The excavations at Umm al-Qanatir were divided into secondary areas. Area A, the spring complex, was excavated during the 2000, 2002 and 2003 seasons by an expedition on behalf of Bar-Ilan University, directed by H. Ben David and O. Zingboym, and are reported here. During the 2004 season, the excavation region was cleaned, and conservation measures were undertaken (Ben David, Gonen and Drey 2007). Area B, the synagogue and its environs, was excavated from 2003 until the present by an expedition on behalf of Bar-Ilan University, directed by H. Ben David, I. Gonen and Y. Drey (Ben David, Gonen and Drey 2007).
 
The spring compound (Figs. 3, 4) comprises a pool and three arch complexes (1–3), each with a trough and an overhead arch, only one of which remained intact (Arch 2). Only two of the arch complexes (1, 2) were previously known. Trough 1, situated beneath the remains of the largely destroyed Arch 1, was cleaned. The trough, which survived intact, was lined on the outside with eleven stones (width 0.25 m). Several stones which had fallen from the arch overhead were found inside the trough. Trough 2, beneath the complete arch (2) is missing several of the stones along its rim. Both troughs were of similar size and there was a connection between them for a water conduit. The remains of the third arch complex (3), which deviated from the line of the other two arches, were found southwest of Arch 1. The spring compound is described below from east to west.
 
Arch Complex 2. Several of the trough’s rim stones and the trough floor, which was built of stone slabs (c. 0.95 m below the top of the rim stones and c. 0.45 m below the water level as measured in the 2001), were exposed below Arch 2 (Fig. 5). Stone slabs (width 0.70–0.78 m) were arranged the length of the front of the trough (0.53 m below the top of the rim stones; Fig. 6). No other building remains were found west of the stone slabs, at the front of the arch and the trough.
 
Arch Complex 1. The floor of the trough, built of stone slabs (c. 0.95 m below the top of the rim stones), was revealed below the remains of Arch 1 (Fig. 7). The stone collapse of the arch and the rear wall of the arch structure were found above the stone pavement and the rim stones. As in Arch Complex 2, stone slabs (width 0.70–0.78 m) were arranged along the length of the front of the trough (0.53 m below the top of the rim stones). Stone collapse belonging to the arch was exposed west of these slabs, as far away as c. 4 m from the arch. A probe (0.8 × 1.0, depth 1 m) was excavated opposite the middle of the trough and 0.4 m west of the stone slabs; no architectural remains were found.
 
Plastered Pool. A plastered pool (3.2 × 10.2 m; Fig. 8) was located northwest of the arch. The floor of the pool was at an elevation of 323.24 m asl and in the past the maximum elevation of the water level in it was 324.54 m asl. The pool was bounded on the west by a stone wall that ran parallel to the line of the arches and continued south and north of the pool. A backhoe was used to excavate a probe west of this wall; however, no architectural remains or a habitation level were found. The pool was delimited on the north by a stone wall (width 0.25 m, height of the stone slabs below the top of the pool’s rim stones 0.4 m). The stones in the wall of the pool were dressed in a similar manner as those of the troughs. Stone slabs (width 1.3 m) were found next to the northern stone wall and also along the entire length of the pool, except close to the western wall. The pool’s pavement was 0.1 m lower than the pavement at the front of the trough and a step was situated where the two were adjoined. The pillar south of the middle arch served as the pool’s eastern wall. A niche was located inside a pillar (bottom elevation of the niche 325.48 m asl) opposite the middle of the pool and above the level of the northern rim stones. The southern wall was wider (0.6 m) and higher than the northern wall and was preserved to an elevation of 324.9 m asl. It is apparent that the wall was constructed similar to the western wall, and it might have rose to a height of more than one meter above the remains that were found and served as a retaining wall for the pavement of Arch 3 (below). Next to the inner northern wall of the pool was a step (or bench; width 0.44 m) built of stone and plastered. Two steps were located next to the pool’s inner eastern wall. The first at the level of the pavement at the front of the trough (324.29 m asl, 0.4 × 2.0 m) and the second below it (323.7 m asl, 0.4 × 1.7 m; Fig. 9). Stone collapse from the pillar was found on the plastered floor at the bottom of the pool.
Numerous pottery sherds, most of which were cooking pot fragments dating to the Byzantine period (fourth–seventh century CE), were removed from inside the pool. No ceramics that predate or postdate the Byzantine period were identified in the pool. The absence of earlier vessels is explained by the fact that the pool was cleaned while it was being used.
The pool split into two parts as a result of an earthquake or landslide which shifted the western part of the pool 0.7 m west, 0.45 m south and 0.44 m down. Two other smaller fragments that belonged to the pool were also evident. The sides of the pool that were broken were reconstructed as part of the conservation measures implemented in 2004.
 
Arch Complex 3. The remains of Arch 3 were found south of the plastered pool (Fig. 10). Only four courses of the rear wall survived. The arch was just north of the hillside and a pillar was built to its south (Fig. 11). Unlike the middle arch, no rear retaining wall was apparently built between the rear wall of the arch and the slope. In the northern part of the arch was a wall that was erected perpendicular to the arch and formed a kind of L-shaped structure (running east and north). A niche (L115) was constructed inside this wall; the elevation at niche’s floor (326.42 m asl) was similar to that of the niche in the pillar above the plastered pool. A trough, delimited by stones like the other troughs was constructed in the front of the arch. The trough was built in an L-shape, next to the arch and the northern wall, and there was probably another western trough that completed the complex rendering it a U-shape. Two rectangular stones (0.3 × 0.3 m) that had a round depression in the top of them were incorporated in the stones of the northern trough. These stones protruded c. 0.55 m above the top of the trough’s rim stones. The excavation did not reach the trough’s floor, thus the nature of its pavement could not be determined. A later wall (L114) was built inside the northern part of the trough and perpendicular to it, below the niche. In this part of the trough there was plaster covering the trough from the line of the arch’s stone collapse until the later wall, on the wall and west of it. Between the later wall and the trough’s northern wall was a built channel for conveying the water at a height such that the water constantly trickled slowly from the channel and collected in the trough (L116). West of the plastered wall was an installation with a step in it that was constructed inside the trough. A plastered channel for conveying water probably emerged from installation and paralleled the entire length of the wall of the large plastered pool. The water in this channel might have flowed via a splitter or similar installation to the large pool in its southern wall.
In front of the arch and west of the trough was an open, stone-paved area, at an elevation 0.75 m below the top of the rim stones (Figs. 12, 13). The rectangular stone slabs were arranged perpendicular to the line of the trough. Above this open area was another pavement that extended as far c. 1 m from the line of the trough stones. A rectangular stone (0.30 × 0.45 m) that was apparently placed on the pavement of the open area was found on this pavement, 0.5 m from the line of the northern trough. On the top of the stone was a depression similar to those on the top of the rectangular stones that were incorporated in the rim stones. An Egyptian scarab that B. Brandl dates to the Late Bronze Age IIC was found on the lower pavement. It is obvious the stone-paved area was damaged during an earthquake or landslide rendering it a downward and northern inclination in the direction of the plastered pool (Wechsler, Marco and Zingboym 2004). The fracture line in the paved area is unlike the fracture line in the pool. South of the stone-paved area, on a level 0.15 m lower than the bottom pavement, was a water channel constructed of terracotta sections that conveyed water from the trough to the southwest. The pipe was located from the line of the trough to the boundary of the excavated area. Next to this pipe was another similar pipe. The pipes were connected in the southwestern corner and included a breather valve that vented each of the pipes. The different elements in Arch Complex 3, namely the orientation of the arch, size of the stones, level of the pavement, the open paved area, shape of the trough, the direction of the fracture and integration of the rectangular stones in the top stones of the trough, are obviously different than the two other arches. All of these might indicate two phases of construction at the site.
 
Other finds discovered included three rectangular stones that have curved facades and a depression on top similar to the finds in the southern arch complex (Fig. 14). A fragment of a basalt sarcophagus (Fig. 15) was also found. It is obvious that the quality of the rock-cutting of the sarcophagus was very good. Evidence was found indicating the sarcophagus was sealed with lead poured in a side recess. The fragment was found ex situ; the cemetery is situated c. 70 m from the spot where the fragment was discovered. In addition to these, a stone basin was found that was plastered on the inside.
 
Buildings dating to the Roman and Byzantine periods, which constituted a rare assemblage in the rural region of Israel, were found at the spring complex. Three hypotheses were proposed for dating the construction of the complex: (1) the complex was built as a single facility in the Byzantine period; (2) the complex was constructed as a single facility in the Roman period, in the second century CE and most of it was used also in the Byzantine period; (3) the complex was constructed in two phases, first the arch and the paved open area were built in the Roman period and it was expanded afterwards in the Byzantine period. The original purpose of the arches and pools is unclear. The hypotheses proposed include a luxurious spring house from the Byzantine period or an industrial complex that required large quantities of water, perhaps as Y. Drey suggested for bleaching linseed oil (Ben David 2007:118–120) or a spring building in a non-Jewish ritual context from the Roman period. The arches and pools ceased to function following the earthquake of 749 CE. Arch 2 and the adjacent trough are still used by local residents to this day.
In recent years a similar complex was excavated at Deir ‘Aziz, located c. 3 km to the west (Zingboym 2011a; Fig. 16). An examination of springs in the region revealed that a similar installation was previously identified in the village of El Ma, c. 7 km to the southeast (Schumacher 1883:172–177). In addition, remains of ashlar stones were found at Jidyeh spring, c. 2 km southeast, on the banks of Nahal El Al. It can be suggested that these buildings indicate a regional construction style in the southern Golan Heights, in a place where the basalt bedrock interfaces with limestone. In the two sites that were excavated, Umm el-Qanatir and Deir ‘Aziz, we did not succeed in dating the foundation of the buildings; but it is clear they were used when Jewish villages existed in both of the sites during the Byzantine period (Zingboym 2011b).