The current excavation area (9 × 9 m) was c. 50 m west of a burial cave, dating to the first–second centuries CE, which was excavated in 1997 (‘Atiqot 23: 49–71). Part of a building was exposed in the current excavation; it was built at the beginning of the twentieth century CE on top of the debris of a pottery workshop from the Late Roman period. Due to the limited excavation area, no remains of the workshop itself were discovered, although it obviously existed nearby.
The building, which belonged to the Arab village, was almost completely destroyed down to the level of its floors (Figs. 1, 2); a section of a room (L2), a porch (L4) and part of an inner courtyard (L3) were exposed. The wall enclosing the structure from the south (W1003/W1004; width c. 0.7 m) was built of rather large fieldstones with smaller stones and mud inserted between them. A concrete threshold was set in the doorway, which was fixed in this wall and led into the courtyard.It seems that two iron beams of railroad tracks, stamped with the year of production (1892), which were found on the floor of the courtyard, were used as a lintel. The doorjambs, like those in the doorway leading from the courtyard to the room, were built of dressed stones, some of which were in secondary use and had been taken from the ancient ruin. The walls of the room (W1001, W1002; width c. 0.4 m) were built of medium-sized fieldstones. Their outer surface was coated with mud plaster while cement-based plaster was applied to the inside and was painted white and decorated with blue geometric patterns. The floor of the room was coated with similar plaster that was made very smooth. The floors in the courtyard and on the porch were a mixture of soil and crushed stones that were placed on a foundation (thickness 0.10–0.15 m) of stones and wadi pebbles. A raised foundation of fill, retained by a wall of fieldstones coated with mud plaster, which faced the courtyard, was placed on the porch.
The northern part of the building was founded on the bedrock, which descended sharply from north to south. The rest of the building was set on fill (L6, L7, L9; thickness 0.85–1.65 m) that was intended to level the area. The fill was composed of soil mixed with small stones, wadi pebbles and debris from the pottery workshop, similar to that discovered beneath the floor of the courtyard (Fig. 4) and the porch. The foundations of Wall 1003 (depth 0.9 m) were embedded in the heap of workshop debris and were founded on the bedrock, which was overlain with the debris.
The workshop debris discovered in the fill on which the building was founded included a large quantity of slag, different kinds of fired mud bricks, ceramic floor tiles and chunks of clay. In addition, numerous pottery vessels were found, some of them flawed and distorted, as characteristic of pottery workshops. The diverse ceramic assemblage included cups (Fig. 5:1, 2), a mold-made footed bowl (Fig. 5:3), plain bowls, some of which are deep (Figs. 5:4–15; 6; 7), many red-slipped bowls with a rouletted pattern made of well-levigated clay (Figs. 8, 9), bowls with a ledge rim (Fig. 10), kraters with a ledge rim (Fig. 11), cooking kraters (Figs. 12:1, 2), cooking pots (Fig. 12:3–5), jars (Fig. 12:6–13), an amphora (Fig. 12:14), jugs (Fig. 13:1–11), juglets with a button base (Fig. 13:12–15), and lamps (Fig. 13:16–18).
Some of the bowls have a rouletted decoration; the bowls with a ledge rim and the juglets with the button base were found flawed, probably due to an error in the production process. These vessels date the activity in the pottery workshop from the late second until the fourth centuries CE. The artifacts found on the surface included a multi-nozzle lamp (Fig. 14:1) and the head of a ceramic zoomorphic vase (Fig. 14:2) that apparently date to the Late Roman or Byzantine periods, as well as a stone bowl on two feet with a pair of horizontal handles (Fig. 14:3) and tools (?) of qirton (Fig. 14:4–6).
J. Magness has suggested in the past that bowls with the rouletted decoration were produced and distributed in the region located close to Jerusalem (Magness J. 1993. Jerusalem Ceramic Chronology [C.200-800 CE]. Sheffield, pp. 186–187); however, until now no finds that could corroborate or contradict this proposal were discovered. A petrographic examination performed on ten wasters, among them bowls with a rouletted decoration, a bowl with a ledge rim and juglets with a button base indicated that they were manufactured locally, of loess soil indigenous to the site. This is sufficient to show that the production region of rouletted bowls is more extensive than was previously suggested.