Baalbek
Lebanon · not visited, written from published excavation reports
Three stones of 750 to 800 tonnes each sit in the podium of a Roman temple. A kilometre away, in the quarry, lie stones that are bigger still, and one of them was only found in 2014. Moving them is mostly understood. Why the temple points where it does is not.
- Latitude
- 34.0069° N
- Longitude
- 36.2039° E
- Temple of Jupiter
- 1st century CE, c. 60 CE
- Largest quarry stone
- c. 1,650 t (estimate)
- Open questions
- 4 logged
What is actually there
Baalbek lies in Lebanon's Bekaa Valley, on the site the Greeks called Heliopolis, City of the Sun. Alexander is said to have renamed the city after his conquest in the 330s BCE. The mound under the Roman sanctuary has been inhabited for something like 8,000 to 9,000 years, and the Jupiter temple probably replaced an earlier temple on the same foundation. That earlier phase is a probability, not a documented fact. The Temple of Jupiter itself was built in the first century CE and was probably finished around 60 CE.
In the western wall of its podium sit three blocks known since at least the early Byzantine period as the Trilithon, though they are not a trilithon in the modern sense. Each is about 19 metres long, 4.2 metres high and 3.6 metres thick, and weighs an estimated 750 to 800 tonnes. They rest on a course of stones of roughly 350 tonnes.
In the quarry, a kilometre away, are stones that never left.
The Stone of the Pregnant Woman (Hajjar al-Hibla) is about 20.5 metres long and was never fully buried. A geodetic survey by a team from Linz in 1996 put it at about 1,000 tonnes. It was apparently abandoned because of poor stone along one edge, which would have made it prone to breaking in transport. The Stone of the South, rediscovered in the same quarry in the 1990s, is estimated at 1,242 tonnes. In 2014 the German Archaeological Institute, with the Lebanese University and the Lebanese antiquities authority, announced a third block in the stone layer beneath the Pregnant Woman: about 19.6 metres long, 6 metres wide and at least 5.5 metres high, estimated at 1,650 tonnes. One narrow side had already been worked smooth. It is widely reported as the largest ancient monolith known, though different reports give different weights, from about 1,500 to 1,650 tonnes, so treat the number as an estimate.
The quarry lies slightly higher than the temple, so the stones did not have to be lifted up to it. The accepted picture is rollers, capstans and temporary earthen banks, and the engineer Jean-Pierre Adam estimated in 1977 that around 500 workers could have moved a block of some 557 tonnes this way. That is a model, not an observed result. The final setting of an 800-tonne block high in a wall is a separate problem, and I found no ancient text that describes how it was done.
The sky here
With a name like City of the Sun, you expect a temple built on a solar date. Giulio Magli's survey (arXiv, 2016, from satellite imagery) gives the temple an azimuth of 75° 30′ and a horizon height of 5°, which puts the declination at about 14° 44′. That is inside the solar range, so the sun rises along the axis twice a year, around 1 May and 12 August. Magli himself says those dates do not seem to have any special meaning for the cult.
His alternative: the Pleiades, whose heliacal rising fell around 5 May in 15 BCE and whose declination matched the axis to within about one degree. Hesiod ties them to the cereal harvest, and the cult image of Jupiter Heliopolitanus carries ears of grain, a kalathos headdress, a whip and two bulls, which Magli reads as a fertility god. He also concedes the “distinct possibility that this alignment occurs just by chance.” Precession would have drifted the match away over the decades. I treat this as one scholar's hypothesis.
I removed a claim that the Temple of the Muses aligns with the summer solstice sunrise. It is not in Magli's paper, and I could not find a source for it.
The name says sun, the axis says something else, and the iconography says grain. What if the Greek label Heliopolis was laid over an older local cult of fertility and the harvest, and the temple's orientation is the one part of the old cult that was never renamed?
Why it is worth asking: because the mismatch between the name and the axis is real, whatever the cause. What supports it: the grain attributes of the cult image, and a site occupied long before Rome. What speaks against it: Magli's own caveat that the match could be chance, and that the axis may simply follow the podium it was set on. What would tell the readings apart: a firm dating of the earlier phase beneath the Roman platform, and a survey of other temples in the Bekaa Valley for the same bearing. What stays unknown is whether anyone ever chose the direction on purpose.
What is still open
How much of the podium is older than the Roman temple?
Magli argues, from the lack of structural connection between the two podia and Herodian-style features in the first, that both belong to a single Herodian-era project, begun around the founding of the Roman colony in 15 BCE. A column drum used in a lower course may belong to that phase, unless it is a later repair. This is an argument from architecture, not a dated sequence. Datable material from beneath the foundations would test it.
How was an 800-tonne block set into a wall?
Rollers, banks and capstans explain the quarry journey on paper. The final placement, with the tight joints visitors have noted, has no surviving description. A full-scale experiment would be the only real test, and none at this scale has been done.
Why were the giant stones left in the quarry?
For the Pregnant Woman the team points to poor stone on one edge. For the 2014 block the German team said it was still investigating, and its waste dumps were being searched for datable pottery. I found no published result. Until then, the date of the quarrying itself rests on assumption, not measurement.
Is the temple's orientation a choice?
Pleiades, harvest, an inherited axis, or chance. A comparison with other temples in the valley, and with the pre-Roman sanctuary if it can be recovered, would show whether this bearing repeats.
Sources
- Magli, G., “Archaeoastronomy and the chronology of the Temple of Jupiter at Baalbek”, arXiv:1606.05888, 2016. arXiv. For the azimuth, the Pleiades hypothesis, and the Herodian chronology argument. I read the full text through a summary tool, not line by line.
- Kropp, A. and Lohmann, D., “Master, Look at the Size of Those Stones!”, Levant 2011, cited by Magli for the Herodian features of the first podium. I did not read it myself.
- German Archaeological Institute (DAI), Baalbek project, announcement of the 2014 quarry find, as reported by HeritageDaily and Smithsonian. Press reports; the two differ on weight.
- Wikipedia, Baalbek Stones and Baalbek, for the dimensions, the 1996 Linz survey, Adam's 1977 estimate and the occupation history. A secondary summary; several of its statements carry no citation.
Not visited in person. Written from published archaeological sources, and marked as such. Last reviewed 12 October 2026. Corrected on this date: the location was given as Türkiye instead of Lebanon, the Temple of the Muses solstice claim was removed as unsourced, and the abandonment flaw was attributed to the correct stone.