I have a world that isn't an exact analogue to medieval technology, closer to the late 1700s in terms of culture and agriculture/metallurgy, only without gunpowder. I have them ahead of real history for this time period in some areas (e.g., medical and farm science), but behind in others (notably military science, they still use enlightenment style militaries - pikemen, heavy cavalry, medieval-style ballistae and trebuchets, crossbows, etc). I have been thinking of a system for rapid mass transit over land and wanted to find out if it was feasible as a replacement for trains as internal combustion engines will not be present in this world for various handwave-y reasons.

Here's what I'm thinking: The fastest and most efficient form of travel before engines was boats - rivers were the primary way to ship things in bulk with any speed over continents, and ocean going vessels are much faster oftentimes than traversing the inland of a continent on foot/horse, even if you end up going way more miles roundabout. So, I have an empire that wishes to build artificial rivers of a sort in order to facilitate this fast travel, connecting their cities. I imagine massive raised aqueducts 100m in width and a few tens of meters in depths. Through a massive set of pumps powered by manual labor / ox teams / wind power, huge amounts of water are pumped up to fill the aqueduct for miles and miles.

The aqueducts are split into half, and each half has a slightly opposite slope: so that the water flows in one direction on each half, so even unpowered ships can just float down the aqueduct to their destination, though oars/sail are typically used to increase the speed. The slope is just slight enough to get the water moving, it doesn't have to be a large incline. The aquifers are raised so that if for instance the destination city is at a higher altitude than the source city, the aqueduct at the source is still higher to maintain that incline. Ignore problems associated with moving cargo up and down these aqueducts.

Assume almost unlimited manpower / political will to accomplish this. Is this a feasible engineering feat for this general level of tech? I imagine the aqueducts will be made from something akin to concrete joining large cut stone blocks but am open to suggestions. My only rule is no gunpowder/combustion engines.

  • 2
    $\begingroup$ Just to clarify, you're suggesting that your people need to build canals that are ~50m wide and >20m deep for their transport needs? $\endgroup$ – Starfish Prime Nov 18 at 16:08
  • 11
    $\begingroup$ Just use standard canals (30 m wide, maybe 4 m deep) and employ the manpower you want to waste on pumping water on towing/rowing boats. Why change a working system? Especially if the architecture almost certainly requires reinforced concrete. This also works with the cities, since a lot of cities are situated at rivers - so no need for canals/aqueducts. $\endgroup$ – Erik Nov 18 at 16:15
  • 2
    $\begingroup$ @Erik manpower is a problem, because you've gotta feed, house and manage all those people. If it needs millions of people to do it, then you must bend the entirely industrial output of a nation towards the effort. If you can handwave away the need for using your entire populace to build your canals, then you can just handwave away the rest of the feasibility issues. $\endgroup$ – Starfish Prime Nov 18 at 16:32
  • 4
    $\begingroup$ Will there be enough pre-industrial cargo to use this vast canal network? Large quantities of artisan-made goods require whole cities of artisans to provide the supply...who could often simply migrate closer to the demand. $\endgroup$ – user535733 Nov 18 at 17:01
  • 2
    $\begingroup$ Medieval (and Renaissance) Europe was full of man-made canals and intricate systems of locks to help to movement of goods. It worker pretty well :) Of course, they weren't used for moving the armies, that would be quite silly. $\endgroup$ – Luaan Nov 19 at 7:24

At 100m wide and 20m deep, you are looking at needing to pump 2000 cubic meters of water for every meter you want a boat to drift. Now, let's say you want your ships to move at the speed of a slow moving sailboat (~4 knots / 8 kph), that means you would need to pump 16 million cubic meters of water an hour. A really good manual bilge pump in the hands of a well rested, and physically fit user can displace about 7 cubic meters of water an hour. Pumping is hard work; so in order to keep your army of manual slave pumpers pumping going at anywhere near peak efficiency, you'll probably need to to have them working in teams of 3 for 3 seperate 8 hour shifts. This equals 9 workers per pump for a total workforce requirement of just a bit over 20 million people. Now, these estimates are for based on what it takes to get water over the side of a boat using a pump made to modern specifications. To raise water 10s of meters using renaissance level engineering, your looking at needing multiple stages of pumps increasing this requirement to well over 100 million laborers.

Since you say "almost unlimited manpower", I would say the answer here will be a pretty solid no. Even major modern cities like New York do not have enough population to perform this sort of work no matter how tyrannically your government tries to make it happen.

That said, your idea of water "highways" is not entirely un-feasible. The hard part about using rivers and canals in a pre-industrial civilization is not that you NEED the current to help you move downstream so much as the current works against you when you go up stream. If you were to build the raised aqueducts to be perfectly level, they could retain rain water and allow for 2-way transit without ever having to work against the current. Then, even simple sailing barges could go both ways at speeds of 4+ knots as easily as if they were traveling in open waters making mass transit doable.

In terms of how fast you can travel using this system: your fastest sailing ships of the era could hit cruising speeds of 8-11 knots covering distances of about 350-500km per day depending on wind conditions. That means that your faster ships could match pony express speeds delivering mail, passengers, and small amounts of priority freight anywhere the country in a matter of days.

Another possible implementation is to use an actually aqueduct. Rome's aqueducts collected melting snow water high up in the mountains and used the elevation difference to carry the water down to the city. If instead you used an aqueduct to carry water half way down, then you could deliver cargo barges to a raised aquifer in the city that would second as a trade port, then you send cargo the rest of the way down back to the 1st town in the foothills near the aqueduct origin. In this way you could maintain two-way flowing water traffic, but only in places where the geography allows for this.

enter image description here

  • 1
    $\begingroup$ Thats a good point about working against the current. It might be better just to have them flat and avoid having to pump altogether. $\endgroup$ – Rick D Nov 18 at 18:29
  • 1
    $\begingroup$ "displace about 7 cubic meters of water an hour" displace by how much? $\endgroup$ – njzk2 Nov 19 at 2:06
  • $\begingroup$ @njzk2 the capacity of pumps is measured by how much they can 'displace', i.e. move from the input to the pump to the output of the pump, in terms of volume per unit time. Pumps usually are connected to pipes and if you keep pushing water into a pipe it eventually comes out the other end; how far away the other end of the pipe is is not a feature of the pump. $\endgroup$ – Pete Kirkham Nov 19 at 17:24
  • $\begingroup$ @PeteKirkham Yes, that is true. I would also specifically note that this is a measure of "horizontal" displacement. If you destination is higher or low than you are, then the work the pump can perform is affected which is why the amount of labor required goes up so much when you are trying to pump up into a higher reservoir. $\endgroup$ – Nosajimiki Nov 19 at 19:20
  • $\begingroup$ @PeteKirkham ok. I was referring to the elevation, as since the canals go down, the water has to be brought back up $\endgroup$ – njzk2 Nov 21 at 7:15

I imagine massive raised aqueducts 100m in width and a few tens of meters in depths

Ok, for starters: scale this back a bit. Not just for practicality reasons, but for the simple fact that it is so ridiculously excessive.

The real world Panama Canal has a standard called Panamax, which allows ships to have a 12m draft, 32m beam and 290m length. That's the size of the locks they need to fit into. They can carry over 52000 tonnes of cargo (excluding the dry mass of the ship). Panamax boats are serious, big, oceangoing vessels.

Needing drafts of over 20m for a pre-industrial society seems frankly implausible. For comparison, HMS Victory, a 100-gun first rate warship of the 1700s, had a draft of 5.5m.

Next, lets think about aqueducts. The biggest aqueduct in the world is the Roquefavor Aqueduct, and the water channel on top of that is merely 30ft wide by 7ft deep (google books link, often doesn't work for people other than the first person to click them. Page 431 of "Engineering in History"). It too is something of an engineering marvel, and holds over 90000 tonnes of water. Your aqueducts would hold the same volume of water in a mere 45m length instead of 300m, and the pressure the walls of the aqueduct would need to withstand at the bottom is over 9 times greater and you couldn't use blasting or power tools to prepare your masonry and so on and so forth.

Really, making country-spanning stone aqueducts is probably impractical, even assuming you could find and cut and fit enough stone to do the job. What you'll have is more like giant levees to raise the level of the canal above the surrounding terrain, which brings in a whole new set of problems to do with leaks, damage and flooding but at least the construction problem has been reduced to one of semi-plausibility.

Ignore problems associated with moving cargo up and down these aqueducts.

Well, that seems a bit of a strange request, when you're asking about the feasibility of the whole enterprise.

The slope is just slight enough to get the water moving, it doesn't have to be a large incline. The aquifers are raised so that if for instance the destination city is at a higher altitude than the source city, the aqueduct at the source is still higher to maintain that incline.

You're inventing a really complex engineering solution which requires a colossal building effort to handle a problem that has already been solved with the aid of locks. You're already happy to build huge structures and create massive pumping systems powered by animals, so you'd be better off replacing your original idea with a perfectly sensible minimal-flow canal with animal-drawn barges. You solve the issue of having to keep your brobdingnagian artficial rivers filled and flowing, and you reduce the need for huge embankments and aqueducts in order to get the vertical elevation you need. What's not to like?

(As an aside, I bet you haven't considered the curvature of the earth in this plan, which will require your endpoints to have a much greater vertical separation than they would on the discworld).

Now lets think about landscape.

The Grand Canal, as referenced by Priska's answer, sensibly involved no aqueducts and no tunnels, helped in no small part by the fact that there's lots of fairly flat, broad terrain available for building a conventional canal, which in turn means lots of farming and food and a large labour pool available to work on such constructions. If wikipedia was to be believed, the labour force for some sections of the canal was as large as five million people back in 600CE, no mean feat.

the empire building these is fairly dry and thus their major cities aren't built on rivers/lakes as we're used to

Well, where on earth are you going to get all the water from in the first place? And how do you expect it to remain in the canals, given problems with evaporation? And even if you've solved that, how are you going to source and feed your colossal work crews? You'll need an empire-sized logistics operation just to get food and water out to thousands or millions of people digging ditches in the desert; this is a vastly harder problem than the builders of the Chinese Grand Canal had to deal with.

Honestly, you'd do better to build a decent road network, or some kind of railway (presumably horse drawn) under the circumstances.

(As another aside, there's a fascinating sort of aqueduct that can be found in some very arid environments in the form of the qanat. Not mass-transit scale, or even tiny-transit scale, but potentially relevant to your interests and should give you some idea of what sort of problems you might have to think about when dealing with water in the desert.)

  • $\begingroup$ Thanks for the heads up on the qanat, very interesting! You're right, it looks like I probably need to scale down some things $\endgroup$ – Rick D Nov 18 at 18:26

It will not be feasible to work with that much water. Let's start with some numbers.

Based on the original specs of the Erie canal, we'll aim to have a canal 12m wide and 1.2m deep, flowing at a rate of about 2m/s (which is about as fast as you could travel on a canal by animal power). Since a canal is two-way, we'll cut that in half for each channel - 6m by 1.2m by 2m/s. That's a flow rate of 14.4 cubic meters per second (about a hundredth of a Nile) or 14,400 kg of water per second.

Assume that the aqueduct is tall enough to walk under - so we're pumping water about 4m off the ground to get it in the highest point. It requires 10 N of force per kg to lift, which adds up to 576,000 joules per second, or 576 kilowatts of power to pump water into the initial end of each canal.

One horsepower is notionally about 750 watts, so your pumping station would need about 770 horses, working around the clock, to haul enough water into the near end of the aqueduct to beat the kind of speeds you could achieve using a handful of oxen and a barge on a conventional flat canal. Obviously, if you want to go faster, you need to pump more water (or your aqueduct will go dry), which will require more power.

  • 2
    $\begingroup$ I note that if you could get a flowing canal of this type, the amount of stuff you could float down it on barges would absolutely dwarf the amount you could haul down it with your 770 horses (about 30 tons apiece, apparently). Thing is though, I'm not sure a pre- or peri-industrial civilisation needs the capacity to ship millions of tonnes of stuff around. What is all that stuff? How did it get there in the first place? How are you going to make use of it all? etc etc. $\endgroup$ – Starfish Prime Nov 18 at 18:33
  • 1
    $\begingroup$ @StarfishPrime Overland, a 2-horse drawn cart maxes out at 4-5 tons; so, 770 horses can do ~1700 tons when fully burdened, but more like 1100 tons if you want to go long distances. While this is a lot of goods for a pre-industrial society that mostly uses local resources, once you make exchanging resources more affordable, you'll see a lot more trade start to happen. If you can suddenly start trading heavy stuff like flagstone and pine 100s of km apart, you'll see even mid-sized towns quickly blow past what those wagons can do. $\endgroup$ – Nosajimiki Nov 18 at 19:25
  • 3
    $\begingroup$ @Nosajimiki I think the better comparison is the one Starfish presented, comparing the tow capacity on a flat canal to the capacity of flowing canal. Building the canal and leaving it unpowered with no flow in either direction will give you a capacity at least 10 times the overland capacity you're comparing against. On an unpowered canal, we're still talking tens of thousands of tons of cargo, comparable to a modern container ship. Given city sizes of the day, this means every citizen is shipping a ton of cargo roughly every week. $\endgroup$ – Nuclear Wang Nov 18 at 19:50
  • 3
    $\begingroup$ @MichaelShopsin no-one is disputing the usefulness of canals. The issue is a) the sheer size of system that the OP is proposing, and b) the use of a flowing artificial river to deliver cargo instead of using a more-or-less static canal and draught animals. $\endgroup$ – Starfish Prime Nov 18 at 20:05
  • 2
    $\begingroup$ The main issue is the elevation and flowing water. If water flows out, and it's not along a natural watershed (which presents its own issues) you have to keep replacing it - a lot of it, all the time. $\endgroup$ – Cadence Nov 18 at 20:25

All due respect your knowledge of transit technologies of this era is quite limited, and you are wild-guessing a lot, e.g. the dimensions of the canal.

You're quite close. Push to the railroad.

In late 1700's you are only 40 years from the critical elements of the railroad. And this is a time when technology did not move along at a breakneck speed! Since you have resources to spare, have your society do a Manhattan Project push to get there quickly.

Canal is nope

You propose wild-guess 100m x 20m size. But even if we wind back to a contemporary/late 17th century 2.2m x 0.8m UK standard... As others discuss, it is simply unfeasible to pump that much water for a flowing canal, due to the sheer number of human or animal pumpers required, and the impossibility of the land within transport distance feeding that many pumpers given agricultural yields of the age.

Further, you would need to maintain a perfect grade. That either means lots of curving to follow the lay of the land, which means lots of canal navigation accidents (keep in mind that a boat being swept by currents has no rudder authority since it has no moving water past its rudder) ... or lavish bridge structures impossible to build and easily attacked or sabotaged.

This reality drags you back to a "normal" standing water canal where animals provide towpath towing, almost certainly with locks to follow the terrain. Extreme engineering could reduce the frequency of those locks, but not their lifting height.

Given your lack of abundant water, it will be all the pumpers can do to make up for leakage and lock consumption. This right here is what makes the canal concept folly.

Also, canals can best about 4 mph. A mounted cavalry can move faster than that, and could outrace a retreating army and delay them with sabotage.

Railroads without engines?

Again your knowledge of railways is limited. Internal combustion engines did not become a factor on the railroad until post-World War II. Really. Prior to that, they relied on the high pressure, "compact" steam engines proposed by George Stephenson (as opposed to James Watt's brobdignandian low pressure engines, quite suitable for canal pumping). But Stevenson's engine merely "put railroads on the map" by providing ludicrous soeeds. Long before this, railways used animal tow. They were around well prior to 1800, though in short range or within-a-site usage for material handling.

If you have 40 tonnes of goods to move on canal vs railway, and only animals to tow them, you are much, much better off on the railway. Friction is close to nil, and you fight only grades. You can change animals as needed, and have extra animals stabled near the notable grades.

Speed is up to you; do you assign a mule, or a team of horses? Not much different than teaming any wagon, but you carrying capacity is much, much larger for the same speed.

The fact of having grades is an important sidebar. It means you can follow the land, which vastly reduces your structural engineering. A simple line can be tossed down with a washboard gradient and put in service, and engineering works can come later to level it out. Can't do that with canals.

As said, railways already were coming into being with animal power by your time. Your "Manhattan Project" is to scale them up into wide loading-gage, well-groomed and fast conveyances fit for inter-city travel.

And once available, engines allow a leap to unthinkable speeds on the same trackway. Adding engines to canals barely increases their speed over horse teams; the UK canals are still 4 mph.


That sounds a pretty large undertaking, even now!

As Erik mentioned in the comments, why not use "normal" rivers, and link them with canals? The UK still has a lot of the canals it had prior to the rail network. For the vast majority of the time they were in use, the boats would have been pulled by horses - canal boats could carry thirty tons at a time with only one horse pulling.

What it's not though is quick! Great for bulk transport, not so good for speed. Prior to the invention of motors, the fastest transport over distance was the pony express. The horses were ridden at a fast trot, canter or gallop, around 10 to 15 miles per hour. Still not very quick, though you could get a letter from the East to West Coasts of America in 10 days, which is not too shabby! You couldn't pull a boat at that speed though.


If you are looking for transit alternatives other than water: unpowered air gliders are feasible with medieval technology, can be manufactured from wood/cloth/paper/ropes. It was not done then because the optimal wing shape and method of steering was not known and it was hard to discover by trial and error.

It may not be considered "mass" transit as it's not feasible to build large glider for more than one or two people, but many gliders can be built and launched. Another disadvantage - it is even more dependent on fair weather than shipping.

They need elevated places to launch from, rising currents to sustain the altiutde and some flat place to land.

If there are no elevated launch places, launch towers could be feasibly built or existing towers used. Horses are too slow to tow the glider, however if faster animals can be harnessed this is an option too.

Rising air currents occur either natural or can be engineered where convenient by creating patch of sand or rocks that is heated by sun more than surrounding terrain.


Are you familiar with the Grand Canal? It's not exactly "land transit", but maybe this is what you're looking for.

  • 3
    $\begingroup$ Worth noting a few things, like the size of the labour force, the topography of the land it was built in (no need for tunnels and aqueducts), and perhaps most importantly: the fact that it wasn't made as a megalithic stone aqueduct, but a sensible and conventional canal. $\endgroup$ – Starfish Prime Nov 18 at 16:37
  • 1
    $\begingroup$ Please add some additional context to what relationship the Grand Canal has to do with the OP. As it is, this is a link only answer. $\endgroup$ – Nosajimiki Nov 18 at 22:20

As @Harper said: railroads without locomotives. Your society should create rail tracks, made of iron, and pull the wagons using animals. In flat terrain you can move a surprisingly large amount of cargo using this method because rail reduces friction. If the wagons are well built, with well-fitting, well-oiled, standard parts, you get another boost to the amount of mass you can move. Speed will top around 15 km/h, the max speed a horse wagon can go (horses can go faster but not for long, so you want a sustainable speed). Since your tech level is around 1700, you can manufacture standardized parts in large scale because you alredy have the technology (efficient water mills) to build factories.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.