out of water into a horizontal position when the weir is open. The girders carry a foot bridge, from which the frames and hinged shutters are raised and lowered, and rest on masonry piers dividing the river into seven bays, the two navigable passes being 106i feet and the five shallower passes 99 feet wide. The girders spanning the two navigable passes leave a dear headway of 17i feet above the navigable high water, whilst the rest of the girders are merely placed above flood level. The fall at the weir is 13 feet. The system is costly, with its girders and piers, but it secures all the movable parts of the weir from injury in flood, and enables the weir to be worked with perfect ease and safety.
Sliding panels have been adopted for closing the frame weir across the Rhone at Mulatiere, near Lyons, erected in 1882; and a comparison is being made of the relative durability of sliding panels and hinged shutters by placing the two systems side by side at Suresnes weir just below Paris.
The earliest form of skutter weir consisted of a gate, or shutter, 1 turning on a horizontal axis at the bottoni, supported by a prop • when raised against the stream, and falling flat on the apron of the weir when the prop was withdrawn. As considerable force would be required to raise such a shutter against a strong stream, a second up-stream gate is usually provided, which, rising with the stream and being retained by chains, relieves tbe pressure on the down-stream gate, and enables it to be readily raised and propped up. The waterlevel is then equalized on both sides of the upper gate, which is then lowered ; and the lower gate forms the actual weir, which can be opened by merely releasing the prop. In India, where this form of shutter vreir has been adopted on a large scale, the strain on the retaining chains was so great, when the upper shutter was raised in a strong current to shut off the river from the irrig,ation canals, that hydraulic brakes have been substituted, by Mr Fouraeres, for controlling the motion of the up-stream shutters. A closed cylinder full of water is fixed on the apron of the weir above the upper shutter, in which a. piston, attached to the upper side of the shutter, is fitted. Directly the current tends to lift the shutter, the piston is drawn against the cushion of water in the cylinder, which controls its motion. The pressure of the piston forces the water gradually out of some small orifices along the side of the cylinder, so that the piston is enabled to travel slowly along the cylinder. In its progress, however, it passes by some of the orifices, whereby the rate of efflux of the remaining water is reduced, and a greater resistance offered to the motion of the piston. Whilst therefore the shutter in rising presents a greater surface to the stream, and consequently exerts a greater pull upon the piston, the retarding force is similarly increased, and the shutter is thus gradually raised without any jar.
At Brulee Island weir, on the Yonne, the up-stream shutter has been dispensed with; and the shutter forming the weir is raised against the stream by a piston working in a hydraulic press (fig. 4), the water pressure being supplied from an accumulator which is charged by means of a turbine worked by the fall of water at the weir. The weir is closed by seven shutters, feet long and 6 feet high, which can be raised in five minutes ; and the power for damming up the stream is actually obtained from the stream itself.
The form of shutter weir most commonly adopted in France is shown in fig. 5, representing a section of the largest and most recent weir of this type, erected at Port-ii-FAnglais weir on the Seine, two or three miles above Paris ; and weirs on the same system have been erected across the Great Kanawha river in the United States. The shutter revolves upon a horizontal axis placed just above the centre of pressure on the down-stream side of the shutter. The axis is fastened on an iron tressel hinged to the apron of the weir ; and the shutter and tressel are supported in position by a wrought-iron prop resting against a east-iron shoe fixed on the apron. When the weir is closed, the shutter butts against a sill at the bottom, as is shown on fig. 5. The weir can be more or less opened, from the foot bridge, by means of chains fastened to the top and bottom of the shutter ; and it can be completely lowered in flood time by releasing the prop from its shoe, when the prop, tresscl, and shutter fall flat upon the apron, their fall being regulated by aid of the chains ; the frames also supporting the foot bridge, being hinged to the apron, can be lowered as in the ordinary frame weir. The weir is raised by firstl reinstating the foot bridge, and then raising the shutters, with the connected tressels and props, by means of tho chains. Each shutter of the navigable pass at Port-h-l'Anglais is 3i feet wide, and. rises 12i feet above the sill of the weir. Smaller shutter weirs, of a similar form, are placed on the top of overfalls to regulate the discharge, being so adjusted that the shutters dip when the water attains a certain height above them. Sometimes the regulation of the flow is effected through the large shutters by means of small shutters fixed in their upper panels, called butterfly valves, which open spontaneously when the water rises above the required height. The smaller shutters on the overfalls are entirely lowered in flood-time, like the large shutters, but, being un-provided with a foot-bridge, they are raised by aid of boats on the approach of the dry season. The earlier shutters erected across the navig,able passes were similarly raised; but, though a foot bridge adds considerably to the cost of a shutter weir, it greatly facilitates its working.
The navigable passes in French rivers, which are always closed either by frame or shutter weirs, serve for the passage of vessels during floods when the locks are submerged.
With the exception of the priinitive movable stanches, there is only one example of a movable weir in England. This self-acting shutter weir has been erected across the Irwell at Thostlenest near Manchester ; and a section of it is given in fig. 6. The weir consists of a series of shutters turning on a central horizontal axis. When the weir is closed, each shutter is inclined at an angle of 35° to the vertical, as shown in fig. 6, and revolves t,o a horizontal position for opening the weir. It resembles in fact the French shutter weirs, except that the shutters and their supports are not removed from the channel, so that the waterway at Thostlenest is not so unimpeded as in the French system. The shutters are so adjusted that they open when the river rises 2i feet above its ordi-nary level ; but, lest the rush of water, which would result from their sudden opening, should injure the river bed, an arrangement has been made for opening any of the shutters by means of a set of chains worked by crabs from each bank, so as to release the pent-up waters more gradually.' This weir, desig,ned by lir Wiswall, consists of fourteen shutters, each 10 feet wide above the axis and 9 feet below, and 12 feet long. The actual height of the weir above the floor is only 10 feet, owing to the inclination of the shutters, so that it presents a surface of 1400 square feet to the stream when closed, which is reduced to 293 feet when open.
The drum weir, which has been adopted. in several instances on the river Marne, consists of an upper and au under iron paddle capable of making a quarter of a revolution round a horizontal central axis. The upper paddle forms the weir, and the under one revolves in a closed recess, shaped like the quadrant of a cylinder, laid below the sill of tho weir, from which the term drum is derived (fig. 7). The under paddle and the drum are so formed that a space is left between the upper face of the paddle and the top of the drum when the paddles are horizontal ; and a similar space exists between the down-stream face of the paddle and the vertical wall of the drum when the paddles are vertical, as represented in fig. 7. These spaces serve as sluiceways by which water can be admitted into the series of drums on the upper or under side of the under pad.dles. The weir is closed by placing the upper sluiceway in communication with the upper pool, when the pressure of water on the upper faces of the under paddles overcomes the pressure of the stream upon the upper paddles, causing them to rise, and closing the weir against the stream. The weir is readily opened by shutting off communication between the upper pool and the upper sluiceway, and opening communication with the lower pool ; the stream then depresses the upper pad.dles, or the action can be quickened by opening communication between the upper pool and lower sluiceway. The pressure on either side of the under paddles can be easily adjusted with the utmost precision, enabling the paddles to be placed at any angle, so that the most absolute control is obtained over the discharge at the weir. The largest example of this type of weir is at Joinville, on the Marne, only a few miles above Paris. The weir is formed by forty-two paddles, 3i feet high and 41 feet wide, which are worked with great ease by rneans of sluice gates on the left bank, being opened or closed in three or four minutes by one rnan. This type of weir has the defect of not being suited for navigable passes, owing to the depth of the foundations for the drum below the floor of the weir having to exceed the height of the actual weir.
Movable weirs possess the great merit over other forms of weirs of offering little or no impediment to the passage of floods ; and this advantage is still further enhanced by the system of warnings, organized for the Seine, the Loire, and other French rivers, whereby tixnely information of the approach of a flood is telegraphed to the various weir keepers, so that they may fully open the weirs before its arrival, and thus aid in facilitating its descent. By teleg,raphic intimation of the rise in the upper tributaries, and of the rainfall in the basin, it is possible to predict with remarkable accuracy the probable rise of the river at places lower down, and to afford valu-able warning of a coming flood to the riparian proprietors.
Rivers may be broadly divided into two classes in respect of the lower portion of their course, for the tide is propagated up some rivers to a considerable distance from their mouth, commingling with the fresh water and pro-ducing an ebb and flow far into the interior ; whilst rivers flowing into tideless seas descend with an unimpeded current to their _outlet. Tidal and tideless rivers, possess-ing very distinct' physical characteristics, necessarily pre-sent different featur6s and require different methods of improvement, and will be therefore separately considered. Tidal rivers are the more numerous, owing to the greater extent of tidal seas. The great differences also in the tidal rise introduce numerous variations in the tidal influ-ence on rivers, and the rise of the river bed detertnines the distance to which the tidal flow extends. Accord-ingly, tidal rivers exhibit a greater diversity in their natural condition than tideless rivers, which are only affected by the volume of their fresh-water discharge, the amount of sediment carried in suspension, and the inclina-tion of their bed. These latter conditions affect also the state of tidal rivers, but their influence is greatly modified by the ebb and flow of a large volume of tidal waters. The effect of the tidal ebb and flow is most readily per-ceived in contrasting the mouths of tidal and tideless rivers. The mouths of the Mississippi, the Nile, the Danube, and the Rhone present very marked differences to the outlets of the St Lawrence, the Seine, the Thames, and the Severn. Tideless rivers divide into a number of mouths, whereas tidal rivers are confined to a single out-let ; and the effect of tidal influence on this difference is still further confirmed by the instance of the Maas, which, with a very slight tidal range, exhibits a tendency to deteriorate into the dispersion of mouths of a tideless river. The value of tidal flow in maintaining a river is fully manifested by comparing the navigable condition of the Thames or of the muddy Humber with the delta of the Nile or the Rhone, though the latter rivers possess a much. larger fresh-water discharge. The tidal rise also frequently allows of the access of vessels to a river whose entrance is barred at low water.
The general improvements of the upper portions of both tidal and tideless rivers may be carried out on similar principles, though on approaching their mouths they need a totally different treatment. To give a river a uniform depth, its channel and flow require regulation. Hard shoals may be permanently removed by dredging; but silty shoals, even when dredged away, will re-form unless the channel is contracted. Formerly rivers were regulated by building out jetties at intervals at right angles to the banks, especially in wide shoal places, in order to contract the channel and concentrate the stream, so as to scour a deeper central channel. These cross jetties, however, whilst effecting a deepening in front of their extremities, caused irregularities in the flow of the current in the intervals between them, thus producing differences in depth. Continuous longitudinal jetties, or training banks, though more costly, are much more efficient in regulating a river, and are now generally adopted for procuring a uniform width, and, consequently, a regular depth. Where the fall alters, a corresponding variation must be made in the section of the channel ; and, in the case of tidal rivers, the section should gradually increase as it approaches the sea, so as to admit the increasing volume of sea water which enters but does not pass far up the estuary whose upper portion has been filled by the earlier flow.
Most rivers, whether tidal or tideless, are more or less impeded for navigation by a bar at their mouth. A bar is a ridge or shoal extending across the navigable channel, over which there is a less depth than either above or below it. The lowering of such a bar forms one of the main objects of river improvement, as upon the depth that cau be obtained over the bar depend the class of vessels that can enter the river, and, in tidal rivers, the period of time during which the entrance can be navigated. A bar may result from the action of the sea, which tends to form a continuous beach across any inlet, and would obliterate the mouths of rivers if the channels were not maintained by the ebb and flow of the tide and the fresh-water dis-charge; or it may be formed by the conflict of the sea and river water, which checks the current at the mouth and causes the river to deposit the sediment which it held in suspension. The bars at the mouths of the Mersey, the Liffey, and the Adour are due to the first cause; Whilst the bars at the mouths of tideless rivers, such as the Missis-sippi, the Danube, and the Rhone, are mainly due to the second.
Improvement of Tidal Rivers.
Tidal rivers differ greatly in their natural character-. istics, owing to the variety in the different conditions which affect them. Thus the Mersey, with an extreme tidal rise of 30 feet at its mouth, is only tidal for 46 miles, whilst the Seine, with a rise of 22 feet, is tidal for. 91 miles, and the Scheldt, with a rise of only 13i feet, is tidal up to Ghent, a distance of 105 miles. These differ-ences are mainly due to the different falls of the river-beds, but they are also affected by the facility of entry afforded to the flood tide, and the form of the channel up which it flows. The tidal capacity of a river depends on the rise of the tide and the configuration of the banks. The tidal flow into the Mersey amounts to 710,000,000 cubic yards at a high spring tide, whilst the flow into the Scheldt at Flushing, with less than half the tidal rise, reaches 475,000,000 cubic yards. The tidal capacity of the Seine, together with its estuary, formerly exceeded that of the Mersey, but it has been greatly reduced by the training works which have been carried out on it since 1848.
e The tidal ebb and flow passing and returning through the entrance channels of a river twice a day exercise a very important influence on its maintenance. The effect of tidal scour is manifested in the history of the harbours of Calais and Ostend,1 which in old times possessed deep outlet channels, but were injured by reclamation, and in the deterioration of the outfalls of the Fen rivers as soon as the tidal flow was curtailed by the erection of sluices. The power of the tidal scour necessarily varies with the volume of water producing it, and therefore one of the first principles of tidal river improvement is that the tide should be admitted as far up a river as possible, and all obstructions to its flow removed. If, however, the main-tenance of an estuary depended solely upon the tidal ebb and flow, the estuary would gradually silt up, for the flood tide brings in matter in suspension which it washes from the adjacent shores and sandbanks, especially during rough weather, when the waves stir up the sand and silt. The impetus of the tide running up the rising bed of an estuary is gradually checked, till at last slack water occurs, and the silt begins to deposit, which the ebb tide, enfeebled by the friction of the tidal water in its passage up and down the estuary, would of itself be unable com-pletely to remove. The erection of any obstructions to the tidal flow, such as sluices and weirs, increases the period of slack tide, and consequently not only reduces the volume of ebb and flow but also promotes the deposit of silt, to the further detriment of the estuary. The main-tenance of estuaries is secured by the aid of their freshwater discharge, which, being penned up during the flood tide, reinforces the ebb and preserves an equilibrium.
The fresh-water discharge of a river, depending upon the area of the basin and the available rainfall, naturally varies greatlyin different rivers - being, for instance, greater in the Tyne and the Clyde than in the Mersey, though these rivers have little more than one-thirteenth of the tidal capacity of the Mersey. The Seine, with a drainage area of 30,500 square miles, nearly six times the size of the Thames basin, has a discharge of 28,000,000 cubic yards, on the average, each tide (about twenty-eight times that of the .Mersey), though this volume sinks into insigni-ficance when compared with the flow of the Danube with a basin of ten times the size, or still more of the Missis-sippi with a basin forty times as large. A large fresh-water discharge is of great value to the maintenance of an estuary, especially when, as in the case of the Seine, it carries little silt in suspension; whilst a small discharge in proportion to the size of an estuary, of which the Mersey is a notable instance, renders the state of the estuary very delicate, and necessitates great vigilance in maintaining its tidal capacity, to which its existence is almost wholly due.
The silt brought down by tidal rivers, instead of being carried to their mouths and there deposited, is met by the incoming tide at points varying daily with the states of the tides; and, moreover, except during slack tide, it is maintained in constant motion up a,nd down the estuary, till at length it gets to the sea. Accordingly, though the sediment of tidal rivers is more or less deposited wherever the velocity of the current is checked, it does not tend to accumulate in one particular part, as in the case of tideless rivers, and therefore the formation of a bar at the mouth is mainly due to the drift by waves along the beach. A flood also, though more largely charged with silt, by giving additional power to the ebb, scours the channel and lowers the bar. The Humber, whose waters are densely burdened with mud which is readily deposited in still water, is nevertheless free from a bar.
The best form of estuary for a tidal river is when it enlarges gradually as it approaches the se,a, thus affording an increasing capacity for the admission of the tide, and promoting a regular flow. The estuaries approximating to such a form are generally free from bars - as, for instance, the Tharnes, the Severn, and the Scheldt. When, how-ever, a river expands abruptly into a wide estuary on a sandy coast, it winds through the enlarged estuary in an unstable shallow channel, owing to the reduced velocity of the ebb in expanding out, and the checking of the flood tide on reaching the head of the wide estuary. Thus the Seine, with a deep stable channel from Rouen to La Mailleraye, had formerly a shallow shifting dangerous channel from thence to the sea ; the Ribble, with a good depth at Preston, has a shoal irregular channel towards its mouth ; and the Dee, with a moderate channel at Chester, is almost barred to vessels, except at high tide, below Connah's Quay. These estuaries do not possess a well-defined bar, but their long shallow winding channels offer a still more serious impediment to navigation. The worst form is a very irregular estuary with abrupt expan-sions and contractions, of which the Mersey is a pro-minent example, for, in spite of its large rise of tide, it possesses a shallow, irregular, and shifting channel above Liverpool, and is encunabered by a wide bar below.
There are three obstructions to which tidal rivers are subject, namely, a bar, a shifting channel, and inadequacy ( of depth; and there are three general methods which may be resorted to for their improvement, namely, jetties, training walls, and dredging, in addition to the regulation of their upper portion by longitudinal jetties, or banks, as previously mentioned.
Jetties.--A bar being caused by the littoral drift, and by the impotence of the expanded current to scour the channel over it to the same depth as elsewhere, it is neces-sary either to arrest the drift or to concentrate the current across the bar. The drift, which comes from the direction of the prevalent winds, might be temporarily arrested by projecting a groyne, or jetty, from the shore on the wind-ward side of the outlet. The material, however, carried along the coast would accumulate against the jetty, and eventually form a bar beyond, or, sweeping round the end of the jetty,. deposit in the channel under its shelter. Accordingly a second jetty is added, on the opposite side of the outlet, to direct and concentrate the current over the bar, and thus increase the depth of the channel, and also to drive into deep water any material that may be carried round the windward jetty, or convey it within the influence of any littoral current farther out. The jetties are either made parallel, or slightly diverging, so as to form a sort of continuation of the banks of the river across the beach, or they are commenced far apart and made to converge towards their extremities, so as to admit a larger volume of tidal water and concentrate the flow into a narrow channel over the' bar. The parallel jetty system has been adopted for the new outlet of the Maas (fig. 13), the mouth of the Adour near Bayonne (fig. 16), and the mouths of the Wear at Sunderland, the Yare at Yarmouth, and the Ouse at Newhaven ; whilst the con-verging jetty system has been carried out at Charleston (figs. 17 and 18) and Aberdeen, and at the mouths of the Liffey, the Tyne (fig. 8), and the Tees.
The ordinary form of jetty is a timber pier resting upon a base of rubble stone, like the jetties of the North Sea jetty harbours of Calais, Dunkirk, and Ostend, so that the solid lower portion may concentrate the ebb, whilst the open upper portion permits the passage of the littoral currents in order that a rapid advance of the foreshore may be pre-vented. Such structures, however, simply delay, and do not stop, the advance of the foreshore, as manifested at Newhaven and Dunkirk, where the accumulation of shingle in the one case and of sand in the other brought low-water mark out to the extremities of their western jetties. The solid northern jetties of the Wear and the Yare have naturally produced a similar advance of their northern beaches, both being exposed to a north-easterly drift.
A special form of jetty has been constructed at the mouth of the Adour to combine the advantages of open and closed jetties. These jetties consist of a row of cylindrical columns placed at intervals and carrying iron girders on the top. Grooves are formed at the sides of the columns, down which panels can be lowered from the roadway above to confine the issuing current, whilst when the panels are open the spaces between the columns admit the flood tide and the passage of the currents. Nevertheless there are indications of an advance of the foreshore ; the sand passing through the spaces in the northern jetty has encroached upon the channel, and the depths are reduced beyond the end of the jetties (fig. 16).
The most important examples of training jetties, includ-ing converging jetties, have been made solid, though sometimes they have not been raised to high-water level at their outer ends, in order to provide for the freer admis-sion of the flood tide. The converging jetties, or walls, at the mouth of the Liffey consist of mounds of rubble stone, and the outer portion of the northern jetty is only raised to half-tide level. The jetties at the mouth of the Maas (figs. 13 and 15) are formed of fascine mattresses ; and the Charleston jetties (figs. 17 and 18) are similar in construction ; in both these cases the outer portions have not .been raised above half-tide level. The converging jetties or piers at Aberdeen, the Tyne (fig. 8), and the Tees are in reality breakwaters,' though they serve the same purpose of protecting the entrance channels from the littoral drift, and promoting scour over the bar, as the less solid structures at Dublin and Charleston. The Tees breakwaters are random mounds of slag; the Aber-deen breakwaters at the mouth of the Dee are upright walls of concrete ; and the Tynemouth piers are masonry and concrete-block walls upon a rubble foundation. The breakwaters afford a much better shelter for vessels and for dredging operations, but the lower fascine-work jetties at the mouth of the Maas are equally effective in directing the current.
Training Walls. - The wandering shallow channel of a river through a wide sandy estuary naay be improved by training the channel, in a suitable direction, by means of longitudinal mounds of rubble stone, commonly termed training walls. These walls fix the channel and prevent the current eroding the sandbanks and thus changing its course. Moreover, by guiding the channel into a more direct line, making the ebb and flow follow the same course, and concentrating the current, the scouring capa-city of the stream is increased and the channel is deepened. The flood tide ascends the trained channel more readily, and therefore is able to extend its influence farther up ; whilst the ebb tide flows out of the improved and deepened channel earlier, and thus lowers the low-water line and increases the tidal capacity in the channel.
The trained channel must be gradually widened out and carried into deep water, otherwise the abrupt expansion which occurs beyond the ends of the training walls would so enfeeble the ebbing current that a shallow shifting channel would be formed only a short distance below. Training walls which stop in the middle of a wide sandy estuary, like the walls carried out on the Seine (fig. 10) and the Ribble, can be only regarded as incomplete works, which sooner or later will have to be extended if the full benefit of a trained channel is to be realized. When the channel is to follow close along one shore of the estuary, a single training wall on the outer side is suffi-cient ; and a single wall is sometimes adequate for main-taining a channel in the middle of an estuary, when placed along the concave side of a bend.
The proper width between the training walls depends upon the fall, the tidal range, and the fresh-water dis-charge, and should gradually increase down stream so as to admit as much tidal water as possible with a steady flow. As the scour of the fresh-water discharge is greater with a contracted channel, the tendency is to place the training walls too close together, which, though improving the depth in the channel between the walls, reduces the volume of tidal water that can get up the channel and thus compromises the maintenance of the outlet beyond the walls. The training walls on the Seinesand on the Ribble, whilst improving the trained channels, have been preju-dicial to the channels beyond; and an extension of the works has been authorized on the Ribble. The widths also a,dopted between these walls are not compatible with an adequate widening out towards their outlet for the free admission of the flood tide, so that these estuaries will eventually be deficient in tidal capacity.
The training of a wandering channel is always beneficial to the maintenance and depth of the channel. The wanderings, however, of the channel, which are thus arrested, though very prejudicial to navigation;are advan-tageous in preventing the silting up of an estuary by the constant erosion and stirring up of the sandbanks which they effect in shifting their position, and which they carry by gradual stages throughout the whole of the estuary. The scouring current predominates at one time in one part and at another thne elsewhere, so that slack water is never permanent in any part of the estuary, and accretion cannot progress for a long period without dis-turbance. When, however, a channel is permanently fixed by training walls, the condition of the estuary is com-pletely transformed. The flood tide, indeed, comes in with its burden of silt as before, rising sooner up the improved channel, and therefore dispersing with a somewhat gentler flow over the rest of the estuary. The ebb tide, however, is mainly concentrated along the trained channel, especi-ally when it attains its maximum scouring efficiency towards low water. Accordingly, whilst the flow in the trained channel is increased, stagnation occurs more or less over the rest of the estuary, and silting-up inevitably occurs, resulting eventually in a large reduction of tidal capacity. The accretion, moreover, is not confined to the portion of the estuary behind the training walls, but gradually creeps down, on each side of the estuary, for a considerable distance beyond the ends of the walls. Low training walls hardly rising above the adjacent sandbanks have been tried with the object of preventing this accre-tion, but the improved flow in the trained channel and the reduced velocity elsewhere still promote accretion behind the low walls, and the deposit, rising first along the shores of the estuary, gradually attains high-water level over a great portion of the area at the back of the training walls, and from thence slopes down to the top of the walls on each side of the channel. Though the pro-cess of accretion is less rapid with low training walls, the ultimate result is only delayed a,nd not prevented, as clearly manifested by observations on the Seine and other rivers, so that the view formerly entertained by some engineers, that if training walls were kept down to the level of the existing sandbanks no accretion would take place, has been proved to be erroneous by the results of experience; and it may be accepted as a general law that training walls, whether high or low, inevitably lead to accretion if the flood tide is charged with silt.
s The most careful consideration should be given to all the conditions of an estuary before training works are commenced, for when once begun they must be eventually carried out to deep water ; and, if ports exist along the shores of the estuary, they are liable to be injured by the accretion resulting from the works unless the trained channel can be led close along them. The training works of the Seine estuary, though very advantageous to the inland port of Rouen, are compromising the approach channels to Honfleur and Havre, and have silted up the port of Harfleur, so that the extension of the training walls to Honflenr has been urged in the interests of that port, whilst large works have been executed to preserve its entrance, and a new direct channel into the sea is being proposed for Havre. Though the training works on the . Dee have not been carried out hitherto in a judicious direction, having been formed mainly with a view to land reclamation, it would be advantageous for the ports of Chester and Connah's Quay to extend these works towards the sea, as there are no ports below the present limits of the training walls to be injured by the effects of their pro-longation, and the navigable channel would be much improved, provided the works were carried out to deep water. The training walls in the Ribble estuary must eventually be extended, even beyond the limits at present authorized, if a good navigable channel is to be secured to Preston, but these works will produce an entire trans-formation in the estuary, of which large portions have been already reclaimed as a consequence of the works already accomplished. The Mersey estuary would need a very comprehensive scheme for its improvement, owing to the very defective natural condition of the estuary, and the situation of the ports along its banks. The mere training of the channel in the upper estuary, for the bene-fit of the up-river ports, would result in the reclamation of the wide estuary between Runcorn and Liverpool, and thus deprive the channels between Liverpool and the sea of their natural scouring reservoir of tidal water ; whilst the training of a channel below Liverpool out to the bar would necessitate very extensive works in deep water and in an exposed situation.
Dredging. - The improvements effected within recent years in the ordinary dredging machinery, and the intro-duction of the sand-pump dredger, have facilitated and cheapened dredging operations to such an extent that some of the most remarkable river improvements have been effected by dredging. The great increase in depth realized on the Tyne and the Clyde has been effected by means of steam bucket-dredgers aided by hopper 'barges, whilst the maintenance of the entrance channel to St Nazaire on the Loire, and the deepening of the approaches to Dunkirk and Calais, have been accomplished by sand pumps, which have the advantage of being able to work when exposed to moderate waves. Dredging merely con-sists in removing material fron: the river bed and thus enlarging and deepening it ; and the extent to which this method of improvement may be carried simply depends upon the economical consideration as to how far the improvement of the traffic on the river by an increase of depth will afford an adequate return for the outlay. Dredging, however, furnishes a cheap method of excava-tion owing to the small cost of carriage by water. Dredg-ing, being a purely artificial means of improvement, generally necessitates regular maintenance ; whereas the improvement from scour effected by jetties aud training walls is permanent, being realized by natural means. Frequently, however, training walls and jetties are supple-mented by dredging, for the walls and jetties render the deepening by dredging easier and more permanent ; whilst, on the other hand, dredging enables a greater depth to be attained, and even maintained, than could have been effected by scour alone.
The improvements on the Tyne and on the Clyde have mainly resulted from very extensive dredging operations, but they have been aided by training walls on the Clyde, and by the Tynemouth piers on the Tyne, which protect the entrance channel from drift and the dredgers from waves, and concentrate the scour over the bar. The three methods of improvement described above have been re-sorted to on the Tees : for training walls have been formed through the wide estuary below Middlesborough for fixing the channel ; converging jetties are being constructed for sheltering the channel from wave-borne sand, and for directing the scour over the bar • and dredging is being employed for deepening the trained and sheltered channel. On the Maas also, and at Charleston, dredging is being used for attaining a depth for navigation which the jetties alone were unable to produce.
River Tyne Improvement IVorks.
The Tyne has a drainage area of 1053 square miles ; its tidal flow extends 18 miles from its mouth, and the range of spring tides at its outlet is 141 feet. Being by nature a small winding irregular river, with little tidal capacity and no estuary, its depth was sinall and variable, and a bar existed at its mouth, ‘shich opens directly on to the sea-coast. The first improvement works, com-menced in 1843, consisted in training the Jiver by cross jetties, subsequently connected by low trainin,,r, walls, so as to tegulate the width and consequently the depth of the river. As, however, the volume of water iu the river was small, the scour was not adequate to effect a great improvement in the depth ; the bed of the river between Newcastle and the sea, in 1860, was in many places above low-water level ; and the depth on the bar at low tide was only 6 feet (fig. 9). The piers at the mouth were commenced in 1856, for the purpose of facilitating the removal of the bar and for sheltering vessels entering the river, and were orig-,inally designed to terminate in a depth of 13 feet at low water ; but eventually a larger scheme was adopted for forming a refuge harbour in com-bination with the improvement of the river, and the piers, which are still in progress, are to extend into a depth of 30 feet. In 1861 extensive dredging operations were commenced for improving the depth of the river, and a maximum of 3,515,000 cubic yards was removed ill ,1866 ; the work has been regularly continued, and the total amount dredged since the commencement, in 1838, reached 49,668,000 cubic yards in 1884, being an average of nearly 2,000,000 cubic yards annually since 1861, when systematic dredging operations were begun. The improvement in depth that was effected between 1860 and 1884 is shown on the longitudinal section of the river (fig. 9); and the deepening of the river between Newcastle and Hedwin Streams, a distance of miles, is in progress, being carried on by six dredgers. The river has also been regulated. by making a straight cut across Lemington Point, and widening the channel from 150 feet to 400 feet opposite Blay-don ; the obstruction offered to the Tyne by the old Newcastle Bridge has been removed by rebuilding the bridge with larger openings ; and a sharp bend in the river has been eased-by removing the high projecting rock at Bill Point (fig. 8). The tidal capacity of the, river has been in-creased by 14,000,000 cubic yards ; the bal. has been lowered 14 feet since 1860 ; and the least depth at low water up to Newcastle is 20 feet, and 18 feet for 3 miles above. The deepening of the channel has produced a very beneficial lowering of the flood line in the river, thereby preservieg the adjacent lands from inundation.' The improvement of the river has led to a great increase in the tonna0.se of the vessels frequenting it, and a large developinent of its trade. The average tonnage of the vessels, whieh was 163 tons in 1863, had risen to 396 tons in 1883 ; and the total tonnage of the vessels entering and clearing the Tyne ports rose from 4,382,000 tons in 1863 to 13,043,000 tons in 1883.
River Clyde Improvement Works.
The Clyde is a sniall river, with a. drainage area of only 945 square miles, and a tidal flow of about 20 miles ; it opens, however, into a deep well-sheltered estuary, or arm of the sea, called the Firth of Clyde, and is free from a bar. The rise of spring tides at its mouth is about 10 feet. The Clyde was by nature an insignificant stream, with numerous hard gravel shoals, and a ford 12 miles below Glasgow which could be crossed on foot. The regulation of the river by cross jetties, and the removal of hard shoals, was commenced in 1773. Early in the present century the jetties were made more uniform, others were added, and their ends were eventually connected by low training walls which were gradually raised as deposits formed behind. The river had to be subsequently widened to accommodate the increasing- trade and a larger size of vessels. As scour alone could only produce a very moderate depth, systematic dredging operations were commenced in 1844, and reached a total of 28,648,000 cubic yards by the middle of 1884, the maximum accomplished in a single year (1878-79) reaching 1,502,000 cubic yards. Dredging is-still being continued with six dredgers in order to maintain as well as deepen the river ; for the channel up to Glasgow has been deepened so far beycnid its natural limit that any matter in suspension which enters the river is readily deposited. Of the 1,041,000 cubic yards dredged in 1883-84, as much as 703,000 cubic yards consisted of deposit, or more than two-thirds of the whole quantity removed. The river has a depth of 24 feet at high water from Glasgow to Port Glasgow, and from 13 to 15 feet at low water. The tide falls 8 feet lower at Glasgow than it did before any works were begun, which not merely adds to the tidal capacity of the river, but also prevents the fresh-water floods which formerly inundated the low-lying portions of Glasgow. The improved depth has caused the average tonnage of the vessels frequenting the port of Glasgow to rise from 199 tous in 1863 to 315 tons in 1883 ; whilst the total tonnage entered and cleared has increased from 1,757,000 tons in 1863 to 5,544,000 tons in 1883.
River Tees Improvement Works.
The Tees was formerly a very irregular winding river between Stockton and Middlesborough, and after passing that town it opens out into a wide sandy. estuary about 6 miles long and. 3 miles across at its widest part. It is tidal for about 17 miles, and the rise of spring tides at its mouth is 15 feet. The improvement of the river between Stockton and the estuary was commenced in 1810 by making a straight cut near Stockton ; another cut was made in 1830, and the river was also regulated by cross jetties. These works provided a more direct channel, and increased the depth by about from 2 to 5 feet. The channel, however, between the jetties was irregular in depth ; and the navigable channel through the estuary was shallow and variable, with a rocky shoal across it, and a bar at the mouth of the river. Accordingly, in 1853 training walls were commenced, both for connecting the ends of the cross jetties, and also for guiding the channel through the estuary. Dredging was commenced in 1854 for removing shoals and deepen-ing the channel ; and the ridge of rock across the estuary channel has been removed by blasting. The total amount dredged up to Oct. 31, 1884, reached 13,145,000 cubic yards, and a maximum of 1,220,000 cubic yards was dredged in 1883-84. Vessels of 3000 tons, drawing 21 feet of water, can leave Middlesborough fully laden ; the average tonnage of vessels frequenting tbe Tees has risen from 169 tons in 1873 to 303 tons in 1883; and the total ton-nage entered and cleared has increased from 1,212,000 tons in 1873 to 2,528,000 tons iu 1883. It is profosed at present to deepen the channel by d.redging, so as to obtain a depth at low water of 12 feet up to Middlesborough, and 10 feet from thence up to Stock-ton ; and it is hoped that eventually 2 feet additional depth may be attained by the same means. Two breakwaters have been designed, starting from opposite sides of the estuary and con-verging over the bar, in order to protect the entrance, to facilitate dredging, and to keep out drifting sand. The southern break-water has been completed, and the northern breakwater is in pro-gress. The breakwaters and the training walls are constructed of slag obtained free of charge from the neighbouring iron-works.
Training Walls on the Tidal Seine.
The Seine has a very winding course between Rouen and the sea, as well as in its upper portion, but it possesses a good natuml depth between Rouen and La Mailleraye, a distance of about 37 miles. Below this point, however, the natural condition of the river was very unsatisfactory, for the channel through the estuary was constantly shifting, and high shoals existed Aizier and Villequier with a depth over them of only 10 feet at spring tides, so that vessels of from 100 to 200 tons found the passage difficult, and even dangerous at times. Training walls, formed of mounds of chalk obtained from the neighbouring clilfs, were commenced in 1848, and were gradually extended along both sides of the channel, as shown in fig. 10, and were terminated at Berville in 1869, a further extension of the northern wall for lf miles having been refused in 1870 for fear of endangering Havre. Those works have effected a remarkable increase in depth in the channel between the walls, as shown by a comparison of the longi-tudinal sections of the river in 1824 and 1875 (fig. 11); for dredging has only been used for deepening the worst shoals. The improvement, however, ceases beyond the termination of the works ; and the channel between Berville and. the sea is still changeable and shallow. High walls were for the most part adopted down to Tancarville on the right bank and to La Roque on the left bank. These walls, however, produced such rapid accretion behind them that low walls, raised only from 3 to 5 feet above low water of spring tides, were adopted below these points. This precaution, however, has not arrested the accretion in the estuary, which is still proceeding, though sixteen years have elapsed since the works were stopped. The accretion resulting from the training works had reduced the tidal capacity of the estuary by 274,000,000 cubic yards in 1875 ; and a survey in 1880 showed that a still further loss of over 40,000,000 cubic yards had occurred between 1875 and 1880. More than 28,000 acres of land have been reclaimed in the upper part of the estuary, as shown by cross lines on the plan ; whilst large tracts have been raised to high-water level, as indicated by dotted lines, etending as much as 8i miles below the ends of the training walls. The walls were originally merely rubble chalk mounds ; but the bore and the currents injured the mounds, so that the walls are being strength-ened by pitching or concrete on the river slope, with an apron of concrete, and piling at the toe (fig. 12). The regulation of the river, by bringing deep water about 25 miles nearer the sea, enables vessels of about 2000 tons, and drawing about 20 feet, to pass the shallow estuary between the sea and Berville at high tide and thus reach Rouen. The prolongation of the training walls has been frequently urged ; but the fear of injuring Havre, and the difficulty of devising a suitable channel which would • effectually serve both Honfleur and Havre on opposite sides of t.he estuary, have hitherto prevented the continuation of the work. The Tancarville Canal is in progress for connecting Havre with the Seine at Tancarville (fig. 10), so as to enable river craft to avoid the clangers of the lower estuary ; and the canal is being made so as to be capable of being readily converted into a ship-canal if the growing accretions should impede access between Berville and the sea.
TVorks at the Month of the Macts.
The Seheur branch of the Maas, which forms the most direct channel to Rotterdam, gradually silted up at its outlet, so that vessels had to seek more southern and circuitous channels. The length of the deepest channel was shortened in 1829 by the construction of the Voorne Canal, but even this course became inadequate for the increasing draught of vessels. Accordingly, iu 1862 works were commenced for providing a new direct outlet for the Scheur branch of the river by a straight cut across the Hook of Holland, with fascine-work jetties for training and maintaining the channel across the sandy beach into deep water (fig. 13)'. The cut, three miles long, was only partially excavated, its conTletion being effected by &mining up the old channel and directing the fresh-water discharge and tidal current through the new cut. The scour soon deepened the narrow cut ; but some of the sand washed from the cut settled in the wider channel formed by the jetties. The cut also, being scoured deeper than the adjacent channel above and below, was not adequately widened, and consequently impedes the entry of the flood tide up the river (figs. 13 and 14). The desired depth of 23 feet at. high water not having been attained as anticipated from the works, dredging has been resorted to for deepeninfrrthe outlet ; and the widenins out of the cut to the proper full width would improve the tidal influx. The jetties consist of fascine mattresses secured by piles and stakes and weighted with stone ( fig. 15) ; and their outer portions have been kept down to half-tide level to promote the freer admission of the flood tide, whilst serving equally to concentrate the latter part of the ebb.
The largest jetty works in the world for lowering a bar in front of a tidal estuary are being constructed at the entrance to Charles-ton harbour (fig. 17). The jetties are being forined after the type of the Maas jetties, with log and fascine mattresses weighted with rubble stone (fig. 18). They start from the shore about 2f miles apart and converge to a width of about 2000 feet over the bar, which stretches across the entrance to the harbour at miles from the shore. The object of the works is to concentrate the tidal and fresh-water current from the land-locked estuary, having an area. of 15 square miles, into which the Ashley and Cooper rivers flow, and whose mouth forms the entrance to Charleston harbour. The northern jetty, which was conamenced in 1878, attained the present length of 14,860 feet in 1881 ; and the southern jetty had been carried out 14,130 feet towards the end of 1883. The outer portions, lying in the direction of the flood cur-rent, are to be raised ; but the inner portions are to be kept low, in order to interfere as little as possible with the littoral drift and thus avoid an advance of the foreshore, and also in order to admit freely the flood tide. Though scour has already commenced, it will be necessary to aid it by dredging in order to attain a depth of 21 feet in place of the former depth of feet.
Improvement of Tideless Rivers.
Tideless rivers on entering the sea have their velocity checked, and consequently deposit the silt which they previously carried in suspension. In process of time this accumulated deposit forms a tract of low-lying land protruding into the sea, through which the river flows in several shallow channels to the sea owing to the impedi-ments offered to its flow by the sediment which it deposits. The form which these diverging channels assume has led to the term delta, being applied to the mouths of tideless rivers and the tract of land which they create (figs. 19 and 22). These deltas are always advancing, and con-sequently reducing the very small fall of the channels through them by prolonging their course. The rate of advance varies with the amount of sediment brought down, the depth of the sea in front, and the extent to which the delta sprea,ds out. The Rhone delta has at present a yearly average progression of 140 feet ; the Kilia mouths of the Danube delta have been estimated to advance 200 feet annually ; whilst the Mississippi delta, extending 220 miles into the Gulf of Mexico, is supposed to have taken four thousand four hundred years in forming, which would be equivalent to an average annual advance of 264 feet, the present advance being about 207 feet in a year.
The only method of improving the outlet of a tideless river is to concentrate the current flowing through one of the channels, and to prolong the banks of the regulated channel into deep water by means of parallel jetties.
The only other way of remedying the impediments to navigation at the mouths of tideless rivers is by avoiding the delta channels altogether, and constructing a canal connecting the deep river above the delta with the sea at some suitable place beyond the influence of the river alluvium. This expedient has been resorted to for the trade of the Rhone ; for, though the discharge of the river was concentrated into a single outlet by forming embank-ments on each side, between 1852 and 1857, which shut off the other three outlets and extended into the sea at its mouth, the increased discharge brought down the whole sedhnent of the river, and a bar formed again farther out. Accordingly the St Louis Canal was formed, between 1863 and 1873, going from the head of the delta into the Bay of Foz beyond the limits of the delta. A similar plan was proposed for the Mississippi, but was abandoned in favour of the jetty system. The canal constructed by the emperor Claudius for connecting the Tiber with the harbour of Ostia proved a failure, as its outlet at Ostia was not sufficiently removed from the delta so that it gradually silted up and is now 2 miles from the coast.
The success of the jetty system depends upon the exist-ence of a littoral current to carry away the sediment in the stream conveyed by aid of the jetties into deep water, or upon the gradual prolongation of the jetties to keep pace with the progression of the delta. The outlets of the Danube and the Mississippi have both been improved by training one of their minor delta channels into deep water ; and hitherto the depth over the bars has been maintained. In the case of the Danube, the southerly current sweeping across the outlet carries away a portion of the issuing sediment. The Mississippi jetties direct the discharge into such deep water, and with so much velocity, that it may be premature to decide to what extent the maintenance of the depth may be due to a westerly current in the gulf ; but hitherto from one cause or the other, or probably from both combined, the bar has not formed again in front of the jetties.
Sulina Piers of the Danube Della.
The delta of the Danube commences about 45 miles from the Black Sea, and has an area of 1000 square miles. The river divides into three main branches; the northern or Kilia branch convoys more than three-fifths of the discharge, but it forms an independent delta near its outlet and is consequently unsuitable for improvement (fig. 19). The southern or St George branch is the next largest, and possesses the best channel, but it divides near its outlet into two channels, which are both barred. The central or Sulina branch, though narrower and less good than the St George branch, and conveying less than one-thirteenth of the total discharge, possessed the only navigable outlet in 1858, and was therefore selected for the provisional iniprovernent works begun in that year. The works designed by Sir Charles Hartley consist of piers, starting on each side of the Sulina outlet, whic-h converge till the width between them is 600 feet, and are then carried parallel across the bar (fig. 20). The piers were at first constructed of rubble mounds with piles carrying a platform strengthened at intervals by timber cribs ; but subsequently they were consolidated with concrete blocks (fig. 21). These piers serve to concentrate the discharge across the bar, and increased the least depth at the outlet from 9 feet in 1857 to 20 feet in 1872, a year after the final completion of the works ; and this depth has been since maintained. The sediment-bearing current, moreover, is carried within the influence of the southern littoral current, which, diverting a portion of the deposit, has reduced the rate of advance of the &dins. delta from 94 feet to 44 feet in a year. This progression, however, will event-ually necessitate the extension of the piers.
The Mississippi Jetties.
Though the Gulf of Mexico is in cominunication with the Atlantic Ocean, it is almost tideless for the average rise of tide in front of the Mississippi delta is onfy 14 inches, and there is only one tide in a day. The Mississippi, accordingly, is a tideless river, and forms a delta which has an area of 12,300 square miles, and has three main channels, or passes, leading the discharge of the river into the gulf (fig. 22). The general features of the delta hare been already described, and the various schemes for improving the outlet referred to (see Mississim). In flood time the river brings down in suspension 2800 cubic feet of solid matter per second ; and before the jetties were commenced the annual advance of the bars at the mouths of the passes was 300 feet at the South-West Pass, 260 feet at Pass h l'Outre, and 110 feet at the South Pass.
feet. The South Pass, however, was selected by Congress for improvement, partly on account of the smaller cost of the works required, and partly because it was the pass which had been pre-ferred by a previous commission. The depth of the South Pass through the delta, a distance of about 13 miles, was 30 feet; so that the obstructions to be removed were restricted to the head and mouth of tbe pass. Dredging had been tried for deepening the outlet of the S.outh-West Pass, and had necessarily failed in pro-ducing any permanent improvement, for sedintent soon filled up again the portion of the channel which had been enlarged beyond its natural limits of maintenance. The object aimed at in the South Pass was. to contract the width of the channel at the head and outlet, so that the current might be forced to regain its required section of channel by scorning out iu depth what it lost in width. At the mouth of the South Pass this result could be effected by concentrating the current over the bar with parallel jetties, thus prolonging the banks of the pass artificially into deep water, and contracting the current sufficiently to ensure the requisite depth. At the head, however, of the South Pass, the conditions were more complicated • for, as soon as the entrance channel there was contracted, a porttion of the discharge tended to desert the impeded pass for the other more open passes, and the head works by themselves would therefore have deprived the South Pass of a portion of its natural flow, which would have necessarily led to a reduction of the channel below the head. Accordingly the entrances to the other passes had to be correspondingly reduced, so as not to absorb more than their former proportion of the dis-charge and thus leave the discharge through the South Pass un-diminished.
The works consist mainly of willow mattresses, which are specially suitable where osiers are abundant, and where settlement on the soft alluvial bottom is inevitable. The funnel-shaped entrance at the head of the South Pass was contracted across the shoal into a uniform channel, 800 feet wide, by means of mattress dykes ; and Mattress sills, 30 to 40 feet wide, 60 to 70 feet long, and 2 feet thick, were laid right across the entrances to the other two passes to restrict the volume of their discharge to its normal amount.
The parallel jetties on each side of the mouth of the South Pass consist of tiers of willow mattresses, 100 feet long, from 20 to 50 feet wide, and about 2 feet thick, consolidated with rubble stone, and capped at the outer ends with concrete blocks to secure them against waves (fig. 23). They have been raised to flood-tide level to within 1000 feet of their extremities. The east jetty is 2i miles long, and the west jetty 1 miles ; they terminate at the same distance out, in a depth of thirty feet. They are placed about 1000 feet apart, and are curved slightly towards their extremities, so as to bring the channel at right angles to the westerly littoral current in the gulf. The jetty channel has been contracted to a width of 700 feet by mattress spurs, in order to promote the scour in the central channel, which was lessened at first by leakage through the mattresses owing to their want of consolidation from difficulties in the supply of stone.
The works were commenced in 1875, and completed in 1879. The channel was to be made according to the terms of the con-tract, 26 feet deep and 200 feet wide, with a central depth of 30 feet. According to a survey made in May 1884, the least central depth through the jetty channel is 33 feet, and the width of the 26-foot channel is 290 feet ; and beyond the jetties the least central depth is 31.8 feet, and the least width of the 30-foot channel is 70 feet. At the head of the South Pass there now exists a straight channel with a minimum central depth of 35 feet ; and the 30- foot channel has a minimum width of 275 feet. There is 110W a depth of 30 feet right through the South Pass.
The latest surveys indicate only a slight shoaling beyond the outlet, showing that hitherto the accelerated current, being pro-truded into deep water and aided by the littoral current, has not created. a fresh bar by the deposit of its sediment. As the littoral current cannot be expected to convey away raore than a portion of the sediment brought clown, the material must be accumulating beyond the outlet, and after having filled up the deep places in front, will gradually rise beyond the ends of the jetties, where the dispersed river current will be unable to carry it away. This shoaling, however, may be delayed for a considerable time by the distance to which the velocity of the trained current carries the sediment, and the depth in which the sedintent is deposited ; and when it becomes prejudicial to navigation, it can be readily removed by an extension of the jetties.
.- The method of lowering a bar at the mouths of rivers by means of jetties has been applied, as indicated above, to both tidal and tideless rivers ; but the systems employed for each type of river are based. on different principles which must not be confounded together. A tideless river is maintained solely by its discharge, and therefore the more its channel is contracted the greater is the depth attained. As, however, a tidal river depends largely for its maintenance on its tidal flow, a contraction at its mouth checks the entrance of the flood tide and reduces the tidal flow. The contracted width between the Adour jettiee (fig. 16) has reduced the tidal rise at Bayonne in spite of the openings formed in them for the admission of the tide ; and the narrow width of the cut at the Maas outlet (figs. 13 and 14), whilst affording an improved depth in the cut,%is prejudicial to the depth elsewhere. A parallel channel with high jetties is suitable for tideless river mouths ; but a slightly diverging channel, with the outer ends of the jetties below high-water level, is expe-dient for tidal rivers. Converging jetties, like those of Dublin and Charleston (fig. 17), would be perfectly use-less at tideless outlets ; but these jetties, by not being unduly contracted at their extremities, and by being kept low towards their ends, freely admit the tidal flow, whilst the increased capacity obtained by their enlarged form increases the tidal scour at the outlet. The comparatively small improvement in depth over Charleston bar, as com-pared with the magnitude of the works, may be due to the want 'of concentration of tidal scour, owing to the small elevation of the inner portions of the jetties, which allows of the dispersion of part of the ebb.
The greatest difficulty in training tidal rivers is so to adjust the width of channel that the free admission of the flood tide may be secured whilst affording adequate scouring power for the current. If the influx of the tide is checked by a sufficient reduction of width to ensure improvement in depth by scour, the capacity of the estuary is eventually reduced, and a portion of the scouring power is lost, as in the case of the Seine. It is better, therefore, to regulate the width so as to ensure a free admission of the tide, and to provide for any deficiency in scour and depth by dredging. Deepening by dredging can be easily and economically effected to any desired extent, as shown by the Tyne and Clyde improvements ; whereas tidal capacity in an estuary, when once lost, can never be regained.
For further inforniation about the works described, reference may be made to Rivers mul Canals, by L. F. Vernon-Harcourt ; RiVCT Tyne Improvements, by P. J. Messent ; " The River Clyde," by James Deas, Proc. Inst. C. E., vol. xxxvi.; " The Delta of the Danube," by Sir Charles Hartley, Proe. Inst. C. E., vols. xxi. and xxxvi.; A History of the Jetties at the Mouth of the Mississippi River, by E. L. Corthell. (L. F. V.-H.)