Long-distance trade in the Middle Chalcolithic of the southern Levant: The case of the olivine beads from Tel Tsaf, Jordan Valley, Israel

Eight olivine beads found at the Middle Chalcolithic site of Tel Tsaf (ca. 5,200–4,700 cal. BC), Jordan Valley, Israel, underscore a new facet of interregional exchange for this period. The current paper presents the olivine beads assemblage, its morphometric and technological characteristics, and chemical composition. The results of the chemical analysis suggest that all eight beads derive from the same source. By means of comparison with the chemical characteristics of known olivine sources, we argue for a northeastern African–western Arabian provenience and cautiously suggest Ethiopia as a probable origin. Finally, we discuss the significance of the assemblage, its possible origin, and the mechanisms that may have brought the beads to the site.


Introduction
In antiquity, beads and other items of personal adornment appreciated by the community were among the principal devices for transmitting social and economic information over large distances. By token of their small size, they could change hands easily, while their aesthetic and symbolic value rendered them economically effective [1]. Notably, beads fulfill various functions [1][2][3][4] and are often considered markers of identity, social status, and even economic progress and stability [5,6].
In the prehistoric southern Levant, exchange and trade in beads was common practice for more than a hundred thousand years [7,8], and while most were made from locally available raw materials, exotic rocks imported from farther away were also recorded (e.g., obsidian, turquoise, amazonite, amethyst, serpentine [7,[9][10][11]).
Among these exotic minerals, olivine (sometimes called peridot or chrysolite [12][13][14]) is one of the rarest. This mineral is common in igneous rocks, measuring 6.5-7 on the Mohs hardness scale, and is usually yellow-green or green in color. Chemically, olivine is a continuous  [11,49,52], including the gradually increasing significance of olives alongside the use of dairy products [59]. Tel Tsaf is also notable for its numerous non-local exotic finds, indicating the site's participation in a far-reaching exchange network [11,52]. A partial list includes beads from Transjordan, beads and Nilotic shells from Egypt, tokens, and figurines from the northern Levant, obsidian from Anatolia, a copper awl from an indeterminate northern origin [60], Ubaid style pottery from Mesopotamia, and seashells from the Mediterranean coast.
Specifically concerning the occurrence of beads at the site, a massive assemblage of over 2,500 ostrich eggshell beads is notable [44]. Otherwise, dozens of beads were also recoded; many were made from non-local rocks and minerals, while others were produced from clay, bone, wood, and Theodoxus Jordani shells [61,62].  Tables 1 and 2). Seven were found during the current excavation project and one during the 2004-2007 expedition. Six were retrieved from Square AR16, a deep cut in Area C. Of these, one was in a pit, and the others were in disparate accumulations. Of the remainder, one bead was recovered from accumulations north of Room C70 in Building complex CI, and the other was found in fills of a Byzantine/Early Islamic grave.
All eight beads are whole and characterized by a dark to light green color with a tint of yellow hue (Fig 3). Optical examination indicated that all beads are translucent and lack foreign inclusions. They have an oval-lenticular cross-section, and their shape varies from round, through oval, to quasi-triangular. The beads' perforations are biconical, and their two faces are sometimes even.
The olivine beads weigh 0.015-0.2 g and are 1.0-3.0 mm thick. Their smallest measurements across are 3.0-6.0 mm, while their maximum measurements are 3.0-8.0 mm. All perforations have round outlines (Fig 3) and seem to have been drilled from both sides (bidirectional drilling). Their minimum diameter is 0.3-0.6 mm, and their maximum diameter ranges between 0.5 and 1.0 mm across (for both faces).

Methods
All beads underwent attribute analysis at the Laboratory for Ground Stone Tools Research (LGSTR) at the Zinman Institute of Archaeology, University of Haifa. First, the beads' morphometric and contextual properties were documented, including circumstances of discovery, degree of preservation, shape, measurements, color, and profile. Next, aided by a Dino-light, edge 3.0, digital microscope (magnification 10-150×), microwear patterns indicative of production and use were recorded.
At the Institute of Earth Sciences at the Hebrew University of Jerusalem, the beads were subjected to further optical examination with Hirox RH-2000 and Nikon SMZ800 binocular microscopes. Preparing for geochemical analyses, we immersed the beads in epoxy and mildly polished 1 μm off one side, using diamond polishing powder. We completed the treatment by cleaning the beads with ethanol and distilled water in an ultrasonic bath.
All specimens were subjected to Raman spectroscopic analysis at the Center for Nanoscience and Nanotechnology, the Hebrew University of Jerusalem, to determine the beads' mineralogical compositions. We used a Renishaw InViaTM Raman microscope,  equipped with Argon-Ion Laser 50 mW (514 nm) and a Leica DM2500-M microscope. Each spectrum was acquired for 10 seconds of illumination, using a 20× objective and the Renishaw CCD camera (1040×256 px). Laser power for all analyses was set at 5%, and spectra were collected between 100 and 1500 cm -1 . The spectral resolution was 1 cm −1 , and the spectrometer was calibrated against the 520 cm −1 peak of a synthetic silicon plate. Mineral identification was made against spectra of the Hebrew University mineral collection and the online RRUFF™ project spectral data for minerals (https://rruff.info/). The beads were also subjected to elemental analysis at the Institute of Earth Sciences at the Hebrew University of Jerusalem. We employed a JEOL JXA-8230 electron probe microanalyzer (EPMA) equipped with four wavelength dispersive spectrometers. Across all analyses, the electron beam used was 1 μm in diameter with an acceleration voltage of 15 kV and a current of 120 nA. Peak counting rates were 30 s for the major elements (Si, Mg, and Fe) and 180 s for minor elements (Ca, Ni, Cr, Mn, and Al), while background counting rates were half the time for each background position (i.e., 15 s and 90 s, respectively). A set of reference materials were used to calibrate the EPMA for specifiable elements: San-Carlos olivine (NMNH# 111312-44) for Si, Mg, and Fe, SPI standard diopside (MgCaSi 2 O 6 ) for Ca, chromite (FeCr 2 O 4 ) for Cr, rhodonite (MnSiO 3 ) for Mn, pentlandite ((Fe, Ni) 9 S 8 ) for Ni, and an in house HUJI spinel (MgAl 2 O 4 ) for Al. These standards were analyzed as unknowns for calibration control and to monitor instrument stability. The ±2σ analytical uncertainty of 75 analyses of San Carlos olivine produced ±0.46 Fo# [where Fo# is the molar ratio of 100Mg/(Mg+Fe)], ±158 ppm Ni, ±41 ppm Ca, ±127 ppm Mn, ±18 ppm Al and ±24 ppm Cr.

Microwear traces and the production sequence
We observed four types of microwear traces on the beads: scratches, flake scars, pits, and edge rounding. Micro-scratches are widespread and observed on practically all parts of the bead: perforation walls (including both horizontal and vertical scratches; Fig 4A), perforation rims ( Fig 4B and 4G), leveled surfaces (Fig 4E and 4F), leveled surfaces' circumferences ( Fig 4C), and areas between the perforations and leveled surfaces (Fig 4D). Slightly less extensive micro- flake scars were recorded along beads' perforation rims (Fig 4B), the leveled surfaces (Fig 4E), and the leveled surfaces' circumferences ( Fig 4C). Lastly, edge rounding ( Fig 4H) and micropits (Fig 4E and 4F) were observed on leveled surfaces only (Fig 4I).
Drawing on these observations, we may carefully infer the beads' chaînes opératoires. We presume that the scratches were produced by grinding with an abrasive powder [63], whereas micro-flake scars were created by percussion. These would have been the first operations applied to the olivine crystals. Next, the beads' leveled surfaces were produced, removing most abrasion and flake scars, although some remained along the edges (Fig 4C, 4D and 4F). While these surfaces (Fig 4A) seem aesthetically desirable, they could have served a technical function as drill-facilitating platforms [64]. The surfaces' shine might suggest that heat treatment was also applied. Such treatment has been shown to render siliceous matrixes finer [65,66] and a stone's color more intense without altering its translucence or hardness. Further support for heat application in the beads' production process is provided by the observation of luster in fresh breaks of platforms (Fig 4I: a) [65] and the occurrence of intensive red particles in some beads' fabrics (Figs 3: 5-6, 3: 8 and 4I: b).
The beads' perforation was carried out by drilling from both sides, producing a biconical shaft. This procedure was employed to avoid damage or breakage of the bead. The drill was probably made of a hard stone such as flint or quartz [64,67,68], and its application presumably entailed the use of a lubricant [63,69] (Fig 4A).
The last step of the chaîne opératoire was smoothing a bead's surfaces (Fig 4E). The rounded grains and low micro-topography observed on these surfaces suggest that a soft material was used, possibly wood or linen [63,64]. Abrasive dust produced in the grinding stage of the manufacturing process might have also been used for smoothing, a possibility suggested by ethnographic and experimental observations [64,70,71].
We hypothesized above, on the grounds of crystal size, that the olivine beads of Tel Tsaf derive from mantle rocks. Now, our chemical analyses provide further support: the beads' Fo#, Ni, and Mn contents are well within the ranges of mantle olivine, which are 89%-94%, 2200-3400 ppm, and 450-1050 ppm, respectively [73][74][75]. Moreover, Ca and Al concentrations in the beads are below the maximum values for these elements in mantle olivines (~650 and   Table 3. Chemical composition of olivine beads.

Tsaf-Olivine8
Tsaf   150 ppm, respectively). They are also far below their values in other igneous olivines, usually above 1000 and 100 ppm, respectively [74,75]. Thus, in all likelihood, the olivine beads of Tel Tsaf originate in mantle rock fragments. On the Earth's surface, such rocks occur in one of two ways: as xenoliths-where mantle fragments are incorporated in ascending magmas-or as ophiolites-exposed segments of oceanic crust and upper mantle.

Discussion
The eight olivine beads from Middle Chalcolithic Tel Tsaf are the earliest specimens of their kind in the Near East and a rare incidence in the southern Levant. As noted above, six potential olivine sources have been identified: Turkey, Pakistan, Egypt's Eastern Desert and Zabargad, Ethiopia, Tanzania, and Saudi Arabia. Of these, the Fo# of Turkey's and most of Tanzania's olivines are higher (92.5%-94.5% and 93.5%-93.5%, respectively; see Fig 7) than those of the beads from Tel Tsaf and, thus, are unlikely to have provided the raw crystals (Fig 6). Similarly, Pakistan can be set aside due to its olivine's comparatively low Fo# between 77%-85% and Ni/ Mg ratios. The statistical variations of the olivine populations' chemical compositions further reinforce these observations (Fig 7). Mn/Ni and Ca/Al ratios clearly distinguish the olivines from Pakistan from those of Tel Tsaf, while Turkey's and Tanzania's olivines have too high Fo# to be viable candidates for the beads' provenance. None of the remaining three potential sources (Saudi Arabia, Ethiopia, and Egypt) can be considered a substantially more probable source than the others. Nevertheless, it is notable that, statistically, the olivines from Ethiopia are closest to the beads from Tel Tsaf. Either way, one of the most important observations is that the beads are chemically almost identical. This, in turn, suggests they derived from a single source and most probably arrived together at Tel Tsaf, possibly as a single chain or necklace. If indeed the Tel Tsaf olivine beads originate from one of the three potential sources mentioned above, long-distance trade was inevitably implicated (the Ethiopian plateau and the main Ethiopian rift area are over 2,000 km away). The beads had to travel via a long chain of hubs across far-flung and diffused systems of trade and exchange.
In all likelihood, the olivine beads constituted personal objects and were entwined with themes of value, desire, and social status. As a product, they boast high production value in terms of technology and distribution [90]. Furthermore, embodying distance, transportation, and far-reaching contacts conferred onto them high use-value.
The abundance of long-distance traded artifacts and substances at Tel Tsaf indicates that it accommodated a community capable of obtaining and managing surplus to foster economic and social connections. Moreover, along with the circulation of tangible goods, exchange networks also provide an infrastructure for the flow of intangible cultural assets communicated by artifacts that convey coded information, interpreted or read as 'value' or social/economic significance. In the case of the olivine beads, the transmitted coded information pertains to their value that derives from acknowledging their distant origin and the difficulty obtaining them. (c) 100×Mn/Fe against Fo# spanning 86%-95%; (d) 100×Ni/Mg against 100×Mn/Fe. The data for the possible olivine origins were acquired from the following sources: For Turkey [29][30][31], for Pakistan [32][33][34], for Eastern Desert of Egypt [24], for Zabargad Island [21][22][23], for Saudi Arabia [25][26][27][28], for Ethiopia [35][36][37], and for Tanzania [38][39][40]. Several samples are not presented here; they include specimens from the Eastern Desert of Egypt with Fo>97% and Ni<2000 ppm and a few specimens from Turkey and Pakistan with Fo = 100. https://doi.org/10.1371/journal.pone.0271547.g006 However, modeling this multi-layered interaction system is nearly impossible due to numerous indeterminate agents and ambiguous factors. Nevertheless, we have good reason to presume that the value bestowed on exotic materials [91] is likely to have fed the motivation to obtain these beads. The properties of olivine (i.e., color, translucence, and reflectivity) coupled with its distant origins might have been instrumental for its far-flung circulation. The olivine beads' significance should be appreciated not only for their rare exoticism but also as indicators of Tel Tsaf's constitution as a pan-regional hub that engaged in (and probably controlled) superregional exchange during the late sixth-early fifth millennia cal. BC.