Salvage excavations at the Neolithic site of Amqa were directed by M. Shemer on behalf of the Israel Antiquities Authority (Permit n^o A7696/2016) as an extension of a larger salvage excavation directed by Y. Alexandre (Permit no A7380/2015). The site plan and sections were drawn by R. Mishayev and R. Liran. Studio photography of selected finds by C. Amit and D. Gazit. Flint drawings courtesy of M. Smilansky. The identification of the green mineral in the Amqa assemblage as nephrite is based on XRF and FTIR examinations conducted by Y. Maor in the Analytic Laboratory of the Israel Antiquities Authorities.

Introduction

1The transition between the Pre-Pottery Neolithic A (PPNA) and Pre-Pottery Neolithic B (PPNB) in the Mediterranean zone of the southern Levant has always seemed somewhat obscure. The Early phase of the PPNB has been poorly documented for many years, and associated sites are few. On the one hand, there are clear differences in site settings between the PPNA and EPPNB, a trait that has been regarded as an indication of a chrono-cultural break (and see discussions in Goring-Morris et al. 2009; Birkenfeld 2018: 215–216). In the aspect of lithic technology, the appearance of the bidirectional core technology in the EPPNB was seen as representing a clear break from previous lithic traditions, further supporting the notion of a cultural “gap” (Bar-Yosef 2007, and references therein). On the other hand, these attributes seemed confined to southern Levantine sites. In comparison, north Levantine sites, particularly along the Middle Euphrates, showed a continuous, gradual, and uninterrupted PPNA–EPPNB transition (Kuijt 2003; Akkermans 2004; Edwards et al. 2004; Goring-Morris and Belfer-Cohen 2016 and reference therein), assuming primacy in the emergence of the PPNB “way of life” (e.g., Cauvin 1994; Gopher 1994; Kuijt and Goring-Morris 2002; Gopher et al. 2013; Kuijt 2003; Edwards 2016). These differences were used to support models of northern development that spread southward.

2The discovery of numerous EPPNB sites in the southern Levant in the past decades, alongside ongoing research and publications of previously known locales, has contributed significantly to the growing mass of available data on the EPPNB in the Mediterranean zone. Our current database is geographically diverse, with main sites including Tell Aswad in the Damascus Basin, Tell Qarassa in the Leja Plain, Mujahiya in the Golan Heights, Horvat Galil, Kfar HaHoresh and Aḥihud in the Galilee, Nahal Oren and Sefunim in the Coastal Plain, and Motza in the Judean hills (Ronen 1984; Gopher 1990, 1997; Cauvin 1994; Garrard et al. 1994; Stordeur 2003; Khalaily et al. 2007; Goring-Morris et al. 2008; Ibáñez et al. 2010; Stordeur et al. 2010; Paz and Vardi 2014; Vardi et al. this volume). These exhibit various sizes and functions, including settlements and special activity sites, and indicate much more complex and intricate dynamics in the region during the EPPNB than previously assumed. Together with an ever-expanding bank of ¹⁴C dates, these new data suggest a more parallel, perhaps even simultaneous, development of PPNB traditions in the northern and southern regions of the Levant.

3This paper presents a newly discovered EPPNB locale at the site of Amqa in the upper Galilee, Israel. It details the findings from the site and discusses their contribution to our understanding of the EPPNB from a regional perspective.

The upper Galilee

4Located in the north of modern-day Israel, the upper Galilee is a well-defined geographic unit: To the west, it is bordered by the Mediterranean Sea; to the north, by the Lebanon mountains; and to the east by the upper Jordan Valley, the Sea of Galilee and the Golan Heights. To the south, the wide Beit Hakerem Valley marks the border between the upper and lower Galilee (fig. 1). In general, the upper Galilee is characterized by a diversified, mountainous landscape, with peaks reaching up to ca. 1,200 m asl. This landscape, as well as the main streams that transverse it from east to west, were created by extensive tectonic faulting activity, which is also the source of the diversified lithology of the region, comprised of chalk, lime, dolomite, and basalt.

5The climate in the region is typical Mediterranean, with annual precipitation ranging between 700 and 1,000 mm. This, and the relative abundance of natural springs, has given rise to plentiful vegetation—mainly Mediterranean forests. Such forests are characterized by Tabor-oak woodlands in lower elevations (ca. 400–500 m asl), which are replaced by Atlantic terebinth and wild almond woodlands (Pistacia atlantica/Amygdalus korschinskii) and then by common oak and Palestine terebinth (Quercus calliprinos/Pistacia palaestina) maquis as the elevation rises. Cyprus oak (Quercus boissieri) is present on the highest summits alongside evergreen oak (Zohary 1973). During the Younger Dryas and the end of the Levantine Epipaleolithic period, palynological records indicate a decrease in temperatures, together with an increase in precipitation and a significantly less distinct wet-dry seasonality (Hartman et al. 2016; Langgut et al. 2021). This started to shift towards the present-day seasonal contrast around ca. 11,500 cal BP with the beginning of the Holocene and the onset of the Neolithic cultures in the region (Cheddadi and Khater 2016).

6The archaeological evidence for the Early PPNB in the upper Galilee is sparse and somewhat patchy. For many years, the site of Horvat Galil, excavated in the late 1980s, was the primary data source in the region (Hershkovitz and Gopher 1988; Gopher 1989b, 1997). Located ca. 13 km east of the Mediterranean shoreline (fig. 1), Horvat Galil was interpreted as a settlement site exhibiting “typical” rectangular PPNB architecture and included human burials and a rich, diversified material culture.

7Other EPPNB sites in the upper Galilee include Nahal Betzet I, where a small test excavation exposed a multi-phased PPNB settlement, which was suggested to have been initially settled during the EPPNB (Gopher 1989b). Another suggested locality is Horvat Turit, where a PPNA/EPPNB component was observed during a survey (fig. 1; Birkenfeld 2018; Barzilai and Getzov pers. comm.).

8A significant addition to the study of EPPNB in the upper Galilee is the discovery and consequent salvage excavations of the site of Aḥihud in 2013. The site is located ca. 10 km east of the Mediterranean shoreline, on a small hill within the Wadi Hilazon floodplain (fig. 1). EPPNB remains were found in a vast area, including various installations, plaster floor segments, and human burials. The rich material culture indicated long-distance connections alongside the well-established exploitation of domesticated legumes (Paz and Vardi 2014; Caracuta et al. 2017; Vardi et al. this volume).

The site of Amqa

9The site of Amqa is located in the western part of the upper Galilee, ca. 7 km east of the current Mediterranean shoreline (figs. 1–2), however, during the PPN, the shoreline was somewhat further to the west (Sivan et al. 2001; Galili et al. 2020). The site is situated on a high hill, ca. 110 m above sea level. The northern slopes of the hill drop steeply into the Beit Ha’emek Wadi, which is nowadays a seasonal stream. The hill marks the highest topographic locality in the immediate surroundings of the site and a vantage point from which, on a clear day, a full view of the Akko Valley is visible, reaching the fort of Yodfat in the east and the Mediterranean Sea in the west. It is composed of chalky bedrock, which is often exposed on the surface.

10The EPPNB of Amqa is the oldest in a sequence of occupation layers identified on-site, of which most prominent are those of the Hellenistic, Roman, Byzantine, and Early Islamic periods (Alexandre and Cohen in press). The nowadays elongated shape of the hill summit is probably the result of earthwork and bedrock exploitation practiced during those periods and during the 20^th century (Alexandre and Cohen in press). Typical products of EPPNB lithic industries were identified in isolated localities, primarily associated with natural depressions in the bedrock. These depressions, reaching up to one meter in depth, probably protected the ancient sediments during the earthworks and construction activities of the later periods. EPPNB artifacts are primarily associated with a distinct change in sediments- from loose, dark brown to compact yellowish. The main concentration of EPPNB artifacts was found in a single bedrock depression, encompassing an area of ca. 25 m² at the northwestern part of the site, on the edge of the steep slope (figs. 3–4). A dense midden layer was exposed near the bottom of the depression, under 0.4–0.6 m accumulation of characteristic yellowish sediments, and ca. 0.8 m below the surface (fig. 4).

11Activity during the later periods was evident on the bedrock outlining the depression: a rectangular structure was built during the Byzantine period on the outer edges of the bedrock depression, and a foundation trench was dug into the EPPNB sediments at its northern edge. Underlying quarrying marks associated with the Hellenistic period were also identified (Alexandre and Cohen in press). Additional wall fragments indicate activity during the Roman period and in the early 20^th century. Nevertheless, “clean”, and undisturbed EPPNB contexts were excavated in the center and southern parts of the depression, yielding rich assemblages in a pristine state of preservation. In addition to a typical flint industry, the assemblage included obsidian artifacts, bifacial tools made of exotic raw materials, and a few bone tools.

Methodology

12Excavations at the site of Amqa were conducted using a 5 m² grid. Upon reaching Neolithic remains, these were further divided into 2 m² sub-squares and excavated in 5–10 cm spits. All sediments were dry sieved using a 5 mm mesh. Artifacts were processed in the lab, recording technological and typological attributes for each of the cores, tools, bidirectional blades, and core trimming elements. A few artifacts, made of green stone, were further subjected to XRF examination in the Analytic Laboratories of the Israel Antiquities Authority to determine their mineralogical composition. Flint and obsidian general tools classification was based on Crowfoot-Payne (1983) and Gopher (1985, 1989a, 1994), while detailed typology of bifacial tools was based on Barkai (2005). Sickle blades typology is based on Brailovsky-Rokser (2015). Products of the bidirectional industry were defined and sorted into technological categories following Barzilai (2010).

Results of the lithic analysis

13The flint assemblage from Amqa contains a total of 2022 items (table 1). Overall, artifacts present a good state of preservation, showing sharp edges and prominent scar ridges. Raw materials are varied and include both local types, such as grey-brown flint typical of the Timrat/Maresha/Adulam Formations (Ekshtain et al. 2014) and non-local pink and purple flint types (perhaps brought from further away, possibly the Transjordan; Quintero 1996; Delage 2007; Rollefson et al. 2007) as well as beige flint type from unknown origins (fig. 5). About 20% of the analyzed artifacts showed pronounced luster and a slight change of coloration, possibly indicating that heat treatment was applied for some of the raw materials (after Crowfoot-Payne 1983; Nadel 1989, 1997: 72–75; Yegorov et al. 2020, 2022).

Blanks

14The basic breakdown of the assemblage shows the presence of all debitage categories, indicating on-site knapping activities. Flakes are the most common among the debitage, alongside a notable bladelet component (table 1). High frequencies of primary elements (ca. 18% of the debitage) imply the on-site execution of preliminary reduction stages, including core shaping and decortication. The abundance of spalls (e.g., burin spalls, chanfrein spalls, tranchet spalls, and bifacial thinning spalls) indicate that the final stages of blank modification and recycling were also conducted on-site.

15Despite the dominance of flakes, the assemblage from Amqa shows a distinct preference for the use of laminar items (i.e., blades and bladelets), as demonstrated by blank selection (table 2). Products of the bidirectional industry form ca. 8% of the assemblage, 64% of them are tools fashioned on bi-­ directional blade blanks. Notably, most bi-directional products were made of non-local exotic, pink-purple, or beige flint or indicated possible heat treatment. The presence of characteristic spalls indicates the shaping and maintenance of bifacial tools on site, including the use of tranchet removals (table 1). Gloss-bearing burin and chanfrein spalls show the occasional reuse and recycling of discarded sickle blades.

Cores

16The assemblage includes 42 cores, constituting ca. 2% of the flint artifacts from Amqa. Cores are relatively small, with an average length, width, and thickness of 3 × 2.6 × 2.3 cm (length range: 1.70–5.5 cm; width range: 1.50–4.80 cm; thickness range: 1.10–4.20 cm). Intensive use is implied by the abundance of exhausted cores (n = 16; 38% of the cores) and cores that were abandoned due to a deep hinge on the removal surface (n = 16; 38% of the cores).

17Single platform bladelet cores dominate the category (ca. 43% of the cores), demonstrating bladelet production from prismatic or sub-circular removal surfaces. Bladelets were also produced from bi-directional cores using two opposed striking platforms; however, this method seems less common (table 3).

18Flake cores indicate the use of various methods to produce non-laminar artifacts, of which the most frequent is the use of multiple platform cores (ca. 12% of the cores), followed by a single platform (ca. 7%) and two, perpendicular or opposed platforms (ca. 5% for each category). Core: flake ratio is 1:20.6, including debitage and tool blanks (n = 247).

19Notably, none of the cores found in the assemblage could be ascribed to the bi-directional blade industry. Therefore, it seems that bi-directional blade blanks and tools were introduced to the site from elsewhere as end products.

Tools (table 4, figs. 6–10)

20Tool frequencies are relatively high, constituting ca. 15% of the assemblage. About half of the tools are “formal” tool types or tools made on “formal” bi-directional blade blanks, while the rest are non-formal, ad hoc types. Products of the bi-directional blade industry constitute ca. 35% of the tool blanks. They dominate in several of the tool categories, with the highest frequencies shown in the sickle blades category (ca. 85%), followed by the retouched blades, projectile points, and burins categories (ca. 63%, 53%, and 52% respectively). Thus, indicating a preference for the use of these items both in “formal” and “non-formal” tool categories.

21The projectile point category (table 4; n = 15, 5.1%) is comprised of several subtypes, the majority are typical EPPNB Helwan points, while the rest are earlier types, some exhibiting morphological properties characteristic of projectile points from distant regions:

22Khiam points (n = 2, 0.7%; fig. 6.1–6.2); one is relatively small (20 × 9 × 2.50 mm), made of homogeneous beige flint. It was fashioned on a blade or a bladelet; the tip and the medial portion were shaped by finely applied abrupt retouch. The tang is stubby with a straight bottom, fashioned by pressure retouch and delineated by abrupt retouch at the final stage. The notches were created by bifacial retouch (fig. 6.1).

23The second point is missing its tip, showing signs of damage due to fire exposure. Yet, it is larger than the first one (minimal length 32 mm, 12 × 3.50 mm). It was made on a blade/bladelet of homogeneous brown flint. The lateral edges were shaped by a combination of semi-abrupt and pressure retouch, creating finely denticulate edges. The notches and tang are morphologically similar to the first point (fig. 6.2).

24These two Khiam points differ from the “typical” Khiam point typology in the sense of Gopher (1994), more resembling the Helwan points in their overall fashioning. However, similar items were reported from the south-Levantine EPPNB sites of Motza (Khalaily et al. 2007: fig. 12.12–12.14) and Nesher-Ramla (Ulman and Brailovsky 2020: fig. 5.5.4, where a fragmented small point was perhaps mistakenly identified as an intrusive PN point rather than an integrative Khiam point), where they were ascribed as Khiam points; as well as south-Levantine PPNA sites such as Gilgal 1 (Dag et al. 2010: fig. 5.13.1), and the recently excavated site of Barkai (Brailovsky-Rokser personal observations).

25“Foliate”/“Nemrik” point (n = 1, 0.30%; fig. 7.1) made on a bladelet of homogeneous brown, semi-translucent lustrous flint (heat treated?), with rounded basal modification, created by bifacially applied retouch. The tip was gently modified by a non-invasive, dorsal, fine retouch, mainly on the left lateral edge. It is not symmetrical and perhaps unfinished, measuring 32 × 12 × 3 mm. Such points are typical of the PPNA in the northern Levant and were reported mainly from sites in northern Iraq, Anatolia, and Cyprus (Kozłowsli 1994; Manning et al. 2010: fig. 7; Miyake et al. 2012: fig. 7; Maeda 2018; Altınbilek-Algül et al. 2022; Maeda et al. 2022). An intriguing appearance of these points was also noted in the Epi-Paleolithic Harifian phase of Abu Madi I in southern Sinai (Pomerantz-Grinblat 2014).

26Helwan points (n = 6, 2%; fig. 6.3–6.4) exhibit variability in shape, size, and fashioning manner. The first item was intensively shaped by bifacial pressure retouch and has two pairs of notches on its medial part (fig. 6.4). The wings are very pronounced, and the tang is stubby and resembles in shape the tangs of the Khiam points described above. Stubby wide tangs are typical of “Aswad points”, a more northern variant of Helwan points (Cauvin 1974; Arimura 2007 and references therein): It was fashioned on fine-grained homogeneous flint, perhaps heat treated; the blank is unidentifiable as the retouch removed all scars. It was later recycled into a drill; the tip was re-shaped and narrowed down, and it exhibits discoloration due to friction. It measures 39 × 14 × 3 mm.

27The second point is missing its distal end, perhaps due to the impact. Nevertheless, it was quite long (minimal length 33 mm, 13 × 4.50 mm). It was fashioned on a bi-directional blade of fine-grained, very lustrous pale beige flint (the luster is probably a result of heat treatment). It has two pairs of notches: One pair on the medial part and another at the distal end, close to the tip, probably serving as the failure point where the tip broke. This item exhibits far less investment in its modification than the first item. The lateral edges were not retouched; bifacial retouch was applied only for notch shaping. The tang is also stubby and very wide; it was shaped by a bifacial retouch combined with a dorsally applied semi-abrupt retouch truncating the proximal end.

28The third item was made on a bi-directional blade; it was burnt, so the raw material type is unidentifiable. It has one pair of bifacially shaped notches in its medial section, and the tip was fashioned by extensive bifacial pressure retouch (fig. 6.3). It measures 32 × 11 × 3 mm.

29The fourth item was made on a bi-directional blade. It is missing both extremities due to fire exposure, but it seems to have had two pairs of notches created by bifacial retouch, which survived only on the right edge. Only the right-wing survived from the entire tang portion; it was also shaped by bifacial retouch. The minimal width and thickness of this item are 15 × 3.50 mm, respectively.

30The fifth item was fashioned on a bi-directional blade of lustrous homogeneous beige flint. Its distal part is missing. No notches are evident on the item, and it is possible that they were placed on the missing portion of the tool. This item was classified as a Helwan point due to its stylistic similarity to the rest of the points in this category. The lateral edges were shaped by inverse pressure retouch, removing the gloss. The tang is very wide and was fashioned by bifacial retouch. This item is quite wide: 17.50 mm. Its thickness is 4 mm, and its length is a minimum of 27 mm.

31The sixth item was also fashioned on a bi-directional blade knapped of fine-grained lustrous flint of beige-pinkish shades with purple veins. The distal part is broken at the point where the single pair of notches, shaped by bifacial retouch, was placed. Both lateral edges were modified: the right by fine retouch and the left by invasive, inverse pressure retouch. The tang differs from the rest of the points, Khiam and Helwan both. It is slightly longer and somewhat narrower and was shaped using bifacial pressure retouch. This item is quite large, with a minimum length of 31 mm, 16.50 mm in width, and 3.50 mm in thickness.

32Hollow-based point in preparation (n = 1, 0.30%; fig. 7.2). This unfinished item was made on a bladelet, with some cortex retained on its proximal end. It was probably broken during initial shaping and, therefore, discarded unfinished. Only its right edge was modified, using bifacial retouch. No notches are present. The item may have been abandoned before notches were created, or perhaps notches were never intended. It is, therefore, unclear whether this was intended to be a Khiam or Helwan point. The tang is very distinct; it is hollow-based, modified by fine abrupt retouch. It is quite small (minimal length of 30 mm, width 10.50 mm, and 3.50 mm in thickness). Hollow-based points were found in Nevalı Çori (Schmidt 1994), Shubayqa 6 (Richter et al. 2016: fig. 6), and Abu Madi I (Pomerantz-Grinblat 2014)

33Indeterminate projectile fragments (n = 5) include two tips and three medial sections. All were fashioned on bi-­directional blades. The two tips were retouched on their ventral face by extensive pressure retouch, on one occasion applied semi-abruptly.

34Sickle blades comprise 6.7% of the tool assemblage (n = 20). Within this category, noticeable are high frequencies of reaping knives (i.e., complete blades or long segments, with a shaped tang and gloss; fig. 8.10–8.11), which form almost a third of the items (n = 6). Four of these are finely denticulated, and two are plain. The tangs of the reaping knives are varied in shape and fashioning methods: Rounded, pointed, square, and irregularly shaped tangs are present, fashioned by obverse-abrupt or inverse pressure retouch.

35Eight additional blades displayed characteristic gloss, of which four are plain (i.e., bear no marks of working edge modification), three show fine denticulation, and one bears light retouch (table 3). As none of these blades show signs of truncation, it is impossible to determine whether they represent sickle segments or fragments of reaping knives (i.e., fig. 8.8). A single item presented a shaped back. As this kind of shaping is uncommon for reaping knives, this item is considered a sickle segment. Five additional blades were too fragmented to allow further classification.

36The Amqa sickle blades and reaping knives are mostly single-edged (i.e., presenting characteristic gloss only on one edge). The denticulations are ultra-fine and medium-fine (Brailovsky-Rokser and Goring-Morris 2019) and were mainly achieved by obverse fine retouch.

37The bifacial tools category includes 12 items (4%): four “leaf-shaped” axes (three bear a tranchet scar; fig. 8.12–8.13), two chisels, and one adze. Five bifacial fragments could not be further classified. All bifacial tools in the Amqa assemblage are quite small and flat, complete items reaching maximal measurements of 74 mm in length, 26 mm in width, and 13 mm in thickness.

38Of note is a chisel that was fashioned on a bi-directional ridge blade, using the ridge as part of the initial shaping and modifying the item mainly ventrally. Another notable item is an adze fashioned on a large flake or blade of fine-grained, homogeneous, pink-purple flint, with some cortex left at its central part. The shaping involved a combination of bifacial pressure retouch with dorsally applied semi-abrupt retouch. It is distally broken and shows signs of re-use after it was broken. Thus, while at its final stage of use, this tool was indeed used as an adze, it is possible that initially, this was a large bifacial knife or dagger. Such daggers are typically associated with the Late–Final PPNB (Goring-Morris et al. 1994; Gebel et al. 2022; Vardi et al. 2022); however, several earlier examples have been found in Anatolia (Zimmermann 2015 and reference therein).

39Among the perforators (n = 29; 9.80% of the tools), borers are the most common subtype (n = 14; fig. 8.5). They were primarily made on laminar blanks with a preference towards bi-directional blades (n = 6). Awls are also quite common (n = 11), yet these tools were mainly fashioned on flake proportioned blanks (i.e., flakes and CTEs).

40Microlithic tools constitute a relatively large group (n = 25, 8.40%; table 4; fig. 8.1–8.4). Items in this category most often display backing created by abrupt retouch and accompanied by proximal or distal truncation. Alternatively, fine, partial retouch of one of the lateral edges, with no back treatment, was also commonly used (table 4). Of note is the occasional shaping of a concave back (table 4; fig. 8.4).

41Burins (n = 66, 22.20%; figs. 8.6–8.7, 9) present high variability when divided into subtypes (table 4), showing a high tendency for the recycling of discarded tools from various categories—including bifaces (n = 2) and sickle blades (n = 10). Of note is the presence of large dihedral burins (n = 3; fig. 9) made from non-local flint.

42Obsidian artifacts. The obsidian assemblage from Amqa includes 36 artifacts (1.70% of all items in the assemblage, including flint, nephrite, and basalt artifacts), all products of blade/bladelet industries. A subdivision into three main types was conducted based on the artifacts’ color as it appears to the naked eye. Homogeneous grey obsidian (fig. 10.4–10.5) is the most common, comprising ca. 53% of the assemblage, followed by stripped black/grey (ca. 28%; fig. 10.6–10.7) and homogeneous black obsidian (ca. 17%). One item (bladelet) could not be classified into a type class due to heavy patination.

43Bladelets comprise 50% of the obsidian assemblage (n = 18), while blades constitute 8.3% (n = 3). Obsidian tools, i.e., bladelets with fine, partial retouch, comprise 25%. Maintenance items (ridge, side ridge, and overpass bladelets) constitute a total of 14% of the assemblage (5.60%, 2.80%, and 5.60%, respectively). One obsidian fragment could not be classified (2.8% of the assemblage). While the origin of the obsidian from Amqa is yet to be determined, obsidian in the region during the period was suggested to usually derive from central Anatolia, specifically Göllü Dağ (Ibáñez et al. 2016; Birkenfeld 2018: 161–163).

44Other finds. In addition to a rich flint assemblage and a relative abundance of obsidian artifacts, the EPPNB layer in Amqa yielded two bone tools (an awl and a spatula; fig. 10.1–10.2), and a miniature chisel and axe fragment made of a green mineral. XRF and FTIR examinations conducted in the Analytic Laboratory of the Israel Antiquities Authority showed a correlation to the mineral nephrite (fig. 10.8–10.9). Other finds include a polished basalt axe (fig. 10.10) and a marine shell (bearing no clear shaping/perforation marks; fig. 10.3).

45Although composing only a small part of the EPPNB assemblage of Amqa, these items imply contacts, possibly by trade, with various regions. The marine shells and the basalt axe can represent local foraging of uncommon raw materials. Shells could be collected from the Mediterranean shore ca. 7 km to the west of the site. Basalt could be collected from the limited outcrops of Tell Shimron near the Nazareth hills, the Sakhnin or Carmel mountains ca. 10 km to the south or 25 km to the southwest, respectively, or from the richer formations of the Jordan Valley and the Golan Heights, ca. 30 km to the east (Sneh et al. 1998).

46On the other hand, nephrite tools and obsidian artifacts are indicators of long-distance connections, as these raw materials are not naturally found in the southern Levant. The closest significant nephrite deposits known today are located in western Europe or Asia (China and Siberia; Harlow and Sorensen 2005; Hsu et al. 2015), and an additional source was recently suggested in Lawdar, Yemen, in the southwestern edge of the Arabian Peninsula (Al Kindi et al. 2021).

Discussion

Chrono-cultural attribution

47The site of Amqa provides a glimpse into the Early PPNB occupation, cultural dynamics, and exploitation patterns in the upper Galilee. Massive construction and bedrock modification during later occupation phases impede complete discussion of the site’s depositional history and character during the EPPNB. Nevertheless, a basic deduction could be made regarding site function. The material culture is diversified, indicating a wide array of activities. The flint assemblage indicates on-site knapping directed into three main trajectories: 1) Full on-site production of flakes, unidirectional blades, and bladelets from locally available flint; 2) On-site production of bifacial tools from locally available flint; 3) Final shaping, maintenance, and recycling of imported items—mainly bi-directional blade blanks and bifacial tools made of non-local, pink/purple/beige flint.

48The tool assemblage includes types associated with harvesting and woodworking activities, as well as types associated with hunting and carcass processing. While Amqa did not provide a faunal assemblage, the record from other sites in the region, such as Horvat Galil, indicates a hunting-based economy dominated by gazelle, with no evidence for domesticated animals (Gopher 1989b, 1997).

49The presence and use of non-local flint types, basalt, obsidian, and nephrite indicate the existence of a network of trade or social contacts with both near and far regions. These features, bearing some resemblance to the sites of Horvat Galil and Motza (Gopher 1997; Khalaily et al. 2007), led to the ascription of Amqa as a settlement or residential context rather than a site of special activity.

50While absolute chronology was not achieved in Amqa due to poor preservation of charcoal and collagen, techno-typological attributes suggest that the site should be affiliated with the EPPNB. This is supported by the dominance of bi-directional blade blanks among the tools, Helwan points, reaping knives, and tranchet bifaces. The appearance of miniature bifacial tools, often made of green minerals, is also considered typical of PPNB industries. Nevertheless, some elements are more characteristic of earlier industries, typically associated with the preceding PPNA period. These include the presence of Khiam points, a “Foliate”/“Nemrik” point, and a Hollow-based point, in addition to the relatively high frequencies of microlithic tools in the Amqa assemblage. Notably, microlithic tools are quite scarce in most EPPNB sites, often not exceeding 1% (e.g., Burian et al. 1999; Khalaily et al. 2007; Ulman et al. 2021), whereas their presence is more pronounced in some PPNA assemblages such as Gesher and Netiv Hagdud (Nadel 1997; Garfinkel and Dag 2006 and reference therein).

51The geographical selection of the site location fits well with other EPPNB sites in the Galilee, as most were established in previously uninhabited locales, on elevated terrain, and directly on the bedrock, taking advantage of natural depressions. Selected spots are usually located in vantage points, providing generous access to the woodlands that are typically found in the upper reaches of the Galilee, albeit having rather limited extents of arable land in the immediate vicinity (Birkenfeld 2018). This is in contrast to the following MPPNB period, in which sites are almost always located at the foothills, in close proximity to the rich valley soils, most probably reflecting the economic shifts related to the consolidation of agriculture in the region.

52Notably, the EPPNB site of Aḥihud, where a plentiful botanical assemblage consisted mainly of domesticated fava beans was discovered, does not fit the described pattern. While the extensive use of bedrock depressions was indeed noted, the site itself is located on the northern slopes of a small hill at the northeastern edge of the wide Akko plain, where arable lands are rich and plentiful (Paz and Vardi 2014; Caracuta et al. 2017; Vardi et al. this volume). Similarly, the site of Nahal Betzet I is also located relatively low, near the present-day Wadi Betzet, at the edges of its flood plain (Gopher 1989b).

53Data on site settings during the PPNA in the Galilee are scarce but, when available, indicate a pattern somewhat similar to that presented by EPPNB sites. The site of ‘Ein Dishna is located on a high terrace overlooking Wadi Tzalmon (Birkenfeld et al. 2019), and the site of Bir el-Maksur is situated on the top of a west-sloping hill, presenting vast exploitation of natural crevices and depressions in the local bedrock (Malinsky-Buller et al. 2013).

54In addition, the fairly extensive use of non-local pink and purple flint types for bi-directional blade production and, subsequently, as tool blanks is a well-recorded and acknowledged phenomenon both in the PPNA and EPPNB industries of the southern Levant (Marder et al. 2007; Barzilai 2010; Malinsky-Buller et al. 2013; Birkenfeld 2018: 142–143).

55With no absolute dating available, it is difficult to determine the chrono-cultural attribution of Amqa. The finds indicate an intriguing combination of fully fledged PPNB industries and some earlier PPNA features. Conversely, the excavations of the site showed no indication of the existence of two distinct occupation levels. The Neolithic sediments were homogeneous, and PPNA elements were found deposited together with EPPNB elements, with no clear spatial distinction. Therefore, the Neolithic finds from Amqa are believed to represent a coherent assemblage. The combination of EPPNB attributes and PPNA features can thus indicate a very early phase within the EPPNB, implying a time when PPNA traditions were declining but still present in the region.

Amqa from a regional perspective

56The lithic assemblage from Amqa displays an intriguing combination of stylistic traits. On the one hand, the fashioning style of the Khiam and Helwan points from Amqa closely resemble the “Aswadian” tradition typical to the Damascus Basin (i.e., wide short tangs and elongated medial sections; Cauvin 1994), implying contacts with the northern part of the southern Levant. The “Foliate”/“Nemrik” point and the Hollow-based point indicate contacts even further north, as they reflect typologies characteristic of southeastern Anatolia and its neighboring regions. Contacts with the Anatolia region are further implied by the presence of obsidian items in the assemblage. The presence of nephrite artifacts is intriguing in its potential indication of contacts with the southern Arabian Peninsula or further into the central and eastern parts of Asia.

57On the other hand, the Amqa assemblage also contains clear southern Levantine attributes. For example, the simultaneous coexistence of Helwan and Khiam points was reported from multiple sites in the region, such as Motza (Khalaily et al. 2007), Horvat Galil (Gopher 1997), Mujahiya (Gopher 1990) and Tell Qudadi (Krispin 2015). In addition, the presence of bifacial tools seems to be a southern Levantine tradition, as these tools are frequently absent from northern assemblages (Bar-Yosef and Belfer-Cohen 1989; Kozlowski and Aurenche 2005: 22–23, 43, 140, fig. 1.3.3.1).

58The different flint types used in the Amqa assemblage imply a typical south Levantine exploitation network, including both local and non-local, probably eastern sources. Notably, however, very few similarities, technological, stylistic, or other, can be drawn between the material culture of Amqa and sites located in the arid and semi-arid areas of the southernmost part of the Levant (e.g., Nahal Zahal in the Negev Desert; Birkenfeld and Abulafia this volume).

59Considering a wider perspective, the data presented here seems to support the notion that three main lithic variations existed during the EPPNB west of the Rift Valley, largely following a north-to-south geographic/environmental trend: 1) A northern, Mediterranean variant, including the Galilee and perhaps the Golan heights; 2) A southern variant, encompassing the arid regions such as the Negev Desert; 3) An “interim” variant, encompassing the Judea mountains and their foothills (somewhat corresponding to the regional “networks” suggested by Barzilai et al. [2010]). While there are clear connections and similarities between the three regions, some clear differences appear as well in both technology and typology.

60This regional distinction represents a massive step forward regarding our understanding of local dynamics during the EPPNB, especially considering that not long ago, the almost complete lack of sites raised the argument for a hiatus in the region during this period (e.g., Kuijt 1997; Edwards et al. 2004, 2007). Nevertheless, we are a long way from an exhaustive discussion of these lithic entities. While several assemblages are now known (e.g., Birkenfeld and Abulafia this volume), most, especially in the Mediterranean zones, are part of a longer occupational sequence, superseded by later Neolithic assemblages (e.g., Gopher 1990, 1997; Khalaily et al. 2007). The inability to fully isolate the EPPNB component impedes our understanding of these assemblages. With this in mind, it seems that the Amqa assemblage, with all of its imperfections, represents a unique opportunity to explore a relatively insulated assemblage. One can only hope that future discoveries will amend and supplement our ever-growing record of the period, enabling a much more detailed narrative of the neolithization processes in the southern Levant and their inter-regional dynamics.