The Enigma of the High-Level Aqueduct to Jerusalem and the Mamilla Water System

Abstract

In the Early Roman period the High-Level Aqueduct conveyed water to Jerusalem. The widely accepted view has been that before reaching the city, the aqueduct made a detour to the Mamilla Pool and then merged with the Mamilla Street Aqueduct on its way to the city. The article argues that this route is implausible. It presents data from excavations that set a consistent dating of the Mamilla water system to the Byzantine period. The Mamilla Pool and the Mamilla Street Aqueduct constituted a stand-alone water supply system that merely collected run-off water outside the urban area. The final stretch of the High-Level Aqueduct remains unknown.

Keywords:

• Jerusalem
• High-Level Aqueduct
• Mamilla Pool
• Mamilla Street Aqueduct
• Water systems
• Roman aqueducts

Acknowledgments

The author wishes to thank David M. Jacobson and Tsvika Tsuk for reading the first draft of this article and making valuable suggestions. The author expresses his gratitude to the editor and anonymous reviewers for their remarks. The author is grateful to the Israel Antiquities Authority for permission to publish the results of the excavation of the Mamilla Pool, with special thanks to Judith Ben-Michael and Nataly Zack for assisting in the preparation of archival materials for publication. Figures 3–4 are based on the draft sketches from the archival file: IAA Archives/Excavations/permit A-3100/1999, after R. Abu-Raya and Y. Billig,

Notes

1 All quotes from publications in Hebrew presented here were translated by the author.

2 Herod’s palace occupied a large area south or southeast of the Ottoman Jaffa Gate, but its exact location is still disputed. The highest elevation in this area is 771 m. Nevertheless, it was not necessary to convey water to the highest spot in the palace. For a topographical map of the area, see Sion and Puni 2010: 114 (note that the palace’s location there is a theoretical suggestion). Re’em (2010: 99; 2018: 38–53) exposed a large Herodian podium in the Kishle police compound, which he regarded as a possible foundation for the palace. The limited scope of his excavation does not allow for the drawing of definite conclusions regarding the nature of the remains.

3 The aqueduct was exposed north of the ‘Pillbox junction‘, between Asher Viner Street and Revadim Street, ca. 150 m east of Hebron Road (Billing and Dolinka 2012: 247–248).

4 Note that Billig and Dolinka (2012: 248) quoted the elevation of 762 m; Amit (2009a: 99) quoted 760 m.

5 Solimany (2012), Gibson (Amit and Gibson 2014: 26, 28, n. 89); contra, see Billig and Dolinka (2012: 248); Amit (2009a: 99). Amit and Gibson (2014: 28, n. 89) suggested that St. Claire’s conduit was only a side branch of the High-Level Aqueduct.

6 The width of the floor was measured in Area B. The floor was not found in Area A.

7 In fact, the destination of an inverted siphon or stone pressure pipe was usually lower than its header due to practical physical reasons, such as friction in the pipes that results in loss of energy (Hodge 1992: 153).

8 According to Vitruvius, a siphon must be set on a built foundation (De Arch., 8.6.5). Indeed, such a foundation is preserved in the siphon section near Bethlehem, but nothing similar has been discovered on the suspected valley-crossing route to the Mamilla Pool.

9 Figure 5 here is based on Amit (Fig. 6 in 2009a: 99). Note that on the original Amit drawing, the conduit is connected with a much longer presumed channel leading to the Mamilla Pool. Another presumed end of the conduit enters into the mapped area from Solomon’s Pools. However, the actual excavated conduit constitutes only a short fragment of the illustrated water channel (as shown in the figure).

10 Remains of an Iron Age irrigation channel, which collected run-off water, was unearthed in the same area (Amit 2009a: 107), hence such terrain-dependent functionality could continue accross many periods.

11 The lowest elevation of the bottom of the Mamilla Pool is ca. 767.5 m; the elevation of the bottom of Hezekia’s Pool is 760 m. The Mamilla Pool–Hezekia’s Pool water system functioned until the early 20th century. In 1864–1865, the Ordnance Survey of Jerusalem photographed Hezekiah’s Pool filled with water up to its rim (Wilson 1865:30).

12 Abbels and Arbit (1995:41–44) suggested that the Mamilla Pool and the Mamilla Street Aqueduct were a separate water system, though they dated it incorrectly to the Hasmonean period (2nd–1st centuries BCE).