How Were Aqueducts Constructed?

By Roberto Ferrari - originally posted to Flickr as Pont du Gard #3, CC BY-SA 2.0,

The aqueduct that fed the city of Saldae (modern Bejaïa, Algeria) can still be traced through the landscape. Of particular interest is the section of the aqueduct that cuts through a mountain. You can still walk down the passageways the soldiers cut through the solid rock. The engineer responsible for the surveying of the tunnelling and the organisation of the units designated to do the work in about AD 153 was a chap by the name of Nonius Datus, and we know that because Nonius set up an inscription telling us that he did it.

That's not uncommon in the Roman world, where public benefaction was part of the process by which one advanced through public office or where pride dictated that you show off a little once you'd engineered some spectacular new edifice, but Nonius isn't so much bursting with military pride, as seething with frustrated rage.

He was a military engineer who had been seconded from his legion, the III Augusta, to organise the construction, probably because he was an experienced hand and the III Augusta were good at building things. Nonius had done his job, told everyone what he wanted, set them digging and then promptly trotted off back to Italy to take up a well-earned retirement. That is, of course, until the people he left behind decided to fuck his nice aqueduct up royally.

The procurator of Mauretania Caesariensis promptly sent for Nonius to come back and fix the mess. Nonius takes up the story:

"I set out and on the way endured an attack by bandits. Although stripped and wounded, I got away with my team and reached Saldae. I met Varius Clemens [the procurator]. He took me to the mountain where they were crying over a tunnel of doubtful workmanship, which they thought had to be abandoned because the penetration of the digging of the tunnel had been carried further than the width of the mountain. It was apparent that the digging had strayed from the line, so much that the upper tunnel turned right, to the south, and likewise the lower tunnel turned north, to its right. So the two ends were out of line and had gone astray ... When I assigned the work, so they knew who had what quota of digging, I set up a work competition between the marines and the auxiliary troops. And so the linked up where the mountain was pierced ... When the water flowed, Varius Clemens dedicated the completed work.

(ILS 5795)

The tunnel is still there, 428 metres long.

This is a very interesting inscription. Not only is it evident that travelling across rural North Africa was a dangerous journey, fraught with angry bandits who would steal your clothes, but it tells us who built the aqueducts - the military - and that specialist units from relevant legions could be called up to supervise work across the empire. Much like the British Army's Royal Engineers, the Roman army would have specialist construction experts who could be moved where they were needed without having to move the whole legion. They would rely on local muscle, namely bored auxiliaries and marines, who, once they had finished bashing the locals into submission, otherwise had nothing much to do except to hack their way through mountains.

The construction of Roman aqueducts began with meticulous planning and surveying, processes that required a deep understanding of topography, hydrology, and urban needs. Roman engineers, often working under the direction of the state, were tasked with identifying suitable water sources and designing routes that could deliver water to cities efficiently. The primary considerations included the elevation of the water source, the distance to the city, and the terrain that needed to be traversed.

Roman engineers employed a range of sophisticated tools to survey the land and plan aqueduct routes. One of the most important instruments was the chorobates, a levelling device used to measure gradients. Vitruvius, the Roman architect and engineer, described the chorobates as a wooden frame with a flat surface and a water-filled channel, which allowed engineers to determine horizontal levels with precision (Vitruvius, De Architectura, 8.5). Another critical tool was the groma, a surveying instrument used to align straight paths and right angles. These tools enabled engineers to calculate the necessary gradient for the aqueduct, typically aiming for a slope of between 1:2000 and 1:5000 to ensure a steady flow of water without excessive pressure or stagnation (Hodge, 2002).

The planning process also involved careful consideration of the water source. Engineers preferred springs, which provided clean and reliable water, over rivers, which were more susceptible to contamination and seasonal fluctuations. Once a source was selected, engineers mapped out a route that minimised construction challenges while maintaining the required gradient. This often involved navigating complex terrain, including valleys, hills, and rivers, which required innovative engineering solutions such as bridges, tunnels, and siphons.

As to how they were constructed, fortunately, we have a detailed account telling us how to do it. Vitruvius' Architecture goes into great length on the subject:

" There are three methods of conducting water, in channels through masonry conduits, or in lead pipes, or in pipes of baked clay. If in conduits, let the masonry be as solid as possible, and let the bed of the channel have a gradient of not less than a quarter of an inch for every hundred feet, and let the masonry structure be arched over, so that the sun may not strike the water at all. When it has reached the city, build a reservoir with a distribution tank in three compartments connected with the reservoir to receive the water, and let the reservoir have three pipes, one for each of the connecting tanks, so that when the water runs over from the tanks at the ends, it may run into the one between them.

From this central tank, pipes will be laid to all the basins and fountains; from the second tank, to baths, so that they may yield an annual income to the state; and from the third, to private houses, so that water for public use will not run short; for people will be unable to divert it if they have only their own supplies from headquarters. This is the reason why I have made these divisions, and also in order that individuals who take water into their houses may by their taxes help to maintain the conducting of the water by the contractors.

If, however, there are hills between the city and the source of supply, subterranean channels must be dug, and brought to a level at the gradient mentioned above. If the bed is of tufa or other stone, let the channel be cut in it; but if it is of earth or sand, there must be vaulted masonry walls for the channel, and the water should thus be conducted, with shafts built at every two hundred and forty feet.

But if the water is to be conducted in lead pipes, first build a reservoir at the source; then, let the pipes have an interior area corresponding to the amount of water, and lay these pipes from this reservoir to the reservoir which is inside the city walls. The pipes should be cast in lengths of at least ten feet. If they are hundreds, they should weigh 1200 pounds each length; if eighties, 960 pounds; if fifties, 600 pounds; forties, 480 pounds; thirties, 360 pounds; twenties, 240 pounds; fifteens, 180 pounds; tens, 120 pounds; eights, 100 pounds; fives, 60 pounds. The pipes get the names of their sizes from the width of the plates, taken in digits, before they are rolled into tubes. Thus, when a pipe is made from a plate fifty digits in width, it will be called a "fifty," and so on with the rest.

The conducting of the water through lead pipes is to be managed as follows. If there is a regular fall from the source to the city, without any intervening hills that are high enough to interrupt it, but with depressions in it, then we must build substructures to bring it up to the level as in the case of channels and conduits. If the distance round such depressions is not great, the water may be carried round circuitously; but if the valleys are extensive, the course will be directed down their slope. On reaching the bottom, a low substructure is built so that the level there may continue as long as possible. This will form the "venter," termed Κολία by the Greeks. Then, on reaching the hill on the opposite side, the length of the venter makes the water slow in swelling up to rise to the top of the hill.

But if there is no such venter made in the valleys, nor any substructure built on a level, but merely an elbow, the water will break out, and burst the joints of the pipes. And in the venter, water cushions must be constructed to relieve the pressure of the air. Thus, those who have to conduct water through lead pipes will do it most successfully on these principles, because its descents, circuits, venters, and risings can be managed in this way, when the level of the fall from the sources to the city is once obtained.

It is also not ineffectual to build reservoirs at intervals of 24,000 feet, so that if a break occurs anywhere, it will not completely ruin the whole work, and the place where it has occurred can easily be found; but such reservoirs should not be built at a descent, nor in the plane of a venter, nor at risings, nor anywhere in valleys, but only where there is an unbroken level.

But if we wish to spend less money, we must proceed as follows. Clay pipes with a skin at least two digits thick should be made, but these pipes should be tongued at one end so that they can fit into and join one another. Their joints must be coated with quicklime mixed with oil, and at the angles of the level of the venter a piece of red tufa stone, with a hole bored through it, must be placed right at the elbow, so that the last length of pipe used in the descent is jointed into the stone, and also the first length of the level of the venter; similarly at the hill on the opposite side the last length of the level of the venter should stick into the hole in the red tufa, and the first of the rise should be similarly jointed into it.

The level of the pipes being thus adjusted, they will not be sprung out of place by the force generated at the descent and at the rising. For a strong current of air is generated in an aqueduct which bursts its way even through stones unless the water is let in slowly and sparingly from the source at first, and checked at the elbows or turns by bands, or by the weight of sand ballast. All the other arrangements should be made as in the case of lead pipes. And ashes are to be put in beforehand when the water is let in from the source for the first time, so that if any of the joints have not been sufficiently coated, they may be coated with ashes.

Clay pipes for conducting water have the following advantages. In the first place, in construction:—if anything happens to them, anybody can repair the damage. Secondly, water from clay pipes is much more wholesome than that which is conducted through lead pipes, because lead is found to be harmful for the reason that white lead is derived from it, and this is said to be hurtful to the human system. Hence, if what is produced from it is harmful, no doubt the thing itself is not wholesome.

This we can exemplify from plumbers, since in them the natural colour of the body is replaced by a deep pallor. For when lead is smelted in casting, the fumes from it settle upon their members, and day after day burn out and take away all the virtues of the blood from their limbs. Hence, water ought by no means to be conducted in lead pipes, if we want to have it wholesome. That the taste is better when it comes from clay pipes may be proved by everyday life, for though our tables are loaded with silver vessels, yet everybody uses earthenware for the sake of purity of taste.

But if there are no springs from which we can construct aqueducts, it is necessary to dig wells. Now in the digging of wells we must not disdain reflection, but must devote much acuteness and skill to the consideration of the natural principles of things, because the earth contains many various substances in itself; for like everything else, it is composed of the four elements. In the first place, it is itself earthy, and of moisture it contains springs of water, also heat, which produces sulphur, alum, and asphalt; and finally, it contains great currents of air, which, coming up in a pregnant state through the porous fissures to the places where wells are being dug, and finding men engaged in digging there, stop up the breath of life in their nostrils by the natural strength of the exhalation. So those who do not quickly escape from the spot, are killed there.

To guard against this, we must proceed as follows. Let down a lighted lamp, and if it keeps on burning, a man may make the descent without danger. But if the light is put out by the strength of the exhalation, then dig air shafts beside the well on the right and left. Thus the vapours will be carried off by the air shafts as if through nostrils. When these are finished and we come to the water, then a wall should be built round the well without stopping up the vein.

But if the ground is hard, or if the veins lie too deep, the water supply must be obtained from roofs or higher ground and collected in cisterns of "signinum work." Signinum work is made as follows. In the first place, procure the cleanest and sharpest sand, break up 'lava' into bits of not more than a pound in weight, and mix the sand in a mortar trough with the strongest lime in the proportion of five parts of sand to two of lime. The trench for the signinum work, down to the level of the proposed depth of the cistern, should be beaten with wooden beetles covered with iron.

Then after having beaten the walls, let all the earth between them be cleared out to a level with the very bottom of the walls. Having evened this off, let the ground be beaten to the proper density. If such constructions are in two compartments or in three so as to insure clearing by changing from one to another, they will make the water much more wholesome and sweeter to use. For it will become more limpid, and keep its taste without any smell, if the mud has somewhere to settle; otherwise it will be necessary to clear it by adding salt. In this book I have put what I could about the merits and varieties of water, its usefulness, and the ways in which it should be conducted and tested; in the next I shall write about the subject of dialling and the principles of timepieces."

(Vitruvius, Architecture, viii, vi)

That's an awful lot of text, but it does pretty much comprehensively answer the question of how aqueducts were constructed and how they worked.

It also, you may have noticed, gives a recipe for opus signinum (signinum work.) or, to give it its more common name - Roman concrete - and one of the most persistent myths about Roman construction techniques is that nobody knows how Romans made such long-lasting concrete. Not only do we know, we have the recipe right there; you just need some volcanic rock.


References and Further Reading

Evans, H. B. (1994). Water Distribution in Ancient Rome: The Evidence of Frontinus. University of Michigan Press.

Fabre, G., Fiches, J.-L., & Paillet, J.-L. (2000). L’aqueduc de Nîmes et le Pont du Gard: Archéologie, géosystème, histoire. CNRS Éditions.

Frontinus, S. J. (97 AD). De Aquis Urbis Romae.

Hodge, A. T. (2002). Roman Aqueducts & Water Supply. Duckworth.

Leveau, P. (1991). Le pont du Gard et l’aqueduc de Nîmes. Éditions du CNRS.

Vitruvius.. De Architectura.

Thanks for reading! If you’re stuck for a gift for a lost loved one, or for someone you hate, or if you have a table with a wonky leg that would benefit from the support of 341 pages of Roman History, my new book “The Compendium of Roman History” is available on Amazon or direct from IngramSpark. Please check it out by clicking the link below!