Across north-western Saudi Arabia, long rectangles built from dry stone form an extraordinary prehistoric landscape. They are called mustatils, from the Arabic word for rectangle. Their outlines were recognisable before their purpose was understood. Ground survey, excavation, animal-bone analysis and radiocarbon dating have since changed the question: researchers can now make a substantial case for ritual activity, while the beliefs, participants and wider organisation behind it remain partly unknown. The development of that interpretation is more informative than treating the monuments as permanently unexplained shapes.
2017: shapes on satellite images
David Kennedy’s 2017 paper introduced “gates” as a distinctive archaeological site type. The name referred to their appearance, rather than a demonstrated role controlling passage. His abstract described how freely accessible high-resolution satellite imagery had transformed the mapping of Saudi Arabia’s stone-built archaeology, particularly in the lava fields. Around Harrat Khaybar, low, roughly constructed walls formed arrangements unlike other already familiar monument types.
Kennedy could identify and map a recurring form, and its relationships with other structures suggested an early position in the sequence. He could not discern an obvious purpose and explicitly called for fieldwork. This was a methodological boundary: a plan visible from above cannot reveal everything buried inside a platform, nor automatically distinguish successive uses of the same stones. The 2017 publication’s abstract is available through the publisher; it is the basis of this account of that initial interpretation, rather than a claim that its complete field catalogue has been examined here. Read Kennedy’s original abstract.
2020: a date and a closer look
Huw Groucutt and colleagues studied 104 mustatils along the southern margins of the Nefud Desert. Their 2020 paper combined satellite mapping with ground observations. The ground views showed that apparently enclosing walls could be very low, while the substantial ends were platforms. A painted stone facing the interior of one platform suggested attention to people inside the defined space. The scarcity of ordinary occupation debris also helped distinguish the structures from settlements.
A charcoal sample came from a section exposed by an unauthorised hole in a platform. The authors described its association with the original structure and reported a calibrated range of 5052–4942 BC. That is a date for a specific sample and context, not a construction certificate for every rectangle across Arabia. They also documented extensive later reuse of stone, complicating appearances from above. Their published survey gives coordinates for a 616-metre example: this is the representative southern Nefud location used by our directory pin, rather than the structure in the lead photograph. Read the original 2020 study.
2021: architecture and offerings
Hugh Thomas and colleagues’ 2021 Antiquity paper brought together aerial and ground investigations in AlUla and Khaybar. Its descriptions identified head platforms with chambers, elongated courtyards, low connecting walls and, in many examples, narrow entrances through the opposite base. Some entrances had been deliberately blocked. The investigators suggested that movement through these restricted openings could involve a procession towards the head, although a reconstructed route is not a witnessed prehistoric ceremony.
Excavation at a mustatil east of AlUla recovered horns and cranial material from domestic and wild animals, clustered around an upright stone interpreted as a betyl. Radiocarbon measurements supported a Late Neolithic date. These finds supplied a more concrete basis for ritual interpretation than rectangular shape alone.
The same paper estimated construction effort using experimental comparisons: small groups might build substantial monuments within weeks or months. These are modelled possibilities, sensitive to assumptions about stone transport, labour and provisioning. They neither require an otherwise unknown technology nor establish the actual workforce. Read the 2021 architecture and excavation paper.
2023: return visits to one monument
Melissa Kennedy and colleagues’ detailed PLOS ONE study examined a 140-metre mustatil about 55 kilometres east of AlUla. Its head lay beneath a rock overhang. Excavations revealed selected animal remains, with cattle particularly prominent, alongside goats and a gazelle element. Exceptional preservation of some horn sheaths helped show what had been deposited; poorer preservation elsewhere remained a limitation.
The arrangement of offerings and evidence for more than one depositional episode supported a ritual interpretation. The authors discussed “pilgrimage”, cautiously allowing the less specific possibility of returning to a shrine. A later human burial near the head added another episode to the place’s history. Returning, burying and originally building are different acts, even where they leave evidence close together. The study did not recover a written description of the ceremonies or identify a named deity. Read the detailed 2023 excavation report.
2024: comparing monuments and everyday life
A 2024 Open Quaternary paper by Kennedy and colleagues widened the excavated comparison. It described a repeated preference for upper cranial elements of horned ruminants, while the proportions of cattle, sheep, goats and wild species varied between monuments. A broadly shared practice could therefore coexist with differences between the communities making deposits. Similar architecture did not demand an identical inventory at every site.
The comparison with domestic occupation was particularly useful. Cattle were much more prominent in the monument deposits than in the domestic assemblages discussed by the authors, where caprines dominated. This warns against reading a ritual deposit as a straightforward menu or census of local herds. Preservation and anatomical selection also affect what can be counted. The authors explored relationships between herding, territorial resources, collective identity and ritual building; those interpretations connect several kinds of evidence rather than translating the stones into a complete prehistoric religion. Read the original 2024 comparative study.
2026: revising the cattle-cult label
The September 2026 PLOS ONE report investigated a concentration of 28 mustatils south-east of Al-Thamad on Harrat Khaybar. The team surveyed the group and test-excavated three structures. Selected upper cranial remains again supported the interpretation of offerings, extending the excavated comparison beyond AlUla.
Crucially, the authors revised an earlier suggestion that cattle themselves represented the central divine focus. Further finds showed that horned ruminants more broadly were selected. They distinguished animals offered to an unknown recipient or purpose from animals necessarily worshipped as divinities. The change concerns what the evidence warrants, rather than the discovery of an inscription naming a god.
They also discussed a possible chronological spread from AlUla towards Khaybar, and possible connections with changing water and pasture. These remained hypotheses based on a small fraction of the recorded monuments. Repeated forms and selective deposits support a coherent archaeological argument for ritual use; they leave substantial questions about origins, movement, meaning and the differences between particular communities. Read the 2026 findings and explicit revision.