The contribution made by the various types of nineteenth-century brick and masonry finishes to the visual experience of British railway travel can be appreciated by taking a journey on the Channel Tunnel Rail Link. Here the carriage-window view is repeatedly of surfaces of concrete, especially the finely textured, fair-faced concrete of modern civil engineering, in whose smooth grey surfaces no weed or shrub can find lodging. The same finishes continue through the tunnel and on through the Continental high-speed rail network. It is impressive, if rather boring. On older lines, new technology manifests itself in the form of geosynthetics: ugly black long-life membranes and netting pinned down by steel fixtures, guarding the permanent way against collapses from the cutting sides.
By directing the route to avoid the steeper slopes, a railway company could reduce its outlay on earthworks and tunnels, but at the cost of other drawbacks in the longer term. For one thing, the finished line would be slower to operate (because curves cannot be taken at high speeds and because the line itself became longer), without necessarily being cheaper to maintain, at least to a high standard. An extreme instance is the slow horseshoe curve that takes the West Highland Railway in a loop around Ben Odhar, Ben a Chaistel and Ben Doran, thus avoiding the need for a monster viaduct. The curve is a highlight for those travelling the line for its scenic pleasures, but for Highlanders heading for the shops of Helensburgh or Glasgow it is also a cause of much cumulative delay. A different approach is represented by Brunel’s Cornish lines, with their abundant use of viaducts for valley crossings. These may look extravagant, but they came about because the selection of inexpensive routes had kept cuttings to a minimum, which left too little spoil to build grand embankments instead.
It is often possible to distinguish the different approaches to railway building on the sheet of an Ordnance Survey Landranger map. Many lines appear relatively straight, with an irregular border along much of their length composed of the little triangular marks that indicate a cutting (triangles pointing inward) or an embankment (triangles pointing outward). There is a good chance that a line of this type will belong to one of the routes built before the 1860s, when companies were highly capitalised and locomotives less equal to climbing steep gradients. On the London & Birmingham, for example, the curves were rarely of less than a mile in radius. For later branch lines and secondary routes, constructed more cheaply and without fast travel in mind, a gently meandering course was considered acceptable. Now that many such lines have closed, their presence on the modern-day map dwindles with each new edition, having left relatively little in terms of great markings on the earth’s surface. In agrarian regions especially, much land has been taken quietly back by the agricultural communities which the railways once served, and which forsook them in the twentieth century in favour of transport by road.
When a deep valley intervened between higher ground, as in hilly Cornwall, there was often no alternative but to throw a multiple-arched viaduct across the gap. There was a precedent in the aqueducts employed by the second wave of canal builders, men such as John Rennie and Thomas Telford, from the 1790s onwards. So when George Stephenson provided viaducts for the Liverpool & Manchester Railway in 1828–30, at Sankey and at Newton-le-Willows, he was merely repeating a form of bridge construction that already existed elsewhere in Britain (the distinction between a bridge and a viaduct, it should be said, is inexact).
The Sankey Viaduct – of nine arches and rising to 60ft – was grand enough to gobble up almost half of Stephenson’s budget for bridges along the line. Even so, it was a midget by comparison with the great viaducts that began taking shape in the 1840s. At Welwyn in Hertfordshire, at Balcombe in Sussex, at Stockport, the new trunk lines soared across on multiple-arched structures that would have stupefied any emperor of the Romans. Welwyn has forty arches each up to 98ft high; Balcombe comprises slightly fewer arches and is slightly less lofty, but has extra grace-notes in the form of a cornice, balustrade and little pavilions at each end; Stockport has twenty-two main arches only, but each spans fully 63ft and the overall height is 111ft. Viaducts of this type – round-arched, with uniform right-angled spans, built of brick or stone – continued to appear until the end of the century and few have had to be replaced. The passenger of today looks out over a different landscape, but shares with his counterpart of 1850 or 1900 the experience of lofty isolation as the ground suddenly and marvellously drops away from alongside the moving train.
This resilience is all the more amazing when the vastly increased speed, frequency and weight of the trains that cross these structures are taken into account. Railway engineers knew that their viaducts and bridges would have to stand firm against years of vibrations from fast, heavy loads, unlike bridges built to carry roads or canals, and made their calculations on the safe side. That they so consistently got these right is one of the marvels of railway history. It is almost as if the turnpike roads of Telford’s day had proved equal to the thundering traffic of the modern motorway system.
Repeated arches of standard width were well suited to valleys and plains comprising mostly solid ground. Other situations required special measures. The best-known of these stone- or brick-built viaducts is the beautiful double-arched bridge of 1837–9 which carries the Great Western Railway across the Thames at Maidenhead. This was another feather in Brunel’s cap. The Thames Commissioners insisted on sufficient clearance for sailing barges, and it would not have been difficult to provide for this by means of a slight gradient on either side. But Brunel had his own ideal to defend, of an exceptionally level railway suited to fast and economical running. So he pushed the technology of brick construction as far as he dared, to a ratio of 5.28:1 between the breadth of each elliptical span and the rise of the arch. No builder in brick anywhere in the world has achieved a flatter arch since. Nor are the spans small: each measures 128ft. The audacious design looked suddenly like a dreadful mistake when one of the arches shifted after its temporary wooden centering was taken away. Partly set mortar was blamed and Brunel ordered repairs. Sceptics waited for a fresh disaster, but it never came, and express trains use the bridge to this day. Aptly, too, it is Maidenhead Bridge that Turner portrayed in his canvas ‘Rain, Steam and Speed – the Great Western Railway’ (1844), one self-fashioned genius paying homage to another.
Even greater spans of brick and stone were possible. The British maximum was reached as early as 1848, again in the form of a bridge that still stands and still carries trains. It is Glasgow, Paisley, Kilmarnock & Ayr Railway’s Ballochmyle Viaduct near Mauchline in Ayrshire, which soars over the flowing Water of Ayr in a single stone arch 181ft wide and 175ft high. Its designer was John Miller (1805–83), who served as engineer to most of the principal lines of central Scotland, making enough money along the way to retire early and buy himself two country estates. Yet no one would call Miller famous, nor is his greatest bridge nearly as well known as it deserves. For posthumous renown, it helps to have a memorable name, and perhaps also to have worked closer to London.
Scotland also witnessed a late flourish of the British arched viaduct. These were the concrete structures built around 1900 for the West Highland Railway’s Mallaig extension by the contractor Robert McAlpine (‘Concrete Bob’), founder of the construction firm of that name. McAlpine’s viaducts were not of the reinforced type with ferrous rods or bars running through, still less of the even stronger pre-stressed or post-tensioned forms. Rather, they were simple mass-concrete castings of the technique the Romans knew (though the Roman world never attempted to make bridges this way). The material was chosen for its cheapness: McAlpine claimed a saving of 10–30 per cent over masonry construction, once the intractable hardness of the local Highland stones was allowed for. By adopting naturally strong arched profiles, McAlpine’s bridges have endured, and some are magnificent: the Glenfinnan Viaduct has twenty-one arches, built on a steady curve; the Borrodale Bridge has a main span 127ft 6in wide. Had large-scale railway building continued across the wilder parts of Britain, there would
surely have been more concrete bridges of this type.
Not every prodigious viaduct was of heroic height or span. In towns and cities, where land costs were high and many existing rights of way had to be negotiated, it was often easier to lift up the line on a low viaduct rather than taking it along ground level, or attempting an embankment. This type of urban railway was perfected very early, in the form of the London & Greenwich Railway’s route of 1833–6 between London Bridge and Deptford. Here was a single brick-built structure almost four miles long, made up of a mind-numbing 828 arches and with a pedestrian tollpath along one side. The path did not last, and now that the line has been repeatedly widened it is difficult to sense its original appearance from a moving train. The Bermondsey pedestrian may do better: here and there, where a road passes beneath the line, massive Grecian colonnades of cast iron still separate the roadway from the paved walkways on either side, an echo from the last days of Georgian London.
Another achievement of the early railway bridge-builders, a source of wonder in its time, was the skew bridge. For simplicity’s sake, most arched bridges are set ‘square’, that is at right angles to their springing points, like the arched head of a doorway in a straight wall. In other cases, the road, river or other obstacle takes an oblique course and the railway crosses at an angle by means of brick or stone coursing that follows the line of the skew. The form was not invented by the railways, having been achieved by at least one canal builder as far back as the late eighteenth century, but it was hugely multiplied for the new transport network. This in turn speeded up the development of the type.
The railways’ earliest skew bridges were of stone and required each course to be cut – expensively – to unique dimensions. Taking no chances, the skew bridge at Rainhill on the Liverpool & Manchester was tested first by means of a timber mock-up erected in a field near the line. In its completed form, the arch of this red sandstone bridge vaults across the railway at a sharp thirty-four degrees. Later in the 1830s, a young engineer named Charles Fox (later a key man in the construction of Euston station and the Crystal Palace) developed a cheaper system, by which the arch courses could be formed of standard-sized bricks. Both methods use slanted courses that spring from the horizontal abutment of the bridge at an angle. When viewed from below, this conveys an eloquent sense of movement, as though of great forces held in check. As skew bridges tend to be elliptical rather than round-arched, the summit of the arch is often relatively close overhead and the experience is intensified. If these spaces belonged to the National Trust, the guidebooks would instruct visitors to pause and marvel; being public, numerous and grubby, they generally pass unregarded.
Skew arches were often built at impressively sharp angles, as at the London, Brighton & South Coast Railway’s viaduct at Westhumble, Surrey, opened in 1867
A skew bridge was not essential every time a railway crossed a road obliquely. As every driver knows, a minor road that approaches a railway line at an angle will often swerve to the perpendicular at the last minute, before bending back to its original course on the far side. Sometimes the road passes under the line, sometimes over (which gives better sightlines). The road has not always had this kink in it; its old alignment has been tweaked to make the bridge builder’s task easier. Buying the extra land required for the small diversion worked out cheaper for the company than making a skew bridge so that the road could keep its straight line.
Most railway bridges are of this simple type, with a single perpendicular span. Huge numbers remain, which makes it easy to forget that bridges were a relatively rare sight before the railways came. Older bridges were almost all bespoke structures, and most were maintained at public expense. Only when the canals arrived were bridges multiplied to standard engineers’ designs, and even these fell well short of the railways’ needs for crossings over and under. In national terms, it took railways to make bridges truly commonplace.
Other bridges used a material not commonly associated with railways: timber. This amounted to a revival, for the use of timber for bridges carrying heavy loads had effectively been given up on the roads. Timber was abundant and it was cheap. There was a good case for its use as a stop-gap on a new line, to be replaced in more lasting materials when funds permitted. Confronted with a terrain of repeated flat crossings of waterways and broad ditches, several fenland lines adopted this policy. Unfortunately, money was not always available to upgrade their crossings when required. The Board of Trade’s inspector in 1855 reprimanded the Eastern Counties Railway for the state of its viaduct across Sir William Beauchamp’s Navigation: ‘in many parts these timbers could be dug out with a spade like garden mould’. The railway’s own engineer resigned two years later, protesting that a safe speed limit of 2 mph should be imposed on one of the worst stretches.
By the standards of Brunel, this was shoddy stuff. His own exploration of timber crossings started in 1839 with a couple on the Great Western main line. One was a skew bridge made of laminated timber arches, conspicuously placed in the middle of Bath. The other was built of ordinary timbers and carried a road across the cutting at Sonning. It had trestle-like uprights, with struts splaying out from the top. Drawings make it look flimsy, but it introduced the basic principle used by Brunel for dozens more bridges and viaducts. By working out a set of standard pinewood sizes and a variety of span types, he was able to make robust timber viaducts of appealing cheapness. Some were built with stone piers for the uprights, others stayed true to the Sonning-style trestle type. From the 1840s to the 1860s these viaducts spread across South Wales and down through Somerset, Devon and Cornwall. Thirty-four were erected during the 1850s for the Cornwall Railway alone, together carrying the route for about one-tenth of its total length of fifty-three miles. Designs were made on the spare-parts principle, so that worn-out timbers could be replaced piecemeal without having to renew the whole. They could be burdensome to maintain: for years after its opening, the Cornwall Railway employed nine teams of ten men each, working to replace strut after strut as they became unsound.
Brunel’s viaduct at Ivybridge on the South Devon Railway, in a lithograph of 1848. Here the spans alone are of timber, supported on stone piers
These bridges must have been nerve-wracking to cross. The Landore Viaduct near Swansea (thirty-seven timber spans on stone and timber piers), an extreme case, looks in illustrations like an enormously long, straightened-out fairground ride. Trains could only proceed across them at a slow pace. A few people today may even remember the experience, as the last timber viaduct on a public line was replaced as late as 1934; it was on the Falmouth branch. Even at this date its successor was made of masonry, laid course by course, as if it were 1834 (or 1134, for that matter), not 1934. That the timber viaducts had a share in prolonging an older, obsolete technology in this way is a pretty paradox.
In Wales it is still possible to take a long wooden ride by train. The bridge in question is the Barmouth Viaduct that carries the Cambrian Railways’ line across the half-mile Mawddach estuary. Opened in 1867, it is constructed for most of its course from timber piles and beams, like a gigantic jetty. The line may be the only one in Britain to have survived a closure threat arising from the burrowing action of Teredo navalis, or shipworm. That was in 1980, since when the bridge piers have been sheathed in inedible concrete.
Some of Brunel’s timber spans were replaced by girders of iron or steel. These had a long ancestry. The potential of iron had been demonstrated most famously at Ironbridge in Shropshire in the 1770s – the first in the world – and the material was taken up before long for plateway bridges and aqueducts. The first iron railway bridge, the Gaunless Bridge of 1825 on the Stockton & Darlington designed by George Stephenson, spaced its trusses just 12ft apart. It was a spindly affair, the ferrous equivalent of a pioneer biplane. The reassembled metalwork now stands outside the National Railway Museum at York, where it looks enigmatic without the heavy stone abutments of the original structure.
The Ironbridge crossing was of cast ir
on, but the Gaunless Bridge included some wrought-iron members too, in recognition of its different properties. Cast iron, having a large proportion of carbon in its make-up, was strong in compression (good for piers and columns), but weak in tension (unsuitable for struts and ties). Wrought iron, like tools hammered out by a blacksmith, was much more resistant to tension and stress. The further development of iron bridges was steered by these constraints.
Arched bridges of cast iron were built for the railways too. Britain’s broadest-spanning example is the 200ft-wide Victoria Bridge at Arley in Worcestershire, constructed for the Severn Valley Railway in 1862 and still in use as part of the preserved line of that name. Its ironwork was supplied by the same Coalbrookdale foundry which had made the mother of all cast-iron bridges, at Ironbridge. By the time the Victoria Bridge opened, another type of cast-iron bridge, using large girders of flat span, had already been and gone. Its drawback was fragility. Robert Stephenson’s record was blotted by the failure of one such bridge, across the River Dee at Chester, one evening in 1847. It had three skewed spans, each just short of 100ft and each composed in turn of three castings bolted together. The structure was just six months old when one of the spans failed as a train was crossing it. Five deaths resulted, including the fireman; the driver survived only because he began to accelerate flat out as soon as he sensed something amiss. The investigation found that repeated flexing had weakened a girder, which had fractured in two places; also that the wrought-iron re inforcement rods that were fixed to the cast beams were structurally worse than useless. Witnesses described how they had seen the spans deflect by several inches every time a train crossed.
The Railways Page 45