The Atmospheric Railway

By Angela Marks

The 'atmospheric railway' was an experiment in locomotion on the South Devon Railway between Exeter and Newton Abbot in the early 1840s which ended in failure.

At the beginning of the railway era, individual railway companies undertook to construct lines between two specified points. Each company required an Act of Parliament to be passed, which allowed for a railway to be constructed between those points, with provision for branch lines to be built as specified. In the West of England, the first enterprise was the construction of the Great Western Railway between London and Bristol. The great Western Railway Act was passed in 1835, and construction began almost immediately Two years previously, the newly formed Great Western Railway Company had appointed the 27 year old Isambard Kingdom Brunel to be the principal engineer. The route had already been surveyed, and within a few weeks of the Act being passed, construction began. The new railway was completed in 1841.

It differed from other railways in the width of its track - the broad gauge, 7 ft wide, as opposed to the standard gauge of 4ft 8¼" used by the other railway companies. This was the width of track used by the earlier horse-drawn colliery railways, it was said that it came about after George Stephenson, the engineer of the very first public railway, the Stockton & Darlington took a number of farm carts, measured them and took their average width. Brunel chose to use broad gauge as he considered it to be safer and more comfortable. Trains rarely left the track after a collision or derailment, as was usual with standard gauge, and the width allowed for larger coaches and a smoother ride.

By the time the GWR opened in 1841, plans had already been laid for extensions to take the railway deep into the South West. The Bristol and Exeter Railway, South Devon Railway and Plymouth and Cornwall Railway Companies all gained the required Act of Parliament during the late 1830s and early 1840s, although other companies had previously existed to plan the routes. The first train to Exeter ran in May 1844.

With the completion of the line to Bristol, Brunel rather lost his enthusiasm for building railways. With Bristol being one of the main ports for America, he now had bigger fish to fry than building a railway across the flatlands of the vale of Taunton to a sleepy Devonshire cathedral city. He had set his sights on reaching New York, and to do that he had to build a steamship to connect with his railway. So he set about doing just that, and concentrated on building the SS Great Western, leaving others to build the Bristol & Exeter. However, once the railway had reached Exeter, a new prospect opened up, and once again, Brunel's imagination was fired. He had been appointed Surveyor to what was then the Exeter, Plymouth & Devonport Railway Company in 1836. Plans had already been drawn up for a railway between Exeter and Plymouth and the idea of running the track along the sea-front wasn't even considered. Brunel suggested two routes, both of which were the same as far as Newton Abbot, running behind Starcross and keeping inland behind Dawlish and Teignmouth. These plans were dropped because of opposition from landowners, mainly the Earl of Devon and the company was dissolved in 1837. Eventually the South Devon Railway Company won the contract, and as the Earl of Devon's son, Lord Courtney was a director of the company and a keen supporter, the Earl changed his mind about the railway running across his land. Brunel decided to run the railway down the river and alongside the estuary to Dawlish Warren (then called Exmouth Warren) and beside the sea to Teignmouth. When Brunel arrived in Dawlish, the locals became alarmed at the prospect of the railway cutting the town off from the sea and spoiling its amenities. So when he presented the proposal to Parliament in 1844, Brunel fudged many of the plans for dealing with the many headlands and inlets between the Warren and Teignmouth. For instance, Langstone Point was a headland with grazing land above it. Brunel told Parliament he was going to tunnel through it, although he always intended it to be the cutting it became. Blasting the rocks to form this cutting made the Langstone into the isolated rock it is today.

The original plan was to build the line even nearer the sea than the one actually built. It was moved inshore slightly, to allow some beach to remain and to preserve Boat Cove, Coryton Cove and Shell Cove. In view of the disruption caused by heavy seas now, one wonders what it would have been like if Brunel had followed his original plan. In fact these problems with the sea are nothing new. The seawall was frequently breached, notably, and nearly disastrously, at Smugglers Cove in 1855.

The South Devon Railway Act received the Royal Assent on 4 July 1844 and track laying began in the same month. The line started at Exeter Depot, then a terminus, with the B & E finishing at the bank of the Exe, so the first section had to be a bridge across the river, still called the South Devon Bridge. Exeter Depot was renamed St David's Station after the opening of the London & South Western Railway's Queen Street (now Central) Station.

Work on the South Devon bridge did not begin until 1845, so there was no rail link to supply the workers further down the line. A second bridge was built alongside Brunel's original bridge in 1861 and both were replaced in 1896. This in turn was replaced in 1997. While the line started at St David's, this was still a Bristol & Exeter Railway station. The first official South Devon Railway station was St Thomas's, and the company's offices were built here in a splendid Italianate building, which still survives, although now converted into a restaurant. Alas, the station itself doesn't survive. It was a magnificent structure, with a glass roof over the track, unusually set into a wooden framework. It was pulled down some years ago and replaced with the present bus shelters.

During the process of getting the S.D.R Act through Parliament, local landowners had the opportunity to lodge objections. There were the usual coach operators, and people living near the proposed line, but one early NIMBY was particularly unfortunate.

James Powell, a businessman from the Midlands, had moved to Dawlish because his previous property had been blighted by the building of the London-Birmingham railway, which passed through the grounds of his house. He set out to buy a house in a location where he felt certain this could never happen again. Unfortunately he chose Sea Lawn House, a charming property almost on the seashore near the Dawlish Coastguard Station. The grounds of the house extended from Coastguards to the Rockstone, and the railway ran, not through his grounds, but right alongside his house. As part of the sea defences for the railway, the S.D.R built a sea wall, along which people could promenade, and Mr Powell gained one concession: the level of the sea wall was dropped outside his property so that people walking along it could not look straight through his windows. In 1846 Mr Powell was awarded £8,000 compensation, but died before the end of the year. However, his widow seems to have forgiven the railway company, for she is recorded as having given a reception for railway staff at the house the following year. Sea Lawn House was demolished in 1888 and Sea Lawn Terrace was built on the site. This gap in the sea wall remains to this day.

The original intention had been for the trains to be pulled by locomotives, but Brunel was worried by the gradients beyond Newton Abbot, and after a visit to an atmospheric railway in Ireland, he decided this was the solution. The Directors of the S.D.R were sceptical to begin with. Brunel managed to persuade them, but to save money, the originally planned double line was replaced with a single track.

The atmospheric railway was so-called because it uses air at atmosphere pressure as opposed to compressed air. There was an alternative pressurised air railway system, which was used by Royal Mail to send parcels around London underground. In the atmospheric system, the train was pushed along by pumping air out of a long pipe to create a vacuum, and the inrush of air when the train passed over propelled it forward. The steam-powered pumps which sucked the air out of the pipe were contained in eight engine houses at Exeter, Countess Wear, Turf, Starcross, Dawlish, Teignmouth, Bishopsteignton and Newton Abbot. The railway was originally intended to extend further west, and although pumping stations were built further down the line (the one at Torquay survives) the atmospheric railway was finished before the line could be completed beyond Newton Abbot.

Brunel designed the buildings to house the great pumping engines driven by coal-fired boilers, and their chimneys were hidden in towers with ornate Italianate cowls. The only surviving engine house between Exeter and Newton Abbot is the one at Starcross, which has a truncated chimney, having lost its splendid topping.

Each length of pipe began and ended at the pumping station, and here there were valves which kept the pipe airtight. The train could pass from one pipe to the next without losing speed. The vacuum inside the pipe was maintained and released by means of a leather valve, but this vacuum also leached the oil out of the leather, and greasers were employed to keep the valve oiled. Water getting into the pipe also caused problems, and a piston carriage would run along first thing in the morning to clear any water out of the pipe. The composition used to seal the leather to the pipe dried out, and gangs had to follow trains to re-apply it.

The summer of 1844 saw the beginning of the construction of the five tunnels (named Kennaway, Coryton, Phillots, Clerk and Parson) between Dawlish and Teignmouth. Much of this was done by blowing up the cliffs, and these explosions, widely advertised for safety reasons became a form of public entertainment. Brunel himself was supposed to be present for one of these blow ups in October 1844, but was two hours late arriving. His carriage finally arrived at the London Hotel, Dawlish and he walked up the Teignmouth Road, followed by a crowd of onlookers and children. Four charges were set, but the last failed to go off.

The installation of the atmospheric pipes took longer than anticipated, and the first trains were pulled by steam locomotives. The tunnels had been completed in 1845, but when the first test run from Exeter was made in December of that year, it could only go as far as Cockwood, because the viaduct wasn't ready. It had been planned to lay the track on a walled embankment, but the mud was so deep it was impossible to find footings for the wall, and the problem was resolved with a timber viaduct. The viaduct was replaced by the present embankment in 1896/8

With the track completed, the atmospheric service began in September 1847 from Exeter to Teignmouth, and was extended to Newton Abbot in January 1848. The railway in its early days was immortalized in a series of watercolours by W Dawson of Exeter, painted in 1848, and now in the collection of the Institution of Civil Engineers.

No-one really knows what the driving car looked like, as Brunel's plans have been lost. It is sometimes shown as a flatbed truck, probably because drawings of the track survive which just show the base of the train. In some of the Dawson prints, it appears that the driver stood up on a platform at the front of the train, with a little sentry box on one side, where he could sit down. Rather like a modern aircraft pilot, the driver didn't have much to do except when the train was arriving or departing from a station, or something went wrong. As the main form of public transport up until then had been a mail coach, where the coachman sat on the box, exposed to the weather, having the driver exposed to the elements wasn't anything unusual. However, driving a coach at about 12mph and a train at 60mph was a different matter!

The atmospheric railway has been regarded as something of a joke, but in fact it was more successful than it is often given credit for. With scarcely any friction with the track, the ride was much smoother than with locomotive hauled trains, and the trains usually travelled at about 60mph, a phenomenal speed for the 1840s. There were, however, problems with the system. If the valve leaked or became water-logged, there could be a lot of stopping and starting, and speeds could be greatly reduced. There was no room for manoeuvre if the train stopped short of the platform the passengers had to get out and push it, if it overshot (the piston car could not reverse) they had to get out and traipse back along the line. When the leather valve became wet, it could freeze in a hard frost, and as the winter of 1847/8 was a cold one, the valves froze several times, disrupting trains. The location didn't help either: the salt air made the problem of the leather drying out worse.

It's often been said that the atmospheric caper, as it became known in Devon, failed because rats ate the leather. This is nonsense (although rats might have had the odd nibble, compared to the other problems, this one was negligible). The scheme failed because it was so expensive - a lot of coal was burned unnecessarily, as the engines often pumped the air out long before the train actually arrived in their section, and then had to keep pumping.

Brunel believed that a properly installed telegraph between the pumping houses would reduce fuel consumption, but before that could happen, the shareholders decided that the system did not make a sufficient profit. With a little more patience, the problems could have been overcome, but in 1848 the Directors decided to switch to locomotive working.

The up goods to Exeter on the night of Saturday 9 September 1848 was the last atmospheric train to run. That night, the boilers in the engine houses shut down for the last time. The engine house at Dawlish became a coal store and was demolished in 1873. The one in Exeter was converted into a water tower, and survived into the 20th century. Apart from Starcross, the others vanished without trace.

The broad gauge, too, eventually disappeared. It caused too many problems when travelling round the country, as passengers had to change trains when they used another railway company's line. This problem was slightly mitigated by installing a third rail, thus enabling both standard and broad gauge trains to use the track. The last broad gauge train travelled from Paddington to Penzance in 1892.

The track between Exeter and Dawlish, and Teignmouth and Newton Abbot was doubled in 1873, but because of the problems of widening the five tunnels between Dawlish and Teignmouth, this stretch of the line wasn't doubled until 1903.

In 1920, because of the frequent cliff falls between the Clerk and Parson tunnels, the latter was lengthened by an external extension. Trains appear to be going under the cliff face, but for part of the way, they pass through what is effectively a rectangular box attached to the outside of the cliff. This extension, of brick and concrete with a flat roof is only visible from the sea.

Shortly after the 1997 South Devon bridge at Exeter was completed, the driver of the first train across it reported, ashen-faced, that he had almost run over someone on the bridge, who had loomed out of the morning mist. As he appeared to be carrying a set of plans, and inspecting the bridge, the driver demanded to know why a surveyor was wandering on the track without the proper safety precautions. Not to mention the fact he wasn't properly equipped, being strangely dressed in a dark suit, with a tall top hat. The 'surveyor' seems to have approved of the bridge works, he has not been seen there since.

© Angela Marks 2010