Area A. Seven squares were located in the southern part of the excavation, south of the residential quarter of the kibbutz. The excavation, which was conducted in alluvium devoid of finds, reached bedrock. On the bedrock were three gray hearths (each c. 1 m in diam.), containing pottery sherds, animal bones and several flint items. The hearths were not hewn in bedrock, but rather utilized natural pits. No architectural remains were discovered in the vicinity of the hearths. The hearths, which date to EB IA, resemble those from other sites located to the north of Nahal Soreq: Teleilat Batashi, at the bottom of the Hafetz Haim reservoir, in a section along a road at Yesodot and in trial trenches at Mishmar David.
Area B (Fig. 4). Fifty-five squares were excavatedslightly to the south of the hilltop, the highest of the excavation areas. The earliest construction dates to the Byzantine–Early Islamic period (Stratum VI). Several wings of a large building surrounding a central courtyard were revealed in this area. An installation paved with a mosaic floor—either a winepress or a bathing facility—was found in the corner of the courtyard. The walls of the building (thickness c. 0.6 m) were founded on bedrock, on a terraced slope descending from north to south. A structure that included three longitudinal halls divided into several rooms by partition walls was built on the northern terrace, north of the courtyard. The walls on its northern side were built without openings. The southern hall was connected by way of openings and arches in its southern wall to the central courtyard. A building extending to the south of the courtyard comprised two long halls separated from the courtyard by a wall with windows. A long hall with partition walls was revealed to the east of the courtyard. The eastern hall connected the northern and southern wings. In Stratum VI, no evidence was found indicating that the western side of the building was closed; hence, it seems that the structure was U-shaped.
Stratum VI was destroyed in an earthquake (possibly in 749 CE), after which a number of new walls were built in the area (Stratum V). Three massive walls were constructed on the western side. They enclosed the building and changed its plan from a U-shaped structure to a square structure, comprising rooms surrounding the central courtyard of Stratum VI. Stratum V is ascribed to the Abbasid period and the beginning of the Fatimid period. This immense building was damaged, possibly in another earthquake, either that of 1033 or of 1068 CE.
Following the destruction, several massive walls were erected inside the structure, apparently in order to reinforce it (Stratum IV). This stratum is ascribed to the end of the Fatimid period. The building with massive walls was used over a long period of time. Apparently, it was gradually abandoned, until it was finally deserted. The floors of its final phase were slightly burnt.
Only one wall (Stratum III) was constructed on top of the burnt layer; it adjoined one of the massive walls from the previous stratum. This stratum is ascribed to the Early Mamluk period. Simultaneously with the construction of the, floor levels were raised in the rooms, and several stone installations were built in the area. These included stones set in circles and rows, the nature and purpose of which are unclear. Several Ottoman-period pottery sherds were collected on the surface where remains of the foundations of farm buildings belonging to Kibbutz Mishmar David (Stratum I) were found.
The ancient buildings in Area B were probably part of an administrative complex/building or a khan. The building is quite unlike the dwellings found in the other excavation areas. Most of its rooms were apparently used for storage. The public function of the building continued until the earthquake that destroyed Stratum IV, after which the inhabitants occupied what remained, reinforcing or reusing its walls.
Area C was divided into three sub-areas (C1–C3).
Sub-Area C1 (Fig. 5). The earliest building remains found in this area date to the Late Roman or Early Byzantine period (Stratum VII). A wide entrance between two massive walls was exposed in its western part. The threshold, with two sockets, was hewn in bedrock and the doorjambs were built of large ashlar stones. Towers belonging to Stratum VI and V (below) were founded right over the wide doorway of Stratum VII, dating it to the Roman or Early Byzantine period.
During the Byzantine–Early Islamic period (Stratum VI), construction in this area was regular and of high quality. Remains from this stratum were exposed only in probes, and the architectural remains are fragmentary, but the general planning of the area is clear. In this Stratum, the area was divided into two parts: an eastern area with private dwellings and a western area with a large free-standing tower set atop a massive stone base. The two areas were separated by an alley or a street running north–south. The houses were built in rows separated by alleys, some of which were paved. The walls of the houses were either constructed of fieldstones set on top of walls from the previous stratum or rock-cut. Similar construction was also noted in the corresponding strata in Area C2 and C3. Due to excavation constrains, the tower itself was not unearthed, but parts of its massive stone base could be discerned. The buildings and tower of Stratum VI were destroyed by an earthquake, perhaps in 749 CE.
Following the earthquake, a new quarter of private houses (Stratum V) was built above the previous dwellings. An almost square tower (Stratum V; wall thickness 0.9 m) was built on the remains of the destroyed tower of Stratum VI. The tower’s long side, oriented east–west, was exposed for 13 m, and its short side for 12 m. Its eastern wall adjoined a square room that had rounded supporting beams (diam. c. 1 m) in three of its corners; the fourth corner, in the southwest, was destroyed by agricultural activity. The purpose of these supporting beams is unclear, as they did not support corner towers. They may have served to support the building’s corners following the earthquake; alternatively, they may have provided the structure with a more aesthetic appearance. The Stratum V buildings were destroyed by a second earthquake, either the one that struck in 1033 or that of 1068 CE. The partially destroyed buildings were renovated (a second phase in Stratum V).
Stratum II, dating to the Late Ottoman period and the beginning of the British Mandate, was noted only along the eastern fringes of the area. Several wall foundations ascribed to this stratum were exposed.
Sub-Area C2. Sixty-six squares were opened. In the southeastern part of the area work was suspended once a small children’s cemetery was discovered (Stratum I). In the northeastern corner of the area, two stepped buildings (Stratum VII) were constructed on three bedrock terraces moderately descending from east to west. The walls of the buildings were in part built on the bedrock and in part rock-hewn. A similar phenomenon was observed in the other sub-areas in Area C. On the lower, westernmost terrace, the walls continued westward beyond the excavation boundary, but were destroyed when the kibbutz buildings were constructed. Architectural remains, drainage channels and two ritual baths (miqva’ot) were found on this terrace; one of the baths was located on the northern part of the terrace (Fig. 6), the other on the southern side (Fig. 7). Both miqva’ot were integrated into rooms in two large buildings, only partially excavated. The walls of these structures were built of ashlars and dressed fieldstones. The southern parts of these buildings were connected to a structure that extended south and adjoined the northern buildings of Sub-Area C1. This connection made it possible to discern a uniform and corresponding stratigraphy for all of the remains in Sub-Areas C1 and C2.
Sub-Area C3. Forty-five squares were opened (Fig. 8) somewhat remotely from the other areas, near the northern margin of the site. Several well-planned buildings were exposed (Stratum VI). Some of the eastern walls of the buildings were rock-hewn while others were constructed on bedrock or abutted bedrock outcrops. Like the structures belonging to this stratum in the other areas, these buildings were destroyed in an earthquake. The plan of wall foundations and construction quality suggest that they belonged to private dwellings.
A round (diam. c. 9 m), unique structure was unearthed in the middle of the area; it consisted of a curved wall (width 0.45 m) with a massive base built of ashlar stones; most of its circumference was built of two courses set on bedrock. Approximately 10 m of the northern circumference of the base was built on top of a solid foundation, c. 3 m deep. The center of the building was paved with a colorful mosaic. Judging by its pattern, the mosaic dates to the seventh century CE, which corresponds to Stratum VI in Sub-Areas C1 and C2 (Late Byzantine–Early Islamic period). A strip of large fieldstones bonded in mortar (width c. 0.8 m) separated the circumferential wall and the mosaic pavement. The width of this strip and the differences in elevation suggest that the mosaic was surrounded by two rows of stone seats that were robbed. This building might be explained as a landmark and assembly place of religious or historical value.
Pottery sherds from the Roman period were found in the probe that was excavated below the mosaic; finds from this period are unique to Sub-Area C3. These Roman-period potsherds raise the possibility that the mosaic was installed in a building that originally belonged to the previous stratum (Stratum VII). Evidence for this was found at the bottom of the solid, round foundation, on its northern side, where a curved wall of an earlier round building can be discerned. This supposition is reinforced by a white plaster pavement (c. 5 m-wide) visible between the round building and its surrounding structures. It seems that this earlier round building, dating to the Roman period, was free-standing. It bears no architectural similarity to the private residences surrounding it, and its unique ashlar construction is unparalleled in the any of the other excavated areas. However, like the tower in Sub-Area C1, located c. 200 m to the south, it is built slightly above a paved road that passed c. 50 m west of them (Area D).
Area D. Forty squares were opened. The upper part of the area is contiguous with Area C. Since the Byzantine–Early Islamic period, the rocky area was neither built upon nor was it cultivated; it was evidently used solely for grazing because the wall remains that protruded above the surface were not damaged prior to the excavation. A road, tombs and agricultural installations were exposed here.
The Road (width c. 7 m; Fig. 9). A section of an ancient road was uncovered by chance, during initial earthwork with a backhoe. It was located at the foot of the built-up area (Sub-Area C3). The road, running in a north–south direction, followed the topography of the western bend of the site. Its western side lay on a foundation of fieldstones in secondary use and was bounded by a stone wall; its eastern side was hewn in the slope’s bedrock; gravel fill the space between them. The road and the remains in the area help us understand the settlement plan of the site. The exposure of the road in the western part of the site explains the round monument in Sub-Area C3, which was built beside the road; the concentration of massive construction in Sub-Areas C1 and C2 alongside the road; and the tower in Sub-Area C1 that stood at the edge of the road and controlled it.
Tombs. Four tombs were found, each of which consisted of a square, rock-hewn anteroom (c. 4 × 4 m), rock-cut steps in its western side and an opening facing east. A roll-stone was found in each of the tombs, some on the floor of the anteroom and some in situ. The excavation did not continue beyond the anterooms. Pottery sherds found in the anterooms date to the Roman and Byzantine periods.
Agricultural Installations. A small winepress, a winepress complex and a quarry were exposed. The small winepress was hewn in bedrock in the upper part of the area. It included a treading floor and two small collecting vats. Some of the original plaster was preserved on its floor and walls. In the middle of the area was a large hewn winepress that was expanded into two winepresses in several phases (Fig. 10). In the early phase, a very large winepress was hewn above one of the Byzantine-period burial caves; this burial cave might have originally been used as a ritual bath in conjunction with the winepress. This winepress had a work surface with a screw base at its center, and a square collecting vat to its north; the treading floors were on its southern side. In a later phase, when the early winepress went out of use, two new winepresses were built: one built in part atop the early winepress, the other hewn to its east. The eastern winepress had a work surface that was operated with a screw. To its north was a settling pool, from which the must was gathered into two collecting vats. Hewn and built treading surfaces surrounding the work surface where the screw was located and drained into it. The western winepress had a work surface for a screw built atop of the earlier rock-hewn work surface. The earlier collecting vat was filled up, and a settling pool was built inside it. The must was collected inside a new, circular collecting vat; its construction required breaking down the eastern wall of the early collecting vat. The two later winepresses seem to have operated simultaneously, and were used to produce wine in several methods. Similar complex winepresses were found at Kefar Sirkin (Sidi, Amit and ‘Ad 2003), Tel Ifshar (Yannai 2009), El‘ad-Mazor (Amit 1998) and elsewhere (Avshalom-Gorni, Getzov and Frankel 2008). The pottery sherds found in and around the installations date them to the Late Byzantine or beginning of the Early Islamic period (seventh century CE).
Area F (Fig. 11) extended across the western part of the excavation, along the eastern bank of Nahal Shaham, at the bottom of the valley and on the rocky ridge west of the built-up area of the kibbutz. A pine forest was planted in this area. Pottery sherds from the Middle Bronze Age were found in several trial trenches prior to the excavation. Three squares were opened in places where sherds were discovered. One or two damaged tombs were found in each square; it was impossible to reconstruct their outline. Three of the tombs date to the MB IIB; a jar found in one of the tombs dates to the Intermediate Bronze Age. Remains of an MB IIB building—possibly two buildings separated by a hallway—was discovered in the western part of the area. Two building phases were discerned: the earlier building was constructed on bedrock; in the second phase, a wall, founded on bedrock as well, was built along an earlier wall in the northern room. In this part of the area was a large workshop for the knapping of sickle blades, most probably from the MB IIB. Throughout the excavated area, numerous pottery sherds dating to the Intermediate Bronze Age were found, as well as a fragment of an EB holemouth rim and a potsherd from the Pottery Neolithic A. In addition, an intact jar from the Iron Age IIB was found.
These remains suggest that the settlement in the lower part of Area F was inhabited as early as the Pottery Neolithic A period; the hearths found in Area A might have belonged to this settlement, which was probably associated with the settlements from this period along Nahal Soreq. The EB rim fragment indicates activity during this period as well, when Tel Gezer was the most important site in the region. Links between the site in Area F and Tel Gezer probably continued into the MB IIA and MB IIB, when the site seems to have been an unfortified settlement associated with the main city at Tel Gezer.
Conclusion. The excavation at Mishmar David was carried out over an extensive area, and the excavation areas were situated far apart from each other. The earliest settlement at the site dates to the Pottery Neolithic period. It was probably part of an array of settlements, to which others in the area—Teleilat Batashi, Yesodot and Hafez Hayim—belonged to. This suggests that already in early antiquity the site was integrated into the regional web of settlements and roads. The settlement was restricted to the lower western side of the site, where it was supplied with water from springs in the wadi that descended from Tel Gezer in the north toward Nahal Soreq in the south. The settlement existed, albeit not continuously, from the Neolithic period until the end of the Middle Bronze Age. The settlement’s cemetery was located on a rocky hill slightly to its south.
The nature of the settlement changed in the Roman period (Stratum VII): the lower part of the settlement was abandoned and a new settlement was constructed on the rocky spur where Kibbutz Mishmar David was subsequently established. The architectural remains of Stratum VII were found only in Sub-Area C1; it is possible that the round building in Sub-Area C3 was also built during this period.
In the Byzantine–Early Islamic period (Stratum VI), most of the area of the hill was densely built. The structures ascribed to this stratum were founded on bedrock. Some of their walls were rock-hewn while others were stone-built. This stratum was probably destroyed by an earthquake in 749 CE, after which the settlement was rebuilt (Stratum V). The remains of the previous stratum were covered over, although some of its walls were reused. The settlement expanded into new areas to the east. This settlement was apparently also destroyed by an earthquake in 1033 or 1068 CE. In the Fatimid period, retaining walls (Stratum IV) were built on top of the destroyed levels of Stratum V; above them were the remains of meager buildings and repairs that are ascribed to the Mamluk period (Stratum III). It seems that most of the settlement was deserted or gradually abandoned. On the surface, there was evidence for secondary use of walls, robbing of earlier walls and other indications of later activity in the area (Stratum II). This activity was carried out by during construction in the village of Khulda and in Kibbutz Mishmar David.
The importance of the excavation is reflected by four significant discoveries. The first is the discovery of a city or rural settlement from the Umayyad period in the Emmaus–Ramle region. The second is the discovery of two archaeological layers that were destroyed in earthquakes. It is very difficult to date precisely ceramic assemblages and other objects of material culture from the Late Byzantine period until the end of the Early Islamic period due to the long-standing conservative design tradition of the pottery vessels during these periods and the protracted use of Byzanto-Arabic coins during the Umayyad period. Thus, the identification of well-dted earthquakes, which are clearly identified as the cause of destruction of Strata VI and V, may contribute in dating material culture from in situ assemblages. The third discovery of importance is the unique, circular building in Sub-Area C3. The two complete ritual baths (miqva’ot) in Sub-Area C2 comprise the fourth important discovery, which as far as we know are the latest in date ever found in archaeological excavations in Israel. If these are indeed ritual baths, they serve as tangible proof of the existence of a non-Muslim community at the site under Umayyad rule. This has important implications, since ritual baths were in use by Jewish and Samaritan communities. Credible evidence for the existence of a Christian community at the site was also found. Thus, one may conclude that the settlement’s community was composed of a mixed population of Jews, Christians and Samaritans. The date of the baths suggests that these communities continued to live and maintain their previous way of life, which involved observing the laws of purity, even after the Umayyad conquest of Israel (Amit and Adler 2010:136). This conclusion is extremely important in our effort to learn about the process and timing of the Islamization of the residents of Israel, as there is a reasonable probability that the laws and customs associated with purity continued to be observed after the residents of Khulda converted to Islam. The two ritual baths may suggest that the residents of Khulda did not see any contradiction between upholding this ancient commandment and alongside the faith of Islam; after all, Muslim religious law does not prohibit the practice of this commandment.