The tall arcades that survive in Italy, France and Spain give a misleading picture of Roman aqueducts. Most of the length of a typical aqueduct ran in a covered channel at or below ground level, following the contours of the land. Arches appeared only where the line had to cross a valley, and those are the sections most likely to be standing today. Behind them lay a system of surveying, construction and management that supplied large cities with water without pumps.
The first aqueducts
Rome had used local springs, wells and the Tiber for centuries. Tradition dates its first major aqueduct, the Aqua Appia, to 312 BCE, and ancient writers credit the censor Appius Claudius with it. It ran largely underground. More followed over the next centuries, and by the late first century CE the city was served by nine aqueducts, according to the account of Sextus Julius Frontinus, who was put in charge of the water supply around 97 CE and wrote a handbook on it. Aqueducts were later built across the provinces, in towns of every size. The pace of this building tracked Roman expansion, as new cities were founded or enlarged.
Surveying a gentle slope
Water in an open channel flows only if the channel slopes continuously downhill, so the central task was to maintain a very shallow, steady fall over distances that could run to many kilometres. Surveyors used simple instruments to do this. Vitruvius, writing in the first century BCE, describes the chorobates, a long levelling bench with a water channel on top, and the dioptra, a sighting device. The groma helped lay out straight lines and right angles.
Engineers needed a consistent gradient, and Vitruvius recommends a small minimum fall, though real aqueducts varied widely depending on the terrain. The aqueduct to Nîmes in southern France, which includes the Pont du Gard, is a well-known example: over its whole length of roughly 50 kilometres its average fall was well under a metre per kilometre, which shows how fine the work could be. Surveyors could not always follow the shortest line, so channels wound around hills, passed through tunnels and crossed valleys on bridges.
Building the channel
The channel, or specus, was usually lined with waterproof mortar and covered with a vault or slabs to keep out dirt, leaves and sunlight. Much of it was built in a trench, and tunnels were dug from both ends with shafts providing access for workers and airflow. Where a valley was shallow, the channel crossed on a substructure of masonry; where it was deeper, builders raised tiers of arches. The Pont du Gard stands close to 50 metres high, and the aqueduct at Segovia in Spain is another prominent survivor. Concrete and stone were used together, and the aqueduct builders drew on the same techniques as Roman concrete used in other large structures.
Construction labour is poorly documented in detail. Soldiers from the Roman legions are known to have built some aqueducts, civilian contractors and local labour built others, and costs were met by emperors, local councils or wealthy donors.
Siphons, tanks and distribution
Roman engineers did sometimes use pressure. Where a valley was too deep or wide for a bridge, water could run down into pipes, usually lead, cross the valley under pressure in an inverted siphon, and rise on the other side, though the far side had to be a little lower than the start. Several siphons have been identified in the provinces, including at Lyon and Pergamon.
Along the route and at its end, settling tanks let sediment drop out. In towns the main flow arrived at a distribution tank, the castellum, and from there was divided into separate pipes. Priority generally went to public fountains and basins, then to the baths and public buildings, and then to private houses that paid for a connection. Frontinus describes standard pipe sizes used to meter private supplies and records complaints about illegal taps. Pompeii, which had a distribution tank and many street fountains, gives a rare view of how water reached ordinary neighbourhoods, and it is an important part of the evidence for Pompeii daily life.
What the water was for
Fountains were the main source of drinking and household water for most residents, particularly for people living in upper floors of Roman apartment blocks without a piped supply. The great Roman public baths consumed large quantities, and so did latrines, ornamental pools and some industrial workshops, since flowing water also flushed drains. The water supply of a large city was therefore both a practical and a political matter, and emperors and magistrates advertised their aqueducts as public gifts.
Estimates of how much water Rome received are very uncertain, because ancient figures are hard to convert into modern units and the flow varied through the year, so they are best treated as order-of-magnitude guesses.
Maintenance and decline
An aqueduct needed constant care. Mineral deposits built up inside channels and pipes, springs shifted, tunnels collapsed and bridges cracked. Rome had dedicated teams of workers, and inscriptions and Frontinus’s handbook show that inspection and repair were routine. Water theft and tapping by landowners along the route were persistent concerns.
When administration weakened, the system suffered. During the siege of Rome by the Ostrogoths in 537 CE, the aqueducts were cut, and the decline of the city’s population over the following centuries meant that most were never fully restored; the story is part of the wider picture of the fall of Rome. Some channels did continue to serve their cities, and the Aqua Virgo, first built under Agrippa in the late first century BCE, still feeds the Trevi Fountain in a restored form.