Area 2000 (Sqs SE 6/6, 6/8, 7/8; Figs. 1, 2)
Two new squares were opened in Area 2000 to clarify the layout of the structure in the ancient village: one in the northwest corner and one in the northeast corner. Excavations also resumed in Sq SE 7/8, which had been partially excavated in 2011 (see Magness 2012; Magness et al. 2013). By the end of the 2013 season, six squares had been excavated in Area 2000, with the balks removed during the course of the season to expose the connections between the walls and floors of the rooms. This process revealed that the rooms might not belong to a single structure but perhaps to at least two side-by-side units with a common wall (W212). Both units have at their core two rooms, one behind the other: Unit 1 (the western unit) consists of Rooms 1 and 2 (separated by W217), and Unit 2 (the eastern unit) consists of Rooms 3 and 4 (separated by W226). There are additional rooms to the west (Room 5), north (Rooms 6, 7, 8), and east (Room 9), whose relationship to Units 1 and 2 is unclear (see below).
The main entrance to Unit 1 was through a doorway in the southern wall of Room 1 (W208), which had monolithic doorposts on either side of a threshold block. The original passage from Room 1 to Room 2 was through three side-by-side openings in W217, separated by monolithic pillars (which perhaps supported arches). Later, the two side openings were blocked, leaving only the central passage; a north–south row of standing stone slabs was installed along the eastern side of Room 1, perhaps creating a trough for a stable (see Magness et al. 2013). However, the original floor of Room 1 (L2146) seems to have continued in use. Excavations in Room 2 ended at a floor (L2152) that was preserved only in the southeast corner of the room, in which were embedded large pieces of pottery and a three-legged grinding stone lying at an angle.
No entrance into Room 3 of Unit 2 was visible in W209. Perhaps the unit was originally entered through W222 on the north (Room 4; see below). Finely hewn threshold blocks in W226 were part of a doorway that provided access from Room 3 to Room 4. As in Unit 1, the floors of Unit 2 were made of compacted dirt with plaster flecks (floor of Room 3: L2157; floor of Room 4: L2156).
Stones found in fills above the floors attest to the gradual collapse of the walls after the units were abandoned. There is no evidence of burning or violent destruction. The collapse layers (down to the floors) contained large quantities of pottery, including many large fragments but very few whole or partially restorable vessels; pieces of a ceramic vessel with copper wires from an ancient repair; grinding stones; and shards of glass vessels.
A threshold for a doorway in W222 could have provided access from Room 4 (in Unit 2) to Room 8 (floor: L2141) to the north. However, the threshold is approximately 0.6 m above the floor level in Room 2, and 0.35 m above the floor level in Room 8, without any sign of steps. It is possible that the threshold was associated with later floors at higher elevations in Rooms 2 and 8, but no evidence of such floors was discerned in the excavations. Indeed, only one floor was found in each room in Area 2000, but the floor levels in the surrounding rooms (floor in Room 5: L2134 and L2144; floor in Room 6: L2147; floor in Room 7: L2139) were at least 0.3–0.5 m higher than in Rooms 1–4. For this reason, the relationship between Units 1 and 2 and the surrounding rooms remains unclear. Furthermore, the foundations of W223 and W211 are nearly two meters below the floor in Room 5 (L2134 and L2144).
Room 9, which may have been an open courtyard, had a large stone threshold block (W220), perhaps in secondary use, partially embedded in the east balk. The top of the threshold block is at a much higher level than the plastered floor in this room (L2124), suggesting that it is part of an installation associated with the floor, or belongs to a later occupation phase for which no floor or surface was found. This later phase within the Late Roman–Byzantine period is well-attested architecturally throughout the building, as for example the threshold cut into W222, the construction of W219 on top of the floor in Room 5, and the blocking of two of the openings in W217. No floors associated with this later phase were found.
Despite the disparities in floor levels and evidence of modifications to the units, the pottery and coins found in the fills below the floors and in their make-up, as well as on top of the floors and in the collapse, indicate that occupation lasted for only about one century, with construction dating no earlier than the early fifth century CE, and the final abandonment before the mid-sixth century CE.
In addition to Late Roman–Byzantine pottery, the fills below the floors in all of the rooms yielded earlier pottery, especially from the Late Iron Age, Persian period and Early Hellenistic period, as well as other finds such as fragments of stone (chalk) vessels (which are typical of the late Second Temple period but continued until the third–fourth centuries CE). Earlier walls (W216, W221, W227) with a northeast–southwest orientation were discovered beneath the Late Roman–Byzantine walls and floors, but they could not be dated as there were no associated surfaces.
 
Area 3000 (Figs. 3–5)
Sq SW 2/7 (Figs. 3, 4). Excavations in the modern village of Yakuk continued in Sq SW 2/7, in the northeastern part of a room bounded on the south by W301, on the east by W308, and by W317 in the north balk (Magness 2012; Magness et al. 2013). W317 is visible on the surface of the ground continuing west beyond the boundaries of the square.
The foundation trench of the pilaster (L3105) abutting W317 yielded pottery dating to the 19th–20th centuries CE. A single course of large field stones (W327) was found at the base of W308, from which it deviates slightly to the south.
Two successive surfaces were associated with W327 in the southern third of the square: a plaster floor (L3902) and a dark, compact floor (L3905). The unevenness and slope of both surfaces suggest that they may have been in an open-air area. W327 and the associated surfaces belong to a phase in the nineteenth–twentieth centuries CE village that predates the latest period of occupation, which ended in 1948.
 
Square SW 4/5 (Fig. 6). Excavations began with a balk trim that extended the square by 0.5 m to the west, followed by the removal of the medieval cobble and plaster floor (L3106; same as L3044 in Sq SW 3/5) and W311, which Floor 3106 abutted. A grinding stone was found lying upside down on top of the floor. The make-up of the floor—which was made of triangular basalt cobbles set in a 10–12 cm thick bedding of plaster, pebbles, and crushed lime—contained pottery dating to the thirteenth century CE. As in Sq SW 3/5, the fill under the cobble and plaster floor (L3122, L3129, L3135) consisted of alternating layers of soft yellow and brown soil mixed with fine yellow building chips, which appear to have been deposited deliberately on top of the synagogue’s floor. All of these layers of fill contained Late Roman–Byzantine pottery, and the uppermost layer (L3122) yielded Crusader–Mamluk (twelfth–thirteenth centuries CE) pottery.
The excavation of L3129 and L3135 brought to light two rows of large ashlar blocks (L3138) at the south side of the square, which apparently were part of a paved courtyard outside the synagogue. Along the north side of Pavement 3138 are the foundations of the synagogue’s southern wall (W340), which was robbed out. In the medieval period, a new southern wall (W315) was built further to the south, enclosing Pavement 3138. To the north of W340 was a rubble collapse (L3137) on top of the mosaic floor (L3139; see below). The collapse contained a large number of cut building stones and architectural fragments, an iron nail, chunks of burnt plaster and ash which had charred the mosaic floor underneath. As the rubble and fire remains are limited to this area, they presumably are associated with the destruction and robbing out of the bema rather than the collapse of the synagogue’s roof.
Portions of a mosaic floor (L3139) were discovered under the layers of fill and collapse, laid in a hard plaster bedding (L3155) that was intact and ran up to W310. Three main sections of the mosaic survive (Fig. 7). To the east and running along W310 is a white border with a guilloche pattern like that uncovered in Sq SW 3/5. The border turns a corner at the southern end of the square, just in front of the bema, where the mosaic ends, indicating that the mosaic and bema are contemporary. The border frames a large scene in the middle of the floor showing Samson carrying the gate of Gaza on his shoulders (see Judges 16:1–3; Fig. 8). Samson is depicted as a giant with his hands wrapped around the sides of the gate. Another patch of mosaic below (to the west) preserves Samson’s red and blue belt, white tunic, and red cloak. In the northwest corner of the square (continuing into the balks) is a third patch of mosaic showing a man with a horse, presumably a Philistine (Fig. 9). Above them, the boot of another male figure and the hoofs of a horse are preserved. Like Samson with the foxes in Sq SW 3/5, here Samson is oriented with his head to the east. Because the mosaic showing Samson carrying the gate of Gaza did not extend into the balks, and because the bedding was unstable, this piece was removed for conservation and temporary storage at the IAA.
Excavations were also conducted outside (to the east) of W310, on either side of a narrow step made of stones laid upright (W312): L3169 to the north and L3168 to the south. These loci consisted of unsealed fills (which contained pottery up to the modern period) above a hard bedding, apparently belonging to the robbed-out pavement of the medieval building’s courtyard. The fills to the north of the partition yielded almost fifty coins, clustered together above the stony bedding. The loci to the south of the partition yielded only two coins.
 
Square SW 2/5. This is a new square opened in 2013, to the north of Sq SW 3/5 (with the mosaic of Samson and the foxes and the inscription flanked by female faces). Just below ground level, which sloped steeply from west to east, were two occupation levels of the nineteenth–twentieth centuries, the upper with remains of walls and floors, and the lower ending with an ashy surface (L3126) associated with tabuns.
Below L3126 was a thick layer of loose fill (L3134). The tops of W310 and W311 appeared, both continuing into the north balk. One of the stones in the uppermost (preserved) course of W310 was recut as a 2 m wide threshold block in the medieval period, apparently providing the main entry into the building at that time. This is also indicated by the fact that W311 does not continue along the inside of the threshold. The same cobble and plaster floor found in Sqs SW 3/5 (L3044) and SW 4/5 (L3106) was also found in Sq SW 2/5 (L3147), abutting W311 and running up to the threshold block in W310 (Fig. 10). At the northern end of the square, immediately adjacent to W311, a small patch of white mosaic decorated with a colored triangle or diamond pattern was preserved on the cobblestone floor (Fig. 11). This piece of floor was left in situ because of the importance of probable medieval mosaic, while excavations in the rest of the square continued below (Fig. 12). Apparently W311 is not a wall but a bench made of reused ashlars from W310, as it sits on earth fill without foundations. If this is the case, the medieval building reused W310 and added a floor with a bench approximately 0.70 m above the level of the original (Late Roman) synagogue floor. Perhaps this is the synagogue with a mosaic floor that Ishtori Haparchi reported seeing at Yakuk in the early fourteenth century CE, although no archaeological evidence has been discovered yet indicating the building’s function.
As in Sqs SW 3/5 and 4/5, layers of fill consisting of brown and yellow soil and building chips had been deposited under the medieval floor down to the level of the synagogue floor. The removal of these layers of fill exposed the original white plaster still adhering to the inner face of W310. Large portions of mosaics were still preserved in the synagogue floor (L3164; plaster bedding = L3178), cut by a later pit (L3171) in the southeast part of the square, just inside the medieval threshold. A small piece of mosaic in the northwest corner of the square contains unidentified linear motifs that might represent parts of building (Fig. 13). A large portion of mosaic is preserved in the southwest and central parts of the square. It is divided into three horizontal (north–south) registers, one above the other, each containing figured scenes (Fig. 14). Beginning on the left, the top register shows three men clothed in tunics and mantles marked with gammata or gammadia (here the Greek letter eta [H]) moving towards a dark-skinned soldier standing in the center (Fig. 15). The forelegs of an animal with cloven hoofs are visible to the right of the soldier. Farther to the right stands an elephant with a decorated collar or harness and shields tied to his sides. His trunk is raised and his eye is large and prominent. Part of another elephant is visible above (behind) the first elephant. The register below this shows an arcade with lighted oil lamps above the arches (Fig. 16). Each arch frames a single figure: young men grasping sheathed swords arranged around a large seated elderly man holding a scroll. Although all of these figures wear tunics and mantles with gammata, they are differentiated by details of their hairstyles. The register below this shows a bull pierced by spears, with blood gushing from his wounds, and a fallen and bleeding soldier grasping a shield (Fig. 17). The three registers are encircled by a delicate wavy ribbon pattern. This mosaic differs in style, quality, and content from the Samson scenes. Its meaning is obscure, as it has no parallels in other ancient synagogues and is only partially exposed. The upper register might depict a triumphal parade, in which case the shields tied to the elephants are the spoils of war. Elephants are not mentioned in the Hebrew Bible, but they occur in 1–4 Maccabees, raising the possibility that these scenes are connected with stories about the Maccabees.
Excavations outside W310 extended below the foundation level to the bottom of the wall (which ended at 27.10 m), but unlike in Sq SW 3/5, there was no thick or compacted layer of building chips. Instead, this part of the foundations appears to have been disturbed by a large rubble collapse.
 
Square SW 5/5. This is a new square opened in 2013 to the south of Sq SW 3/5 (with the mosaic of Samson carrying the gate of Gaza). Because W315, which was in the balk between Sqs SW 4/5 and SW 5/5 appeared to lie on the line of the synagogue’s southern wall, Sq SW 5/5 was opened in the hopes of exposing the synagogue’s main façade. A twentieth-century room bounded on the north by W318 and on the east by W319 was at the ground level.
W318 and W319 rested on an ashy plastered surface (L3142), below which was a hard yellow plaster surface (L3160) that continued under W324 that runs north–south on the western side of the square. Both surfaces (L3142 and L3160) were associated with tabuns of the nineteenth century CE, but the latest pottery from the make-up of Surface 3160 is Crusader–Mamluk. As in Sq SW 2/5, the latest (twentieth century) occupation phase has architectural remains, while in the nineteenth and early twentieth centuries the area around the cistern appears to have been filled with tabuns used for food preparation.
Below Surface 3160 were layers of loose brown and yellow soil (L3173) that resemble the fills below the medieval floor inside the synagogue, and contained pottery dating up to the early seventh century CE. Excavations stopped when a layer of hard plaster bedding was reached (L3181).
This square was heavily pitted, apparently due to the extensive robbing out of the walls, especially the southern wall of the medieval building (W315). The upper part of W315 appears to have been rebuilt at a later date, as indicated by the incorporation of the northern end of W324 into W315. The excavation of the pits revealed an earlier east–west wall (W326) along the south side of the square. The lowest courses of W315—built of ashlar stones resembling W310—terminate on the east at a threshold. The bottom part of the threshold is roughly level with the Pavement 3168 exposed immediately to the north (Sq SW 4/5, above). An apparent foundation trench (L3180) on the south side of W315 cuts through Surface 3160 and L3173, and contained seventh-century CE pottery.
 
Excavations in all of the squares in Area 3000 have yielded the following stratigraphic sequence. (1) In the twentieth century, this area was occupied by architectural remains, apparently belonging to village houses. (2) In the nineteenth–early twentieth centuries, this was an open area with a series of tabuns around the cistern. (3) In the medieval period (the Crusader–Mamluk periods; twelfth–thirteenth centuries), a monumental public building (perhaps a synagogue?) with a cobble and plaster floor originally paved with mosaics was constructed above the Late Roman synagogue. It reused the Late Roman synagogue’s eastern wall and was lined by a bench made of reused blocks from other parts of the Late Roman synagogue. (4) Below the floor of the medieval building is a fill consisting of deliberately dumped layers of yellow and brown soil and yellow building chips, which was deposited on top of the collapsed ceiling above the mosaic floor of the Late Roman synagogue.
It is unclear whether there is a gap in occupation following the medieval period (the Crusader–Mamluk periods) up to the nineteenth century. How long the synagogue remained in use after the Byzantine period also has yet to be determined.