Exactly

HOW WAS ROMAN CONCRETE MADE?

Burn the limestone, slake it or don't, add volcanic ash and pack it in with rubble. How Roman concrete was made, from Vitruvius to the walls still standing.

Baiae, where Vitruvius says the powder is found, in a photograph by Giorgio Sommer: the ruins the print calls the "Tempio di Venere & Diana", by the water. Part of the Roman town now lies under the sea. Rijksmuseum, Amsterdam.

How, exactly? · Lime, volcanic ash, rubble and not much water: the proportions Vitruvius wrote down, the ones in the walls still standing, and what is quick about quicklime.

The short answer. Vitruvius wrote down the recipe for Roman concrete, and it can be made again. You need lime, volcanic ash, rubble and a little water. For harbours, two parts of ash to one of lime, mixed with seawater; in the concrete that has survived there is a little more ash, about two and three quarters. The lime was burnt in a kiln and used slaked, or else as quicklime, mixed dry with the ash. Nobody poured it. It went in stiff, packed between the two faces of a wall or into a timber box driven into the seabed, and the rubble was pressed into it. Time did the rest.

That is the short answer. The one you usually read fits on a single line.

The textbook answer

Lime, volcanic ash and seawater, plus a few lumps of volcanic rock: that is the recipe you usually find, for instance in the press release the University of Utah put out in 2017 for a study of Roman piers. For harbours it is right. Everywhere else it leaves things out: on land seawater has nothing to do with it, the proportions matter, the lime could go in quick or slaked, and the rubble was not always volcanic. And Vitruvius, who actually wrote the recipe down, says nothing about water.

The recipe

Vitruvius puts it down in black and white, the way a recipe is written at home: so many parts of this, so many of that, and the water left to whoever does the mixing. Here it is, with his proportions and, where he gives none, the ones scholars have found by measuring the walls and piers that are still standing.

The recipe

Roman concrete, two ways

For a wall, in Rome

Proportions measured in the Roman concrete that survives

Slaked lime: one measure.

Red pozzolana from the Alban Hills: a scant two and three-quarter measures.

Water: just enough for a stiff mix.

Rubble (tuff, chips of lava, broken brick), in cobble-sized pieces: more than twice as much as the mortar, up to three times.

Two faces for the wall, tuff or brick, built up as the core rises.

For a pier in the sea

Proportions from Vitruvius; the water and the rubble from the piers that have been studied

Dust of Pozzuoli, the kind found from Cumae to the promontory of Minerva: two parts.

Lime: one part.

Seawater: just enough.

Tuff rubble: less than the mortar, two thirds to four fifths as much.

A box of oak stakes with ties between them, driven into the seabed.

Vitruvius's everyday mortar, without pozzolana, is one part of lime to three of pit sand, or to two of river or sea sand. With river or sea sand he suggests adding a third part of burnt brick, pounded up and sifted: cocciopesto, the cook's own touch.

Method. Burn limestone in a kiln at 850 to 900 °C. White, if you can find it, because Cato disapproved of lime from stone of mixed colours. Slake it in water and leave it to rest for a long time: Pliny writes that the old building laws forbade contractors to use a lime slurry less than three years old, and that this is why old plasterwork never cracked. Or don't slake it at all, and mix it with the pozzolana as quicklime, dry, before any water goes in: that is how they did it on a building site in Pompeii left half finished, and the white lumps quicklime leaves in the wall are one of the reasons it lasts. Mix it in a mortar, the trough where the mixing was done: Vitruvius writes exactly that, in mortario. Don't pour. Pack the mix between the two faces of the wall and press the rubble into the mortar. At sea, level and dredge the bottom inside the box, working from beams laid across it, then fill it until the wall takes up all the empty space. Leave it to cook on its own: at the heart of a large mass the reaction between lime and ash heats the mix, and a calculation for one block in the sea in the Bay of Baiae gives 85 to 95 °C for one or two years, before it cools to the temperature of the sea. Leave to rest for two thousand years.

As with any family recipe, the written version and the one actually made don't match. Vitruvius writes two of pozzolana to one of lime; in the Roman concrete studied today there are about 2.7 parts of pozzolana for every part of lime. In 2004 the ROMACONS team remade Vitruvius's recipe in the harbour at Brindisi, in the form of a pila, as the Romans called it: a block about two metres a side, small by the standards of a Roman harbour. They lowered the mortar into the box in wicker baskets, in layers of about twenty centimetres tamped down with rakes, and threw the tuff in on top, piece by piece; by the next day it had set enough to walk on. In the upper part, though, they had corrected Vitruvius the way the Romans did, with 2.7 parts of pozzolana, and it was the upper part that came closest to the ancient samples. Perhaps, they wrote, Vitruvius was following an ideal rule, the kind that belongs in a treatise, rather than what was done on site: lime was the most expensive ingredient, and builders in the first century may have skimped on it a little and found that the concrete came out better. It did not always end so well. The main reason buildings collapse in Rome, Pliny writes, is the purloining of lime.

The ingredients, one by one

Lime. It is cooked stone. The heat of the kiln drives the carbon dioxide out of limestone and gives the stone back the same size, only lighter: it loses weight in the cooking, like a roast. Weighed, Vitruvius writes, it has lost about a third, and he thought the fire had boiled the water out of it. What comes out is quicklime, living lime. Why it should be living is not known for certain: perhaps because, like a fire, it has not yet been quenched, or because of the force with which it reacts. Wet it and it wakes up: it gets hot, very hot. To Pliny this was a remarkable quality: once burnt, it grows hotter still with water. When it has drunk and gone quiet again, it is called slaked. The words are Roman, calx viva and calx exstincta, living lime and quenched lime, and they are already in Vitruvius; quicklime says the same in English, from the days when quick meant alive. In the wall, the lime that has not reacted with the ash comes full circle: it takes back from the air the carbon dioxide the fire drove out, and slowly turns into calcium carbonate again. The stone that went into the kiln becomes stone again in the wall.

An etching of a big square stone kiln beside a river: men work on its flat top around a column of smoke, another stokes the fire at a glowing opening under a tiled roof, and in the foreground a crowd of workers rests and eats beside a donkey with baskets and two cows, with a bridge and a church across the water — A large lime kiln on a riverbank, etched by Cornelis Visscher after Pieter van Laer between 1638 and 1658. Not Roman, but the job is the same: burn the stone, and what comes out is quicklime. Rijksmuseum, Amsterdam.

Pozzolana. It is volcanic ash, and it is the ingredient that makes the difference. In Rome it is the red pozzolana of the Alban Hills, erupted almost half a million years ago by the same volcano the Appian Way runs over, and it is the main ash in the mortars of the city's monuments. For harbours the ash came from the Bay of Naples instead, and it travelled, some of it a very long way. The Romans called it pulvis puteolanus, dust from Pozzuoli, and the word pozzolana still carries the town's name. Vitruvius knew it from around Baiae and the towns round Vesuvius, a powder that "from natural causes produces astonishing results": with lime and rubble it sets hard under water, hard enough to build piers in the sea. Vitruvius thought a fire hidden under that land made it work. The real reason is that the ash contains volcanic glass, molten rock quenched in an instant, and the glass reacts with lime and water to make new minerals that bind everything together, at sea as well.

Rubble. The Romans called it caementa, chips of split stone. Vitruvius calls this kind of building caementiciae structurae; opus caementicium, the name in the textbooks, turns up only once in ancient sources, in an inscription from a second-century sanctuary in what is now north-eastern Greece. It was split stone or fired clay: tuff, travertine, lava, broken brick. On land there was a lot of it, about three quarters of the volume; in the piers, a little under half.

Water. Not much, because the mix had to stay stiff. And at sea, seawater: that is what the Romans mixed the ash and the lime with.

Two recipe cards on a beige background. The left one, headed For a wall, in Rome, shows one measuring jar of white slaked lime, two full jars and one jar filled to about three quarters with red pozzolana, a heap of tuff, lava and brick rubble, a wall in section with two brick faces and a core of stones in mortar, and a drop of water. The right one, headed For a pier in the sea, shows two jars of grey dust of Pozzuoli, one jar of lime, a drop of seawater, a smaller heap of tuff rubble, and a box of oak stakes standing in the sea, filled with mortar and stones. — The recipe written out like a card kept in the kitchen: one measure of slaked lime, a scant two and three quarters of red pozzolana, the rubble, and the two faces of the wall rising with the core. Beside it, the version for the sea: two parts of Pozzuoli dust to one of lime. Schematic, not to scale.

A detail nobody notices

The recipe is written into the words, too. Cement comes from caementa, and caementa from caedere, to cut: at first it was the name of the stones, not of the glue. Mortar comes from mortarium, the trough it was mixed in, the same word as the mortar in a kitchen: the name passed from the bowl to what was made in it. The Latin for lime was calx, which the Romans had taken from the Greek chálix, a pebble. Calx also gave calculus, the little pebble used for counting, and so our calculations.

Where to see it

Under water, at Baiae. The underwater archaeological park, between Bacoli and Pozzuoli, can be seen from a glass-bottomed boat, usually from April to November, or with a mask or with scuba tanks, through the authorised diving centres. A few metres down, carried down by bradyseism, the slow rise and fall of the ground in the Campi Flegrei, lie the Roman villas and Portus Iulius, the harbour Agrippa built in a hurry around 37 BC, with its piers and its concrete pilae: it is one of the harbours where the ROMACONS team took its cores in 2006.

Under hazy blue water, a tall rectangular pillar crusted with shells and weed stands on a seabed of scattered stone blocks — A pillar of Portus Iulius, the harbour Agrippa built around 37 BC, now a few metres under the sea in the underwater park at Baiae. Photo Ruthven, 2018, CC0, via Wikimedia Commons.

On land, in Rome, the same recipe with red pozzolana is in the walls of Trajan's Markets. Why it is still there after two thousand years is another question, with an answer of its own. [[ LINK: Why Does Roman Concrete Last? → /why-does-roman-concrete-last/, on "another question" ]]

…, Exactly? answers one question about a building, properly. If there is one you have always wondered about, reply to any Slow Letter and ask.

Sources: Vitruvius, De architectura II.4–6, V.12 and VIII, in Morgan's translation; Pliny the Elder, Natural History XXXVI.174–177, in Eichholz's translation; J. P. Oleson et al., the ROMACONS pila at Brindisi (2004); E. Gotti et al., "A Comparison of the Chemical and Engineering Characteristics of Ancient Roman Hydraulic Concrete with a Modern Reproduction of Vitruvian Hydraulic Concrete", Archaeometry 50 (2008); C. J. Brandon, R. L. Hohlfelder, J. P. Oleson, "The Concrete Construction of the Roman Harbours of Baiae and Portus Iulius, Italy: The ROMACONS 2006 field season", The International Journal of Nautical Archaeology 37 (2008); M. D. Jackson et al., "Mechanical resilience and cementitious processes in Imperial Roman architectural mortar", PNAS 111 (2014); L. M. Seymour et al., "Hot mixing: Mechanistic insights into the durability of ancient Roman concrete", Science Advances 9 (2023); Vaserman et al., Nature Communications 16 (2025); M. D. Jackson, M. J. Heap, V. L. Pavlovics, A. G. Sinner, "Caementiciae Structurae: Ancient Roman Concrete Structures Fabricated with Reactive Volcanic Rock", Annual Review of Earth and Planetary Sciences 54 (2026); Parco Archeologico dei Campi Flegrei; Lewis and Short, A Latin Dictionary; Online Etymology Dictionary.