Determining the contents of bronze jars from an Ancient Greek shrine in southern Italy.

In 1954 archaeologists excavated a shrine to an unknown deity at the Greek settlement of Paestum in southern Italy, which was dated to the sixth century BC. Within this shrine they found six bronze hydriai (storage jars) and two amphorae, arranged around a large iron bed. The jars contained a pasty residue with a wax-like aroma. Traces of this substance were also found on the outside of the jars, which were originally sealed with cork, leading the archaeologists to conclude that it was originally a liquid, although possibly a fairly viscous one. It was interpreted that the bed was intended to be the residing place of the unknown deity, with the contents of the hydriai and amphorae were intended as offerings. 

Honey, in the Greek and Roman worlds, was a substance of some significance. As the only practical way to sweeten food and drinks, it was economically important, but it also had spiritual significance, being associated with wisdom and immortality, and therefore a suitable offering to the gods. With this in mind, it seemed highly likely that the original contents of the Paestum hydriai was honey.

With this in mind, the Bee Research Association in London arranged for an analysis of the residue to be carried out. The substance was found to be insoluble in water, but soluble in organic solvents, and to contain trace amounts of Plant and Insect remains, Fungi, and pollen, which led the scientists carrying out the analysis to conclude that it was probably a remnant of the wax which had originally sealed the jars.

In 1970 scientist at the Istituto Centrale del Restauro in Rome carried out an analysis on residues found around the neck and in the bottom of one of the amphorae from Paestum. This was found to comprise a saponifiable substance (substance which will react with an alkali to form a soap), such as a wax, fat, or resin, but contain no detectable sugar or protein (the major components of honey).

The residue was tested again in 1983 by the Laboratory of the Rome Chamber of Commerce, who again found that it was a saponifiable substance insoluble in water but soluble in organic solvents, and did not contain any sugary or starchy substances. On this occasion gas chromatography was also used to analyse the residue, concluding that it was 77.4% palmitic acid, 6.1% oleic acid, 5.2% stearic acid, 1.0% heptadecanoic acid, 1.0% linoleic acid + arachidic acid, 0.4% linoleic acid, and 6.5% unidentified substance. Since triglycerides of palmitic acid are extremely common in nature, the researchers concluded that the container had held animal fat or a vegetable oil.

In 2019, the residue from the Paestum hydriai was loaned to the Ashmolean Museum in Oxford, for an exhibition, 'Last Supper in Pompeii', and permission was obtained to carry out a new analysis of the biomolecular composition of the substance, using modern equipment not available when the previous tests were carried out.

In a paper published in the Journal of the American Chemical Society on 30 July 2025, Luciana da Costa Carvalho and Elisabete Pires of the Mass Spectrometry Research Facility at the University of Oxford, Kelly Domoney of the Ashmolean Museum, Gabriel Zuchtriegel of the Parco Archeologico di Pompei, and James McCullagh, also of the Mass Spectrometry Research Facility at the University of Oxford, present the results of this new analysis, and confidently identify the original material within the Paestum hydriai.

(A) Underground shrine in Paestum. (B) One of the hydriai on display alongside a Perspex box containing the residue at the Ashmolean Museum in 2019. (C) Graphic representation of the arrangement of the bronze jars inside the shrine. (D) Sample from the core of the residue.  Carvalho et al. (2025).

The residue arrived at the Ashmolean Museum in a non-hermetically sealed Perspex box, in which it had apparently been displayed at the Paestum Museum. In order to reduce the chances of modern contamination affecting their results, Carvalho et al. took samples from 40 mm below the surface for analysis, as well as from each distinct colour zones observed on the exterior of the material; black, orange, and green, colours which suggest some sort of interaction with the bronze vessel itself over the past 2500 years. Carvalho et al. also obtained modern beeswax, honey, and honeycombs from locations in Italy and Greece in order to compare these to the residue sample.

Carvalho et al. first used Fourier Transform Infrared Spectroscopy to obtain an overview of the chemical substances present within the sample. This yielded a spectrum almost identical to that of modern beeswax for the interior sample, strongly supporting the idea that this was the original substance. They also compared the spectra of new and artificially aged honeycombs from both Greece and Italy, establishing that there was little difference in these, and that these structures appear to be chemically stable at least 

Fourier Transform Infrared Spectrum  of the core sample of the archaeological residue superimposed on beeswax (a) and honey’s (b). Carvalho et al. (2025).

Next, Carvalho et al. carried out a Gas Chromatography coupled to Quadrupole Time-Of-Flight Mass Spectrometry analysis of the sample, along with samples of honey, beeswax, and fresh and artificially aged honeycomb. This produced almost no results for the beeswax, suggesting that the bulk components of this material were broken down by the high temperature at which this method operates (over 300°C), but did produce results from the sample, as well as from the honey and honeycomb controls, suggesting that the sample was never pure beeswax.

Electron ionisation chromatograms from the Quadrupole Time-Of-Flight Mass Spectrometry analysis of beeswax (a), the residue core sample (b), honey (c), and honeycomb from Greece, fresh (d) and aged (e). Compounds identified: (1) Hexadecanoic acid, (2) Heneicosane, (3) Octadecanoic acid, (4) Pentacosane, (5) Heptacosane, (6) Nonacosane, and (7) Hentriacontane. Carvalho et al. (2025).

Anion exchange Ion-Chromatography coupled to Mass Spectrometry identified seven hexose sugars, hexadecanoic acid, heneicosane, octadecanoic acid, pentacosane, heptacosane, nonacosane, and hentriacontane within the sample at levels higher than would be expected in beeswax, but lower than would be expected in honey. It also found significant levels of the sulphur amino acid taurine, which was not present in any of the control samples.

These hexose sugars were also recovered from aqueous extracts of honeycomb (i.e. the liquid obtained by soaking honeycomb in water)along with gluconolactone (a derivative of glucose) and galacturonic acid, and low levels of succinic, malic, and citric acids. All of these compounds were yielded at higher levels by the fresh honeycombs than by the artificially aged honeycombs.

Finally Carvalho et al. used a proteomic approach to try to identify specific proteins within the sample, as well as the beeswax and honeycomb controls, which were compared to the UniProt All Proteins database. The reesidue sample produced matches for three proteins derived from the royal jelly produced by the Western Honeybee, Apis mellifera, several Bacteria-derived proteins, and a number of common contaminant proteins, including keratins, caseins, lysyl endopeptidase, and trypsin. Encouraged by this, they then compared the sample to the Bee-specific UniProt Honey database, a search which yielded eight matches, including some associated with the Eastern Honeybee, Apis cerana cerana,

Proteins were also recovered from the modern honeycomb samples, but not the beeswax, indicating that they were derived from the honey portion of the comb. Notably, the royal jelly protein signature from the Greek and Italian honeycombs was quite different, although this was not completely unexpected as the two looked different. Carvalho et al. note that factors such as climate and the floral sources from which nectar is obtained can affect protein expression in Bee products, so this difference does not necessarily mean the Bees were particularly different.

Finally, a sample of the surface residue showing orange, black, and green colouration was subjected to X-ray photoelectron spectroscopy. This determined that the green areas of this residue were composed of 74.98% carbon, 20.78% oxygen, and 4.24% copper, with the copper portion largely dominated by Cu²⁺ ions, while the black areas were composed of 77.96% carbon, 20.12% oxygen, and 1.92% copper, with the copper dominated by Cu⁺ ions, and the orange areas were comprised of 86.50% carbon and 13.50% oxygen. The discoloration in these areas is, therefore, presumed to be derived by interactions between the original substance and the copper portion of the bronze vessel.

Honey is comprised primarily of sugars (typically 79% of the total mass, including 39% fructose), along with water (typically 18% of total mass), acids (0.17-1.17% of total mass) and trace amounts of other substances, such as vitamins, enzymes, flavonoids, and phenolic compounds. Over time this mixture undergoes Maillard reactions ('browning') as the amino acids of the proteins react with the sugars. This will occur more rapidly if the honey is stored at a warmer temperature. Eventually, the honey will take on a dark hue, as the sugars break down into furans and the acid content rises.

Previous studies of the Paestum hydriai residues concluded that this was a wax, most likely used to seal the vessels. The most commonly use wax in the ancient world was beeswax, a substance with quite different properties to honey. Beeswax is typically comprised of about 64% esters, 14% odd medium chain alkanes, 12% free acids, 2% acid polyesters, 1% acid monoesters, 1% free alcohols, and 6% other materials. Beeswax is much more stable than honey, but over time the acid and alcohol contents will increase as the was esters hydrolyse and the shorter chain alkanes are eliminated. 

The Fourier Transform Infrared Spectrum obtained by Carvalho et al. yielded results very similar to beeswax, suggesting that this may have formed a significant portion of the material from which the residue was derived. However, the Gas Chromatography coupled to Quadrupole Time-Of-Flight Mass Spectrometry analysis carried out suggested that the substance could not be pure beeswax. A study of the proteins present within the sample found several associated with honey production in the Western Honeybee, Apis mellifera, with a subsequent search of the Bee-specific UniProt Honey database yielded proteins associated with the Eastern Honeybee, Apis cerana cerana. Also produced were proteins associated with the wood-decay Fungus Armillaria gallica, and the parasitic Mite Tropilaelaps mercedesae, which targets Honeybees.

Eastern and Western Honeybees are closely related, and even where divergence has occurred, their proteins tend to be very similar. Furthermore, there is ample evidence for the cultivation of the Western Honeybee in ancient Italy, but none for the Eastern Honeybee. For these reasons, Carvalho et al. conclude that the readings suggesting the Eastern Honeybee as a source of proteins are probably erroneous, caused by the software trying to 'best fit' ancient degraded proteins.

The presence of proteins associated with the Mite Tropilaelaps mercedesae is also interesting. This Mite originated in Asia, and has a long historic connection with the Eastern Honeybee, but is generally thought only to have begun to infect Western Honeybees in the past few centuries. Carvalho et al. observe that it would be tempting to interpret this as evidence that the material at Paestum originated in Central Asia, but that a more likely scenario is that the proteins in question are common to a range of Acarid Mites.

Carvalho et al, conclude that a bulk composition similar to beeswax combined with the presence of proteins and other molecules found in honey make it likely that the material within the Paestum hydriai was almost certainly a honeycomb. This matches well with ancient literature, which frequently cites honey and other Bee-products as being suitable offerings for the gods, but contradicts earlier studies which were unable to obtain this result. Carvalho et al. emphasise that this underlines the importance of revisiting samples which have previously been analysed with less modern techniques, thus allowing our imptoving technology to improve our understanding of the past.

See also...

The death of a hydrothermal vent community.

The East Pacific Rise runs from the Southern Ocean to the Gulf of California, where it passes into the subduction zone beneath the North American Plate, becoming the San Andreas Fault. A number of deepsea hydrothermal vent communities have been discovered along the East Pacific Rise, some of which have been studied extensively. One of these was the site known as Bio9, located at 9°50' North, which was discovered in 1991, and studied extensively over the next few years. The site thrived until 1995, but appeared to sicken and die over the next two years.

In a paper published in the journal Geochemical Transactions on 27 January 2012, Michael Hentscher and Wolfgang Bach of the Department of Geosciences at the University of Bremen discuss the history of the Bio9 hydrothermal vent community over the period 1991-1997, and theorize about how changes to the chemistry of the water emerging from the vent may have brought about the demise of the vent community.

Diagram showing the chemical cycle of a healthy hydrothermal vent community. From Hentscher & Bach (2012).

The Bio9 community was dominated by tubeworms of the genus Riftia when it was first discovered. These worms obtain nutrition from symbiotic bacteria that live within their bodies, bacteria that oxidize hydrogen sulphide from the mineral rich waters. These worms were thriving in 1991, but in 1994 the colonies appeared to have developed rusty red spots. In 1995 the colonies were largely covered by the rusty spots, and by 1997 the colonies had largely died off.

Healthy Riftia tubeworms. NOAA.

Hentscher & Bach theorize that the rusty material was colonies of red, iron-oxidizing bacteria, and that their growth reflected an increase in the amount of iron ions in the water emerging from the vents; they do not believe the bacteria directly harmed the tubeworms.

The death of the tubeworms between 1995 and 1997 would therefore not be directly related to the spread of the red rust in 1994-5. Hentscher & Bach suspect that this was due to a fall in the level of hydrogen sulphide within the vent water, resulting in the death of first the symbiotic bacteria within the worms, and then the worms themselves.

The chemicals emitted by terrestrial volcanic springs and fumaroles change over time, and there is no reason to suspect that deep-sea hydrothermal vents would behave any differently. If this is the case then the death of this vent community would be a part of the natural life-cycle of the tubeworms, and hydrothermal vent communities in other parts of the world, with different biological communities, probably suffer similar die-offs.

See also A deep-sea hydrothermal vent community discovered in the Mariana Trench, Deepest hydrothermal vent communities yet found discovered in the Caribbean and New deep-sea hydrothermal vent ecosystem discovered in the Southern Ocean.

Cooking the primordial soup; did the first life emerge in volcanic pools?

The blood plasma and lymph of modern animals is similar in chemical composition to seawater, strongly supporting the idea that animal life began in the oceans, but the liquid inside our cells has a quite different chemistry, suggesting that cells themselves first arose in a different environment, since the first cells are unlikely to have shared modern cells ability to maintain an interior chemistry very different to the liquid outside their membranes.

In a paper published online in the Proceedings of the National Academy of Sciences on 13 February 2012, a team of scientists led by Armen Mulkidjanian of the School of Physics at University of Osnabrück and the A. N. Belozersky Institute of Physico-Chemical Biology at Moscow State University, describe a review of our understanding of the chemistry of the earliest cells and the environment on the early Earth in these cells are thought to have inhabited, and the conclusions drawn from this.

The oldest known rocks on Earth are about 4 billion years old, but life is thought to have originated earlier than this (life does not pre-date rocks, just any rocks still around), in a period known at the Hadean. This makes the chances of finding direct evidence of the earliest life effectively nil, but does not stop us speculating, as we are still able to make inferences about the environment on the Hadean Earth.

An artist's impression of the Hadean Earth. The continents were still assembling from thousands of volcanic island arcs, the moon was closer, the sun dimmer, the planet was being constantly bombarded by meteors, the atmosphere rich in Carbon Dioxide but entirely lacking in Oxygen, consequently there was no Ozone layer and the planet was constantly bathed in ultra-violet radiation. From the Palaeos website.

The oldest, conserved, proteins found in all organisms tend to use Zinc and Manganese, but not Iron, which is used by more modern proteins that have evolved separately in different groups, implying that the first cells evolved in an environment rich in Zinc and Manganese, but poor in Iron. Water from hydrothermal vents tends to be rich in Zinc, Manganese and Iron, but the iron precipitates out of solution rapidly.

All known cells maintain an interior environment richer in Potassium than in Sodium; the reverse of the situation found in the modern ocean. We do not have any evidence that suggests the situation in the earliest oceans would have been any different. Hydrothermal vents in the seas have similar Sodium/Potassium ratios to seawater, since this is where the water in them derives from, and the contribution of Sodium and Potasium ions from the seawater outweighs the volcanic contribution. Terrestrial hydrothermal springs derive their water from precipitation (rain and snow), which lacks Sodium and Potassium ions. The Sodium/Potassium ratio in such pools is variable; those which are dominated by water that is emitted in a liquid tend to be rich in Sodium, but those where water is emitted as a gas and then condenses tend to be rich in potassium.

All cells maintain a high interior Phosphate concentration, but the oceans are not rich in Phosphates, and there is no reason to believe the early oceans were any different. Water from volcanic vents tends to be rich in Phosphates.

Previous theories have suggested that life may have originated around deep-sea hydrothermal vents, but these have steep chemical gradients, and chemistry unlike that found inside cells. Hydrothermal pools on-land tend to have stable chemistry, but are very acidic, and so have been discounted as a likely source of life by many studies. However this acidity is caused by the reaction of Sulphur-compounds with Oxygen in the atmosphere, something that could not have happened in an ancient environment lacking atmospheric Oxygen.

The other objection to terrestrial hydrothermal pools as a place of origin for life is the high level of ultra-violet radiation that would have bathed the early Earth and which is harmful to life. The modern world is protected from ultra violet radiation by the Ozone Layer, but Ozone (O₃) is a form of Oxygen, so this would not have existed prior to the evolution of an Oxygen rich atmosphere. Water also protects against ultra-violet radiation, but there needs to be enough of it; the sea would protect early cells, but shallow pools would probably not. However Sulphur from volcanic vents, if it was not reacting with oxygen from the atmosphere, would probably react with Manganese and Zinc if they were present in the same pools of water. Sulphur compounds of Manganese and Zinc are good at absorbing ultra violet radiation, offering protection to any life living in these pools.

See also A deep-sea hydrothermal vent community discovered in the Mariana Trench, The oldest animals - Pre-Ediacaran Sponges from Namibia(?) and Did the first land plants cause an Ordovician glaciation?