One square (3.3 × 7.0 m; Figs. 1, 2) was opened on a slope, just west of the Mē Golan building, and remains of a vault and a trough below it, built of basalt ashlars, were exposed. The vault was composed of a wall (width 7.5 m, max. preserved height 2 m; Figs. 3, 4) that was adjoined from both sides by piers, whose bottom parts were only preserved. A retaining wall of large fieldstones was built at the rear of the wall, to prevent its stones from rolling down the slope. Spring water accumulated in the trough built beneath the vault (2.5 × 7.5 m). The trough was enclosed and paved with basalt slabs, which were coated with hydraulic plaster (Fig. 5).
It was ascertained in the excavation that part of the trough shifted c. 8 cm to the south during an earthquake or landslide (Fig. 6). The structure probably collapsed in the earthquake that struck the region in 551 CE. It was also determined that one of the slabs that enclosed the trough was missing, may be because this spot incorporated a pipe or channel that conveyed water to cultivation plots or possibly to the nearby pottery workshop. In a later phase, probably after an earthquake, the trough was sealed and a floor of stones, which were taken from the collapsed vault, was placed above it (Fig. 7). The excavation of the trough, below the level of the later stone pavement (L103), revealed numerous fragments of pottery vessels that dated until the mid-sixth century CE, including bowls (Fig. 8:1–6), cooking vessels (Fig. 8:7, 8), jars (Fig. 8:9–11) and fragments of roof tiles (Fig. 8:12, 13). Based on the ceramic finds, the late floor is dated to the second half of the sixth century CE. It seems that the early phase of the floor should be dated to the second quarter of the sixth century CE and the late floor to the third quarter of the sixth century CE. The two building phases uncovered in the excavation may correspond to Phases 1 and 2 of the synagogue exposed at the site.
Studying the data from the surveys and from photographs taken in the 1970s show that another vault with a trough below it (Figs. 9–11) was originally built east of the vault exposed in the current excavation. The eastern vault and the trough beneath it were destroyed and a small pumping pool of the Golan Heights Water Association was built on the base of the trough; it did not deviate from the dimensions of the trough, except for its eastern side. A pier (width 1.8 m) was erected between the two vaults; part of it was exposed in the excavation and another part is visible today in the pump building. Originally, the spring structure probably included another vault situated to the west of the excavated one; however, it has not yet been exposed.
The spring structure exposed at the site resembles the one excavated at the Umm el-Qanatir site (Fig. 12), located c. 4 km to the east (Ben-David H, Gonen I. and Drey Y. 2007. Umm el-Kanatir. Qadmoniot 132:118–120 [Hebrew]). The spring structure at Umm el-Qanatir includes two vaults and a trough below each. It is possible that these two structures reflect a building tradition, practiced in the southern Golan Heights, where the basalt and limestone bedrocks meet. It is obvious that the construction of these massive buildings is expensive, takes time, and requires engineering knowledge and skill in hewing basalt. In both excavations the construction of the buildings was not dated; however, they were apparently used when the Jewish settlement at the site was present in the Byzantine period. In all likelihood, the two buildings were constructed in the first or second quarter of the sixth century CE.