The Railway That Ran on Nothing

In 1848, a railway train could cross the Devon coast without a locomotive. Nothing pulled it. A row of stationary steam engines beside the track pumped air from a long iron pipe, and the pressure of the atmosphere pushed a piston beneath the first carriage. The train moved because there was less air in front of it than behind it.

The idea sounded like a way to make railways cleaner, quieter and stronger on hills. It also made the train dependent on a narrow leather flap running for miles through the weather. When the flap leaked, the invisible engine became visible as a stalled train, an overheated pump and a bill no one wanted to pay.

The force already waiting in the sky

The atmospheric railway was not a train travelling through a tube. Its carriages ran on ordinary rails. Between those rails lay a cast-iron pipe with a slit along its upper side. A piston inside the pipe was connected by a bracket to the carriage above. Stationary steam engines at pumping houses removed air from the section ahead. Atmospheric pressure then pressed on the back of the piston and drove the train forward.

The arrangement separated power from the vehicle. A conventional locomotive had to carry its boiler, fuel, water and machinery wherever it went. An atmospheric railway left the heavy engines beside the line. The carriages would be lighter, the gradients easier, and the countryside free of smoke from a locomotive chimney. On paper, the system offered the advantages of steam without asking every train to carry its own steam plant.

The crucial phrase was “on paper.” The piston needed to move through the pipe while the pipe remained nearly airtight. A continuous leather valve covered the slot behind it. The leather had to flex when the piston passed, then lie flat again before the next section lost its vacuum. It was a seal, a hinge and a piece of trackside weatherproofing at once.

The system’s early demonstrations made the difficulty look manageable. By 1845, the House of Commons had appointed a Select Committee to examine its merits, using the Kingstown and Dalkey railway in Ireland as an important practical example. Atmospheric propulsion had moved beyond a laboratory trick. It was now a question for railway bills, investors and Parliament: not whether air pressure could move a train, but whether it could do so every day at a cost that made sense.

Brunel puts his faith in the pipe

Isambard Kingdom Brunel was not the inventor of the atmospheric principle, but he was its most consequential British advocate. In 1844 he recommended it for the South Devon Railway, whose route west from Exeter included steep gradients and difficult coastal engineering. The promise was especially attractive there. If atmospheric power could handle hills without relying on the adhesion between iron wheels and iron rails, the line might be built and operated with fewer compromises.

Brunel’s decision was not irrational. The principle had been demonstrated at small scale and in regular service elsewhere. The South Devon railway was also being designed at a moment when railway technology was changing faster than experience could govern it. Engineers were still discovering how heavy trains would become, how frequently passengers would expect them to run and how much maintenance a line could absorb.

Construction created its own machinery landscape. Engine houses had to be placed along the route, with boilers, pumps, water supplies and staff ready to create the vacuum in the right section at the right time. Peter John Margary, Brunel’s chief assistant on the Bristol and Exeter Railway, later became responsible for building the South Devon engine houses. His diaries, preserved in the National Railway Museum’s archives, record visits to those houses and meetings about the engines needed to operate them.

That is the part of the invention a drawing tends to hide. A locomotive is a machine that arrives with the train. An atmospheric railway is a machine distributed across the landscape. Every mile of pipe is part of the engine, and every engine house is part of the timetable.

The weather enters the timetable

The first atmospheric train on the South Devon line ran experimentally in September 1847. Full atmospheric operation between Exeter and Newton Abbot began in February 1848. For a while the railway did what its advocates had promised. The train moved without a locomotive, and the steep country no longer seemed to belong exclusively to the steam engine.

Then the seal began to fail.

The leather flap had to remain flexible, but rain soaked it, frost stiffened it and dry weather shrank it. The system used grease to preserve the leather, which introduced another problem: vermin found the treated material worth eating. Any damage allowed air into the pipe. The pumping engines then had to work harder to restore the vacuum, consuming more fuel while producing less useful motion.

Brunel’s report to the South Devon directors on 19 August 1848 dealt with the atmospheric apparatus, the longitudinal valve and the failures caused by heat and rain. The problem was not one dramatic break that an engineer could repair and forget. It was the repeated conflict between a long flexible seal and an outdoor railway. The train needed the pipe to behave like a laboratory instrument. The Devon weather kept reminding it that it was infrastructure.

The economics were no kinder. At the time atmospheric working cost 3 shillings and 1 penny per mile, compared with 1 shilling and 4 pence for locomotive working. The system that had promised to save the weight of locomotives was requiring a chain of pumping stations, fuel, operators and repairs. The air itself was free. Making the air behave was not.

On 10 September 1848, atmospheric operation between Exeter and Newton Abbot ended. The more difficult section towards Totnes had never entered atmospheric service, although pipes had been laid and an engine house built. The abandoned iron tubes remained as a physical record of a railway whose power had been spread along the route and lost at its weakest seam.

A failure with a useful shape

The atmospheric railway did not fail because the atmosphere was too weak. It failed because a working principle had to survive scale, weather and traffic simultaneously. A short demonstration could prove propulsion. It could not prove that a seal would remain sound for miles, that pumping stations would coordinate with trains, or that maintenance would stay cheaper than attaching an engine to the front.

Steam locomotives improved while the atmospheric system struggled. That changed the comparison. The vacuum railway was not competing with the crude locomotives Brunel had known when he made his recommendation; it was competing with better locomotives arriving during the same decade. A technical advantage can disappear without being disproved. It only has to be outpaced.

The lesson is visible in the surviving pipe. The machine was elegant in isolation: pressure behind a piston, motion along a rail. The railway was not an isolated machine. It was a promise made repeatedly across distance, in frost and rain, at every joint and every timetable. The locomotive carried its weakness with it, where mechanics could see it. The atmospheric railway hid its weakness under a strip of leather stretching through the landscape.

The train ran on nothing, until the nothing got in.