During July–August 2004, a salvage excavation was conducted near Giv‘ati Junction in the western Negev (Permit No. A-4212; map ref. NIG 16816/61799; OIG 11816/11799; Fig. 1), prior to the widening of Highway 35. The excavation, undertaken on behalf of the Israel Antiquities Authority and underwritten by the Department of Public Works, was directed by N.S. Paran, with the assistance of Y. Haimi and E. Aladjem (administration), V. Essman and V. Pirsky (surveying) and C. Hersch (drawing).
The excavation was located c. 20 m north of a potter’s kiln (‘Atiqot 42:43*–50*) and c. 1 km southeast of Giv‘at Arnon (Kh. ‘Ajjis er-Ras; Permit Nos. A-854, A-1251), where a Byzantine settlement was partially exposed; it included an industrial winepress that was excavated and reconstructed.
Two areas were opened; a winepress that probably dated to the Byzantine period (sixth–seventh centuries CE) was uncovered in Area A and in Area B, c. 20 m west of Area A, a number of walls that did not form a coherent plan, the base of a tabun and a refuse pit of a pottery workshop from the latter Byzantine period (sixth–seventh centuries CE), were exposed.
An industrial winepress that consisted of several components was exposed (Fig. 2).
The pavement in the two cells (L128, L142; 1.5 × 3.0 m) consisted of mortar bedding on a lime base, with stone slabs above it that were mostly robbed. The floors slanted to the east at a 3% gradient. The walls of the cells, built of fired mud bricks (5 × 25 × 25 cm), were coated with hydraulic plaster whose lower part overlapped the margin of the paving slabs.
The northern cell (L142; Fig. 3) was enclosed within Wall 15 on the north (length 3.75 m, width 0.25 m, height 0.2–0.3 m), Wall 14 on the west (length 1.7 m, width 0.25 m, height 0.20–0.25 m) and Wall 10 on the south (length 3.75 m, width 0.25 m, height 0.15–0.25 m). The walls were preserved three–six courses high.
The southern cell (L128) was similar in construction and measurements to Cell 142 and its walls (W9, W17, W18) were preserved four–nine courses high (0.20–0.45 m). The two cells were bordered by Wall 12 on the east, preserved three–five courses high (length 5.75 m, width 0.45–0.50 m, height 0.15–0.25 m), in which the remains of two channels that connected between the cells and the treading floor were discerned.
The space between W10 of Cell 142 and W9 of Cell 128 was blocked with a wide wall of medium fieldstones (W11; length 3.6 m, width 1.2–1.3 m), whose western end was incorporated within the exterior wall of the winepress, preserved three–six courses high (W4; length 8.8 m, width 0.50–0.65 m, height 0.4–0.8 m).
Soil fill (L144) to the north of Cell 142 contained numerous potsherds. A concentration of fieldstones at the western end of L144 was probably wall remains, although the exposed section was too small to determine if it extended beyond the excavation area.
The southwestern part of the treading floor (Loci 124, 133; presumed measurements 4.7 × 6.5 m, 6.5 × 6.5 m) was exposed. The northern part was not excavated and the eastern part had been destroyed by earthmoving works. The pavement of the floor was robbed, but the bedding survived with imprints of the stone slabs. The treading floor sloped southward (2.5% gradient). Wall 12 in the west and Wall 13 in the south, preserved five–eight courses high (length 4.2 m, width 0.50–0.55 m, height 0.25–0.40 m), were built of fired mud bricks and coated with hydraulic plaster whose lower part overlapped the margins of the floor slabs.
A round recess (L130; Fig. 4) in the center of the treading floor was intended for a stone screw base (not found). A plastered channel (L141; length 3.5 m, width 0.2–0.3 m, depth 0.4–0.6 m, 3% gradient) extended from the recess to the settling vat, crossing W13 where it was damaged. It seems that originally, he channel was covered with the pavement slabs of the treading floor. Another opening at the base of W13 was probably used for a pipe that also linked the treading floor to the settling pit.
The settling pit was only partially exposed in its northern part (L121; 1.4–1.6 × 1.5 m). It had a trapezoid plan and its floor consisted of stone slabs that were coated with hydraulic plaster (thickness 4 cm), probably as a repair. A circular recess in the center of the pit (diam. 0.2 m, depth 0.15 m) was the lower part of a clay jar, embedded in the floor. The settling pit was enclosed within Wall 8 on the west (length 1.75 m, width 0.70–0.75 m, height 0.5 m), Wall 7 on the east (length 1.8 m, width 0.45 m, height 0.4 m), preserved seven–nine courses and a mud-brick built step along W13 on the north. The pit’s walls were coated with hydraulic plaster.
A collecting vat on each side flanked the setting pit. The channels leading to the collecting vats were not exposed, probably because they were in the southern part of the settling pit that was not excavated. Only the northern part of the western collecting vat (L127; Fig. 5) was excavated. It had a rectangular shape (2.4 × 3.0 m, depth 1.3 m) and its floor and walls were coated with hydraulic plaster. Two concentric and plastered depressions (L136; diam. 0.3 and 1.2 m, depth 0.3 and 0.4 m respectively) were cut in the center of the floor. A step was built around the vat, continuing the top level of W8. At the western end of the vat was a working surface (1.2 × 1.4 m), whose elevation corresponded to that of the step.
Only the northwestern corner of the eastern collecting vat (L122) had survived; it was apparently similar to the Vat 127.
Between the step around Vat 127 and Cell 128 was a thick wall (W5; length 3.4 m, width 1.1–1.2 m, height 0.6 m), built of fieldstones. Two circular niches (Loci 140, 146; diam. 0.9 m, height of base 0.2 m; Fig. 6) were in the southern side of the wall; the western (L140) was plastered and the eastern (L146) was blocked with stones.
The ceramic finds from Area A, mostly dating to the end of the Umayyad and the Abbasid periods (end of the eighth–ninth centuries CE), included small bowls and bowls with incurved or upright rims (Fig. 8:1–5), bowls with an upright wall (Fig. 8:6–9), kraters with an inverted thickened rim and a wavy pattern below it (Fig. 8:10, 11),
ampulla of cooking-ware fabric (Fig. 8:12), Khirbat el-Mafjar jugs and juglets (Fig. 8:13–15), jars (Fig. 8:16–19) and a flask (Fig. 8:20). However, a small number of potsherds, mainly jars from the Byzantine period, point to a mixed fill that penetrated into the winepress from the adjacent site after the press was no longer in use.
The area (Fig. 7), located to the west of the winepress, was severely damaged by earthmoving works, hence no coherent plan of the remains could be formed. A number of walls and between them—the base of a tabun and a refuse pit, which contained numerous potsherds that apparently derived from a pottery workshop, were discovered.
A wall (W1; length 3.4 m, width 0.65–0.75 m, height 0.15–0.25 m) at the northern side of the square was oriented east–west. Another wall (W2; length 9.25 m, width 0.6–0.7 m), preserved up to two courses high, was parallel to and c. 1.75 m south of W1.
The base of an elliptical tabun (L112; 1.0–1.2 m) full of ash, whose perimeter consisted of small stones, was found between the western ends of these walls.
A pit (L110), adjacent to the center of W2 northern side, was filled with jar fragments to its top, most likely the debris of a pottery workshop.
Abutting the eastern part of W2 from the south was another wall (W16; length 1.8 m, width 0.75 m, height 0.25 m), preserved two courses high. The western side of W16 and the southern side of W2 were coated with hydraulic plaster, indicating perhaps that this was a levigating pool of a pottery workshop. The eastern end of W2 intersected with W6 (length 1.3 m, width 0.7 m, height 0.1–0.2 m) that continued north and the refuse pit (L110) seems to have reached it.
The ceramic finds from Area B included a large number of Gaza jars from Refuse Pit 110 that dated to the latter part of the Byzantine period (sixth–seventh centuries CE).
The exposed building remains in the current excavation constituted part of an industrial compound that most likely belonged to the Byzantine settlement on Giv‘at Arnon. It seems that Refuse Pit 110 in Area B was associated with the kiln that had been exposed in the past. It also seems that at least some of the installations, particularly the winepress, continued to be used in the Early Islamic period.