Venice has no standing Roman ruins. The Roman lagoon survives underwater and on the mainland: a first-century well-cistern submerged in the San Felice channel in the north, and the buried town of Altino at the lagoon's landward edge. Both have been excavated and dated in the last twenty years, and together they show a lagoon already in use centuries before the city existed.
The clearest single object is the brick structure known locally as the Torrione San Felice, near Lio Piccolo. Underwater survey in the summer of 2020 reinterpreted what earlier work had called a Roman defensive tower, reading it instead as a well-cistern: a brick shaft inside a squared box packed with river sand to filter rainwater (Dilaria et al., 2024, p. 4). Analysis of its mortar published in 2024 found volcanic ash from the Phlegraean Fields near Naples, the first time that material has been identified anywhere in the lagoon of Venice (Dilaria et al., 2024, pp. 1, 12).
What Roman remains survive in the Venice lagoon?
Nothing Roman stands above water in the historic centre. The islands of Venice were settled later, and the dated Roman material in the lagoon proper is submerged in the tidal channels of the northern basin, in the shallow belt of salt marshes and mudflats between Lio Piccolo and Torcello.
The best studied example is the box-like brick structure in the San Felice channel. Earlier studies read it as a defensive tower of the Roman imperial period, which is where the local name Torrione comes from. Underwater surveys carried out in the summer of 2020 by a Ca' Foscari University team reinterpreted it as a well-cistern, and a comparable structure survives in better condition at Ca' Ballarin along the same channel (Dilaria et al., 2024, p. 4).
On the landward side the picture is different. Altino, a Roman municipium from 49 BC, sits at the northern edge of the lagoon at what is now Quarto d'Altino, and parts of it are excavated, published and open to visitors (Paiano and Delpozzo, 2024, p. 21). For scale, the lagoon these sites lie in covers 550 square kilometres, of which 390 are open lagoon and 70 are salt marsh, with an average depth of about one metre over the shallows and channels exceeding 15 metres (Madricardo et al., 2019).
The Roman well-cistern in the San Felice channel
The structure is a flat, almost square box measuring about 6.7 by 7.7 metres, built entirely from local bricks in both the masonry and the floors. The bricks measure around 43.5 to 46 by 29 to 29.5 by 6 centimetres, the standard sesquipedale size of the Po Valley (Dilaria et al., 2024, p. 4). Associated pottery places the construction in the first century CE or slightly later, so the date is archaeological rather than radiometric.
The design explains itself once you see it in section. A circular well shaft lined with brick or stone sits inside the squared masonry box, and the space between the two is filled with river sand. Rain falls on the sand, the sand filters it, and clean water collects at the bottom where the shaft draws it (Dilaria et al., 2024, p. 4). Given the size of the thing, the study suggests it was built to supply fresh water to ships working the lagoon route rather than to serve a single household (Dilaria et al., 2024, p. 23).
Anyone who has crossed a Venetian campo has walked past the descendant of this design. The authors describe the structure as an ancestor of the pozzo alla veneziana, the sand-filtered rainwater well developed in the city during the medieval and modern periods (Dilaria et al., 2024, p. 4).
Volcanic ash from the Bay of Naples in a lagoon wall
The mortar is what made the site worth publishing. It contains pyroclastic aggregates, pumice and volcanic glass that are inconsistent with the local geology and were added deliberately as natural pozzolans, reacting strongly with the lime binder (Dilaria et al., 2024, p. 1). The obvious local candidates, the volcanic rocks of the Veneto quarried in Roman times, were tested and ruled out (Dilaria et al., 2024, p. 12).
Discriminant analysis of the pumice chemistry returned a first-probability correlation of 75 to 100 per cent with post-Neapolitan Yellow Tuff products of the Phlegraean Fields, the volcanic district west of Naples (Dilaria et al., 2024, p. 20). That is the pulvis puteolanus Vitruvius recommended for building in seawater, named after Puteoli, modern Pozzuoli. Its identification here is the first anywhere in the lagoon of Venice, and it suggests the lagoon was plugged into the supply network for the most specialised building material the Roman world had, probably carried as ship ballast on the inland route linking Ravenna with Altinum and Aquileia (Dilaria et al., 2024, p. 23).
The masons knew what they had. Concentrations of volcanic aggregate vary between samples, with the highest amounts at the nodal joints, which argues for careful calibration rather than ballast tipped into the mix (Dilaria et al., 2024, pp. 23-24). The mortar also carries ground iron slag from metalworking done near the site, reused as an aggregate (Dilaria et al., 2024, p. 12).
How the Romans built underwater, and why the mortar lasted
The first course of bricks now sits at about 4.20 to 3.75 metres below mean sea level. Even allowing for the lower Roman sea level, the base stood at roughly 2.0 metres below the surface when it was built, which means at least the foundations and probably much of the structure went up underwater (Dilaria et al., 2024, p. 20).
Laying bricks on mortar joints at that depth is hard to imagine without a dry working space. The study argues for cofferdams, probably of the double-bulkhead type Vitruvius describes, with machinery for lifting water out of the enclosure (Dilaria et al., 2024, pp. 20-22). This is also why the imported ash mattered. A lime mortar that sets only by drying in air is useless in a flooded trench, while a pozzolanic mortar hardens in contact with water.
The chemistry then worked in the building's favour for two thousand years. In the submerged, alkaline and low-oxygen conditions of the brackish lagoon the pozzolanic reaction continued while ordinary carbonation was suppressed, and the phases that formed kept adding strength over time (Dilaria et al., 2024, p. 22). The structure lasted because it stayed wet.
How much has the sea risen in Venice since Roman times?
The well-cistern is useful beyond archaeology because it is a datum. It was built to work at a particular height relative to the water, and it has not moved since.
Today it lies at a depth ranging from 4.5 to 3.0 metres below mean sea level, with its top standing about 1.5 metres above the floor of the channel (Dilaria et al., 2024, p. 4). Reconstructing the Roman relative sea level for this part of the northern lagoon puts it about 2.15 metres lower than today, which places the base of the structure at roughly 2.0 metres below the surface in the first century CE (Dilaria et al., 2024, pp. 4, 20).
Two things are worth keeping straight about that figure. It is relative sea level, so it combines the rise of the sea with the sinking of the ground, and the 2.15 metres comes from earlier work on this stretch of lagoon rather than from measurement on the structure itself. It is also local. The lagoon floor around it has been heavily reshaped in the modern period, with about 7.8 square kilometres, roughly 24 per cent of the tidal channel surface, directly modified by dredging and other human activity (Madricardo et al., 2019).
Where you can see Roman Altino today
The lagoon sites are not visitable. The San Felice well-cistern is a submerged structure in a working tidal channel, recorded by divers rather than exposed for visitors, and there is nothing to see from a boat.
Altino is different. The Museo Archeologico Nazionale di Altino stands at Via San Eliodoro 56 in Quarto d'Altino, on the mainland at the top of the lagoon, and holds the finds from the Roman town. It opens Tuesday to Saturday from 8am to 7pm and on Sunday from 2pm to 7pm, closed on Monday, with full admission at 5 EUR and a reduced ticket at 2 EUR. Two archaeological areas lie about 500 metres from the museum, one with a city gate and an urban road, the other a residential quarter including the Domus della Pantera and its black and white mosaic floor. They open from April to October and close in bad weather (Ministero della Cultura, 2026).
What excavation found there was a water city. A canal on average 25 metres wide ran through the monumental quarter for at least 360 metres, a quay of squared sandstone blocks stepped down to the water with mooring posts tree-ring dated to between 80 and 70 BC, and the harbour basin measured about 139 by 41 metres in its main arm (Paiano and Delpozzo, 2024, pp. 26-29). The see moved from Altino to Torcello in 639, a date fixed by an inscription found in the crypt there in 1895 (Paiano and Delpozzo, 2024, pp. 30-31).
