The famous nineteenth-century engineer for the Great Western Railway, Isambard Kingdom Brunel, flirted with a novel means of train propulsion in the company's early years. This was 'atmospheric' propulsion. Brunel didn't invent it, but he thought he saw ways to improve it and make it practical for real-life use.
So far as I understand it, a pipe with a continuous slot in the top side was laid between the tracks (there was plenty of room, those being Broad Gauge days). The slot was sealed with a lubricated leather flap. The air inside the pipe was sucked out by coal-fired pumping stations at intervals along the line. Below each train was slung a piston that fitted the diameter of the pipe. As the train moved forward, the shaped rod that connected the piston to the train momentarily unsealed the flap, allowing air to rush in and push against the piston, propelling the train. When working properly, air pressure would push the piston (and the whole train) forward smoothly, and at speed.
I don't quite see how one got the train to start moving in the first place, nor what happened where lines joined or diverged, as that would mean smoothly combining two slotted pipes into one, and yet somehow preserving the vacuum. Tricky.
Brunel was however so confident of his atmospheric system's power to propel trains up quite steep gradients that he built the main line through South Devon with hillier sections than would otherwise be feasible for a single conventional locomotive. This of course saved money in construction costs, and must have been attractive to the GWR, and instrumental in their agreeing to adopt the system. It was up and running by 1847.
You'll have spotted the fatal flaw already: the leather flaps. At the time no better flexible material existed, and no weatherproof lubricant either. The pipes leaked air in, reducing the vacuum and therefore the system's efficiency. The leather quickly split with wear and tear. To cap it all, the fixed pumping stations consumed enormous quantities of coal - the vacuum had to be maintained whether or not trains were running - so what was saved in construction costs was more than cancelled out by mounting coal costs. Brunel had to admit defeat, and in 1848 reverted to conventional steam locomotives. The system had lasted only one year. Unwanted material was sold off at a colossal loss. And South Devon was now stuck with sections of steep railway line that were always a challenge until the diesel era.
Most of those pumping stations are long gone, but one remained at Starcross, on the main line south of Exeter, on the very banks of the River Exe estuary. I went to see it.
Approaching Starcross from the north, I was held up by roadworks. But a dark green GWR train (that's the modern company, not the original one), probably bound for Plymouth and beyond, passed by and relieved the tedium.