Area A
An area (10×15 m) was excavated where wall segments were visible on the surface and a large, poorly preserved winepress was exposed (Figs. 1, 2). The main elements of the winepress that were preserved make it possible to reconstruct it and estimate its size (c. 14×15 m). The depth of the stratified deposit was examined throughout the excavation area and in and around the winepress; it was determined that the winepress was built in a single phase on top of sterile sand. The components of the winepress consisted of a treading floor, filtration pits and a collecting vat, plastered channels, and remains of possibly another treading floor.
Treading Floor (L107; c. 7×8 m; Fig. 3). The floor was composed of three levels: a foundation (thickness c. 0.6 m) of a mixture that contained small stones (max. length 0.1 m) and white mortar and was placed directly on the ground (Fig. 4); bedding for floor tiles, consisting of a layer of white cement (thickness 8–10 cm) and the stone floors that did not survive. Although the pavement was not preserved, rows of floor-tile negatives were left on the surface of the floor bedding (c. 0.5×0.8 m, thickness 8 cm; Fig. 5). A limestone screw base (L111; 1.2×1.2 m, height c. 0.8 m) with a square hole in its center (0.4×0.4 m) was set in the middle of the treading floor. A channel (width c. 0.15 m; Fig. 6) was visible in the bottom of the base. A robber trench (width c. 1 m) surrounded the screw, which severed the base from the treading floor (L152; see Fig. 3); this was probably a failed attempt to extract it. Farther away from the middle of the floor, toward its edges, the remains were more disturbed, particularly in the east where not even the foundation was preserved.
A rectangular complex (4.0×7.7 m) delimited by four walls (W129, W134, W140, W158), which contained two filtration pits (L135, L136), a collecting vat (L138) and a plastered cell (L139), was adjacent to the northeastern side of the treading floor.
Filtration Pit 135 (1.30×1.75 m; Fig. 7) was enclosed on the north, east and west by walls (W130–W132); the southern wall, which separated it from the treading floor, was completely destroyed as a result of removing the building stones, which left Robbery Trench 152 in its wake. The walls, built of a heap of small stones bonded with mortar, were coated with thick white plaster. Plaster remains were also found on top of W130 in the east (width 0.35 m, height 0.4 m), which seems to have been preserved to its original height and was used as a bench. The bench was next to W129 (width 0.35 m, preserved height 0.7 m), which was built of two–three courses of stone (max. length c. 0.3 m) on a foundation of small stones bonded with mortar (width 0.5 m, height 0.3 m) and was preserved to a height of 0.3 m above the top of the bench.
Two perforations were cut in the walls in the northwestern corner of the room. One hole (L142; diam. 0.1 m; Fig. 8) was cut in the west of W131, into which a terracotta pipe was inserted connecting Filtration Pit 135 with Collecting Vat 138; the second hole (L143, diam. 5 cm) was cut in Wall 132 in the north, connecting Filtration Pit 135 with Filtration Pit 136. The bottom half of a store jar (diam. c. 0.4 m, depth 0.3 m), apparently used as a settling pit, was found embedded in the floor beneath the two holes.
The floor of the pit consisted of small stones’ bedding bonded with white mortar. Remains of limestone paving tiles preserved along the edges (Figs. 9, 10) indicate that the filtration pits were paved with tiles similar to those of Treading Floor 107.
Filtration Pit 136 was square (1.75×1.90 m); its floor and walls were coated with a thick layer of white plaster (Figs. 11, 12). The pit was delimited on the north by a wall (W157; height c. 0.6 m), whose top was plastered; this too was probably a bench, preserved to its original height and adjacent to a higher wall (W148), which served as an outer enclosure. Half a store jar (L146; diam. c. 0.4 m, depth 0.3 m) was embedded in the floor of the southwestern corner, c. 0.2 m below Perforation 143.
Collecting Vat 138 was round (diam. 2.6 m, depth 1.1. m; Fig. 13). Its walls were coated with a thick layer of white plaster and a sump was cut in its floor (diam. 0.7 m, depth 0.5 m). The collecting vat was situated within a square enclosure (2.9×2.9 m) surrounded by walls (width 0.4 m). Thick white plaster was also applied to the level at the top of the collecting vat (L137) and the tops of the enclosure walls, which were 0.1 m higher than the latter. Two perforations opened into the vat: Perforation 142 from Filtration Pit 135, and Perforation 143, which was discovered blocked with white cement, from Filtration Vat 136 (Fig. 14). A section (length c. 0.8 m) in the western wall of the collecting vat (W127; width 1.1 m; Fig. 15) was missing where a perforated hole was located (L141; diam. c. 7 cm), connecting the collecting vat with the plastered cell (L139).
Plastered Cell 139. The poorly preserved remains of a plastered cell were located to the west of Collecting Vat 138; a section of the cell was excavated (0.9×1.6 m) and it was not possible to estimate the depth of the cell. Wall 140, whose bottom part was only of preserved (max. height 5 cm), delimited the western side of the cell.
The poor preservation did not facilitate a satisfactory reconstruction of the architectural remains; however, in light of Cell 139’s location, filtration pits, which received liquid from the treading floor on its way to a collecting vat, were probably located west of the collecting vat, similar to its eastern side.
Plastered Channels. Wall 134 (width 0.8 m, height 0.4 m), which delimited the winepress on the north, was built of small stones bonded with mortar. At the top of the wall, next to the outside of Wall 148 that delimited Collecting Vat 138 from the north, was a channel (exposed length 3 m, width 0.4 m, depth 0.1 m; Fig. 16), coated with a thick layer of white plaster. Wall 140 enclosed Plastered Cell 139 from the west and was poorly preserved. However, its remains included small bonded stones, at whose top was a straight strip of plaster, which raise the possibility that this too was a plastered channel that ran perpendicular to the channel on top of W134.
Another Treading Floor (?). A square was excavated northwest of the winepress. The remains of walls and an eroded foundation of small stones, set in white mortar (L123; thickness c. 0.15 m; Fig. 17), which was similar to that found below Treading Floor 107, were discovered. Foundation 123 was surrounded by walls (W125, W149, W150), built of dressed kurkar (max.  length 0.7 m), which ran parallel to the directions of the winepress’ other walls. Only a few stones of W149 (length c 3m) were preserved on its outside and the others were apparently robbed; however the straight line ending of Foundation 123, parallel to the outside of Wall 149, makes it possible to reconstruct the width of the wall (c. 0.7 m). Wall 125 (preserved length c. 1.5 m, width 0.7 m) was perpendicular to W149 and enclosed Foundation 123 from the north. A channel (W124; width 0.2 m, depth 5 cm) whose direction corresponded to the continuation of Channel 134 was built on the top of the wall.
The elevation of Foundation 123 was 0.2 m lower than that of Treading Floor 107, but it can reasonably be assumed that the original elevation of the two floors was identical and that Foundation 123 is the remains of another treading floor. This is corroborated by the alignment of the walls around Foundation 123 and the course of Channel 124.
A single course of stones was preserved from W150 (length c. 2 m, width 0.6 m), overlaying a foundation course built of different size stones. A little gray floor material, without finds was discerned on both sides of the wall (L154, L155), which was probably not part of the winepress, but rather that of an adjacent structure.
Evidence of Later Use in the Winepress. An earthen floor mixed with ash (L109; min. dimensions 3×3 m; thickness c. 0.1 m; see Fig. 15) was founded on the collapse that filled the winepress and above it was a tabun (L147; diam. 0.4 m, preserved height 0.1 m; see Fig. 11) built of reddish mud-brick material. Next to the tabun, Floor 109 abutted a round installation (L156; max. diam. 0.7 m) built of fieldstones (max. length 0.3 m) whose inside was coated with light pink plaster.
These finds suggest that after the winepress was abandoned, the place was used again, possibly as a temporary dwelling; the poor preservation and meager finds preclude the understanding of its essence or period.
Area B
Four squares were opened c. 0.3 km northeast of the winepress and four cist graves were found (not excavated; Fig. 18). One grave, exposed in its entirety (T205; c. 0.8×2.0 m; Fig. 19), was covered with stone slabs (length c. 1 m) placed widthwise; a shattered jar was found at its western end (Fig. 20). A second grave (T207) was located west of T205 and a third one (T208) was east of it; the ends of both of these graves protruded from the sides of the square (see Fig. 19). The fourth tomb (T206) is located c. 7 m to the east and its end also protrudes from the balk of the square (Fig. 21). 
Numerous fragments of pottery vessels were recovered from the excavation, all dating from the Byzantine period and including Late Roman C bowls (Fig. 22:1–3) dating to the fifth–eighth centuries CE; a Fine Byzantine Ware bowl (Fig. 22:4) dating to the sixth–seventh centuries CE, but no later than the Early Islamic period; kraters decorated with wavy combing (Fig. 22:5) that are characteristic of the Late Byzantine period and frequently found in industrial winepresses of the kind excavated here; a cooking pot (Fig. 22:6) and a cooking-pot lid (Fig. 22:7). Most of the finds are Gaza type wine jars with a low rim, which date to the fifth–sixth centuries CE (Fig. 22: 8). A few jars of other types included an early Ashqelon jar (Fig. 22:9) that dates to the fourth–fifth centuries CE, a baggy-shaped jar (Fig. 22:10) from the sixth century CE and an imported jar (Fig. 22:11) known to occur in the pottery repertoire of the Byzantine period. Other pottery types included amphorae (Fig. 22:12) and juglets (Fig. 22:13, 14) that are ascribed to the fifth–seventh centuries CE.   
The ceramic artifacts date the winepress to the late phase of the Byzantine period, the sixth–seventh centuries CE, and do not include any diagnostic finds from the beginning of the Early Islamic period. The finds and their dating are consistent with finds from similar sites in the northern Negev, many of which have been excavated, e.g., Horbat Shelah (HA-ESI 121) and Horbat Qarqar (Permit No. A-4398). Earlier potsherds were not found in any of the sites and it seems that these winepresses should be dated only to the Byzantine period.
Industrial winepresses of the kind excavated here are a common phenomenon throughout Israel, particularly along the coastal plain, in the Northern Negev and the Byzantine cities of the Negev, in different size cities, in churches, monasteries, and sometimes even in sites that have no connection to an adjacent settlement. The winepress described here is close to Khirbat Namus, where potsherds from the Byzantine period were found, and it can reasonably be assumed that there is a connection between them.
Mostly fragments of Gaza type wine jars that have a low rim (see Fig. 22:15) and date to the fifth–sixth centuries CE were discovered in the excavation between the cist graves in Area B; it is very likely that most of the graves were used by the inhabitants of Khirbat Namus in the Byzantine period.