INTRODUCTION
The Southern Levant’s Early and Middle Bronze Age populations witnessed the rise, abandonment, and rejuvenation of fortified towns atop this region’s mounded tells. Levantine archaeological syntheses historically have characterized this roughly two-millennium era in terms of urban growth and recession (e.g., Cohen Reference Cohen, Steiner and Killebrew2014; de Miroschedji Reference de Miroschedji, Steiner and Killebrew2014; Greenberg Reference Greenberg2019; cf. Falconer Reference Falconer, Mathers and Stoddart1994; Chesson and Philip Reference Chesson and Philip2003; Savage et al. Reference Savage, Falconer, Harrison, Levy, Daviau, Younker and Shaer2007). Many Bronze Age communities may be considered “peripheral” by virtue of their small populations and distant locations (geographically and/or politically) from larger “centers” (e.g., see discussions in Mohr and Thompson Reference Mohr and Thompson2023). Despite long-standing recognition of the importance of peripheral communities in early urbanized societies (e.g., Rowlands et al. Reference Rowlands, Larsen and Kristiansen1987; Champion Reference Champion1996; Maeir et al. Reference Maeir, Dar and Safrai2003; Haour Reference Haour2013; Cohen Reference Cohen2022), marginal settlements in the Southern Levant have not received the detailed chronometric attention that would improve their integration into Bronze Age social reconstructions. Bayesian analysis of a newly-expanded suite of calibrated AMS ages from Zahrat adh-Dhra‘ 1 (ZAD 1), Jordan documents this settlement’s occupational history, thereby providing a detailed chronology for a unique peripheral Middle Bronze Age community amid the urbanization that has otherwise attracted the focus of archaeological attention (Figure 1). We also correlate the habitation at ZAD 1 with ongoing revisions to regional Middle Bronze Age chronology as an example of how such settlements may be integrated in larger chronometric interpretations of marginal, as well as central, components of ancient Levantine complex societies.
Figure 1 Map of eastern Mediterranean showing Zahrat adh-Dhra‘ 1, Zahrat adh-Dhra‘ 2, Dhra‘ and Bab edh-Dhra‘ on the Dead Sea Plain, Jordan, plus key sites that contribute radiocarbon ages for Levantine Bronze Age chronologies.
The pottery excavated from ZAD 1 conforms with typological parallels from the Levantine Middle Bronze Age, with no material culture evidence indicative of earlier or subsequent periods. More specifically, stylistic distinctions between pottery in lower and upper stratified levels indicated phases of occupation in Middle Bronze I and II (Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Czarzasty, Day, Meadows, Meegan, Sayej, Swoveland and Westaway2002a; Berelov Reference Berelov2006a:92, Reference Berelov2006b; Fall et al. Reference Fall, Falconer, Edwards, Levy, Daviau, Younker and Shaer2007). Spatial analysis of pottery deposition also suggested multiple phases of occupation in which some structures (e.g., at the northwestern end of ZAD 1) were occupied in Middle Bronze I, while others across the site were occupied in Middle Bronze I and II (Berelov Reference Berelov2006a, 2006b; Fall et al. Reference Fall, Falconer, Edwards, Levy, Daviau, Younker and Shaer2007).
Over the last two decades, Southern Levantine Bronze Age chronology has experienced significant revision with the emergence of a radiocarbon-based “High Chronology” (e.g., Regev et al. Reference Regev, de Miroschedji, Greenberg, Braun, Greenhut and Boaretto2012; Höflmayer and Manning Reference Höflmayer and Manning2022; Fall et al. Reference Fall, Richard, Pilaar Birch, Ridder, D’Andrea, Long, Hedges-Knyrim, Porson, Metzger and Falconer2022), which features repositioned subdivisions for the Early, Middle, and Late Bronze Ages, and offers the opportunity to build regional chronologies and societal interpretations independent of assumed interregional historical correlations (Table 1). Until recently, both the timing and explanation of Southern Levantine Bronze Age urban dynamics have been derived from the “export” of Egyptian dynastic chronology (Bietak Reference Bietak, Shortland and Bronk2013:81), especially as derived from Tell el-Dab‘a (Bietak and Höflmayer Reference Bietak and Höflmayer2007; Höflmayer and Manning Reference Höflmayer and Manning2022). In the Northern Levant, Bronze Age chronologies have been influenced analogously by rich textual evidence, especially from Kültepe, Turkey and Mari, Syria (e.g., Barjamovic et al. Reference Barjamovic2012; Sasson Reference Sasson2015), and by dynastic sequences from the Hittite Old Kingdom and Mesopotamia’s Old Babylonian Period (e.g., Pruzsinszky Reference Pruzsinszky2009; Roaf Reference Roaf2012).
Table 1 Traditional and revised Early and Middle Bronze Age chronologies for the Southern Levant. (Traditional chronology based on Dever Reference Dever1992; Levy Reference Levy and Levy1995: fig. 3; revised chronology based on Regev et al. Reference Regev, de Miroschedji, Greenberg, Braun, Greenhut and Boaretto2012; Fall et al. Reference Fall, Falconer and Höflmayer2021; Höflmayer and Manning Reference Höflmayer and Manning2022.)
In the Southern Levant, an era of town abandonment in Early Bronze IV has been both correlated with and attributed to Egyptian political disintegration during the First Intermediate Period, ca. 2200–2000 BCE (e.g., Dever Reference Dever1992:2; Sharon Reference Sharon, Steiner and Killebrew2014:52). Following a step-by-step historical rationale, the subsequent renewal of Southern Levantine town life in the Middle Bronze Age (starting about 2000 BCE) has been associated with the improved political and economic milieu of the Egyptian 12th Dynasty (linked to Middle Bronze I) and 13th Dynasty (paralleling Middle Bronze II) (e.g., Bietak Reference Bietak1991; Dever Reference Dever1992; Sharon Reference Sharon, Steiner and Killebrew2014:54). In turn, the end of the Middle Bronze Age has been attributed to the disruptions associated with Egypt’s 18th Dynasty ca. 1550–1500 BCE (Mumford Reference Mumford, Steiner and Killebrew2014). These synchronisms stem traditionally from seriation of Levantine pottery excavated from Tell el-Dab‘a (ancient Avaris) in the Nile Delta (e.g., Bietak Reference Bietak1991), chronologically diagnostic Egyptian artifacts (e.g., scarab stamp seals) excavated from stratified Southern Levantine sites (e.g., Weinstein Reference Weinstein1992), and from the chronological correlation of major periods of socio-political fragmentation and coherence in Egypt in tandem with those in the Southern Levant (Mumford Reference Mumford, Steiner and Killebrew2014; Sharon Reference Sharon, Steiner and Killebrew2014).
An ever-growing body of radiocarbon analysis (e.g., Regev et al. Reference Regev, de Miroschedji, Greenberg, Braun, Greenhut and Boaretto2012, Reference Regev, Paz, Greenberg and Boaretto2019; Höflmayer Reference Höflmayer2021; Höflmayer et al. Reference Höflmayer, Dee, Genz and Riehl2014, Reference Höflmayer, Kamlah, Sader, Dee, Kutschera, Wild and Riehl2016a, Reference Höflmayer, Yasur-Landau, Cline, Dee, Lorentzen and Riehl2016b) has transformed Southern Levantine Bronze Age chronology by disarticulating it from Egyptian dynastic history. Among its most significant changes, the recession of town life in Early Bronze IV now stretches over a half-millennium or more (Höflmayer et al. Reference Höflmayer, Dee, Genz and Riehl2014; Regev et al. Reference Regev, de Miroschedji, Greenberg, Braun, Greenhut and Boaretto2012; Lev et al. Reference Lev, Bechar and Boaretto2021; Fall et al. Reference Fall, Falconer and Höflmayer2021, Reference Fall, Richard, Pilaar Birch, Ridder, D’Andrea, Long, Hedges-Knyrim, Porson, Metzger and Falconer2022), rivaling the lengths of the urbanized Early Bronze II/III and Middle Bronze eras, and thereby drawing attention to the chronometric investigation of non-urban settlements in the ancient Southern Levant. Likewise, these studies show that the constituent subperiods of the Middle Bronze Age no longer coincide with the convenient dynastic alignments of the Egyptian Middle Kingdom. In the context of this newly-independent interpretation of Levantine chronology and society, our Bayesian modeling of radiocarbon ages elucidates the habitation of ZAD 1, a unique peripheral Middle Bronze Age settlement, in conjunction with this era’s revised regional chronology.
Archaeological Setting of Zahrat adh-Dhra‘ 1 (ZAD 1)
The ancient settlement of ZAD 1 lies at about –180 msl on the hyperarid Plain of Dhra‘, a highly eroded landscape perched above the eastern shore of the Dead Sea (−410 msl). ZAD 1 is marked by more than 40 semi-subterranean rectangular stone structures varying from single wall remnants to multiple room compounds with adjoining enclosure walls (Fall et al. Reference Fall, Falconer and Porson2019). The archaeological features of ZAD 1 cover about 6 ha along a narrow 600-m-long ridge bounded by the Wadi Dhra‘ and Wadi Wa‘ida (Figure 2), two spring-fed tributaries of the Wadi Kerak, which flows west from the Transjordanian Plateau to the Dead Sea. Truncated structural walls along the southwestern edge of this ridge, as well as 200 meters across the Wadi Dhra‘, suggest that ZAD 1 originally extended over roughly 12 ha prior to stream down cutting. Three earlier sedentary agrarian settlements on the Plain of Dhra‘ attest to the presence of springs on the Dead Sea Plain: Pre-Pottery Neolithic A Dhra‘, upslope near the spring of ‘Ain Dhra‘ (Kuijt and Mahasneh Reference Kuijt and Mahasneh1998), PPNA Zahrat adh-Dhra‘ 2, just east of Zahrat adh-Dhra‘ 1 (Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001, Reference Edwards, Meadows, Sayej and Metzger2002b; Edwards and House Reference Edwards and House2007), and Early Bronze Age Bab edh-Dhra‘, about 4 km downstream along the Wadi Kerak (Rast and Schaub Reference Rast and Schaub2003). For example, the remains of freshwater mollusks (Melanopsis praemorsa) were excavated from Zahrat adh-Dhra‘ 2 (Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001, Reference Edwards, Meadows, Sayej and Metzger2002b). During the Middle Bronze Age, local wadis similarly would have flowed at or near the plain level of ZAD 1, as indicated by ubiquitous evidence of hydrophilic wild sedges (Cyperaceae) recovered in flotation samples (Fall et al. Reference Fall, Falconer and Porson2019). Freshwater mollusks at the PPNA site of ZAD 2 and an abundance of sedges at Middle Bronze Age ZAD 1 both suggest the presence of springs on the Plain of Dhra‘ between about 9000 and 2000 cal BCE. The lack of subsequent Late Bronze or Iron Age sites on the Plain of Dhra‘ and the presence of Middle Bronze Age structures at ZAD 1 on both sides of the Wadi Dhra‘ suggests that wadi downcutting occurred during or immediately following the Middle Bronze Age occupation of ZAD 1 and may have forced its abandonment (Figure 3). Wadi incision on the Plain of Dhra‘ could have resulted from long-term lowering of the Dead Sea in the mid-Holocene (Frumkin et al. Reference Frumkin, Kadan, Enzel and Eyal2001; Bookman (Ken-Tor) et al. Reference Bookman, Enzel, Agnon and Stein2004; Migowski et al. Reference Migowski, Stein, Prasad, Negendank and Agnon2006; Torfstein et al. Reference Torfstein, Goldstein, Stein and Enzel2013; Neugebauer et al. Reference Neugebauer, Brauer, Schwab, Waldmann, Enzel, Kitagawa, Torfstein, Frank, Dulski, Agnon, Ariztegui, Ben-Avraham, Goldstein and Stein2014; Guillerm et al. Reference Guillerm, Gardien, Waldmann, Brall, Ariztegui, Schwab, Neugebauer, Lach and Caupin2023).
Figure 2 Site plan of Zahrat adh-Dhra‘ 1, Jordan, showing visible wall lines, enumerated structures and lettered excavation units.
Figure 3 Photo of Wadi adh-Dhra‘, Jordan showing Zahrat adh-Dhra‘ 1 illuminated at top of photo and down cutting of wadi; see crew member crossing wadi bed at bottom of photo.
Chronologically diagnostic pottery from ZAD 1 includes vessel forms most closely paralleled in Middle Bronze I and II assemblages from sites across the Southern Levant (see discussions in Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Czarzasty, Day, Meadows, Meegan, Sayej, Swoveland and Westaway2002a; Berelov Reference Berelov2006a:92, Reference Berelov2006b; Fall et al. Reference Fall, Falconer, Edwards, Levy, Daviau, Younker and Shaer2007). The ZAD 1 ceramics also have a dearth of transitional Early Bronze IV/Middle Bronze I forms (e.g., as seen in Tell el-Hayyat Phase 5; Falconer and Fall Reference Falconer and Fall2006:46–49) and lacks several hallmark Middle Bronze III types altogether (e.g., ovoid cooking pots, highly profiled jar and bowl rims, high-footed bowls or chalices; also seen in Tell el-Hayyat Phases 2 and 1; Falconer and Fall Reference Falconer and Fall2006:56–60). These factors led to an original estimation of occupation between mid-Middle Bronze I and mid-Middle Bronze II (Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Czarzasty, Day, Meadows, Meegan, Sayej, Swoveland and Westaway2002a; Berelov Reference Berelov2006b).
Analysis of excavated plant remains and animal bones provides a portrait of lifeways and landscape at ZAD 1. Charcoal evidence shows that the settlement’s occupants relied on fuel wood from desert and riparian trees (e.g., Acacia spp. and Tamarix spp.), complemented with dung fuel from animals browsing on wild and domesticated plants (Klinge and Fall Reference Klinge and Fall2010). A modest faunal assemblage reflects sheep/goat husbandry, based on 323 bone specimens, of which 30 of 31 identifiable elements represent domestic sheep (Ovis aries) or goat (Capra hircus), while one identifiable bone comes from a domestic pig (Sus scrofa) (Metzger in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001; Fall et al. Reference Fall, Falconer, Edwards, Levy, Daviau, Younker and Shaer2007; Fall et al. Reference Fall, Falconer and Porson2019). The seed evidence reveals more abundant deposition of barley than wheat, suggesting arid-adapted cereal cultivation, while the presence of figs and grapes and a lack of olive form a crop combination more common at marginal agrarian settlements elsewhere in the Levant (Fall et al. Reference Fall, Falconer and Lines2002, Reference Fall, Falconer and Porson2019). Highly variable ubiquities among both cultivated and wild plant taxa, and increasingly sporadic seed deposition across structures and through time attests to habitation by small numbers of scattered households practicing non-intensive irrigated agriculture in an isolated, extremely dry environment where inhabitants exploited desert and riparian trees and utilized water from local springs. In overview, the botanical and environmental evidence from ZAD 1, combined with its rectilinear semi-subterranean structures and decidedly marginal setting, make this settlement a unique example of a peripheral agrarian community on the verge of urbanized Southern Levantine Middle Bronze Age society.
METHODS
During winter 1999/2000, excavation of 24 units sampled the interiors, exteriors, and enclosures associated with nine structures (Structures 36–44) (Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001). Material culture and organic remains were recovered primarily from interior deposits consisting of upper and lower stratified sediment layers. In some cases, these sediments were separated clearly into earlier and later layers by thin lenses of fine-grained aeolian sediment (Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001), which may have been deposited in single- to multiple-season episodes. The interior concentration of trash deposits (including restorable whole vessels) and the apparent absence of exterior middens, despite the excavation of promising midden locations (e.g., units O, S, and T), suggest occupation by small populations who periodically inhabited and vacated these structures, leaving interior trash behind (Schiffer Reference Schiffer1985; Wilson Reference Wilson1994).
All archaeological sediments with visible burned organic remains were processed by non-mechanized water flotation to recover plant macrofossils (Fall et al. Reference Fall, Falconer and Porson2019). Each flotation sample received a unique spatial identifier consisting of a structure number, an excavation area, a locus number associated with a three-dimensional archaeological feature, and a bag number. All plant remains larger than 0.25 mm were sorted using a binocular microscope at 6–40× magnification and were identified on the basis of reference material and published keys in accordance with established methods of archaeobotanical recovery and analysis (Fall et al. Reference Fall, Falconer and Porson2019; Porson et al. Reference Porson, Falconer, Pilaar Birch, Ridder and Fall2021).
Twenty-two seed samples recovered from seven structures at ZAD 1 were selected for AMS analysis by the University of Arizona Accelerator Mass Spectrometry Laboratory, the University of Georgia Center for Applied Isotope Studies, and the Australian Institute of Nuclear Science in the Australian Nuclear Science and Technology Organisation (ANSTO) (Table 2). These seeds were pretreated prior to analysis with an acid/alkali/acid protocol. The uncalibrated radiocarbon age for each sample is indicated in radiocarbon years BP based on the 14C half-life of 5568 years. The error for each uncalibrated date is shown as one standard deviation and reflects both statistical and experimental errors. These dates have been corrected for isotope fractionation using δ13C values.
Table 2 Seed samples from Zahrat adh-Dhra‘ 1, Jordan submitted for radiocarbon analyses. Excavation context shown according to Excavation Unit, Locus and Bag, (e.g., A.012.41 = Unit A, Locus 012, Bag 41).
The AMS samples from ZAD 1 were grouped into three phases based on a priori stratigraphic information (see Berelov Reference Berelov2006a:23–53; Fall et al. Reference Fall, Falconer and Porson2019). The seven structures that provide these ages have stratigraphically defined upper and lower sedimentary layers according to which seven samples from upper layers were grouped into Phase 2, and 10 samples from lower layers were grouped in Phase 3. Structures 37 and 40 had additional, lowermost sedimentary layers that provided three samples from basal earthen floors (Falconer in Edwards et al. Reference Edwards, Falconer, Fall, Berelov, Davies, Meadows, Meegan, Metzger and Sayej2001; Berelov Reference Berelov2006a:26, 35) that were grouped in Phase 4. Phases 4–2, which were defined a priori by stratigraphy and its associated ceramic typology in our previous chronological modeling of ZAD 1 (Fall et al. Reference Fall, Falconer and Porson2019), also included an anomalously late AMS age from a disturbed context (OZH-759), which we designated originally as Phase 1 (see discussion in Supplementary Material).
The radiocarbon ages from ZAD 1 were calibrated using OxCal 4.4.4 (Bronk Ramsey Reference Bronk Ramsey2009) and the IntCal20 atmospheric curve (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey, Butzin, Cheng, Edwards and Friedrich2020; van der Plicht et al. Reference van der Plicht, Bronk Ramsey, Heaton, Scott and Talamo2020). OxCal 4.4.4 also was used for Bayesian modeling of the calibrated dates. Bayesian analysis accommodates the non-normally distributed probabilities of calibrated 14C ages and enables probabilistic modeling of calibrated 14C determinations using prior stratigraphic information (Bronk Ramsey Reference Bronk Ramsey2009). Agreement values (A, Amodel) provide a means of assessing the reliability of the individual calibrated ages in Bayesian models and the quality of overall models. Values of A calculate the likelihood of overlap of the non-modeled distribution for each calibrated age with its posterior Bayesian modeled distribution. Values of A > 60 approximate values of p < 0.05 for a χ2 significance test (Manning Reference Manning2013:496, fig. A5), and values of Amodel > 60 identify statistically robust Bayesian models. The radiocarbon ages from ZAD 1 were organized for modeling using the “Phase” function in Oxcal, in which each phase consists of a group of unordered events. The “Difference” function in Oxcal was used to assess the statistical probability of gaps between phases.
RESULTS
A suite of 22 calibrated AMS radiocarbon seed ages from ZAD 1 was considered for Bayesian modeling of site occupation (Table 3). Our preferred model incorporates 20 ages in three sequential phases of occupation and excludes two additional ages (Amodel=101.9; Figure 4). This model groups these ages in sequential phases since they come from non-contiguous strata in varying combinations of seven distinct structures through time. We do not assume that these phases of occupation equate from start to finish between structures. Rather, we envision these phases as chronological windows within which varying combinations of structures were inhabited during each phase. Further, our three-phase model produces more parsimonious results than those from alternative one-phase or two-phase models, which make less effective use of this prior stratigraphic information. (Alternative phasing schemes are presented in Supplementary Figures 1–3, and their results are summarized in Supplementary Table 1.)
Table 3 Unmodeled and modeled calibrated AMS radiocarbon seed ages from Zahrat adh-Dhra‘ 1, Jordan. Uncalibrated 14C ages are indicated in BP with their 1σ uncertainty. Calibration based on OxCal 4.4.4 (Bronk Ramsey Reference Bronk Ramsey2009, Reference Bronk Ramsey2017) using the IntCal20 atmospheric curve (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey, Butzin, Cheng, Edwards and Friedrich2020; van der Plicht et al. Reference van der Plicht, Bronk Ramsey, Heaton, Scott and Talamo2020). Stratigraphic phases at Zahrat adh-Dhra‘ 1 run from Phase 4 (the earliest phase) to Phase 2 (the latest phase). Samples are tabulated by phase and ordered chronologically according to conventional 14C age within each phase.
Figure 4 Bayesian sequencing of 20 calibrated 14C ages for seed samples from Phases 4–2 at Zahrat adh-Dhra‘ 1, Jordan. Amodel = 101.9. Light gray curves indicate single-sample calibration distributions; dark curves indicate modeled calibration distributions. Calibrations and Bayesian modeling based on OxCal 4.4.4 (Bronk Ramsey Reference Bronk Ramsey2009) using the IntCal20 atmospheric curve (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey, Butzin, Cheng, Edwards and Friedrich2020; van der Plicht et al. Reference van der Plicht, Bronk Ramsey, Heaton, Scott and Talamo2020).
Phase 4, which documents the earliest phase of occupation at ZAD 1, is indicated by three samples from two structures (Table 4) whose modeled 2σ distributions lie between about 2100 and 1900 cal BCE. Modeled boundary 1σ distributions frame this phase between about 2050 and 1900 cal BCE. The modeled boundaries for the end of Phase 4 and the start of Phase 3 indicate a gap between these phases at the 1σ confidence level (e.g., as shown by their disjunct 1σ probability distributions and by the “Difference” function in Oxcal).
Table 4 Structures at Zahrat adh-Dhra‘ 1 shown according to phases of occupation and numbers of radiocarbon dates in each phase, based on three-phase Bayesian model of 20 AMS ages (see Figure 4).
Phase 3 is characterized by a set of 10 AMS ages from seven structures with 1σ and 2σ modeled distributions ranging from just after 1900 to just after 1800 cal BCE. These remarkably consistent modeling results suggest a well-defined phase of settlement during the 19th century cal BCE. Our model does not reveal a gap between Phases 3 and 2, as shown by overlapping 1σ and 2σ probability distributions for the modeled boundaries at the end of Phase 3 and the start of Phase 2. Phase 2 at ZAD 1 is represented by another consistent sequence of seven modeled dates from five structures, in this case suggesting settlement in the early 18th century BCE, based on 1σ probability distributions between about 1790 and 1740 cal BCE. Phase 2 is estimated to conclude by 1700 cal BCE, as shown by the 2σ probability distribution for its end boundary.
In sum, our 20-age Bayesian model for ZAD 1 commences with three ages from lower sediments in two structures suggesting occupation in Phase 4 most likely between about 2050 and 1950 cal BCE. Phase 3 signals occupation probably within the 19th century cal BCE based on 10 calibrated ages from lower sediments in all seven sampled structures. In Phase 2, seven ages from upper sediments in five structures provide particularly focused evidence for settlement in the early to mid-18th century cal BCE, prior to the apparent abandonment of ZAD 1 by about 1700 cal BCE.
DISCUSSION
Our new Bayesian analysis of calibrated AMS ages from ZAD 1 (a) revises and clarifies the chronology for the occupation of ZAD 1 within the Middle Bronze Age, (b) provides the only radiocarbon-based model for a peripheral Bronze Age settlement in the Southern Levant, and (c) bolsters the emerging radiocarbon-based Middle Bronze Age chronology and its implications for more independent interpretation of the development of Levantine complex society. The importance of peripheral communities has been noted in archaeological syntheses of growing and receding town life in Levantine Bronze Age societies (e.g., Prag Reference Prag, Steiner and Killebrew2014; Cohen Reference Cohen2022). Most notably, the Early Bronze IV Period offers evidence of seasonal encampments, upland cemeteries and small villages that figure prominently in long-standing characterization of Early Bronze IV society in terms of urban abandonment (e.g., Dever Reference Dever1980, Reference Dever and Levy1995; Prag Reference Prag, Steiner and Killebrew2014). Nevertheless, only a handful of Early Bronze IV marginal settlements (e.g., Be’er Resisim, Ein-Ziq, Nahal Refaim, Ha-Gamal) provide a modest corpus of largely charcoal radiocarbon ages that has been used to initiate a discussion of a general chronology for the period (e.g., Regev et al. Reference Regev, de Miroschedji, Greenberg, Braun, Greenhut and Boaretto2012; Fall et al. Reference Fall, Richard, Pilaar Birch, Ridder, D’Andrea, Long, Hedges-Knyrim, Porson, Metzger and Falconer2022), but which is insufficient for formal modeling of occupation at any given site. During the Middle Bronze Age resurgence of fortified towns, marginal populations remained important social and economic elements of Levantine society (e.g., Cohen Reference Cohen2002). To date, however, ZAD 1 provides the only example of a peripheral settlement that offers a Bayesian model specifying its chronological articulation within the Middle Bronze Age.
The growing evidence for a high Southern Levantine Bronze Age radiocarbon chronology features a lengthened Early Bronze IV beginning by 2500 cal BCE and continuing after its traditional end about 2000 cal BCE (Fall et al. Reference Fall, Falconer and Höflmayer2021, Reference Fall, Richard, Pilaar Birch, Ridder, D’Andrea, Long, Hedges-Knyrim, Porson, Metzger and Falconer2022). Likewise, growing radiocarbon evidence questions the assumed correlation between the beginning of the subsequent Levantine Middle Bronze Age and the accession of the Egyptian 12th Dynasty ca. 2000 BCE (e.g., Höflmayer Reference Höflmayer2021; Höflmayer et al. Reference Höflmayer, Kamlah, Sader, Dee, Kutschera, Wild and Riehl2016a). Accordingly, the timing for the transition between Early Bronze IV and the subsequent Middle Bronze Age has been shifted as late as 1925 cal BCE (Cohen Reference Cohen2002) or even later (e.g., as modeled about 1900 cal BCE at Tell el-Hayyat; Fall et al. Reference Fall, Falconer and Höflmayer2021). The interval modeled here for Phase 4 at ZAD 1, therefore, might seem unexpectedly early, unless we consider the possibility that the newly-flexible Early/Middle Bronze Age transition might have been time-transgressive, as well as less dependent on Egyptian history, potentially appearing in southern locations like ZAD 1 earlier than in more northerly settlements. Following a possible gap between Phases 4 and 3 (modeled at a 1σ confidence level), Phase 3 habitation in the 19th century cal BCE accords with current modeling of late Middle Bronze I, including the Middle Bronze I/II transition about 1850–1800 cal BCE (e.g., Höflmayer and Manning Reference Höflmayer and Manning2022). Thereafter, Phase 2 occupation at ZAD 1 most likely in the early to mid-18th century cal BCE is consistent with current chronological definition of Middle Bronze II and ends clearly before the proposed Middle Bronze II/III transition about 1700 cal BCE (Höflmayer and Manning Reference Höflmayer and Manning2022). In the larger context of regional chronology, Phase 4 at ZAD 1 lies early in Middle Bronze I, potentially beginning as early as, or slightly before, the traditional outset of this period around 2000 cal BCE. Phase 3 accords with the MB I/II transition as now dated between about 1850 and 1800 cal BCE, based on data from sites across the Southern Levant (Höflmayer and Manning Reference Höflmayer and Manning2022), while Phase 2 lies squarely in the midst of Middle Bronze II in the early 18th century cal BCE in keeping with continued emergence of a high chronology for the Levantine Bronze Age. ZAD 1 thereby provides a uniquely detailed occupational history for a peripheral settlement during the burgeoning town life of the Southern Levantine Middle Bronze Age.
Modeling of the Middle Bronze Age habitation of ZAD 1 appears amid increasingly ambitious efforts over the last decade to synthesize radiocarbon-based chronologies across regions of the Eastern Mediterranean and Near East previously separated according to geography and archaeological tradition (e.g., Lebeau Reference Lebeau2011; Peltenburg Reference Peltenburg2013; Finkbinder et al. Reference Finkbinder, Novak and Sconzo2015). Among recent breakthroughs, Manning (Reference Manning2022) models the chronological linchpin eruption of Thera in the early to mid-16th century cal BCE, thereby providing another instance of shifted historical correlation between the ancient Mediterranean world and Egypt. Similarly, Herrmann et al. (Reference Herrmann, Manning, Morgan, Soldi and Schloen2023) use modeled AMS ages from Zincirli, Turkey to connect Mesopotamian and Eastern Mediterranean chronologies at the close of the Middle Bronze Age in the decades following the abandonment of ZAD 1. In this context, ZAD 1 illuminates the life history of a Southern Levantine peripheral community on the verge of the ensuing emergence of Late Bronze Age empires.
CONCLUSIONS
The excavation and analysis of ZAD 1 have illuminated this unique, non-intensively occupied settlement on the cultural and geographical margin of Middle Bronze Age urbanized society in the Southern Levant. Bayesian modeling of 20 calibrated AMS ages provides an occupational chronology unique to the Middle Bronze Age and unparalleled in detail among peripheral sites from other periods, most notably Early Bronze IV. Three sequential phases of occupation at ZAD 1 and their associated pottery articulate well with the emerging regional radiocarbon chronology for Middle Bronze I and II, thereby bolstering increasingly independent interpretation of the Levantine Bronze Age. The radiocarbon chronology presented here also exemplifies the potential for detailed chronological investigation of peripheral sites and their integration into increasingly nuanced archaeological chronologies for the development of early complex societies in greater Southwestern Asia and the Eastern Mediterranean. In particular, the radiocarbon evidence from ZAD 1 illuminates the prehistory of a unique Middle Bronze Age settlement as a means of broadening archaeological comprehension of the full range of communities that constituted Bronze Age society in the Southern Levant.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2023.99.
ACKNOWLEDGMENTS
Fieldwork at Zahrat adh-Dhra‘ 1 was conducted under permit from the Department of Antiquities, Hashemite Kingdom of Jordan, in affiliation with the American Center of Research, Amman. We thank two anonymous reviewers for their insightful comments and suggestions.
FUNDING
Funding provided by grants from the US National Science Foundation (#1850259, #2114406), the Australian Research Council (with P. Edwards), the National Geographic Society (#6655-99), the Wenner-Gren Foundation for Anthropological Research (#6608), and the Australian Institute for Nuclear Science and Engineering (with P. Edwards).
AUTHOR CONTRIBUTIONS
Fall: conceptualization, methodology, investigation, resources, writing original draft and editing; Ridder: methodology, analysis, visualization, data curation, manuscript revision and editing; Pilaar Birch: methodology, analysis, investigation, manuscript revision and editing; Falconer: conceptualization, methodology, investigation, resources, writing original draft and editing.
DECLARATION OF COMPETING INTERESTS
The authors declare no conflicts or competing interests.
Open access
