/' Jf**-^-**- -^v X '/ DAWSON /l^^,, 0.V .J -4 THE EASTERN CANAL OF FRANCE. X) %^ 8 z THE EASTERN CANAL OF FRANCE; Fon KSTABLTSIIING A LINE OF WATER-COMMUNIOATION FROM THE SAONE TO TOE MEUSE. DY WILLIAM liELL DAWSON, li.A., B.A.So., Stud. Inst. C.E., AnCIEN t^r.fcvE EXTERNE BE I-'I^IJOI.E DES I'ONTS ET ( 'lIAT'SSKKH By novmisaion of the Council. Excorpt Mimitofl of ProcoodingR of Tlio Institution of Civil Enj^inoors, Vol. Iv. Session 1878-7U.— Part i. . '.'"J.it'-vl Kv JAMr.H j'.riuii':H'r,'.A.Sc., Stud. hist. C.E., Aucicn Elovo Externe de I'^ficolc des Ponts ct Chaussees. Befouk the war of 1870, two important lines of water-com- munication united with each otlier at Strassburg; one, passing between tho Vosges and tho Jura, connected tho northern tribu- taries of tho Rhone with tho Ikhiuo ; tho other, leaving tho Seine at Paris, took a direction almost due east, and mot tho Rhino at tho same town. Tho loss of Alsace and Lorraine rendered it necessary to unite these two lines by a now canal, running north and south on tho west side of the Vosges, in order to prcservo their connection upon French territory. It was also decided tliat this new lino should bo prolonged to tho coal-fields of Belgium, A navigable route in this direction had long been asked f(»r by tho Eastern Departments, and a scheme for its execution had boon drawn up before the war, which had to bo abandoned on account of tho new position of tho German frontier. Tho request itself became more urgent, as many of tho largo manufacturers of Mulhouse and elsewhere, who had re-established themselves upon French soil, were in great need of tho cheap mode of conveyanco which water-communication alone can afford, both for the fuel they required, and for tho heavy articles of their manufacture. Tho valley of the Mouse forms a natural channel, lying precisuly in tho direction wished for ; so that the improvement of its navigation, from tlie point at which tho canal from Paris to Strassburg crosses it, furnishes an obvious solution to this part of tho question. As several towns of importance are situated upon that river, thoir interests would also bo met by its improvement. The navigation of the Rhone is prolonged to tho north alo*^.g the Saone, which has been canalized to within a short distance of its ' Tliis coniinuuicr\tiou wuh read and diseiiBHod lit a meeting of tho Students on tho 29th of Maroh, 1878, and was awardwl tho Millor Bcholarsliip in the Soiisiou 1877-78.— Sue, Iimt. V.K B 2 55997 4 OTHER SELECTED PAPERS. source, and forme, therefore, the nearest and most accessible of the tributaries of the Rhone to the canal from Paris to Strassburg. To reach it from the canal, either the Meuse or the Moselle might be chosen, as both these rivers are crossed by the canal. The Meuse, howcvir, soon dLninishcs to a (small stream above this point ; and the tributaries of the Saono nearest to its source are but ra})id brooks for a oonsiderablo distance, while the water-shed separating them is formed of hills difficult to cross. The Moselle is much larger, and, after some distance, runs parallel to the water-shod separating it from the basin of the Suone a short way from its left bank. A choice is thus affoided, as the Moselle is still a large stream, and the tributaries of the Saone aro numerous. The one selected was a small stream called the (^oney, whose source is within 5 miles of the important town of Epinal, on the Moselle. The connection of the two rivers can be effected without tunnelling — which is certainly remarkable,, considering that they are separated by one of the primary water-sheds of Europe ; and it shows the natural advantages of the point chosen. The scheme under consideration consists, then, in the construc- tion of a line of watei'-communication upwards of 280 miles in length, having a general north and south direction, and running through the five Departments of the Haute Saone, Yosges, Meurthe- et-Moselle, Meuse, and Ardennes. It is comprised undei' the name of the Eastern Canal {Canal de VEst) (Plate 7), and it affords a navigable route to Belgium and the North Sea on the oue hand, and to the Ehone and the Mediterranean on the other. The depth of water in the Marne-Rhine ( 'anal, as the canal from Paris to Strassburg is called, is 5 feet 3 inches. Tlie locks aro 17 feet wide, and 125 feet long. 'l"he Saone is already canalized as far as Povt-sur -Saone, up to which the draught of water and width of the locks is the same as in the Marne-Rhine Canal. Their length, however, is only 114 feet. On the Eastern Canal also the locks are to be 17 feet wide, but the depth of water will be G feet 6 inches. Their length is 131 feet 3 inches between the mitre-sills (Plate 8, Figs. 1 and 2). The lift-wall is double, the floor of the head-bay being at a level inter- mediate between the bottom of the reach above and the floor of the lock. The object of this arrangement is that the sluices iu tho head-gates may always be under water, oven when the tail-gates are open, so that tho boats may not be inconvenienced by the water potiring down upon their decks while tho lock is being filled. This increases somewhat the amount of ashlar masonry, as well as the height of the head-gates (Plate 8, Figs. 4, 5, and G) ; but there THE EA8TEBN CANAL OF FRANCE. 5 is economy on tho whole, as ordinary gates with sluices are thus given all tho advantage, without the expense, of side culverts. In accordance with Acts passed on the 15th of April and tho 7th of August, 1871, tho Minister of Public Works directed tho engineers in his department to commence preliminary surveys, and to draw up reports upon the proposed scheme. The estimated cost of the works was £2,600,000 ; and as the Government was unable, so soon after the war, to furnish this sum, it was arranged that the amount should be made up by the five Departments interested, and that they should be repaid in annual instalments. The whole work of surveying, location, and construction was undertaken by tho Government, and the canal is eventually to become Govern- ment property. Tho date fixed for the completion of the work is May 1882. Tho scheme consists naturally of two parts ; the first to be considered is the plan for connecting the Marne-Iiibine Canal with the Saone, as it has been proposed and partly carried out. The Canal fuom the Moselle to the Saone. Tb'J Marne-Ehine Canal is already connected with tho Moselle near Toul by a short branch, which serves at the same time to supply tho canal with water. This forms a starting-point for the new canal, which then ascends the Moselle as far as Epinal. The Moselle here flows in a narrow valley, varying in general from a ^ to ^ mile in width, and enclosed between hills from 200 feet to 300 feet in height. The current is rapid, as the fall averages 6 feet in the mile, and the volume of water passing Toul is 118,000 gallons a minute. Several mill-dams and weirs, built partly across the river, occur in this part of its course, although the greater portion of the water-power derived from tho river is in the vicinity of Epinal, and above that town. Nothing had previously been done to render tho river navigable, and in the Government reports it is merely clasiaed as Jlottahle, i.e. suitable for tho conveyance of firewood in logs. It was found advisable to abandon tho channel in many places ; and the plans, as finally approved, t^how by much the greater part of the work constructed as a lateral canal. On reaching Mcssein, a branch cros-es to the valley of tho Mcurthe, and unites with the Marne-ltliiue Canal a little above Nancy. This is called the IS'ancy brmn h, and is of great import- ance to that town, as it makes tho communication with the upper part of tho Moselle much more direct. Its length is 7;^ miles, 'i'he summit reach on this branch is 1-11 I'cct above the level of tho H OTHER SELECTED PAPERS. Mourtho at Nancy, while it is only 69 feet above the Moselle at Messoin. This latter river is, therefore, more suitable as a source of supply, and was resorted to, as there are no lakes or streams of importance in the neighbourhood of the summit. The water is derived from the Moselle, a short distance above F] avigny, where a mill-dam raises the level of the watei- in the river. A now dam will bo constructed to increase the height of the water ; from here to Messein an aqueduct, consisting of an open trench 16 feet 5 inches wide at the bottom, and with slopes at li^ to 1, is being excavated. It follows the right bank of the Moselle, and passes under the first arch of the bridge on which the canal crosses the river at Flavigny, below which there is a branch leading to the factory supplied by the former mill-dam. It then runs between the canal and the river, and enters one of the reaches, which is brought purposely to the same level. This part of the aqueduct is navigable for small boats, and serves to join the canal with tho Moselle, thus enabling wood cut along the banks to be brought into the canal. A guard-lock on the aqueduct near the point where it meets the canal, with a pair of gates opening in each direction, severs the connection with the river when the canal is emptied for repairs. From the lower end of this reach, the aqueduct is cimtinued along tho side of the hill above tho canal; crossing the Nancy branch on the level, above the first lock, it arrives at Messein, where a portion of tho outfall is used to work hydraulic engines, which raise the remainder of tho water to the summit reach. The bridge at Flavigny, for carrying the canal across the Moselle, is of masonry. The level of the canal is kept as low as possible, to avoid heavy embankment, as far as the bridge, where it ascends 16j^ feet by means of a lock, continuous with the masonry of tho bridge. The hills on the other bank are close to the river, so that the reach, of which tho bridge forms part, can be extended along them in cutting without difficulty. The foundations of this bridge are rather peculiar. The danger of scour was considered so great, on account of the rapidity of the current in time of flood, and the gravelly nature of the bottom, th.'it a flooring of con- crete, level with tho bottom, was laid completely across the river. Its width is about 25 feet greater than the length of the piers, and the concrete flooring is so placed as to extend 15 feet be- yond them down stream, and 10 feet on the up-stream side. It was built in two parts, in coffer-dams extending from the banks. The depth of the excavation was 10 feet below the summer level of the water. The piers, 0^ feet thick, stand upon this slab of TH£ EASTKfiN CAMAL OF FRANCE. 7 concrete. The rise of the arches is one-tenth of the span. When they are completed and the centres struck, the extrados will be coated with a layer of cement to prevent leakage, before the channel of the canal is commenced, and it also will be lined with cement. This mode of foundation has already been successfully employed several times in France. As far back as 1750, M. de Regemortes built a bridge in this way across the Allior, whose rapid current and sandy bed had previously caused the destruc- tion of several bridges. The edges of the concrete were protected above and below by several lines of heavy sheet-piling. The same method was again adopted in the construction of a bridge carrying a canal across the same river; and also for a railway- bridge crossing a little higher up. The thickness of the concrete in these cases varies from 5 to 6 feet.^ Above Flavigny the canal follows the left bank of the Moselle till it arrives at a point 2 miles below Epinal. Very little has yet been done on this part of the line, and no works occur of special interest. At this point the main line of the canal turns abruptly from the river, its direction being prolonged as far as Epinal by a short branch, which connects this town with the canal, and derives a supply of water from the river. A portion df the canal now follows, in which the question of the supply of water and of economy in its use becomes of primary importance. The hills lying between the Moselle and the sources of the Coney form a spur of the Vosges, and consist for the most part of soft red sandstones, known as the variegated sandstone (Gres higarre), on account of their streaked and spotted appearance, 'i hese form the lower member of the Trias. On the slopes towards Epinal, beds of a considerably harder grey sandstone occur. This is the Vosgian sandstone (Grcs cles Vosges) of the upper Permian. They are together a conformable series, dipping at a low angle toward the north. The summit reach is 1,1 82j feet above sea-level,'* and is a little over 7 miles in length. There is a considerable amount of rock cutting upon it, and at the end furthest from the Moselle, where the water-shed is actually crossed, the depth of cutting is 48 feet measured to the water-level of the canal. The level of the water in the reach common to the Epinal branch and to cho main line of the canal is 1,036^ feet above sea-level, leaving a height ' For drnwiii^'s of these bridgoa see ' Morandierc, Traile de la Constructiou des I'oiitset ViiidiicH,' riato 1]. ' Tliu luvcl of tiiu Mcditurruncau it; taken ub datum. 8 OTHER SELECTED PAPERS. of 146 foet to bo ascendod in a distance of less than 2 miles. To reduce as far as possible the amount of water required to take a boat up this ascent, it was decided to make the locks separate and not in a flight ; and further, that they should all be of equal lift, and that the lift should not be groat. The lift chosen was 9 feet 10 inches, which gives fifteen as the required number of locks. Less water would bo consumed if the amount of lift were lessened, but this would bo difficult to effect, as it would necessitate an increase in the number of the locks. Most of the reaches arc short, and it was considered essential that the locks should be situated upon tangents at least 160 yards in length, so that boats might not be inconvenienced by coming upon the commencement of a curve before they were completely out of the lock. The curves are rather sharp as it is, and any further increase in the number of the locks would render it difficult for the boats to pass them, as the valley by which the ascent ia made is a narrow one. The reaches between the locks are so short, and their surfaces, if the usual width of 52^ feet at water-level w«re given to them, would Ix) so small, that when the quantity of water necesf'ary to fill a lock was drawn from them they would be no longer n'.vrigable. To obviate the difficulty no bank was constructed on the side opposite the tow-path, but the water of the canal was allowed to extend across the valley, till it met the sloping groiind forming its opposite bank. This part of the canal consists, therefore, of a series of ponds, one above the other, occupying the bottom of tho valley. 'I'hey are separated from each other by dams of earth across the valley, terminating at one end against the outer side of the lock walls, and extending to the slope on the right side of tho valley. They are placed at the lower end of the lock, in order that the area of the ponds may be as great as possible. This plan, for which the natural shape of the valley was particularly well adapted, gave to the reaches surfaces which vary from 9,500 square yards to 24,000 square yards. The smallest of the ponds has, therefore, an area ei^ual to a length of nearly 550 yards of canal of the usual width ; and when water is drawn off on a boat passing through the lock below, its level will be lowered only 3 inches. The draught of water in tho channel adjoining the tow-path is here increased from 6 foet 6 inches to 7 feet 4 inches, so that tho water may be allowed to fall 10 inches without interrupting tho circulation. This corresponds, in the smallest pond, to the quantity of water necessary to fill a look three times. The level of tho water can also be raised 12 inches, if water be required for the supply of the pond below. Several of the locks hive been placed in the position THE EASTERN CANAL OP FRANCE. 9 they occupy in order that the dam extending from them might serve to re-establish the communication along roads crossing the valley. The position of the remaining locks has been determined by the slope of the ground, combined with the condition of heir being on a tangent of given length, as already mentioned. The dams retaining the water in the ponds are 9 feet 10 inches wide on the top, and extend 3 feet 3 inches above the level of the water, a height which allows for the level being raised occasionally, and for the action of waves, as sudden gusts of wind are frequent in this hilly region. The thickness increases to 21 feet 7 inches at the level of the water, and the slopes continue below this level at 1.^ to 1. A small waste-weir in eacli dam, at the end furthest from the lock, allows any surplus water to pass into the pond below. Tho work on this part of the canal is now actively in progress, and groat care is being taken to render the dams water-tight. Much of the earth used in their construction is brought from the cuttings on the summit icach. It is conveyed in wagons running upon rails, and when tip[ted, is spread with shovels in thin layers and well rammed, as required by the specification in all such cases. The embankments which form the sides of the canal towards tho lower ends of tho reaches are treated in the same way. In calculating the amount of water required for the summit reach, the annual tonnage passing through the canal is taken at 1,000,000 tons, of which one-half is assumed to be local traffic for tho conveyance of stone and tire-wood from the summit to the navigable rivers on tho one side or the other, and the rest through traffic. This tonnage is nearly twice the present annual amount on the Marno-Ehine Canal, and must be considered as allowing an ample margin for the future development of commerce. The number of days in tho year suitable for navigation is esti- mated at three liundred, as there is always a considerable period towards the spring when floods render the rivers impassable, and in summer their level may be too low. This gives 1,700 tons a da}' for each kind of traffic. With boats carrying 100 tons, there would therefore be 34 -j- 17 = 51 lockfuls of water drawn olF each day from tlie summit roach. The loss by percolation and evapora- tion is estimated at 100 gallons a day per lineal yard. On multi- pl}'iiig this by tho length of tho canal on tho two 8lo])es, and adding tho amount of water required for lockage, the total daily consumption is found to be 9,418,000 gallons — equivalent to 0,000 gallons per minute (half a cubic metre per second). Tho original plan, suggested in 1872, for obtaining the necessary amount of water was to construct an a(iueduct which would dorivo 10 OTHEU SELECTED PAPERS. its supply from the upper waters of the Moselle and from its tribu- tary the Vologne, and following the hillside forming the left bank of the Moselle would fall into the summit reach at the end nearest Epinal. The level of these rivers where the supply would be obtained is such as to aflford a sufficient fall for the aqueduct, and a volume* of water double that required could be taken from them during six months of the year without infringing upon the rights of the mill-owners in the valley. For the remaining six months an additional supply could be obtained from the lakes of Gerardmer and Longemer near the sources of the Vologne, by storing up the winter rainfall for use in summer. Water would be let out uuiing the dry season, and, finding its way down the natural channel of the Vologne, would be conveyed by the aqueduct to the canal. Much of the water in passing down the Vologne would bo drawn off for the irrigation of meadows, and in this way considerable loss might be expected to occur. The essential defect of the project, however, is found in the circumstance that the supply of water is stored at so great a distance from the canal. The summit reach contains 132,000,000 gallons, and if emptied for repairs it would take eight or ten days to refill it ; whereas on the Marne- Ehine Canal only three days are needed for re-establishing the navigation. There is always a delay in the arrival of the supply, which, in the case of a press of traffic or of any accident, might prove to be a serious inconvenience. To meet the difficulty it was determined to construct in addition a reservoir near the summit reach. A suitable position was found in the upper part of the valley of the Aviore, a small stream tributary to the Moselle. It was called the Bouzey reservoir, from a little village in the neighbourhood, and derives its supply from the basin, 4,000 acres in extent, in which it was situated. The annual rainfall is 9 '8 inches, and the reservoir was accordingly calculated to hold 907,000,000 gallons. This capacity could easily be given to it by the construction of a retaining wall across the valley ; but as there are important starch works in the lower part of the valley, their claims to remuneration for being thus deprived of a large part of the water xipoii which they depended would have to be satisfied. Soon after the necessity for the construction of the Bouzey reser- voir was admitted, it became apparent that it might be extended HO as to serve for the storage of all the water recpiired during the dry (season. The ^supply would thus be kept in the immediate neighbourhood of the canal, which would be an immense advantage, and do away with tlie necessity of having recourse to the far- distant lakes at the sources of the Vologne. It was found that the THE EASTERN CANAL OP FRANCE. 11 capacity of the reservoir could be nearly doubled, while the expense of doing so luerely increased in the ratio of the volumes of water impounded. The raising of the water-level in the rtjser oir made it necessary for the beginning of the aqueduct to be plpoed further up the Moselle. The supply is therefore derived from the mill-dam of the factories of St. Etienne, near RiremoAt, about 8 miles above the point at which it was originally proposed to draw off the water for the canal. The aqueduct commencing here follows the left bank of the Moselle, parallel to the direction marked out for the former one, but at a higher level, till it falls into the Bouzoy reservoir. The quantity of water for the supply of the canal being 9,418,000 gallons a day, the amount required annually is 9,418,000 X 300 = 2,825,000,000 gallons. Adding the quantity necessary to fill the summit reach after the annual interruption in the navi- gation, the total annual amount readies 2,957,000,000 gallons. It remains, then, to be ascertained whether this amount can be taken from the Moselle at Riremont without injury to existing rights. In summer the St. Etienne factories use the whole volume of water which the Moselle furnishes to their mill-pond ; but when the water overflows the dam, all rights are satisfied, and the surplus is not used. For the supply of the canal one hajf of the surplus is counted upon and appropriated. From careful obser- vations, it was found that the numbar of days in the year during which the overflow of the dam would exceed 52,000 gallons per minute could safely be assumed to be one hundred and sixty. Taking half the amount, the quantity of water which could bo conveniently drawn from the river is 5,7t)0,000,000 gallons per annum. The length of the aqueduct is very nearly 27 miles, and the loss from leakage is assumed to be 60 gallons a day per lineal yard, which is half that/ of the canal, and must bo considered an ample allowance. The annual loss from this cause is therefore 380,000,000 gallons, BO that 5,410,000,000 gallons reach theBouzey reservoir. It will be observed that this quantity i« nearly double the amount calculated to be sufticient for the canal. An ample supply of water can therefore be deiived from the Mosidle without interfering with the rights of the mill-owners, and can be stored in the immediate neighbourhood of the summit reach. An aqueduct giving an outfall of 20,400 gallons per minute is found to bo sutTicient for its conveyance ; and although this aqueduct carries double the amount of water and is one-third longer than the one originally pnijeeteil, and falls into a rof^ervoir of nearly twice the capacity of that first proposed, the cost of carrying out the scheme 12 OTHKR SELECTED PAPERS. is within 3 per cent, of the former estimato. The longest con- tinuous interruption to the supply, as shown by the observations, was in 1868, when it amounted to a period of one hundred and thirty-six days. Taking one hundred and forty days as a maxi- mum, and allowing 0*3 inch per diem as the rate of evaporation, the capacity of the Bouzey reservoir is 1,470,000,000 gallons. The crest of the St. Etienne mill-dam is 1,201 feet above sea- level, and the total fjJl available for the aqueduct is therefuns 43 feet. The whole fall is not utilised ; but the higher level of the aqueduct thus secured facilitates the passing of the villages on the banks of the Moselle. The aqueduct is mostly an open trench cut in side-long ground, and having a longitudinal incli- nation of 1 in 1 0,000. Its dimensions are : — Width at the bottom 5 feet 7 inches. Slope of the sides IJ to 1. Depth C feet 7 inches. Widtli at the top 25 foot 3 iuches. Depth of water 4 feet 11 ixiehes. Sectional nri a of water GIJJ square feet. Velocity of current' 1 • 131 foot jier second. Outfall '2G, 985 gallons per minute. At one point where the hillside forming the left bank of the Moselle stands at a slope of from 30° to 40°, and consists of sand and gravel, a cylindrical brick conduit is laid at the same incli- nation as the rest of tho aqueduct, so that the water occupies only part of its section. It was constructed in a shored trench, whicli was afterwards filled, so as to leave the natural slope of the ground unbroken. Where secondary valleys are met with, they are crossed either by making a detour or by an inverted syphon. The syphons consist of two cast-iron pipes, laid side by side, having an internal diameter of 3 feet 9 inches. The difference of level between their upper and lower ends is reckoned at the rate of 1 foot for every 500 feet of their length, measured along the pipes. At the end near Ej)inal, two tunnels are being constructed for the conveyance of the aqxxeduct in order to avoid long detours, which, in addition, would pass close to the town, where the price of land is high. They are nearly of equal length, and traverse for the most part tho Gres des Vosges. In the estimate it is assumed that they will require a masonry lining throughout their entire length ; but, except at the extremities, this will scarcely bo neces- ' Calculated from tho tables of IJazin given in 'Rochercliea c.xperimcntulca Bur I'tiuoulemoat du I'eau dans los cunuux decouvortcN.' i THE EASTERN CANAL OP FRANCE. 13 sary. In BeCtion the lining is egg-shaped. It is semi-circular licloAV, with a diameter of 4 feet 3 inches, and pointed above, the whole height being 6 feet, with a thickness of 14 inches. The longitudinal inclination to ensure the necessary outfall is 6 in 10,000.1 Where the masonry is not built, it is expected that the increase in the sectional area will more than compensate for the irregularity of the walls, so far as the flow of water is concerned. The tunnels have been commenced from both ends, and as the longer of the two is only 1,5G3 yards in length, it is not intended to sink intermediate shafts. The drifts are similar to those excavated in the hills of the neighbouring Departments in mining for iron ore. They are timhered as far as the work has yet extended, as the rock for some distance from the surface is softened ]iy atmospheric action; and as it is much fissured and jointed, timbering may be required throughout. The rock is blasted with powder, and its consistency is such that the explosion disintegrates it into sand. A small rotating fan, worked by hand, is placed near the entrance to the tunnel, and forces air through a light sheet-iron pipe, about 1 foot in diameter, to the head of the work. By giving this a few turns after each blast the men can return at once to their work. A tramway serves for the conveyance of the excavated material. The quantity of water met with is consider- able, but it is easily disposed of; as even at the upper end, where the inclination of the tunnel is in the direction of the work, side ditches can be cut with a counter-gradient sufficient to carry the water back to the entrance. The rate of advance, based upon measurements made in the mining districts, is taken to be 16 yards per month at each heading. The Houzcy reservoir, into which the aqueduct discharges, is in the valley of the Aviere. It is no longer necessary to prevent the water flowing in this stream from leaving the reservoir, as was intended in the original scheme, for the snjjply is now independent of the rainfall in its basin. It was desirable to place the retaining dam as mar as possible to the canal, in order to increase its area; but as an old quarry exists near the canal, it was thought piudent to put the dam at least 60 yards above it. The valley in question is formed of a streaked reddish sandstone (^Gres higarrc), much jointed; but as ponds were found in the valley, it was inferred ihat the leakage would not bo so great as to make the establishment of a reservoir impossible. It ' Tlio Kriidioiit of fi in 10,000 umountH to a full of loss tlmn 17 im hoa on Imlf tlu' loiif^th of the tiiiinol. 14 OTHER SELECTED PAPERS. was ascertained by borings that the sandstone maintains its fissured and jointed character for a considerable depth, below which the beds become solid and ocntlnnons. The height of the dam required is 68 feet at the centre of the valley As it would be difficult to con- struct a substantial earth dam of such dimensions, and little earth was to be had in the vicinity, it has been decided to build it of masonry. The wall will not be completed to its full height at first, as the capacity of the reservoir is sufficient for the probable traffic of the canal for some years to come ; but the dimensions will enable it to be raised to its final height by simply adding the part left unbuilt. According to the designs prepared for this retaining wall, the width of the top is 13 feet. This width, though less than usual, is considered sufficient 'to resist the action of any waves likely to arise on a surface which has only an extent of 247 acres. Its height above the level of the water is nearly 2 feet, and there is in addition a parapet 3 feet 3 inches high on the side next the water. The inner face is vertical. The curve of the outer face is so calculated that the pressure may nowhere exceed 156 lbs. on the square inch. The sandstone used fur the masonry has a specific gravity of from 2-0 to 2-2. When soaked with water, this is increased to from 6 to 8 per cent., and, by taking these higher values of the specific gravity, the maximum pressure is found to be less, as the curve of pressure approaches the vertical. The stone crushes under a pressure varying from 2 to 4 tons per square inch, and will be built with hydraulic lime from Theil. The foundation of the wall presents much difficulty. On the right side of the valley the ground is the more favourable, as there the rock is solid and impermeable at a depth of from 10 feet to 20 feet below the surface. At the centre of the valley there is a bed of peat 28 feet thick, but below it the rock seems very com- pact. On the left side of the valley the upper beds are completely broken up into loose blocks, separated by wide cracks, and the foundations will there have to be taken down to a considerable depth. It is not necessary, however, that the whole thickness of the wall should rest upon the impermeable beds, as the main body of the masoniy can have as a foundation any of the beds above, whether they be fissured or not. A guard wall, 6S feet thick, will be carried down vertically below the inner face of the wall till it reaches the impermeable beds. At the right side of the valley the masonry will end abruptly against the rock, which is here highly in- clined, and there will bo no difficulty in making thu connection ; but on the other side it will bo necessary to prolong the wall by tunnel- THE EASTERN OANAL OP PRANCE. 15 ling into the hillside to a distance of 180 feet, in order to prevent the water from passing round its end. At this point the surface of the impermeable beds is only 6 feet below the level of the water in the reservoir, and as the rock is covered with a thick layer of com- pact soil, it is not thought likely that there will be any leakage worth considering. The reservoir is to be provided with a waste-weir, and the usual tower contairing the outlet and emptying valves. The canal, on leaving the summit reach, descends the valley of the Coney, PI. 7, till it reaches Port-sur-Saone. The distance is nearly 56 miles, and the locks have the same fall (9 feet 10 inches) as on the other side of the Bummit. The Canalization ok the Meuse. The second part of the scheme under consideration consists in the canalization of the Meuse, from the point at which the Marne- Rhine Canal crosses it to the Belgian frontier, a distance of 172 miles. The Meuse in this part of its course flows through two regions of very dissimilar character. As far as Charlovillo the valley fur about 1 mile in width is occupied by meadows whose surface is nearly level. Beyond these on each side the ground rises rather abruptly to a height of 200 feet or 300 feet. The channel of the Meuse winds through these meadows, and its summer level is only about 3 feet below their surface. The volume of water is not very variable, and yet the least flood inundates the whole valley from side to side. Below Charleville the river enters a mountainous region, and the valley becomes narrow and circuitous. The slope throughout is uniform, varying from 2^ to 3J feet per mile, while above Charleville the winding of the river reduces the fall to an average of 1 foot in the mile. The Meuse above Verdun is still in a natural condition No pro- ject for improving it was ever drawn up until it was decided that the Eastern Canal should follow its valley. Here the nature of the valley makes the system of weirs and ponds, so much employed in France, inapplicable. The construction of a lateral canal has been preferred, as there is every facility for it in so wide a valley. The improvement of the channel of the river would bo more expensive, and it has therefore been avoided, except in passing through the towns of St. Mihiel and Verdun. Below Verdun the width of the valley is not quite so uniform. In several places an appreciable slopo commences immediately upon each side, of the river, and the hills bounding the valley 16 OTHEB SELECTED PAPERS. approach each other. It is at these narrow points that the prin- 1 cipal towns alon^i^ the river are situated. An additional difficulty | is thus presented to its canalization, as the channel of the river j has to bo followed in passing through them. The rebuilding of i biidges, and the construction of retaining-walls, is thus often ; rendered necessary ; while there is little space for the tow-patli ; along the bank. These remarks apply to St. Mihiel and to Verdun i as well as to the towns below. In almost every instance mill-dams, j some of which are very ancient, exist at these towns. This cir- i cuiustance suggests the probable reason for their having grown j up ou the sites which they now occupy, as the intervening parts ! of the river are extremely unfavourable to the construction of ; dams, thus preventing the establishment of mills and manufactories. | The mountainous region of the Ardennes, which the Mouse j enters at Charleville, continues to the frontier and extends into I Belgium. It is supposed to be of Silurian age, and consists of highly- ■■ inclined schists and slates which are extensively quarried. This j part of the river, although circuitous, has always been navigable ; i and at certain seasons large boats can ascend it, while in summer ] the shoals reduce the depth in some places to 18 inches. Mill- : dams no longer occur, as the river is rather large to utilise in \ this way. The country is thinly peopled, there being no villages j except in the valley of the river ; little can be done in tho way of \ farming in the country around. The hills do not produce much j timber, as the trees are usually cut for poles, and these and tan- j bark are the principal vegetable productions. There are important i industries, however; and in addition to the slate quarries, and lime- ! stone quarries in the vicinity of Givet, the manufacture of iron is i cariied on, for which Belgian coal is used. i Although the canalization of the Meuse has been discussed since I the beginning of the century, it was not until 1861 that a pro- I ject was authorised for the construction between Verdun and j Sedan of side-cuts and locks, by which the mill-dams could bo \ passed, and tho mill-ponds, which are 6 feet to 10 feet deep, could I be utilised for navigation. The draught of water was to bo 3 feet l 11 inches. It was intended at the same time to construct a canal | to connect the Meuse and the Moselle by the valley of the Ome, | a small tributnry of the latter river, a little below Metz. VV hen ; the war of 1870 broke out, the side-cuts and locks along the Meuse \ had been completed, but the intervening portions of the channel i of the river had not been deepened. In 1871 when the prosecution i of the works again came uuder notice, it was found that tho latter j part of the project would havo to be abandoned,. as a portion of ] THE KAblTSBN CANAL OP PRANCE. 17 the country which the canal would pass through was lost to France. The now position of the Gornan frontier, together with other changes brought about by the war, nd to the scheme of the Eastern Canal, as has already been explained, and of this project the canalization of the Mouse formed a part. The Marne-Rhine Canal crosses the watershed between the Marne and the Mouse, near the village of Mauvages, after which the summit reach is named. The slope thence is all in the same direction as far as Toul, as the canal is carried across the valley of the Mouse by a reach 11^ miles in length, on which there is a masonry bridge over the Meuse. This is called the Pagny reach from a small village in the vicinity. By this arrangement a summit is avoided between the Meuse and the Moselle; and the question of water-supply is much simplified, as a group of springs found upon the west bank of the Meuse near the village of Vacon became available for the supply of the Pagny reach, and the surplus was led by it to the series of locks descending towards Toul. The most convenient point at which the Eastern Canal can leave the existing one is on this long reach, where, after following the right bank, it turns abruptly fo cross the valley of the Meuse. liience it descends by a rapid succession of locks to the river, about 4 miles off. As the section of the Eastern Canal which connects the Marne-Rhine Canal with the Saone commences at Toul, a portion of the existing canal is common to both, and the Pagny reach becomes a summit reach on the new line of canal. From it is drawn off the water required for the Eastern Canal as far as the Meuse. It has been found necessary to increase the depth of water from 5 feet 3 inches to 6 feet 7 inches on this reach and along the canal as far as Toul, which not only makes the leakage of the canal greater, by augmenting the pressure upon its sides, but also increases the volume of water necessary to iill the locks ; and as the traffic of both canals will pass along this portion, the demand made upon the Pagny reach becomes very great. The supply of water for this reach and for the summit at Mauvo.o;e8 has never been more than barely sufficient, so that the whole question is again opened up, and it has been determined to give it a complete and satisfactory solution. The summit reach upon the Marne-Rhine Canal at Mauvages, and the slope towards Toul as far as the Pagny reach, together with a length of 3 miles on the slope towards the Marne, are supplied exclusively from the head waters of two streams called the Ornain and the Meholle. According to calculations made at the time the canal was opeiud, based upon observations during the [1878-79. N.S.] c 18 OTHER SELECTED PAPERS. dry Boason whioh occurred in June 1854, the volumes whicli these streamB could be counted upon to furnish were : — TheOmain. .... 9,540,000 gallons per twenty-four hours. „ Mehollc .... 660,000 „ „ „ Totnl . . . 10,200.000 which amount is exactly equal to the consumption on this part of the canal.' It was found, subsequently, that the Mehollo is at times completely dried up, and that the volume of the Ornain may diminish to little over half that here supposed. This leaves ii deficit of 6,500,000 gallons a day. Observations show further that the period of deficiency might reach six months, in which case a total volume of 1,012,000,000 gallons would bo required. As it is impossible to construct reservoirs in the neighbourhood of the summit, the only way by which this supply can bo obtained is by raising it by artificial means from some available source below. In all the schemes proposed for the complete solution of the problem, the supply for the summit at Mauvages is derived from the Pagny reach. The difference of level between them is 124^^, feet, and the water is raised by steam engines situated at the village of Void. The volume of 1,012,000,000 gallons must therefore be added to the quantity supplied to the Pagny reach . The amount of water supplied by the springs at Vacon was found to be sufficient at all seasons for the Pagny reach and for the descent towards Toul, although during five months in summer there was no surplus. The additional leakage caused by raising the level of the water is estimated at 2,200,000 gallons a day, which, in five months, amounts to 352,000,000 gallons. On the sec- tion of the new canal from the Pagny reach to the Mouse, the loss by leakage will bo very great, on account of the nature of the soil. It is estimated at 250 gallons per lineal yard a day. This, with the amount required for lockage, gives a total of 484,000,000 gallons for tho period of five months. After the canal reaches the level of the Mouse, use is made of a mill-pond near Commorcy, in which the depth of water is sufficient, and part of the water in tho pond is thus lost to tho mills. As compensation 204,000,000 gallons are allowed to pass down the canal from the Pagny reach during this period. By the addition of these various items, the total quantity of water which has to le furnished to tho Pagny ' Vide an article written 1)y M. Mdlc'zieux in the ' Annalcs des Fonts at Chausscos,' Srd series, vol. xi. , THE EASTERN CANAL OP FRANCE. '19 roach is found to bo 2,112,000,000 gallons; and assuming that tho whole of this amount has to be supplied during the five summer months, tho volume rotjuirod is 14,000,000 gallons a day, or 9,800 gallons a minute. It ^ is been thought more prudent to call this 10,500 gallons. The diflFerence of level between the Mouse and the water in tho Pagny reach is only 18 feet, and by ascending the river a few miles it would not be difficult to construct an aqueduct which would lead its water directly into the canal. Several schemes for effecting this wore, in fact, drawn up before tho construction of the Eastern Canal was proposed. The discharge of the Mouse, however, is known to fall below 12,000,000 gallons per twenty- four hours, as in tho summer of 1870, so that its whole volume is insufficient for present requirements, and to draw from it would entail great expense in indemnities to tho riparian owners of mills. It was then suggested that reservoiis might be established in the valley of tho Mouse or of its tributaries, to store up and render tho flood-water available both for tho mills and the canal during tho summer season ; but suitable sites wore only to be found far up tho valley. The valley, almost from the source of tho river, is cut through the Jurassic formation, in which many of the beds are oxtr^^'^ely permeable, so that it would bo difficult and expensive to make tho reservoirs water-tight. It was also found impracticable to establish an adequate reservoir in the vicinity of the Pagny reach, and as no other choice was left, the Moselle had to be resorted to for the necessary supply. Tho volume is ample, but unfortunately its level is much below that of the Mouse. At Toul it is 130 feet 6 inches below tho water in the Pagny reach. A considerable amount of water-power is available at the locks along the Moselle immediately above Toul, and it has boon decided to utilise this for tho purposes of supply. Three pumping stations will bo established near these locks. It is found from the data in regard to the volume of water in the canal and tho fall at the locks, that 7,300 gallons per minute could be raised to tho necessary level. This volume will bo led by an aqueduct along the left bank of the Moselle and the right bank of tho Ingressin, and will fall into the Pagny reach at the village of Foug, near its eastern end. The buildings in which the pumping engines are to Ix) erected are close to tho canal on tho side next the river just above tho locks, and the water is led to them by arched masonry conduits passing through tho bank of the canal. The current produced in tho canal by tho water being thtis drawn off is less c 2 $0 OTHER 8BLE0TED PAPERS. than J mile an hour. The pipes conveying the water up the slope on the loft bank of the Moselle to tho aqueduct are 4 feet in diameter, and cross the canal either on a bridge over the lock at its lower end, or by being laid under the canal. They ascend the line of steepest slope, except at the second station from which the pipe deviates to the right up the side of the valley in order to fall into the aqueduct below a siphon which occurs in this vicinity, and which would otherwise require to be of larger diameter. The points at which the aqueduct is supplied divide it into three sections. In calculating its dimensions, the velocity of the flow is fixed at 1 foot per second. If it were less, the discharge would be liable to be much decreased by an accidental obstruc- tion ; and a greater velocity would necessitate a greater slope, and an increase of the height to which the water would have to bo pumped in the first instance. The given discharge on each section was increased by one-third, and from it and from the velocity the sectional area was computed in each case. Assuming the side slopes to bo 1^ to 1, the width of tho channel and the corresponding depth of water can be determined, which will give a minimum value to the mean radius. The necessary longitudinal inclination was then calculated from the tables of Darcy and Bazin. It wa s found to be as follows : — Ist section 35 iu 100,000 2nd „ 20 „ 100,000 3rd „ 15 „ 100,000 The aqueduct is carried across the secondary valleys by inverted cast-iron siphons. The diameter of those on the first section is 2 feet, and of the two others which occur on the second section, 2 feet 7 inches. There is a short tunnel on the third section, through which the aqueduct passes under the village of Choloy. This tunnel is a circular brick conduit 6 feet 6 inches in diameter in gravelly clay. The water only occupies a depth of 4^ feet. The aqueduct passes over beds belonging to the Lower and Middle Oolite and to the Oxford clay. On tho Oolite formation, it is in places lined with a layer of concrete to prevent leakage. With this precaution it is thought possible to reduce tho loss to an average of 20 gallons per lineal yard a day. Tho aqueduct is therefore capable of conveying 6,900 gallons per minute to tho Pagny reach. It has been stated that the total amount of water required is 10,600 galhms a minute, and that after employing the whole THB EASTEIiN CANAL OF FItAKOE. SI available water-power of the Moselle there still remain 3,G00 gallons a minute to be furnished in some other way. The diffi- culties of obtaining a supply from the Meuso ha^e already been explained, and although in general its basin is extremely un- suitable for the construction of reservoirs, it was found possible to establish one about 30 miles above Pagny in the valley of the Vraine, a stream falling into the Vaire, which is in turn a tri- butary of the Mouse. It could bo placed partly upon the clays of the Lias, and partly upon the outcrop of the Lower Oolite. An embankment 52 feet high would retain a volume of 1,518,000,000 gallons. The reservoir can bo filled by leading to it the water of a brook in the vicinity, by a tunnel 1,630 yards long, as observations show that the annual discharge of the brook is always superior to the volume cdntained in the reservoir. The embankment retaining the water is of earth, of which a most suitable quality is found in the immediate neighbourhood. It is protected on the inner face by a covering of masonry, built in steps so as to give a general slope of 1.^ to 1. The water when let out will follow the valleys of the Vraine, the Vaire, and the Mouse, of which the discharge will be sufficiently increased to allow of 3,600 gallons a minute being abstracted at Pagny during the summer months without encroaching upon maiiufac- tuiing interests. During five months this amounts to 778,000,000 gallons, or only half the amount supplied by the reservoir; so that the loss occurring on the way down the river is amply allowed for. The water will be raised to the Pagny reach by steam-engines on the bank of the Mouse. The Eastern Canal from Pagny to Verdun is constructed as a lateral canal to the Meuse, as the river is small and the width of the valley affords every facility for this system. Immediately on leaving the Pagny reach there is a lock upon the new canal, and the following reach continues along the hillside on the right bank of the Meuse. The soil consists of angular pebbles, vary- ing from the size of a pea to sand ; but the quantity of sand is not sufficient to till the interstices between the pebbles. The formation belongs to the Coral liag, and allows much water to escape ; especially as it was necessary to form one side of the canal in embankment, to separate it from an old borrow-pit. In making this embankment the natural soil was mixed with one-fifth to one-fourth of vegetable earth ; but the loss of water, ascertained by gauging the inlet from the Pagny reach, still amounts to 40^ gallons per lineal yard in twenty-four hours, and the water can bo seen bubbling up in the bottom of the borrow-i)it. Fine sand 22 OTHER SELECTED PAPERS. is now being thrown in to stop the leakage. Beyond the next lock the soil becomes clayey, and this with the lowered level, renders the leakage very slight. After crossing the railway from Paris to Strassburg the canal follows the right side of the valley of the Mouse, and for a short distance keeps above the line to which floods extend. It then descends and crosses the meadows forming the bottom of the valley, and enters the head of the mill-pond of Commercy. In crossing the valley the tow-path is placed at the natural level of the ground, so that the flow of the river in time of flood may not bo interfered with. The summer level of the Mouse and of the Avater in the canal is 3 feet below this ; and by the time the water rises to the level of the tow-path the navigation has in any case to be suspended, on account of the current in the parts of the river made use of. The interruption thus occasioned usually lasts about a fortnight. In excavating in this part of the valley there is, first, a depth of about 3 feet of silt, which has been deposited by the river in time of flood, and which now forms a moderately stiff blue clay. Below this, gravel is met with, which extends to an indefinite depth, and is always full of water up to the level of tlie river. In making tlio cutting for the reach which lies across the valley, it was first taken down to the surface of the water in the gravel along its whole length. Below this it is being com- pleted in short lengths, in which the water is kept down by pumping. For this purpose a double hand- pump, raising about 200 gallons a minute, has so far been sufficient. Between Commercy and Lerouvillo the railway from Paris to iStrassburg skirts the "left side of the valley, and sometimes obliges the canal, which also follows the left side, to take the channel of the riA'er. At Lorouville there are extensive (juarries of white crystalline limestone belonging to the Middle Oolite, from which all the ashlar for the locks, bridges, and other structures, as far as Charleville, and even beyond it, has been obtained. It can be had in large and perfect blocks, and is easily cut. From Lorouville a line of railway has recently been constructed, which follows the valley of the Mouse into Belgium. It is kept along the left bank as far as Sedan, and both it and the canal are on the same side for about 5 miles below Lorouville. Here there is a long detour on the Meudo, caused by a spur extending from the hills on the west side of the valley, Tho railway, which has as yet only one line of way, crosses tliis ridge in a cutting, about 45 feet above tho level of tho river ; and tho canal will pass under tho railway in a tunnel. It is being constructed by making an open cutting acrosB THE EASTERN CANAL OF FBANCE. 23 tlio ridge, which is of limestone, in which the arched roof of tlie tunnel will be built, the railway being meanwhile diverted. When completed the cutting will be filled in again. On reaching St. Mihiel the canal enters the river, passes under one of the arches of an existing stone bridge, and below the town is con- tinued as a lateral canal all the way to Verdun. The passage of Verdun presents several points of interest. The town occupies the whole width of the valley, here very narrow, and the fortifications with which it is surrounded extend to the hills on each side. The river passes through a series of arches, resembling a bridge in appearance, connected at each end with the general line of the fortifications. By a suitable arrangement these arches can be stopped, causing the river to inundate the whole valley for miles above, and so to prevent the approach of an enemy. The canal passes through the fortifications by a tunnel now completed, which can be stopped. This is done by placing across it a double barrage of beams, laid horizontally, and retained at their ends by grooves in the side-walls. By filling the interval between them with earth a cofierdam is formed, which reaches almost to the roof of the tunnel. The side-walls extend vertically down from the springing of the arch, and the tow-path, which is of wood, is supported by timber uprights, so that it can be removed in case of war. This disposition is favour- able to the passage of tlie boats, as it increases the sectional area of the water-way, and so allows the water to pass around the sides of the boat much more readily. It was proposed at first to place a lock in the tunnel in order to save masonry ; but as this would have rerpiired the foundations to bo 'carried down to a much greater depth, a matter of diflRculty in so confined a poeiticm, and as it would also have caused inconvenience in the working of the lock, it was decided to ])lace it immediately outside the tunnel, at the lower end In constructing the tunnel an open excavation was made as far as the springing of the arch; and below this, timbered trenches were carried down on each side, in which the side-walls wore founded. The earth left between tlu^m was used to BUpi)ort the centering, and when the arch was conipieted it was removed and the paving of the floor put in. None of the three bridges in Verdun permitted the passage of the canal. They are now replaced by new bridgis, two of stone and one of iron. In soHKi cases the existing piers have betni made use of. In the middle of the town, the Mouse meets the canal St. Vanne, which is the lail-raeo of some mills; and it here makes an abrupt turn (juito uiisuitalile for a navigable channel to follow. 'I'o round otf the 94 OTHEB SELECTED PAFEBS. sharpness of the angle, and to lay out a curve among houses crowded together to the edge of the water was attended with diffi- culty. The intersection of the curve with each of the property lines was determined, and by joining these points a straight front was given to each, while the curve on the outer edge of the tow-path was made continuous. All those works are now com- pleted. They have been carried out in a most thorough manner, and the improvement to the town itself and to the appearance of the river is marked. The level of the water in the town is raised by a barrage situated about ^ mile below, near the village of Belleville, after which it is named. It may be taken as a typo of all the barrages being erected along the Meuse, as the system is the same throughout. (Plato 8, Figs. 7, 8, and 9.) Between Verdun and Sedan the sysstem of canalization is modified to adapt it to existing works. At each of the towns, which occur at intervals of about 10 miles, there is a mill-dam increasing the depth of the water ; and to pass these there are the side-cuts and locks constructed since 1861. Although the depth of water in them is only 3 feet 11 inches, they are taken advantage of as far as possible ; and in all cases the mill-ponds form a most convenient way of passing the towns. They sometimes extend as much as 2 or 3 miles up the river. Where the natural features of the valley admit, a barrage is constructed below the dam, to increase the depth of water on the lower lock-sill. From this to tho head of the next mill-pond the canal leaves the river, as its depth is often less than 2 feet, and skirts its right bank to avoid tho rail- way on tho other side. At Sedan tho existing side-cut is not made use of, as tho surrounding mills render the circulation upon it inconvenient. The town consists of two | parts, which are con- nected by a long stone bridge, passing over a meadow which is covered with water when tho river is in flood. In this meadow tho new cut has been made, and the lock on it has been placed under tho bridge, as it is necessary to replace tho part over tho canal by an iron span ; a saving of masonry has been effected by onrrying up tho side- walls of tho lock to form tho abutments. Between Sedan and Charleville the system of barrages is more employed; and by moans of them the water in the river is suificiently raised to enable tho canal to use its channel for thrce- iifths of tho total distance. Tho barrages, which can be entirely removed while the river is in flood, arc placed about 3 miles apart, and raise the water to a height averaging (J feet above Slimmer level. Below Chiii-levillu this system is canied out iu a most complete THE EASTBEN CANAL OP PRANCE. 25 way, as the valley of the river is particularly adapted for it. The general fall of the river is about 2 feet in the mile, and the barrages are placed along it at an average of 3^ miles apart, giving 6 feet 8 inches as the mean height to which they raise the water. The whole fall, however, is not taken up by them, as the raised level of the water does not necessarily extend to the foot of the barrage above, although it is usually made to do so. The barrages are passed by means of side-cuts called "derivations," usually I less than 1,000 yards long, as the valley is too narrow in most cases to allow of their being extended further. In time of flood the water rises 18 feet above its summer level, ^ and the bank I of the derivations on the side next the river is carried up to this height, to prevent the destruction of the earthwork which would result from the water running over it. The locks are always placed in these derivations, and are near their lower end, the side- walls of the head-bay being raised to meet the earthwork. They are built with a double lift-wall, as in the case of those further up, and the head-gates come up to flood level also, to prevent a current from passing through the derivation, 'J'his increases their weight, and so makes the lockage a little more troublesome, but the difference is unimportant. The side-walls of the lock- chamber and the lower gates are only a f sv inches above the level to which the water is raised by the barrage ; for when the level of the river rises above this after the barrage has been removed, the circulation is no longer possible, and the head-gates have to be kept shut. In all cases the earthwork is faced with a rock protection as far up as the water ever rises. In the longer deri- vations a guard-lock is placed at the upper end, so that they can be used as harbours of refuge by boats during floods, or while ice is in the river. The barrages are all similar to each other, and, as shown in Plate 8, Fig. 7, consist of a set of sticks called " needles," placed side by side, and extending across the river. The needles are 2^ inches stjuaro, 14 feet long including the handle, and are slightly inclined up-stream. They are 8Ui)portcd by light wrought- iron frames placed parallel to the current, and connected above by cross-bars, against which the upper ends of the needles rest. They servo also to support a light foot-bridge, from which the needles can be inserted or removed. The franios are U feet 7 inches apart, and are placed on the top of a masonry foundation which extends across ' Tlu) water toho to this Iiri;,'lit all along llio river in IHKi. The grcatoat flood oil rcconl, wiiioli (lecurrcd iu 1784, uttuiuoil a hoiglit of iiii Ictt. 26 OTHER SELECTED PAPERS. i the river. (Plate 8, Fig. 8.) The barrage is divided into three equal 1 passes by two piers resting upon the masonry foundation, of which < one serves as a fish-ladder, as shown in Plate 8, Fig. 10. The { horizontal bar composing the lower side of the frame can turn in < iron sockets, anchored firmly to the masonry, and when the cross- - bars and needles are removed, the frames can be folded over flat on \ the top of the masonry. This system of barrage was invented by | M. Poiree, and was tried successfully in 1834 at Basseville on the \ Yonne. It has the advantage of enabling the height of the water • to be regulated with great precision, by the removal or replacement of some of the needles. The water is not permitted to flow over it, ; as this would render the foot-bridge impassable ; and considerable | watchfulness is necessary in case of floods after storms. At each j barrage there is a house occupied by the men who attend to the I lockage and watch the water-level ; a line of telegraph unites i these houses, whereby notice can be given from above of the time at j which floods may be expected to arrive. The cross-currents, which ] are occasioned by the partial removal of the needles, render it in- 1 advisable to place the locks immediately at the end of the barrage, : as it would be diflicult for the boats to enter the locks cither from ; above or below. This is the principal reason for the construction j of derivations for passing them. The movable part of the barrage \ is in all cases of the same height, so that the frames and needles ; may be identical all along the river. The level of water required | is obtained by adjusting the height of the masonry, which is always \ somewhat above the bottom. j This system is convenient for a river of the size of the Mouse i at Charleville ; but in Belgium, where its volume becomes greater, \ it has been found necessary to make part of the barrage in such i a way that storm-floods may flow over it. Two of the passes | are constructed according to the system of M. Chanoine, in ' which the water is retained by wooden shutters, framed like a | door. These are supported by iron trestles placed Ujxm a masonry | foundation, whose coping is considerably above the bottom, in | order that the height of the shutters may not be too great. They ; aio BO arranged that they can be laid back upon the top of the \ musoniy. To gain access to them it is necessary to have a foot- j bridge on the ui)-Btrcam side, supported by frames which can also ! be folded over upon the masonry, and whi(;h are similar to those i used for the needle barrage. This increases the width of the | masonry foundation, and makes this kind of barrage more than j twice as expensive as the other, in which the frames servo the i double purpose. In the remaining pass, the masonry is j)laced at I THE EASTERN CANAL OF FRANCE. 27 10 level of the bottom, so that by opening the barrage the pass oiui be used for navigation at certain seasons of the year. It is dosed by needles, as the height of the movable part is necessarily considerable, and they are much easier to handle than the shutters. The masonry, being at the level of the bottom, is liable to bo encumbered by pebbles and stones, which might seriously interfere with the working of the shutters, for the ironwork connected with tliem is somewhat complicated, and movable parts occur close to tJio surface of the masonry. Shutters of this kind have been used to close navigable passes, but only in rivers in which the current iii gentle and the bed is of fine sand or earth.* The above-mentioned combination is now used in Belgium, and scorns to be admitted to be the most satisfactory for large rivers, after the experience of a number of years. It also has the advan- ^igo of allowing the lock to bo placed in the river next to the navigable pass, as the overflow takes place at the other side of tlio river ; as the water rises, its level is regulated by manipulating tlio shutters, until the whole barrage has to be laid down. tThe bod of the Meuse below Charleville consists of sand and 7-avel, below which the rock is found at no great depth. The Biasonry foundation is, in most catses, built upon the rock ; but ♦ hero the rock is too deep, a layer of concrete is laid upon the gravel. Jn constructing a barrage, a cofl dam is built from the ■liDro half way across the river, enclosing a space adjoining the (|1 lore which can be pumped out. The disposition of the cofferdams employed will bo seen on referring to Plato 8, Fig. 10. Con- •idorablo caution is required in their construction as the scour in often great. When the current is very strong, a dyke of loose Blonos is first formed across the middle of the enclosure to break its force. The stones can thus bo used subsequently in the masonry. After the first half of the masonry is completed, an enclosure is built out from the opposite shore, which has to cross a part of the Diasuiiry already built. I'ho water then flows over the top of tlio masonry, so that the level of the water is higher than during ' For tlio Rroatcst height of water which ahuttcrB have yet been employed to tiiin. SCO 'Nnuvelle i)usHe navigable e'tablio en 1870 ilan.s le barrage do Port-K- Aiigliiis,' in the ' AiiiialcH dos PontH et Chnusscos,' 2° Seniestro, 1873. Also two fuiuphleta by M. Boule entitled 'Notice sur uno nouvelle piiMso navigable etablio «ii 1870 dans lo barrage do Port-h. I'Anglais,' and ' Menioire sur les barrages J nliiloa a Ibrtc cliiite.' (Dnnod, 1870.) ( Vldi' Minutca of Proceidirigs Inst. C.E., Jill. xliv. p. 2(JS). A geniTid dcscriiition of tlie dilUrcnt kindu of barrages will bo f iiiid in an article by Lagrcnc, ' Ktudo conii)arativo Bur divers Bystemes do bur- fiigea mobiloa' in tlio ' Annalea dea Ponta ut Olmuaeef,' 1" Sonioetrc, 18G6. I 28 OTHEE SELECTED PAPEB8. the construction of the first part. It is usual to extend the first enclosure rather more than half way across the river on this account. The barrages are sometimes placed whore the river is divided into two branches by an island. Their construction is then easier, as the two parts of the river can bo dealt with separately. When the water within the enclosure has been removed by pump- ing, the work is set out with extreme nicety, as is essential when ironwork, consisting of movable parts, has to be placed upon it. The principal reason for the width of the river being divided into sections, by the construction of piers, is that any slight error in the spacing of the frames may be allowed for. If the sections were too long it would become diflScult to place the cross-bars which con- nect the frames at the top, and the stability of the barrage would be impaired. The enclosure is made wide enough to include the sloping aprons (Fig. 8), which are built of dry rubble, and connect the top of the masonry with the natural bed of the river above and below ; they are hand-laid, and the surface stones require to be fitted with care. The apron on the lower side of the barrage is the part of the work most liable to yield, as during the construction of the second half of the barrage it is exposed to a violent current. The barrage of St. Joseph, near Fumay, is the highest upon the river, and is the only one in which the summit of the masonry is above the natural summer level of the water. The surface of the lower apron was grouted with cement ; but notwithstanding, a considerable portion of it was carried away during the construction of the second part of the work. It is to be observed, however, that this does not comprouiiso the solidity of the masonry foundation of the barrage itself, unless the damage extend so far as to allow it to become undermined. It is not diflicult to replace the stono thus washed away ; though it has been found, in more than one instance, necessary to do this several times before the slope of the surfac-o became gentle enough to resist the erosive action of the water. Two tunnels on the river below Charleville re(iuire to be enlarged. The first, at Fumay, has boon heightened and widened in order to allow of the construction of a tow-path along one side. The rock is solid, and no masonry is required. The other tunnel is just above Givet. The derivation of which this tunnel forms a part, cuts off" a long bend of the river, and upon it there are two locku, as the fall on the river is considerable. One of those locks is situated above the tunnel, which is remarkable, us it serves merely to increase its length and the diificulty of construction. The part of the derivation below the tunnel is not long, but THE EASTERN CANAL OF FRANCE. 29 there would be no inconvenience in the locks being placed end to end. However, this cannot be altered now, and to obtain as much advantage as possible from the existing works, a lock has still to be placed above the tunnel. It was first proposed to make use of the existing lock ; but this plan would render it necessary to heighten the tunnel, to secure sufficient head room above the raised level of the water, which would be difficult as the rock in question is much jointed and often treacherous. It was found preferable to deepen it on the floor, and to construct a new lock with a greater fall above. The roof has never been in a satis- factory state, and a masonry lining is now being put in along the greater part of its length. The lock below the tunnel will continue to be used. Below Givet there is one derivation before reaching the frontier. The lock on it is made as long as those in Belgium, and like them is built without a lift-wall. In conclusion it may be observed that the Eastern Canal, when finished, will be one of the finest linos of water- communication in France. This communication is accompanied i numeroua drawings, from which Plates 7 and 8 have been compiled. • • • • • ' • • • • • .••. ... : ' .♦•. : •• ... ■f HI £ A 5 -J" I n j'i '>1mut^:Jt ot" H-u'^epthn^iJ ot Tnv luHUUiU/Jii I CAnM D J* f nAU'Cl IKi,, \ i^ '•t\'%. {-m'Y^ v.-i-' •K* ''-»* ii^. ■j > ;li€ ,,w.*«^ ?!ArE 7 ./ »/ '"■■"%, lIUiMONTW'J .» V.'' *w y leliw -H« ^^^^ fj^ip*" ,\)\i: Ji i* y^ 7f ■! i|;. r l-vi; 4 '^) ..*/^ '"•'•"V,,! // Jl"*'! JT,— ^ ,. ;i;.«'»#■*"•„,/;* ^'♦J^ -^ 7 i\ frnm /X i«« « III HI V ^,.,^^ 1^'' ¥ ■%*'W' ij' il.3tT r -^ Nf* >i\^ I im.ssi'N' -^ ,,r*,;#,,«if««**««t««.^ ^ s^'" Ni I T ■>!.vn:AC (^^ i:^"^'^ 'y^. ^%^ j*"a,( '^ %L I* 'I'W'^ ^, f^, ■««■"♦ 25 <^ i S: \ l"f*/, ■■*j:j!« f^ s %^ '-('r- ( -7,™' Xiki: f^l JS*^ i\ )^,» 1 •%, ^j^ .iif^K '^/imn _^v ^fl<^ ISIRSAI^ -^^^4 jT U- 1 i'n£jHi..itn Vi)i iV o'-jBuai ih/t.-y'.) i ,ict, i THO< KEI.L LITH V KINC STKFET ■QVFKI ■ AliliJiS iiijn, RRAGE. VERTICAL SECTIONS OF ASSEMBLAGES. Ficj-:-: a. DETAILS. UPPER SOCKET AT N. \ s^^fey^- ^"tZT" L IRON AT Q. ••;'•.••>•■';.■^^;.•••v.■;■.•■■^^^■;a:^.■•;. .V •T..v.v%;i8^ LOWER SOCKET AT 0. '•"■''" ':[....•( Q 1—1 1 2 ±- /■'.•(■/ , J /■•..,■ ,ai>f Fiii : 8 . SECTION OF MASONRY FOUNDATION. .:\.0v i i^jio i a.iiQ 4^0^ I ailliiiiiliiiiiiiiiil .V•^V■'*^V.■.;■>:.:/••;.t;,v■A•?^V•\^^^■•■/:V;^:/■■''.■.:^V.^■;ti^v;■■yi,^^ •.'^''•^t•■ .o.«.i. •i.3S .I.AO. ■"?■•>.■ '-.iiSS .A..QO. in90nj.Vol:LV. S«HBiun 1B78.-7!), fart i, lucltfS (« « « .1 (I L-j I 1.1 .1.1 k— 1 — i — 1 — 1 — i Fuj-:-: 9 Fifj.lO. PLAN OF BARRAGE AT LES DAMLS DE MEUSE. PLATK %, '^^"^'^^ *"' ' ,,,,,™«T'n*W«^^^ Sriile Ifil' Delailtl. /Vr'