The Roman Siege Camps and Circumvallation Wall

The 2024 survey measured the circumvallation wall at 4,290 metres, and about 6,270 metres of built wall once the camp walls outside it are added. Where both faces still stand and are not buried, the wall is between 1.8 and 2 metres wide. Its original height was not measured but calculated: the 3D model gave the volume of stone in a segment and of the collapse lying beside it, and dividing by length and width produced a weighted mean of 2.2 metres, with a few stretches at 2.4 to 2.5.

Two of the camps are much larger than the rest. Camp B on the east and Camp F on the west are given in the World Heritage nomination file as 135 by 170 metres and 125 by 150 metres, described there as very similar in size and plan to the classic camp of the Roman legions, with Camp F named as Silva's headquarters. Both sit outside the siege wall, with the wall between them and the mountain. Inside all eight camps are hundreds of contubernia, the mess units of eight or nine men who shared a tent: low rubble walls about 1 to 1.5 metres high, built as a footing over which the tent was stretched.

The towers are the feature that has been counted most often and agreed on least. The 2024 team identified fifteen, every one of them on the eastern and northern sections of the wall, protruding about 0.9 to 1.5 metres from both faces, with gaps between towers 2 and 13 ranging from 61 to 110 metres and averaging 90. Earlier counts differ: Gichon and Yadin identified twelve, Arubas and Goldfus roughly thirteen, Richmond fourteen.

• Circumvallation wall: 4,290 m measured, 1.8 to 2 m wide, about 2 to 2.5 m high when built
• Camp walls not incorporated into it: about 1,980 m more
• 15 towers, all on the eastern and northern stretches
• Camp B (135 × 170 m) east and Camp F (125 × 150 m) west, both outside the wall
• Hundreds of tent bases inside the camps, each for eight or nine men

Jerusalem fell in 70 CE. Masada was among the last places in Judaea still held against Rome, and in 73 or 74 CE the governor Lucius Flavius Silva brought the Tenth Legion Fretensis against it. The parks authority's own site brochure, written by Guy Stiebel, puts the force at 8,000 troops including auxiliary units, and records that captive Jews carried water to them, apparently from Ein Gedi, along with food. There is no spring anywhere near the mountain, a difficulty Josephus notes about the Roman camp himself.

Josephus describes the ring in one clause and gives no measurements at all: Silva "also built a wall quite round the entire fortress, that none of the besieged might easily escape; he also set his men to guard the several parts of it." The figures came much later. The nomination file the State of Israel submitted to UNESCO gives a siege wall about four kilometres long encircling the entire mountain; the 2024 survey put a number on it, 4,290 metres.

How long the work took has moved a great deal, and the old picture of a fortress holding out for years has gone. Joseph Roth argued in 1995 for a siege of four to nine weeks; he and Amnon Ben-Tor separately reckoned the wall could have been finished in under five days, a figure resting on a speech of Hadrian's claiming a soldier could build a cubic metre of wall a day, which the 2024 team considers exaggerated. Their own calculation gives about 26,700 cubic metres of stone, 50,000 to 80,000 workdays, and 11 to 16 days for 5,000 men. Cross-checked against the siege of Jerusalem, where a force five times larger built seven kilometres of wall and thirteen camps in three days, the same method gives about nine days.

Even the date of the fall is unsettled. The parks authority's brochure says 73 or 74 CE. The 2024 paper says the siege probably took place in 73/74 CE. The ICOMOS evaluation states 2 May 73, and Britannica gives 15 April 73 CE. Nobody has reconciled them, and the disagreement traces back to when Silva took the governorship, which the ancient evidence does not fix.

The nomination file states it plainly: "The Roman camps haven't been dug (except of probes)." ICOMOS, assessing the same property, called the siege works the most complete in the world and almost completely unexcavated. The preservation is a product of the climate, the remoteness, and the fact that nobody came afterwards to rob the stone. So the honest answer to what is original here is: nearly all of it. That is why the site was inscribed under criterion iv, whose wording calls the camps and fortifications encircling the monument the finest and most complete Roman siege works to have survived to the present day.

The exceptions are worth knowing. Camp A and Camp F have been excavated, along with the siege ramp: the nomination file records a season at Camp F and the ramp led by Gideon Foerster in 1995, and the 2024 paper credits that excavation to Benny Arubas, Haim Goldfus and Jodi Magness. Camp A was dug by Shmarya Gutman in the 1960s and its walls reconstructed in the process, which produces a small paradox: the tidiest-looking camp on the ground is the one whose walls are least original.

The modern intrusions have mostly been taken away rather than added. The nomination file records that the former entrance buildings, which had been scattered among the Roman camps and impaired their appearance, were demolished during the development project of the 1990s so the camps could be seen without modern construction in the way. ICOMOS noted at inscription that a service cable lift from the 1960s excavations passed directly above the ramp and through the breach, and that a project was under way to move it.

One thing has been lost to nature. The 2024 survey found long stretches of the eastern wall buried under sediment washed down from the mountain, which is why segments look missing on the ground. The authors use that finding against the argument, advanced from the Camp F and ramp excavations, that the siege ramp has barely eroded since the first century.

• Original and untouched: the circumvallation wall, the towers, the tent bases and six of the eight camps
• Excavated: Camp A, Camp F and the siege ramp
• Reconstructed: the walls of Camp A, rebuilt by Shmarya Gutman during his excavation
• Removed: the old entrance buildings that stood among the camps until the 1990s
• Buried: long stretches of the eastern wall, under sediment washed off the mountain

At about two metres wide and two to two and a half metres high, in fieldstone and without a parapet, the circumvallation could not have withstood a serious assault, and it is easily scaled. It is also taller than a person, so a soldier could not fight from behind it. The 2024 authors conclude that it worked as an obstacle to slow a sortie, and as a fighting platform only when soldiers stood on top of it, expected to counterattack rather than hold. Gutman found five staircases running up to the tops of the walls at Camp A, two of them on the rear wall.

Compare the works Rome built when it expected trouble. At Numantia in 133 BCE and Alesia in 52 BCE the siege lines had walls up to 4.7 metres wide, parapets, ditches, traps, and towers set within a bowshot of each other, about 31 metres. Masada's towers stand 61 to 110 metres apart, beyond mutual bowshot though still within catapult range, and there are no ditches or traps at all. The 2024 reading of that is straightforward: the Romans did not expect heavy fighting, from the rock or from anywhere else.

Where the towers do stand tells its own story. All fifteen are on the eastern and northern sections, every one beside a wadi or a deep gully, and several sit on stretches that bulge outwards rather than facing the mountain. That makes little sense for a wall watching Masada and a good deal of sense for a wall watching the desert behind it, against infiltration, deserters, or a relief force coming up the ravines from the east.

The strangest finding concerns the stretches that were never meant to stop anything. Above the western cliff and along the northern edge of Mount Elazar, where a 200-metre precipice already made the line impassable, the wall was built only in the lower dips, exactly where a viewer on the summit would need masonry to see an unbroken wall top. The 2024 authors take this as deliberate psychological warfare, and add that the audience may not have been only the besieged: a ring of works like this also had a job to do on the morale of the legionaries building it.

The nomination file makes the practical point better than any guide: of the siege system, "there is an excellent view of it from the mountain." That is the whole trick. From the plain the camps are grey stone on grey ground and hard to pick out at all; from the summit lookouts the plan lies out flat below, wall, towers, gateways and tent rows together. Early and late light is what separates them, because everything here is a shadow line about half a metre high.

The two big camps are on opposite sides of the mountain. Camps A, B, C and D and the towered stretch of wall lie on the marl plain to the east, below the cable car and the Snake Path. Camp E, Camp F with its later garrison enclosure, Camp G and Camp H are west and southwest, Camp H on the top of Mount Elazar and Camp F at the top edge of the western cliff, close to where the ramp path begins. The park brochure warns that there is no road between the two entrances and that driving round takes about an hour and a quarter.

None of the camps is a museum. There are no signs, no rails, no interpretation and no reconstructed anything except the walls of Camp A. The parks authority's site rules are to stay on marked paths and not to climb walls or damage antiquities, and that applies here more than anywhere on the mountain, because these low rubble courses are the actual first-century surface. The western trail forecourt, on the Arad side, has the closest thing to a display: the parks authority lists parking, toilets, a visitor service centre, a picnic awning and models of Roman weapons there.

One piece of the story sits above the ramp rather than below it. The breach the battering ram made in the casemate wall is still identifiable at the top, and the nomination file says it is clearly seen today. Josephus describes the bank raised to 200 cubits on an outcrop he calls the White Promontory, a platform of great stones 50 cubits square on top of it, and an iron-plated tower of 60 cubits set on that.

• East side: Camps A, B, C and D, the fifteen towers and the best-preserved wall
• West side: Camps E, F, G and H, plus the ramp and the breach above it
• Best viewing: the summit lookouts, in low sun
• No road between the two sides; about an hour and a quarter by car