NEW BAYESIAN RADIOCARBON MODELS AND CERAMIC CHRONOLOGIES FOR EARLY BRONZE IV TELL ABU EN-NI‘AJ AND MIDDLE BRONZE AGE TELL EL-HAYYAT, JORDAN

ABSTRACT We present two new Bayesian 14C models using IntCal20 that incorporate 17 new calibrated AMS ages for Early Bronze IV Tell Abu en-Ni‘aj and Middle Bronze Age Tell el-Hayyat, located in the northern Jordan Valley, Jordan. These freshly augmented suites of carbonized seed dates now include 25 AMS dates from Tell Abu en-Ni‘aj and 31 AMS dates from Tell el-Hayyat. The modeled founding date for Tell Abu en-Ni‘aj strengthens an emerging high chronology for Early Bronze IV starting by 2500 cal BC, while the end of its habitation by 2200 cal BC may exemplify a regional pattern of increasingly pervasive abandonment among late Early Bronze IV settlements in the Southern Levant. In turn, our modeled date for the Early Bronze IV/Middle Bronze Age transition at Tell el-Hayyat around 1900 cal BC pushes this interface about a century later than surmised traditionally, and its abandonment in Middle Bronze III marks an unexpectedly early end date before 1600 cal BC. These inferences, which coordinate Bayesian AMS models and typological ceramic sequences for Tell Abu en-Ni‘aj and Tell el-Hayyat, contribute to an ongoing revision of Early and Middle Bronze Age Levantine chronologies and uncoupling of their attendant interpretive links between the Southern Levant and Egypt.


INTRODUCTION
The Early and Middle Bronze Ages of the Southern Levant are characterized traditionally in terms of early urbanism, its abandonment and redevelopment, between about 3500 and 1500 BC. Incipient walled settlements emerged in Early Bronze I (Joffe 1993;Gophna 1995;Philip 2003), followed by larger and more numerous fortified towns in Early Bronze II-III, time periods often combined to reflect a lengthy initial era of Levantine urbanism (Greenberg 2002(Greenberg , 2014Philip 2008;de Miroschedji 2009de Miroschedji , 2014. The ensuing Early Bronze IV Period (or Intermediate Bronze Age) witnessed the pervasive abandonment of these towns in the Southern Levant. Traditional social interpretations have emphasized seasonal transhumant pastoralism (e.g., between low elevation encampments and hill country cemeteries; Dever 1980Dever , 2014, while more recent studies have elucidated a growing number of sedentary agrarian villages (Palumbo 1991;Cohen 2009;Fall 2009, 2019;Richard et al. 2010;Prag 2001Prag , 2014D'Andrea 2014). The subsequent Middle Bronze Age has long been celebrated as the apex of pre-Roman urbanism in the Southern Levant (e.g., Dever 1987;Ilan 1995). Walled cities reappeared rapidly in Middle Bronze I (known previously as Middle Bronze IIA) and grew in size, number, and fortification during Middle Bronze II and III (formerly Middle Bronze IIB and IIC) (Burke 2008;Bourke 2014;Cohen 2014). In overview, Levantine urbanization has been viewed as a long-term social transformation that unfolded over most of the third and second millennia BC, aside from a brief punctuation during the abandonment of towns in Early Bronze IV (hence this period's alternative "intermediate" nomenclature).
The relative chronology of the Southern Levantine Bronze Age traditionally derives from seriated material culture sequences and stylistic parallels (especially in pottery vessel morphology and metal weaponry) with Syria and Lebanon (Cohen 2002(Cohen , 2014Bourke 2014;de Miroschedji 2014;Prag 2014;Richard 2014). Major junctures in Levantine absolute chronology and their associated socio-political interpretations have relied on correlations with Egyptian dynastic history: (1) Early Bronze II-III town life roughly paralleled the rise of the Egyptian Old Kingdom (Stager 1992;Sowada 2009); (2) Early Bronze IV town abandonment is correlated with political decentralization during the Egyptian First Intermediate Period ca. 2300/2200 to 2000 BC (Stager 1992;Dever 1995;Prag 2014); (3) the rapid reestablishment of Middle Bronze Age towns is both inferred and explained as a response to political reunification starting with the ascension of the Egyptian 12th Dynasty ca. 2000 BC (Dever 1987;Stager 1992;Cohen 2016; see critique in Bruins 2007); and (4) the apex of Middle Bronze urbanism (MB III) is correlated with Hyksos rule in Egypt, ending with the Hyksos "expulsion" ca. 1550/1500 BC (Bietak 2013;Burke 2014;Sharon 2014; see discussion in Höflmayer 2019).
As a result of an ongoing radiocarbon revolution (e.g., see Manning et al. 2014) involving sitespecific and regional Bayesian radiocarbon modeling (Bronk Ramsey 2009a), the chronology and interpretation of early Levantine urban growth and abandonment are undergoing substantial revision. Critical examination of Levantine radiocarbon chronologies now reveals that former correlations with, and interpretations based on, Egyptian dynastic history may no longer be assumed axiomatically (Kutschera et al. 2012), whereas the Egyptian historical chronology itself has proven to be compatible with radiocarbon dating (Bronk Ramsey et al. 2010). For example, earlier beginning and end dates have been proposed for Early Bronze II and III as part of an emerging "high" Early Bronze chronology (e.g., Golani and Segal 2002;Bourke et al. 2009;Regev et al. 2012aRegev et al. , 2014Höflmayer et al. 2014;Falconer and Fall 2016), while a shortened duration has been suggested for the Middle Bronze Age (e.g., Höflmayer et al. 2016a;Fall 2016, 2017).
At the heart of these revisions lie three crucial temporal junctures: the earliest evidence for Early Bronze IV village settlements, the Early Bronze IV/Middle Bronze Age interface between eras of town abandonment and rejuvenation, and the end of the Middle Bronze Age II-III apex of town life. Excavations at the village sites of Tell Abu en-Ni'aj and Tell el-Hayyat, located in close proximity in the northern Jordan Valley (Figure 1), provide detailed stratified settlement histories and fine grained radiocarbon chronologies through Early Bronze IV and the Middle Bronze Age. New Bayesian models based on expanded suites of seed ages from both sites illuminate these junctures in the northern Jordan Valley, as well as their larger implications for independent explanation of Bronze Age societal dynamics in the Southern Levant.

SITE DESCRIPTIONS Tell Abu en-Ni'aj
Tell Abu en-Ni'aj (32º24 0 52 00 N; 35º34 0 6 00 E) incorporates the remains of an Early Bronze IV agrarian village in the northern Jordan Valley, Jordan. This mound sits at about 250 m below sea level (bsl) overlooking the zor, the active floodplain of the Jordan River, from its position near the edge of the ghor, the terrace of agricultural lands to the east. This site's size of about 2.5 ha suggests a likely Bronze Age population of 500-600 people , based on analogies with the population densities of ethnographically documented traditional farming villages in southwestern Asia (e.g., Kramer 1982). Tell Abu en-Ni'aj rises 3.30 m above the surrounding landscape, which consisted of agricultural fields during the site's excavation. Today, the tell is fenced amid the roads and warehouses of the Jordan Gateway industrial park. The site was reported originally by the East Jordan Valley Survey as being relatively large with predominantly "EB-MB" surface ceramics (Ibrahim et al. 1976: 49, 51;site 64). Tell Abu en-Ni'aj, occupied solely during the EB IV Period, is particularly significant as an uncommon example of a sedentary, continuously occupied, agriculturally dedicated settlement in a time period traditionally interpreted by archaeologists in terms of non-sedentary pastoral society.

Tell el-Hayyat
Tell el-Hayyat (32º25 0 14 00 N; 35º34 0 36 00 E) lies approximately 1.5 km northeast of Tell Abu en-Ni'aj at an elevation of about 240 m bsl in the midst of orchards and cultivated fields in the ghor. This 0.5 ha mound incorporates about 4.50 m of archaeological deposits and embodies the remains of a Bronze Age community of up to 200 inhabitants (Falconer and Fall 2006). Tell el-Hayyat was reported originally among the sites surveyed by Glueck (1951: 259; site 154, "Tell abu Hayet") and by Mellaart (1962: 144-145;site 24, "Tell Abu Hayet"). The subsequent, more systematic East Jordan Valley Survey highlighted Tell el-Hayyat's evidence for occupation during "EB-MB, MB IIA, MB IIB-C" (Ibrahim et al. 1976: 49;site 56) and therefore emphasized the site as having the potential to provide "a stratigraphically controlled sequence from EB-MB to MB IIB-C" (Ibrahim et al. 1976: 54). Indeed, a major part of Tell el-Hayyat's significance lies in the chronological inferences we can draw from its rare stratigraphic sequence over the Early Bronze IV/Middle Bronze Age transition and its record of agrarian village life during Middle Bronze I-III.

Field Excavations
Tell Abu en-Ni'aj was excavated over three field seasons totaling 16 weeks in fall 1985, winter 1996/97 and winter/spring 2000 . Our initial field season involved two weeks of test excavations in ten 4 × 4 m units in Fields 1-3, primarily exposing Phases 3-2 with soundings to Phase 6 ( Figure 2). Two subsequent seasons featured excavation of 27 contiguous 4 × 4 m units in Field 4 through seven stratified architectural phases of occupation, from uppermost Phase 1 to basal Phase 7. Fifteen units reached archaeologically sterile sediments underlying Phase 7. The cumulative area of the excavation units in all four fields represents about 2.5% of the site's area.
Tell el-Hayyat was excavated in three field seasons totaling 24 weeks in 1982, 1983and 1985(Falconer and Fall 2006. Excavation of 16 4 × 4 m units and four 2 × 4 m units proceeded through six stratigraphic phases from uppermost Phase 1 to basal Phase 6 ( Figure 3). Units A and B uncovered an intact pottery kiln and ceramic production debris on the south flank of the tell, while Units C-U revealed village remains centered around a remarkable stratified sequence of four Canaanite temples in antis (Magness-Gardiner and Falconer 1994;Falconer and Fall 2006). The aggregate excavated exposure of these 20 units represents about 8.0% of the site's area.

Material Culture Recovery and Analysis
During the excavation of both sites, larger elements of material culture and animal bones were recovered in the course of excavation, and all excavated sediments (other than those processed by water flotation) were dry-sieved in the field through 0.5 cm wire mesh to insure maximum recovery of smaller sized ceramics, bones, metal and stone artifacts. All excavated remains, except metallurgical evidence, were washed and sorted by material type for preliminary descriptive analysis in our field quarters at the Deir Alla Archaeological Station. All sherds were counted, non-diagnostics were returned to the modern surface of the two sites, and diagnostic sherds were labeled. Following this preliminary assessment, the excavated collections were shipped to the University of Arizona (1982-1985 seasons) and Arizona State University (1996/97 and 2000 seasons) for further analyses.
We employed the same methods of ceramic collection, analysis and interpretation at both archaeological sites (Falconer and Fall 2006: 29-30, 44-64;Falconer and Fall 2019: 13, 75-102). During the excavations, all ceramic remains were washed and sorted to segregate diagnostic sherds, including rims, bases, handles, spouts and all decorated specimens, from non-diagnostic undecorated body sherds. During lab analysis, all diagnostic sherds were quantified according to vessel form and function, decorative techniques were coded and counted, sherd dimensions and rim and base diameters were measured. Vessel forms, and the relative frequencies of vessel sizes, morphologies and decorative techniques were calculated and compared through the stratigraphic sequence at each settlement and compared with evidence from assemblages at other sites to infer chronological relationships and geographical affinities with other Bronze Age settlements in the Southern Levant (see discussions in Fall 2006: 44-64, 118-123, 2019: 75-114).

Botanical Analysis
All sediments with visible burned organic content excavated at Tell Abu en-Ni'aj and Tell el-Hayyat were processed using water flotation to recover plant macrofossils (Fall et al. 1998(Fall et al. , 2002(Fall et al. , 2015Fall 2006: 38-43, 2019: 13-14;Klinge and Fall 2010). To minimize the potential for chronological mixing, samples were selected as often as possible from relatively shallow localized deposits in or on burned surfaces. During the Tell el-Hayyat excavations and the 1985 season at Tell Abu en-Ni'aj, each flotation sample was poured into a metal basket with 3.2 mm mesh screen across its bottom, which was suspended in a metal tub of water. Each sample was agitated to dissolve the sediment and dislodge carbonized plant fragments, which were extracted with a large tea strainer (1.6 mm mesh). The smallest seeds were recovered by placing a piece of cheese cloth in the strainer. The coarse sediment fractions at the bottom of the flotation equipment were checked after processing of each sample for occasional heavier seeds that did not float (e.g., olive stones). During the 1996/97 and 2000 seasons, a Flote-Tech 2000 flotation machine was used to separate organic remains from the sediment matrix mechanically (Figure 4). Following flotation, plant remains were dried indoors for about 24 hours before being packaged for laboratory analysis. Both the coarse and fine fractions were examined for seeds and charcoal fragments. In the lab, each dried sample of carbonized remains was poured through nested 4.75 mm, 2 mm, 1 mm and 0.5 mm mesh sieves Fall 2006: 38-43, 2019: 13-14;Klinge and Fall 2010;Fall et al. 2015;Porson et al. 2019;Fall et al. 2019). All recovered material 0.25 mm or larger was sorted under a binocular microscope at 6 to 40× magnification to separate charcoal fragments from charred seeds. Seeds were identified using Fall's personal reference collection and comparative literature (e.g., Helbaek 1958Helbaek , 1966Delorit 1970;Martin and Barkley 1973;Zohary and Hopf 1973;van Zeist 1976;Zohary and Spiegel-Roy 1975;van Zeist and Bakker-Heeres 1982;Hubbard 1992;Jacomet 2006), counted and categorized taxonomically (Klinge and Fall 2010;Porson et al. 2019;Fall et al. 2019). Analysis of 123 flotation samples from Tell Abu en-Ni'aj has led to the recovery of more than 20,000 carbonized seeds, seed fragments and chaff (Porson et al. 2019), and 152 samples from Tell el-Hayyat have produced nearly 9000 identified seeds (Falconer and Fall 2006: 65-72;Fall et al. 2019).

AMS Analysis
Over the course of our investigations, we have emphasized the importance of generating AMS ages from seeds, rather than charcoal specimens, which are susceptible to the effects of inbuilt age (Waterbolk 1971). Our radiocarbon ages have been determined by the University of Arizona Accelerator Mass Spectrometry Laboratory, the University of Groningen Centre for Isotope Research, the Vienna Environmental Research Accelerator and the Oxford Radiocarbon Accelerator Unit.
Seed samples from Tell Abu en-Ni'aj were selected from well-defined contexts in all seven stratigraphic phases for AMS analysis (Table 1). Our sampling focused on cereals identified as Triticum sp. or Hordeum sp. whenever possible, and also included two samples of Prosopis sp. seeds. Among these 25 samples, 23 are drawn from small, well-defined burned   (Reimer et al. 2020). Stratigraphic phases at Tell Abu en-Ni'aj start with Phase 7 (the earliest, basal stratum) and end with Phase 1 (the latest, uppermost stratum). Samples are tabulated by phase and ordered chronologically according to conventional 14 C age within each phase. Context is indicated according to Excavation Unit, Locus and Bag (e.g., D.016.51 = Unit D, Locus 016, Bag 51); * Outlier Index ≥ 5%. features, including hearths, bins, shallow pits, and thin sediments (e.g., ash lenses) on earthen use surfaces. The remaining two samples come from localized deposits of decomposed mudbrick sediment. Our earlier publication of radiocarbon ages from Tell Abu en-Ni'aj (Falconer and Fall 2016) included three olive pit samples incorrectly attributed to Phase 1 at this site, which have been replaced with four new Hordeum sp. seed samples (AA-113003-113006) from secure Phase 1 contexts at Tell Abu en-Ni'aj. Phases 6-1 are now represented by three to five samples each, while the more limited Phase 7 floral assemblage provides a single AMS sample.
Similarly, 31 AMS seed samples were drawn from all six stratigraphic phases at Tell el-Hayyat (Table 2). Most AMS samples consist of Triticum sp., Hordeum sp., Cerealia and Olea sp. seeds. As at Tell Abu en-Ni'aj, these specimens were recovered from localized burned features, including ovens (tabuns), ash lenses, burned surfaces, and thin occupational sediments deposited on earthen floors. Phases 5-2 provide six to seven samples each, while Phase 6 contributes three specimens and Phase 1 offers a single sample. Our previous discussion of radiocarbon chronology at Tell el-Hayyat (Falconer and Fall 2017) concentrated on ages from Phases 5 and 4, and included two dates from Phase 6, and single dates from Phases 3, 2, and 1. This study contributes a greatly strengthened set of 14 C data, including 13 new ages from Phases 6, 3, and 2 produced by the Micadas AMS at Groningen (prefixed GrM in Table 2).

Bayesian Analysis
The radiocarbon ages from Tell Abu en-Ni'aj and Tell el-Hayyat were calibrated using OxCal 4.4.2 (Bronk Ramsey 2009a) and the IntCal20 atmospheric curve (Reimer et al. 2020;). The analytical tools in OxCal 4.4.2 were used for Bayesian modeling of the calibrated dates. As evidenced by its increasingly common use in archaeology, Bayesian analysis permits probabilistic modeling of large suites of calibrated 14 C determinations from multiple strata at a site or from multiple sites across a region. This method can incorporate prior stratigraphic information and accommodate the non-normally distributed probabilities of calibrated 14 C ages (Bronk Ramsey 2009a). The dates from Tell Abu en-Ni'aj and Tell el-Hayyat were organized for modeling in contiguous stratigraphic phases based on their records of continuous deposition. In each phase, the sequence of the individual samples was taken to be unknown (i.e., an unordered group of events). Outliers were detected using the Outlier Analysis tool in OxCal (Bronk Ramsey 2009b). Samples that are too old (e.g., due to residual material) or too young, and thus do not fit the prior stratigraphic information for the model are often excluded by hand, a process that can be highly subjective. In order to avoid subjective bias, we employed the Outlier Analysis in OxCal, using the General model for all short-lived samples. Outlier Analysis detects determinations that do not fit prior stratigraphic information and weighs their impact on the model accordingly. Thus, outlying dates have very little impact on the overall results (the posterior probabilities). The Outlier Index (O:xx/5) shows the expected (prior) chance of a determination being an outlier.

Stratigraphy and Architecture
Tell Abu en-Ni'aj The mound of Tell Abu en-Ni'aj incorporates seven stratified phases of construction and remodeling of mudbrick and rammed earth architecture . Our excavations revealed no stone-founded architecture, but did uncover flagstone surfaces and  (Reimer et al. 2020). Stratigraphic phases at Tell el-Hayyat start with Phase 6 (earliest, basal stratum) and end with Phase 1 (latest, uppermost stratum). Samples are tabulated by phase and ordered chronologically according to conventional 14 C age within each phase. Context is indicated according to Excavation Unit, Locus and Bag (e.g., L.006.9 = Unit L, Locus 006, Bag 9); * Outlier Index ≥ 5%.    networks of sherd-paved streets. The initial settlement, exemplified by the remains of Phase 7 in Field 4, was founded on archaeologically sterile sediments. The uppermost village is represented by Phase 1 remains, which are found over a limited area just below the modern surface at the top of the mound. The relatively modest extent of Phase 1 evidence may reflect erosion of this uppermost layer over the last four millennia. The archaeological stratification of Tell Abu en-Ni'aj shows no evidence of a stratigraphic gap or settlement hiatus from Phase 7 to Phase 1. The village's characteristically agglutinative architecture featured localized remodeling throughout its settlement history, with more substantial episodes of structural rebuilding and street reconfiguration between phases.
The architecture of Phases 7 and 6 in Field 4 includes rammed earth and mudbrick wall foundations, which are relatively sparse in comparison with those of subsequent phases. Some

Tell el-Hayyat
The history of occupation at Tell el-Hayyat is documented in six stratified phases of archaeological deposits (Falconer and Fall 2006). Phase 6 is a basal stratum lying just above archaeologically sterile sediments at the center of the site. The sediments that constitute this phase are primarily earthen use surfaces containing ceramics that feature forms, fabrics and decoration that are exclusively Early Bronze IV and include no Middle Bronze Age sherds. Other Phase 6 evidence includes substantial chipped stone remains, and modest amounts of animal bones and carbonized seeds, with no preserved architectural features. In light of the repeated comprehensive leveling of Tell el-Hayyat for phase-byphase reconstruction in subsequent phases (see discussion below), the multiple use surfaces in this basal phase, and its substantial ceramic assemblage (nearly 1000 sherds), Phase 6 represents a clearly distinguished Early Bronze IV stratum that may also have been levelled prior to building the Middle Bronze Age structures in Phase 5. The central buildings of Tell el-Hayyat in subsequent phases are four Middle Bronze Age mudbrick temples in antis, with associated standing stones and enclosure walls, which are stratified above Phase 6 at the mound's center (Magness-Gardiner and Falconer 1994;Falconer and Fall 2006: 33-43, 83-110). This remarkable temple sequence begins with the rammed earth foundations of a modest shrine in Phase 5, which are directly overlain by the remains of successively larger and more elaborate mudbrick, plastered and painted temples in Phases 4-2 ( Figure 6). Although no additional Phase 5 structures were uncovered, Phase 4-2 domestic buildings with walled courtyards, which were separated by alleyways, lay outside the temple enclosures. Noteworthy architectural evidence also includes an intact pottery kiln with an interred human skull in Phase 4 (Area A; Falconer and  The architecture and associated community structure at Tell el-Hayyat differs from Tell Abu en-Ni'aj in a number of characteristics. For instance, Tell el-Hayyat's occupants utilized compounds, whether domestic or ritual, that were intentionally segregated from one another by walls and alleys, rather than conjoined in architectural blocks like those of Tell Abu en-Ni'aj. Further, in contrast to the incremental remodeling at Tell Abu en-Ni'aj, the architectural history at Tell el-Hayyat is marked by roughly simultaneous architectural changes across the village between phases. To begin with, Phase 5 is marked clearly by the construction of the earliest temple and its enclosure wall, which lie directly over the nonarchitectural sediments of Phase 6. Subsequently, each of the temples in Phases 5-3 appears to have been intentionally leveled to facilitate its expedient reconstruction. For example, the rammed earth foundations of Phase 5 were shaved flat to provide a level base for the slightly enlarged mudbrick foundations of the Phase 4 temple. The Phase 4 temple walls, in turn, were left standing at a consistent height of five brick courses on top of which the stone foundations for the Phase 3 temple were laid (see Falconer and Fall 2006: figure 3.3). This pattern of intentional leveling continues in Phases 3 and 2, and extends to Tell el-Hayyat's domestic structures as well, such that the entire settlement appears to have been rebuilt en masse in each of the intervals between Phases 5 and 2.

Ceramics and Ceramic Chronologies
Tell Abu en-Ni'aj The most influential interpretive paradigm for Early Bronze IV ceramics stems from Dever's pottery "families" (e.g. , 1970a, 1973, 1980), in which site assemblages (primarily from cemeteries) are grouped according to vessel form and decoration both geographically and Figure 6 Photo of Phase 4 temple in antis at Tell el-Hayyat, facing west. Temple interior has central depression for pedestal, mudbrick curb, and mudbrick altar in northeast corner. Door framed by buttresses, with standing stones surrounding northern buttress; temple enclosure wall to south.
temporally. An associated tripartite chronology of Early Bronze IV A, B and C was predicted to cover 200-300 years at the end of the third millennium BC (Dever 1995) in accordance with a traditional Early Bronze IV time frame. Tell Abu en-Ni'aj provides the prime example of a stratified settlement in the region ascribed to Dever's North Central (NC) family, which he placed in Early Bronze IV B-C (Dever 1980). Among many noteworthy characteristics, Family NC ceramics feature trickle-painted decoration, especially on bowls and jars, which was suggested as a technique introduced from Syria around the middle of Early Bronze IV (Mazzoni 1985).
The ceramic evidence from Tell Abu en-Ni'aj (summarized in Table 3) includes early assemblages (Phases 7-4) characterized by holemouth jars and cook pots, and hemispherical bowls with simple upright rims (Falconer and Fall 2019: 78-86). Many vessel forms (e.g., everted rim jars with smooth shoulder-neck profiles) adhere to Family NC expectations, in contrast to examples to the south (e.g., jars with sharp shoulder-neck profiles) from Jericho (Kenyon and Holland 1983) and Jebel Qa'aqir (Gitin 1975;Dever 2014). However, a number of similarities are apparent between the early phases at Tell Abu en-Ni'aj and more southerly families J, CH and S in holemouth profiles (Palumbo and Peterman 1993), an emphasis on incised, combed and applique decoration (Cohen 1999), and the use of exterior decorative ridges or grooves (Prag 1974(Prag , 1986.
The later assemblages from Tell Abu en-Ni'aj (Phases 3-1) conform more closely with the characteristics predicted for Family NC, including the more frequent appearance of open bowls and everted rim jars (Falconer and Fall 2019: 86-101). The frequency of tricklepainting shows its greatest increase between Phases 6 and 5, continues to rise thereafter, and stands in contrast to the predominance of slipped, burnished and rilled decorative techniques seen in more southerly Early Bronze IV families TR and J, for example at Tell Iktanu (Prag 1974), Bab edh-Dhra' Schaub 1978, 2003) and Khirbat Iskandar (Richard et al. 2010). Phase 3-1 typological parallels emerge most clearly from northerly mortuary assemblages at Tiwal esh-Sharqi (Helms 1983), el-Hammeh (Wightman 1988), el-Husn (Harding and Isserlin 1953), Megiddo (Guy 1938), Beth She'an (Oren 1973) and Tiberias (Tsaferis 1968), as well as the settlement at Tell Um Hammad (Helms 1986;Kennedy 2015). As an additional chronological note, folded envelope ledge handles, the most frequent handle type through all seven phases at Tell Abu en-Ni'aj, do not develop into the vestigial forms expected later in Early Bronze IV (e.g., in Family S; Dever 1970bDever , 1980. While prior assignment of Tell Abu en-Ni'aj to Family NC implied its habitation primarily in Early Bronze IV B-C, the mixture of traits associated with more southerly families in Phases 7-4 suggests its founding earlier in the period (i.e., in Early Bronze IVA), while the lack of some late traits connotes abandonment before the end of Early Bronze IV. Based on traditional pottery chronologies, this evidence would predict occupation within a time frame starting after the beginning of Early Bronze IV (i.e., after ca. 2300/2200 BC) and ending before the conclusion of this period (i.e., before ca. 2000 BC).

Tell el-Hayyat
Among the ceramic evidence from Tell el-Hayyat (summarized in Table 4), Phase 6 deposits at the base of the tell produced exclusively Early Bronze IV pottery typified by hand-built construction, fine grained tempering, and trickle painting as its primary decorative motif, particularly for cups and bowls (Falconer and Fall 2006: 44-46). This assemblage accords Short-necked cookpots: simple everted rounded or tapered rims. Holemouth cookpots: flattened or internally beveled rims, sometimes with internal or external projections, rims commonly incised or thumb-impressed. Shift from short-necked majorities (70-60% in each phase) in Phases 7 and 6 to holemouth majorities (65-85% in each phase) in Phases 5 and 4. Noteworthy characteristics: Combing and incision more frequent than painting (mostly trickle-painting); combing and incision common on bowls (both deep and open) and holemouth jars (including herringbone incision in Phases 6 and 5); applique more common than in Phases 3-1; amphoriskoi (often trickle-painted) in Phases 5 and 4; envelope ledge handles (largely on jars) are most common handle form (50-80% in each phase); small numbers of four-spouted lamps; unusual seven-spouted lamp in Phase 7.
best with Dever's families J, N and NC, which have been assigned to Early Bronze IV B-C (Dever 1980). The pottery from stratigraphically contiguous Phase 5 represents a very early Middle Bronze I assemblage characterized by many vessels whose form, fabric and/or decoration combine elements found in classic Early Bronze IV and Middle Bronze I assemblages. In this sense, these vessels may be considered typologically and chronologically transitional. For example, a globular jar from Tell el-Hayyat Phase 5 (e.g., Falconer and Fall 2006: fig. 4.2: g) has a fabric that accords with most of the site's Middle Bronze Age repertoire, while its thickened rim, exterior rilling, and elaborate incision and painting closely parallel those of a jar type found through the final phases of Early Bronze IV at Tell Umm Hammad al-Gharbi (Kennedy 2015). Likewise, Phase 5 bowls commonly combine classic Middle Bronze Age carinated forms with attributes normally found in Early Bronze IV bowls, including slightly incurved rims, external rilling and hand-built construction (see discussion in Falconer and Fall 2006: 46-49, fig. 4.2:c, h, fig. 4.3). The transitional nature of Phase 5 cookpots is best illustrated by coarse-tempered, bell-shaped vessels with thumb-impressed ledge handles (e.g., Falconer and Fall 2006: fig. 4.2: h), which lie typologically between the ledge handles and holemouth forms of Early Bronze IV and the flat-bottomed, straight-sided cookpots of Middle Bronze I. Thus, in light of its transitional pottery and its stratification immediately below the Middle Bronze I deposits in Phase 4, the occupational date of Tell el-Hayyat Phase 5 is ascribed most prudently to very early Middle Bronze I.
Phase 4 pottery provides a classic Middle Bronze I assemblage across a range of functional types, including bowls, juglets, jars and straight-sided cooking pots (Falconer and Fall 2006: 49-52). Vessel forms are consistent with those found in Middle Bronze I assemblages at a variety of Levantine sites, particularly at Jericho (Kenyon and Holland 1982), Lachish (Singer-Avitz 2004), Nahariya (Ben-Dor 1950 and Pre-Palace Aphek (Beck 2000). The Phase 3 assemblage continues many Middle Bronze I vessel forms, some of which develop in accordance with early Middle Bronze II examples elsewhere (Falconer and Fall 2006: 52-57). Typological parallels continue from a variety of sites, including Jericho (Kenyon and Holland 1982), Pre-Palace Aphek (Beck 2000) and Lachish (Singer-Avitz 2004), plus additional parallels indicative of Middle Bronze II from Shechem (Cole 1984), Megiddo (Ilan 2000) and locally at Tel Beth She'an (Maeir 2010).

AMS Results and Bayesian Models
Tell Abu en-Ni'aj Radiocarbon determinations from Tell Abu en-Ni'aj now include 25 AMS ages distributed through Phases 7-1 (see Table 1). Phase 7 at Tell Abu en-Ni'aj is represented by a single AMS date (AA-94177) that models about 2500 cal BC. Phase 1 now provides four new radiocarbon ages, all of which model in the 23rd century cal BC. Bayesian analysis coupled with OxCal's Outlier Analysis detected three radiocarbon determinations from Tell Abu en-Ni'aj that met the 5% threshold in the Outlier Index [O:5/5]: AA-90075 in Phase 6, AA-90071 in Phase 5 and AA-90067 in Phase 4. These samples produce ages sufficiently early to indicate that they most likely represent residual material. All other radiocarbon determinations generate outlier indices < 5% and are thus treated as representative of their respective stratigraphic contexts. In sum, our Bayesian model indicates that occupation at Tell Abu en-Ni'aj most likely began by 2500 cal BC (based on the start boundary for Phase 7), concluded by 2200 cal BC (based on the end boundary for Phase 1), and spanned up to 300 years (Figure 7).

Tell el-Hayyat
The recent addition of 13 new dates generated by the Micadas AMS at Groningen now raises the number of 14 C ages from Tell el-Hayyat to 31, covering Phases 6-1 (see Table 2). Phase 6 provides three ages that model about 1900 cal BC. Phase 1 is represented by a single age (AA-108793) that models in the late 18th or early 17th century cal BC. OxCal's Outlier Analysis detected two outliers with levels > 5%: AA-1239 in Phase 5 and GrM-11953 in Phase 2. The Outlier Index reaches 100/5 for these samples, showing both to be anomalously recent and therefore intrusive. AA-1237 reaches the 5% threshold [O:5/5] and therefore might also be intrusive. All other dates have index scores well below 5, usually between 0 and 1. These results indicate the earliest deposition at Tell el-Hayyat by 1900 cal BC (based on the Phase 6 start boundary), with early Middle Bronze I occupation starting just after 1900 cal BC (based on the modeled transition between Phases 6 and 5). In conjunction with excavated ceramic evidence, our model suggests a Middle Bronze I/II transition in Phase 3 shortly after 1800 cal BC, and a Middle Bronze II/III transition in Phase 2 between 1800 and 1700 cal BC. These results indicate that Tell el-Hayyat was abandoned by 1600 cal BC (based on the Phase 1 end boundary), following occupation over approximately 300 years (Figure 8).

Occupational History of Tell Abu en-Ni'aj and Tell el-Hayyat
The excavated evidence from Tell Abu en-Ni'aj portrays a continuously occupied Early Bronze IV agrarian village comprised of mudbrick room blocks built and rebuilt incrementally around a network of earthen and sherd-paved streets. Major iterations of this architectural plan are represented by seven stratified occupational phases and their associated material culture. Based on the affinities of its stratified ceramic assemblages with Dever's pottery families and their chronological assignments, habitation at Tell Abu en-Ni'aj would be predicted to commence after the start of Early Bronze IVA and conclude before the end of Early Bronze IVC.
Bayesian analysis coordinates data from 25 calibrated AMS ages to build a probabilistic model of habitation through seven major phases of roughly comparable lengths (estimated at about 25-40 years each; Table 5), which reaffirms the architectural evidence of fairly steady incremental remodeling within phases and cumulative rebuilding between phases. Our optimal model provides conservative estimates for the founding of Tell Abu en-Ni'aj by about 2500 cal BC and its abandonment by 2200 cal BC. Hence, the exclusively Early Bronze IV ceramic repertoire of Tell Abu en-Ni'aj, pottery-based estimation of habitation beginning in the first portion of this period, and both individual and modeled 14 C ages leave little doubt that this village was founded toward the start of Early Bronze IV, no later than 2500 cal BC. Evidence for the latest occupation of Tell Abu en-Ni'aj includes a modeled Phase 1 end boundary prior to 2200 cal BC. Thus, prior pottery-based inference of Tell Abu en-Ni'aj's latest occupation before the end of Early Bronze IV, plus new calibrated radiocarbon determinations, suggest village abandonment by 2200 cal BC.
Tell el-Hayyat contrasts with Tell Abu en-Ni'aj by virtue of its disjunct individual mudbrick structures, its temples as persistent architectural centerpieces, its phase-by-phase rebuilding en masse, and its more variable phase lengths (ranging between about 20 and 70 years each; Table 5). The less consistent phase lengths at Tell el-Hayyat may accord with less incremental, more corporate reconstruction than seen at Tell Abu en-Ni'aj, suggesting the possible influence of central authority (as manifested by the Tell el-Hayyat temples). The relatively modest lengths of Phases 4 and 3 are tempered by a much longer Phase 2, which featured the largest and most elaborate temple. The evidence from Phase 1 includes the fragmentary remains of stone-founded structures on the crown of the tell, but no longer incorporates a village temple.
Based on our Bayesian model, Phase 6 started just before 1900 cal BC and ended shortly after cal 1900 BC. The calibrated dates in immediately subsequent Phase 5 model between 1900 and 1800 cal BC. The distinctly transitional pottery repertoire of Phase 5 marks it as very early Middle Bronze I, which we would expect traditionally to begin about 2000 cal BC.    (Thalmann 2006: 230;, and six ages from Tell Fadous-Kfarabida, in Lebanon south of Tell Arqa, the latest of which has a calibrated median of 2243 cal BC (Genz 2014;Höflmayer et al. 2014). At this point, the best radiocarbon-dated evidence for late Early Bronze IV settlement is limited to the Northern Levant. In contrast, the radiocarbon record for the Southern Levant, and perhaps as far north as Tell Fadous-Kfarabida, features an intriguing dearth of seed-dated evidence later than about 2250 cal BC, and raises the possibility of more common abandonment of southern Levantine villages after that date (see also discussion in D'Andrea and Vacca 2015).

The Early Bronze IV/Middle Bronze Age Interface
Roughly three centuries of Early Bronze IV habitation at Tell Abu en-Ni'aj were followed by a hiatus prior to the initial settlement of Tell el-Hayyat. Based on their temporally distinctive assemblages and their stratigraphic contiguity, a suite of nine radiocarbon ages from Phases 6 and 5 at Tell el-Hayyat offers an unparalleled opportunity to clarify the beginning of the Middle Bronze Age locally, with potential regional implications.  (Garfinkel and Cohen 2007;Bourke et al. 2009;Fall et al. 2019).
The Pella date is followed by four more ages significantly later in the Middle Bronze Age (Bourke and Zoppi 2007;Bourke et al. 2009), while the age from Gesher reflects a standalone charcoal sample, and the seed date from Zahrat adh-Dhra' 1 is a single disjunct age from a settlement occupied discontinuously primarily later in the Middle Bronze Age  A Middle Bronze I/II transition about 1800 cal BC is supported by a detailed comparative assessment of architecture, material culture and AMS seed ages from Tel Ifshar and Tell el-Burak (Höflmayer et al. 2016b), while a Middle Bronze II/III transition before 1700 cal BC is bolstered by a comparative assessment of multiple lines of evidence from these sites and Tel Kabri on the coast of northern Israel (Höflmayer et al. 2016a). The radiocarbon record from Jericho continues later in the Middle Bronze Age (Bruins andvan der Plicht 1995, 2003;Lombardo and Piloto 2000;Nigro et al. 2019), toward a conclusion in the late 17th century cal BC (which would align with Tell el-Hayyat) or in the 16th century cal BC (suggesting asynchronism along the Jordan Rift).
In sum, the chronological inferences stemming from Tell el-Hayyat in conjunction with sites stretching across the Southern Levant and Lebanon provide an empirical basis for beginning the Middle Bronze Age in some locales 50-100 years later than assumed traditionally, an end for some Middle Bronze Age settlements up to a century earlier (Höflmayer 2019), and relocated transitions between the constituent subperiods within this era of rejuvenated town life in the Southern Levant.

CONCLUSIONS
This study presents and interprets new Bayesian models for Tell Abu en-Ni'aj, based on four new Phase 1 samples and their AMS ages for the final occupation of this Early Bronze IV settlement, and for Tell el-Hayyat, now bolstered by 13 new 14 C ages from Early Bronze IV Phase 6 and Phases 3 and 2 in the later Middle Bronze Age. The evidence from Tell Abu en-Ni'aj demonstrates that Early Bronze IV settlement in the Jordan Valley started at least as early as 2500 cal BC, and strengthens the argument for a higher Early Bronze IV chronology for the Southern Levant with a start date multiple centuries earlier than assumed traditionally, at least in the Jordan Valley. Our modeling shows that occupation at Tell Abu en-Ni'aj ended by 2200 cal BC, which may hint at a larger pattern of more pervasive village abandonment across the Southern Levant late in Early Bronze IV.
Although the basal stratum at Tell el-Hayyat (Phase 6) is characterized by solely Early Bronze IV sherds, our new modeling shows a substantial gap between the abandonment of Tell Abu en-Ni'aj and the founding of settlement at Tell el-Hayyat. Our modeled start date for very early Middle Bronze Age occupation at Tell el-Hayyat (Phase 5) falls after 1900 cal BC, and represents a significant departure from the standard beginning date for the Middle Bronze Age in the Southern Levant. Bayesian modeling for the later occupation of Tell el-Hayyat, which incorporates newly-augmented suites of AMS ages from Phases 3 and 2, establishes occupation at Tell el-Hayyat by about 1900 cal BC, and suggests the abandonment of this village in Middle Bronze III before 1600 cal BC, which again represents a substantially early departure from standard dates, which presuppose the end of the Middle Bronze Age ca. 1550/1500 cal BC based on conventional synchronization with Egypt.
In tandem, these Bayesian models and their integration with emerging regional radiocarbon chronologies strengthen a set of inferences for Bronze Age settlement in the Jordan Valley and more broadly across the Southern Levant. These inferences indicate (1) further support for a high 14 C chronology for Early Bronze IV, (2) abandonment at Tell Abu en-Ni'aj and diminished regional sedentary settlement late in Early Bronze IV, (3) a start date for the Middle Bronze Age more likely at about 1900 cal BC rather than 2000 cal BC, (4) Middle Bronze I/II and II/III transitions around 1800 cal BC and 1750 cal BC, respectively, and (5) an earlier than expected end date in Middle Bronze III for the abandonment of Tell el-Hayyat by 1600 cal BC. All of these inferences illuminate the need for independent Levantine chronologies and societal interpretations through the Early and Middle Bronze Ages.