The aqueduct was built into the clay alluvial soil of the valley. Two parallel walls (height c. 0.9 m) were constructed 0.4 m apart from each other on a bedding of fieldstones and mortar. They were built of two courses of dressed limestone blocks with stone wedges inserted between them. A layer of fieldstones that served to level the walls was set at the top of the two courses, and carried the covering stone slabs (Figs. 2, 3). The slabes were coarsely hewn and elongated (length c. 0.8 m, width 0.4 m, height c. 0.15 m). A similar form of covering is known from the Umayyad aqueduct leading to Ramla (HA-ESI 117).The walls were covered with light gray hydraulic plaster of low quality, which crumbles easily. A herringbone pattern was discerned running the length of the wall on the upper part of the plaster, in places where it was better preserved (Fig. 4). This is a constructive element meant to strengthen the plaster’s grip and a characteristic tradition found in the plaster craft. The aqueduct seems to have been built from the inside out. First, the foundations and the walls were set within the foundation trench. Later, the builders filled the gap that remained between the built walls and the trench walls with fieldstones of various sizes. Surprisingly, no pottery sherds or debris were found in this fill; it seems that the ancient builders’ work was better organized and clean.
The few sherds that were found in the ground next to the foundation indicate that the aqueduct dates to the Late Byzantine period or to the start of the Umayyad rule in Israel in the seventh century CE. This dating is similar to the one proposed by D. Syon, who excavated the western part of the aqueduct (ESI 14:46–49). The movement and pressure exerted by the heavy clay soil in which the aqueduct was built caused the walls to collapse inward (Figs. 2, 5), and distorted of whole sections of the aqueduct. The covering slabs were pushed southward, as seen clearly in Area B (Fig. 3), and the aqueduct was rendered a squat, triangular shape (Fig. 2: Section 2–2). The ground movement hindered any accurate calculation of the aqueduct’s gradient, but it presumably followed the general declination of the valley and of Nahal Zippori, running to its south. This ground shift and the damage it caused seems to have led to the cessation of the aqueduct’s use, whether because it was blocked or due to the lack of economic viability in maintaining it.
The excavation results could not provide an exact date for the construction of the aqueduct, and its source and destination remain unclear. But this salvage excavation does contribute to our assessment of the size and extent of the water works and to understanding the course of the forgotten aqueduct.