The excavations were conducted in several places inside a magnificent Herodian structure to the west of Wilson’s Arch, and inside Room 8 of the Great Causeway, 30–50 m west of the Temple Mount and 6–15 m below the surface level of the Street of the Chain (Bab el-Silsileh), near the confluence of the Tyropoeon Valley and the Transverse Valley (Figs. 1–3). The dimensions of the Herodian structure are impressive, and it comprises two splendid rooms (21 and 23), one on either side of an elaborate fountain (22) and a water reservoir (112). The eastern room (23) was discovered as early as 1867 by C. Warren, who named it the ‘Masonic Hall’ (Warren and Conder 1884:193–209). In the twentieth century, it was investigated by W. Stinespring (1967), M. Ben-Dov (1985:178–180), as well as D. Bahat and A. Meir (Bahat 2013:122–128) and was renamed the ‘Hasmonean Hall’ or the ‘Herodian Hall’. In 2007–2008, A. Onn excavated the building’s western room (21), and in 2010–2012 he discovered the fountain (22) in the center of the building and the reservoir (112; Onn, Weksler-Bdolah and Bar-Nathan 2011; Onn and Weksler-Bdolah 2016). The structure was identified as a triclinium—a magnificent public building for the use of municipal dignitaries and their guests (Fig. 4; Onn, Weksler-Bdolah and Patrich 2016; 2019).

Towards the end of the Second Temple period, the structure underwent drastic changes, mainly the cancellation of the fountain, the division of the structure into three separate spaces and the construction of plastered water installations in the rooms, including a large miqveh in Room 21 that was connected to two earlier, south–north vaulted tunnels (230, 240).

After the destruction of Jerusalem in 70 CE, and particularly between the second and the early fourth century CE, the site underwent renewed construction. The plastered water installations of the Second Temple period hewn in the floors of the Herodian building were blocked with soil and stone rubble. The Great Causeway was built above the Herodian building; it comprised two rows of arches, one alongside the other, which together supported a street leading to the Temple Mount compound (see Fig. 2). To the south of the Great Causeway, large buildings were constructed that demarcated the remains of the Herodian structure to the south. The structures from the Late Roman period were aligned to correspond with the axes (orientation) of the Temple Mount, unlike the Herodian structure whose construction preceded the expansion of the Herodian Temple Mount.

In the Late Byzantine and Umayyad periods (sixth–eighth centuries CE), narrow industrial pools were installed in Room 21. During the Abbasid period (ninth–tenth centuries CE), the pools and the entire space of the room were blocked with soil.
The current excavations took place in Rooms 802 and 804 of the Great Causeway, in Rooms 21 and 22 of the magnificent Herodian building, and in Vault 230 adjacent to the building on the northwest (see Fig. 2). The following is a description of the finds from the excavation, based on the stratigraphy established in Onn’s excavations (2007–2012) and revised in the current excavations (2018–2020; there may be slight changes in the final report; Table 1).
Table 1. The Principle Remains Found in the Excavation
Principle remains
Solid cast wall, probably a dam, incorporating aqueducts and tunnels
Roman period (second half of the first century BCE; there may have been several building phases)
15 (or 14)
Magnificent public building—triclinium
Roman period (first half of first century CE)
Renewed construction and alterations in the magnificent building
Roman (first century CE; 35–70 CE)
Blockage of Second Temple installations with soil fills; construction of the Great Causeway and buildings to south of causeway
Roman (second–fourth centuries CE)
Late Byzantine (sixth–seventh centuries CE)
Stratum 16: Massive Cast Wall Incorporating Drainage Channels and Tunnels (Roman period; no earlier than the second half of the first century BCE; Figs. 2, 3, 5)
Cast Wall. The earliest building remains from the Second Temple period discovered so far are those of a broad cast wall that was probably part of a dam, built on a southwest–northeast alignment (estimated width 14 m, height at least 7 m). Non-continuous segments of the wall were discovered along an approximately 100 m stretch to the west of the Temple Mount, beneath the arches of the Great Causeway (excavated by Hamilton and by Onn), and below Wilson’s Arch (excavated by Uziel, Lieberman and Solomon). The wall was apparently longer, as the western wall of the Herodian Temple Mount cuts it at its eastern end (for segments of the cast wall, see Hamilton 1932; Onn, Weksler-Bdolah and Bar-Nathan 2011; Onn and Weksler-Bdolah 2016; Uziel, Lieberman and Solomon 2019). The wall was constructed using the opus caementicium method, in which boulders and variously sized fieldstones were bonded together with hard mortar. The wall’s core was filled with boulders that had partially dressed faces and with building blocks in secondary use, all arranged beside each other in roughly leveled courses and bonded with hard yellow mortar. The top of each course was leveled with a compacted surface of small, flattened stones, bonded with hard mortar. Inside the core of the cast wall, drainage channels (L538, L540) and short vaulted tunnels for water (L230, L240) were incorporated when it was built; they were probably built together with the casting of the wall.
The northern face of the cast wall has not yet been uncovered in excavations by the Israel Antiquities Authority; it was discovered during Hamilton’s excavations. It presumably follows a line along the northern boundary of the Great Causeway, which corresponds to the northern boundary of Vaulted Tunnels 230 and 240. The cast wall’s southern face (W330, W577; Fig. 6) was uncovered in the current excavations in line with the facade of the Fountain wall in the Herodian building, c. 1 m to its south (c. 723 m asl); it lies c. 14 m south of its estimated northern face and parallel to it. The wall was found sealed beneath the paving stones of Room 22 in the Herodian structure.
The core of the cast wall was previously discovered in the bedding of Rooms 5, 6 and 8–10 of the Great Causeway (W5006 in Onn’s excavations; its top reaching c. 730 m asl). The current excavations revealed the highest level of the cast wall’s core (W905; Fig. 7) inside the eastern part of Room 804 of the Great Causeway. The core of the cast wall was also uncovered in the western wall of Vaulted Tunnel 230, beneath the vault’s courses (max. height 723 m asl); the excavation did not reach bedrock in Tunnel 230.
In some places, ‘seam lines’ form a vertical step in the core of the cast wall. These seam lines run from southwest to northeast, along the center of the cast wall’s axis (c. 7 m from the southern wall and 7 m from the northern wall). The cast wall may therefore have been built of separate sections consisting of large cast cuboid blocks (width 7 m; length and height unknown; the dashed line in Fig. 5 marks the estimated borders of the cast blocks). When the wall was built, using the method of casted blocks, the Channels 538 and 540 and the Vaulted Tunnels 230 and 240 were incorporated within it.
Channels 538 and 540 are two impressive drainage channels built one on top of the other, which run through and perpendicular to the cast wall (L538 [upper channel], L540 [lower channel]; width of each channel c. 0.75 m, height at least 2 m; Figs. 3, 8). The channels were not excavated due to safety considerations, but they were documented. Their walls were built of dressed stones, many with marginal dressing on all four sides and a smoothed boss; a few stones have a protruding boss. Some of the stones’ corners were broken and repaired with coarse gray mortar containing grits. The two channels are roofed with massive flat rectangular stone slabs and their appearance is impressive. Both were blocked in their northern part by collapsed building stones at the point of the estimated northern boundary of the cast wall, and they continue in a direct line to the southern boundary of the wall. To the south of the cast wall, the channels continued for another 3 m within a leveled stratified stone fill, whose composition resembles that of the cast wall itself. Approximately 3 m to the south of the cast wall, the two channels were blocked by a later wall. The appearance of the channels, their impressive construction and their incorporation in the cast wall suggest that the channels and the cast wall were built at the same time; further evidence for this can be found in a dressed stone that is both part of the cast wall’s southern face and incorporated in Channel 538’s eastern wall. Had the channels been cut through the cast wall at a later date, their walls would have probably been hewn and plastered, not built of ashlars. As the two channels documented across the cast wall were intact and undamaged, and neither canceled the other, they apparently existed simultaneously.
Vaulted Tunnels 230 and 240 are a pair of short vaulted tunnels (723–726 m asl) that were partially excavated by Onn; the current excavation reached a greater depth in Tunnel 230 (Figs. 3, 9). On removal the plaster from the walls of the vault in Tunnel 230, it became apparent that the built vault was incorporated in the core of the cast wall, with the core of the cast wall lying both beneath the foundation of the vaulted tunnel and above its highest point (Fig. 10). The tunnel walls were built of ashlars with finely drafted margins and a flat boss. Later, probably in Stratum 13, the floor of Tunnel 230 was apparently deepened by quarrying into the cast wall, and an arch (W20) incorporated in the cast wall was built at its southern end. Onn’s excavations found that the top of the cast wall was c. 730 m asl, indicating that the tunnels were incorporated deep within the core of the cast wall, like Channels 538 and 540.
Stratum 15: Magnificent Public Building—Triclinium (Roman period; first century CE; Figs. 4, 11)
Prior to the building’s construction, the top of the cast wall was lowered, and Channel 538—which originally reached higher than the building’s floor—was put out of use (see Fig. 3). A square vertical shaft hewn in the base of Channel 538, near the south face of the Fountain wall (W2203), carried the water from Channel 538 down into Channel 540. A blocking wall was built across Channel 538 to the south of this shaft; the southern part of Channel 538, beyond this blockage, was plastered and turned into drainage channel for water from the building’s fountain (L515; depth 1 m; Fig. 12). The water flowed from the room’s floor (L504) down a chute and into Channel 515.
The current excavations dismantled the later walls along the east and west sides of Room 22, which were built in Stratum 13 (below; Figs. 2, 3: W2202E, W2202W). After they were dismantled, the lower parts of three previously unknown piers (P1, P2, P6; Figs. 3, 13) were uncovered in the building’s Fountain wall (W2203). The building’s original paving stones (L504; Fig. 14), which were well-preserved beneath the dismantled walls, were also discovered in Room 22. The paving stones covered an area of c. 2.5 sq m beneath the western wall (W2202W) and of c. 9 sq m beneath the eastern wall (W2202E). The paving stones found in the current excavation are large (length 1.0–1.7 m, width 0.7–1.0 m, thickness 0.25–0.30 m). Made from local limestone, they were finely dressed to form a perfect fit. Their faces are smoothed with fine combing and without marginal dressing. In the eastern part of the paving at the base of the Fountain wall, several shallow narrow channels were hewn in the paving that drained into an opening leading to Drainage Channel 515. The channels may have been hewn during a later phase of the building, as their appearance detracts from the fine paving. The floor at the base of the eastern part of the Fountain wall slopes slightly southward and ends here in a shallow channel with a low railing hewn as part of the paving. The channel and the railing continue along a parallel alignment with the Fountain wall and c. 3.5 m to its south. They were probably intended to stop water from flowing on the floor (see Fig. 13). The channel floor contained a white limestone sediment that was probably deposited by the standing water.
In the center of Room 22, the eastern corner of a built and plastered installation (L530) was found. The installation may have been a pool for collecting water from the fountain, which was sunk into the stone foundation beneath the building’s floor. The relationship between the installation and the room’s paving (L504), which was preserved to its east and at a higher level, is not clear.
Approximately one meter south of the western part of the Fountain wall, a large, broken railing stone (c. 0.40 × 1.00 × 1.25 m; Fig. 15) was discovered inside the core of the dismantled later wall (W2202W); it was hewn of hard limestone, and its surface was dressed smooth with fine combing typical of the Second Temple period. The railing has a flat face on its southern side and a convex face on its northern side, which is facing the Fountain wall. It is not possible to determine whether the railing stone belonged to a draw basin at the base of the Fountain wall, which was built before the building was divided into three spaces, or if it was brought here in secondary use for the core of W2202W.
A coin of a Roman governor minted in 17–25 CE and pottery fragments dating from the first century BCE to the first century CE were discovered in the bedding of the paving in Room 22. These finds enable the dating of the triclinium’s construction to 17 CE at the earliest.
Stratum 13 (Roman period; first century CE; Fig. 16)
The area inside the magnificent structure underwent considerable changes during this phase, some of which had previously been discovered (Onn and Weksler-Bdolah 2016). The current excavation indicates that the paving of the structure in front of the fountain was damaged. This may have occurred when the railing stones, which may have resembled the only preserved railing stone found at the site, were removed. Later, the pavement was repaired with various stones, including architectural elements that were probably taken from lavish buildings from the Second Temple period that had already been demolished, e.g., two fragments of coping stones with a narrow, rounded edge, similar to the upper courses of the Temple Mount wall known from the destruction rubble at the foot of the Temple Mount (Fig. 17), and a fragment of a Doric cornice typical of the splendid Second Temple Herodian architectural style (Fig. 18). In this phase, the structure’s interior was divided into three separate rooms (Rooms 21–23) that were roofed with vaults, either at that stage or slightly later. The floor of Room 21 was dismantled, and a stepped and plastered pool—possibly a miqveh—was installed in its place. In the southern part of Room 21, above the upper preserved step of the miqveh, was a large ashlar bearing a north–south groove on its top, which may have served as a partition stone in the miqveh (Fig. 19). The removal of the northern wall of Room 21 and the deepening of Tunnel 230 connected the deeper northern part of the miqveh to Vaulted Tunnels 230 and 240 to its north. An arch (W20; see Fig. 9) was built along the line of the northern wall of Room 21. To date, the function of Arch 20 cannot be determined with certainty. It may have been built before the northern wall of Room 21 was dismantled, so as to support the wall when its foundation was partially dismantled.
Strata 8–10 (Roman period; second–third and early fourth centuries CE)
In this phase, the installations of the earlier strata were blocked with soil fills. A probe (1.5 × 2.7 m) in the southern part of Tunnel 230, to the north of Arch 20, revealed a soil fill in which several white surfaces were discerned, probably meager floors abutting the vault’s plastered wall. The soil fill yielded abundant fragments of pottery and roof tiles dating from the second–third and early fourth centuries CE. An exceptional find from the fill is a roof tile impressed with a private stamp: the Latin name of the potter or the workshop owner (CILO) written in Greek letters (ΚΕΙΛΛΟ) (Fig. 20). This stamped tile joins a growing group of tiles, bricks and clay pipes bearing private and municipal stamps that have been discovered in Jerusalem and date from the time of Aelia Capitolina, adding to the well-known corpus of legionary stamps on building material from the legionary workshop at Binyene Ha-Umma (Weksler-Bdolah et al. 2022).
Stratum 7A (Late Byzantine period; sixth–seventh centuries CE)
Remains of this stratum were discovered in Room 802 of the Great Causeway and in Room 21 of the magnificent building (see Fig. 2). A built, plastered water channel (width 0.45 m, depth 0.32 m) discovered in Room 802 ran from north to south and apparently continued along the route of a similar channel from the Byzantine and Umayyad periods, previously uncovered in Onn’s excavations in Room 804. The channel’s alignment corresponded with the eastern wall of Room 802 and with a reinforcing wall built along the eastern wall of Room 804. The channel was found without covering slabs and filled with soil containing Early Islamic pottery.
A pair of long, narrow industrial pools was revealed in Room 21, each of which had at least two phases of use dating from the late Byzantine period (sixth–seventh centuries CE). The eastern of the two pools was built along the entire length of Room 21, while the western pool was about 1 m shorter at its southern end. A narrow flight of steps leading to the bottom of each pool was installed in their corners in a late phase of their use. A coin of Heraclius (629–639 CE) discovered in the core of the wall separating the two pools dates their construction to no earlier than the late Byzantine period.
The remains and finds uncovered in the excavation make it possible to propose more accurate dates for the late Second Temple-period building phases identified to the west of the Temple Mount. Between the second half of the first century BCE (the Herodian period) and the beginning of the first century CE, a long, broad cast wall was built that evidently dammed the streams in the Tyropoeon Valley and the Transversal Valley. A road that was apparently paved on top of the dam wall led to a gate in the wall of the pre-Herodian Temple Mount (probably the Kiponos Gate mentioned in historical sources). Drainage channels and vaulted tunnels were incorporated across the cast wall and inside its core. The construction of the wall, including the channels and tunnels inside it, are probably related to architectural layout and planned drainage of the area to the west of the Temple Mount prior to the expansion of the Herodian Temple Mount. In the first third of the first century CE, a lavish structure was built at the foot of the dam wall and over part of it. The structure included two triclinia with a magnificent fountain between them. The construction of this structure required the lowering of the top of the dam wall and the cancellation of one of the drainage channels that ran across it. The new finds correspond to the overall picture provided by previous excavations.