The excavation took place immediately to the west of the confluence of the tributary and Nahal Ha-Shiv‘a, 900 m southwest of Bet Qeshet and c. 2 km north of the foot of Mount Tabor, on the southeastern foot of a hill locally known as Giv‘Ha-‘Shiv‘a (Fig. 1). The water-pipe trench, running in a general north–south direction, revealed large building stones and numerous pottery sherds. The excavation yielded building remains from the Late Roman period, as well as pottery spanning the Iron Age II through the Mamluk period.
 
A stretch of four squares (A1–A4; Fig. 2) was opened along the western side of the infrastructure trench, near where the large building stones were exposed; Sq 4 was only partially excavated (1.5 × 3.5 m). The sediment overburden in these squares was removed mechanically, reaching the level of the large stones, and exposing the tops of building remains, which were than excavated manually. As the mechanically removed soil contained only scant pottery, it was presumed that most of the sherds uncovered in the infrastructure trench originated in a deeper level. In addition to these four squares, a fifth square (B1; max. depth 1 m), roughly trapezoid in shape, was dug mechanically on the eastern side of the trench to determine if additional building remains could be located; although it did yield some remains, the square was not enlarged due to time constraints.
The excavation uncovered the eastern edge of a building (Sqs A1–A2); the remains of a floor, possibly a courtyard, to the south of the building (Sqs A2–A3); a segment of wall (Sq A4); and possibly an additional wall (Sq B1). The pottery and other finds point to activity at the site during the Iron II, Persian, Hellenistic, Early and Late Roman, Byzantine, Early Islamic and medieval periods, with the most significant occupation apparently during the Late Roman period. Also found were Middle Paleolithic flint artifacts in secondary deposition.
 
Architectural Remains
Squares A1–A3. Sqs A1–A2 revealed the eastern part of a building (Figs. 3, 4): two parallel north–south walls (W1001—preserved length 7 m; W1002—preserved length 6 m); two adjoining east–west walls in the north (W1000—exposed length 2.5 m; W1003—exposed length 1.2 m; Figs. 5, 6); and one adjoining east–west wall in the south (W1004; Fig. 7). The east–west walls continued westward into unexcavated sediments. Both the exterior and interior walls were founded on bedrock and were built of boulders and hewn stone blocks—W1004 has one surviving ashlar—with sediment and medium-sized fieldstones between them. The exterior walls were built in a sturdier manner than the interior walls, for which smaller fieldstones were used. All the walls, possibly except for W1004, include several courses of smaller stones that served as foundations. The building stones of these walls were possibly robbed from earlier structures.
The interior walls (W1002–W1004) enclosed a room (width 4 m; Figs. 4, 8). Between Walls 1001 and 1002 was a rather narrow corridor-like space (width c. 0.75 m), in which two patches of floor beddings of compacted small stones (L109, L208) were found. Floor 109 (Fig. 8), in the northern part of this space, abutted the corner of Walls 1000 and 1001, reaching their upper foundation course. Floor 208 (Fig. 9) abutted Walls 1001 and 1002 at the southern end of the space, reaching the top level of the lower course of W1002 and extending above the row of stones comprising the eastern extension of W1004. The floor may have thus been laid down after the walls of the building were dismantled almost down to their foundations. While the building’s southeastern corner apparently did not survive, a reconstruction suggests that its exposed part had a symmetrical layout (see Fig. 2).
None of the building loci yielded a chronologically homogenous assemblage. Nevertheless, as most of pottery sherds from the deposits associated with the building date from the Late Roman period, it may be suggested that the building was erected during this period. Metal items were retrieved from soil accumulations covering the architectural remains in Sq A2: an unidentified artifact, perhaps a chisel or an arrowhead (Fig. 10:1) and two complete nails with a short body and a semi-elliptical head (Fig. 10:2). These appear to have been associated with late agricultural activities at the site, although the possibility that they belonged to an earlier phase of occupation and were moved vertically closer to the surface due to disturbances cannot be discounted.
Several large stones and stone clusters were found to the east of the building, some at the same level as that of the preserved wall courses and others below their level. These seem to represent robbed-out rather than collapsed walls.
 
Squares A2–A3 (Figs. 3, 11, 12). Part of a floor comprising stones of various sizes was uncovered adjacent to the southern end of W1001 (L207; Fig. 9), just north of the balk separating Sqs A2 and A3; it was disturbed by the infrastructure work. This floor was possibly part of a courtyard that extended south of the building. The floor was covered with stones, some of which may have belonged to the lower course of W1001, while others may have been part of the eastern extension of W1004. An Umayyad fils minted in Egypt (697–750 CE; IAA 145644; Fig. 13) was found on the floor. The accumulations in the area of Floor 207 contained a handful of poorly preserved, and hence unidentified, bone fragments.
Further south, in Sq 3, was an accumulation of collapsed stones of unclear origin, which yielded Umayyad pottery. Two additional Umayyad fulus were found nearby: one minted at Ramla (L303; IAA 145645), the other in Damascus (L301; IAA 145646). Like the fils found on Floor 207, they date roughly from the post-reform period (697–750 CE) and likely date the dismantling of the building.
The uppermost soil accumulation in Sq A3 (L300) yielded four fragments of iron nails with a dome-like head (Fig. 10:3), two blade tips and one tip that resembles the edge of a horseshoe but is more likely part of a pruning hook (Fig. 10:4) and one complete horseshoe with a nail still attached (Fig. 10:5). Also found in this accumulation were a basalt basin and three upper grinding stones, two of the ‘bread-loaf’ type and one cylindrical (not illustrated). As these artifacts were found in soil accumulations closer to the surface than the architectural remains described above, they may attest to agricultural and domestic activities of a later date.
 
Square A4 (Figs. 11, 14) yielded only an east–west row of flat, medium-sized stones laid on a foundation of small stones (W3000; length 1.5 m), possibly a terrace wall. The relationship between this wall and Floor 207, five meters to its north, is unclear, although it is noteworthy that the wall level is below that of the floor.
 
Square B1 (Figs. 15, 16). Two large, elongated stones uncovered at the bottom of the mechanically dug probe seem to have been part of a very poorly preserved east–west wall, overlying a foundation of smaller stones (W4000). It may represent an earlier phase of construction at the site. The deeper part of the probe (L105) yielded abundant Iron Age II (see Fig. 17:1–5) pottery and a handful of weathered Persian-period sherds; it is unclear if this assemblage dates the wall. While similar Persian-period pottery was found in other parts of the excavation, this is the only place where Iron Age pottery was uncovered.
 
Finds
None of the excavated loci yielded chronologically homogeneous finds. These comprised mostly pottery and a small assemblage of glass fragments. Most of the potsherds were body sherds; weathered sherds were quite common; and some of the glass fragments were very small—all indicating that building and agricultural activities continually disturbed the underlying deposits at the site. The three Umayyad fulus found on Floor 207 and in Accumulations 301 and 303, as well as the abovementioned metal artifacts, which were found in the uppermost soil accumulations in Sqs A2–3, attest to these activities.
 
Pottery. The pottery dates from the Iron II (Fig. 17:1–5), Persian (Fig. 17:6–8), early Hellenistic (Fig. 17:9–16), Early Roman (Fig. 17:17–20), Late Roman (Fig. 18:1–5), Byzantine (Fig. 18:6–8), Umayyad (Fig. 18:9, 10; mostly from Sq A3) and Mamluk (Fig. 18:11, 12) periods.
Pottery of the Late Roman period comprised by far the majority of the material in each locus (>90%). The deposits richest in pottery were those associated with the building, up to the level of the topmost course of its walls, suggesting that it was erected in the Late Roman period. The remaining c. 10% of the pottery is mostly from the Hellenistic and Byzantine periods.
Typologically, the vast majority of the Hellenistic–Byzantine pottery consists of jar fragments; cooking pots and especially bowls and jugs/juglets were found in very small numbers. This suggests that the site was a farm or possibly a storage and redistribution facility for agricultural goods. This possibility may fit with the geographic position of the site, near the streambed; the wadi may have been an active stream or perhaps a road during the site’s main period of occupation.
The pottery comprises types that were locally made or traded over a short distance at most: Persian-period Phoenician ‘torpedo’ jars (Fig. 17:8); Hellenistic-period white and pinkish jugs (Fig. 17:9, 10) and jars (Fig. 17:11–15); Roman-period Galilean jars (Fig. 17:19, 20); and Byzantine-period jars of a ‘metallic’ black fabric with white-painted motifs (Fig. 18:7). No pottery attesting to long distance trade was found.
 
Glass. The profiled fragments of glass date almost exclusively from the Late Roman–Byzantine periods, including scant remains of glass-production waste. Only one, colorless glass fragment is dated to the Abbasid–Fatimid periods.
 
Flints. The flint assemblage totals 37 artifacts uncovered in different loci. All the specimens were found in secondary deposition, as evident by the thick patina on many of the items and the heavily abraded state of some of them. The technological and typological traits were mostly nondiagnostic, and the assemblage is clearly of a mixed nature. The only items that could be attributed to a specific industry are two Levallois flakes, one of which is retouched (Fig. 19:1), and one Levallois core (Fig. 19:2). The location of the Middle Paleolithic deposit from which they originated is presently unknown.
 
Although the pottery uncovered in the excavation points to activity at the site during the Iron II, Persian, Hellenistic, Early and Late Roman, Byzantine, Early Islamic and medieval periods, the most substantial remains at the site seem to belong to the Late Roman period. These comprise the vast majority of the pottery and the remains of a building. The activities at this site seem to have been agricultural in nature, accompanied by local trade in pottery vessels. Further conclusions could not be drawn due to the limited extent of the excavation and the mixed nature of most of the exposed contexts.