Secrets of a Roman Sewer: What Herculaneum's Cesspit Says About Diet

Most of what Romans wrote about food was written by and for the rich: banquets, rare fish, sauces flavored with imported spices. Much less is known about what a shopkeeper living over his workshop ate on an ordinary Tuesday. At Herculaneum, the small town on the Bay of Naples buried by Vesuvius in AD 79, one answer came from an unglamorous place: a tunnel under an apartment block where residents had been dropping their kitchen scraps and emptying their latrines for about a decade before the eruption.

The tunnel is known as the Cardo V sewer, after the street it runs beside. Its contents were excavated by the Herculaneum Conservation Project, and the food remains were analyzed in detail by the archaeobotanist Erica Rowan for an Oxford doctoral thesis completed in 2014.¹ What came out of it has changed how scholars talk about the diet of non-elite Romans, though, as later work shows, it is one line of evidence among several.

A sewer with no way out

The sewer ran under Insula Orientalis II, a multi-story complex of shops with living space above and behind them, built alongside the town’s large exercise ground. Rowan, drawing on the excavators’ reports, describes the main north-south branch as 86.3 meters long and 80 centimeters wide, rising from 2 meters high at its northern end to 3.6 meters at the south. The latrines of the shops and apartments line up with it exactly, which shows it was planned together with the building.¹

What makes it unusual is that it had no exit. Unlike another Herculaneum sewer that drained onto the beach, the Cardo V tunnel was walled off at its southern end, so it was not designed to carry away rainwater. Eight latrine shafts and three small drains fed into it, which, in Rowan’s words, “in effect turned the sewer into an enormous cesspit.” Three large access shafts allowed slaves to climb down and clear out the accumulated material from time to time, and a wide side branch acted as a collecting point for waste creeping slowly downhill.¹

Andrew Wallace-Hadrill, then director of the project, explained the arrangement to National Geographic in 2011: “All the foul stuff from the latrines and all the rubbish thrown down the chute accumulate and compost, as in a septic tank.”² Readers who know the site’s piece on Roman latrines and the sponge stick will recognize the pattern: in towns like Herculaneum and Pompeii, most toilets emptied into pits or tanks rather than into a flowing sewer.

From Maiuri’s tunnel to 775 bags of soil

The Italian archaeologist Amedeo Maiuri entered the sewer in 1949. He emptied the side branch and the southern 50 meters of the main tunnel, mapped the rest by looking in from each end, and left a brief published description with no record of method. The middle stretch, sealed under hardened volcanic material, was left untouched.¹

The Herculaneum Conservation Project, a partnership of the Packard Humanities Institute with the Italian heritage authority, has worked at the site since 2001, with the British School at Rome as a partner for a decade.³ Its reason for returning to the sewer was practical: modern drainage pipes were to be laid through it, and the team had to make sure they would not destroy the archaeology. They cleared a meter or more of sand washed in since Maiuri’s visit, then used a power drill on the cement-hard volcanic fill, leaving a protective layer just above the ancient deposit. Rowan writes that the large volume of well-preserved organic material they found was “a pleasant but unexpected surprise.”¹

Because the whole deposit had to come out for the pipes, all of it was bagged: more than 11,600 liters of soil in about 775 bags of 15 liters each. In September 2007 Mark Robinson of the University of Oxford led a team that processed a sample on site, washing it through a series of meshes, floating off light seeds and charcoal, and sorting the heavier residue by eye and under a microscope.¹ Wallace-Hadrill told reporters that the material had no smell at all: “It’s exactly like earth compost.” National Geographic reported that only 70 of the sacks had been examined by then, and Wallace-Hadrill said that looking at everything in detail “would take a lifetime.”² Alongside the food remains came lost jewelry, coins, stones from a gem shop, and broken lamps and pottery.²

Pottery from the deposit dated it to the last ten years before the eruption, so Rowan treats the sewer as about a decade of waste from the people living above.¹

How a seed survives a sewer

Plant remains usually rot. In the Cardo V tunnel most of the seeds survived by mineralization: calcium phosphate from feces, bone and water replaced the soft tissue and left a stony cast. Of the botanical taxa Rowan identified, 84 were preserved this way, 27 as charred remains and about ten by waterlogging (the thesis gives nine in its results tables and ten in its conclusions). Mineralized items were mostly things swallowed whole, such as fig pips, seasoning seeds and weed seeds, which means they arrived inside human waste. Charred items, mostly olive stones, had been burned in cooking fires or as fuel and swept down the shafts.¹

The faunal remains carried their own clues. Some fish vertebrae were flattened by chewing, and some otoliths (the small ear stones of fish) were smoothed by stomach acid, showing that they had passed through a human gut. A few sea urchin spines show the same acid damage and were probably swallowed by accident.¹

The waste also moved. Material drifted south along the tunnel over time, so it was not possible to tie any find to the latrine of a particular shop. Rowan analyzed the whole sewer as a single context: a shared record of one building, not a household-by-household one.¹ In her analysis of six one-meter sections, 220 liters of soil yielded 194 taxa: 94 plants, 45 fish, 53 shellfish and two birds. She counted 114 of them as foods.¹ Comparing the plant list with nine assemblages from Pompeii, she found that 82 of the 94 botanical taxa also occur there, and concluded that the sewer was not unusually biased by its preservation conditions.¹⁴

What was on the menu

The picture that emerges is of a diet built on a few staples eaten by nearly everyone, with a wide range of other foods eaten often but less regularly. The staples Rowan identifies are cereals, grapes and figs. Around them came other fruits, eggs, herbs, fish and shellfish. In all she counts seven kinds of cereal, 12 kinds of fruit and 45 species of fish. Nearly all of it could have been grown or caught locally. Only three items were definitely imported from beyond Italy: dates, probably from the Middle East, Egypt or North Africa; black pepper, probably from India by way of the Red Sea port of Berenike; and sorghum, which may have arrived by accident mixed into other cargo.¹ Press reports in 2011 also mentioned chicken, mutton, fennel, olives and sea urchin.²

Seafood was a regular part of life. A 2018 study of the marine remains by Rebecca Nicholson, Jennifer Robinson, Mark Robinson and Rowan identifies around 70 fish taxa, 48 marine mollusks and three marine arthropod taxa. The fish were mostly small, probably netted in shallow coastal water, and much of the shellfish could have been gathered close to the shore. The authors conclude that both “probably represented cheap and nutritious dietary fare,” with most bought fresh and some eaten salted or as fermented fish sauce.⁵ In the thesis, Rowan points to anchovy, sardine, picarel and mackerel, the species most often used to make fish sauce, which turn up as proportionally more bones than ear stones, a pattern that suggests they often arrived in the kitchen as sauce rather than as whole fish. Sea breams, horse mackerel and anchovies were the fish eaten most regularly.¹

None of this would have surprised Pliny the Elder, who died in the same eruption. In his survey of Italy he called Campania a blessed country and wrote that its seas “are distinguished beyond all others for the superlative excellence of their shell and other fish.”⁶ What Pliny and other elite writers did not do was treat shellfish as worthy food for the likes of the shopkeepers of Insula Orientalis II, a bias Rowan notes in her comparison of the texts with the finds.¹

A richer table than expected

Two older assumptions about the urban poor take a knock from these results. The first is that their diet must have been dull and repetitive, because they could afford little beyond grain, wine and oil. The architecture of Insula Orientalis II, with its cramped rooms and mezzanines, would point to exactly that kind of diet. The food remains do not. As Rowan puts it, “In Herculaneum, even small shop owners could afford fish and seafood on a consistent basis.”¹ She goes further: going by the ancient authors’ remarks on prices, “they should not be able to afford black pepper and sea bass.” Whether the authors exaggerated prices or Herculaneum was simply a well-off town, she leaves open.¹

The second assumption is that people in small apartments without built kitchens bought hot food from street counters instead of cooking. Millet and legumes, which are not usually eaten raw, along with thousands of fragments of charred olive stones used as fuel, suggest that residents cooked at home on portable hearths or tripods.¹

Rowan also found signs of inequality inside the building. Fish variety dropped toward the northern part of the sewer, and rare items like pepper turned up only occasionally, which she reads as differences in wealth between households: most people ate much the same food, but not exactly the same.¹⁴ She cautions that Herculaneum, close to the farms of Campania, the fishing grounds of the bay and the port of Puteoli, may have been unusually privileged compared with other Roman towns.¹

Checking the sewer against the skeletons

A sewer shows what went into people, but not how much, or who ate what. For that, researchers turn to the roughly 340 skeletons excavated since the 1980s on the ancient beach and in the stone vaults along the shore where residents sheltered during the eruption.¹⁷

Rowan compared her findings with earlier skeletal studies, which reported average adult heights slightly above those of modern Neapolitans measured in the 1960s, which she reads, following earlier work, as a sign that people had enough food while they were growing. She also noted that the anthropologists who studied the bones did not always use the same methods and sometimes disagreed about the same skeletons.¹ Her overall verdict was upbeat: “Life in Herculaneum was good, at least where diet is concerned.” She paired it with a reminder that health and sanitation “were far from modern standards.”¹ Wallace-Hadrill had already suggested in 2011 that microscopic study of the deposit might one day reveal parasites and infections.²

Isotope chemistry has since added precision. Earlier bulk isotope work cited by Rowan put marine foods at roughly 0 to 15 percent of dietary protein.¹ In 2021 Silvia Soncin and colleagues measured individual amino acids in the bone collagen of 17 adults. They wrote: “We show that the bulk isotope data underestimate the marine protein component of diet.” Their estimate for marine protein averaged 26 percent, well above the share in 20th-century Mediterranean diets, with cereals and land-based foods still supplying most of the calories.⁷ They also found a gap the sewer could never reveal: men obtained, on average, 1.6 times more of their protein from seafood than women did.⁷

Taken together, the two kinds of evidence agree that fish mattered far more in this town than the old image of a grain-and-oil diet allowed. They also show the limits of each approach. The sewer gives a long list of foods for one building but no portions and no names. The skeletons give individual diets but cannot say which fish or which grain. The rest of the soil, kept in storage at Herculaneum and most of it not yet sorted, may one day add more.¹

Footnotes

1. Rowan 2014, DPhil thesis, chs. 1, 2.4, 2.9, 4.2, 5.15, 6, 7.6-7.7, 9.1. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵

2. Owen, National Geographic News, June 24, 2011. ↩ ↩² ↩³ ↩⁴ ↩⁵

3. Herculaneum Conservation Project, herculaneum.org. ↩

4. Rowan 2017, Environmental Archaeology 22(3), abstract. ↩ ↩²

5. Nicholson et al. 2018, Journal of Maritime Archaeology 13(3), abstract. ↩

6. Pliny, Natural History 3.9 (3.60), Bostock and Riley translation. ↩

7. Soncin et al. 2021, Science Advances 7(35), introduction, results and discussion. ↩ ↩² ↩³