Venice was built by driving short timber piles through soft lagoon mud until they reached a hard clay layer the Venetians called caranto, then laying stone and brick on top of them. The piles are small: 1 to 3.5 metres long and 100 to 250 millimetres across, set close together with sand packed into the gaps between them (Pagella et al., 2025, p. 1035).
Those piles are still in place. Wood sampled from foundations in six districts of the city was dated by its builders' period to between the tenth and the fifteenth century, and short piles of this kind sit under almost every historical building in Venice (Pagella et al., 2025, pp. 1038, 1039). The technique itself was ordinary medieval carpentry. What was unusual was the ground it had to work on, and the reason the timber survived is that it has spent its entire life below the water table, where almost no oxygen reaches it.
How was Venice built?
The sequence runs from the bottom up. Masons drove timber piles through the soft upper lagoon deposits until the points met firmer ground, filled the spaces between the pile heads with sand so the group acted as one block, and then built masonry on that platform. The piles are friction piles, which means they carry load through contact with the soil around them as well as at the tip, and they were deliberately crowded together (Pagella et al., 2025, p. 1035).
Building in the lagoon was not a Venetian invention. A submerged brick structure in the northern lagoon near Lio Piccolo, roughly 6.7 by 7.7 metres and dated to the first century AD, was put up with mortar containing volcanic ash from the Phlegraean Fields near Naples, the material Vitruvius recommends for work in seawater (Dilaria et al., 2024). Roman builders in the lagoon were already importing specialist binders for wet ground.
The city that grew on the Rialto islands from the ninth century was itself almost entirely built of wood, with the exception of churches, monasteries and the ducal palace, and it needed constant infilling both to keep floor levels dry and to widen the habitable ground (Gasparri & Gelichi, 2022, p. 24). Much of the Venice a visitor walks across is reclaimed tidal flat and filled channel (Tosi et al., 2013, p. 6).
What is Venice built on? The mud and the caranto
Under the paving there is very little solid ground. In the subsoil of the city the lagoon deposits, a mixture of silty clay and sandy infills of old tidal channels, are only 2 to 6 metres thick, and the oldest of them dates to about 4,670 radiocarbon years before present (Mozzi et al., 2003, section 1). Across the central lagoon the subtidal mud flat deposits average around 5 metres, thinning towards the mainland, while channel scours can exceed 10 metres (Mozzi et al., 2003, section 1).
Beneath that soft fill lies the layer that holds the city up. Venetian farmers on the mainland used the dialect word caranto for the hard, cemented crusts that turned up in their fields during ploughing, and early geotechnical studies of the Venice subsoil borrowed it: the foundations of the city rest on timber bearing piles driven into the resistant caranto (Mozzi et al., 2003, section 2). The layer is a buried soil, continuous with the calcisols of the plain behind the lagoon, and it marks the surface of a Late Pleistocene alluvial plain that was drowned as sea level rose through the Holocene (Mozzi et al., 2003, conclusions).
In other words, the hard layer Venice is nailed to is a field. Below it, the fluvial deposits laid down since the last interglacial run to about 55 metres (Mozzi et al., 2003, section 1), and the whole Quaternary sequence is roughly 1,000 metres thick (Tosi et al., 2013, p. 6).
What wood are the piles made of, and how long are they?
Venetian piles are short and straight, 1 to 3.5 metres long with diameters of 100 to 250 millimetres, and in some foundations up to about 5 metres (Pagella et al., 2025, pp. 1035, 1038). The comparison with Amsterdam is instructive: Dutch piles of the sixteenth century onward were tapered, 10 to 12 metres long, with head diameters of 180 to 200 millimetres, driven down to dense sand as end-bearing piles (Pagella et al., 2025, p. 1035).
There is no single Venetian pile species. Across 69 samples from six historical sites, including the bell towers of the Frari and Santo Stefano and the church of Santa Maria Maggiore, the wood is mostly alder and larch, with oak, elm and Scots pine appearing less often, and no spruce at all (Pagella et al., 2025, p. 1039). A later campaign of 19 drilled cores from the Arsenal, Riva di Biasio, Rio S. Giustina and Rio dei Greci identified larch, oak, alder, elm and willow in the same role (Pagella et al., 2025, p. 1046).
The timber sits well below the pavement. Those 19 cores came from depths of 1.6 to 2.0 metres at the three canal sites and 3.7 to 7.0 metres at the Arsenal (Pagella et al., 2025, p. 1041). The Arsenal piles themselves can only be placed within a building period of about 1100 to 1600, because repeated rebuilding makes any individual pile impossible to date (Pagella et al., 2025, p. 1045).
Why does the wood under Venice not rot?
Oxygen is the missing ingredient. Wood submerged below the water table in waterlogged soil can be attacked either by soft rot fungi, which need at least a little oxygen, or by bacteria, which work in anoxic conditions. Fungal attack on fully submerged piles progresses very slowly for want of oxygen, so the decay that does happen is mainly bacterial, and bacteria erode the cell walls of the wood fibres far more gradually (Pagella et al., 2025, p. 1035).
That difference in rate is the whole story. Because degradation in waterlogged anaerobic soils is dominated by the slower bacterial process, fully submerged piles can keep working for centuries before their load-bearing capacity falls substantially (Pagella et al., 2025, p. 1035). The popular claim that the piles turned to stone is wrong; they stayed wood, in an environment where almost nothing could eat them quickly.
The qualification matters for anyone reading about Venice and water. Where canal levels fluctuate, the tops of the piles are occasionally exposed to air, and fungal decay then speeds up (Pagella et al., 2025, p. 1038). The system depends on staying wet. Sampling is normally done about 30 centimetres below the pile head, which is itself submerged and below ground level, so none of this is visible from the street (Pagella et al., 2025, p. 1038).
How much strength is left in the piles?
Less than most accounts suggest. Of the Venetian samples tested, most had a basic density between 0.1 and 0.25 grams per cubic centimetre and a residual basic density of 30 to 50 per cent, which indicates significant decay with extensive destruction of the wood tissue (Pagella et al., 2025, p. 1039). Compression tests under saturated conditions gave strengths of 0.6 to 5 megapascals, a reduction of 80 to 90 per cent against a reference value of 23 megapascals for fresh alder (Pagella et al., 2025, p. 1039).
Decay is spread fairly evenly along a pile, and no correlation was found between the depth a sample came from and how degraded it was (Pagella et al., 2025, pp. 1039, 1045). Species matters more than depth. Larch, a softwood with non-durable sapwood, was consistently worse than the hardwoods: all the oak, elm and alder samples in the 19-core set showed only light decay, while moderate to severe bacterial decay was confined to larch and willow (Pagella et al., 2025, p. 1045).
The consequence shows up in tall buildings. Slender Venetian structures carry a risk of structural instability from uneven foundation settlement, and the bell towers of Santa Maria Gloriosa dei Frari and Santo Stefano are the standard worked examples (Pagella et al., 2025, p. 1038). Both lean, and both have been studied for it.
What sits above the piles: Istrian stone, brick and render
On top of the pile heads comes stone, and it is chosen for one job. A typical Venetian Gothic palace, with Ca' Foscari on the Grand Canal as the published example, has a basement of Istrian stone carrying full brick masonry, rendered with cocciopesto and finished with marmorino (Falchi et al., 2018, p. 120). Cocciopesto is finely crushed brick or tile mixed with lime and water; marmorino is powdered marble in a lime base.
The reason for the stone course is capillary water. Measurements pooled from 65 brick masonries in 25 historical buildings across the city centre, covering thirty years of restoration records, show that Venetian brickwork is normally soaked with water up to 120 to 150 centimetres above sea level, with an evaporation zone where salts crystallise out between 200 and 350 centimetres (Falchi et al., 2018, p. 117). A dense stone base is what keeps that wet band from starting in the brick.
Cocciopesto is used throughout Venice as a storage layer for salt, because its pore structure and breathability let it hold salts and moisture without failing (Falchi et al., 2018, p. 120). The system reaches a steady state, since permeable masonry in permanent contact with the lagoon settles at a stable height, although subsidence and sea-level rise have made the conditions harsher over time (Falchi et al., 2018, pp. 118, 124). How wet a given wall is also depends on how close its facade stands to the water (Falchi et al., 2018, p. 117).
Is Venice still settling? What satellite radar finds
It is, slowly. Satellite radar interferometry over 1992 to 2010 and 2008 to 2011 puts the natural subsidence of the historical centre at 0.9 plus or minus 0.7 millimetres a year, while short-term movements attributable to human activity in 2008 ranged from minus 10 to plus 2 millimetres a year, local and concentrated along the margins of the islands (Tosi et al., 2013, p. 1).
The human part is no longer industrial. Groundwater pumping in the mainland industrial zone ended decades ago, and what remains comes from restoration of individual buildings, repairs to canal embankment walls, and the deformation of historical structures themselves (Tosi et al., 2013, pp. 2, 6). Boat and ship wake probably contributes as well, by eroding masonry canal banks and building foundations (Tosi et al., 2013, p. 6).
The foundations explain where the city settles fastest. Local zones sink at 1 to 3 millimetres a year because the shallow subsoil varies, because the city grew by reclaiming tidal flats and channels, and because historical palaces differ in foundation depth and load. Buried sandy channel levees match the sectors that settle least, while compressible silt and clay settle most, and the city as it stood before about 1500 is the most stable part (Tosi et al., 2013, p. 6). The historical centre stands only about 90 centimetres above mean sea level, and the twentieth century cost it some 25 centimetres, 15 from pumping and 10 from sea-level rise (Tosi et al., 2013, p. 1).
