The irrigation system was cleaned and documented, and a rock-hewn burial cave was documented below the reservoir.
Spring. A cavity that was intended to reach the layer of the aquifer was hewn inside what was probably at first a natural cave (L3; height 2.5 m, width of opening c. 5 m; Fig. 3). An elliptical cavity (2.0 × 3.3 m, height 1.35 m) was hewn in the southern end of the cave, thereby deepening its interior.
Tunnel. A channel (L2; width 0.75–0.80 m, height 2.2 m; Fig. 4) hewn northward from the opening of the spring cavity was lined with medium-sized (0.2–0.3 × 0.3 m) and large (0.30–0.50 × 0.50-0.65 m) fieldstones. The stone lining probably also served to support the ceiling of the spring cavity. A vault was built c. 3.5 m north of the beginning of the channel. The tunnel, which was hewn from this point, was lined with medium-sized and large fieldstones and was covered with a vault. The tunnel turned c. 30° to the west approximately 4.5 m north of the beginning of the vault. The floor of the tunnel was leveled nari bedrock, but the bedrock at its northern end was higher; a hewn channel conveyed the spring water to the reservoir (Fig. 5). The beginning of the channel was inside the tunnel and extended c. 2.1 m beyond the tunnel’s northern end to the point where the channel turned sharply eastward at an almost 90° angle and continued from there a further 1.3 m to the southwestern corner of the reservoir.
The northern end of the tunnel (4 m) was not stone-lined and the rock through which the tunnel was hewn was visible there. Weathering of the soft rock may have caused the tunnel to collapse, which may have led to the walls being lined with fieldstones and the construction of a vault in its ceiling. Those hewing the tunnel presumably took advantage of the soft, easily worked marl bedrock, and when the level of hard limestone was reached, at the elevation of the spring, they utilized that for the floor of the tunnel. They quarried a channel where the hard bedrock was higher than the level of the spring to maintain the gradient.
Reservoir. The reservoir was almost square (L1; 3.0 × 3.4 m, depth 1.7 m; Fig. 6); its bottom part was hewn in the bedrock of the slope, and its upper part was built except for the rock-cut southern wall. The walls of the reservoir were gray-plastered (thickness 1–5 cm) over a white foundation layer. The gray plaster was probably a modern renovation. A hewn step, used for cleaning and maintenance, was in the southwestern corner of the reservoir. The channel that conveyed water from the tunnel reached this corner. The elevation of the channel and its relationship to that of the walls shows that the walls of the reservoir had survived intact. At the bottom of the reservoir, 1.1 m from its northwestern corner, was a drainage hole (diam. 3 cm).
Burial Cave. A rock-hewn burial cave was documented below the reservoir, c. 5 m to its west. The rectangular cave (2.24 × 3.43 m, height c. 1.5 m) had a vaulted opening (width 0.64–1.01 m, height 1.2 m; Fig. 7) hewn in a rectangular frame (1.0 × 1.4 m). Two loculi (0.50–0.52 × 2.33 m, height 0.55 m) were hewn in its western wall; they were found plundered.
No datable finds were discovered in the excavation.
The ancient irrigation system was part of a large irrigated agricultural complex that extended over a long narrow strip of land (more than 3 km in length) on the northern slope of Har ʽAminadav, between ʽEn Saʽadim in the west and ʽEn Sarig in the east. The system took advantage of the aquifers that formed between two layers of rock (marl and limestone), using tunnels, reservoirs and irrigation canals. Intensive irrigated farming added much to the appearance of the landscape in the Jerusalem Hills, and on Har ʽAminadav, and was the prevailing farming method practiced in this area until the establishment of the state. This form of traditional agriculture still exists in villages such as Batir and Wadi Fukin, in the Judean Hills and around Hebron.
The burial cave was a type of loculus cave that was prevalent in the ​​Jerusalem area during the Second Temple period (see Kloner and Zissu 2003).