The Paestum Shrine Jars: Seventy Years of Tests

In 1954, archaeologists working at Paestum, the Greek city on the coast south of Naples, opened an underground shrine and found bronze jars more than two thousand years old, once sealed with cork. Inside was a waxy paste. The excavators decided it had been honey. Since then the residue has been analysed at least five times that we know of, and the laboratories have not agreed. This page sets out what each one found, and why the question is still open.

What was found at Paestum in 1954?

An underground shrine, dated to about 530–510 BC. Around an iron bed at its centre, as the excavators described it, stood eight bronze jars: six hydriai, the three-handled water jars, and two amphorae. The jars had been sealed with cork disks, and they held a paste-like residue with a strong smell of wax; traces of it were found on their outsides.

The excavators read the empty bed and the closed shrine as signs that a deity was present, and the residue as honey, in the words the 2025 paper reports, “a symbol of immortality,” offered as honeycombs of which only the beeswax remained. The shrine’s owner is itself disputed. The 2025 paper calls it a shrine “to an unknown deity”; the 2023 study calls it a heroon, the shrine of a hero, and reads the iron at the centre as spits laid on a wooden table rather than a bed.

What did the first three laboratories find?

None of them found honey. The 2025 paper sets out all three.

• 1957, Liebefeld, Bern. A bee laboratory, working at the request of London’s Bee Research Association, found the residue would not dissolve in water. What did dissolve in other solvents was a fatty substance similar to wax, and the little pollen in the sediment did not fit honey.
• 1970, Istituto Centrale del Restauro, Rome. Two samples from one amphora, from its neck and its base. The tests found no sugars or proteins.
• 1983, the laboratory of the Rome Chamber of Commerce. The residue would not dissolve in water but did in ether, and its fatty acids were 77.4% palmitic acid. The report concluded that the residue had once contained fats and phospholipids, and found no sugars.

What did the drying-oil study find?

A study published in 2023 by Dodinet and colleagues, part of a research programme involving the Centre Jean Bérard in Naples, took five samples of the residue from two of the hydriai, material that had been stored in the museum in plastic boxes since the excavation. Gas chromatography–mass spectrometry “did not reveal any wax or animal fat, nor oleoresin or plant pitch, but rather the markers of a siccative oil”: a drying oil, the kind that hardens as it oxidises. The authors ruled out honey, wine and olive oil among other contents. Most of the pollen they counted, 163 of 212 grains, came from the cannabis and hop family, and they describe the oil as compatible with a hemp oil. They do not say what it was for: in their words, “the evidence is overall too scarce to hypothesise the precise role of hemp”.

What did the 2025 analysis find?

In 2025 a team led from Oxford published a new analysis, in the Journal of the American Chemical Society. Their sample came from a mass of the residue that arrived at the Ashmolean Museum in Oxford for display in 2019, and they took it from 40 mm inside, to reach material least touched since excavation. For the first time, hexose sugars were detected in the residue’s water extract, at ten times the level in an extract of modern beeswax. The authors note that sugars are not conclusive on their own, since fruit contains them too.

The firmer evidence is protein. Sequencing found peptides of two honeybee proteins, MRJP-1 and MRJP-2, matched to the western honeybee. The authors conclude that their analysis “presents the first direct molecular evidence supporting the presence of honey, likely offered as honeycombs”.

Fig. I · Five readings of one residueDiagram

One residue from sealed bronze jars, five laboratories, seventy years

1. 1957

   Bee laboratory at Liebefeld, Bern, for London’s Bee Research Association

   Solubility tests; microscope

   Not soluble in water; a fatty substance similar to wax. The pollen in it did not fit honey.

   No honey

2. 1970

   Istituto Centrale del Restauro, Rome

   Solubility tests; extraction and hydrolysis

   Two samples from one amphora. No sugars or proteins.

   No honey

3. 1983

   Laboratory of the Rome Chamber of Commerce

   Solubility tests; gas chromatography–mass spectrometry

   Fats and phospholipids; the fatty acids 77.4% palmitic acid. No sugars.

   No honey

4. 2023

   Centre Jean Bérard programme, Naples (Dodinet, Garnier, Barbier-Pain)

   Gas chromatography–mass spectrometry; pollen count. Dated by publication

   No wax, animal fat, resin or pitch; markers of a drying oil. Most of the pollen from the cannabis and hop family.

   A drying oil

5. 2025

   Oxford-led team (da Costa Carvalho and colleagues)

   Infrared, chromatography and protein sequencing on a core sample

   Hexose sugars, and the honeybee proteins MRJP-1 and MRJP-2.

   Evidence supporting honey

Why do “royal jelly proteins” not mean royal jelly?

MRJP stands for major royal jelly protein, and the name misleads. These proteins are secreted by worker bees’ glands, and the 2025 paper describes them as “the principal proteins in honey” as well as part of the food the bees give their larvae. Finding them shows that something made by honeybees was in the jars. It does not show royal jelly, and the authors do not claim it: their conclusion is honey.

What is still open?

More than the headlines suggest.

• The two recent teams disagree. One found a drying oil and no wax; the other found wax-like material, sugars and honeybee proteins. The 2025 paper does not mention the 2023 study, and calls the 1983 analysis “the most recent”.
• How the residue was kept is not recorded in detail. The 2025 team write that they “do not have a detailed history of its collection or its storage conditions since excavation”, and the 2023 team list contamination during excavation or storage among their own limits.
• The samples differ. Each laboratory took its own sample, some from named jars and some of unrecorded origin, so the five results may not describe the same material.
• Neither result has been repeated by another laboratory that we could find.
• Other contents are open. The 2025 authors write that “other commodities, other bee products (e.g., propolis), plant oils, or milk, cannot be excluded”.
• One co-author directed the site. One of the 2025 paper’s authors was director of the Archaeological Park of Paestum and Velia during the research; the paper declares no competing financial interest.

What do the jars not show?

That honey keeps forever. What sits in the museum is a waxy paste that the laboratories cannot agree on, not a jar of honey anyone could eat, and even the 2025 paper argues only that honey was one of the things once placed there. Our guide to whether honey goes off covers how honey keeps, which is a different question.

Where THYMELIA stands

THYMELIA is a UK house of Greek foods: thyme honey from Chania in western Crete, and early-harvest olive oil from the hills above Kalamata. The Paestum jars are a lesson in what a laboratory can and cannot say about a residue, and the same questions, what a test shows and what it leaves open, are the ones we mean to put to our own jars; what we will measure and publish is at testing. There is no shop yet, and we name no date for one.

The longer history of honey in Greece is at honey in the ancient world, and the rest of the Greek table at Greek food.