Two sections of the Lower Aqueduct (Area A—length 34 m; Area B—length 15 m; Fig. 3) and a agricultural terrace retaining wall (W20) to the north of them were revealed in the two excavations. Four construction phases (1–4; Figs. 4, 5) were discovered in the aqueduct; these were built above each other rather than next to each other, as is the case on the western side of the valley. They range in date from the Roman period to the Ottoman period. The walls of the aqueduct from the early phase (4) served as a foundation for the later phases that were built on top of them.
Phase 4. A short section of the earliest phase of the aqueduct (L27; length 1.4 m, width of aqueduct’s channel 0.2 m, depth 0.16 m; Fig. 6) was exposed beneath the conduit’s later phases. This section of the aqueduct was built in a straight line, along an east–west axis. It was founded on chalk bedrock that was quarried level. The walls of the aqueduct (W5, W6) become wider toward the outside, rendering them a conical cross-section. They were built of white limestone and were treated on the inside with white plaster mixed with gravel and grog inclusions. The channel floor (L38) consisted of a layer of hard plaster applied on top of small fieldstones. No datable finds were discovered in the aqueduct, yet it is clear this section of the installation predates Phase 3. Therefore, the aqueduct was presumably constructed in the Early Roman period.
Phase 3. A section of this aqueduct’s channel (length 3.75 m, width 0.4 m; Figs. 6, 7) was exposed above Phase 4. The aqueduct was rectangular in section and its width was double that of the earlier aqueduct. The channel floor (L25) was founded on a casting of hard, light gray plaster (L30) mixed with grog inclusions, charcoal, lime and small stones; this casting was poured into the channel of the aqueduct of Phase 4. The level of the channel’s floor was c. 0.13 m higher than that of the channel in the Phase 4 aqueduct. The channel was bounded by two walls (W1, W2), both exposed for the entire length of the excavation above the walls of the aqueduct of the previous phase (Fig. 8). Walls 1 and 2 were built of various-sized fieldstones and were treated on the inside and outside with gray and white plaster. A low retaining wall (W3; Fig. 9) was discovered c. 0.5 m north of and parallel to the aqueduct, where the conduit crossed the valley in the western part of Area A. The retaining wall was built of medium and large fieldstones, whose dark color stood out prominently in contrast with the light color of the aqueduct. This retaining wall was also exposed in the eastern part of the valley (Solimany 2012; W1). It seems that the wall was meant to prevent erosion from accumulating against the aqueduct, and may have also demarcated its boundary. A Greek inscription from the Byzantine period, which was discovered in the vicinity of Bethlehem, tells us about the marking of the Lower Aqueduct. The inscription states that it is forbidden to plant a tree or to sow within a distance of fifteen feet from the aqueduct, and that anyone who disobeys this order is liable to be executed and his property confiscated (Abel 1926; Mazar 1989:189).
Several non in-situ coins were discovered on the surface and in the soil accumulation near the aqueduct. These included several coins minted by Alexander Jannaeus or imitations (IAA 143130, 143131), a coin minted by an unknown Hasmonean ruler (IAA 143129) and a coin dating to the reign of the Byzantine emperor Marcian (450–457 CE; IAA 143318). Several potsherds dating to the Early Roman, Late Roman and Early Byzantine periods were discovered in the foundations of Retaining Wall 3 and W1 of the aqueduct, indicating that Phase 3 of the aqueduct was not constructed prior to the Byzantine period.
Phase 2. In this phase, a high, broad retaining wall (W4; width 1.4 m, max. height 2.1 m; Figs. 10, 11) built of large stones bonded with white mortar was exposed. It was founded on brown soil fill (L29) meant to level the inclined surface and maintain the minimal slope of the aqueduct. A layer of white plaster containing flint gravel and grog inclusions was applied to the outer face of the wall. Walls 1 and 2 of the aqueduct in Phase 3 were renewed above the retaining wall. In this phase a black terracotta pipe (L14b; see Figs. 7, 9) was inserted into Channel 26 of the aqueduct from Phase 3. The pipe (inner diam. 0.15–0.17 m, length c. 0.41 m; Fig. 12:1) was composed of sections, one side of which was wide and the other narrow. The sections were set in place with their narrow side facing northwest and their wide side toward the southeast, while the water flowed to the southeast. The pipe connections were sealed with white plaster that hardened after the sections were assembled. The pipe was covered on all sides with mortar consisting of light gray earth, lime and pipe fragments (L34; see Fig. 7), which held the pipe sections in place. No inspection openings were discovered in the pipe. Travertine sediment was found inside the terracotta pipe, evidence of the prolonged use of the pipe and the difficulty in cleaning it. It was presumably due to this difficulty that the pipe was blocked and the water supply to the city was disrupted, leading to its replacement by a wider one that could be cleaned (Phase 1).
A coin of the Mamluk ruler el-Mansur Mahmud (1361–1363 CE, Hammat mint; IAA 143128) was discovered between the second and third courses in Retaining Wall 4, indicating that this section of the wall was built or renovated after the mid-fourteenth century CE. The historical sources from the Mamluk period do not mention that a terracotta pipe was inserted into the aqueduct to Jerusalem, and therefore it is difficult to determine if the pipe was installed during this period or at the beginning of the Ottoman period. The sources do state, however, that several renovations of the Lower Aqueduct were carried out during these periods (Salama and Zilberman 1986; Mazar 1989:189–190; Amit 2013). Phase 2 might have been associated with the last renovation implemented in the aqueduct during the reign of the Mamluk ruler Qa’itbay in March 1483 CE, by the governor of Syria Kanazawa el-Yahyawi, who was also a high official in Jerusalem (Ashtor 1955:82–83; Salama and Zilberman 1986:93).
Phase 1. A brown terracotta pipe (14a; see Figs. 6, 7, 9) that replaced the pipe of Phase 2 was exposed several centimeters above Pipe 14b. It was installed above a fill of fine-grained soil (L31; see Fig. 7) that was deposited on the covering of the earlier pipe (14b). The pipe was covered with mortar consisting of whitish-yellowish soil, small stones and lime. The pipe was assembled from sections that were wider and shorter than those of the previous pipe (inner diam. 0.21–0.26 m, length c. 0.38 m; Figs. 12:2; 13–15). The connections between the sections were sealed with white plaster. A channel (L5; Figs. 16–18) leveled with white plaster (L14 in the west, L19 in the east) was constructed directly above the terracotta pipe. The floor of the channel was 0.88 m higher than the level of the floor of the earlier aqueduct from Phase 4 and 0.75 m higher than the level of the floor of the aqueduct from Phase 3. Several sections of Walls 1 and 2 of the aqueduct’s earlier phases were raised. In the curved sections of the aqueduct, round inspection openings (diam. 9–10 cm) were drilled in the floor of the upper channel and in some of the terracotta pipe sections beneath it. These were meant to unclog blockages and release air pressure from the pipe (see Fig. 16). These openings were sealed with round fieldstones and pipe fragments affixed with white plaster. When it was necessary to repair the pipe or unclog a blockage, the inspection openings were opened in the pipe and the water flowed in the channel above it until the blockage was freed. This way, an uninterrupted water supply was ensured at all times, and it was also possible to use the channel without the pipe.
Large stone covering slabs (L21; see Figs. 16, 17), which were designed to guarantee the cleanliness of the water, were discovered on one of the sections of the channel. Similar stone coverings were found at many point along this aqueduct. Numerous fragments of pipe sections (14a) were discovered beneath the covering slabs (L6) in the alluvium that had accumulated inside the channel.
A section of the terracotta pipe from this phase was discovered in the Lower Aqueduct west of the Sultan’s Pool. This pipe supplied water to a sabil constructed on the bridge on the Hebron Road in 1536 in honor of Suleiman I (the Magnificent); it may have been used even earlier (Be’eri 2011). The renovation of the Lower Aqueduct and the construction of the brown terracotta pipe were completed five years later (1541) by Muhammad Çalbi en-Nakash, the Ottoman official responsible for maintaining the aqueduct that conveyed water from Solomon’s Pool to Jerusalem (Salama and Zilberman 1986:94).
Terrace Wall. An agricultural terrace retaining wall was discovered in Area B, c. 15 m north of the aqueduct (W20; Fig. 19). It was built of dark flint stones, along a north–south axis, and was preserved to a maximum height of five courses. The terrace it supported was evidently destroyed in the modern era. Fragments of pottery vessels from the Late Roman and Byzantine periods were found in the alluvium that had accumulated east of the wall and at its foundation, indicating the wall was not erected before those periods.