Titus inaugurated Rome’s Flavian Amphitheatre in AD 80. The building, now known as the Colosseum, was intended for gladiatorial contests and animal hunts, yet the inauguration also included naval spectacles. Cassius Dio, writing later, says that Titus flooded the amphitheatre and brought ships into it. His account leaves a difficult question: how could a stone arena built for combat on dry ground accommodate a battle on water?
These staged naval battles were called naumachiae, a term drawn from the Greek for a sea battle. Roman rulers had arranged them for more than a century, generally in excavated lakes where vessels had space to move. The Colosseum was a very different setting. Establishing whether it briefly held water also means asking what ships and fighting could have fitted inside its confined arena.
The evidence supports a brief early period of water spectacles, which the authority responsible for the Colosseum also associates with its inauguration. The spectacle itself is better attested than its workings. The water depths, rapid filling mechanisms, miniature warships and choreographed boarding fights often described in modern accounts are largely attempts to reconstruct a process that the ancient authors never explain in technical terms.
Rome was staging naval battles long before the Colosseum
Julius Caesar staged a naval battle during his celebrations in 46 BCE, long before the Colosseum existed. A basin had been dug in a low area of Rome for the occasion. Suetonius, Life of Julius Caesar 39.4 records Tyrian and Egyptian opposing sides and describes ships with two, three or four banks of oars, familiar classes of Mediterranean warships. Their appearance does not establish that the performance reproduced one specific historical engagement. Caesar had emerged victorious from civil war and other campaigns, and the resources required to assemble crews, vessels and spectators advertised his power. Suetonius says visitors came from distant places, some sleeping in tents beside the roads, and that people were crushed in the crowds.
Constructing Caesar’s basin was a substantial operation in its own right. Organisers needed vessels, a supply of water and space for an enormous audience. Cassius Dio’s account of Caesar’s games places an excavated and flooded site in the Campus Martius. Neither Dio nor Suetonius tells us exactly how deep the installation was or gives a complete plan of its position. What the episode establishes is that Romans could arrange large-scale naval entertainment before the city possessed a permanent stone amphitheatre.
In 2 BCE Augustus, Caesar’s adopted son and Rome’s first emperor, staged an even larger display across the Tiber. Augustus gives the basin’s measurements in his Res Gestae, section 23: 1,800 by 1,200 Roman feet, roughly 530 by 355 metres. He also reports thirty ram-equipped biremes or triremes, smaller vessels and around 3,000 fighting men in addition to rowers. These numbers belong to Augustus’s basin, decades before the Colosseum’s opening. His decision to record them among his achievements shows how heavily such spectacles depended on imperial expenditure, organisation and control of manpower.
Augustus counted fighters separately from the men at the oars. A trireme relied on a coordinated rowing crew to steer and manoeuvre, while armed men fought from its decks or boarded an opponent. Those distinct roles could make a staged encounter resemble naval warfare. The ancient accounts, though, seldom reveal who received training, which participants were forced to appear, or how much practical seamanship was expected of them.
Claudius provided an especially revealing precedent in AD 52, when he arranged a sea battle on Lake Fucinus east of Rome. The occasion was associated with the attempt to drain the lake through a tunnel. Tacitus, Annals 12.56 says Claudius assembled 19,000 men and equipped triremes and quadriremes. Guards and rafts prevented easy escape, although there was still room on the lake for vessels to row, turn and strike one another. Tacitus identifies the combatants as condemned men and reports that many were killed before the survivors were spared additional slaughter. His account makes clear both the human cost of these performances and the advantage of a lake over a narrow amphitheatre floor.
Under Nero there was also a naval show in an amphitheatre. Suetonius, Nero 12 describes salt water and marine creatures as part of the entertainment, held in Nero’s temporary wooden amphitheatre in the Campus Martius. It cannot have taken place in the Colosseum, which had not yet been built. These precedents help explain why spectators in AD 80 might have regarded a flooded arena as an ambitious extension of an established form of imperial display. They do not settle how the new building was adapted to hold water.
What happened at the Colosseum in AD 80?
Cassius Dio, Roman History 66.25 offers the clearest ancient statement that the Colosseum itself was flooded. According to Dio, Titus filled the newly opened amphitheatre with water, displayed trained horses and bulls there, and then staged a fight between ships representing Corcyraeans and Corinthians. Corcyra, now Corfu, was a Greek island whose fifth-century BCE quarrel with Corinth helped lead towards the Peloponnesian War. The historical reference gave the performance recognisable opposing sides, whether or not the Roman audience knew the earlier conflict in detail.
Dio describes a second naval spectacle as well, but places it outside the city at Augustus’s earlier basin. That event involved 3,000 combatants, fighting between groups representing Athenians and Syracusans, and a subsequent battle on land. The events at the two venues are easily confused. Dio does not put the 3,000 fighters, the island assault, or the Athenian and Syracusan encounter inside the Colosseum. He treats the animal displays in the amphitheatre and entertainments at the older basin as separate parts of a larger inauguration programme.
Dio was writing more than a century after Titus and drew on earlier materials that are not all extant. For a nearer account, Suetonius, Titus 7.3 also records a naval battle among the opening spectacles. He locates it in the old naumachia, however, meaning the purpose-built basin, and does not explicitly say that Titus flooded the amphitheatre. Both narratives can fit together if celebrations took place at two sites, as Dio describes. Suetonius’s account by itself provides no confirmation of a Colosseum flood.
Martial adds the testimony of a poet who lived in Flavian Rome and celebrated the games. In On the Spectacles 24, he speaks to a late-arriving spectator who finds ships floating where dry ground had stood shortly before, and says the ground will soon be dry once more. The poem shows that a sudden change between land and water formed part of the spectacle’s appeal. Martial was also praising an emperor, using the transformation to magnify imperial generosity and power. He says nothing about pipes or depths and does not identify the venue with enough precision to settle the argument on its own.
Naval entertainments formed part of Titus’s inaugural games, and Dio explicitly places some inside the Colosseum. They tell us far less about the sequence of work involved. Full-sized war galleys circling the oval, followed minutes later by gladiators fighting on perfectly dry sand, belong to modern visualisations rather than to a documented programme.
Domitian and the last water spectacles in the arena
Titus died in AD 81, and his brother Domitian succeeded him. Suetonius, Domitian 4 says Domitian mounted a naval battle in the amphitheatre. The same passage mentions separate naval displays in a newly excavated basin near the Tiber, fitted out with seating. As in Dio’s description of Titus’s celebrations, an amphitheatre performance was only one possible setting for a Roman naval show.
The dates of these early displays do not allow us to trace a long-running tradition of flooding the Colosseum. Construction continued after AD 80, and Domitian’s changes transformed the space beneath the arena. The Colosseum Archaeological Park’s account of the underground levels describes how an earlier timber arrangement gave way to permanent masonry. Fourteen corridors, working rooms, lifts, ramps, trapdoors and other machinery eventually made it possible to raise animals and performers into view. The park connects the original converging passages with the inauguration’s naval spectacles and places the end of water performances at the introduction of the permanent masonry system.
A removable timber floor could permit temporary arrangements that a fixed underground complex would prevent. Removing sections might expose a basin below, while masonry walls, equipment and storage rooms left little room for a large body of water. The surviving hypogeum, the network beneath today’s arena, belongs to a later phase and cannot simply be projected backwards to AD 80. Nor do those later underground walls prove that a waterproof deck originally covered them all. Such a platform remains a proposed reconstruction.
It is sometimes suggested that the flooded battles ended because spectators lost interest or emperors decided they were too expensive. Neither explanation is securely established. We do have evidence for an architectural change that made flooding impractical, and the archaeological park associates that change with the end of the shows. Romans continued arranging naval spectacles elsewhere. The disappearance of water displays from this building was not the end of the naumachia as an entertainment.
How could the Romans have filled and drained it?
Where the Colosseum held its water remains an engineering problem. Popular descriptions imagine sealing the upper arena floor and pouring water across it. Such a pool would place enormous weight on the floor while needing to keep the rooms below dry. Archaeologists have also proposed removing parts of the original timber decking to fill a basin lower down within the building’s foundations. That possibility makes the difference between the early wooden structure and the later masonry hypogeum especially important. Neither arrangement survives in enough detail to reconstruct a complete and verified ancient plan.
Heinz-Jürgen Beste’s 2019 report for the German Archaeological Institute, drawing on investigations in 2017 and 2018, sets out the fullest specific proposal. Beste recorded thirty-two vertical shafts opening into niches in the inner perimeter wall, linked through subsidiary passages to a circular drainage channel. He also describes traces of an earlier movable timber support system and water-resistant cocciopesto plaster on part of the perimeter. The shafts, channel and plaster are physical observations. Explaining them as parts of one managed flooding system is Beste’s interpretation of their purpose.
In that reconstruction, water would enter along the ring channel and move through thirty-two passages to outlets around the basin. Adjustable barriers might let the pool fill before water was drained away, while jets from the openings could themselves contribute to the display. Using assumed gradients and flow capacities, Beste calculated that a depth of about one metre might be reached in roughly twenty-two minutes. No Roman account gives that filling time. It is the result of a modern model, and the original system has not been recovered in working condition.
The proposed mechanism still lacks a securely identified source of water. The Colosseum had fountains, demonstrating a supply, but nobody has established a definite aqueduct connection feeding a battle basin. Beste considers possible routes from the Aqua Claudia on the Caelian Hill and from the vicinity of the Baths of Titus. Nor is there a surviving high-water mark to settle the pool’s depth. He estimates that, to avoid flooding adjoining corridors, his arrangement could hold no more than about 1.2 metres. Shallow-draft river boats might operate at that level; it would be unsafe to assume that full seagoing war galleys could.
Estimates of the filling time differ sharply. Martin Crapper’s 2007 hydraulic study assumed different capacities for the basin, water supply and drainage and found that filling might have taken roughly two to five hours, with draining taking a similar period. That conclusion is reported in the study’s repository record and published abstract. Crapper and Beste did not measure an intact Roman system under identical conditions, so their figures cannot be combined into one established rate. They show ways in which flooding might have been feasible, without identifying what Titus’s workers did in AD 80.
The Colosseum replaced a lake in Nero’s palace grounds, the Domus Aurea. Water management was already part of this valley’s infrastructure. Paolo Storchi and Ilaria Trivelloni, in a 2023 archaeological study of water in Roman entertainment buildings, discuss that setting alongside traces of hydraulic mortar. They caution against treating every nearby watercourse or drain as proof of naumachiae: amphitheatres also required water for fountains, cleaning, rainwater disposal and routine use. A workable drainage system is relevant evidence, but it does not independently confirm a naval battle.
Whatever the design, a change from dry arena to pool would involve substantial work. The timber floor had to be moved or adapted; craft and participants needed access; water flow had to be managed without cutting off essential passages; and the space eventually had to be made ready for another kind of performance. Ancient writers celebrate the effect upon spectators, but provide no crew numbers, conversion timetable or frequency of flooding. Claims that the operation took place weekly, or was routinely completed within an hour, cannot be traced to secure ancient evidence.
What kind of ships and fighting could fit inside?
The floor of the Colosseum measured about eighty metres in length and under fifty in width. Those dimensions readily accommodated gladiators and animal hunts, yet left little room for warships to manoeuvre. Augustus’s basin and Lake Fucinus were large enough for approaching fleets and collisions. A sizeable vessel in the Colosseum would have struggled to turn, and the oars of smaller boats could readily foul the walls or one another.
Dio mentions fighting from ships without describing their hulls. If Beste’s shallow-water reconstruction is broadly right, small craft drawing little water would be the practical choice. They might have been decorated to suggest larger warships, bringing armed men close enough to fight on board or at the edge of the pool. To the spectators, such a close-quarters encounter might have looked like a naval battle. The number and size of the vessels, the reach of their oars and the presence of any special boarding bridges cannot be established from the surviving accounts.
The men who fought in Titus’s water shows remain largely unidentified. Other naumachiae could involve captives or condemned criminals, as Tacitus’s account of Claudius’s spectacle illustrates, but that does not tell us the legal or social status of every Colosseum rower and fighter. Even Roman gladiators differed in training, status and exposure to danger. No roster of these particular naval combatants survives, and neither a standard death rate nor promises of pardon or a specific recruitment system can responsibly be attached to them.
Tacitus’s description of Lake Fucinus includes armed guards, siege machinery and men fighting violently, so some Roman naval spectacles certainly used weapons capable of killing. Whether archers, thrown spears, swords or some combination appeared in the Colosseum is unclear. Missiles, attempted damage to enemy ships and fighting at close quarters belonged to real ancient naval warfare, but all were constrained by the size of a flooded arena. Depictions of ships ramming at full speed or combatants crossing elaborate mechanical bridges go beyond the available evidence unless supported by a specific source.
Stories about sharks swimming around the Colosseum have no support in ancient testimony. Dio describes trained horses, bulls and other animals; Suetonius’s reference to marine creatures concerns Nero’s earlier salt-water performance in a different building. Combining the two creates an episode neither author records. The same care is needed with named battles. Roman naval re-enactments sometimes invoked Salamis, the Greek victory over Persia, but Dio assigns the amphitheatre encounter to Corcyraeans and Corinthians, not Salamis.
What would spectators have seen?
The Colosseum’s stepped seating, controlled entrances and assigned places allowed an immense audience to look down on a single arena. A handful of boats on water could be startling in a building usually associated with solid ground. Dio’s account supports a scene with moving vessels, armed people above the water and trained animals performing within the transformed space. It does not supply the sound of a particular oar stroke, the colours of the banners or an emperor signalling a sentence of death. Those features belong to the imagination of later storytellers.
Martial shows why the sudden change of setting was worth celebrating at court. The poet presents an emperor capable of turning an arena from land to water and back again. Behind the performance stood the Flavian dynasty’s recent political history. Vespasian and his sons gained power after the civil wars of AD 69, and their new amphitheatre occupied land previously devoted to Nero’s private palace pleasures. Opening that space for enormous public entertainments, including water shows, advertised the dynasty’s ability to remake Rome and spend its resources on the people.
Imperial resources also included the power to compel people to take part. The state could gather engineers, sailors, builders and crowds, and it could force men to risk their lives as entertainment. Claudius’s lake show provides the clearest evidence of the suffering and coerced status of some naumachia participants. It helps establish what this Roman form of spectacle could entail, while leaving the identities and experiences of the people in Titus’s Colosseum show unresolved.
Later artists had a powerful influence on how these events have been pictured. Seventeenth- and nineteenth-century images show vast warships crashing together in front of packed amphitheatre stands. They document the tastes and historical imagination of the artists’ own periods. They cannot tell us how many boats entered the Colosseum in AD 80 or how its arena looked at the time, centuries before any of these images were made.
Why the flooded Colosseum remains difficult to reconstruct
The surviving writers describe different parts of the festivities. Dio puts a naval battle inside Titus’s flooded amphitheatre. Suetonius places Titus’s naval entertainment at the older purpose-built basin, while also mentioning an amphitheatre battle under Domitian. Martial praises an apparently sudden transformation from dry ground into water. These writers were recording history or celebrating imperial achievement, rather than leaving engineers precise instructions. Their accounts provide valuable testimony without resolving every venue and technical detail.
The surviving masonry and fittings help test the written accounts. Early timber supports, perimeter shafts, drainage channels and water-resistant plaster are genuine features of the structure. Beste and other investigators have used them to propose a workable flooding arrangement. The inlet, water depth, exact method of sealing, boat types and working schedule remain uncertain. Domitian’s later masonry hypogeum changed what could happen beneath the arena and destroyed or obscured parts of its earlier configuration.
The official archaeological account accepts that the Colosseum hosted water spectacles in its early Flavian years, and the surviving evidence makes that interpretation credible. The surviving record leaves the battles’ scale and hydraulic engineering open to question. For spectators used to gladiators and animals performing on sand, a temporary basin with armed boats inside the newly completed amphitheatre would itself have demonstrated the extraordinary resources Titus could command, whatever the actual size of the vessels.
Discussion
Comments