The excavation focused exclusively on Area A where the western wall of a synagogue had been discovered in 2010 (Fig. 1). Work continued in Sqs AC 29/30 that had been left unfinished at the end of the 2010 season (Fig. 2). In addition, the area was expanded to the south (AC 31, not entirely excavated, and AD 31), to the east (AD 29/30, AE 30/31 and AF 30/31) and to the north (AC/AD 28). A cistern in Sq AC 28 was found sealed and initial exploration was undertaken by Y. Shivti’el.
The aim of activities in this area was to establish the size and layout of the synagogue, to retrieve sufficient data on its chronology and to gather as much information as possible on its decoration, inventory and accompanying finds.
Stratigraphic excavation in Sqs AC 28–30 and careful comparison with the results from the 2010 work in the same area produced important data about the accumulation of artificial fills between the western wall of the synagogue (W7018) and the 'stylobate wall' of the portico to its west (W7073). This fill was bound by a single-faced retaining wall (W7148) in the south and by a terrace wall (W7114) in the north (Fig. 3). The fill consisted of three distinct layers: A lower, brownish gray layer, placed on top of a thin, natural layer of reddish brown soil directly above bedrock; the gray layer contained large fieldstones that meant to raise the surface level, few tesserae, potsherds and hardly any coins (L7125, L7190); An intermediate, hard, grayish layer containing much plaster, many pebbles, potsherds, several coins and very many tesserae (L7157 ;'tesserae layer'); An upper, soft brownish layer covered with cobbles and containing many tesserae, some potsherds and very many coins (L7156; 'coin layer').
All potsherds down to the lowest layer, as well as the readable coins, predominately date to the early Byzantine period (c. 380–425 CE; Fig. 4). However, a few coins can be dated to the reign of Marcian (450–457 CE; four coins) and Justinian (527–565 CE; two coins), while others can most probably be identified as imitative issues (see Callegher 2007, Pp. 49–51). At least, two possible Umayyad potsherds from Coin Layer 7156 supplement the picture, combining a vast majority of early Byzantine finds with meager finds dating more than 100 years later. The intriguing question is how to explain this situation. A satisfactory answer is not an isolated interpretation of the numismatic material, but a comprehensive analysis of the entire archaeological context in which the coins were found.
The large amount of crude, limestone fieldstones, and the large number of tesserae in conjunction with pebbles, plaster chunks and plaster-coated potsherds suggests that all three layers are construction debris, whereas the thin, natural reddish brown layer above bedrock and below the fieldstones once constituted the original surface. The most likely source for this construction debris was the synagogue itself, except for the layer of fieldstones, which was probably deposited when the original mosaic floor was removed in the course of repair works and replaced by a now still well-preserved plastered floor (Fig. 5). Preliminary architectural analysis suggests that the floor removal occurred in the context of a major rebuilding of the synagogue. The removal of the mosaic and its subsequent dumping outside the building resulted in what could be called 'reversed stratigraphy'. At the moment, the following scenario of how this fill was built up seems the most probable:
Initially, the space between the pillars of the portico wall’s first phase was closed by W7114 in the north and W7073 in the west, both built of fieldstones and W7148 in the south, built of much higher quality masonry. These walls now formed a rectangular annex to the synagogue’s western wall (W7018).
Next, the existing bedrock level, i.e., the original surface outside the building, was raised and leveled with a layer of fieldstones (L7125/L7190). Then, the presumably damaged original mosaic floor of the synagogue was broken up, crushed, removed through the synagogue’s western exit and dumped on the layer of fieldstones, creating the 'tesserae layer' (L7157) that contained much plaster.
Furthermore, immediately after the mosaic had been torn out, the builders hit the sub-floor bedding of the original mosaic floor, consisting of pebbles, plaster and many coins and originally deposited just before and during the mosaic’s construction. The builders had dug everything out down to bedrock and dumped the debris, together with chunks of plaster and very few potsherds, right on top of the previous Tesserae Layer 7157 (Fig. 6). Due to the fact that the new Layer 7156 contained c. 1000 coins, it was dubbed the 'coin layer' (Fig. 7). Each coin was carefully noted, enabling us to plot them on a three-dimensional plan that indicates the original location of each coin, thereby helping us to understand how Layer 7156 was built up (Fig. 8). As a consequence of this filling activity, the bedding, once buried below the mosaic, now came to be placed on top of it in the portico, creating the above-mentioned 'reversed stratigraphy'. The dumping of Coin Layer 7156 on Mosaic Layer 7157 was not done carefully, mixing elements of both into a compact and complex agglomeration of soil, potsherds, coins and building material.
Finally, the surface of this artificial fill (L7156) was beaten hard and strengthened with cobbles to create a new walking surface that connected the western entrance of the synagogue (L7094) with the threshold of the portico in W7073.
As mentioned above, the coins from the portico area show two clear chronological clusters: a very large one from the early fifth century CE, and a small one from the second half of the sixth century CE at the earliest. Although the two clusters cannot be easily separated stratigraphically, they find their best explanation in the 'reversed stratigraphy' hypothesis. If seen in the broader context of the history of the synagogue building, the 'reversed stratigraphy' hypothesis fits in well with changes observed in the architecture of the building. The mixing of the two chronologically distinct coin clusters can be understood as the result of a particular refurbishment activity that must be dated, according to the last coin cluster, to the second half of the sixth century CE at the earliest, namely Phase II. The earlier coin cluster, then, gives the terminus post quem for the construction of the first phase of the synagogue, i.e., not longer after c. 425 CE, namely Phase I.
An alternative scenario is also conceivable. If the distinction between two separate chronological clusters is not accepted and the coins are regarded as a single unit, the latest coins will date the entire fill. In this case, one would have to assume that the earlier coins had a much longer period of circulation right into the late sixth century CE at the latest, which theoretically is entirely possible, but this assumption would leave the obvious changes in the architecture, e.g., addition of the portico, replacement of the floor inside the synagogue and the origin of the mosaic accumulation, totally without context. This is the reason why the two-phase hypothesis seems the most viable at the moment. The careful examination of the accompanying pottery will undoubtedly aid in refining/correcting this scenario.
Excavating Sqs AC/AD 28 identified the northern end of the synagogue. Wall 7114, as well as the short section of the bench (W7140), shows that the north–south axis of the synagogue was c. 11 m long. A thick layer of plaster covered the northern face of W7114, whose extension beyond the square limits is yet unknown. The wall line, interrupted by a massive rectangular corner stone, continued further west as the northern perimeter wall of the portico and for another c. 50 m as a terrace wall of the upper Horbat Kur plateau.
Remains of possibly three pools (L7129, L7295, L7160) were found attached to the northern synagogue wall (Fig. 9). The central pool (L7295) was later filled and covered with a cobblestone floor (L7172), into which the pool’s opening was integrated. A cistern (L7187; Fig. 10) to the north of Pool 7172 was found intact and sealed possibly because it was closed sometime at the beginning of the Byzantine period. The accumulation of soil inside the cistern (L7188) contained two complete early Byzantine cooking pots. Palynological analysis of the sediment revealed a large amount of pollen, now under examination. It is likely that the pools and the cistern were used to collect rain water from the spacious roof of the synagogue, although a ritual use of the pools can not be excluded. The eastern pool (L7129) was excavated down to its bottom, as far as the size of the square allowed. Below a thick collapse layer, two sediment layers (L7143, L7154) containing only Byzantine pottery, were found. Two columns were found standing upright on the lowest sediment layer (L7154), suggesting that alterations to the form or function of the pool were carried out sometime during its later stages of use. Only the southern end of Pool 7160 in the west was detected and the rest could not be excavated.
Squares AC/AD 31 were opened to explore the southern end of the synagogue and gain information on a structure beyond it. Both squares, together with Sqs AE/AF 30/31, revealed the full course of the southern synagogue wall (W7131, W7123, W7134 and W7191) and provided data on its construction and layout. It seems that the walls of the synagogue to the south, west and north were built in the same method and were plastered inside-out; the eastern wall of the synagogue is not yet clear enough to draw further conclusions. Although only little plaster is preserved in situ, especially close to the southwestern corner, it is likely that the synagogue was once painted white and the black basalt did not show. Pieces of colored plaster were discovered in the fill, inside and immediately outside of the synagogue, especially to the south; these suggest that the interior of the building might have been partly decorated with red, green and yellow hues on a white background color (Fig. 11). An entrance, having a single stepping stone outside and three steps inside (L7133), was exposed between W7123 and W7134 and enabled access to the synagogue from the south. Compared to the entrance in the western wall (L7094), this was not the main entrance to the synagogue. Scattered architectural fragments in the eastern part of the synagogue area suggest that there may have been another entrance on that side as well.
Outside the southern synagogue wall, a sequence of beaten earth floors (L7138/L7146, L7141/L7169) that covered bedrock, indicate a large open space. To create this surface, a large amount of construction debris was used; it was mixed with plaster chunks, more than in the area of the portico, and contained almost no coins. Where did this debris come from? On the one hand, the lack of coins might suggest that the material was not taken from inside the synagogue, but its contents, mostly roof tiles, wall tiles and large quantities of thick mortar and plaster fragments, indicates that this fill was connected to the same refurbishment phase as the addition of the western portico described above.
Parallel to and south of W7134, three steps of a staircase were found (L7180); these might have led to an upper story of the synagogue or to the roof of another building adjacent to it on the south.
The two central squares (AD 29/30) were excavated to understand the interior layout of the synagogue and the nature of the fill accumulations in it. Both squares were filled with a thick layer of destruction debris, building stones and plaster chunks, mixed with loose, brownish soil, from below topsoil almost down to the plastered floor of the building. The potsherds in this fill were predominately Byzantine, with a few fragments of medieval glazed ware. The latter, albeit few in number, occurs particularly in the layer below topsoil. The glazed ware and the loose, mixed character of the debris layer are considered to be an indication of stone robbing some time during the Middle Ages, perhaps during the Crusader-period activities in Tiberias. This means that the synagogue debris encountered during the excavation does not directly reflect the situation of the immediate collapse, probably during the earthquake of January 18, 749 CE (see HA-ESI 123), but rather what was left of this debris after the plundering of usable stones in the Middle Ages. The presence of Ottoman clay pipes and some very late pottery in the uppermost topsoil further suggests that stone robbing may have also taken place in more recent times (see the placement of Stone Step 7002 in Area AB 29 in HA-ESI 123).
Below the debris layer was a thin layer of sediment with very few finds, overlaying the latest floor of the synagogue (L7167, being the continuation of Floor 7080 in Sqs AC 28–30). Three pillar bases and a line of stones connecting them (W7186, W7176, partly robbed out) separate a higher western area (L7167/L7206) from a lower eastern area (L7205/L7207). This arrangement resembles the aisle-nave-aisle order of a basilica, but the present lack of information from the eastern side of the synagogue precludes any firm conclusions. The step-like relation of the western aisle and the nave may suggest that the floor did not only follow the terrain slope to the north, but also to the east. The fact that the present plastered floor slopes up to the pillar bases, the steps (W7185), and the eastern ‘bema’ wall (W7112) proves that it is later than these features and corresponds to our scenario about the build-up of soil in Sqs AC 28–30 (see above). Fallen debris has caused considerable damage to the plaster floor, especially in Sq AD/30; holes partially revealed a lower layer of gray plaster set on a bedding of pebbles and more plaster (L7208), which directly overlaid bedrock.
It became more obvious during the excavation that W7112, located c. 7 m east of western exterior wall (W7018/W7039), cannot be the eastern wall of the synagogue as anticipated on the basis of aerial photographs from 2010. Its makeup differs considerably from all exterior walls discovered so far, and it did not continue further north. It was therefore decided to open two new squares (AE/AF 30/31) to clarify the character of W7112 and to find the eastern wall of the synagogue. The southern synagogue wall continued as W7134 in Sq AE 30/31, separating an area of tightly packed building stones to the south from the usual debris layer inside the synagogue. A perpendicular wall (W7183) and a decorated entrance (L7200) were exposed parallel to W7112, forming the perimeter of the bema (podium for the Torah shrine). Unlike the rest of the synagogue architecture, the parts of the bema that are still in-situ, consist of carefully dressed limestone of high quality (Fig. 12). The bema’s northern end is still unexcavated, so the exact dimensions of the podium are still unknown; yet, the remains already visible suggest that it is one of the biggest podiums excavated to date.
Little was left of the decoration that once adorned the bema. Some fragments of worked limestone, which were exclusively found in the area between W7112, W7183 and 7200 (a laurel wreath with rosette), in the debris (L7181) south of W7134 and on top of Staircase 7180 (fore part of a lion?), suggest that it once was elaborately embellished (Fig. 13). The narrow Entrance 7200, of which only the lowest course is preserved, could have given access to an interior room under the platform of the bema (cf. Umm el-Qanatir), although it still remains unknown how high that room originally was.
Due to erosion on the top slope of the hill toward the east intact architectural remains were found at deeper levels than in the other western squares. The southern synagogue wall continued for c. 2 m into Sq AF 30/31 (W7191), and created a corner with a perpendicular wall (W7193), which seems to be the best option for the eastern exterior wall. Hence, the size of the synagogue can be estimated on the basis of this evidence to be c. 11 m long and c. 16.5 m wide. Together with the evidence in the central squares (AD 29/30), a possibly in situ column drum in Sq AE 30/31 east of bema wall W7183 and another fallen, plastered column (min. length 1.5 m) next to it, support the assumption that the synagogue, despite its broad layout, was a basilica-type structure.
Apart from W7193, which may possibly be the eastern wall of the synagogue, the situation in Sq AF 30/31 still remains quite unclear at present. A plastered area in the center of the square and walls perpendicular to W7193 from east and west indicate intensive building activity in this area. At least in its northern part, W7193 seems to have been built as a double wall, showing a layer of plaster between two lines of stones in the core of the wall. Wall 7193, just like W7114, very likely served as a terrace wall, delineating also the eastern end of the Horbat Kur plateau.
The synagogue at Horbat Kur turned out to be much wider than Galilean or Golan synagogues; it is somewhat similar to the broad synagogue at Khirbet Shema’. Horbat Kur resembles more the southern, Judean “broadhouse type” synagogues, e.g., Khirbat Susya, Eshtemo‘a, Ma‘on and ‘Anim; however, unlike them, Horbat Kur has columns like a basilica, it is entirely built of local stone and shows peculiar “northern” features, e.g., the side door in the bema, as in Umm el-Qanatir.
On the basis of the two-phase hypothesis, explained above, a date for the construction of the synagogue is proposed. It seems to have been constructed shortly after 425 CE (Phase I) and substantially refurbished during the second half of the sixth century CE (Phase II). The synagogue remained in use until the devastating earthquake of 749 CE, after which permanent habitation ceased on Horbat Kur.
The discovery of the synagogue at Horbat Kur contributes to the ongoing debate about the design and dating of synagogues that formed an important part of rural life in the ancient Galilee.