Rome and Salcott 2008 350


photo – Steve Coe


No resource is more important to any City than a guaranteed supply of clean water. This truism is aptly summed up in the words of the first century engineer and architect, Vitruvius Polio, who said “water is very necessary for life, for delight, for daily use”. Securing an adequate water supply for Rome was a serious problem. Rome is a city astride a large river, but it is also a city of hills and the effort to haul water from the river up to a population estimated to have peaked at over a million and spread over several square miles, would have been considerable and would have required armies of slaves, mules and wagons.


Actually, the city has always been fairly well endowed with springs and wells, but these localised sources would have been insignificant for supporting a large and rapidly growing population. The Tiber was also the recipient of the city’s discharging sewers and there would have been similar polluted run-off from settlements upstream, such as Tivoli. So the waters of the Tiber would have certainly been contaminated and potentially lethal as a  source of supply.  So the ability to import large amounts of clean water from the nearby countryside (the foothills of the Apennine mountains), using the technology associated with the aqueducts, was actually essential to the continued survival, health and growth of the ancient city.



Rome, for all the success of her mighty armies, would have been hard put to grow to a size, in keeping with the demands of administering and providing a focus for its far-flung empire, without the aqueducts and the clean water they supplied. The aqueducts must be recognised, not just as architectural and engineering phenomenons, constructed far ahead of their time, but as the vitally important facilities they represented to the City, (and hence to the empire).



photo – Steve Coe


It is illustrating to get a modern perspective on the scale of the water demand that has to be met in supplying the basic needs of any city of over one million inhabitants (as has been reliably estimated for Rome’s population during the Augustan and Antonine eras).  This Rome of the first and second centuries AD was an incredibly huge city by ancient standards and, unlike the cities of classical Greece, was already a highly urbanised commercial and administrative centre, with a large percentage of the population living in multi-storied tenement blocks (insulae) and most of the city’s supply of food and other consumable goods (clothing, olive oil, firewood, wine etc) already being imported from a considerable distance away. In these aspects and many others, Rome had many of the attributes of a modern city, rather than those of another more typical fair sized ancient city-state such as Athens or Antioch.


In fact, with its large and rapidly growing, recently urbanised population of poor citizens (many unemployed), living in close proximity to an affluent minority, this ancient city must have had a fairly similar dynamic to some of the more prosperous cities of the modern developing world, such as many of those found today in Southern Africa, Latin America and elsewhere. Therefore, while making allowances for technological differences, it could be useful to consider water consumption patterns for this type of modern first-world/ third-world city.


In a country such as South Africa, for example, municipalities would typically cater for domestic water consumption on the statistical basis that a modern middle-class suburban household would be using up to 200 litres of water per resident per day (in the USA 300 – 400 litres per day). Sub-economic township dwellers ( usually having to use public standpipes and fetching water by hand), would be expected to use only around 50 litres per person per day. So at a poor/middle class population mix of a 3:1 ratio in South Africa, a weighted mean of about 80 litres per capita, per day would be expected (this is excluding specific heavy industrial and agricultural demand).

Rome and Salcott 2008 302


photo – Steve Coe



The estimates of the actual delivery capacity available to the Population of Rome from the aqueduct system, is a contentious problem that has vexed scholars for decades and is still not satisfactorily resolved, though most publications have settled on a million cubic metres per day (1000 mega litres) – a figure that I believe is too high. However, it was probably not less than 600 megalitres per day, giving a per capita capacity of between 600 and 1000 litres per day to every man woman and child. So we are still talking about a huge amount of water and with only nine aqueducts supplying the bulk of the supply at this period the aqueducts were certainly hard working facilities.


How was this huge water supply physically brought to the city and managed for subsequent distribution? By the end of the first century AD most of the infrastructure was already in place and many kilometres of solidly built aqueduct channels were bringing water into the city. However, contrary to popular opinion, only a relatively small portion of Rome’s aqueducts consisted of the famous arcade bridges. Most flowed discretely through hidden subterranean culverts, in tunnels cut into the soft tufa of the area, where waters were kept cool and safe, emerging from time to time to make their way along valleys and around the sides of hills, mostly as contour-hugging, roofed channels, resting upon simple masonry or concrete platforms called substructures.



Photo – Steve Coe


The impressive iconographic arcades were hugely expensive to construct (and very vulnerable to potential hostile forces). They were therefore only used near the approaches to the city itself for the specific purpose of maintaining water head for delivery to the higher lying city districts, or otherwise they were used to cross deep valleys that occasionally lay across the aqueducts’ route (although inverted lead piping siphons were also sometimes used for this latter purpose).



The water management system and the distribution network itself should be given some comment. The separation of the better supplies, sourced from the clearest spring water for priority for drinking water was of great concern, as was the difficulty of providing of supply to high lying districts. A major consideration of the planning was to arrange the various secondary supply channels in such a configuration, that alternative routing and sharing was possible, as a form of backup, so that lines could be isolated and cut off for repairs and maintenance.



Usually, whether the sources were from springs or from rivers, the waters were diverted and collected into catching basins, before being led off into the main delivery channels or conduits known as rivus[1] or specus. The Aqueduct conduits themselves had very strict specifications, both for the cross sectional dimensions of their channels (they had to allow head room for maintenance gangs) and for the gradient of their beds in the direction of flow. The specifications also catered for such issues as the provision of aeration shafts to prevent gas build up, as they were mostly underground.


If not cut directly into solid bedrock, the conduits were usually constructed from a combination of cemented masonry and in-situ concrete. Masonry structures and conduits through more porous rock were mostly given a waterproof lining of a specially prepared pozzolanic concrete (opus signinum), though sometimes this was substituted with a skin of baked clay tiles. Subsidiary pipes or conduits (ramus) were also used to share/ supplement water between aqueducts or their reservoirs and also for cooling.


Rome and Salcott 2008 258


Photo – Steve Coe



The water brought in by the aqueducts was managed through various types of reservoirs or tanks that had different functions, although sometimes several of these functions could be combined in a single reservoir for practical convenience. We have mentioned the catchment basins, where the source was ponded and diverted into the aqueduct channels. Secondly there were deep stilling tanks or cisterns (piscinae), used to slow down the rapid water flow at suitable points along the aqueduct route. The drop in water velocity allowed the incoming water’s load of sediment to settle out and be removed, before it reached the distribution network.


Within the city,  there were smaller tanks, used for managing and sharing water between complementary aqueduct systems, where they ran close together. Larger balancing and storage reservoirs (castella) were placed on high land and were used to even out the flows in a particular network these castellae could also function as splitter reservoirs where they were used to divide up the waters between different types of end use – public fountains, state use, baths or for private (tariffed) usage. Typically these reservoirs were compartmentalised by dividing walls, with overflow weirs, set at specific levels to establish consumer priority. Vitruvius explains the law and practicalities involved in the method of apportioning and controlling the various stake holders’ vested rights of delivery (Vitruvius: De Arch. 8.6.1-2).


Copy of Rome and Salcott 2008 339


Photo – Steve Coe



Water for private usage was tapped off, through a restricted tapping nozzle of standardized size and length called a calix and then delivered either through a lead or baked clay pipe (fistula) directly to the household. These calices were set at a constant depth (head) below the water surface, which was kept to a constant level by an overflow weir. As the head (pressure) was constant and the calices diameters were also standardised, the outflow was thus predictable and always the same for any particular size of calix. The flow relative to each size of calix was measured as a quantity of water in a measure called a quinaria. In a major first century AD treatise, Julius Frontinus a famous senator and water commissioner, goes into considerable technical detail in describing and categorising these calices and calculating the amount of water they delivered (Frontinus: De Aquis 1,24 – 63).




Photo – Steve Coe



(Six centuries of construction excellence)


Although all were part of an integrated system, taken individually each one of the aqueducts of Rome may yet be regarded as an entity in its own right. Each had a name, a history and specific characteristics of function, form and specific areas of the city that it served; moreover the aqueducts were also distinguished by their supply capacity and the quality of the water that they could deliver to the city. Construction of aqueducts was spread over the course of six centuries starting from the early republican era.  Aqua Appia was the first, built in 312 BC, by the consul Appius Claudius (who also built the Appian Way).


Construction of aqueducts continued sporadically until Alexander Severus (AD 222-235) commissioned the last one – the Aqua Alexandrina. He built this specifically for his new baths in 226 AD. The longest aqueduct was the Anio Novus, with a length of nearly 100km, sourced from the upper reaches of the river Anio. The construction of this aqueduct, together with its near contemporary, the Aqua Claudia which shared part of its route and arcade, virtually doubled the delivery capacity supplied by the previous seven aqueducts. Less is known of the Alexandrina and the Aqua Traiana (built by Trajan in 109 AD (constructed after Frontinus’s death).


Any modern engineer studying this extremely sophisticated water supply system cannot help but be astonished at the unexpectedly high level of the technological expertise of the ancient Romans water engineers. And he would also be completely taken aback to see how in many ways, so little has changed in the way that we deal with the basic problems of supply and reticulation. They were constructed through every possible kind of terrain, and although lacking the obvious advantages of modern construction equipment and materials, were completed in an incredibly short time, even by today’s standards.


Rome and Salcott 2008 326


Photo by Steve Coe


Rome’s water system was  an incredible triumph of effective project management. The overall strategic thinking; the optimising of resources; the systematic methodology behind the structural designs and the careful planning and routing of the aqueducts and the distribution network they supported, was essentially what we would now deem, in our vanity, to be a “thoroughly modern approach”





The aqueduct system kept Rome clean and healthy for centuries and the Roman tradition was to be closely echoed a millenia and a half later in a sudden revival of interest in the provision of clean drinking water to cities and towns during the early Victorian era, when engineers emulated the Romans closely . Today the availability of clean drinking water to cities and towns, together with another Roman invention – the water bourn sewer, is simply taken for granted. Yet always working away from the limelight,  the humble water engineers, since Roman times, have made a far greater contribution to public health and saved many more lives in total than the much vaunted medical profession has ever done.


AQUA VIRGO –    Trevi Fountain


Rome and Salcott 2008 078

THE TREVI FOUNTAIN - still fed by the crystal clear water of Aqua Virgo after 2000 years - True Magic!

Photo by Steve Coe


At the beginning of the first century AD, a party of  soldiers were out in the countryside, scouting for water for their general, Marcus Agrippa, the famous friend and son-in-law of the first emperor, Augustus. Agrippa, a fine administrator and engineer in his own right, had recently been appointed water commissioner for Rome. According to tradition, a young girl showed the soldiers a hidden spring of the purest water. The aqueduct that Agrippa built to bring this water to Rome was subsequently named the Aqua Virgo, in honour of the young girl. This aqueduct was entirely built underground and so survived intact when the barbarians eventually attacked and sacked Rome. The water still flows strongly today and its water feeds the magnificently sculptored Trevi Fountain (commemorated by the well-known song – “three coins in a fountain“)


AQUA FELICE:    A 16th century ‘add-on’

( still delivering substantial water after five hundred years)



Rome and Salcott 2008 314

AQUA FELICE: Constructed upon Aqua Claudia’s Arcade - Location: Aqueduct Park, near Casale di Roma Vecchia:

Photo – Steve Coe


In 1585 Pope Sixtus V built the ugly but functional concrete aqueduct – the Aqua Felice, called after his family name (Felix). To save on expense, he used large portions  of the ancient arcade structure of the Aqua Marcia as a base to support his new aqueduct and he also tapped into the ancient sources used by the Roman engineers.  It was hugely successful and was largely instrumental in revitalising the city of Rome, which had been in a ruinous state for many centuries. The aqueduct is still delivering a substantial amount of excellent water to Rome to this day.



One can only come away from a study of the unique and wonderful Roman water supply network with a sense of awe at the magnificent achievement of those ancient engineers and administrators, who set up such a sophisticated system and on such a grand scale.

As Rome’s most famous water commissioner, Julius Frontinus, was reported to have said –


Remembrance will endure if the life shall have merited it.





Name Date Built    % Paces (km) Source / Terminus
Aqua Appia 312 BC 7% 11190 Source: Springs near Praenestian
(built by Appius Claudius) 16.2 km               Way on Lucullus  estate
Terminus – Forum Boaricum
Aqua Anio Vetus 272 BC 18% 43000 Source – River Anio
(built by Commissioner Flaccus)  62.4 km               above Tibur
(supplemented by Marcia)
Aqua Marcia 144 BC 19% 61710 Source – springs Valerian Way
(built by Praetor Marcius) 89.5 km
(Supplemented by tributary Terminus – Capitoline
 branch called Augusta)
Aqua Tepula 125 BC 2% 15426 Source – springs Latin Way and on
(Censors Caepio & Longinus) 22.4 km Lucullus estate
(also supplemented by Julia Frontinus does not give length – so taken same as Julia (same springs that later fed Julia -plus warm spings Alban Hills
Terminus – at Capitoline and city
Aqua Julia 33 BC 5% 15426 Source – springs Latin Way
(built by Agrippa) 22.4 km                on Lucullus estate
Aqua Virgo c19 BC 10% 14105 Source – springs Collatian Way
(built by Agrippa) 20.5 km Lucullus estate
Aqua Alsietina * c2 BC 2% 22172 Source – Alsietinian Lakes
(also called Augusta) 32.1km Lago di Martignano and
bad water – only for Naumachia Lago di Bracciano
or irrigation purposes Claudian Way North of Tiber
Aqua Claudia    AD 38-52 19% 46606 Source – Springs Caerulan &
(Started Caligula finished by Claudius) 67.6 km Curtian (also Albudinus) along
supplemented by tributary branch Sublacensian Way
called Augusta)
Aqua Anio Novus AD 38-52 19% 58700 Source – River Anio
(Started Caligula finished by Claudius) 85.1km and springs
TOTALS 100% 418.1 km
Aqua Traiana AD 109 Length? Source – Springs near Treviglani
Built by Trajan North of Lago di Bracciano
Aqua Alexandra AD 226 Length? Source – springs south of Praenestina
Built by Alexander Severus



  • Although the technical descriptions are all mine and were derived from a number of sources, including the one mentioned below,  most of the Latin names used here for specific waterworks items or structures, such as “revus”, “calix” or “castella” were taken from the Bowdoin College Website – introduction to   Archaeology 291 “The Roman Aqueducts and Water Systems”.
  • Vitruvius was well aware of the dangers of lead poisoning from lead pipes – see De Arch Book 8
  • Quinaria were actually units of area not of quantity, although  they were incorrectly used as such by  Frontinus  – this leads to major problems in making realistic evaluations of  capacity.
  •  It took only three years to construct the 62km long Anio Vetus, built 272 -269 BC.

.                                                        MORE ROMAN PAGES  ROMAN BATHS





Your email address will not be published. Required fields are marked *


7 + = eight