The Rotherhithe Tunnel.” by EDWARDHENRY TABOR, M

190 TABOR ON THEROTHERHITHE TUNNEL. [Minutes of

8 December, 1908. JAMES CHARLESINGLIS, President, in the Chair.

(Paper No. 3743.)

“ The Rotherhithe Tunnel.” By EDWARDHENRY TABOR, M. Inst. C.E.

THEconstruction of a tunnelunder the Thames at or nea,r Rotherhithe formed the subject of many proposalsfrom time to timeduring the last century. In 1805work was actuallystarted on a subway between Limehouse and Rotherhithe, but was aban- donedfor u-ant of funds. Parliamentary powers t.o undertake the scheme which forms the subject of this Paper were obtained by the London County Council in the year 1900, but it was not until the beginning of 1904 that a contract for its execution WRY made with Messrs. Price and Reeves, of Westminster. The design of the work follows closely that of the Blackwall tunnel,‘ which was opened in 1897, the principal differences being an increase of 3 feet in the diameter, and the greater length of the subaqueous portion, necessitated by the fact that the riveris crossed obliquely. The choice of the bestposition for thetunnel was somewhat limited,owing tothe fact that, at Rotherhitheand Shadwell respectively, thereare entrances to the Surrey Commercial and London docks, situated almost exactly opposite each other. If the tunnel were to pass under the docks on either side of the river its length would be increased undesirably, and to obviate this difficulty an oblique crossing was adopted. In this way, as may be seen from theplan (Fig. 1, Plate 3), thetunnel passes west of theSurrey Commercial and east of the London docks.

1 D. Hay andM. Fitzmaurice, ‘‘ The Blackwall Tunnel.” Minutesof Proceedings Inst. C.E., vol. CXX, p. 50.

The levelsunder the river were governed by the fact that the Thames Conservancy have the right to dredge a width of 450 feet to a depth of 45 feet below Trinity high-water mark. A thickness of 4 feet was left between this level and the top of the tunnel, the actual minimum cover during construction being about 7 feet.

GENERALDESCRIPTION. Directcommunication between two importantstreets is provided by the tunnel and its approaches. These streets are, on the south side of the river, Lower Road, Rotherhithe, and on the north side,Commercial Road, Stepney. The openapproach leaves the former street at its junction with Union Road, and the latter near StepneyStation, which twopoints are more than 3 miles apwt by road ; by the tunnel this distance is 14 mile. Commencing onthe south side, the openapproach leaves the street on a gradient of l in 36.5 and is 930 feetlong. Near its lower endthe East LondonRailway is crossed bya bridge, the roadwaybeing here about 25 feet below ground-level. Thenext

530 feet is brick tunnel, constructedby the " cut-and-cover " system, andterminating in No. 1 shaft. At thisshaft the cast-iron-lined tunnel begins, and,still descending onthe same gradient fora distance of 890 feet,attains its maximumdepth in No. 2 shaft, which is situated on the river-bank. At this point the roadway is 75 feet below ground-level. From No. 2 shaftto No. 3 is the subaqueous portion of the tunnel, 1,535 feet in length. The roadway rises 1 in 800 for the greaterpart of this distance, the up-gradient of 1 in 37 beginning about 270 feet from No. 3 shaft. Between this shaft and No. 4, a distance of 1,155 feet, the tunnel passes round a curve of 800 feet radius, and is under buildings for nearly the whole of the way. Beyond No. 4 shaft isa length of 600 feet of cut-and-cover, followed by 1,186 feet of open approach, at the end of which the street-level is once more attained.

OPEN APPROACHES. The open approaches consist of two side walls and an invert, all of 640-1 concrete. The outer sides of the walls are vertical, and the inner sides have a batter of 1 in 4. Down the back of the walls, and through the invert, isa layer of plastic asphalt, applied in three coats, to ensurewater-tightness. The side walls were built in trenches, thedumpling between them beingafterwards removed andthe invert laid. On thetop of the retaining-walls arebuilt

brick parapet walls, 8 feet high. In the southern approach a work of some difficulty was encountered inthe crossing of theEast London Railway at RotherhitheStation. The railway runsap- proximately at right angles to the lineof the approach, the roadway of whichis about 19 feetabove rail-level. Theretaining-walls of the station, which extend to a depth of 45 feet below ground-level, had to be pulled down for a length of 35 feet on each side of the railway and new abutments for the bridge had to be built behind them. A length of 25 feet of the tunnel-arch which covers part of the station was also cut away, and the street above is carried on girder-work, so as to admit more light to the platforms. The old retaining-walls, built about 40 years ago, have a backing of clay puddle, 2 feet thick, to keep out the water in the surrounding gravel which is saturated for a depth of about 22 feet above the clay. Sumps werefirst sunk oneach side of the railway and the gravel was drained ; the stratabelow consisted of clay and greensand whichcontained no water. Trenches were then excavated behind the old walls and thenew abutments were built in them, the old work being subsequently removed. Considerable dificulty was experienced in making a water-tight junction between the old and new walls. The bridge consists of two plate girders with a span of 64 feet, the roadwaybeing laid on steel-trough flooring which was riveted to their lower flanges. The whole work was carried outwithout interruption to the railway-traffic, which was continuous except for one interval of 4 hours in each week.

CUT-AND-COVERWORE.

The cut-and-cover construction consists of a brick barrel, 27 feet in internal diameter, surrounded with concrete. The brickworkis five rings in thickness, and between it and the concrete is a water- tightlayer of asphalt.The circular shape was adopted so that the subway, which is formed in the space under the roadway in the cast-iron tunnel (Fig. 3, Plate 3), would extend also through the cut-and-cover portions. The excavation was taken out to the full width and depth, the latter being 56 feet as a maximum. Heavy timbering was necessitated, and great carehad to be taken to avoid settlement in adjacent buildings,which in some cases stoodonly a few feet from the trench.Notwithstanding this, a small amount of damage was caused, theground apparently settling towards the trench on a 1 to 1 slope fromthe bottom of the excavation. Water was met within the Thames ballast on both sides of theriver, but in

moderate quantities only, theamount pumpedbeing about 15.0 gallons per minute on the south side and considerably less on the north. Ornamental arches of red graniteare built at the ends of the tunnel where they join the open approaches.

SHAFTS(Fig. 4, Plate 3). Thefour shafts are all of similarconstruction and differ only indepth. Eachconsists of a caisson of two skins of steel plating, 60 feetin external and 50 feet ininternal diameter, the space between theskins being filled withconcrete. The thickness of the plates varies from ;. to 4 inch and the inner skin is tapered out in the usual way at the bottom to form a cutting edge. The outsides of the caissons are cylindrical,each ring of plating being battered outwards to bring the top edges of all the rings to onediameter. All the joints in the plating werecaulked, making the caissonswater- andair-tight. In each shaftthere are two openings for the tunnel, 32 feet in diameter, and lined with $-inch steelplating rigidly attached to both skins by angles and plate diaphragms. During sinking the openingswere temporarily closed with plugs, composed of steel plates and angle-bars in thecase of the twodeeper shafts, and of wood sheetingsupported by steeland timber girder-work in the others. About 13 feetabove thecutting edge in each shaft isfixed a permanent air-tight floor, consisting of a system of six plate girders, 8 feet6 inches deep. Three of thesegirders are parallel and cross the other threeat right angles, all being in the same horizontal plane.Each girder is made effective throughoutits length by cover-plateson the flanges where theyare cut through by other girders. These cover-plates are of equal sectional area to the flange plates and angles cutthrough. The webs of thegirders are not continuous, butare strongly connected attheir intersections by 6-inch by 6-inch by &inchangles. The spaces between the main girdersare bridgedby smaller girders 2 feetin depth, and by buckled plates about 3 feet 6 inches square. The air-tight floors were designed for an upward or downward pressure of 20 lbs. per square inch, any excess of air-pressure over this which might be required being counterbalanced by loading the floor. Provision was made in each caisson for the attachment of a temporaryair-tight floor, of similardesign, above thetunnel openings. [THE INST. C.E. VOL. CLXXV.] 0 Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. 194 TABOR ON THE ROTHERHITHE TUNNEL. [Minutes of The shafts are lined with brickwork with a white glazed facing ; andspiral staircases,giving access tothe adjoining streets, are provided inthe two deepershafts. Brick buildings covered with glazed domes are built on the tops of these two shafts. The other two are open to the atmosphere.

CAST-IRONTUNNEL. A crosssection of the completed tunnelis shown i- Fig. 3, Plate 3. Each ring of the cast-ironlining consists of fourteen ordinary and two special segments and a key. Details are shown in Figs. 5 and 6. The metal in the body of the segments is 2 inches thick under the river and 12 inch in the land tunnels. The flanges are 14 inches deep, and have recesses 2 inches by g inch in their inner edges forrust-jointing. The bolts are 14 inch in diameter and number 164 per ring, the longitudinal joints having five bolts each, intwo rows, andthe circumferentialbolts being equally spaced allround, so thatthe ringscan be builtin any relative position to break joint. The plates are all 2 feet 6 inches wide, and those of the heavier section weigh 19 tons per ring. All the flanges were machined on the faces and the rust-joint recesses referred to were caulked,first with lead wire, and then with theusual mixture of cast-iron borings and sal-ammoniac. Pneumatic hammers were used for caulking in the lead, but were not found satisfactory for the rust-jointing. In orderto prevent leakage the boltswere fitted with soft lead washers, or grummets, under the heads and nuts, the holes in the flanges being bevelled or countersunk, so that the lead is squeezed between the iron washers into the bevelled recesses in screwing up the nuts. In each plate there is a l+-inch tapped hole which was used for grouting and afterwards plugged.The flanges were faced in specially-constructedmilling-machines. As may be seen on the plan (Fig. l), a portion of the tunnel is curved to a radius of 800 feet. For this curve each ring of lining has a taper of l&inch in the diameter, the two faces of thering being, of course, intvo convergingplanes which meet at a distance of 800 feetfrom thecentre of the ring. In order to obtainthis result,a lathe was speciallybuilt, capable of facinga complete ring at one operation. The longitudinal joints were first machined in the ordinary way and the ring was bolted together ; it was then set in the lathe and faced on both sides at the same time. Similar taper rings were used in a few places in the tunnel where the face of the lining was found to be getting out of square. All the lining wits cast and machined at theStaveley Ironworks, near Chesterfield.

ROADWAY,LINING, ETC. The roadway is 16 feet wide between the curbs, and the two footways in the tunnel have each a width of 4 feet S$ inches. The paving isof Aberdeen granite sets on the gradients, of and compressed asphalton the comparativelylevel portion underthe river. The footways are paved with York stone flags. A 9-inchbrick archsupports the roadwny, the surface of the latter being 8 feet 6 inches above the invert of the tunnel, in order to obtain the desiredwidth. The spa,ce between the arch and the invert forms a subway, 13 feet 6 inches wide and 6 feet 11 inches high, which is not utilized for the drainage of the tunnel. This is provided for by the two pipes, shown in the crosssection, which receive the water from thcgullies in the roadway and lead to a sump in No. 2 shaft, from which the water mill be raised to thesurfaceby electrically driren three-throw pumps. Thecast-iron tunnel is lined with concrete, faced withwhite glazed tiles ; theinner face of the tilesis 4 inchesinside the flanges of the segments, thusgiving a clear internaldiameter of 27 feet. The brickwork of the cut-and-coveris faced with similar tiles, and the shafts are lined with whiteglazed bricks. Thetunnel is lighted by three parallel rows of electric glow- lamps, spaced 30 feet apart in each row. No special provision is made for ventilation, it being found that under similar conditions at the Blackwall tunnel the natural ventilation from thc shafts and open ends is sufficient.

RIVER-WALL. Shaft KO. 2 is sunk with its centre on the line of the old river- wall, the scheme providing for the construction of a new wall about 40 feet farther out than the old one. This work was executed before the shaft was taken in hand, so that the cnieson might be sunk behind the new mall. The wall isbuilt of 8-to-l concrete, faced with brickwork. A single-pile coffer-dam was first constructed and strutted back to the old wall, and the water was pumped out. The mail was then built in the dry. Considerable pumping was required while getting in the foundations, as it was impossible to make a wtter-tightjunction between the coffer-dam andthe oldwall at each end. The piles of the coffer-dam, after the wall was finished, were cut off at thc level of the foreshore. No sign of movement 02 Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. 196 TABOR ON THE ROTHERHITHE TUNNEL, [Minutes of

has been observed in the wall due to the sinking of the caisson at a minimumdistance of only 10 feetfrom it, or tothe subsequent tunnelling underneath.

STRATA. The strata met with are indicated on the section, Fig. 2, Plate 3. The Thames ballast is about 25 feet thick at the southern end of the works, and thins out gradually towards the river. On the north side the ballast is only 2 feet thick at the river-bank, and 18 feet at the upper end of the approach. The London clay is absent on the south side and under the river, but attains a thickness of 38 feet at No. 4 shaft. With theexception of the portion adjoining this shaft, the tunne passes almost entirely through the formationknown as the Woolwich and Reading beds. Theseconsist of blue ancl mottled clays, sands with and without shells, hard silt or sandy clay with shells, pebbly gravelwith greensand, and a layer of rock. The clays overlie the snncls, which cont:rin water at moderate pressure.

PLANT. In view of the uncertainty as to the exact nntnre of the ground under the river, and of the importance of the buildings over and near the line of the tunnel, the contract for the works stipulated that the shnft-sinking and tunnelling should be carried out under compressed air. Underthese circumstances the plant employed in the execution of the work becomes of great importance. Near the site of No. 3 shaft, space was available for the contractor’s yard, and upon it, as soon as the works were put in hand, were built engine- and boiler-houses, repzir-shops, stores, etc. Sufficient room was left for the storage of about 4,000 tons of cast-iron tnnnel-lining. The engine-roomcontained six cross-compound condensingair-compressors, with a total capacity of 1,000,000 cubic feet of free air per hour, at a pressure of 2%lbs. per square inch above the atmosphere. The air was drawn from outside the engine-house through a large steel pipe with branches to each engine, and after compression it passed to the receivers through pipes on the roof, upon which water was allowed to trickle for cooling purposes. Besides the largeengines there were six small compressors for supplying air at about 80 lbs. per square inch for grouting and for working pumps, hauling-engines and pneumatic tools. A small air-compressing plant was also installed at No. 2 shaft on the other

side of the river.Two compound hydraulic pumping-engines delivered water at a pressure of 1,100 lbs. per square inch to two accumubtors.There were also four electric generator-sets for lightingand power. Steam at a pressure of 120 lbs. persquare inch was supplied by seven boilers of the Scotch or marine type. The exhaust steam was condensed in independent condensers and the water was cooled in a tower. The fact that all this machinery worked throughout the contract without any hitch or breakdown reflncts great credit on those responsible for its installation.

AIK-LOCKS.

As it was foreseen that a considerable amount of tunnelling would have to be done while No. 3 shaft was under air-pressure, before bulkheads could be built in the tunnel, the air-locks for this shaft were designed so as to deal with large quantities of material. Two steel nir-shafts or tubes, 8 feet 9 inches in diameter, were erected side by side upon theair-tight floor, to which they were rigidly attached. At their upper ends, which were about12 feet above gronnd-level, they were connected with a common air-chamber into which the air-locks opened. The air-shafts were fitted with timber guides extending to the working-floor at the bottom of the shaft; in the quidesran cages, each holdingone wagon and worked by hydraulic hoists also fixed in the shafts. The skips or wagons ran straight from the cages into the air-locks, each of which held two wagons. The locks were 16feet long and 5 feet 9 inches in diameter,and were fittedwith cast-steel doors. The excavated material was tipped from the working-stage at the level of the air- locks into barges in the river alongside. The same combination of air-chamber and locks was used forsinking the caisson, theair- shafts being then attached to the lower or permanent air-tight floor. A similar :trr:rngement was used in No. 2 caisson, a single air-shaft inated of two being fitted.

BULKHEADS.

In nearly all the tunnels hitherto constructed under compressed air the bulkheads for retaining the air have been built of brickwork or concrete. Theirconstruction in thesematerials is simple, but their removal is slow and costly. Steel bulkheads were therefore preferred by the contractors as being more e8sily erectedand removed,

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. 198 TABOR OX THE ROTHERHITHETUNNEL. [Minutes of The bulkhead shown in Figs. 8, Plate 4, consisted of a diaphragm of #-inch steel plating supported by rolled joists and built-up beams crossingeach other. The load onthese was taken by fourmain girders, 2 feet deep, radiating from thecentre. Each of these girders took a bearing on one of the flanges of the iron lining at its outerend, while atthe centre the combined load fromall four girders was transmitted to the tunnel behind by four steel struts which extended diagonally, two in a vertical, and twoin a horizontal direction,to the iron lining. These strutshad their outer ends connected in pairs by vertical and horizontal ties, so that no bursting pressure was exerted on the lining. The air-pressure for which the bulkheads were designed was 35 lbs. per square inch, the total load on the structure at this pressurebeing over 1,500 tons.The skin plating of thc diaphragm was bolted round its outer edge to a ring of ;-inch plates which had been inserted between two rings of segments at thetime the liningwas erected. The ringof plates projected about 4 inches inside of the inner edges of the cast-iron flanges of thetwo rings.Three air-locks were provided,a small emergency lock near the top of the bulkhead, a material lock, 5 feet 9 inches in diameter and 36 feet long, and a men’s lock, 7 feet in diameter and 18 feetlong, which was also used formaterials. The greater portion of the length of the air-locks was arranged to be on the atmospheric side of the bulkhead, so :\S to put RS much as possible of the material of the locks in tension. In the length of tunnel under the river a raised gnngway, about 3 feet wide, was fixed on brackets to the side of the lining, and extended from the travelling stage at the ren.r of the shield to thc emergency air-lock. A hanging diaphragm of wood, with its lower edge at a level below that of the floor of the emergency lock, was kept about 100 feet behind thc shield, being moved forwnrd periodically.This diaphragm was builtup of twothicknesses ofwood sheeting, with tarred felt between them, and was caulked wherc it abutted against the flanges of the iron lining. Its provision ensured that, in theevent of an inrush of water, the portion of the gangway behind the hanging diaphragm would never be submerged as long as the air-pressure was maintained.

SEIELDS. Two shields were constructed from the contractors’ designs, one being used for driving the tunnel under and southof the river, and theother for the remaining length of tunnel on the north side. The latter shield was about 5 feetshorter than the other, as

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. Proceedings.] TABOR ON THE ROTHERHITHETUNNEL. 199 it had to work rounda longcurve of 800 feetradius. In other respects the two were similar, and the longer one, which was built and started first, will be described. The shield, shown in Figs. 7, Plate 3, weighed in working order about 380 tons. It was built of cast-steelsegments, sixtyin number, in three rings, the segments in each ring breaking joint with those in the next. The flanges were 2 feet deep and the metal generally was 3 inches thick. They were strongly bolted together, and in order to increase the resistance to the boltsshearing and the segments sliding upon one another, steel keys, 2 inches by 1 inch in section, were fitted in corresponding grooves in the flanges. The front ring of segments formed the cutting edge, the longitudinal flanges of thering tapering outwards towards thefront The tail of the shield, 7 feet 3 inches long, was built up of three thicknesses of &inch steel plates, riveted together, and was bolted to the rearmost ring of segments, which it overlapped, and which was reduced in diameter to receive it for a length of 2 feet 9 inches. The front portion of the shield was divided into sixteen compartments for working purposes, by three vertical and three horizontal partitions. These partitions consisted of three thicknesses of 1-inch steel plate, riveted together, and were bolted to the flanges of the segments, and connectedby 6-inch angles attheir intersections. The front edges of the partitionswere chamfered or bevelled off, and were all in the same plane as the circular cutting edge. Face-rams, 5 inches in diameter and of 2-foot 6-inch stroke, were arranged in pairs in the different compartments, working in cylinders bolted to thevertical partitions. In each compartmentthere was awater- seal or trap, consisting of a hanging diaphragm of steel plnte, with its lower edge about 3 feet above the floor and 4 inches below the upper edge of another diaphragm which rose from the floor about 3 feet behind it. The main shield-rams, of which there were forty, had a diameter of 9 inches, and a stroke of 3 feet 6 inches, and were fitted with small internal draw-backrams. Bothrams and cylinderswere machined out of forged steel, and were constructed for a working- pressure of 3 tons per square inch. At this pressure, and with all the rams at work, the total force exerted was about 6,000 tons. On the ram-heads were boltedgrouting-ribs consisting of steelbars, 5 inches wide and 3 inches thick, which just clea~ed the inside of the tail of the shield, and almost abutted against each other at their ends, so as to form a nearly continuous ring round the shield when all the rams were out to the same extent. For building the lining, two hydraulic erectors wereused, consisting of the usual arrangement

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. 200 TABOR ON THE ROTHERHITHE TUNNEL. [Minutes of of sluing rack and pinion and sliding arm. They were fixed to the backs of the two inner vertical partitionsof the shield. A working-stage on wheels, about 50 feet long and 25 feet wide, was attached to and travelled on temporary rails behind the shield. The stage had two floors and carried thehydraulic pumpswhich suppliedpressure tothe rams in the shield.These pumps were actuated by compressed air, and were supplied with water from the accumulators on the surface at a pressure of 1,100 Ibs. per square inch, which they intensified to the pressure required to move the shield, up to a maximum of 3 tons per square inch. The grouting- pans, which were of the ordinary pattern, with agitators worked by hydraulic motors, were also fixed on the same floor. In addition to carryingthis machinery, thestage served asa working-platform for bolting up and caulking the lining, and as a receptacle for the spoil which was cast out of the upper half of the shield on to its lower floor, and thence fell through shoots into wagons below.

SIXKINGTHE SHAFTS. As it was arranged to start the tunnelling from No. 3 shaft on the north side of the river, arrangements were made to sink this shaft as early as possible. Work was begunupon the site on the 1st October, 1904, and the cutting edge of the caisson had sunk about 22 feet when work was suspended until compressed air could be applied. The cutting edge was then in a bed of peat, and it was thought that settlement might be caused in the buildings adjoining the site if further excavation were attempted without compressed air. The compressingmachinery, air-tight floor, air-locks,etc., having been installed and completed, sinking under compressed air was started on the 32ndMay, 1905, andthereafter proceeded steadily until the 13th August, when the caisson reached its final level,having sunk a distance of 67 feet in 63 working-days. The air-pressurerequired todry the bottomvaried from 5 Ibs. to 18 lbs. per square inch, the water encountered being land-water, of which the level was notaffected by thetide in the river. No difficulty was experienced in keeping the caisson sufficiently level, which was principallydue, no doubt,to the parallel sides. The concrete was filled in between the skins as the erection of the steel- work and sinking proceeded, and it was found that the weight of the structureitself was sufficient to cause it to sink. The usual procedure was to excavate about 2 feet depth of ground at a time, starting in the centre and working outwards, more or less ~f 8 berm being left all round, inside the cutting edge, according tp

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. Proceedings.] TABOR ON ROTIIERHITHETHE TUNNEL. 201 the hardness of the material.About every 2 days,advantage was taken of a meal-interval, when no men were in theworking-chamber, to lower the air-pressure by a few pounds; the caisson then sank the 2 feet, occasionally going down suddenly and raisingthe air-pressure by 3 lbs. or 4 lbs. per square inchin a few seconds. The skin-friction was found to be 38 cwt. to 4 cwt.per square foot. Thestrata passed through werepeat, sand, gravel, clity, sandand clay with shells, hard mar1 or rock, and greensand and pebbles. As soon as the caisson had reached its final level, the space under the air-tight floor was filled with concrete. The air-lockswere then raised and connected to thetemporary air-tightfloor which had been fixed in the c:tisson, so thatt tunnelling undercompressed air could be commenced. Shaft No. 2, as already mentioned,was sunk behind the new river- wdl on the south bank. The cutting edgewas erected on a level bed, which was prepared partly on the old foreshore and partly in a recess excavated inthe old qmy-wall.Sinking wasbegun in November, 1005, air-pressure was applied in May,1906, andthe shaft reached its final level on the 19th July following. The skin- friction in this shaft was found to be greater than in the case of No. 3, attaining a maximum of nearly 6 cwt. per square foot. The material passed through was very similar, although somewhat harder; hut probably some of the additional resistance to sinking was due to the fwt that the caisson was built slightly oval about the middle of its height. The lower air-tight floor had to be loaded with about 2,300 tons of ballast to cause the caisson to sink to thedesired level, the air-pressure under the floor being reduced by 5 lbs. per square inch at the same time. Nos. 1 and 4 shafts were sunk without the use of compressed air, the excavation being kept dry by pumping. As soon as the Thames ballast was passed, little water was encountered.

TUXNELLIRG. It was wrl-anged that the tunnel should be driven first from No. 3 shilsft southwardsacross theriver and on to No. 1 shaft,the remaining portion north of the river being also started subsequently from the same shaft. From borings on both banks and from dredgings in the river, it was thought probable that a bed of clay extended across the river. Since, however, the strata in other places on the Thames have been found very irregular, this was by no means certain, and it was con- sidered likely that partof the tunnel-face would be in gravel or sand in communicaticm with the river. In order to Qbtais defrnite

information, the contractors elected to drive a small pilot-tunnel or heading across the river in adva,nce of the main tunnel. The pilot- tunnel was 12 feet 6 inches in external diameter andwas lined with cast iron, the top of this lining being about 2 feet below the top of the main tunnel. The small tunnel was driven with a shield fitted with a rotary excavator of the pattern introduced by Messrs. Price and Reeves, which has been successfully used in the construction of several of the London tube railways. The arrangement isshown in Figs. 9, Plate 4. The shield was of theordinary type with a cast-iron cutting edge and body, andsteel-plate tail. It was providedwith ten 7-inch hydraulic rams, six of which were below its horizontal axis. A little above this level, two stronggirders .were fixed horizontally, carryingin bearingson their lowerflanges a steel shaftwith its centreon the axis of the shield. On thefront end of this shaft was keyed a heavy casting, to which were bolted sixradial arms, built up of steelchannels andextending out to, andabout 1 foot infront of, thecutting edge. A circularrack, 10 feet in diameter, with internal teeth, was fixed to the backs of the radialarms and was rotated,through treble-reduction gear- ing, by an electricmotor of 52 HP. enclosed inthe shield. Projectingabout 6 inches in front of theradial arms were steel cutters, which,when thecutter headrevolved, scored concentric grooves inthe materialforming the face. Thepoints of the outercutters describeda circle 1 inch less in diameterthan the shield,leaving thisamount to be removed by thecutting edge itself. Behind the cutters, dredger-buckets, also fixed to the radial arms, scooped up the spoilbroken down by thecutters and discharged it intoa shoot inthe upperpart of the shield. The revolvinghead advanced simultaneously with the shield,being forced intothe face by the hydraulic rams ; therear bearing of thecentral shaft was fittedwith athrust-block. A travelling belt conveyor received the excavated mxterial from the shoot into which it had been discharged by the dredger-buckets, and delivered it into wagons in the tunnel about 25 feet behind the shield. The method of working was as follows. A ring of lining having been putin place and bolted up, the conveyor, which hadbeen temporarily run back out of the way, was put into position under the shoot and started. The motordriving the excavator was then switched on and any loose material in the face was removed by the dredger. Hydraulic pressure was next admitted to therams, and the shield and excavator were fed into theface. The amount of pressure on the rams was regulated during working according ts the power

required to drive the excavator, the driverof the pumpwhich supplied pressure to the rams having the hydraulic gauge and ampere-meter constantly under observation. The usual time occupied in excavating the length of a ring (1 foot 8 inches) was about 20 minutes in clay, increasing to 45 minutes or more in hard material. As soon as the shield had advanced a sufficient distance, the machinery was stopped, the conveyor was run back, and the ringof lining was erected by hand in the ordinary way. As a precaution, the shield was fitted with a water-trap,consisting of the usualpair of half-diaphmgms ; the spoil was delivered through a hole in the upper one, which could be closed by a sliding door at a moment’s notice. As soon as the shield and excavator, which had been put together on a cradle in the proper position in the shaft, were ready to start, air-pressure was put onunder the upper, or temporary, :+tight floor. A circular opening, 13 feet in diameter, was then cut through the temporary plug in the side of the shaft next the river and the shield started.This took placeon the12th October,1905, and 4 weeks later the excavator had advanced far enough to enable a bulkhead to be built in the small tunnel. This bulkhead, which was built of brickwork and accommodated two air-locks,was brought into useon the 9th November, and the air-pressure was taken off the shaft, so that thelarge shield could be erected and got ready for work under atmospheric conditions. The pilot-shield made steady progress, the presence of a bed of rock, 3 to 5 feet in thickness, preventing it from advancingas rapidly as similar machines have done in clay. This bed of rock extended right across the river,and was overlaidby clay and sand, the materials under it being sand a,nd pebbly gravel. The average rateof progress was 13 feet 6 inches per day of 24 hours, and the shield was driven continuously until the 22nd January, 1906, when it was stopped, being then distant only 150 feet from No. 2 shaft, which had not been sunk at that date. It was found in driving the small tunnel that an air-pressure of 12 to 21 lbs.per square inch was enough to keep the face dry. This pressure nearly counterbalanced thehydrostatic head, dueto the river-water, at thetop of the shield, but verylittle water came in from the lower part of the face. In the meantime the main shield had been built in No. 3 shaft, and was ready for starting by the time the progress of the small tunnel was stopped. Air-pressure was again put on under the air- tight floor, and the shield wasmoved close upagainst the plug, which was then removed. This was effected without difficulty, and the shield entered the ground. An air-pressure of 16 to 21 lbs. per

square inch, according to the height of the tide in the river, was maintained, and it was found that this was sufficient to keep the face nearlydry. A littleair escaped intothe surrounding strata, and appeared in the form of bubbles over a considerable expanse of the river. This gradually decreased as the shield moved away from the shaft. Progress at first was slow, but was maintained steadily ; starting on the 17th February,1906, the shield advanced80 feet in the first month.The face stood well, and as a rule needed littlesupport. Horizontal boards strutted off' the face-rams wereused when required. A length of 2 feet6 inches, enough forone ring, was usually excavated infront of the shieldexcept roundthe outside, where some materid was left to be brokendown by the cutting edge. The bed of rock, which appearedfirst in the lower part of the face, considerably delayed progress. This material was at times extremely hard, and as much as 5 feet thick, and it was necessary to excavate it quite clear to the cutting edge. Blasting with small charges was adopted for a time, but was abandoned when the cover over the tunnel became thin, for fear of disturbing the river-bed. After each advance of 2 feet 6 inches, the ringof lining was built and the space left by the tail of the shield, outside the iron, was grouted up at once. From 2h to 3 tons of lias lime, which was used neat, wererequired for each ring.The removal of the sndl tunnel- lining, of which two rings had usually to be taken out each time, was left to the last, so as to retain the benefit of its support to the face as long as possible. Thefact that so large a portion of the upper part of the face was occupied by this tunnel was found to be a considerable advantage, as the unsupported area of the face was SO much reduced. The small tunnel had been grouted throughout, partly to prevent leakage, but. chiefly to avoid settlement of the groundabove, which mightbreak upthe crust of the river-bed. The grout was found in excellentcondition when the lining was removed. In somecases the ground was soft and had fallen down upon the iron, but the grout had found its way into the interstices left above and apparently filled all the voids. The distance which lime grout will travel through fissures in this way is remarkable, as was noticed later when the shield was approaching No. 1 shaft. The timber plug in the side of this shaft had been grouted to make it watertight, and when the shield was still at a distance of 60 feet thin seams of limewere found inthe face. These fissures were previouslyknown to exist, for when thetunnel was as much as 1,700feet distant under the river, a small quantity of air was observed to be blowing through the sheeting of the same plug iuto the shaft,

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. Pi'oceediq~..] TABOR ON THE ROTXERHITHI! TUNNEL. 205 As soon as the travellingstage, with the hydraulicand other machinery installed upon it, was in thorough working order, progress gradually improved, A bulkhead was built in the tunnel 320 feet from the shaft, but owing to various delays was not brought into use untilthe end of July, when the shield had advancednearly 700 feet. After the bulkhead was in operation, the tunnel behind it, as well as No. 3 shaft, was opened to the atmosphere, and some of the plates of the air-tight floor were removed, so that the partsof the second shield could be lowered forerection. Thejoints in the lining of the length of 320 feet of tunnel opened to the atmosphere had all beenpreviously caulked, and were found to be practically watertight. On the 14th August, 26 weeks from the start, half the tunnel had been drivenand the shieldhad only about 8 feet of cover. The clayabove, however, was almostimpervious, very few air- bubbles beingvisible inthe river. The air-pressure was adjusted as nearly as possible to balance the head of water at the top of the shield at all states of the tide, so that the risk of blowing up the river-bed wasminimized. The leakagefrom the lower part of the facewas slight. The rockwas atthis time in about the middle of the face and couldbe excavatedwith comparative ease. Steel piles were driveninto it by the shield,breaking it down so as to beeasily removed. Timberpiles werealso used for the same purpose in the softer parts of the face. In the secondhalf of the tunnel,partly owing tothe rock beingmore easily dealt with, andpartly to the greater capacity of the air-locks inthe bulkhead,rapid progresswas made, reaching a maximum of 12feet 6 inches in 24 hours, and of 267 feet in the month of October. By this time the shield had passed the deep portion of the river a.nd had a substantial cover, so that there was little likelihood of com- municationbeing opened withthe river. No claywas actually deposited, but during the time when the cover was thin, a hopper- b:wge loaded with puddled clay was kept moored near the tunnel for use in emergency. As soon as No. 2 shaft was ready, air-pressure was put on under the upper floor and a small timbered heading, 6 feet by 4 feet, was driven through the plug back to the face of the small tunnel. This proved very useful as a ventilator, the surplus air from the main tnnnel being allowed to escape at No. 2 shaft. On the 21st November the shield arrived at the side of the shaft and a week later had advanced to the middle. It was found to be in perfectcondition, and,after slight repairs to the hydraulic apparatus, was ready to proceed wit11 the remaining lengthof tunnel.

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. 206 TABOR ON THE ROTHERHITHE TUNNEL. [Minutes of Little difficulty was experienced in keeping the tunnel sufficiently nearto line and level. The line,having been established from plumb-bobs hanging in No. 3 shaft, was carried forward into the small tunnel, and on checking it after the heading had been driven through,no error of importance was discovered. Theiron lining showed a slight tendency to settle at the top during erection, but did not move after being bolted up. The inner flanges of the liniq are nowhere more than 2 inches out of their correct position, and there being, as designed, a thickness of 3$ inches between the tile lining and the iron, the former can be set absolutely true to line. Tunnel between Slmfts Nos. l and 2.-Compressed air was kept on, and the shield was moved up to the remaining plug in No. 2 shaft at the beginning of December,1906. This being removed, tunnelling mas resumed under conditions very similar to those already experienced. At a distance of 100 feet from the shaft a bulkhead was built, which was brought into use on the 5th March, 1907. Simul- taneously with this bulkhead, another had been constructed in the tunnel under the river, about 80 feet from No. 3 shaft, to confine the compressed air in the tunnel north of that point. When both bulkheads mere put under pressure, therefore, the greater part of the tunnel under the river, No. 2 shaft, and 100 feet of tlle south approach tunnel were opened to the atmosphere. Asthe tunnel between shafts Nos. 1 and 2 is chiefly under private property, special care was taken with the grouting, a mixture of lime and cement being used. The air-pressure, no longer varying with the tide, was kept at 15 lbs. per square inch. The works of the South Metropolitan Gas Company were passed under without any settlement being observed. On the 30th April theshield encountered the ballast for the firsttime, being then under a metal-refining works. Some escape of air took place, which was noticeable on the surface. The ballastgradually increased in depth as the shield worked up the gradient, attaining a maximum of 6 feet from the top of the shield. The face waspoled and allowed to come back into the shield as it advanced. Under these conditions the pressure required to move the shield naturally increased, as the cutting edge was forcedbodily intothe ballast.All the rams were usually required, with a hydraulic pressure of 2 to 2$ tons per square inch, equal to a total force of 4,000 tons exerted. The shield arrived at No. 1 shaft on the 16th July. Allowing for a stoppage of 7 weeks while the bulkhead was built and caulkingwas carried on,the average rate of progress was 38 feet per week of 6 days, almost exactly the same as under the river. Tunnel between shafts Nos. 3 and 4.-The second shieldbeing

Downloaded by [ University of Liverpool] on [19/09/16]. Copyright © ICE Publishing, all rights reserved. Proceedings.] TABOR ON THE ROTHERHITHE TUNNEL. 207 ready in No. 3 shaft, the air-tight floor was again plated up and air- pressure was applied on the 9th January, 1907. This put the whole length of tunnel between the shaft and theface of the southern tunnel under air-pressure, and workat this face was somewhat hampered by all Dhe material having to pass through the air-locks at the top of the shaft. Tunnelling was therefore suspended until a bulkhead was ready, and the men were transferred to the northern tunnel. The remaining plug in No. 3 shaft having been removed, the face was exposed and was found to be very much broken up, the strata lying at all angles between horizontal and vertical. To keep out the water an air-pressure of 19 lbs. per square inch was required. London clay appeared in the upper part of the face about 50 feet from the shaft, and little or no trouble was experienced thereafter. The clay stood up well, and could be got out the full 2 feet 6 inches in front of the cutting edge. A bulkhead \vas built about 100 feet from the shaft, andthe air-pressure was maintained at 14 Ibs. persquare inch. Progressimproved rapidly, and was not affected by the shield having to be drivenround the curve. The best week’s work was twenty-eight rings, or 70 feet, in 6 days. The air-compressors were stopped on the 16th August, and 2 days later the shield arrived at No. 4 shaft, all tunnelling being then completed. Theaverage rate of progress, allowing for 3 weeks’ stoppage for building the bulkhead and other causes, was 45 feet per week. Owing to its being 5 feet shorter, considerably less pressure was required to move this shield than the other. The reduction in length, however,involved some loss of strength,and it became slightly oval whenbeing driven round the curve,recovering its circular shape when the tunnel again became straight. For the last 500 feet of its course this tunnel passes under the buildings on the north side of Broad Street, Ratcliff, and although the top of the tunnel is as little as 16 feet below the cellar-floors in some cases, no claims for damage due to subsidence have been substantiated. The temporary roads for the spoil-wagons were carried on timber raised about 3 feet above the invert of the tunnel. Electric winches were used forhauling. At times,when rapid progresswas being made, the capacity of the air-locks in the bulkhead was taxed to its utmost, &S many as 1,000 wagons in each direction having been dealt with in 24 hours. Conapressed-Air Wor1c.-All work under compressed air was carried on in three shifts of 8 hours, and as a rule about sixty men were employed on the pressure side of the air-locks. At one time, while both shields were at work, there were as many as 450 men in he three shift>s. When the whole tunnel between the two shields, as

well as the two shafts, was under air-pressure, the volume occupied by the compressed air was 1,500,000 cubic feet. Owing chiefly, no doubt, tothe low pressurerequired, very little compressed-air illnessoccurred. All the menwere medically examinedbefore starting work and periodically afterwards. Lining, etc.-As soon as the tunnel was opened to the atmosphere, the construction of the roadway was taken in hand.The subway arch was built, the concrete above it was laid, and the granite curb and channel was fixed. Thecurbs were then taken asguides for setting thecentres for the concrete lining and tile facing. As already stated,the iron lining was found to be fairlywatertight on being opened to the atmosphere. Such leakage as existed came from the bolts and was effectively stopped by re-grummeting. After the lapse of some months, however, when the concrete lining was put in hand, a large number of small leaks developed in the caulking. On these being cut out and made good, the caulking was found to be in a sound condition ; and there can be little doubt thitt the movement of the lining, due to changes of temperature, causes the joints toopen. It appears, therefore, that absolute watertightness is unattainable with rust-jointing. The contract-price of the works was 21,088,484, which includes the cost of a considerable amount of incidental street widenings. It is anticipated that the actual cost will be a little less than this sum. The workwas designed by, and cnrriedout under thesupervision of, Mr. Maurice Fitzmaurice, C.M.G., M. Inst. C.E., the Chief Engineer of the LondonCounty Council, for whom theAuthor acted as ResidentEngineer. The contractors were represented by Messrs. J. H. Price, Assoc. M. Inst. C.E., and James Brown, M. Inst. C.E. Thetunnel wasopened for traffic on the12th June, 1908, having occupied a little more than 4 yearsin construction. The period allowed by the contract was 53 years.

ThePaper is accompanied bytwelve tracings, from which the il1ustr:ttions reproduced in Plates 3 :II~4 Imve been selected.

l CROSS SECTION OF TCNNEL UNDER RIVER. A ! HALF ELEVATION BACK OF SHIELD. ' i LONGITUDINAL SECTION OF SHIELD.