Mechanical equipment was used at first to remove a layer of topsoil (thickness c. 0.2 m) in the southwestern part of the excavation area. A poorly preserved rectangular installation consisting of two construction phases was exposed. In the early phase, the installation included a plaster floor (L106) that was delineated on the north, south and west by walls (W109, W115, W122). Foundations built of roughly hewn limestone and kurkar blocks were all that had survived of the walls. The western wall (W109; length c. 1 m, preserved height 0.25 m) formed a corner with the northern wall (W122; length c. 1.5 m, preserved height c. 0.2 m); the eastern end of W122 did not survive. Wall 115 (length c. 1.5 m, preserved height 0.65 m) may have enclosed the installation on the south, although it could have been an internal partition within the installation. Floor 106 was made of whitish pale pink plaster, and it was mainly preserved in the northern part of the installation
The floor was laid on a bedding of small fieldstones and plaster (L107) that was set on soil. Several rectangular ceramic tiles (0.20 × 0.27 m; Fig. 4) were incorporated in the center of the floor. These tiles were evidently in secondary use and were taken from an earlier installation in the area. A channel aligned in a north–south direction (L118; Fig. 5) ran through the middle of the installation’s floor. A terra-cotta pipe extending in a north–south direction ran through the bottom part of W122; it served to convey liquid northward, away from the installation (Fig. 6).
The floor and two of the walls were rebuilt in the later phase of the installation. A new plaster floor (L100) was built over the floor of the previous phase and covered Walls 109 and 122 of that phase. Because of the proximity of the floor to today’s surface, only sections of it survived. Wall 122 was replaced by a new wall (W108; length 4 m) constructed along its northern face. The terra-cotta pipe that ran through W122 was canceled as well. Only the foundation of W108, built of fieldstones (height 0.5 m), was preserved; its eastern part was destroyed. A new wall (W105; preserved height 0.25 m) was erected in the second phase above W109, but only its foundation, built of coarsely dressed stones, was preserved. Walls 105 and 108 formed a corner. Three coins were discovered in the soil that accumulated above Floor 100 of the late phase: a Late Roman coin (351–361 CE; IAA 154903); a Byzanto-Arab C coin (691–697 CE; Yavne mint; IAA 154901); and an Umayyad coin (c. 737 CE; Ramla mint; IAA 154904). 
 
A surface constructed of fieldstones bonded with white plaster (L112, L119) was discovered northeast of the excavation area (Fig. 7); it was probably the continuation of Foundation 107, which was discovered in the southwest of the area. A square, stone-lined pit (L120; 0.8 × 0.8 m, depth 0.4 m) treated with white plaster was constructed in the center of the surface. In the western part of the surface was another stone-lined pit (L121). Only the southern and eastern walls were exposed and only partially (to a depth of 0.5 m); they too were treated with white plaster. A coin dating to the fourth century CE (IAA 154902) was found deep in the pit. No walls were discovered that delimited the surface in this part of the excavation area. A refuse pit containing a concentration of Byzantine-period Gaza jar fragments was discovered east of the surface.
 
Only a few artifacts were discovered in the excavation. Four coins were found: two date to the fourth century CE, and two that are from the seventh–eighth centuries CE. Fragments of pottery vessels dating from the Late Byzantine period to the beginning of the Early Islamic period were also recovered. Since the area was disturbed as a result of cultivation, it was impossible to separate between the ceramic assemblages from the two construction phases of the installation.
 
The ceramic finds in the excavation included a FBW bowl (Fig. 8:1) that was common in the seventh–eighth centuries CE (Magness 1993:197–198, Form 1F); two kraters with a thickened rim that are decorated with a combed design (Fig 8:2, 3), characteristic of the sixth–eighth centuries CE (Magness 1993:209–210, Form 3); Gaza jars (the bulk of the finds: Fig. 8:4–6), dating to the sixth–seventh centuries CE (Majcherek 1995:169, Pls. 3:4, 7:2, Form 4); a bag-shaped jar with a thickened rim and a neck of medium height (Fig. 8:7; Magness 1993:225, Form 4B); a saqiye jar (Fig. 8:8), which dates from the fifth–seventh centuries CE (Ayalon 2000:223, Fig. 3:4); a funnel rim, possibly of a jar, for which no parallel was found (Fig. 8:9); and a wick-hole of wheel-made sandal lamp (Fig. 8:10), which was common in the fifth–seventh centuries CE (Rosenthal and Sivan 1978:122, No. 506–509).
 
The excavation yielded remains of an installation that was probably built in the Late Byzantine period and reused at the beginning of the Early Islamic period. The terra-cotta pipe discovered in the early phase of the installation and the numerous fragments of Gaza jars that were found in the excavation suggest that the installation was built as part of a large industrial winepress, which was subsequently canceled at the beginning of the Early Islamic period.