The excavations focused on the ancient town of Bet She‘arim, extending across the Sheikh Abreik hill, which is crowned by the sheikh’s tomb and the horseback statue of Alexander Zaid. The site is located at the western edge of the Jezreel Valley, in an area of easily quarried chalk and nari bedrock. Today, the Bet She‘arim National Park, Moshav Beit Zaid and the Yoffe family farm are located in the area of the ancient town and cemetery on the hill and its slopes. Bet She‘arim is first mentioned by Josephus as a town bordering on the district of ‘Akko, and as the location of Queen Berenice’s tax-collector’s estate (Josephus Life 24). The Sanhedrin rabbinical court moved to Bet She‘arim in the second century CE, when Rabbi Yehudah Ha-Nasi resided there. After Rabbi Yehuda’s death and burial at the site in c. 220 CE, the necropolis became a favored burial place for Jews from the region and from the Diaspora, from Syria to Yemen. The necropolis functioned mainly in the third and fourth centuries CE, and possibly also later in the Byzantine period.
The site was excavated in the 1930s and 1950s on behalf of the Israel Exploration Society and the Hebrew University of Jerusalem. The excavations focused on the necropolis with the burial caves that were hewn in the northern and western slopes of the Sheikh Abreik hill, yielding inscriptions and rich artistic finds. Some of the town’s hillside structures were also excavated, including a synagogue, residential dwellings, an oil press, a basilica, and rock-hewn installations (Mazar 1958; Avigad 1972; Schwabe and Lifschitz 1974); the excavated buildings were only briefly described in reports and in a few articles, and were never fully published. The excavators’ main conclusions were that the settlement at Bet She‘arim prospered mainly from the second to the mid-fourth century CE, when it was destroyed in the Gallus Revolt.
In the 1980s, a salvage excavation conducted in Moshav Beit Zaid unearthed a mosaic-paved structure dating from the Byzantine period (sixth–seventh centuries CE; Vitto 1996). It was later proposed that the necropolis also functioned in the Byzantine and later periods (Weiss 2010). A book presents a reappraisal of the urban structures unearthed in Mazar and Avigad’s excavations (Tepper and Tepper 2004).
The current excavations were initiated in 2014 with the aim of focusing on the town and understanding its character and history. The main research questions were: when did Bet She‘arim become a Jewish settlement and when did it cease to be Jewish; When did the settlement come to an end and under what circumstances; What kind of settlement was it in terms of size, urban planning, construction etc.; and what was the relationship between the settlement and the necropolis.
During the six excavation seasons, nine excavation areas were opened, six (Areas A–F) near the top of the Sheikh Abreik hill and three (Areas X–Z) on the hill’s northeastern lower ridge (Fig. 1). The earliest remains consist of a few finds from Iron Age II–III, without architectural remains. Although Mazar published some Persian-period pottery sherds, the renewed excavations yielded no pottery from this period. The site’s earliest significant stratum is a settlement layer from the late Hellenistic period (second–first centuries BCE) with rock-cut installations, architectural remains and finds. Significant remains from the Early Roman period (first century CE) were mainly exposed in Area B. The most prominent strata in all the excavated areas are settlement strata dating from the second to fifth centuries CE. In the Middle Roman period (second–third centuries CE), the town was built up across the hill, with buildings and rock-hewn installations, such as cisterns, ritual baths and underground hiding passages. The buildings continued in use with some alterations, until the mid-fourth century CE, when a few rooms collapsed or were blocked up. Shortly afterwards, in the early Byzantine period (late fourth century CE), buildings were erected again throughout the town, continuing to be occupied until the first half of the fifth century CE. The meager architectural remains and finds at the site from the fifth–sixth centuries CE attest to the settlement’s decline, and there is no evidence for occupation from the late sixth century CE until the Mamluk period. In the Ottoman period, the tomb of Sheikh Abreik was erected on the hill summit and a small tenant farmers’ village was built. During the First World War, the villagers were evicted, and the hill served the Ottoman army as an entrenched outpost. The latest stratum yielded recent remains dating from the British Mandate-period settlement built by Alexander Zaid and Moshe Yoffe of the Ha-Shomer Organization, until the establishment of the National Park.
Area A
Approximately twenty squares and half-squares were excavated west of the Sheikh Abreik hilltop and east of the basilica (Mazar 1957), exposing remains dating from the Hellenistic or Early Roman periods until the Ottoman period (Fig. 2). The stratigraphy in this area is complex, with both earlier and later walls, sometimes dating centuries apart, lying on the same level, the earlier building stones in secondary use in the later buildings.
Iron Age and Hellenistic finds were recovered in rock cuttings and soil pockets.
In the Roman period (second–third centuries CE), a part-hewn and part-built water reservoir was constructed in the eastern part of the area. The reservoir was divided along its length by two arches, one of which was fully preserved, and its walls were coated with gray plaster characteristic of the first–second centuries CE. Two openings of an underlying cistern were cut in the floor of the reservoir. A large wall was built inside and across the cistern, probably to support the superstructure. In the central part of the area, an underground rock-hewn installation was divided into two cells with a stepped and plastered eastern side that may have served as a miqveh, a ritual purity bath (Fig. 3). Another miqveh was found in the southwestern part of the area. To the west of the installation, a room with benches that was excavated by Mazar and Avigad, was cleaned out again; it probably dates from the Roman period. In the southern part of the area, water cisterns were exposed. In addition, small pillars discovered in the Mamluk stratum, may also originally have been built in the Roman period, when this was probably a residential area. During the fourth century CE, this area underwent changes: the installations were filled with soil and rubbish, walls were built in the arched reservoir, putting it out of use, and some rooms appear to have collapsed.
In the Byzantine period (mid-fourth–fifth centuries CE), the area was rebuilt with a partial change of plan in the walls and openings, and the construction of irregularly angled walls incorporating stones in secondary use. The water installations were canceled: the arched reservoir was now part of a building, one of the cisterns in the southern part of the area was blocked off with a wall, and the miqveh in the southwestern part of the area fell into disuse. An alleyway in the east of the area was probably built in this period. The haphazard alignment of the building walls may be the result of an earthquake in the fifth century CE. Later, in the fifth–sixth centuries CE, the area of the arched reservoir was paved with stones and the western miqveh served as a rubbish pit. The sixth-century CE remains are relatively limited and are located near the surface, including a fragment of a rare pottery artifact decorated with incised motifs that may be a lamp holder or an incense burner (Fig. 4).
Some Mamluk-period walls were exposed, mainly in the central part of the area, as well as packed floors incorporating small stones, and small installations such as a basin and a tabun. Earlier architectural features were reused for building in the Mamluk period. In some places, the Mamluk building penetrated down almost to the bedrock. Many glass bracelets and a small assemblage of weapons also date to this period.
In the Ottoman period, a structure paved with floors of small, packed stones was built in the northern part of the area. One of the walls is tilted, as if damaged by an earthquake; this may have been caused by the 1837 earthquake that devastated Safed (A. Agnon; pers. comm.).
Area B
About twenty squares and half-squares were excavated near the eastern side of the hilltop, where no archaeological remains were visible on the surface before the ecavation (Fig. 5). The strata here lie on top of each other, as in a tell, and they reach a depth of about four meters from surface down to bedrock. The remains date from the Iron Age until the present day, the earliest finds being a few Iron Age II pottery sherds retrieved in cracks in the bedrock in the central part of the area.
Remains of a structure, whose walls were built in characteristic Phoenician style with roughly dressed upstanding stones interspersed with fieldstones (Fig. 6), date to the Hellenistic period (probably the second century BCE). Pairs of rock-hewn bell-shaped silos had stone lids over their openings. Pottery sherds characteristic of Hellenistic Phoenicia were recovered.
In the next phase, probably in the first century BCE and the beginning of the Early Roman period, a long wide wall, blocking the earlier Hellenistic building on the north, and additional perpendicular walls were built. Some of these walls were constructed over the mouths of the earlier silos (Figs. 7, 8). A bath or a small installation, and first-century CE in situ pottery were exposed in the central part of the area. In this phase, the building may have served as a fort, similar to other forts in the region (Sha‘ar Ha-‘Amaqim, Horbat ‘Aleq), and it may later have been part of Queen Berenice’s estate described by Josephus.
In the Roman period, a large structure built on top of the earlier walls had ashlar-built walls and rooms paved with stone slabs, most of which were robbed in recent centuries. The building may have been an administrative center built in the second century CE and functioning until the third or fourth centuries CE. Rooms near the eastern side of the building yielded rich third-century CE finds, as well as tubuli pipe fragments and marble slabs that indicate the presence of a nearby bathhouse. This area, near the hilltop, was probably the town’s civic center in the Roman period.
After the mid-fourth century CE, the area underwent a complete transformation, and an oil pressing complex was built above the remains of the large building in the western part of the area (Fig. 9). The round base stone of a crushing installation was uncovered, to the east of which were three stone bases of direct-pressure screw-type pressing installations. Two of the bases were preserved in situ, their channels draining westwards into plastered collecting vats; one base was found in secondary use. The pressing bases have a circular groove, similar to the pressing stones of the oil press excavated in the past in the northern part of the settlement. The bases are made of limestone, not from the local nari rock, and must therefore have been brought from elsewhere. In this period, some of the Hellenistic silos were converted into water cisterns. In the Byzantine period, this central part of the Roman town was evidently transformed into an industrial area housing various installations.
The nature of the area in the Mamluk period is not known, since only mixed fills from the Ottoman period yielded Mamluk sherds.
In the Ottoman period, the area was close to the sheikh’s tomb. Deep pits found throughout the area were filled with layers of ash and rubbish from all periods, with no architectural remains. Since some of the pits cut into walls and floors, they were probably the result of dynamite detonated by the Ottoman troops entrenched there during World War I. In the British Mandate period, the Yoffe and Zaid families built a village on the site and leveled the area with soil (A. Yoffe, pers. comm.).
Area C
Fourteen squares and half-squares were excavated in two sub-areas between the sabra bushes on the eastern hillslope; large building blocks and cisterns were visible on the surface prior to the excavation (Fig. 10: C1, C2). The excavation aimed to determine the extent of the settlement and its outer limits on the eastern slope. The area yielded finds dating from Iron Age II until the Ottoman period. In the 2019 season, a probe trench dug lower down the eastern slope, below Area C, did not expose any ancient remains, indicating that it lay beyond the limits of the ancient settlement.
Area C1. The earliest finds were Iron Age II sherds retrieved on the bedrock. A street sloping down the hill, of which only the drainage channel was preserved, is attributed to the Late Roman–early Byzantine period. The rooms of houses on either side of the road were probably built in the third or fourth century CE and destroyed in the early fifth century CE. One of the rooms contained collapsed walls mixed with a large quantity of potsherds and a hoard of coins dated to the second half of the fourth century CE. The collapsed rubble yielded a bifacial Iron Age stamp seal that was in secondary use (Fig. 11; Erlich 2020).
Meager remains were attributed to a later phase, probably the Byzantine period. Above these remains, about fifteen pit graves without diagnostic finds found throughout the area were the late Ottoman-period graves of Muslims who wished to be buried near the sheikh’s tomb.
Area C2. The earliest remains comprise rock-cuttings, including a cistern that was visible before the excavation and that could not be dated; it was breached by several rooms. The surface remains included floor foundation layers, exposed in a few places and attributed to the second century CE. Further down the slope, patches of floors, walls and a section of a street built in the third century CE and in use until the early Byzantine period (fourth–mid-fifth centuries CE), were exposed. In the fifth century CE, stone paving was laid in the western part of the area, overlying earlier walls; the paving continued in use until the sixth century CE. Some of the underground spaces were reused in this period. The area was not settled after the sixth century CE.
Area D
Water installations, streets, rooms and a hiding complex, probably all part of a residential area, were exposed in eleven excavation squares (Fig. 12). Based on the pottery, the remains are dated mostly from the Early Roman to the mid-Byzantine periods, although some earlier sherds were also retrieved.
The northern part of the area. Four Roman-period water storage installations were excavated: an installation comprising a row of gray-plastered hewn chambers—one vaulted—that led to an underground chamber that led into yet another chamber (Fig. 12:1); a Roman-period cistern (Fig. 12:2); a westward-flowing water channel (Fig. 12:3); and a cistern with two entrance shafts, one stone-lined and the other plastered (Fig. 12:4). This area was probably an open area, possibly a courtyard, containing mostly water-collection installations. The cistern and other rock-cuttings, as well as patches of floor bedding, belong to the earliest phase, which probably dates to the second century CE. Two streets forming a T-junction are also attributed to the Roman period (Fig. 12:7). The street paving was not extant, but the drainage channel was found curving round between the streets and leading westwards. Middle Roman-period pottery recovered from the street bedding on both sides of the channel dates the street’s construction to about the second century CE.
In the early Byzantine period, one cistern opening (Fig. 12:2) was blocked up with a wall and rooms were built. A room with a floor overlain by collapsed rubble was exposed here (Fig. 12:5); about 80 coins discovered in the collapsed debris and in the floor bedding, provide a terminus post quem of 363 CE for the floor’s construction. The room fell out of use only one or two generations later, at the beginning of the fifth century CE. A few fifth–sixth-century CE finds were discovered above the early Byzantine habitation level, mainly in stone collapse layers and on the surface. A stone podium and an adjacent basin (Fig. 12:6) in the northeastern part of the area are also attributed to the Byzantine period. Based on the pottery in the street-drainage channel, it probably became blocked up at the beginning of the fifth century CE.
The southern part of the area. South of the street remains, a few rooms were built in the Roman period. One part-hewn and part-built room was divided in two by an entirely preserved arch (Figs. 12:8; 13); its floor exhibited rock cuttings and openings into an underground complex (below). The room was built in the first or second century CE and abandoned in the mid-fourth century CE, when it was almost entirely filled in with rubbish. In the fourth century CE, the passage to the arched room was blocked, and a rectangular plastered pit yielding complete fourth century CE pottery vessels was hewn. In the early Byzantine period, a packed earth floor was lain above the arch, for which purpose the upper part of the arch was leveled and served as a wall. This floor was overlain by a crushed chalk floor from the middle Byzantine period.
To the east of this room, additional rooms were also built in the mid-Roman period and used until the fourth century CE, and then reused in the Byzantine period.
A hiding passage was hewn northward from the southern corner of the arched room; it cut through earlier cavities such as the cistern to the north of the streets’ drainage channel (Fig. 12:2). The passage had several branches, an entrance shaft, a sealing stone, lamp niches, and a few levels and small chambers (Fig. 14). After the passage was hewn, the opening from the arched room was blocked up; the finds from this blockage date to the second century CE. One chamber yielded meager finds that probably date from the third–fourth centuries CE. The passage was probably hewn at the time of the Bar Kokhba Revolt, and it may have been reused later.
The three southernmost excavation squares revealed rooms and architectural remains. One room had several superimposed packed earth floors; the earliest surface, overlying Hellenistic pottery, dates to the Early Roman period, and the latest dates to the fourth century CE. The upper part of another room contained an early Byzantine podium made of small stones (Fig. 12:9). Several construction phases and walls were also discovered (Fig. 12:10), including a mid-Byzantine phase (fifth–sixth centuries CE).
Area E
Two squares were excavated to determine the extent of the settlement on the hill’s western slope. No architectural remains, and only mixed potsherds and other mixed finds were found. This area probably lay beyond the boundaries of the settlement.
Area F
Down the hill’s southeastern slope, no remains were discovered, apart from a remnant of a small wall that is probably recent.
Area X
The excavation was located in a mostly roofed-over area adjacent to a storehouse on private land in Moshav Beit Zaid, c. 100 m southeast of the remains of the previously excavated synagogue. A few years ago, after the Israel Antiquities Authority halted damage to antiquities caused by mechanical equipment removing large ashlars, the damaged area was re-excavated with the landowner’s consent. A layer of modern refuse was removed, and four squares and half-squares were excavated, exposing a large gatehouse from the Roman period in the southwestern part of the area, and an early Byzantine pottery kiln in the northeastern part (Fig. 15).
The gatehouse comprised a wide entrance flanked by two square rooms. The northern room was built of large ashlars with drafted margins on the external faces, with a fill of fieldstones (Fig. 15:1). This may be the foundation course beneath the floor level, or a room containing a stone fill from a later phase. The foundation course of the southern room was only partially preserved (Fig. 15:2). The entrance (width including doorposts 4.3 m, width of opening 3.9 m; Fig. 15:3) was built of two large threshold stones containing a hole to secure the gates, and sockets for the hinges. The northern threshold stone (length 1.85 m) was found in situ, while the southern threshold stone (length 2.4 m) was found in the excavation debris pile towards the end of the season, and it had probably been shifted when the building was damaged; the entrance probably had a double door. The gatehouse led to a small plaza paved with rectangular stone slabs (Fig. 15:4) and continued to a street that was made of crushed chalk in its final phase. Later walls abutted the southwestern side of the northern room, where a cistern was also found (Fig. 15:5). Most of the pottery in the gatehouse dates to the fourth century CE. Although the probes dug beneath the flooring did not yield sufficient pottery to date it, it was probably built in the Roman period and functioned until the fourth century CE. It may be one of the town’s gates, whence a road led eastward, possibly to nearby Simonia (Tel Shimron).
The pottery kiln is round and was built inside a square room (Fig. 15:6). Another room adjacent to the gatehouse, near the western side of the kiln, was probably the potter’s workshop. The kiln’s firebox was preserved, with a round central pillar built of fieldstones coated with a thick layer of clay and plaster, and a stoking vent for fuel on its northeastern side. The kiln’s outer wall was built of fieldstones and coated on the inside in the same manner. The kiln floor was lower near the stoking vent and higher further inside to enable the heat to spread more efficiently from the firebox to the vessel firing chamber. A deep pit outside the stoking vent served to feed the fire. The vessel firing chamber was not preserved, but the extant rubble contained many fragments of tile- or tabun-like material.
The rubble in the kiln yielded a large quantity of potsherds dating from the second half of the fourth century CE, most of them bag-shaped jars fired in the kiln. The kiln probably operated for a short time in the early Byzantine period before it collapsed. Similar kilns, producing mostly jars, are known from elsewhere in Western Galilee, as at Horbat ‘Uza. Limited evidence of a glass-blowing workshop was discovered above the kiln.
Area Y
Four squares and half-squares were opened southeast of an oil press excavated in the past (Avigad 1954) in order to date the press and the structures in its vicinity. The area contained several rooms of a residential building that was built and used in the early Byzantine period and destroyed in a large collapse in the first half of the fifth century CE. A small excavation in the southern entrance of the oil press’s eastern room provided a similar date. The oil pressing installation and its dating are similar to the oil press excavated in Area B. The gatehouse excavated in the past west of the Area Y oil press (Avigad 1954) is similar in plan and size to the gatehouse discovered in Area X. Consequently, its dating, its stratigraphic relationship to the oil press and the question if it served as a gate, or as a passageway between the city and the cemetery on the slope to the north must be reconsidered.
Area Z
A narrow strip of four squares was excavated between the Area Y oil press and the synagogue that was excavated in the past, c. 20 m from the modern houses adjacent to the synagogue. In the center of the area, a stepped miqveh coated with gray plaster, was exposed, partly roofed with a vault that had collapsed and partly underground and roofed with a gabled ceiling (Figs. 16, 17). Based on the plaster, the miqveh dates to the first or second century CE. The soil fill in the miqveh yielded fourth-century CE coins. Southwest of the miqveh, second–third-century CE potsherds were embedded in plaster and in a channel adjoining the bedrock. A large, plastered chamber to the north of the miqveh was probably initially used as a large pool adjoining the miqveh.
The main stratum uncovered in the area dates to the early Byzantine period. In this phase, or possibly earlier, the entrance to the miqveh was blocked by a massive wall and the miqveh was used as a rubbish pit. Stone-built rooms from the previous phase found on all sides of the miqveh and above it, included upside-down stones with drafted margins, and an out of place column drum. Beyond a narrow corridor northeast of the miqveh there was a room with three pilasters and two floors, which reduced the size of the earlier pool. The Byzantine buildings collapsed in a devastating event that is evident throughout the area, which yielded a large quantity of shattered domestic pottery, glass vessels and coins. The pottery assemblage from the early Byzantine stratum (at least 2 m deep) is homogeneous, dating to the mid-fourth–first half of the fifth century CE.
Meager architectural remains and a few fifth–sixth-century CE potsherds were found above the collapsed remains of the main stratum. The area was not reoccupied after the Byzantine period.
The excavation shows that Bet She‘arim was first settled in Iron Age II, probably in the ninth century BCE. The excavated areas yielded no significant finds from the Persian and early Hellenistic periods. The earliest architectural remains were found on the hilltop, where a building was constructed in the second century BCE. In this period, small bell-shaped storage installations and possibly also cisterns were hewn into the soft chalk bedrock. The Hellenistic settlement contained Phoenician-style walls and Phoenician pottery types. In the first century BCE, a substantial building, possibly a fort, was built on the hilltop, in use until the first century CE. This building may have been part of the royal estate and the camp of the Roman commander Aebutius, against whom Josephus fought (Josephus Life 24).
In the Middle Roman period, the late first and the second century CE, an urban settlement developed at Bet She‘arim and continued into the mid-fourth century CE. The town included residential and public buildings, streets, cisterns and rock-hewn water installations and ritual baths. Although it was only a small town, the settlement was built in the style of a Roman polis, with an orthogonal street plan between residential and public buildings, such as the buildings in Areas A–D, the basilica, the synagogue excavated by Mazar, and its adjacent neighborhood. The synagogue, cemetery, ritual baths and chalkstone vessels attest to the town’s Jewish character. This stratum yielded Corinthian capitals and small marble slabs for lining walls and for paving, attesting to the high socio-economic status of the town’s population. Mazar and Avigad suggested that the town covered an area of about 100 dunams, but the current excavations indicate that it was probably slightly larger, although it is still too early to determine its extent. Although no town wall has been found, a second large gate was discovered. It may cautiously be proposed that the gates granted the town its name, Bet She‘arim (‘House of Gates’). During the Roman period, there were slight changes in the plan and use of the various spaces in the stratum: in the second century CE, perhaps in the context of the Bar Kokhba Revolt, a hiding passage was hewn that breached earlier underground spaces; and in the Late Roman period, the ritual baths were converted into rubbish pits. In the middle of the fourth century CE the site apparently underwent a change, attested to by the collapsed buildings and filled-in spaces. However, there is not much evidence for this, and it is impossible to determine the exact date and circumstances of the town’s destruction and decline in the mid-fourth century CE.
In the early Byzantine period, after the mid-fourth century CE, the town was rebuilt on different lines from that of the previous settlement. Although in most places the lines of the ancient walls and the orthogonal plan were preserved, some changes are noticeable: the houses in Areas Y and Z deviate from the orthogonal plan, their construction was simpler and they incorporated older elements in secondary use. The settlement was large, but unlike its former impressive urban style, it was rural and production oriented. Two oil presses, a pottery kiln and a glass industry found at the site were incorporated between and beside the houses, and one of the oil presses was even built on the hilltop, in a place that was formerly the town’s civic center. At this stage of the research, the identity of the town’s population is still unknown. The early Byzantine stratum was destroyed in a single, violent event, which included large-scale stone collapse and the destruction of buildings, evident in most of the excavated areas. The town’s destruction may be attributed to the earthquake of 419 CE.
After the town’s destruction in the early Byzantine period, the settlement on the site was reduced in size. From the fifth–sixth centuries CE, there are meager settlement remains in some areas. Architectural remains, installations and a few imported vessels are attributed to this period. The architectural remains include the building paved with mosaics, excavated in the 1980s (Vitto 1996), which may have been located in the center of the settlement or may have been part of a farm or monastery.
No remains from the Early Islamic period have been found; the settlement may have shifted to a nearby location. The hill was reoccupied in the Mamluk period, probably in the fourteenth century CE. Remains of a building, possibly a fort or a farm, were discovered in Area A. In the Ottoman period, the tomb of Sheikh Abreik was built on the hilltop and a small hamlet of tenant farmers was built on its slopes. In the First World War, an Ottoman military post was built on the hill summit.