Phase II
A large, 'complex type' industrial winepress was exposed in Square A1 (Fig. 2). It consisted of an upper treading floor/work surface paved with mosaic, storage cells; a large main treading floor with a socket for anchoring a screw; a settling pit and a collecting vat.
The upper treading floor/work surface (L232; min. dimensions 3.0 × 4.4 m) was paved with white tesserae (50 tesserae per sq dm) that had been repaired several times. According to D. Amit, additional storage cells were probably constructed on this surface. Storage cells were located in the western and northwestern part of the surface, five of which were preserved (L233, L239, L241, L243, L248; diam. 0.8 m, depth 0.5 m). The sixth storage cell, which did not survive, was located in the northwestern side, based on the remains of its foundation. Cell 241 was intentionally blocked at some point. The cells, which were coated with a thin application of plaster (thickness c. 0.5 cm) and built on a bedding of small stones and gray mortar, were covered with a plastered, barrel-shaped vault. The floors of the cells were paved with mosaic and a cavity was cut in the center of each. No gutters that used to transfer the liquid from the storage cells to the main treading floor were visible in the storage cells, as was the case in other complex winepresses, i.e., at Mazor (El‘ad), where the gutters were made of lead (HA-ESI 18:60). The cavities in the floors of the cells may point to ladling vessels that aided in removing the liquid from them.
Below and to the west of the upper surface, in the middle of the winepress, was a large mosaic surface (L246; 4.4 × 4.4 m), with a perforated cavity for anchoring a screw (L240). The sides of the surface (height 0.4–0.8 m) were coated with two layers of plaster; the bottom layer was a pink shade of hydraulic plaster. For some reason, the stone for securing the screw was not positioned exactly in the center of the surface; an effort to remove it seems to have been made, perhaps in an attempt to rob it. A terracotta pipe (L245) that served as a gutter and led to the settling pit (L250) was discovered in the northwestern corner of the surface.
Pit 250 (0.9 × 0.9 m, depth 0.5 m) was also damaged by the development work, but could be reconstructed. The pit was paved with a white mosaic, most of which did not survive. A gutter that conveyed the liquid to the collecting vat was discovered in the western side of the pit. The collecting vat (L230; 1.8 × 1.8 m, depth 1.6 m) was coated with hydraulic plaster and paved with a mosaic. Its southeastern side was a plastered vault (Fig. 3), which was built to increase the capacity of the vat and perhaps resulted from engineering considerations. A sump (L252), also paved with tesserae, was cut in the floor of the vat, east of the southern corner. A small basin where a jar could be placed was revealed at the highest point, on the northwestern side of the collecting vat.
The winepress was delimited on the southwest by a massive wall (W235), abutted by a pavement (L251), and by another wall (W256) on the northwest.
The potsherds discovered in the foundation of the winepress and in the accumulations overlying the installation included mostly jar fragments (Fig. 4) that indicate it was used in the Roman period (first–second centuries CE).
 
Phase I
An earlier wall (W253) was exposed beneath W256 in Square A1. The ascription of the remains in the rest of the squares relied on differences in elevation between them and the winepress in Square A1.
A wall built of dressed fieldstones (W212) was exposed in Square A3; it was abutted from the west by a floor of small stones and crushed chalk (L231).
Two walls (W217, W249), aligned northwest-southeast, were exposed in Square A5; they formed a kind of alley (L216; Fig. 5). The walls were built of roughly worked fieldstones and parts of them were founded on bedrock. Wall 217 survived three–four courses high (max. width 0.8 m, height 1 m).
Two walls (width 0.5–0.7 m) built of roughly hewn stones were exposed in Square A7; they formed a corner of a massive building founded on bedrock. The inside of the building was paved with small fieldstones (L221; Fig. 6).
Two rock-hewn shafts, probably shaft tombs originally, were exposed in Square A11, at the western end of the excavation. The western shaft (L228; depth 1.5 m; Fig. 7) was circular and opened into a bell-shaped cavern, probably a cistern; a kind of funnel for conveying rainwater was hewn east of the cistern. The eastern shaft (L227; 0.6 × 0.7 m, depth 1.3 m; Fig. 8) was rectangular and opened into a cavern, probably into the same cistern. The inner spaces of the cisterns were not excavated. The shafts could not be dated with accuracy; however, the potsherds discovered inside them suggest they were used in the Roman period or earlier.
 
A built and plastered vault was exposed in the collecting vat, which is a construction method that is uncommon in winepresses known to us. Another complex winepress was recently exposed c. 20 m west of the current excavation. This installation has a collecting vat with an almost identical vault (Permit No. A-6061; Y. Elisha, per. comm.). This is apparently a construction method characteristic of the region, and possibly even by the same builders.