Disaster befell the first great viaduct across the Firth of Tay, but the engineers, undaunted, succeeded in building a new bridge, more than two miles long, across the windswept waterway
SEVENTY-FOUR SPANS, excluding the approach spans, form the second bridge across the Firth of Tay. The bridge has a total length of 10,711 feet, or 2 miles 53 yards 1 foot. Its building occupied nearly five years and it was opened for passenger trains on June 20, 1887, having cost £670,000, or about £62 10s a foot. This photograph was taken shortly after the bridge had been opened.
WHEN engineers first considered, as a practical possibility, the spanning of the Firth of Tay at Dundee, they were venturing into the unknown, for in those days the Menai Bridges of Telford and Stephenson represented the last word in bridge construction, and the Firth of Tay was a different proposition from the relatively narrow Menai Strait.
It was known that the Scottish firths presented difficulties additional to that of distance. Any ferry-boat skipper could tell of the terrific gales which swept down during winter and early spring. Far-sighted road planners had contemplated a great bridge across the Firth of Forth at the Queensferry Passage, near Edinburgh, in the early years of the last century, but it was not for many years that anything of the kind came to fruition.
The Firth of Tay at Dundee is wider than the Firth of Forth at Queensferry, yet it was the Tay which was first spanned by a giant bridge. In 1854, Thomas Bouch brought forward a plan for bridging both the firths. He was the Engineer and Traffic Manager of the Edinburgh, Perth and Dundee Railway, and his plans had been inspired by an accident to one of the ferries in 1849. His directors were astonished by his proposals, and even impressed, but they would not allow him to carry them out at the time. The idea was revived periodically, and several possible sites were chosen.
A final decision was reached in 1869. A bridge was to span the firth between Wormit, in Fife, and a point some way to the west of Dundee, the enterprising Thomas Bouch being charged with responsibility for the design. All kinds of craft pass up and down the firth at this point, so, in addition to being over two miles in length, the great viaduct had to have a considerable elevation above the fairway. At first this was to be not less than 100 feet, but the promoters of the first Tay Bridge successfully applied to the Board of Trade for leave to reduce this to a height of 88 feet above high-water level.
In due course Bouch prepared his designs. Preliminary borings showed the existence of an underlying stratum of rock, gravel and hard clay. But here occurred the first of a number of errors of judgment, errors that were fated, in due course, to culminate in disaster. The borings were not carried out on as extensive a scale as prudence demanded.
Bouch’s original design incorporated eighty-nine spans resting on piers of solid masonry and brickwork, each pier consisting of a twin cylindrical pillar. All went well until the builders had reached the fifteenth pier from the Wormit side. After that the underlying material showed a consistent deterioration, and was no longer capable of bearing the weight of solid piers such as Bouch had designed. Lighter piers, resting on a wider base, were necessary if the bridge was to remain immune from settlement.
Bouch adopted various methods for sinking these larger-based piers, finally using huge caissons. The light, upper parts of the piers, above high-water level, were built up of slender cast-iron columns. Such piers might well be able to bear the heavy superimposed weights of the superstructure and of trains passing over the bridge, but what their designer had not taken into account was the great pressure exerted on the side of the bridge by gales blowing at right angles to it. He had not overlooked wind pressure, but he had underestimated it.
At the same time as he made the alteration in the design of the piers, Bouch made modifications in the arrangement of the spans. The bridge, as finally built, had thirteen spans of 215 feet over the fairway instead of fourteen 200-feet spans. These central spans were so arranged that the rails were carried inside the girders, giving a clear headway below rail level, whereas on the other spans the rails were carried on top of the girders. This was a perfectly good arrangement in itself, and exists in the present bridge, but in combination with various weak points it was to prove one of the deciding factors in the ultimate failure of the work. Later, too, experts had something to say about the quality of the wrought iron and of the cast iron used in the piers.
The bridge was completed and opened on May 31, 1878. In due course Queen Victoria travelled over it and knighted its designer. The bridge was undeniably one of the wonders of the year, and excited a great deal of admiration. When we look at old pictures of it, however, we cannot help observing that the whole superstructure looked strangely slender for so great a work, and that the visible change over from the brick pillars to the light cast-iron columns after the fifteenth pier was not reassuring. The bridge was still young when a number of weaknesses became only too apparent. After the iron columns had been erected, they were filled with Portland cement concrete. As this set it swelled, causing certain of the columns to split and crack. These cracks were remedied with wrought-iron bands by the engineers in charge of the maintenance of the bridge. The cross-bracing between the cast-iron columns also tended to slacken, and there is no doubt that this cross-bracing was insufficient for the strains which it had to undergo from lateral wind pressure on the superstructure.
Lessons of a Disaster
The remainder of the bridge’s history is brief and tragic. On the evening of Sunday, December 28, 1879, a fierce storm was in progress. The wind was not continually at its maximum force, but it rose by short and rapid intervals to gusts having the power of a tropical hurricane.
Persons who braved the streets of Dundee and other Scottish towns on that terrible evening were often forced to cling to railings and lamp-posts to pre vent themselves from being swept away Slates and chimney pots came hurtling from the roofs, and all shipping had fled for shelter. Stark across the Tay stood the great bridge, with the wind whistling through the girders as it was buffeted by gusts estimated as ranging in speed from 90 to 100 miles an hour. In spite of the weather, there were fitful bursts of moonlight, and at first it seems surprising that so little was seen of what occurred. On that night, however, nobody who could help it was out of doors.
Most was seen from Wormit, on the south side of the Tay. The evening mail train from Burntisland to Dundee stopped there for the driver to pick up the train staff (the bridge carried a single track only) at 7.13 pm. Having restarted, the train ran on to the bridge and steamed slowly but steadily across. It was watched by the signalman, Barclay, and by a permanent way ganger named Watt. They reckoned that the lights of the carriages had reached the high girders over the fairway when they and the red tail-lamp disappeared abruptly. There was a flash of falling fire, a shower of sparks, and that was all. At the Dundee side, nothing was known at the time. The train failed to appear, and at first it was thought that the officials on the Fife side had refused to allow it to cross the bridge in such a gale. Attempts to call up Wormit were rewarded by complete silence, because of the breakage of the telegraph wires. This might have happened in any great storm, but additional evidence that something had befallen the bridge existed in clouds of spray visible in the occasional bursts of moonlight. These came from the Newport water main, which had been laid beside the rails on the viaduct on its way from Dundee to the Fife shore.
PLANT FOR MOVING GIRDERS from the old bridge to the new bridge, whose piers can be seen in the background. Mounted on large pontoons are telescopic lattice-girder supports. The pontoons, rising with the tide, lifted the spans of the old piers and, falling with the ebb, lowered the girders on to the new piers.
Two officials, with great courage, began to make their way along the bridge from Dundee, and although nearly blown into the river again and again, they were able to proceed far enough to establish beyond doubt there was a huge gap in the structure.
What had happened was this. When the train had passed into the high girders over the fairway, it and the girders offered a solid broadside resistance to the gale, which was blowing full on to them from west-south-west. The high spans consisted of continuous girders in three sets, one set of five spans and the others covering four spans each. The continuous girders were mounted on rollers, being fixed, in only one place, to a pier near the middle of each set. When the wind struck the train and the girders together, it tilted the girders over on their sides. The frail, cast-iron columns snapped, and the high girders, with the train and seventy-three persons, were hurled into the waters of the Tay.
The terrible accident produced a number of valuable lessons. The official findings, that the bridge was “badly designed, badly constructed and badly maintained”, sound damning enough, but they showed that not nearly enough allowance had been made for the greatest strain which a bridge can be called upon to bear, that of lateral wind pressure. After this nobody was likely to use cast-iron columns filled with Portland cement for the support of a giant viaduct in such a situation.
Settlement Tests
Great as had been the shock occasioned by the destruction of Bouch’s Tay Bridge, it was essential that the structure should be either rebuilt or replaced, and numerous proposals were brought up before the necessary Parliamentary Committee. One was to build a new bridge on the site of the old one, sinking new caissons beside the original ones and uniting each pair, the new and the old, with a brick arch, thus spreading the weight of the new superstructure over a wider area. On these duplicated caissons, the projectors of the new bridge intended to erect masonry piers supporting a superstructure of wrought iron. There were various disabilities. To begin with, Bouch’s original plan was being reverted to. Thus there was no clear evidence that the original caissons had not moved out of position in places, and it was doubtful whether the old caissons would be capable of bearing the much greater weight of the new masonry piers.
There were other considerations adverse to the adoption of this plan. It was known that some of the old caissons had not been sunk nearly far enough, and that considerable scour had taken place. Large quantities of rubble had to be dumped round their bases periodically to prevent them from becoming undermined and dangerous. The Committee urged the building of an entirely new bridge on different foundations, and the railway company commissioned W. H. Barlow to report on the feasibility of restoring the bridge connexion between Wormit and Dundee.
FOUNDATIONS FOR THE SUPERSTRUCTURE of the piers for the new Tay Bridge. Above each pair of cylinders a brickwork foundation was laid for the wrought-iron superstructure. Alongside are the remnants of the first Tay Bridge, destroyed in a gale.
W. H. Barlow and his son Crawford Barlow, who were destined to become the designers of the new Tay Bridge, first undertook a programme of elaborate soundings of the river bed, and experimented to determine the amount of scour sustained by the piers of the old structure. They carried out their borings at intervals of 500 feet along a line parallel to the centre line of the old structure. In the worst material they sank a trial cylinder, loaded it heavily and watched results. Their experiments showed that the building of the old bridge had caused the bed of the river to sink near the piers, especially where the current of the ebb tide was most powerful. The trial cylinder was sunk into silt to a depth of 20 feet, a concrete bottom was put in, and the cylinder was loaded until it exerted a gross weight of 7 tons to the square foot - twice the maximum anticipated for the new bridge.
The settlement amounted to 5¼-in. The weight was kept on for ten weeks, but no further settlement occurred. In addition to this, the Barlows loaded two of the old piers with 1,500 tons of old railway metals, thereby exerting a gross pressure of 3-3/5 tons a square foot. One of the piers stood in clean sand and settled only a quarter of an inch; the other was in micaceous sand and settled a little over 2 inches. In the course of their soundings the Barlows ascertained that on and near both shores the substratum consisted of rock. For a distance of 900 feet from the Fife shore there extended a bed of sandstone, the shore material consisting of whinstone. From the edge of the rock bed on either side extended a layer of hard clay for a further 900 feet, and between the two clay beds lay an area of sand, mixed in places with veins of clay over 70 feet thick.
The Barlows duly prepared their plans, which they deposited towards the end of 1880. Here they encountered a new difficulty. They wished to use what was left of the old bridge, all of it, that is to say, except the high girders which had fallen, as a staging for the building of the new viaduct. The Board of Trade required the old bridge to be removed completely before any work was carried out on the new one.
After much wrangling the old piers were allowed to remain in position so long as they were cut down to high-water level and were provided with visible warnings for passing shipping. The old piers may be seen to this day, grim reminders of the bridge that failed, 60 feet down-stream from the present bridge.
In the matter of the layout, the Barlows were faced with the same problem as Bouch had been. The approach from Fife is at a considerable elevation, the shore being bounded by high cliffs. On the Dundee side they had to bring the final spans down to the level of the existing railway, which was only a few feet above high-water level, and also ran at right angles to the centre-line of the bridge. For these reasons both bridges were shaped in the manner of a letter J with the tail cut short, coming into Dundee on a bold curve with a downward gradient.
Irregular Spans
The total length of the present bridge is 10,711 feet, or 2 miles 50 yards 1 foot, the length of the straight section accounting for 1 mile 1,038 yards 2 feet. The structure begins at the south side with a brick approach viaduct consisting of four 50-feet spans known as the Wormit Arches. The bridge proper consists of eighty-five piers supporting girder spans. The height at the southern abutment is 83 feet, and this is maintained over a level section as far as the fourth pier. Thence to Pier 28 the deck of the bridge follows a down grade of 1 in 762, bringing it down to a height of 79 feet. The section from Pier 28 to Pier 32 is level again; after that the bridge follows a continuous down grade on an incline of 1 in 113¾, the height at the northern abutment being only 26 ft 6-in above high-water level. The curve at the northern end has a radius of 21 chains, and bears eastwards until the line of the bridge is nearly parallel to the river.
The spans which lie over the water vary considerably in length, and in this respect the bridge is identical with its predecessor, certain of the old girders having been used over again, though considerably reinforced by new girders. First comes a single span of 118 feet, then ten of 129 feet each, then thirteen of 145 feet, eleven of 245 feet, two of 227 feet, one of 162 feet, eleven of 129 feet, twenty-four of 71 feet and one of 56 feet. These measurements are taken from centre to centre of the piers. The landward spans, at the end remote from the Wormit Arches, are the Esplanade spans, which connect the main part of the bridge with the original brick arching leading downhill into the railway yard. They comprise two wrought-iron skew arches, four 66-feet girder spans, a curved-top girder span of 108 feet and a short brick arch connected with the old arching.
GIRDERS FOR THE CENTRAL SPANS of 245 feet were assembled on a staging at the southern end of the bridge, the various members having been made and drilled in the contractors’ works at Glasgow. Eleven spans were thus assembled, each weighing about 514 tons.
The engineers responsible for the Tay Bridge adopted the following method in sinking the cylinders for the piers. They used four pontoons, each pontoon consisting of five watertight tanks. Two of the tanks were long and the remaining three were short, being sandwiched in between the longer tanks. The completed pontoon thus resembled a letter H with the top and bottom openings filled in. Through the spaces in the middle, the cylinders were sunk, the pontoon being made secure meanwhile by legs passed down through its corners to the bed of the firth.
Men on shore riveted the cylinders together in suitable lengths, and then boats carried the cylinders out to the pontoons. As each length was lowered through its respective well in the pontoon, it was bolted to its predecessor through an internal flange. Their upper parts were then lined with brickwork, and they were slowly lowered by hydraulic jacks until they rested on the bed of the river.
For excavating the silty sand in the cylinders, the engineers arranged two 6-in hose-pipes in the bottom of each cylinder.
Twenty feet from their open ends, these hoses united to form a single 12-in pipe, the upper end of which connected with a powerful centrifugal pump on the pontoon. Divers manipulated the pipes so that one sucked up sand while the other took in only clear water, thus preventing the choking of the pump by diluting the flow through it. This arrangement was capable of removing as much as 40 cubic yards in one hour, the consequent subsidence of the cylinder being 2 feet.
When each cylinder had reached the required depth, it was cleaned out down to the cutting edge, and then concreting took place. Having been completed as far as low-water level, the cylinders were tested. Each was subjected to a weight one-third more than it would be expected to bear in eventual conditions, assuming that each of the two lines of railway was completely occupied by locomotives. The heaviest of the company’s locomotives at that time weighed in working order about 78 tons each. Thus the test allowed a wide margin of safety for the considerable increase in the weight of railway rolling stock which has since taken place. The test weight on each of the largest piers was as much as 2,438 tons.
Provision for Climatic Extremes
After the testing of the cylinders, the builders attached temporary caissons of wrought iron to them, before completing the masonry work and building the foundations of the superstructure. Each pair of cylinders was joined above high-water level by a horizontal connecting member; above this the pier rose in the form of two octagonal pillars of iron, hollow, and united at the top by a semicircular arch of the same material. The design varied slightly in different places, and the relatively short piers at the northern end were of simpler construction altogether, the octagonal pillars and the arches being dispensed with and a simple rectangular structure being superimposed on the iron cylinders. Except where the piers rested on hard rock, the foundations were carried down to a depth of at least 20 feet lower than the lowest part of the bottom of the firth in the vicinity. This acted as a precaution against the scouring action to which the piers of Bouch’s bridge had always been subject.
Barlow and his son had to make special provision for expansion and contraction of the structure in climatic extremes. Scotland is a country of changeable weather, and its eastern counties have known the coldest weather ever experienced in Great Britain. At intervals of approximately 500 feet, the ends of the girders were supported on rocker bearings at one end of each length, the other end being fixed to the piers. An observation made over a period of twelve months showed that the variation in a length of 516 feet during a range of temperature of 55 degrees Fahrenheit amounted to 1·65-in.
WROUGHT-IRON SUPERSTRUCTURE of the piers for the second Tay Bridge. The piers rest on sunken cylinders, each pair of which is connected above high-water level by a horizontal member encased in brick.
Several distinct operations attended the erection of the girders on the piers. The first was concerned with the removal of sound and undamaged girders from the old bridge to the new. For this process the engineers prepared large pontoons, on which were mounted telescopic lattice-girder supports, which could be adjusted to any required height. They floated their pontoons under the old girders at low tide, and allowed them to rise with the tide beneath the girders, the telescopic members being pushed up and made secure underneath the section to be moved. This gradually rose clear of its bedplates and thus floated with the pontoon. The pontoon was then manoeuvred upstream until it was between two piers of the new bridge, and secured again, the girder being carefully lowered into position on the new piers. The tops of the new piers were slightly lower than those on the old bridge.
This process was not feasible with the final spans at the northern end because of the curve. Here the builders transferred each girder in turn by steam cranes.
Meanwhile erection of the new girders for the northern and southern sections was taking place in the contractor’s yards adjacent to the bridge, preliminary work having been undertaken at Glasgow. Rails had been laid on the tops of the old girders, which were by now all in their new position, and the new girders were run out in series on traversers running on the rails, and duly lowered into position between the old girders. The flooring of the bridge was built up of corrugated steel, which eventually supported an ordinary ballasted road bed bearing a double line of standard cross-sleepered track. [The old bridge had carried a single track only.]
With the high spans of the central section, the builders had to adopt a totally different series of operations. In the old bridge it was this section that had been destroyed, and there was no question of using old girders over again. Further, as the tracks run inside the girders over the central section, there would have to be only two great girders to each span with a clear space between them. The various members of the girders were made and drilled ready in the works at Glasgow, and the girders were then erected complete from these parts on the south side of the firth; the only parts left unfinished were a small section of the connecting members at each end. On completion, each girder span, weighing approximately 514 tons, was lowered on to a pontoon arid towed out to its future site by four steam tugs. Here the engineers lowered the span on to beech blocks on the piers, the pontoon being meanwhile securely anchored to the old and to the new piers. The pontoon was floated away while the tide was on the ebb, and was generally free of the span some two and a half hours after high tide.
Completed in Five Years
The next step consisted of erecting the superstructures of the piers, on which the spans were to rest, and only when these had reached their full height did the process of raising the spans themselves begin. To effect this, the engineers mounted a steam boiler and a set of powerful pumps on the middle of the span, the pumps being connected to a set of four hydraulic rams, two at either end. A system of stopcocks enabled these pumps to operate alternately on either pair of rams. Thus, stage by stage, the great central spans rose to a level above their final position, when the engineers inserted the connecting cross-girders, rockers and bearings beneath them. Finally each span was lowered gently into place and secured to the fixed bearings or adjusted to the rocker bearings as required.
The building of the Tay Bridge accounted for 13,452 tons of wrought ironwork in the parapets and girders, 7,626 tons of wrought iron in the cylinders and superstructures of the piers, 2,705 tons of cast iron in the cylinders, 3,588 tons of steel in the flooring, 37,024 cubic yards of concrete, and 26,419 cubic yards of brickwork.
Nearly five years were occupied in construction. Operations on the ironwork began in the Glasgow yards on June 22, 1882. On July 6, 1883, the engineers on the site began operations on the sinking of the cylinder foundations in the bed of the firth. Three hundred men were employed in the Glasgow yards, and the number at work on the site of the bridge was at times as much as 900.
On June 12, 1887, goods trains began to cross the bridge. On June 18, Colonel Rich and Major-General Hutchinson inspected the bridge for the Board of Trade. For the Board of Trade test, the railway company supplied sixteen locomotives, having a total weight of 955 tons. These were placed in groups of eight, one on either track, and ran backwards and forwards over the bridge. The slight movements of the structure under the weight were observed, first by levels and later by stretched steel wires. The only permanent set observed was a deflection of one-hundredth of a foot on the 245-feet span between Piers 30 and 31, and this took place on the west side only.
So the new Tay Bridge was passed, with honours, and the triumph of the Barlows and of William Arrol, the contractor, with all their assistant engineers and workmen, was complete. The bridge had cost £670,000, or about £62 10s. a foot. It was opened for public passenger traffic on June 20,1887. Since then it has given irreproachable service.
TOWING ONE OF THE 245-FEET SPANS on pontoons from the assembling staging to its position between the piers in midstream. Hydraulic rams raised the ends of the span alternately as the piers were being built, until the span had reached the desired height.