Abstract
Between 2004 and 2008, a wealth of ancient agriculture-related finds was uncovered during a survey throughout the loess plains of the Beer Sheva Valley region, in the arid northern Negev, Israel. The survey was conducted under the framework of an archeological study, aimed at assessing the similarities and dissimilarities of the valley’s ancient agricultural systems to those of the Negev Highlands to the south and of the Judean Lowlands and Southern Hebron Mountains to the north. Data collection from selected sites included detailed mapping of settlements and their hinterlands. Ancient runoff farming systems, comprised of relatively uniform stone terraces transecting the wadis (ephemeral stream channels), and other agriculture-related structures, such as tuleilat el anab (spatially patterned stone mounds erected on hillslopes), were revealed throughout the region. Other archeological finds included a variety of structures, including livestock pens, square watchtowers, rock-cut water cisterns, and others. This study indicates that like the agricultural systems in the neighboring southern and northern regions, the systematic terracing of wadis across the Beer Sheva Valley region was affiliated with the monastic settlement system, which was centrally managed by the church in the service of the Byzantine Empire. The significance of this settlement system stems from its highly capable central organization, aiming to achieve geo-political control of space. Despite peaking in the mid-6th century CE, this system persisted throughout the Early Islamic period, until the mid-8th century CE. Results of this study contradict the conclusions of previous studies, which negated the viability of ancient runoff farming across the loess plains of the Beer Sheva Valley region. Insights of this research highlight the need for interdisciplinary studies when assessing the interactions between human populations and the natural environment in ancient times.
Keywords
Introduction
An archeological mapping project of the Beer Sheva Valley (central datum point: 31°14′ N, 34°50′ E; mean elevation: ~280 m a.s.l.; Figure 1) and its surrounding loess landscape, southern Israel, took place between 2004 and 2008. This article reports on the wealth of ancient runoff harvesting systems and related agricultural installations revealed during the survey, which were located at the margins of settlements dated to the 6th–8th century CE (Haiman, 2012: 43–45, 2018). The finds include systematically terraced wadis (ephemeral stream channels), comparable to those of the ancient runoff farming systems prevalent south of the Beer Sheva Valley – across the Negev Highlands – as well as in the northern Judean Lowlands and Southern Hebron Mountains. These terraces – as studied and classified (as types 1 and 2) by Evenari et al. (1971: 97–117) during the 1950s and 1960s – played a significant agro-hydrological role by tremendously increasing water provision for the agricultural crops.

A map of Israel, with the Beer Sheva Valley marked in gray.
Apart from terraces, the tuleilat el anab (vine heaps in Arabic) systems are the most abundant. These structures, constructed on hillslopes, are comprised of spatially patterned stone mounds ~1 m in diameter and ~0.5 m high. It has been suggested that these systems aimed to increase (up to sevenfold) the runoff ratio into the downslope terraced wadis (Bruins et al., 2019; Evenari et al., 1971: 127–147), facilitate soil erosion in hillslopes (while increasing its deposition in the downslope wadis; Kedar, 1967: 45–54), or increase dew condensation, thus increasing water availability for vines planted at the mounds’ center (Kogan and Trahtman, 2003). Our survey revealed that these systems are common throughout the entire region, near wine presses and square watchtowers.
In the Negev Highlands, grapevines were a significant crop during the Byzantine period (324–638 CE), allowing the wine industry to play a central role in the Empire’s economy (Mayerson, 1992). During the Umayyad period (639–750 CE), grains became the most significant crop across the region (Bruins, 1986). Indeed, a growing body of evidence from the Negev Highlands suggests that runoff farming systems enabled significant production of cereals, legumes, grapes, and fruit trees during the late antiquity (Bar-Oz et al., 2019; Dunseth et al., 2019; Evenari et al., 1971; Fuks et al., 2016; Kedar, 1967; Mayerson, 1962). Yet, a clear distinction should be made between small-scale agriculture for domestic consumption, and the large-scale wine industry that existed mostly during the 6th century CE. Indications of this industry include the many wine presses across the Negev and further to the north, the extensive production of standard wine jars in the north-western Negev (Israel, 1993), and wine export from the ports of Gaza and Ashkelon (Mayerson, 1985, 1992).
Our main interest in these agricultural systems stems from their similarity to those surrounding the ancient towns of Avdat (30°47′37″ N 34°46′26″ E), Shivta (30°52′52″ N 34°37′46″ E), and Nessana (30°52′31″ N 34°25′57″ E), which have been dated to the 5th–7th centuries CE (Haiman, 2018). Some towns, for instance, Avdat, have earlier histories as desert caravanserais between the late 2nd century BCE and the mid-3rd century CE (Negev, 1988, 1997), but lack evidence for agricultural activity during this early period (Erickson-Gini, 2010, 2012). This topic was initially raised in the 1960s (Ilan, 1980) and further studied in emergency excavations during the late 1980s.
Most early studies of the agricultural systems surrounding Byzantine towns did not consider the interactions between the historical background, natural environment, and archeological artifacts (e.g. Negev, 1988, 1997). Furthermore, even the pioneer study by Evenari et al. (1971) on ancient desert agriculture in these towns – which focused on climatic, hydrologic, and geomorphic features of the agricultural systems – lacks research and insights on the interrelations between the agricultural structures and the nearby settlements. Specifically, Evenari and colleagues attributed advancements in ancient agriculture to ‘natural’ human adaptation to the desert environment and ignored possible influences of the Byzantine Empire. A significant influence of the Empire and the church is indicated by the establishment of settlements, which were colonized by mixed populations arriving from other regions across the empire, such as Egypt and Syria (Di Segni, 2004; Mayerson, 1986; Rubin, 1990; Stroumsa, 2008). Yet, the pioneer study by Evenari and colleagues included detailed mapping of archeological artifacts in the towns’ agricultural hinterlands and yielded early publications simultaneously with fieldwork implementation (Aharoni et al., 1960; Evenari et al., 1958).
Later studies that focused on archeological aspects of the agricultural systems also ignored the interdisciplinary approach. Some studies chronologized farmhouses, shedding light on the age of the agricultural systems themselves. For example, Nevo (1991) identified two types of farmhouses in the Avdat hinterlands. The first is a ~5 m × 5 m square watchtower made of large stones, with a low, arc-supported ceiling. This house type was dated to the late 5th century CE or early 6th century CE, and therefore, was no longer being constructed before the end of the Byzantine period (early 7th century CE). The second type is a more complex farmhouse with dwelling rooms and open courts. Its architecture is considerably different, and dates to the Umayyad period. These two types of farmhouses have been identified across the central Negev and Beer Sheva Valley (Haiman, 2018). The first type is commonly found surroundings towns and monasteries, and the second reflects the distribution of agricultural lands throughout the region, regardless of towns and monasteries (Haiman, 2020).
Regarding the chronological context of settlements, an early stage of the local church existed during the 4th and 5th centuries CE and was prevalent throughout the Judean Desert (central datum point: 31°55′ N 35°35′ E; Hirschfeld, 2002; Patrich, 1995) and Gaza (central datum point: 31°85′ N, 34°75′ E) (Bitton-Ashkelony, 2010; Di Segni, 2004; Hirschfeld, 2004). The agricultural lands surrounding these monasteries were based on simple terraces constructed in large wadis, mildly intervening with the natural environment. In the mid-6th century CE, agro-hydrological technologies considerably progressed (Aharoni et al., 1960; Evenari et al., 1958; Haiman, 2020) and massive terraces were established in wadis of various sizes. Based on the structure of the nearby complex farmhouses, it seems that these systems represent the Umayyad’s agricultural settlements, which thrived until the mid-8th century CE (Haiman, 2012; Nevo, 1991).
The runoff farming systems across the loess plains of the Beer Sheva Valley and its vicinities are attributed to the same archetype as those of the Negev Byzantine towns, based on the structural and morphological similarity they exhibit. No similarities between these systems and settlements of other historical periods were noted (Haiman, 2012). Attempts to date these systems by optically stimulated luminescence (OSL: Avni et al., 2013) or radiocarbon (14C: Bruins et al., 2012) revealed a wide range of results. These attempts were severely criticized (Gibson, 2015; Shahack-Gross and Finkelstein, 2017), highlighting the need for a more comprehensive approach to correctly date these structures (Haiman, 2019: 191).
Regardless, during the late 1980s, various research groups studied the potential for runoff farming in the Beer Sheva Valley (Lavee et al., 1997; Yair, 1987, 1994). However, these studies did not thoroughly evaluate the relations between the archeological artifacts and the historical background, and negated the potential for runoff farming throughout the region.
The objective of our research was, therefore, to assess the ancient runoff farming systems and related agricultural structures across the loess plains of the Beer Sheva Valley region using an interdisciplinary approach. It was expected that extensive assessment of the archeological artifacts may yield new insights regarding the ancient use of these farming systems in this region.
Methodology
Regional settings
In this project, we focused on agricultural systems surrounding ancient settlements across the Beer Sheva Valley and its nearby loess region. The valley, ~300 km2 in size, is located between the Southern Hebron Mountains to the north and the Negev Highlands to the south, and between the Arad Valley to the east and the Western Negev to the west (Figure 2). The climate of the valley is arid, with average annual precipitations (P) of ~200 mm (Israel Meteorological Service website), which varies interannually between 180 and 230 mm (Kafle and Bruins, 2009). The potential evapotranspiration (PET) is ~1,900 mm (Israel Meteorological Service website), yielding an aridity index (P/PET) between 0.10 and 0.12, and classifying the region as arid (see Cherlet et al., 2018). The mean daily temperature in the coldest (January) and warmest (July) months are 12.4°C and 28.0°C (Israel Meteorological Service website). The lithology is dominated by chalk of the Eocene and alluvium (Hofstetter et al., 2013). The distinctive topography is undulating to flat landforms, transected by wadis of different orders. The predominant soil is composed of Pleistocene aeolian loess deposits, attributed to the calcic xerosols group. In many places, the loess – with sandy loam to silty clay loam texture – has been re-deposited by floods (Sneh, 1982; Yaalon and Dan, 1974). The loess is characterized by moderate infiltration capacity, available water capacity of ~20%, limited swelling capacity, and good aeration, and its thickness is highly variable and depends on the local topography (Singer, 2007).

A map of the Israeli Negev, with indications of sites mentioned along the article: 1 – Kh. Baqar; 2 – Kh. Karkor; 3 – Nahal Ashan A; 4 – Nahal Ashan B; 5 – Goral Hills; 6 – Beer Sheva Ridge C; 7 – Beer Sheva Ridge B; 8 – Kh. Abu Omaria; 9 – Nahal Kovshim; 10 – Beer Sheva Ridge A; 11 – Kh. Abu-Mahfudh; 12 – Kh. Ytan; 13 – Tel Masos.
Mapping of ancient agricultural systems and data collection
The mapping methodology was described in detail by Haiman (2012, 2018, 2019). The major methodological components included (1) identifying key sites publicized in the Archeological Survey of Israel (http://survey.antiquities.org.il/index_Eng.html#/). For most of these sites, some information is provided for the settlements themselves, while only partial (if any) information is provided for their surroundings; (2) detailed field mapping of selected sites. In these sites, the settlements and their surroundings were mapped, with an emphasis on runoff farming and other agricultural structures across the settlements’ hinterlands. The mapping was conducted using a GPS (Trimble®, GeoExplorer 3). The collected data were uploaded to Arc Map 10.3.1 GIS (ESRI®) software for producing an ‘archeological layer’; (3) for each site, the archeological finds were classified, and the relations between artifacts from settlements and their surroundings were assessed; and finally, (4) these data were used to generate interdisciplinary knowledge and insights.
Results
Artifacts discovered during the past half-century
Since the 1960s, extensive excavations revealed ancient agricultural systems throughout the Beer Sheva Valley. Among other artifacts, the excavations revealed a Byzantine church (31°15′10.39″ N, 34°47′27.30″ E) that functioned until the Umayyad period (Israeli, 1967), farmhouses, and rock-cut graves (31°15′31.98″ N, 34°48′29.88″ E: Cohen, 1969), as well as many terraced wadis. Since the late 1980s, comprehension of the documented agricultural systems, dated to the 6th–8th centuries CE, significantly improved. Here, we describe selected data obtained over the past half-century, which have been utilized to determine the chronology of the artifacts surveyed by the authors of this paper.
Beer Sheva Ridge A (Figure 2): three structures, 150–250 m from each other, separated by agricultural terraced wadis (31°15′59.76″ N, 34°48′48.15″ E: Sontag, 2012). One of the structures was identified as a 17 m × 27 m farmhouse with five rooms, a yard, and a water cistern. The second structure is similar to the square watchtower prototype that was prevalent throughout the Negev during the Byzantine period, aimed at protecting the agricultural fields. Pottery fragments discovered in this site were dated to the late Byzantine period.
Nahal Kovshim (Figure 2): six structures along Nahal Kovshim (31°16′26.30″ N, 34°48′23.46″ E: Ustinova and Nahshoni, 1994), of which one is quadrangular with a low arcs-supported ceiling, and the rest contain rooms and yards. All structures were dated to the 6th–7th century CE. The first is similar to structures of the Negev’s late Byzantine towns. The other structures were built in the style of the abovementioned Umayyad period farmhouses. Pottery fragments located onsite indicate its existence until the mid-8th century CE.
Goral Hills (Figure 2): two square watchtowers next to a terraced wadi (31°16′48.64″ N, 34°48′25.77″ E: Fabian and Goldfus, 2004). The terraces were dated to the late Byzantine period. An extensive tuleilat el anab system covering an area of 150 m × 450 m in a nearby hillslope was uncovered.
Beer Sheva Ridge B (Figure 2): a complex of structures (31°15′35.35″ N, 34°48′75.55″ E: Israel et al., 2013). Two structures, dated to the late 6th century CE and early 7th century CE, and located ~120 m from each other, are worth noting. The first, 20 m × 30 m, has outstanding architectural properties, including an arc-supported ceiling, tablet-paved floors, and a central kitchen. In this structure, a jar fragment with Greek script, which is characteristic of the Byzantine period, was found (see Figueras, 2004a). In the second structure, 25 m × 25 m, a kitchen and stoves were found. Both structures were identified by the excavators as farmhouses. Yet, according to the characteristic kitchens, it is plausible that the structures were utilized as a monastery (see Hirschfeld, 2002) whose residents managed the surrounding agricultural systems. An additional structure, 8 m × 20 m, was suggested to be an Early Islamic farmhouse. The structure is architecturally similar to the farmhouses dated to the 7th–8th centuries CE that are common throughout the Negev. Next to this complex, another 10 m × 20 m farmhouse (31°15′42.22″ N, 34°48′45.70″ E), encompassing a water cistern and a granary and dated to the late Byzantine period or Early Islamic period, was probably in use until the mid-8th century CE (Nikolsky, 2014).
Nahal Ashan A (Figure 2): a 20 m × 25 m structure, encompassing rooms and a courtyard (31°16′38.35″ N, 34°45′12.42″ E), dated to the 8th century CE (Eisenberg-Degen and Kobrin, 2016). The architectural properties of this structure are identical to those of many farmhouses built throughout the Negev during this period.
Finds obtained in the recent mapping project
Beer Sheva Ridge C (Figures 2 and 3): a terraced section of the wadi, surrounded by a 0.5-m high stone wall (31°16′20.29″ N, 34°48′43.01″ E: Figure 3). The stone terraces are multilayered (composed of 2–5 stone layers). The surrounding hillslopes are covered by a tuleilat el anab system.

Beer Sheva Ridge C: 1 – terraced wadi; 2 – surrounding wall; 3 – tuleilat el anab.
Khirbat Abu Omaria (Figures 2 and 4): a complex of structures, extends throughout a 250 m × 300 m expanse (31°16′15.47″ N, 34°50′12.49″ E: Figure 4). Among the structures, a church with a southern-oriented apse, unlike the common eastern-oriented apses (Tahal, 1998), was recorded. An apse with a similar orientation was observed in a church in Khirbat Yattir (31°21′10.77″ N, 35°00′59.61″ E: Bordowich, 2007). The latter sheds light on many structures that were discovered across the Negev, whose purpose was previously misinterpreted. These structures included elements characteristic of churches, such as high-quality constructions, pillar fragments, mosaic stones, and marble tablets, but their orientation ranged between NE and SE (45–135°) (Meshel et al., 1987), negating their identification as churches. Specifically, the found chapels imply that many structures across the Negev, which were previously characterized as estate houses, were actually used as monasteries. The K. Omaria’s complex includes three structure-aggregates, as well as several stand-alone structures, and two wells. Unfortunately, modern construction and forestry activities have destroyed extensive areas surrounding the complex, demolishing many terraces in the nearby wadis (these structures were therefore omitted from Figure 4).

Kh. Abu Omaria: 1 – church; 2, 3 – wells; the dashed lines represent clusters of structures in various states of preservation.
Khirbat Abu-Mahfudh (Figures 2 and 5): a complex of many structures (31°14′46.73″ N, 34°50′11.40″ E: Figure 5). Remnants of characteristic Byzantine walls and floors were found in one of these structures, as well as clay jar fragments with Greek script (Israel, 2008). Another structure – assumed to be a monastery – includes seven rooms, a yard, and a rock-cut water cistern. The yard was identified as a livestock pen. Other structures, identified as farmhouses, were dated to the Early Islamic period according to their characteristic architectural properties. Thirteen agricultural terraces, either uni- or multilayered, were mapped in a nearby wadi. Additional rock-cut water cisterns were found in the wadi’s banks. A corner of a large wine press, containing mosaic stones and pieces of plastered stone-walls, was located next to the wadi.

Kh. Abu-Mahfudh: 1, 2, 3, 8, 9, 10, 16, 17, 18 – structures in various states of preservation; 4, 13, 15 – terrace walls; 5 – terraced wadi; 6 – assumed monastery; 7 – livestock pen and a water cistern in the north-east corner; 11, 14 – water cisterns; 12 – corner of a large wine press.
Khirbat Ytan (Figures 2 and 6): a complex encompassing several structures (31°14′90.08″ N, 34°51′55.85″ E: Figure 6), among them, a 12 m × 18 m structure with several rooms and yards. Like many monasteries found throughout the northern Negev, the structure was constructed with large chiseled stones, suggesting it functioned as a monastery. A 3 m × 10 m structure with three rooms was also discovered. Walls of this structure were constructed with double-layered small stones, characteristic of farmhouses of the Early Islamic period. Three livestock pens, spread over a total area of 2450 m2, as well as a water cistern, were also discovered.

Kh. Ytan: 1 – assumed monastery; 2 – livestock pens; 3 – a recent Bedouin compound consists of a tent’s contour and a livestock pen; 4 – water cistern; 5 –terraced wadi; 6 – Early Islamic (assumed) farmhouse.
Nahal Ashan B (Figures 2 and 7): a complex agricultural system (31°17′38.89″ N, 34°46′30.15″ E: Figure 7), which includes terraced wadis in a total area of 0.15 km2. Part of this area is surrounded by a low stone wall. An extensive tuleilat el anab system covers an area of ~3000 m2. In addition, the complex includes four square watchtowers and two rock-cut water cisterns. Similar complexes are characteristic of the fringes of Byzantine towns in the Negev Highlands.

Nahal Ashan B: main features across the site’s area.
Khirbat Baqar (Figures 2 and 8): various structures within a 3 km2 land unit (31°21′46.08″ N, 34°45′13.23″ E: Figure 8). This complex includes a church with pillar fragments and a mosaic floor, and an aggregate of structures, some of which are constructed in a style characteristic of the Early Islamic period. A rock-cut, roofed water cistern was located at the footslope next to this aggregate. In addition, a 20 m × 25 m bi-stage structure was discovered (Paran, 2009) and dated to the 6th–7th century CE. The first stage of this structure encompasses a 5 m × 7 m watchtower, which was constructed of large stones. In the second stage, simpler rooms and yards were added. A 6th–7th century CE dated monastery, 33 m × 33 m in size, with a mosaic-floor chapel, an industrial wine press, three rooms, and a yard, was also discovered. Tens of agricultural terraces and several square (~5 m × 5 m) towers characteristic of the Byzantine period were located in the nearby wadis. A low stone wall surrounds some of the terraced lands. Nearby hillslopes, with an area of 0.03 km2, are covered by a tuleilat el anab system.

Kh. Baqar: 1 – assumed boundary of the ruin. The red cross represents a church; 2, 3, 6, 10 – terraced wadis; 4 – wadi bed; 5 – ruin; 7 – simple terraces along a wadi; 8 – wall marking part of the wadi; 9 – four groups of tuleilat el anab; 11 – excavated watchtower; 12 – excavated monastery.
Discussion
This study focuses on runoff farming throughout the loess plains of the Beer Sheva Valley region, located between the warmer and drier (arid) Negev Highlands to the south and the relatively temperate and moister (semi-arid) Hebron Mountains and Judean Lowlands to the north. The settlements and associated agricultural systems uncovered during this study are structurally similar to those of the Byzantine and Early Islamic period that were previously documented in these regions. Results of our study negate the conclusions of some earlier studies (e.g. Lavee et al., 1997; Yair, 1987, 1994), which ignored the archeological finds of the loess plains of the Beer Sheva Valley, and concluded that runoff farming in this region was inviable. This emphasizes the need for an interdisciplinary approach when studying such complex themes.
This study demonstrates that archeological methodologies that focus only on settlements have a limited contribution to the comprehension of human–environment interactions in ancient times. Moreover, detailed mapping of dwellings and habitats, churches and monasteries, and even watchtowers and wine presses, while ignoring the surrounding terraced wadis and other agricultural systems in nearby hillslopes, misses the ‘big picture’. Thorough surveying of all archeological finds in the target land units, while using an interdisciplinary approach, yields new insights regarding these interrelations. In this study, we thoroughly recorded all the discovered artifacts in the above detailed sites across the Beer Sheva Valley and its vicinities, aiming at shedding light on the interactions between the ancient human populations and the natural environment.
The aim of this manuscript was to emphasize the crucial importance of Byzantine runoff farming across the Beer Sheva Valley and its vicinity. Over the years, despite the wealth of farming-related artifacts, the area’s agricultural systems were not thoroughly researched compared with areas such as the Negev Highlands to the south (Rubin, 1990; Shereshevski, 1991), the Besor region to the west (Kofsky and Bitton-Ashkelony, 2004), and the Judean Lowlands and the Samaria region to the north (Dayan, 2015; Taxel, 2008, 2013). These studies emphasized the central role played by the monasteries in leading and managing the agricultural domain. Yet, some of the studies identified monasteries as estate houses, such as in the case of western Samaria (32°60′37.03″ N, 35°01′42.54″ E: Shadman, 2016), where the many uniform agricultural systems in their vicinities (terraces, watchtowers, and winepresses: Haiman, 2012) were ignored. Above all, the elements characteristic of churches found in these structures suggest they were not estate houses.
During surveys and before implementing detailed excavations, monasteries are typically identified based on outstanding architectural properties, such as hewn stones, stone mosaic floors, marble fragments, evidence of chancel screens, and pillar fragments (Hirschfeld, 2002). Sometimes, a structure is only correctly identified as a monastery after the excavations are complete. For example, the Tel Masos monastery (31°12′51.15″ N, 34°57′57.27″ E: Figure 2) in center of the Beer Sheva Valley was identified as a monastery only during the final stages of its excavation (Fritz and Kempinski, 1983). Similarly, a structure in K. Baqar was originally defined as a farmhouse, while its excavation – revealing a mosaic-floored chapel and a large, industrial-scale wine press – enabled its classification as a monastery (Paran, 2009). The hinterlands of these two example sites demonstrate the agricultural systems that were previously mentioned. These sites and others enabled us to identify some structures in the Beer Sheva Valley as monasteries. For example, structures in the western Beer Sheva Ridge, K. Abu-Mahfudh, and K. Ytan, which were not identified as monasteries in previous surveys are now confirmed as such. The importance of correct identification of structures as monasteries rather than as estate houses or farmhouses, stems from their socio-cultural and geo-political significance. The populations of the Byzantine Empire settlements congregated from different places and represent a society that was highly capable of organizing and control its surroundings, an ability reflected by the uniformity of the agricultural systems.
Overall, it seems that settlement of the southern Byzantine Province of Palestine took place in two stages. The earlier, which is less relevant for our study, started in the 4th century CE in the Judean Desert (Bitton-Ashkelony, 2010; Hirschfeld, 2002; Patrich, 1995), in the Gaza region (Hirschfeld, 2004; Kofsky and Bitton-Ashkelony, 2004), in some of the Negev Highlands’ towns (Tepper et al., 2018), and in the Beer Sheva Valley (Khirbat Karkor ’Illit; 31°19′21.83″ N, 34°45′26.32″ E: Figueras, 2004b; Figure 2). It is evident that the region was spontaneously populated by immigrants from Egypt, Syria, and other places across the Byzantine Empire, while only a small portion of the population was of native origin. This early stage of settlement reflects relatively modest adaptation of the natural environment, and relied on agriculture in flood-prone wadis, where simple terraces were constructed to harvest runoff.
The later stage was reflected by a tremendous increase in number of monasteries and churches throughout the provinces of Palestine, centrally organized terracing of wadis, building of watchtowers and industrial wine presses, and the standard manufacturing of wine jars. Such jars were found in sites distanced 20–30 km from Gaza (Israel, 1993), which accord with historical sources that indicate the extensive wine export from the ports of Ashkelon (31°39′49.24″ N, 34°32′51.42″ E) and Gaza (Mayerson, 1985, 1992). This stage is dated to the mid-6th century CE and coincides with the reforms imposed by Justinian the Great (527–565 CE), who called for reorganization of the Palestine provinces. Under this framework, the Gaza Monastery adopted the Chalcedonian Logic and in return, won the right to join the Orthodox leadership in the Holy Land (Bitton-Ashkelony and Kofsky, 2000). This reunion of monasteries explains the sharp increase in settlements throughout the southern Byzantine province of Palestine, as opposed to the early stage, where settling was sparse and sporadic. It has been suggested that the main reason churches and monasteries, alongside flourishing agricultural systems, were established during this period was to convert the local populations to Christianity (Bar, 2005). A more plausible explanation was the motivation of the Byzantine Empire to control the inter-settlement spaces (Bucking, 2017; Ilan and Amit, 1989).
Alongside this organized settlement system, two additional groups settled throughout the wider region. The first was a group of families or small tribal groups, which inhabited villages in the Yattir area, such as Khirbat Lower Anim (31°21′10.08″ N, 35°03′50.15″ E), Khirbat Upper Anim (31°21′35.94″ N, 35°04.44.07″ E), and Tel Qrayot (31°21′27.93″ N, 35°08′49.44″ E). This type of village is limited to the semi-arid region north of the Beer Sheva Valley, where annual precipitations exceed 250 mm. These villages were found with adjacent small agricultural plots surrounding the settlements in a radius of ~300 m. These agricultural practices did not involve significant structural modifications of the natural environment (Haiman, 2012). The second group encompassed nomad clans that inhabited the Western Negev. These nomads’ camping sites were identified along Nahal Besor (31°10′27.69″ N, 34°31′34.32″ E) and dated to the 4th century CE onward (Gazit, 1996).
In light of the above evidence for church-dominated agriculture in the Beer Sheva valley during Late Antiquity, it is pertinent to briefly mention the collapse of this system. One of the main suggested reasons for this collapse is climatic change (Hirschfeld, 2006). However, it has become apparent that this argument was based on short-term climatic fluctuations, rather than on long-term climate changes (Avni, 2014: 287, 329–331). Furthermore, it is evident that the Byzantine runoff farming was well adapted to desert conditions. Moreover, during the Umayyad period, runoff farming practices expanded further south to the Negev Mountains (30°39′54.08″ N, 34°38′23.92″ E), which was deprived of any agricultural land-use during the Byzantine period (Haiman, 2012). An alternative suggested reason is the Islamic conquest. However, it has become clear that this conquest did not cause any changes in the Byzantine’s province administration (Avni, 2014; Magness, 2003). Another related reason is the prohibition of wine production and consumption, apparently imposed by the new Islamic rule (Mayerson, 1985). However, this prohibition was enforced only in the beginning of the Abbasid period (mid-8th Century CE), about a century and a half later than the collapse of the Byzantine Empire (Sharon, 1985: 34*–35*), or even later (Amar, 2000).
Therefore, an alternative explanation for the diminution of the immense, church-affiliated agricultural domain should be sought. Studies on the collapse of the Byzantine Empire suggested that it began to dwindle in the late 6th century. Several reasons for this process have been suggested, of which an economic crisis is predominant (Cameron, 1993). Such a crisis may have led to the considerable decline of the region’s wine industry, which was mainly export-oriented. Recently, it has been proposed that the population in Shivta and other church-affiliated settlements decreased during the late Byzantine period. Yet, Shivta persevered into the Early Islamic period, until the mid-8th century CE (Bar-Oz et al., 2019; Tepper et al., 2018). It is worth noting that while the Negev towns continued declining during the Umayyad period, populations of rural settlements simultaneously increased. These were managed by the Umayyad’s central authorities from Gaza and covered areas not settled during the Byzantine period.
Overall, our study supports the approach of a wide-scale decrease of the agricultural domain before the end of the Byzantine period. The excavations confirm that the watchtowers and industrial wine presses went out of use before the end of the Byzantine period (Ayalon, 1997; Taxel, 2013), although the settlements themselves prospered until the end of this period and beyond. We hereby propose that the region’s agricultural domain diminished because of the considerable decline of the export-oriented wine industry, rather than the collapse of the entire organized settlements.
The main insight of this study is that the similarity between the agricultural systems of the Beer Sheva Valley loess plains and those of the Negev Highlands, Judean Lowlands, and Southern Hebron Mountains, indicates their affiliation to the same centralized Byzantine-state framework, which was led and managed by the church. This study highlights the benefits of the interdisciplinary approach when assessing the interactions between human populations and the prevailing natural conditions, while considering the central role-played by the state. On the one hand, as revealed by the immense runoff farming systems across the region, the state can considerably advance agricultural development to fulfill certain geo-political purposes. On the other hand, weakening of the state could lead to the considerable decline of the centrally organized agricultural domain.
Conclusion
To understand the capacity of rural development in ancient times, the relevant social and geo-political frameworks must be clearly defined. This study emphasizes the significance of the organized settlements across the Beer Sheva Valley and its vicinities during the Byzantine period, which peaked in the mid-6th century CE. The regions’ settlements and agricultural systems are similar to those known to the south (Negev Highlands) and to the north (Judean Lowlands and Southern Hebron Mountains).
Settlement of this area was made possible by the centrally established and maintained runoff farming systems, enabling agricultural practices across this arid region. These complex farming systems, managed by the church in the service of the Byzantine Empire, demonstrate the high organizational capacity of the state’s authorities. The runoff harvesting agricultural systems were in use until the end of the Umayyad period, during the mid-8th century CE. As opposed to the proposed mechanism of human reaction to climatic change that dictates the spread of agriculture (Hirschfeld, 2006; Rosen, 2007), our study emphasizes the concept according to which the increase and decline of drylands settlements was predominantly regulated by geo-political powers. A more thorough discussion of this topic is beyond the focus of this study. One way or another, results of this study contradict those of some previous studies, which negated the viability of runoff farming systems across this region in ancient times. Insights of this study demonstrate the need for an interdisciplinary approach when studying such complex themes.
Footnotes
Acknowledgements
The authors are grateful to two anonymous reviewers, whose comments resulted in a considerable improvement of a previous version of the manuscript. The authors also thank the Ministry Of Science and Technology for their general support in this study.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
