Intermediate Bronze Age Crescent-Headed Figures in the Negev Highlands

Rock Art in the Western Negev Highlands

The survey area is located in the western Negev Highlands. It includes the basins of Naḥal (Wadi) Zin and Naḥal Nizzana, Ramat Matred (the Matred Plateau), Ramat Barnea (the Barnea Plateau), and the Nafḥa and Ḥamran Ridges that rise above these plateaus (Fig. 1). Most of the region’s landscape is hilly, typical of the Avdat geological formation that dominates the scenery. The limestone hills of the formation are dotted with clusters of dark crusted rocks, which appear as layers, outcrops, or separate boulders scattered randomly from wadi beds to mountaintops and on the slopes in between (Fig. 2). The patina, or desert varnish, that forms the dark crust on the limestone is composed of a thin layer of manganese, iron, and clay, which develops on an intermediate range of porous cavities made by microbiological activity that attacks the rock surface (Fig. 3). The microbiological process, a key element in the creation of the patina, depends on a range of factors including geological formation, climatic conditions, rock components, solar radiation, water, etc. (Krumbein and Jens 1981; Dorn 2013; Goldsmith, Stein, and Enzel 2014). In this unique constellation, when the patina is scratched, the soft cavity below the crust gives way and the light color of the inner stone is revealed. Ancient inhabitants of the region were familiar with this feature, and using pecking, incising, or carving, they removed the patina to delineate various designs. The etched scenes are thus clearly visible due to the contrast between the dark patina and the light-colored stone beneath it.

Petroglyphs in the region were incised over an extended range of time, and, in many cases, single rock panels were incised in multiple events at different times. Since the new patina is continually formed on the rock face, the older engravings gradually darken until they regain the original color of the entire rock. In cases of multiple engravings on particular rock surfaces, the different hues of the etched scenes and their superposition—when they overlap each other to some degree—enable the elucidation of the sequence in which the scenes were engraved on the panel. Also, the appearance of specific images in individual engravings offer clues as to their absolute dating. For example, since the camel was domesticated in this region during the late 2nd or early 1st millennium b.c.e., carving depicting men riding camels can only be from that time onwards.

The History of Petroglyph Research in the Negev

The first petroglyph survey in the Negev was carried out by Anati as part of his archeological survey of the region in 1954. Initially, some rocks bearing inscriptions and carved images were found in Wadi Ramalia (later renamed as Naḥal Avdat). These were followed by the discovery of additional petroglyph sites in Wadi el-Ajram (Naḥal Nizzana) and at Jebel Idid (Har Karkom), which by 1956 had yielded 600 documented panels (Anati 1956). Contemporaneous with Anati’s work, Nelson Glueck also documented numerous petroglyphs in his survey of the Negev, along ancient routes and near burial sites, and dated them according to inscriptions appearing in some, from the Nabataean times to the present (1956). In the 1980s, Anati returned to Har Karkom and carried out a systematic petroglyph survey of the mountain and its surrounding valleys (1999: 22–34).

As data from these surveys accumulated, and based on comparisons with petroglyphs from neighboring deserts (Sinai and southern Jordan), Anati developed a chronological framework for the Negev rock art. Studying the appearance of certain representations and styles, he divided the petroglyphs into seven chronological stages ranging from “early hunters,” dated to prehistoric times, to “herders and traders” from the early Islamic period. This division, and further subdivisions, were linked with the archeological timetable of the area (Anati 1965). Anati revised his scheme in 1998 (1999: 22–34), adding the data collected at the Har Karkom survey. The latter scheme proposed a more detailed framework, to account for the fluctuating frequency of significant motifs, and the changes in economic traits witnessed by the various styles and substyles.

During the 1980s and early 1990s, a broad archeological survey of the Negev Highlands was conducted under the auspices of the Archaeological Survey of Israel, then a subsidiary of the Israel Department of Antiquities (which later transformed into the Israel Antiquities Authority [IAA]). The Negev Emergency Survey was initiated in 1979 in response to the planned redeployment of the Israel Defense Forces from Sinai to the Negev after the Egypt-Israel peace treaty. This survey added a vast number of archeological sites to the regional record (Haiman 1986, 1991, 1993), however, except for the Har Nafḥa survey, which included a special appendix of petroglyph inscriptions (Lender 1990), almost no petroglyph sites were documented in the other subregions.

It was only in 2008 that a specialized “petroglyph survey” was conducted again in the Negev Highlands. This work, led by Eisenberg-Degen, focused on Har Miḥia on the northeastern part of the Ramat Matred. Nine hundred sixty-five petroglyph panels were discovered, and the superposition of their elements was documented. As part of her work, Eisenberg-Degen devised a new chronological sequence of the main representations she recorded. The sequence is based on engraving superposition, differences in patina hues, the occurrence of Old North Arabian and Arabic inscriptions, and the manifestation of well-dated episodes in this region, such as the domestication of the camel or the use of long swords. One of her essential conclusions is that the creation of petroglyphs in the Negev started in the Timnian culture (that is no earlier than the 6th millennium b.c.e.; Rosen 2017: 161), with the introduction of herding into the region, and not earlier as was previously proposed by Anati (Eisenberg-Degen 2012; Eisenberg-Degen and Rosen 2013: 240–41).

The “Crescent-Headed Figures” in the Western Negev Highlands Rock Art Survey

The Western Negev Highlands Rock Art Survey was conducted from 2012 to 2020. The area (Fig. 1) was surveyed on foot with significant attention given to areas with patinated rocks. Each concentration of petroglyphs was defined as a site, and these were, by definition, limited to a maximal size of 25 × 25 m. Thus, a continuous spread of petroglyphs larger than 25 × 25 m was artificially divided into several sites. Within the sites, any rock surface containing petroglyphs was defined as a “panel,” while the images included in a panel were defined as “elements.” Altogether, 1,454 rock art sites containing 9,320 panels were documented.

Out of this massive corpus, this paper focuses on a certain design element that has only rarely been described in the region before, the “crescent-headed figure” (henceforth CHF). The basic shape of a CHF is that of a finely pecked human image with an hourglass-shaped body, wearing a lunate-shaped head-cover with a decoration at its tip, and bearing a dagger with a lunate pommel secured to its waist (Fig. 4).

We are not the first to report such a design in the region; Anati described one such figure from Ezuz, and dated it to his Style VI-A, which belongs within his so-called “Bronze Age Complex” that stretches from the 4th to the 2nd millennium b.c.e. (2015: 146–62). However, the regularity of this representation is apparent only with our broad-scale and detailed survey, which has revealed the recurrence and distribution of the CHFs, as well as their link with a specific context in the region’s history. In addition, we have noticed a consistent clustering of unique designs in close association with the CHFs, such as engravings of lions, bulls, or both. This fact is remarkable, especially because these animals are rarely depicted in the Negev.

In our dataset of CHFs, we have included elements with at least two of the following three components: 1) crescent-shaped headgear, 2) a dagger with a crescent-shaped pommel, and 3) an hourglass-shaped body (which probably represents in its lower part a knee-high garment). The CHFs often contain additional recurring attributes, such as a plume attached to the headgear, a rod/staff/torch carried in one hand, and occasionally a small decurve bow (see the distribution of these attributes in Fig. 5). To date, 21 CHF sites displaying a total number of 48 figures have been discovered in the survey area (Table 1).

The CHFs do not appear as stand-alone figures; rather, almost always, they are displayed as actors taking part within more complex scenes. In order to present this artistic rendering, three selected panels are described below: the Har Nafḥa Panel (Site 8, Panel 2), the Ramat Matred Panel (Site 44, Panel 3), and the Upper Laʿana Panel (Site 116, Panel 4).

The Har Nafḥa Panel (Fig. 6), with motifs common to many petroglyphs panels in the Negev Highlands, was incised in a few different events, resulting in an accumulative scene comprised of six superimposed layers of different ages.

The latest layer on the panel (Layer 1) is represented by not-yet patinated bullet holes that probably hit the rock during recent military training in the area. Layers 2–4 portray images of riders and hunting scenes with a patina hue ranging from light to dark brown (Fig. 6b). The lowermost, oldest layer (Layer 6), shows seventeen CHFs, a lion, canine predators, and ibexes. Two of the human figures are kneeling, while three seem to be holding a lightning-shaped natural fissure in the rock. Some of the figures hold small decurve bows and some grasp torch-like objects (Fig. 6c). A part of the Layer 6 scene was erased or damaged not long after its creation, as the patina hue of the disturbances (labeled as Layer 5) is identical to that of Layer 6.

The Ramat Matred Panel (Figs. 7–8) exhibits four different layers of engravings. The latest (Layer 1) consists of Bedouin tribal marks in a light hue. Layer 2 comprises several etching episodes from the Roman and Byzantine periods that display symbols, footprints, and a part of a short ancient North Arabian inscription.

The lowermost, thus oldest, layers (3–4) portray three CHFs. The figure at the bottom of the panel is shooting an arrow from a small decurve bow while standing next to a bull (Fig. 8). A second figure, in the middle of the scene, is depicted next to a lion. This figure is engraved horizontally, a posture that might suggest its death. Layer 3 includes some engravings with CHF features and was defined as a separate layer since its patina is lighter. As such, it probably represents later copies of the older images.

The Upper Laʿana Panel (Figs.9–10), located at a junction of paths leading from Ramat Matred to Naḥal Resisim, consists of four layers. The first and youngest layer of the panel displays ibexes, a cross, and two human figures, one of which is holding a B-shaped composite bow similar in shape and size to the early Islamic bows (Zutterman 2003: 148). This layer can thus be dated to the second half of the 1st millennium c.e. The second layer is much older and portrays a typical Negev rock art scene, with a predator, probably a dog, chasing a male ibex. Images of this type and their interpretations are a fundamental component within the Negev rock art corpus (e.g., Avner, Kolska Horwitz, and Horowitz 2016). In the third layer, below the ibex, a CHF is pecked. Its crescent pommel dagger, and crescent head-cover with a plume on top, confirm that this layer is contemporaneous with the Nafḥa and Matred Panels. The fourth and artistically richest layer is older than the CHFs, its elements are entirely covered by patina. This scene shows eleven human figures, most of whom are males, some armed with adzes on long sticks in shapes that are common in the southern Levant during the late Neolithic and the Chalcolithic periods (Barkai 2011). The scene is complex. Some of the figures are lying horizontally, a posture that might represent death; one of the standing figures has horns and exaggerated open hands and might represent a chief or a healer. Currently, we are not aware of any comparable scene.

Dating the CHF Scenes

Since no archaeological material accompanies the CHF panels, which have been exposed to the elements since their production, their dating poses a methodological difficulty. The scenes certainly represent an old phase within the rock art history of the Negev, as the CHFs’ petroglyphs almost always appear in the lowermost (oldest) layers of multi-period panels. Supporting this hypothesis is the fact that their hue is always very close to the original dark patina covering the natural face of the rocks.

In addition, the fact that there are no inscriptions incorporated within the CHF compositions supports the estimate of their early date, since the oldest inscriptions of Negev rock art are written in Old North Arabian, dated to the 4th century b.c.e. to 3rd century c.e. (Macdonald 2000).

However, these date estimates are still very broad. In order to gain a more accurate evaluation, we hereby use a combination of arguments, which taken together offer a firmer date-range for the CHFs. These considerations include:

In the following sections, these considerations will be described in detail, and on their basis, precise dating of this socio-artistic endeavor will be proposed.

Presence or Absence of Specific Animal Species in the CHF Scenes

The Negev Highlands rock art scenes contain a large variety of animals: ibexs, camels, horses, dogs, donkeys, bulls, lions, and more. The occurrence of some, like the ibex, is so broad that it cannot offer any specific hint for a date. Similarly, since dogs have accompanied humans for at least the last 15,000 years, their presence in etched scenes is not useful for dating purposes. However, the presence or absence of other species is more chronologically indicative.

To begin with, it is highly revealing that camels and horses never appear in combination with the CHFs. Camels were domesticated and introduced into the Negev in the Iron Age, in the early 1st millennium b.c.e. (Sapir-Hen and Ben Yosef 2013: 283–77), and from then onwards they are a very common species in Negev rock art (Schwimer 2015). Noteworthy, there are pre-1st millennium b.c.e. petroglyphs of camels in the Near East, specifically in Saudi Arabia. There, they even appear occasionally next to a certain kind of CHF, which Anati includes within his “realistic dynamic” style of “late hunters” (Anati 2015: 84). However, these are not domesticated camels, but wild ones, as can be deduced from their depiction as prey animals within a hunting scene (Curci, Carletti, and Tosi 2014: 219).

Horses are also absent from CHF compositions. Although horse bones appear in the region as early as the Chalcolithic period (Grigson 1993: 645–55), it is generally accepted that domesticated horses were not present in the Levant before the Middle Bronze Age (Moorey 1986: 198). Textual references and artistic depictions of horses as riding beasts exist in Mesopotamia and the Levant from the 18th century b.c.e. onwards, for example, in letters from Mari that deal with the operation of stables and the training of horses (Moorey 1986: 198). Therefore, it might be concluded that CHFs date before the Middle Bronze Age.

Unlike horses and camels, we do find occasional depictions of domesticated donkeys incorporated in CHF scenes (Fig. 11). Although it is not always easy to distinguish between different kinds of equids in ancient petroglyphs, in some of the linear styles of the Negev rock art, donkeys have longer ears and are smaller than any anthropomorphic figures when these appear in the same scene. Also, horses are drawn with longer necks and legs, and with an accentuated chest. Donkeys were the first domesticated equids in the Levant, appearing in local households at least from the 4th millennium b.c.e., both in faunal assemblages or as cult figurines (Milevski 2009: 133–36). Thus, we can narrow down the range within which the CHFs appeared in the Negev to sometime between the Late Chalcolithic and the Middle Bronze Age.

Other animal species that appear in combination with the CHFs are bulls and lions. Both these animals are rare in the Negev Highlands rock art. Bulls appear in 3.2% of the rock-art sites in the region (in 47 out of a total of 1,454 sites), and lions in 1% (15 sites). However, their frequency within the CHF sites in the region is tenfold higher. Some 75% of the CHF sites include depictions of bulls, lions, or a combination of both (Table 1). The high correlation between CHFs and bull and lion depictions is reflected, too, in the similar patina of the three features, all ranging from dark brown to black.

Anati recorded depictions of bulls in his rock art surveys in the region as well. He dated their presence to his “Period 3/late hunters,” that spans the 6th–3rd millennia b.c.e., and to “Period 4/the age of pastoralism,” dated to the 3rd–2nd millennia b.c.e. He notes a rise in the number of bull and lion depictions in his “sub-period VI-A,” which he describes as a time when a new population of a Mesopotamian origin entered the Negev, at the beginning of the 3rd millennium b.c.e. (Anati 2015: 145).

Bulls are uncommon in arid regions, as the amount of food and water they require is well beyond those available in deserts like the Negev. Similarly, the Negev environment is not suitable for lions since there is not enough game to sustain them. Accordingly, it is reasonable to assume that the people who incised the lions and bulls next to the CHFs either depicted sights and memories from faraway places, or depicted these animals as symbols of political power (as in the Narmer Palette; see Yadin 1955), religious icons (representing the moon or the storm deities; see Ornan 2001), or astronomical signs (the bull and lion constellations; see Hartner 1965)—all common in formal and non-formal ancient Near Eastern art.

Particular Attributes of the CHFs and Their Occurrence within Dated Archaeological Sites across the Ancient Near East

In addition to the presence or absence of certain animal species, the recurrence of particular sets of garments, tools, or weapons associated with the CHFs provides another clue for understanding their socio-chronological settings. In a similar vein, Anati (2015: 81) claimed that the depiction of weapons or tools combined with the presence or absence of domesticated animals’ representations in rock art can be considered as a solid chronological anchor.

The significance of repeated portrayals becomes even more meaningful when they are compared with similar attributes in broader geographical areas. In this way, Eisenberg-Degen and Rosen (2013: 247–50) offered valuable insights while studying the temporal-spatial distribution of long swords that they had identified in Negev Highlands petroglyphs. Similarly, Beno Rothenberg identified figures engraved on the Chariots Panel in Timna as New Kingdom Egyptian soldiers, based on parallels of their gear with weapons discovered in Egypt (1972: 121–24).

In the following pages, certain iconographic elements appearing in the CHF panels will be discussed and parallels for them found in the ancient Near East and beyond will be noted. On this basis, we shall attempt to offer chronological and socio-cultural interpretations for these scenes. The elements to be discussed are the crescent headgear, daggers with crescent-shaped pommels, and small decurve bows.

The Crescent Headgear

Headgear, crescent in shape (Fig. 12), is the main attribute that defines the petroglyphs labeled as CHF (only two images out of 48 occurrences of these figures were bareheaded). This headgear is sketched as an almost complete circle, opened on its bottom. Its width is approximately two-thirds that of the shoulders. The figures’ necks are long, and neither faces nor heads are indicated. The curved outline, which we interpret as headgear, can also denote a hairstyle, but the frequent addition of a plume on its top favors its identification as headgear. The plume is sometimes longer than the headgear itself, and it usually twists to one side.

Although some helmets are known in the regional archaeological record of the Bronze Age, such as the Meskalamdug gold helmet from Ur (Moorey 1982), or the soldiers’ helmets on the Stele of the Vultures from Tello (Winter 1985), none of them resembles in any way the CHF headgear, which in principle could have been made of perishable materials like leather, feathers, or textile. Not only that, such helmets are also uncommon in the region’s (Israel, Jordan, Egypt) rock art. However, a headgear style that resembles the CHFs appears further away, on engraved Bronze Age burial stelae from megalithic burials in Yemen dated to the end of the 3rd millennium b.c.e. (Vogt 1997: 30–33), and on rock art from the same region (Jung 1991: 67). Interestingly, in the latter cases, the very abstract rendering of the figures also includes crescent-pommel daggers (Fig. 13).

Daggers with Crescent-Shaped Pommels

Over 60% of the CHFs from the Negev Highlands carry a dagger with a crescent pommel, which makes it one of their most common attributes (Fig. 14). These daggers are usually equal in length to the shoulder; thus, in reality, they might have been 40–50 cm long. Some examples are longer, perhaps drawn in an exaggerated way to emphasize the importance of their bearers, or they might have been swords. The daggers are attached horizontally to the waists, the pommels are drawn mostly on the right side of the figures, indicating that the daggers were pulled out with the right hand.

Daggers with crescent-shaped pommels, similar to those carried by the Negev Highlands CHFs, have been recovered from burial contexts in the ancient Near East, from the mid-3rd to the mid-2nd millennia b.c.e. The earliest examples are from Mesopotamia and Anatolia, later ones appear in the Arabian Peninsula and the Levant, and the latest is found in Egypt and the Nile Valley.

The earliest example of a dagger with a crescent-shaped pommel that we came across was found by Sir Leonard Wooley in the Royal Tombs of Ur, in the tomb of Meskalamadug dated to the mid-3rd millennium b.c.e. (Moorey 1982: 58). Six daggers were found in that tomb, one of which had an ivory crescent pommel (Fig. 15a).

A second parallel comes from Alaca Höyük in central Anatolia. It is a gold-plated crescent pommel dagger with an iron blade (Fig. 15b),¹ found in Tomb K of the Royal Tombs (Koşay 1938). The tomb was among a series of elite burials that consisted of fourteen shaft tombs yielding a considerable number of finds, including crescent-headed anthropomorphic female figurines and bulls’ heads. The excavators originally dated it to 2300 b.c.e., a date which was later revised according to ¹⁴C readings to the mid-3rd millennium b.c.e. (Yalçın and Yalçın 2013). Another dagger with a crescent-shaped pommel was found in Tumulus 422 at the necropolis of Dilmun in Madinat Hamad, Bahrain (Fig. 15c), dated by its contents to 2200–2000 b.c.e. (Lombard 2000: 55).

Additional examples of crescent pommel daggers come from Israel. One such item was discovered in shaft tomb No. 2 at a cemetery south of Tel Reḥov in the Jordan Valley. The tomb was a burial of a “warrior,” ca. 50 years old, with a duck-bill ax, a spearhead, four-spouted lamps, and a dagger with a crescent pommel made of alabaster placed next to him (Fig. 15d). The cemetery functioned from the Intermediate Bronze Age to the first stage of the Middle Bronze Age, while Tomb 2 is dated to the latter period, hence within the range of 2000–1900 b.c.e. (Yogev 1985: 107). The excavators identified the dagger as Type 25B of Rachel Maxwell-Hyslop’s typology (1946: 26) of daggers and swords from the Bronze Age Levant, which is dated to the Middle Bronze Age.

Also, a copper coating of a crescent pommel was excavated in a shaft tomb at the Intermediate Bronze Age cemetery of Qanat el-Jaʿar in the Bet Shean Valley. It was a burial of two adults and two children, next to whom were placed a short sword and a few spearheads (Horowitz 2016: 83). In the same general region, more examples were discovered: a crescent pommel made of limestone at Tomb T3 at the Intermediate Bronze Age cemetery in Naḥaf (Getzov 1995: 13), a crescent pommel in Shaft Tomb 65 at the Intermediate Bronze Age cemetery of ʿEinot Shuni (Peilstöcker 2008: 2040), and a bronze coating of a crescent-shaped pommel in Shaft Tomb 3 at the Maʿabarot Intermediate Bronze Age cemetery (Dar 1977: 63).

Further south, and later in date, are crescent pommel daggers discovered in Egypt, which Yigael Yadin (1963: 127) assigned to the 2nd millennium b.c.e. (Fig. 15e). Such a dagger was also discovered in a tomb at Tell e-Dabʿa Layer H, dated to the beginning of the 2nd millennium b.c.e. (Bietak 1996: fig. 10, pl. 20).

Some studies of ancient Near Eastern weaponry make more general observations about daggers with crescent pommels. For example, Graham Philip (1989: 116, 2006: 145) in Metal Weapons of the Early and Middle Bronze Age in Syria Palestine deals with crescent pommel daggers, which he labels “type 12.” He asserts that this type originated in west Asia during the Intermediate–Middle Bronze Age and that it appeared first in the Levant and later in Egypt. This observation complements, in his opinion, the accepted views concerning the adoption of Asiatic traditions by the Egyptian ruling class at the beginning of the 2nd millennium b.c.e.

Besides actual daggers, there are many depictions of crescent pommel daggers in ancient Near Eastern art. A beautifully engraved example appears on a limestone inlay from Palace G in Ebla, dated to 2350–2250 b.c.e., portraying a war scene where a warrior uses a crescent pommel dagger to stab his opponent (Aruz 2003: 177). Daggers with similar shapes also appear on tombstones from south Arabia and Yemen. These tombstones, or stelae, are up to 1.5 m high. They usually portray upper torsos of bearded figures, who carry daggers with a crescent pommel handle (Fig. 13). Burkhard Vogt (1997: 33) suggests that these decorated stelae should be dated to the 3rd and 2nd millennia b.c.e., while Lynne Newton and Juris Zarins (2000: 165–70) date them to a general “Bronze Age” culture.

Similar daggers also appear in rock art from Saudi Arabia, Yemen, and Jordan. A few panels that the Philby-Ryckmans-Lippens Expedition documented at Jebel Qara in Saudi Arabia in 1951, which were later published by Anati (1968), depict human figures carrying crescent pommel daggers. Despite the weapons’ resemblance to the Negev examples, the human forms are sketched differently than the Negev CHFs. The former demonstrates dynamic postures (Fig. 16), while the Negev CHFs are mostly static. Anati (1968: 75) dates the Jebel Qara figures to the 3rd and early 2nd millennium b.c.e., but Robert Bednarik and Majeed Khan (2009: 17) reexamined some of the panels and concluded that certain figures should be dated to a later period. Crescent pommel daggers appear also south of Jebel Qara. At Jebel Haid in Yemen, Pavel Červicek and Fritz Kortler (1979: 250) photographed a panel portraying a figure carrying a dagger with a dramatically large crescent pommel that according to Michael Jung (1991: 67) could be an indication of its status, a chief or a leader (Fig. 17a). Noteworthy, a similar depiction of an “exaggerated” crescent pommel was noted in our survey as well, at Har Nafḥa (Fig. 17b).

Discussing the overall phenomenon of crescent pommel daggers, Newton and Zarins (2000: 157–61) note their appearance in burials all over the ancient Near East, their pictorial depictions on tombstones in Yemen, and on rock art in Saudi Arabia. Accordingly, they propose that elites used them as status markers indicating warriors or an upper class. They date the occurrence of this dagger from the mid-3rd to the mid-2nd millennium b.c.e.

Combining all these data enables an appraisal of the origin and spread of this type of weapon, an estimate of the date in which such depictions appear in the Negev Highlands, and a speculation concerning their meaning.

It is highly plausible that the form of a crescent pommel dagger originated in Mesopotamia or Anatolia around the mid-3rd millennium b.c.e. Shortly afterward this style spread to Arabia and the southern Levant as well. At the beginning of the 2nd millennium b.c.e. crescent pommel daggers occur in Egypt, where they are usually interpreted as a result of Asiatic immigration into the Nile Delta. Hence, from a southern Levantine point of view, it looks as if a style of weaponry, or a group bearing this distinct kind of daggers, spread from the north to the south. While one trajectory led into the Arabian Peninsula, a second arrived in Egypt, through the Negev Highlands.

The broad distribution of crescent pommel daggers in the region, described above, goes hand in hand with an unprecedented increase in the number of metal weapons deposited in southern Levantine tombs next to the deceased at the end of the 3rd millennium b.c.e. (Greener 2012: 34). This situation may be an indication for increased inter-personal violence at that time (although there is no supportive evidence for this hypothesis), or a fashion in which a weapon was considered a compulsory component in personal outfits, in life, death, or both.

Small Decurve Bows

Another conventional weapon carried by the CHFs is the decurve bow (also sometimes labeled as a double-convex bow). It appears in 20% of their images, yet the depiction is consistent throughout, hence it probably reflects a standard type. The bow is small, and by comparing its relative size to the CHF bodies, it can be inferred that it was about 50–70 cm long (Fig. 18).

Regarding bows in general, it appears that most bows documented in the Western Negev Highlands Rock Art Survey are not associated with CHFs. However, these are different in shape, much larger, and of the simpler curved type. Bows that are similar in shape to those carried by the CHFs, but slightly bigger, appear in Egyptian artistic depictions from the 4th millennium b.c.e. (e.g., the Hunters Palette, Spencer 1980), to the Middle Bronze Age (e.g., the Beni-Hasan tomb painting, Shea 1981: 225). Bows similar in shape and size are rare; one such example is depicted in a war scene relief from Mari, dated to the 3rd millennium b.c.e. (Yadin 1972: 89).

The Fit between CHF Panel Locations and the Distribution of Dated Archaeological Sites in the Region and the Routes Connecting Them

A final indication for the dating of the CHFs and for understanding their socio-cultural association can be inferred from the relationship between their findspots and well-dated archaeological sites. Two such contexts occur in our case: the discovery of a CHF scene within a well-dated excavated site, and the spatial distribution of panels along possible routes connecting dated points of activity (settlements, mines, smelting areas, etc.).

The ʿEin Ziq CHFs

Major support for the proposed association between the CHF panels and Intermediate Bronze Age points of activity (sites, mines, etc.) was recently discovered at the Intermediate Bronze Age site of ʿEin Ziq.

ʿEin Ziq is the largest Intermediate Bronze Age site in the Negev Highlands, located ca. 10 km southeast of Sde Boker near the channel of Naḥal Ziq, at the southwestern edge of Biqʿat Zin (literally the Zin Valley; Figs.1,23). The site was recorded in a survey conducted in the early 1980s by Yigal Israel and Dov Nahlieli and excavated in the mid-1980s by Rudolph Cohen. Covering ca. 2 ha, it consists of more than 200 circular structures arranged in clusters, with small courtyards and alleys in between them (Cohen 1984, 1986, 1999). The finds and ¹⁴C dates indicate that ʿEin Ziq was a single-period Intermediate Bronze Age site, occupied within the range of 2500–2200 b.c.e. (Avner and Carmi 2001: table 2; Dunseth et al. 2017: 161; Finkelstein et al. 2018: 70). The finds at the site included a large number of pottery and stone vessels, hammerstones, stone anvils, copper ingots, various copper objects, and more (Cohen 1984, 1986, 1999). Mordechai Haiman (1996) proposed that these finds, as well as the size and design of the site and its structures, indicate that copper processing took place there. Zach Dunseth, Israel Finkelstein, and Ruth Shahack-Gross (2018), who re-excavated and conducted a series of analyses at certain localities at the site in 2014–2015, indicated that they found no trace of copper processing at the site and therefore suggested that it was a trading post on the way to the coastal plain and Egypt.

A passing visitor, walking through ʿEin Ziq in 2018, noted incised figures carved on one of the site’s stones and brought this to our attention.² This engraving consists of three images of CHFs. Contrary to all other known panels, it is engraved on a white stone, making it less visible at first sight (Fig. 19). The lowermost figure on this panel holds a crescent pommel dagger on the left side of its hourglass-shaped body and has a crescent head-cover. The middle figure is quite similar to the lower one, but is damaged on its upper part. The image on top, engraved with sharper pecking marks that may indicate later retouching, has a large crescent pommel weapon, which is worn in the opposite direction, with the pommel in the right side. The head of this figure is damaged, too.

When this panel was discovered we questioned its authenticity, since, to a certain degree, it seems strange that it was not noticed before. Yet we do believe it is authentic, especially due to the pattern of the incisions’ damage (see below), and the consistency of the details which conform with the common attributes of the CHFs.

In order to examine the authenticity of this panel, we consulted the field diaries and photographs of the 1980s excavation at the Israel Antiquities Authority’s archives.³ The diaries did not indicate that anyone noticed the incisions, and unfortunately, all photographs of the structure were taken from angles that do not allow one to see the incised face of the stone. But a few important facts did appear.

First, we could delineate the pre-excavation surface level at the spot (labeled by the excavators “structure 11”) and reconstruct it in relation to the incised panel (black line, Fig. 20). The uppermost two CHFs were damaged in a line that is parallel and about 10 cm below the previous surface (yellow line, Fig. 20). This situation probably represents the lower level of the soil A Horizon, the layer which is most prone to erosion and mineral wash (Bridges 1990: 37), or an old surface level. Accordingly, we propose that although the structure was covered up to the black line before the new excavation, the rock surface broke off due to soil processes down to the yellow line, creating the current state of the panel (Fig. 20).

Support for this assumption derives from Dunseth et al. 2017 (table 2, figs. 3, 8) who dated the immediate accumulation of aeolian dust over Intermediate Bronze Age deposits at ʿEin Ziq, which occurs approximately at the probable A Horizon line (as seen in Fig. 20). They discovered it to be 3.8 ± 0.3 kya by OSL (optically stimulated luminescence) to ca. 2100–1500 b.c.e., lending more support to the idea that the CHF figures were carved before that date.

Second, the examination of the excavation photographs revealed that the stone on which the CHFs are incised is one of the building blocks of a wall within “structure 11,” exposed in the 1985 excavations at ʿEin Ziq (Cohen 1986: 75). More specifically, it is the right-hand side frame-stone of the opening leading from Room 87 to Room 78 (Fig. 21). Notably, adjacent to the threshold of that specific opening, within Room 78 and about 50 cm away from the engraved panel, a cache of copper ingots was hidden just below floor surface (Fig. 22).

We propose that the close proximity between the engraving and the cache is not a coincidence, and that the engraving actually served as a marker for those who deposited the cache, so they could locate it when they, or their messengers, returned to the site. The connection between the cache and the engraving suggests a practical association between the engraving of the CHF scenes and the handling of copper.

Routes

As findspots of CHF panels have accumulated, their distribution reveals that they delineate two lines that cross the Western Negev Highlands in a southeast to northwest direction: a “northern route” going along the northern edge of Ramat Matred, and a “southern route” passing south of Har Nafḥa (Fig. 23). The two routes converge at two places, 5 km east of Beʾer Resisim, and again at the northwestern corner of the survey area, near Beʾer Milka.

Most of the panel sites are found adjacent to features resembling ancient paths that pass through the main valleys or cross the low rising ridges that separate them. This observation strengthens the assumption that they were engraved along ancient routes.

Expanding the map to include the areas of Edom in the east (within the current borders of Jordan), and Sinai in the west (within the modern borders of Egypt), reveals that the two routes follow a general direction running from Edom, through the Negev Highlands, and towards the north Sinai coast (Fig. 23). Furthermore, within the survey area, the routes pass through the major Intermediate Bronze Age sites of Beʾer Resisim and Naḥal Nizzana. Beyond the surveyed area, their extrapolated lines continue east through ʿEin Ziq, the largest Intermediate Bronze Age site in the Negev Highlands, towards Faynan, the large copper mining area that operated in the Early Bronze and Intermediate Bronze Age, and west towards the Jiradi area of North Sinai, where a large concentration of small Intermediate Bronze Age sites was discovered (Oren and Yekutieli 1990: 10).

In order to examine whether the proposed routes have a topographic advantage over other possible tracks connecting Faynan with North Sinai through the survey area, we conducted several GIS-generated least-cost-path tests for certain sections: (a) ʿEin Ziq to Beʾer Resisim, (b) Faynan to Beʾer Resisim, and (c) Ramat Matred (at a point where a concentration of CHF panels is found) to Beʾer Milka,⁴ which returned two options. The results of these tests display a good match between the CHF site locations and the most efficient paths (Fig. 24), thus supporting the conclusion that the panels are spread along routes connecting Intermediate Bronze Age points of activity (sites, mines, etc.).

Notably, in the region of Har Karkom, Anati (1986: fig. 60) recorded a petroglyph portraying an unarmed CHF. In a visit to ʿEin Hameʿara, a small spring about half way between Har Karkom and the Western Negev Highlands Rock Art Survey, we identified a rock art site that overlooks the main valley, with classical engravings of CHF with lions and bulls next to them (Fig. 25). These two observations may hint at the existence of another route joining our proposed ones from the south, leading to the ʿArava Valley, and even perhaps towards the Timnaʿ copper mines.

The CHFs in Context: Intermediate Bronze Age Copper Distribution Networks

The proposed scenario for the spread of the crescent pommel dagger motif, coupled with the presence of bulls, lions, and domesticated donkeys in the CHF scenes and the absence of camels, horses, and inscriptions on them, as well as the occurrence of CHFs incised onto a building stone of a structure within the single period site of ʿEin Ziq, enables dating the CHF compositions in the Negev Highlands to the second half of the 3rd millennium b.c.e. In the regional terminology, this chronological range matches the Intermediate Bronze Age.

Before our current research, a fairly detailed reconstruction of the Negev Intermediate Bronze Age social landscape was proposed by numerous scholars. Haiman (1996) had divided the hundreds of Intermediate Bronze Age sites discovered in the Negev Highlands into large “permanent” sites and small “temporary” ones. Within the first category, he defined three groups: the first, which includes ʿEin Ziq, Beʾer Resisim, Mashabbe Sade, and Beʾer Hayil, was labeled the “ʿEin Ziq group.” He described these sites as large settlements comprising of circular, stone-built structures, with pillars in their centers to support the roofs, but lacking enclosures for animal husbandry (Haiman 1996: 3). The second group, which he called the “Har Yeroḥam group,” with the sites of Har Yeroḥam and Har Tsayad, included reasonably large sites, usually situated on hilltops, and consisting of rectilinear structures and small enclosures (Haiman 1996: 3). The third, the “Naḥal Nizzana group,” with the Naḥal Nizzana and Wadi Fauqiya sites, had large structures with multiple rooms and large enclosures for animal husbandry (Haiman 1996: 3–4). Haiman suggested that a population of a non-Negebite origin built the first two groups, while the third group, whose architecture preserved old local traditions, probably belonged to an indigenous population living in the region since the Early Bronze Age, if not earlier. He noted that the attributes of those three groups could be distinguished within the much smaller-size “temporary” sites’ class as well (Haiman 1996: 5).

By examining the pottery assemblages of the Intermediate Bronze Age sites in the Negev Highlands, Yuval Goren (1996) concluded that they contain ceramics from varied origins in a radius of more than a hundred kilometers around the region. Specifically, in regard to Beʾer Resisim and ʿEin Ziq, which belonged to Haiman’s first group, and which, according to our spatial analysis, show a direct association with the CHF findspots, Goren noted that more than 50% of their ceramic assemblages originated in southern or central Jordan (1996: 50).

Beyond architecture and pottery, a consistent component within the material remains of the Negev Highlands Intermediate Bronze Age sites are copper ingots. Such ingots were found at ʿEin Ziq, Har Tsayad, Beʾer Resisim, Mashabbe Sade (Cohen 1999), and Har Yeroḥam (Kochavi 1969: 40). Isotope analysis revealed that the primary source of this copper was the Faynan region. This observation was later supported by the discovery of molds that fit the Negev ingots in shape at Khirbet Hamra Ifdan, in the heart of the Faynan copper extraction district (Levy et al. 2002; Hauptmann et al. 2015).

The latter discoveries confirmed Haiman’s (1996: 21–22) proposal that the large Intermediate Bronze Age sites in the Negev Highlands were involved in copper trade that originated in Faynan and moved west, probably using donkey caravans. Incorporating the results of excavations at the ʿEin Yahav smelting site with Haiman’s observations, Yekutieli, Sariel Shalev, and Sana Shilstein (2005) reconstructed a complex copper production, processing, and distribution network, covering the Edom-Negev-Sinai region during the Intermediate Bronze Age.

Recently, Finkelstein, Matthew Adams, Dunseth, and Shahack-Gross (2018) and Dunseth, Finkelstein, and Shahack-Gross (2018), following additional excavations and analyses at ʿEin Ziq, agreed with the idea that the main reason for Intermediate Bronze Age settlement in the Negev Highlands was related to the copper trade, but narrowed down the time frame for this activity to the first half of the Intermediate Bronze Age, and noted that although ʿEin Ziq was involved in the copper trade, no copper processing took place at the site.

The evidence of the CHF engravings, presented above, supports these previous reconstructions and supplements them with additional nuance. We now have a graphic notion of how some of the Intermediate Bronze Age community conceived of their appearance—as individuals with a distinctive headdress or headcovers—occasionally with a feather on top, wearing knee-high skirts, and consistently wearing a specific type of dagger attached to their belts. Also, they often carried small decurve bows. Their iconographic repertoire also includes references to animals that were not common to the region, such as lions and bulls, probably reflecting collective memories or myths.

Putting the new information within already known aspects of the Intermediate Bronze Age in the Negev and beyond allows for additional insights. First, our discoveries further support the notion that one of the main reasons for the abundance of Intermediate Bronze Age activity in the Negev Highlands is linked with the copper production and distribution network. Second, the spatial spread of the CHF panels clearly outlines Intermediate Bronze Age routes leading from Faynan, through the Negev Highlands, towards North Sinai. These routes were probably used, among other things, for transporting copper to Egypt. Supporting evidence for this notion lies in the discovery of Egyptian late Old Kingdom pottery at the North Sinai Intermediate Bronze Age sites at the northwestern end of the proposed routes (Fig. 23; the Jiradi sites, Oren and Yekutieli 1990), as well as some pottery of an Egyptian origin at the Negev Intermediate Bronze Age sites (Goren 1996: 57). Third, it has long been known that the Intermediate Bronze Age in the southern Levant witnessed an unprecedented increase in the number of copper weapons—daggers, swords, and javelins—deposited in tombs next to the deceased. Our analysis indicated that daggers, almost as a rule, appear as a part of the CHFs’ dressing kit. Combining the two observations, we propose that the daggers were a part of the typical representation of people, probably men, in the southern Levantine Intermediate Bronze Age. In other words, we propose that in that period, a weapon was a compulsory component of the idealized outfit, in life, death, or both. Fourth, the study of the crescent pommel daggers and its numerous parallels in the ancient Near East in general, and the southern Levant in particular, reflects a pattern of spread across a wide geographical area, to such a degree that this specific dagger became a must-have item, or even a social symbol of certain groups, including the Negev copper traders. Finally, from the point of view of Negev Highlands rock art research, our study had managed to date a recurring artistic style and propose the historical context of its formation. This can now be used as a chronological anchor for future works on Bronze Age Negebite rock art. From the perspective of southern Levantine Intermediate Bronze Age research, we have managed to shed more light on a dynamic regional socio-economic system that encompassed broader regions than just the Negev Highlands, and connected centers and peripheries, people, and ideas from various origins.