Photograph of grassy hills with rocky features outlining an archaeological site

Discovering a 4,500-year-old olive oil factory in Jordan

By Jamie Fraser, Curator for the Ancient Levant and Anatolia and Caroline Cartwright, Senior Research Scientist

Publication date: 26 November 2020

Curator Jamie Fraser and scientist Caroline Cartwright run you through the archaeology and science involved in tracking down an ancient olive oil factory, and explain how the process of making modern olive oil would be familiar to people in the past.

Discovering a 4,500-year-old olive oil factory in Jordan

When I (Ed. Jamie) first visited the archaeological site of Khirbet Um al-Ghozlan in the north of Jordan, I dismissed it as nothing more than an ancient farming village. Arabic for the 'Ruins of the Mother of the Gazelles', the site sits on a small knoll above the steep-sided Wadi (Valley) Rayyan. The slopes are covered in oak forests, wild flowers and olive groves. You can see west to the Jordan Rift Valley (200m below sea level) and east to the Transjordan plateau (rising 1400m above sea level in the distance). The view is spectacular.

A photograph of an archaeological site with stones and boulders on top of a grassy hill. A sandy track passes through on the right, at the end of which a car is parked
Khirbet Um al-Ghozlan in the Wadi Rayyan, Jordan. Photo by Bashar Tabbah.

Archaeologists would normally classify the site as a rural farming village because of its size. Only one acre (0.4ha) in area, Khirbet Um al-Ghozlan could easily fit into the Great Court at the British Museum and still leave room for the coffee shops. It was also occupied in a time of urban collapse, at the end of the Early Bronze Age (2600–2000 BC). During this period, people in the southern Levant (modern-day Jordan, Israel and the Occupied Palestinian Territories) abandoned their cities and dispersed into small, rural communities. 'Nothing special', I mused as I walked back to the car, 'at least the view was nice'.

So imagine my surprise when I stumbled across a constructed line of massive stone boulders hidden by the long grass. With mounting excitement I followed them around the entire knoll, slowly realising they formed a monumental enclosure wall protecting the site. This didn't make sense: in the Bronze Age world, walls were built to fortify large urban centres, not tiny rural hamlets. What had motivated people 4,500 years ago to defend such a remote site?

An ancient olive oil factory

Back at the British Museum I sought out Dr Caroline Cartwright in her microscopy lab in the Department of Scientific Research. As an archaeobotanist, Caroline has spent decades studying the relationship between people and their environment by analysing ancient plant remains. She was the perfect person to discuss one possible explanation for the site's mysterious defensive wall – that it protected a Bronze Age olive oil factory.

Photograph of grassy hills with rocky features outlining an archaeological site
The hills of the Wadi Rayyan, Jordan. Photo by Bashar Tabbah.

The key, I suggested, lay in the site's upland location. The well-drained slopes of the Wadi Rayyan are ideal for cultivating olive trees, especially compared to the flood-prone Jordan Valley floor where their woody roots tend to rot. If Khirbet Um al-Ghozlan served as an olive oil factory and storehouse, then maybe ancient farmers enclosed it to protect their seasonal stockpiles of olive oil – one of the most precious commodities in the Bronze Age. 'Don’t think of it like a permanent settlement behind a formal fortification wall', I said, bouncing on my stool, 'more like a bank-vault in the landscape, where people stashed their olive oil during the month-long harvest'. Caroline raised an eyebrow, agreed it was a reasonable idea, then responded with sharp scientific acumen: 'but where's the evidence? We need archaeological data to prove it'.

We needed to dig. And so, with permission from the Jordanian Department of Antiquities, a joint team from the British Museum/University of Sydney undertook excavations at Khirbet Um al-Ghozlan in March 2017 and November 2019. Three Jordanian archaeologists joined us from the Department of Antiquities, and we employed ten men from the local village to help.

The origins of olive oil

A black-figure amphora showing the olive harvest. People shake olives from a tree using long poles. Another is collecting the fruit from the floor, and another is sitting in the branches.
Scene depicting the olive harvest in Greece by the Antimenes Painter. Attic black-figure amphora, Greece, around 520 BC.

The story of olive oil starts around 6000 BC, when people in the eastern Mediterranean learned how to extract oil from the bitter fruit growing on wild olive trees. By 5000 BC, people were cultivating domestic trees in orchards. Olive oil quickly transformed the way people lived: it was burned as fuel in lamps, it helped preserve food longer, especially dairy products such as yoghurt and cheese, and it enabled a broad cuisine to flourish, providing the foundations of what we know today as the 'Mediterranean diet'.

The Middle East soon developed an international oil economy – albeit one based on the olive. The first cities in the southern Levant emerged between 3700–3000 BC, partly driven by trade with Egypt, where olive oil was prized as one of seven 'sacred oils' for embalming the dead. However, this fledgling urban experiment soon faltered and collapsed, possibly when Egyptian merchants shifted the oil trade north to ports in Syria and Lebanon. The southern cities were gradually abandoned and people disbursed into the countryside. This so-called 'Dark Age' marks the end of the Early Bronze Age (2600–2000 BC). It is to this period of urban recession that Khirbet Um al-Ghozlan dates.

Producing modern olive oil

To test our hypothesis we needed to understand what the signatures of olive oil production would look like as archaeological remains. Although modern olive presses are mechanised, the stages involved in harvesting the fruit and extracting the oil have remained largely unchanged over the last 8,000 years. To learn more, we returned to Jordan in late 2019 to observe the annual harvest.


The harvest occurs every October to November, usually after the first autumn rains have washed the summer dust from the ripening fruit. Farmers shake or rake the branches causing the fruit to fall onto sheets below. Groups of people remove twigs and leaves before gathering the olives in sacks.

Top: Two people harvest olives while standing in the branches of a tree with a ladder.Bottom: Three people sit on a blanket and sort the fallen olives into containers.
Olives being shaken from the tree (above) and sorted on the ground (below) in Jordan, November 2019. Photos by Adam Carr.


The sacks are taken to local olive factories that contain the pressing equipment. The fruit is emptied into a large basin and crushed by rotating mill-stones into a tapenade-like paste. A worker presses the paste onto circular fibrous mats, which are stacked high like pancakes onto a trolley ready to be pressed.

Two large mill-stones inside a green metal hopper. The stones are used for grinding olives.
Mill-stones crushing olives in the Toubneh olive mill in Jordan, November 2019. Photo by Adam Carr.


Traditional olive presses used stone weights or screwed vices to apply pressure to the mats. Today, the stacked mats are pressed in a motorised steel vice, forcing the oil to exude in golden rivulets, which are drained into barrels or tins. Once the pressing is complete, the olive paste is scraped from the mats and collected for fertiliser or fuel.


As oil collected from non-mechanised presses contains a significant proportion of water, oil was traditionally stored for several weeks in jars or vats in cool, dark storehouses to allow the oil and water to separate. The lighter oil was then scooped or decanted away. The process was complete.

Producing ancient olive oil

If our hypothesis was correct, then excavations at Khirbet Um al-Ghozlan should yield archaeological evidence illustrating all stages of this process. The site did not disappoint. Strikingly, the monumental enclosure wall appears to have protected only two or three buildings, each containing several stone bins and storage installations. Were these compounds the storehouses used to cache oil jars accumulated throughout the harvest? Clues lay in the materials they contained.

A computer rendering of the archaeological site showing the outline of buildings with stones. Trenches are marked with red lines and numbers.
A partly-excavated storage compound at Khirbet Um al-Ghozlan. Orthophoto by Ehab al-Jariri.


The most compelling evidence lay in the charred remains of ancient plant remains – known as the archaeobotanical record. Tiny fragments of charcoal found in the soil represent the remains of wood that people brought to the site to burn as fuel. Using a scanning electron microscope, Caroline identified the original tree species of each sample by examining its anatomical structure from different angles. We know from other excavations that people in Bronze Age farming villages typically burned a variety of wood species gathered from around the site. Not so at Khirbet Um al-Ghozlan. 'It's quite extraordinary', Caroline observed, 'every single sample is olive wood. They weren't burning anything else.'

This pattern suggests that, like farmers today, Bronze Age people pruned their olive trees as part of the harvest and stockpiled the wood for fuel. The picture became clearer when Caroline discovered fragments of small shoots and suckers, usually removed during pruning to shape the tree.

A scanning electron microscope image showing olive wood structure in detail
Scanning electron microscope image of olive wood. Image by C R Cartwright.
A scanning electron microscope showing an olive sucker close up
Scanning electron microscope image of an olive sucker. Image by C R Cartwright.
A scanning electron microscope of a crushed fragment of an olive stone close up revealing structure
Scanning electron microscope image of a crushed fragment of an olive stone. Image by C R Cartwright.

This inference tallied neatly with the discovery of several beautiful flint blades that may be the remains of ancient pruning saws. Known as Canaanean blades, these tools are about as long as your hand, from the base of your palm to the top of your middle finger – the perfect length for pruning slender olive branches. Each blade is serrated on both sides and notched at one end where they were hafted to a handle.

A collection of ancient flint blades organised with the tallest blades in the centre. They have jagged outlines and smooth centres.
Ancient flint blades used as olive-pruning saws. Photo by Adam Carr.
A modern metal saw cuts a branch from an olive tree.
A modern metal olive-pruning saw. Photo by Adam Carr.


But the smoking gun was Caroline's discovery of miniscule fragments of crushed olive stones, known in Arabic as jift. These fragments are unambiguous by-products of olive oil production that prove olives had been crushed for oil at Khirbet Um al-Ghozlan. The fact they were charred indicates that, once pressed, the jift had been collected and burned as fuel, just as people burn jift as fuel today. Several limestone mortars excavated in and around the storehouses were probably used to crush the olives into paste.


The crushed paste was probably pressed in one of six olive presses found hewn into the limestone bedrock nearby. Each comprised a rectangular basin into which mats laden with the crushed olive paste was pressed under heavy stones. The oil would flow through a circular channel into a storage jar nestled into a collection tank below.

An olive press cut into stone at the site. Two deep depressions can be seen and a channel runs from the tank at the top to the collection vat at the bottom.
Olive press cut into limestone bedrock at Khirbet Um al-Ghozlan. Note the circular channel draining the oil from the pressing tank (top) to the collection vat (below). Photo by Adam Carr.


The storage compounds themselves contained at least 23 storage jars lined up in several rooms. Although broken, Jordanian conservator Naif Zaban at The American Center of Research in Amman has been slowly restoring these jars to their original form. Remarkably, 83 per cent of all pottery sherds were from storage jars. This is a strikingly high proportion, as we would normally expect other types of domestic vessels to be well represented, including cooking pots, bowls and jugs. We hope to analyse these jars for organic residues: if we can detect lipids associated with vegetable fats, then the jars probably contained olive oil – stayed tuned for updates! Although this last piece of the puzzle is still being researched, the evidence overwhelming points to a site for specialised industrial production rather than a typical farming village. Together, the plant remains, storage jars, olive presses and stone tools all support the theory that the site was used for the seasonal production of olive oil.

An archaeologist digs at the site. Stone walls and structures outline the shape of the building and broken pottery shards are being excavated.
Archaeologist Beau Spry excavating broken storage jars at Khirbet Um al-Ghozlan, Jordan, March 2017. Photo by Adam Carr.

Olive oil and the city

Perhaps the best analogy for understanding the site is the olive oil factories scattered through the hills of Jordan today. For 11 months of the year, these isolated structures are locked-up, silent and dark, their pressing equipment stored under plastic. Then, in mid-October, their doors are opened, their lights switched on, and they become some of the most dynamic places in the country, as farmers bring their olives to be pressed. The key question for us is where the people who visited Khirbet Um al-Ghozlan came from. Did they live in nearby villages, or did they venture to the uplands seasonally, from larger settlements on the Jordan Valley floor?

An archaeologist digs at the site. Rocky pieces and broken pottery shards are being excavated.
Archaeologist Sarah Carter carefully excavating a broken storage jar in the back of the storage compound. Photo by Adam Carr.

The storehouses at Khirbet Um al-Ghozlan attest the timelessness of the olive oil industry. The sounds of the olive harvest have echoed through the hills for millennia, and genetic studies indicate that trees cultivated today are the descendants of those harvested in the Bronze Age. The modern olive harvest would certainly be familiar to someone from antiquity – a fixed point in an eternal agricultural cycle.

From the earliest Neolithic villages to the modern international economy, we cannot understand the development of cultures in Jordan without understanding the role of olive oil. The excavations at Khirbet Um al-Ghozlan help articulate the close relationship between olive oil and the rise and fall of early urban societies. The site illuminates how a resilient olive industry maintained strong economic sinews during a period of urban collapse at the end of the third millennium BC. Indeed, many archaeologists believe it helped promote the recovery of urban societies in the early second millennium BC. This rejuvenated civilisation would come to be known as the 'Canaanites'. Although later biblical texts would describe the Canaanites as occupants of a 'land of milk and honey', this sobriquet overlooks one of the most significant and enduring aspects of the region – that it was a land of olive oil as well.

You can hear Jamie talk about the excavations as part of the July edition of the British Museum's podcast series.

You can read more about the project in online reports for the Palestine Exploration Fund and the American Schools of Research.


The 2017 dig was funded by the Wainwright Fund (University of Oxford) and the Curtiss T. and Mary G. Brennan Foundation. The 2019 dig was funded by the Palestine Exploration Fund, the Curtiss T. and Mary G. Brennan Foundation, the American Schools of Oriental Research and the Australian Academy of the Humanities. The SEM analysis of the archaeobotanical remains was funded by the Mediterranean Archaeological Trust.