Roman aqueducts were built through a combination of precise surveying, massive organized labor, and ingenious engineering that moved water entirely by gravity across distances of up to 100 kilometers. The process started with finding a water source at a higher elevation than the destination city, then carefully mapping a route that maintained a gentle downhill slope, typically around 20 centimeters of drop per kilometer of length. From there, thousands of workers cut channels through rock, built arched bridges across valleys, and lined everything with waterproof concrete to deliver a continuous flow of fresh water.
Surveying the Route
Before any stone was cut, Roman engineers had to solve a fundamental problem: finding a path from a spring in the hills to the city that dropped just enough to keep water flowing, but not so steeply that the current would damage the channel. The target slope was roughly 0.02%, or a drop of about two centimeters for every ten meters. Pliny the Elder wrote that the slope “should not be less than twenty centimeters per kilometre of length.” Achieving that kind of precision over dozens of kilometers, through mountainous terrain, required careful measurement at every stage.
Engineers used three key instruments. The groma was a cross-shaped device with plumb lines hanging from each arm, used for sighting straight lines and establishing right angles. It worked alongside leveling rods and measuring chains to calculate horizontal distances and vertical differences using basic right-triangle geometry. The dioptra was more sophisticated: a sighting table that could be adjusted to a perfect horizontal, allowing surveyors to read elevations off distant rods where assistants moved a black-and-white disk until it aligned with the sight line. For confirming level over short distances, the chorobates used plumb lines hanging across marked scales to verify whether the instrument was perfectly horizontal.
For measuring distance across the ground’s surface, engineers used chains, cords, and sometimes a hodometer, a wheeled device that counted rotations to calculate distance traveled. A surveying team would work outward from the city toward the spring, measuring elevation changes between successive leveling rods, tallying horizontal distances, and recording the total drop from source to destination. This data determined not just the route but whether the project was feasible at all.
Organizing the Workforce
Aqueducts were built primarily by the Roman army, which provided both the engineering expertise and the disciplined labor force needed for a project that could stretch across 60 miles of terrain. Work camps were established along the entire construction route, from the spring source to the city walls. These camps functioned like small temporary towns, complete with day laborers, construction engineers, and even cafes and restaurants to feed the workers. When the route passed through an existing town, locals were hired as supplemental labor.
The military organization was critical. Construction could proceed simultaneously at multiple points along the route, with different teams cutting tunnels, laying channel beds, or building bridge foundations all at once. Coordinating these segments so they connected at the correct elevation required the surveying work to be completed well in advance and communicated clearly to each team.
Building the Channel
The main water channel, called the specus, was the backbone of every aqueduct. Most of the system ran underground or at ground level, not on the dramatic arched bridges that survive as ruins today. The typical specus was a rectangular channel lined with waterproof morite, a type of hydraulic concrete that the Romans made by mixing lime with volcanic ash. This lining prevented water from seeping into the surrounding earth and kept the channel smooth enough to maintain steady flow.
Where the terrain was relatively flat, workers dug a trench, built the channel walls from stone or brick, applied the waterproof lining, and covered the top with stone slabs. This buried design protected the water from contamination and the channel from weather damage. Where the ground rose, tunnels were cut directly through hillsides. Roman tunneling crews worked from both ends toward the middle, and also sank vertical shafts from the surface at intervals along the route. These shafts served multiple purposes: they allowed debris to be hauled out, provided ventilation, and let teams dig in both directions from each shaft, multiplying the number of work faces. Vertical shafts were preferred over angled ones because they made it easier to maintain accurate alignment underground. When two digging teams approached each other, they sometimes used acoustic communication, listening through the rock to guide their final connection.
Crossing Valleys
Valleys presented the biggest engineering challenge. When a valley was shallow enough, the Romans built the iconic multi-tiered stone bridges (arcades) to carry the channel across at the correct elevation. These bridges used semicircular arches built on stone piers, with the water channel running along the top.
When a valley was too wide or too deep for a bridge, engineers turned to a different solution: the inverted siphon. Water from the aqueduct channel flowed into a sealed tank at the top of one side of the valley, then entered a closed pipe that ran down the slope, across the valley floor, and back up the opposite side. At the top of the far side, it emptied into a receiving tank set slightly lower than the entry tank, and from there continued in an open channel again. This worked on the principle of communicating vessels: water in a sealed pipe naturally rises to nearly the same level it started from.
The pressures involved were enormous. In the lowest section of the valley, the weight of the water column above created forces that could reach extreme levels. The Hellenistic siphon at Pergamon in modern Turkey, which predated Roman examples, reached pressures of 19 bar at its deepest point, corresponding to a depth of 190 meters. Roman siphons used heavy lead or stone pipes to withstand these forces. In the valley bottom, the pipe was often mounted on a low “siphon bridge” to keep it above any river or stream crossing below.
Water Distribution at the City
The aqueduct’s journey ended at a structure called the castellum aquae, a large distribution tank usually built just inside the city walls. At Pompeii, this took the form of a circular basin that split the incoming water into three separate channels, each controlled by a gate system. This design let operators prioritize where water went during shortages. The primary recipients were public fountains spread throughout the city. Secondary lines fed public bathhouses, and a third set of pipes served private homes and businesses.
From the castellum, water traveled through the city in lead pipes. These were standardized by size, with the basic unit called the quinaria, a pipe formed by rolling a flat lead sheet five digits wide into a tube. Larger pipes were named by the same system: a six-pipe was six quarter-digits in diameter, a seven-pipe was seven, and so on up to a twenty-pipe. Pipes were cast in sections no less than ten feet long and joined together to form the urban distribution network.
Keeping the Water Flowing
Building an aqueduct was only the beginning. Mineral-rich water steadily deposited layers of calcium carbonate (limescale) on the channel walls, gradually narrowing the flow. Left unchecked, these deposits could reduce an aqueduct’s capacity dramatically over the years. Roman maintenance teams periodically entered the specus through access shafts and scraped the buildup away by hand using mattock-like tools with flat blades about 3 to 4 centimeters wide.
Evidence of this maintenance survives in aqueducts across the former empire, from Rome and Nîmes to Istanbul. Researchers studying the aqueduct at Divona in France found tool marks preserved in the carbonate layers, along with microscopic signs of physical impact: calcite crystals near the cleaning surfaces show deformation patterns that only form under pressure or shock, confirming the deposits were chipped away manually rather than eroded naturally. The channel floors were often cleaned more thoroughly than the walls, suggesting maintenance crews prioritized keeping the bottom clear to maintain water depth and flow rate. Debris from cleaning and repair work has been found embedded in later layers of deposits, creating a geological record of each maintenance visit.