Antequera · Spain · 37.0244, -4.5464 · PYR record in the living atlas
For most of the last century the guidebooks said the great capstone of the Dolmen de Menga weighs about 180 tonnes. In December 2023 a team led by José Antonio Lozano Rodríguez, with Leonardo García Sanjuán and Francisco Martínez-Sevilla among the authors, published the first proper petrological and volumetric study of the monument in Scientific Reports and put the figure at 149.59 tonnes, give or take 9.17. That is still the largest stone ever set in a composite megalithic monument in Neolithic Europe, beaten only by the Grand Menhir Brisé in Brittany at around 335 tonnes, which is a single standing stone rather than part of a building. It is also 30 tonnes lighter than the number most sites still print, and the higher figures in circulation do not rest on any published structural analysis.
Somebody finally weighed it
The same study identified five distinct rock types in the monument and traced them to outcrops within about a kilometre, principally Cerro de la Cruz roughly 850 metres away and about 50 metres higher than the hill Menga stands on. The route is downhill at an average slope of around 22 degrees. That is not a small mercy: it is the reason the project was possible at all.
Published measurements of the chamber itself do not agree, and the difference is worth naming rather than smoothing over. The 2024 study measures the chamber at 24.9 metres long, 5.7 metres at its widest, and between 2.50 metres at the entrance and 3.45 metres at the back. Popular accounts, including the English Wikipedia entry, give 27.5 by 6 by 3.5 metres. The larger figures are the older ones and they include more of the structure; the smaller ones come from direct survey of the chamber. Where they conflict we would follow the survey. The chamber is lined by fourteen orthostats, seven a side, closed by a single backstone, roofed by five capstones and carrying three large pillars along its axis.
Soft stone, and what that forced them to invent
Menga's uprights and capstones are calcarenites and bioclastic calcirudites, classified in the 2024 paper as soft to moderately soft rock. This is the fact everything else follows from. You cannot haul soft calcarenite on rollers or lever it up an earth ramp without shearing pieces off it, so the builders did something else, and the traces of what they did are still in the ground.
The foundation sockets were cut at about 77.5 to 78 degrees rather than vertical, and roughly a third of each upright sits below floor level: about 2.75 metres of burial at the entrance, about 3.20 metres at the back of the chamber. The sockets themselves form a stepped terrace descending towards the entrance, which is what stabilises the whole structure on sloping ground. The stones lean inward, averaging 85.2 degrees on one side and 84.0 on the other, so the chamber's cross-section is a trapezoid, narrower at the roof than at the floor, which means every capstone has less span to cover. The lateral faces of adjacent uprights were dressed at matching angles so that they lock against each other in both directions, an arrangement the authors say has no precedent in contemporary megalithic building. The upper surface of capstone C-5 was shaped into an arch, which the paper describes as the first stone discharge arch ever built by human beings. And in place of external ramps the authors reconstruct a system of counterweights, tilting each upright smoothly into its socket to what they call millimetric precision.
The detail we find hardest to forget is smaller than any of that. The bioclastic stone contains fossil bivalve shells, and the overwhelming majority of them sit with their concave sides facing the inside of the monument. The builders were tracking which way up each block had lain in the quarry bed and setting it accordingly. That is a quarrying convention, invisible from three paces away, followed across a building of 32 stones weighing something like 1,140 tonnes in total.
The mound over the top is structural, not decoration. A 2022 multimethod study in Quaternary Research, which combined 12 radiocarbon dates, 11 optically stimulated luminescence measurements and six uranium-thorium dates for 29 new ages in all, describes a tumulus 48.3 by 44 metres with a lower layer of dark organic soil capped by about three metres of interlocked sandstone slabs and rammed clay. It braces the chamber, sheds water and keeps the interior dry, which is the only reason porous calcarenite has survived inside it for well over five thousand years. That study places construction in the Late Neolithic, broadly 3900 to 3400 BC; the 2024 paper narrows it to about 3800 to 3600 BC.
The face in the mountain
Almost every dolmen around the Mediterranean has its axis somewhere in the arc of sunrise. Michael Hoskin's long series of Iberian orientation surveys is what established that pattern, and it is also what made Menga's exception visible. Menga's axis runs at about 45 degrees, north-east, at the northern cliff face of La Peña de los Enamorados, a limestone crag whose profile is read locally as a human face lying on its back. That reading is a local and modern one rather than a documented ancient one, and the two are worth keeping apart.
Its neighbour Viera, about 21 metres long inside, faces sunrise near the equinoxes, which is the ordinary Iberian pattern rather than an exception to it. El Romeral, the later tholos of the group, stands about four kilometres from the other two and has two successive false-cupola chambers reached by a masonry corridor some 26 metres long. It is commonly reported as facing the El Torcal massif, and that much can stand as reported. A winter solstice midday alignment for El Romeral is repeated widely, and this article repeated it until now, but it is not established in the specialist literature, which describes the monument's intended target as anomalous and unresolved. We are withdrawing it rather than restating it.
The strongest evidence that the mountain was the point, rather than a coincidence of terrain, has arrived recently. Between September and December 2020 García Sanjuán and colleagues excavated a megalithic tomb at Piedras Blancas, at the foot of La Peña, and published it in Antiquity in 2023: a chamber 4.5 metres by 1.45 metres cut into bedrock and edged with slabs, in use from around 3185 to 2935 cal BC through to around 1900 to 1630 cal BC, with an arrow-like arrangement oriented on summer solstice sunrise for roughly 3400 cal BC. The same cliff carries the Matacabras rock shelter, whose schematic red paintings are dated to around or before 3800 BC, which is at or before Menga's own construction. The thing Menga looks at was already a painted place before Menga existed. People were not only looking at that mountain from the dolmen; they were building on it and marking it.
It is worth resisting the tidy version, though, and the tidy version cuts both ways. The 2024 study notes that at the summer solstice the left side of Menga's chamber stays in shadow while much of the right is lit, so the sun was not ignored. More than that: because La Peña lies to the north-east, the same 45-degree axis falls close to summer solstice sunrise anyway, and several researchers argue the builders chose a mountain silhouette through which the midsummer sun would rise. A 2026 archaeoastronomical study goes further and proposes that the design also captures moonrise at the northernmost lunar standstill. Mountain, sun and moon are all now on the table for one bearing, and which of them was primary, if any single one was, is not settled. What can no longer be said is the exclusive version, that Menga points at a landform and at no celestial target at all. A monument can do both, and this one appears to.
One caution about the landscape argument as a whole. No comprehensive area excavation of the wider site has ever been carried out. Cerro Marimacho, a hill about 200 metres east of Menga and Viera, has long been suspected of holding further prehistoric features; limited investigation during road construction turned up pits with pottery and flint, and nothing systematic has been done there since. The case for a deliberately planned ritual territory here is a good one. Most of that territory is still in the ground.
Nobody ever closed it
Menga has stood open on its hill through Roman Spain, through al-Andalus and through every century since. The first recorded investigation was Rafael de Mitjana's, a Málaga municipal architect who dug from about 1842 and published in 1847, identifying the monument as a druidic temple and reporting that he found no human remains at all. Other nineteenth-century accounts report the remains of several hundred people from inside the chamber. Those two statements are hard to reconcile, and we are not going to smooth them into one; what is certain is that nobody was recording to a standard that would let the question be settled now. Antonio and José Fuentes brought the Dolmen de Viera to notice in February 1903.
The paperwork is older than the archaeology. Menga was made a Spanish National Monument in 1886, Viera in 1923, and El Romeral entered the National Archaeological Artistic Treasure in 1926; a 2009 decree placed the whole area in the General Catalogue of Andalusian Historic Heritage, and UNESCO inscribed the site on 15 July 2016, at the 40th session of the World Heritage Committee in Istanbul. Modern excavation is thinner than that list suggests. The University of Málaga worked on the orthostat foundation sockets from the late 1980s into the early 1990s, and the University of Granada dug inside Menga in 2005 and 2006.
That 2005 and 2006 campaign found the strangest thing in the building. A well drops 19.4 metres through the floor at the back of the funerary chamber, reaching the water table, a hydraulic feature with no parallel in any other European dolmen, and apparently prehistoric. Apparently is carrying weight in that sentence. Proposed functions run from ritual or oracular access to water, through a structural or symbolic role in the monument, to the possibility that it was sunk much later by treasure hunters. Its age and its purpose are both unresolved in the published record, and the site's own management authority says so. It is currently the most active research question at the monument and it has no answer.
The best evidence for what the place meant to later people came from the 2005 excavation of the atrium, and has only recently been published. Two burials were found there, radiocarbon dated to between the eighth and eleventh centuries CE, when Antequera was part of al-Andalus. Both bodies were laid face down and carefully oriented on the monument's central axis, which is not how the Islamic cemeteries of the period arranged their dead. Only one, an adult man labelled Menga1, preserved enough DNA to sequence. His paternal lineage is a common European one, aspects of his maternal lineage sit close to modern North African populations, and genome-wide he resembles other Iberians of the Roman and medieval periods. Silva and colleagues published the analysis in the Journal of Archaeological Science: Reports. Four and a half thousand years after it was built, somebody still thought this was the right place to be buried, and still thought the axis mattered enough to line up with it.
What our engines found
We ran the same test here as everywhere else, with the settings fixed in advance and identical at every one of the fifty sites we cover: a 200 km search radius, at least 12 km between sampled sites, 15 sites, a 2 per cent tolerance, and 2,000 Monte Carlo runs against a null model resampled from real catalogued sites in the same region. The settings are frozen because the answer moves if you tune them, and tuning them after seeing the answer is exactly the trick we object to elsewhere. Antequera returned 50 golden-ratio pairs where chance predicts 38.9, giving p = 0.1759.
That is a little more than chance would predict, and it is not significant. The control points the same way: the arbitrary ratio pi/2 scored 1.38 times its expected count against phi's 1.29, so nothing here singles out the golden ratio.
We publish the misses because a tool that only reports its hits is not measuring anything. Every figure above is reproducible from the same coordinates, the same settings and the same seed.
One limit worth stating plainly, because it is the one most often glossed over. This test compares positions between monuments. It says nothing about any building's own proportions, its height, base, slope or perimeter. That is a separate question, and at several sites a genuinely open one.
That includes the question this article spends most of its time on: the bearing of Menga's corridor is a property of one monument, and a test on distances between monuments cannot speak to it.
Explore it yourself
Open the Antequera record in the living atlas, stand in the chamber in 3D, sight down the corridor to the Lovers' Rock and see the three monuments among their neighbours. The Antequera site-emblem tee carries the great chamber in the Pyramason Chronicle style, with the site name and the exact coordinates.
Sources worth reading
Lozano Rodríguez, García Sanjuán, Martínez-Sevilla and colleagues, "The provenance of the stones in the Menga dolmen reveals one of the greatest engineering feats of the Neolithic", Scientific Reports (2023), the source of the 149.59 tonne figure. Lozano et al., "Early science and colossal stone engineering in Menga", Science Advances (2024), the engineering and geometry study behind most of the second section above. García Sanjuán et al. on the Piedras Blancas tomb, Antiquity (2023). The 2022 multimethod dating study of the tumulus in Quaternary Research. Silva et al. on the Menga1 medieval burial, Journal of Archaeological Science: Reports. Michael Hoskin's Iberian orientation surveys, which established the sunrise pattern Menga departs from and are the origin of the whole question.