The site of Shivta (c. 90 dunams; 350 m asl) is located in a semi-arid region, c. 40 km south of Be’er Sheva‘, on the southern edges of Biq‘at Qorẖa. In the center of the site are two reservoirs, a church (Southern Church) and a mosque, and in the northern part, there are two additional churches (Middle Church and Northern Church). Numerous excavations have taken place at the site, revealing dozens of densely spaced dwellings and alleyways (Baly 1935; Segal 1986; Tepper and Bar-Oz 2018). The settlement of Shivta, mentioned in the sixth-century CE Nizzana Documents (papyrus), was apparently founded in the second century CE, reaching its zenith in the Byzantine period, and was abandoned at the beginning of the Early Islamic period (Segal 1986; Shereshevski 1991; Hirschfeld 2003; Tsuk 2003; Tepper, Weissbord and Bar-Oz 2015; Tepper et. al 2018).
Ten excavation areas (A–J) were opened. In the southern part of the site, a courtyard house and the southern reservoir were excavated (Areas A–C), and in the northern part, near the Northern Church and the adjacent plaza, a dwelling and a complex of channels directing surface runoff to the settlement and to the Northern Church chapel, were uncovered (Areas D–J). The excavation documented two strata. The early and more significant stratum, exposed in all the excavation areas, is dated to the Byzantine period (fifth–seventh centuries CE). The later stratum, whose remains were found in Areas A and G, dates to the Early Islamic period (seventh–eighth centuries CE); pottery from the twentieth century CE was also retrieved in Areas A and G.
Area A. To the west of the northern reservoir, a three-story courtyard house (Fig. 1) was excavated. On the lowest story, there was a rock-hewn plastered cistern, and underground spaces and rooms. A channel built below the courtyard led into the cistern, whose ceiling was supported by a central column, one of whose corners bore some incisions. Finds retrieved in the uppermost fill layer in the cistern, attest to a temporary presence in the British Mandate period (twentieth century CE). Underground spaces were discovered southeast of the cistern. A room with two intact arches supporting the ceiling, was completely excavated to the west of the courtyard (Fig. 2). Some stone beams overlay the arches, bearing the floor of the intermediate story. In two walls of the room, entrances, sealed up with construction, were documented. In an adjacent room, an entrance in one wall, preserved from threshold to lintel, was also sealed by construction; it probably led into another room, or a complex of rooms (not yet excavated). The lintel, decorated with rosettes, was dated to the Byzantine period. Remains dating to the late Byzantine period were found in the soil accumulated behind the walled-up doorway.
On the intermediate story—the central part of the building—a partially paved courtyard surrounded by eight square rooms, was found. The floors of the rooms were made of packed soil, apart from one room that was paved with stone slabs. The meager finds retrieved on the floors attest to an orderly abandonment of the building. Pilasters bearing stone arches were found abutting the walls of the rooms; the stone voussoirs and roof slabs were found in the collapsed heaps of debris in the rooms, amongst which loess had accumulated. One of the voussoirs bore the remains of decorations, as well as red-painted crosses. Signs of fire on the walls of the rooms, and on the courtyard floor, are probably witness to a conflagration. The upper story, although hardly preserved, was identified by the remains of the pilasters that bore the upper level rooms, as well as by the height of the walls in the southeastern part of the building.
The building was constructed in the Byzantine period (fifth century CE), and was abandoned within, or at the end of that period. In the Early Islamic period, the floors were raised, installations were built, and the entrances into the rooms surrounding the courtyard were walled up. A trash pile from the Early Islamic period, excavated in one of the rooms, attests that the room was now not used for habitation. The finds reveal that the building was finally abandoned in the eighth or early ninth century CE. Slightly thereafter, the stone ceilings collapsed in some of the rooms. It seems that the cistern was partly renovated in modern times, serving nomadic populations who camped at the site.
Pottery sherds and glass fragments from the Byzantine and Early Islamic periods were retrieved in the building, as well as scraps of metal items, coins, a Latin metal seal (not yet deciphered), stone items adorned with crosses, stone vessels, including bowls, and grinding and milling implements, botanical remains and animal bones. Soil samples were retrieved for wet and dry sifting, in order to complete the previous season’s research of this building, and to try to date more precisely the chronological phases in its history (radiocarbon and OSL dating).
Area B. The southern reservoir was excavated. In the 1930s, the Colt expedition excavated the southern reservoir to a depth of 2.0–2.5 m, revealing two staircases, a northern and a southern one (Baly 1935). In the trial sections dug in the 2017 season, a floor was uncovered next to the northern staircase, and a foundation layer near the southern staircase (Tepper et al. 2018). In the current excavation, a loess layer accumulated in the upper part of the reservoir, was removed mechanically. On the eastern side, this layer was removed down to an underlying layer of collapsed debris, and on the southern side, the layer was removed almost down to the reservoir floor, to enable conservation. In the section dug next to the southern staircase from the top of the debris layer down to the floor, a wall and the staircase were exposed to a depth of c. 5 m. Another section was dug in the southwestern corner of the reservoir, exposing the floor at a similar depth. Small stones had accumulated on the floor, and loess layers in the upper part, with intervening soil layers containing much gray mortar. In both the sections dug in the northwestern part of the reservoir, a crooked wall was exposed, built from west to east across the reservoir, dividing it into two levels, an upper and lower level. The northern staircase was discovered in the wall, descending to the lower level of the reservoir floor, at a depth of c. 5 m. Metal tools dating to the early twentieth century were found on the steps, evidently from the time of the Colt expedition. Five squares were opened along the southern and western walls of the reservoir, and in its center, where a settling basin was unearthed (2.5 × 4.5 m; depth 1 m; Fig. 3). The reservoir walls, floor and settling basin were coated with gray plaster on a foundation of small stones and gray mortar. The meager finds retrieved in the collapsed piles and soil accumulations in the excavation squares, included pottery sherds and coins from the Byzantine period (dating at the latest, to the seventh century CE). The study of the accumulations and sediments collected from the sections, and their precise radiocarbon and OSL dating, will help understand the history of the reservoir’s destruction.
Area C. In the 2017 season, a section was dug in the southwestern corner of the northern reservoir, exposing a plastered floor at a depth of 3.5 m. In this season, a section was dug alongside the reservoir’s northern staircase, revealing that northwestern part of the reservoir wall is modern, and was apparently built in the course of conservation work undertaken in the past. The excavation of the northern reservoir was stopped due to technical limitations.
Area D. A modern (c. 30-year old) layer covering up the mosaic floor between the baptistery and the nave in the Northern Church chapel, was removed. The mosaic, excavated by the Colt expedition in the 1930s, exhibits multi-colored geometric patterns characteristic of the Byzantine period, and bears a Greek inscription dated to the end of the Byzantine period (Segal 1986; Fig. 4). Probes dug in gaps in the central panel of the mosaic, and to the west of the chancel screen, revealed that the mosaic floor predates the construction of the chancel screen and the apse, and that the mosaic was originally set in a larger building that was later reduced in size. Conservation work accompanying the exposure of the mosaic, focused on stabilizing and protecting the remains. The pottery and coins retrieved dated to the Byzantine period. Samples of organic material from the mosaic makeup were collected for radiocarbon dating, which may pinpoint the construction date of the church, or the renovation of the floor.
Area E. Adjacent and to the south of the Northern Church, a primary water channel, which probably collected surface runoff from the slopes northeast of the site and conveyed it to cisterns in the vicinity of the church, was excavated (Fig. 5). The channel, built of stone slabs, led to a square, plastered distribution pool with two channel openings in its walls. One channel directed water into the cistern behind the baptistery in the southern wing of the church; the other led off to the southwest, possibly to a public cistern in the northern plaza, or to another unlocated cistern. A plastered pool found alongside the cistern behind the baptistry, distributed the water into secondary channels, and the water drawn from the cistern flowed through a hole in the church wall into the baptistery basin. South of the distribution pool, part of another channel was excavated, and next to it, a stone lintel bearing an incised cross, with two valves that served to regulate the flow of the surface runoff, was discovered. Flanking the main channel were additional built and plastered secondary channels, that conveyed water through a series of amphora jar necks in secondary use. The amphorae date to the late Byzantine period (sixth century CE). Pottery sherds retrieved in the soil fill flanking both sides of the channel also date to this period. These finds indicate that a garden was probably cultivated in the area between the church and Insula 1, the complex of structures that was partially excavated by the Colt expedition. In the past, it was proposed that this area was part of a monastery located in the northern part of the site (Hirschfeld 2003). Soil samples retrieved from the sections cut in the channels will enable dating the water system by OSL. Additional samples were sent for pollen analysis to identify the plant species grown in this plot.
Area F. To the east of the Northern Church, a primary wide earthen water channel that conveyed water to the site center, to the area of the Central Church and the Governor’s House, was excavated (Fig. 6). In a previous excavation in this area, a distribution installation exposed at the end of the earthen channel, led surface runoff from the Shivta slope into the settlement. It was proposed that a central valve branching off the water to the Northern Church on the one hand, and to the reservoirs in the southern part of the site on the other (Tsuk 2003). In the current excavation, a double valve system was discovered that may possibly have also channeled water to the Northern Church, although a channel extending northwards from the valve has not yet been found. The excavations have revealed the complexity of the surface runoff channel system providing water to the site, and the significant quantities of water that were collected from the slopes. Soil samples were taken from the channel in order to date the period of its use by precision method (OSL).
Area G. Two probes were excavated in two adjacent rooms in the southern part of Insula 1, one south of the Northern Church (Segal 1986), and the other southwest of the winepress and near the cistern. The northern probe was dug in a room with a doorway intentionally walled-up with ashlars and pilaster arch stones, supported from behind by medium and small-sized stones set in a semicircle (Fig. 7). The southern probe was dug in a room with a doorway that had not been walled up. The aim of the probes was to analyze examine the chronological and stratigraphic differences between the rooms by examining the accumulations in them. Soil accumulations in both rooms were documented from the surface down to the floor, as well as below the floor, down to bedrock. In both rooms, twentieth century pottery sherds were found in the surface layer, attesting to a temporary presence, presumably associated with the nearby cistern. The fills under the floor in both rooms were dated to the Byzantine period (the fifth–sixth centuries CE), evidence that the entire building complex was built at this time. The soil layers in the walled up rooms contained finds from the Byzantine–Early Islamic periods, and were identified as accumulations reflecting habitation in this period. The soil accumulations in the unsealed room contained finds dating from the beginning of the Early Islamic period to the modern era, hence indicating that they were not stratified in a clean archaeological context. Soil samples were retrieved from the probes in both rooms, above and below the floors, with the aim of dating the construction, and the stratification by precision methods (radiocarbon and OSL dating).
Area H. A mosaic floor was previously documented in the northeastern courtyard of Insula 2, west of the Northern Church (Segal 1986; Fig. 8). The entrance into the courtyard was from the northern plaza, in the eastern part of which there were steps, and in the northwestern part, there were entrances into two adjacent rooms. The mosaic floor was made of large white tesserae (2 × 2 cm) set on a stone makeup. In the western part of the courtyard, a probe was dug down to bedrock, revealing the mosaic makeup, which contained debris from the mosaic industry, including glass fragments probably from the preparation of the mosaic. Finds from the Byzantine period were discovered under the courtyard floor. The function of the courtyard is not clear. Conservation work was undertaken on the mosaic immediately upon its exposure, focusing on stabilizing and protecting the remains.
Area I. An excavation square was opened south of the northern plaza and east of the Central Church, where there was a natural depression. The depression extended eastwards along the line of a presumed channel directing floodwaters from the Shivta slope to the main channel in the settlement (see Area F, above). The aim of the excavation was to trace the continuation of the channel in the public domain. A built and partially roofed channel extending westwards (Fig. 9) toward the structures adjacent to the Central Church and the Governor’s House, was exposed. This finding does not support the theory that the floodwaters from the Shivta slope were channeled into the reservoirs in the southern part of the site (Tsuk 2003). Soil samples were taken from the section in the channel to date them by precision method (OSL).
Area J. Two probes were excavated down to bedrock in the Northern Church. The eastern probe was excavated in the interior corner of the narthex, in the foundations of the original walls, and the western probe was dug in the western doorway of the church, next to the surrounding sloping retaining walls (Fig. 10). The aim was to examine the structure of the walls and the supports abutting the exterior, that were built in the later construction phase, as a basis for planning the conservation of the church. It aspired that the original church walls were set on bedrock, and that the exterior surrounded sloping wall was set on fill. Excavation of the eastern section revealed fragments of plaster bearing red paint. This finds in the foundation and fill layers in both probes dated to the Byzantine period. Samples of organic material taken from the stratified fill below the retaining wall, will be dated by radiocarbon precision method to help date the construction of the wall.
The 2018–2019 excavation seasons at Shivta contributed considerably to an understanding of the site and its history. Most of the architectural remains uncovered were dated at the earliest to the Byzantine period; some were dated to the Early Islamic period, and no later architectural finds were discovered. The architectural-stratigraphic analysis of the Area A building, and the chronological data retrieved therein, will enable outlining the settlement sequence in the center of the site around the reservoirs. Moreover, the Area A excavation contributed to an understanding of the extent of the settlement at Shivta, and the stages of its abandonment in the transition from the Byzantine to the Early Islamic period. It is possible that some stages in the destruction of the building, and perhaps also of other public buildings excavated at the site, were affected by a seismic event or events.
The excavations in and around the Northern Church, and those in Insulas 1 and 2, are the first carried out since Colt’s excavation that left many unresolved questions regarding the settlement sequence at the site, its development and the identification of the ruins. The ongoing research, and the laboratory analysis of the biological finds, will add data regarding the economic and agricultural activity around the church. The dating of the accumulated sediments will enable dating the destruction and abandonment of the buildings, and will thus contribute to an understanding of the development phases and the decline of the northern part of the settlement. The channel complex uncovered behind the church is unique, attesting to a high level of technology, hydrological planning and expedient use of water for agriculture within the site. The ongoing botanical-archaeological research may shed light on the nature of the agriculture, which was probably carried out under the auspices of the church or the monastery.
The southern reservoir, excavated in an attempt to understand the surface runoff collection, to and within the site, revealed that its depth and volume were larger than previously estimated. The excavation also documented the construction techniques and functioning of Shivta reservoirs, whilst no evidence indicating that these installations were roofed was found. The excavation along the water channels in the northern area of the site, provided new information on the functioning of the surface runoff collecting system that conveyed water from around the settlement to its center, and the way rainwater was diverted and stored. It seems that surface runoff was collected from the slopes surrounding the settlement in open dug earthen channels, and conveyed to built and roofed channels within the settlement. The water was directed to public cisterns via settling pools and a system of valves installed along the channels. These systems attest to water management at Shivta controlled by a public administrative apparatus, apparently ecclesiastical. They also attest to the large quantities of water that were stored underground within the settlement. The excavation has not so far revealed a connection between the water systems in the northern part of the site and the open reservoirs in the southern part.