This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

The red-stained flint crescent from Gesher: new insights into PPNAhafting technology

Dana Shaham*, Leore Grosman, Naama Goren-InbarInstitute of Archaeology, Hebrew University of Jerusalem, Mt. Scopus, Jerusalem 91905, Israel

a r t i c l e i n f o

Article history:Received 17 January 2010Received in revised form14 March 2010Accepted 17 March 2010

Keywords:PPNAHafting technologyOchreMud plasterGesherLevant

a b s t r a c t

A red-stained flint crescent found in the Pre-Pottery Neolithic A (ca. 9000 BC) site of Gesher, Israel,provides us with a unique opportunity to study the hafting technology of a particular tool type in theLevant. We present here a reconstruction, based on chemical and mineralogical analyses, of the haftingtechnologies and materials employed in the process. Use will also be made of relevant studies of thearchaeological record. Local material was used for the production of an elastic adhesive paste, mudplaster, which was then hardened to hold the crescent securely in the haft. The study contributes insightinto the hafting technology that was most probably implemented in the production and maintenance ofcomposite tools during the Early Neolithic period.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Reconstructing the hafting technologies of prehistoric culturesis a demanding task, as the archaeological record is fragmented andin most cases it is only the hafted stone artifact that is preserved.While flint assemblages can provide unequivocal data pertaining toprehistoric tool production, hafting technologies can only rarely bereconstructed.

Hafting technologies can be classified into three basic types:wedged hafts, where the tool is inserted into a slot in the handle andis retained bymechanical forces; tied hafts, where the implement islashed to the handle; andmastic hafts, where the tool is attached tothe handle by glue, resin, tar, or plaster (Keeley, 1982). Mastic haftsare usually very reliable, allowing virtually no movement of thehafted part. As hardened mastic is more elastic than either theinsert(s) or the handle, it provides a cushioned setting that reducesthe chance of breakage. The preparation of mastics, however, maydemand the investment of more time and equipment than theother arrangements, because they require heating to increase theirviscosity. In addition, many vegetal pitches and gums havea tendency to crystallize with dehydration, which must be coun-teracted by “tempering” with fibrous particles, charcoal, grit, orother materials (Keeley, 1982).

A flint crescent-shaped artifact with red staining, 5.9 cm long,found in the Pre-Pottery Neolithic A (PPNA) site of Gesher, Israel,can provide evidence for mastic hafting technology in the Levant.The results of chemical and mineralogical analyses carried out onthe red stains, together with archaeological data, allow the recon-struction of a specific hafting technology using a particular adhe-sive paste. The crescent was probably part of a composite tool withseveral key elements: a handle made of bone, wood, etc., a haftedimplement or implementsmade, for example, of flint, and a bindingadhesive.

Ethnographic records attest to a wide range of mastic materialsused for hafting: glue obtained from the bulbous root (Krige, 1965),hot porcupine grass resin (Basedow, 1929), boiled pieces of buffalophallus (Ewers,1958), beeswax pastewith ashes (Suárez,1968), andwarmed pitch of bull pine (Pinus sabiniana) (Gifford, 1965), are onlya few examples. They illustrate a great variety of materials, whichare frequently gathered locally, and diverse procurement methods.

Middle Paleolithic cultures furnish data on hafting technologyas a worldwide archaeological phenomenon. The earliest evidencecomes from Campitello Quarry, central Italy, where a birchbark-tar adhesive was found (Mazza et al., 2006). Plant twine, resin,and ochre used as an adhesive are reported from the Middle StoneAge layers of Sibudu Cave and other sites in South Africa (Wadleyet al., 2004, 2009; Lombard, 2005, 2007, 2008; Wadley, 2005).Artifacts with residue of bitumen adhesive are known from theLevantine Middle Paleolithic sites of the el-Kowm basin, Syria(Boëda et al., 1998, 2008). The site of Quneitra, Golan Heights,

* Corresponding author. Tel.: þ972 2 5882118; fax: þ972 2 5825548.E-mail addresses: dana.shaham@mail.huji.ac.il (D. Shaham), lgrosman@mscc.

huji.ac.il (L. Grosman), goren@cc.huji.ac.il (N. Goren-Inbar).

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http: / /www.elsevier .com/locate/ jas

0305-4403/$ e see front matter � 2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.jas.2010.03.006

Journal of Archaeological Science 37 (2010) 2010e2016

Author's personal copy

yielded a Mousterian flint scraper bearing an unusual patinapattern on both faces. Experimental analysis supported thesuggestion that these patterns resulted from the hafting of thetool in a handle (Friedman et al., 1995).

The Middle to Upper Paleolithic transition in the Levant wasprobably accompanied by two typo-technological features affectinghafting: the appearance of the butt-thinned Emireh point, typical ofthe initial Upper Paleolithic (see Marks, 1983), and the beginning ofmicrolithization (Clark, 1959; Bar-Yosef, 1987, 2002). The increaseduse of microliths during the Epi-Paleolithic was clearly associatedwith the invention of a hafting technology for composite tools(Belfer-Cohen and Goring-Morris, 2002).

Pyrotechnology was introduced to the Levant during this periodin the form of the production of, among others, lime plaster(Kingery et al., 1988). The Geometric Kebaran site at Lagama NorthVIII, Sinai (Bar-Yosef and Goring-Morris, 1977), yielded severalmicroliths with adhesive along the longitudinal axis. Examinationof its microstructure and chemical composition, which is a nearlypure calcium carbonate, identified the adhesive material as limeplaster (Kingery et al., 1988). This technology was also identified atthe late Natufian site of Saflulim, central Negev, Israel (Goring-Morris, 1999).

A few examples of hafted sickle blades in handles have beenfound in Natufian sites (El-Wad Cave: Garrod and Bate,1937: pl. XII;Kebara Cave: Turville-Petre, 1932; Oumm ez-Zoueitina: Neuville,1951: pl. XIII: 24; Wadi Hammeh 27: Edwards, 1987, 1991). Thebone handle found by Garrod (Garrod and Bate, 1937) in El-Wadcave was bedded in tough red earth and the blades were held inplace “. by a calcareous concretion. all trace of the originalsticking matter having disappeared” (Garrod and Bate, 1937: 37).Tomenchuk argued (cited in Bar-Yosef, 1987) that the calcareousconcretion was indeed the limestone-based adhesive material ofthis specimen rather than a secondary calcareous deposit.

Backed bladelets, La Mouillah points, and geometric microlithisof the Late Epi-Paleolithic Madamaghan industry in southernJordan often display red or yellow ochre on their retouched edges(Henry and Shen, 1995). Microscopic study confirmed the correla-tion between hafting abrasion polish and the distribution of ochreon the tools. Although traces of resin were not observed, it wassuggested that organic material served as hafting adhesive whilethe ochre was used as a decoration with symbolic meaning (Henryand Garrard, 1988).

The Levantine Epi-Paleolithic cultures provide evidence for bothlimestone-based and possible resin adhesives (Bar-Yosef,1987), andthe European Mesolithic artifacts (e.g., Curwen, 1941) have yieldedevidence for birchbark-tar or resin adhesives (e.g., Keeley, 1982);while the North African Capsian cultures used a blend of organicmaterial and ochre as an adhesive (Inizan, 1976; Beyries, 1983). Thepresence of ochre traces and the microscopic analysis’ results of theochrewere interpreted as part of a structured process, whereby theartifacts were dipped in a liquid made of gypsum plaster andhematite as part of the hafting technology (Inizan, 1976).

The PPNA period is marked by major changes in human cultureand socio-economic behavior such as subsistence modes, as attes-ted by the beginning of cultivation, sedentary lifestyle, and socialcomplexity. Yet no major technological change within the realm ofhafting technology has been observed to date in the archaeologicalrecord. Cauvin (1983) suggested that resin was frequently used inthe Neolithic of the Euphrates to haft sickle blades but was rarelypreserved, hence its absence from the archaeological record. In thesouthern Levant, especially in the vicinity of the Dead Sea, therewas an abundance of bitumen, which explains the presence of blacktraces on flint tools as an adhesive for hafting. Examples are theelongated blades (“Beit Ta’amir knives”), from Salibiya IX and NetivHagdud, both PPNA sites located in the southern Jordan Valley

(Bar-Yosef and Gopher, 1997), and the sickle made of horn(Nissenbaum, 1997) from Nahal Hemar Cave in the Judean Desert,with three blades inserted in a groove, each separated from thenext by a patch of bitumen (Bar-Yosef, 1987). Cauvin herself studiedsickle blades from Tell Assouad, Syria, and showed that half of themretained traces of blackish material suggested to be bitumen(Cauvin, 1973).

The selected cases of hafting technology presented abovedemonstrate both the early origin of hafting and the variability ofthe prehistoric record. The objective of the present study is toexamine in detail the red substance that appears on the Geshercrescent. The various analyses provide further information on thematerials and the technological processes used by the PPNA groups.

1.1. The crescent from Gesher

The PPNA site of Gesher is located in the central Jordan Valleyof Israel. It was excavated in 1986e7 by Y. Garfinkel (Garfinkel,1989, 1990a, 1990b; Horwitz and Garfinkel, 1991; Garfinkel andBelitzky, 2005; Garfinkel and Dag, 2006). It is situated on thesouthern bank of Nahal Tavor and the Jordan River flows about1 km east of the site (Garfinkel and Dag, 2006: 4e6, Fig. 1.1, 1.2). TheNeolithic layer, dated to 9300 calBC (ibid.), accumulated on top ofvarved strata of the Lisan Formation, which is interstratified withsandy layers. This anthropogenic layer is covered by sediments ofthe Faza’el Formation (ibid.: 32).

Architectural remains and a flint assemblage with typical PPNAtool types were identified; they include sickle blades and thecrescent-shaped artifact, together with a variety of other tools(ibid.: Table 4.3).

The crescent-shaped item under discussion is 5.9 cm long and2 cm wide. It is abruptly retouched along its curved edge (Fig. 1).Light-colored encrustation appears in no particular pattern on bothventral and dorsal surfaces. Red staining is present on both faces

Fig. 1. The PPNA flint crescent from Gesher with red staining; left: dorsal face, right:ventral face (scale¼ 2 cm).

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e2016 2011

Author's personal copy

and has a very particular pattern: on the dorsal face it extends fromthe abruptly retouched curved edge to the longitudinal ridge, whileon the ventral face the patterning exhibits a mirror image of thedorsal one, ending in a straight longitudinal line. Thus, the stainingon both faces of the crescent spreads over 1.3 cm of its maximumwidth. While it is currently more intense and homogenous on thecrescent’s dorsal face than on the ventral one, it is quite obviousthat originally it was evenly distributed within the observed“boundaries”.

By its particular distribution, the red staining is considered to beassociated with hafting the crescent in a shaft. In order to test thisassumption and understand the specific technology used by thePPNA people of Gesher, we present below the results of the anal-yses of the composition of this red staining.

2. The red stains: chemistry and mineralogy

2.1. Methods

A variety of analyses were carried out on the crescent, focusingon the red staining, the flint, and the encrustations. These analyseswere aimed to explore the various aspects of the material in orderto investigate the possible hafting technology. In addition to visualexamination, the morphology and local chemical composition ofthe samples were analyzed with Environmental Scanning ElectronMicroscope (ESEM) Quanta 200 (FEI Company, USA) equipped withenergy dispersive X-Ray spectroscopy (EDS) (EDAX-TSL, USA).Analyses of the mineralogical composition were carried out by X-ray powder diffractometer (Bruker AXS, Karlsruhe, Germany) witha goniometer radius of 217.5 mm, Göbel Mirror parallel-beamoptics, 2� Sollers slits and 0.2 mm receiving slit. XRD patternswithin the range 10� to 55� 2q were recorded at room temperatureusing CuKa radiation (l¼ 1.5418�A) with the following measure-ment conditions: tube voltage of 40 kV, tube current of 40 mA,step-scan mode with a step size of 0.02� 2q and counting time of1 s/step.

Scanning Electron Microsope (SEM), environmental SEM, Elec-tron Dispersive Spectrum (EDS), X-Ray Dispersion (XRD), and InfraRed Spectrometry (FTIR) were all used in order to define andcharacterize the red staining.

(1) Scanning Electron Microscope (SEM): Three ground sampleswere examined, originating from: (a) red stains on the ventralface, adjacent to the distal edge; (b) encrustation on the ventralface, adjacent to the distal edge, with no coloring; (c) dark redstains on the dorsal face.

(2) Environmental SEM: The crescent was placed inside the ESEMchamber and analyses were carried out on the red stains, witha control test on the unstained flint surface. Specific exami-nations were performed on various features observed on thesurface of the crescent.

(3) Energy dispersive spectroscopy (EDS): EDS analyses werecarried out under the SEM in order to identify the chemicalcomposition of different features observed. Accordingly,samples were analyzed on the dorsal and ventral faces of thecrescent and from red stains and unstained areas.

(4) X-Ray diffraction (XRD): Ground powder samples from threedifferent areas of the crescent were examined in order to revealthe crystallography and mineralogy phase. These were carriedout on: (a) the bulk flint without any red stains; (b) the flintwith red stains; and (c) a ground powder of the red stains.Subsequently, the same ground sample of red-colored materialwas examined under the SEM to observe its grained miner-alogy patterns.

(5) Fourier Transform Infra Red Spectrometry (FTIR):We examineda sample of ground powder (1 mg) of the red material,extracted from the dorsal face near the distal edge in order totest whether it was burned.

2.2. Results

2.2.1. The red stainingThe red material is composed of a thin bed of vegetal material

placed on a clayey substance, as illustrated by the SEM photograph(Fig. 2). Environmental SEM and EDS analyses demonstrate that thered staining on the dorsal face of the crescent (Table 1:1) iscomposed of silica, iron oxides, sodium, magnesium, and oxygen.

Fig. 2. SEM photos of the red stains: (a) vegetal fibers placed on a substratum (8111, 15KV, WD 37); (b) close-up of the vegetal fibers (8114, 15 KV, WD 37); (c) close-up ofimprinted vegetal fibers (8116, 15 KV, WD 37).

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e20162012

Author's personal copy

The quantitative analysis of these elements shows a high amount ofiron and silica, and a lower amount of calcium carbonate.

Within the redmaterial, some special features were recognized:

a) White inclusions of barium and sulfur (BaeSO4), a naturalmineral usually found in lacustrine sediments.

b) Organic particles of elongated morphology, which are vegetalmaterial with 75% atoms of carbon (Table 1:4, 5), are strewn onthe red substratum and sometimes pressed into it.

c) A rounded feature in the red material which chemicallycontains, among other materials, phosphorus and what couldpossibly be a bone fragment (Table 1:6).

Thus, the red material is composed of a thin bed of vegetalmaterial placed on a deposit of clay made of calcites, silicates, andsome inclusions with iron oxides (Fig. 2, Table 2).

Extensive SEM and chemical analyses of the red substance onthe ventral face of the crescent obtained results identical to thosefor the dorsal face.

The XRD analysis (Fig. 3) furnished a mineralogical identifica-tion of the red material, which originated from the combination ofcalcite and quartz and was identified as calcium carbonate (CaCO3).It also showed that iron is not dominant in the sample but is rep-resented only by small particles.

The FTIR spectrometry analysis of the ground red powderrevealed a composition of an equal ratio of calcite and clay (Fig. 4).Generally, a calculation of two FTIR peaks height ratio (h.874/h.713> 4) allows the determination of whether a sample wasburned, but the current analyses could not indicate whether or notheating had occurred (see, e.g., Chu et al., 2008: 908 and referencestherein). The results of the ground red powder analyses (874/713¼ 2.932) show that neither the calcite nor the clay was burned(Fig. 4). The FTIR analyses of the iron mineral did not indicate thepresence of goethite, but a comparison of the crescent’s spectrumwith a hematite standard shows a similarity between the two, andtherefore the presence of hematite can be assumed (Fig. 4).

2.2.2. The bulk flintSEM examination of the area without the red staining showed

that it has a homogenous chemical composition, devoid of vegetalmaterial. As expected, the Environmental SEM analysis on thedorsal face of the crescent revealed that this surface is composedmostly of silica (Si¼ 48%), carbonate, and low percentages of iron,

Table 1Results of EDS analyses under the environmental SEM.

% Atom of each chemical element

Chemicalelement

Sample number

1 2 3 4 5 6 7 8

C 15.66 14.49 15.24 75.99 75.42 66.27 20.86 24.59O 44.87 18.84 45.73 17.95 19.45 18.67 42.42 41.24Na 0.43 0.44 0.76 0.16 0.19 0.77 0.37 0.46Mg 0.78 0.62 0.24 0.16 0.25 0.42 0.81 1.07Al 2.73 1.12 1.02 0.5 0.6 1.15 2.33 3.34Si 8.94 10.14 7.3 3.09 2.11 5.19 12.38 15.63S 0.18 0.16 13.2 0.08 0.08 0.14 0.15 0.04K 0.39 0.22 0.3 0.15 0.11 0.23 0.48 0.48Ca 1.1 0.67 1.09 0.75 0.8 2.56 3.65 2.59Ti 0.4 0.16Fe 24.51 48.02 1.67 1.04 0.88 3.82 16.05 9.98Cl 0.2 0.04MnBa 13.26 0.1 0.05 0.15 0.28 0.24Cr 5.11 0.18 0.02 0.06 0.03 0.22 0.18P 0.4 0.14

The samples: (1) General sample of the red staining, dorsal face; (2) Generalsample of the red staining, dorsal face; (3) Zone of intense staining (as observedunder the SEM), dorsal face; (4) Particular feature observed under the SEM withmorphology of organic material, dorsal face; (5) Particular feature observed underthe SEM with morphology of organic material, dorsal face. Similar features areobserved on the homogeneous substratum, and sometimes slightly imprintedinside it; (6) Sample of the rounded contour observed on the dorsal face;(7) Encrustation coating, ventral face; (8) Diatom (phytoplankton) sunken insidethe red staining, ventral face.

Table 2Results of EDS analyses under the SEM (ground sample of the red staining).

% Atom of each chemical element

Chemicalelement

Sample number

1 2 3 4 5

Mg 1.94Al 10.405 15.855 15.608 14.999 4.075Si 10.367 16.614 16.946 16.053 10.673K 0.382 0.444 0.4 0.626 0.654Ca 20.08 2.229 2.067 4.989 5.619Ti 0.601Fe 0.172 0.171 0.269 19.119S 0.276 1.048 1.104 0.496 0.318Cl 0.149 0.161O 58.05 63.309 63.455 62.2 56.853

The samples: (1) The fiber mat; (2) Imprinted substratum; (3) Fibers on thesubstratum, general view; (4) General area of fibers and clay (kaolinite); (5) Grain ofochre-like red color.

Fig. 3. Comparison of XRD analyses on the red stains and the bulk flint of the crescent.Red: PDF (Powder Diffraction File) 86-1560, quartz, SiO2; Blue: PDF 86-2339, calcite, Ca(CO3); Green: PDF 89-6538, kaolinite A12(Si2O5)(OH)4.

Fig. 4. FTIR spectrometry analysis: ground sample (1 mg) of the red stain on the dorsalface, adjacent to the distal edge of the crescent.

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e2016 2013

Author's personal copy

calcium, and magnesium. The high frequency of silica is evidentlytypical of flint.

2.2.3. The encrustationAn additional feature, which was also identified by the SEM

analyses of the ventral face, is a post depositional encrustationcomposed of silica, magnesium, calcium, potassium, oxygen, anda low percentage of barium (Table 1:7).

Several conclusions concerning the adhesive material could bedrawn from the results described above:

1. The chemical elements that comprise the red stains are silica,iron oxides, carbonates, calcium, oxygen, sodium, magnesium,and alumina, as well as small amounts of phosphorus andbarium. The XRD and FTIR analyses of the mineralogicalcomposition indicate that these elements are components ofcalcite (CaCO3), quartz (SiO2), and clays. These analyses alsoshow small amounts of iron in the samples (in contrast to itsdominance in the chemical analysis). It can be concluded thatthe adhesive paste was produced from a mixture of calcite andclay (hereafter paste). The manipulation or modification of thematerials required for the production of the adhesive paste willbe discussed below.

2. The SEM photo (Fig. 2) along with the results of the chemicalanalyses (Table 2:1, 3) show that a layer of organic vegetalmaterial was laid over the paste. Thus, the adhesive is composedof two layers: the paste and the vegetal matrix above it.

3. The red color of the paste is derived from iron oxides, whichcould derive either from minerals that occur naturally in theclay or minerals that were added intentionally (seeDiscussion). A clue as to which of the two options is validcould be provided by the varied quantities of iron observed inthe different analyses. While iron oxide was dominant insome of the EDS analyses, in the XRD analysis only lowquantities of it were found. One explanation for these differ-ences could be that the iron oxides in the paste had a granularmorphology and therefore were not distributed homoge-nously. An additional clue is provided by the FTIR analysis,which showed the presence of hematite. While the spectrumband of our sample is not identical to the hematite standardband, as shown by Helwig (1998), this variation is most oftencaused by differences in the size and shape of the hematiteparticles. Furthermore, hematite-based pigments that havebeen processed in different ways can be distinguished by theirparticle morphology (ibid.).

4. The FTIR analyses indicate that the adhesive paste was notburned. This was supported by the results of the EnvironmentalSEM analyses, in which a diatom was visible (phytoplankton,Fig. 5); if the material had been burned, we could not haveidentified the diatom’s morphology (Goren et al., 2001). Onceburning is ruled out, heating of the paste would have been theonly way in which a homogenous red stain could have beencreated.

3. Discussion

In the following discussion we shall attempt to identify andreconstruct the sequence of the activities undertaken by the PPNApeople that resulted in the hafting of the crescent.

3.1. Is it ochre?

The term ochre deserves explanation, due to its wide use andvaried definitions based primarily on a reddish color. The reported

list of minerals defined as ochre includes hematite (Fe2O3), limonite(2Fe4O3 3H2O), mudstone (marl), goethite (F2O3 3H2O), siltstone,and cinnabar (Wreschner, 1983). Although the term ochre is usedfrequently, only one mineralogical archaeological study of haftingadhesive presents the composition of its samples (Inizan, 1976).Two compositions were given there for red stains: one consists ofgypsum (80%)þ hematiteþ a small amount of quartz, and the otheris calciteþ quartzþ hematite. The latter composition is similar tothat obtained by the mineralogical analysis of the red paste of thecrescent. It includes two of the items in the list of ochre mineralscompiled by Wreschner (1983): marl and a small quantity ofhematite. Therefore, mineralogically the red stains on the crescentmay be defined as ochre.

3.2. Hafting

Examples of ochre associated with hafting technology arereported from the Middle Stone Age record of South Africa (Wadleyet al., 2004, 2009; Lombard, 2005;Wadley, 2005), the Epi-PaleolithicIberomaurusian culture of Algeria (Ighilahriz, 1996), and the Capsiancultures of the Maghreb (Inizan, 1976; Beyries, 1983).

Ochre minerals were found to be used for the loading and fillingof adhesives, for example a combination of ochre and organic resinor wax (for experimental studies see Wadley, 2005; Wadley et al.,2009; for an archaeological context see Inizan, 1976). Conversely,the adhesivemade by the PPNA people includes a siltyeclayey mudand ochre-like minerals (marl and hematite) as a pastewith loaderslike quartz and vegetal material atop of it.

3.3. Mud plaster

The Gesher ochre paste served not as a loader but as the mainadhesive component, resembling a mud plaster. Plaster technology,known from the Epi-Paleolithic of the Levant and peaking in thisarea during the PPNB, includes three types: mud, gypsum, and limeplasters (Gourdin and Kingery,1975; Kingery et al., 1988; Goren andGoring-Morris, 2008). The latter two require firing temperatures inthe order of 300 �C to 750e850 �C respectively; both demonstratehardness and high resistance, characteristics that made thema desirable choice for several different functions (Gourdin andKingery, 1975). While the technologies of lime and gypsum plas-ters are fairly well known, that of mud plaster requires furtherstudy, though it is known from as early as the Natufian culture (e.g.,Garrod and Bate, 1937; Valla et al., 2007; Grosman et al., 2008). Ina study of PPNB plastered skulls (Goren et al., 2001), at least some of

Fig. 5. Environmental SEM photograph (HUJ) of the red stains, showing a diatom(rounded morphology) and two flint microartifacts located to its right.

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e20162014

Author's personal copy

the skulls were regarded as mud-plastered. The mud plaster wascomposed of clays, iron minerals (probably hematite), and localwadi sand fillers, all collected from the vicinity of the site. Thesedata further indicate that the mud plaster was heated at a lowtemperature and not fired.

Since the mineralogical composition of the Gesher crescent issimilar to the examples of mud plaster discussed above, its pastemay be defined as mud plaster.

3.4. Manipulating the mud plaster

Though the plaster adhered to the flint crescent is not burned,the treatment required to transform an elastic adhesive clay intoa hard substance is heating (Keeley, 1982: 800;Wadley, 2005: 596).Experiments have shown that the transformation of yellowgoethite to red hematite requires heating at a low temperature(260e280 �C) (Weinstein-Evron and Ilani, 1994; Pomiès et al.,1998). The red color of the crescent’s mud plaster suggests justsuch a transformation, since heating at a higher temperaturewouldhave resulted in black rather than red hues. To avoid overheating,prehistoric people could have heated the goethite either at thefringes of the hearth, where temperatures were lower, or by placingthe object on stones laid on hot ashes (Weinstein-Evron and Ilani,1994).

3.5. Reconstruction of the hafting technological process

Taking into consideration the entire set of analyses presentedabove, we propose a possible reconstruction of the process of thecrescent’s hafting. The inhabitants of Gesher used local availablematerials for hafting and were knowledgeable about the specialcharacter and qualities of the local natural resources, manipulatingthese materials in order to obtain a desired product. The local sil-tyeclayey loam sediments and local iron oxides, either goethite orhematite, were ground together to form a powder that includeda loader/filler of quartz (e.g., local wadi sand) and anthropogenicsediments containing charcoal, bone, and flint microartifacts.Water was then added to obtain a smooth elastic paste (mudplaster). The paste was then flattened and on top of it a fibrousvegetal material was placed, possibly to create friction between thefine-grained mud plaster and the flint. In the next phase the elasticviscose mud plaster was manipulated to partially coat the crescent.

In order to fix the crescent in the adhesive mud plaster, it wasthen heated at a low temperature, either on the fringe of the hearthor with the use of hot stones. The result was a hard yet elasticadhesive, which securely held the crescent in place.

3.6. The nature of the composite tool

The adhesive-coated crescent was either used on its own orincorporated within a composite tool that also comprised a haft.While various materials could serve as a handle in composite tools,the ethnographic and archaeological records show a predomi-nance of wood, horn, and bone. In all of the archaeologicalLevantine examples of composite tools, the handle is made of boneor horn. In connection with the possibility of composite tools, fourbasic tool types have been suggested (Anderson-Gerfaud, 1983:82, figure 2): transverse arrowhead (ibid.: 81), barb (Clark, 1959:171; Büller, 1983: 111), knife (Anderson-Gerfaud, 1983: 86), andsickle (Curwen, 1941: 335; Cauvin, 1973, 1983; Helmer, 1983;Anderson-Gerfaud, 1988: 185; Copeland and Verhoeven, 1996),of which the latter is the most frequent. Bar-Yosef (1987) hassuggested that while the reconstruction of the Natufian lunates astransverse arrowheads is the most reasonable for the Levant, thelarger PPNA crescents should be associated with sickles.

4. Summary

The results of the above analyses provide us with a uniqueopportunity to reconstruct the complex technological chaîne opér-atoire employed in the hafting of the crescent. A detailed recon-struction has been offered of the use of adhesive material made ofmud plaster and a vegetal component. Most probably this tech-nology and its variants were also used for the manufacturing ofnumerous artifacts, including baked clay beads retrieved at Gesher(Garfinkel and Dag, 2006: figure Xa, 6.3e6.4).

There are indications that variants of this technology wereknown as early as the Epi-Paleolithic (e.g., Bar-Yosef and Goring-Morris, 1977; Valla et al., 2007), but it became more widespreadand elaborate during the PPNA period. The latter was a period oftransformation accompanied by significant changes in the subsis-tence strategy, and in particular the beginnings of agriculture inarable areas (Bar-Yosef, 1992). This transformation necessitatedmajor technological innovations in order to supply new demands.Following this initial phase, when lime plaster technology wasintroduced during the PPNB, plaster technology flourished exten-sively and was widely, intensively, and variably used.

Acknowledgment

This work is based on the find unearthed by Yosef Garfinkel,who kindly allowed us to study it and present the results here; weare most thankful for his generosity. We thank Ruth Shahack-Grossfor assisting with the FTIR analysis of samples and the staff of theunit for Nanocharacterization, the HarveyM. Krueger Family Centerfor Nanoscience and Nanotechnology at the Hebrew University ofJerusalem (HUJ), particularly Vitaly Gutkin and Vladimir Uvarov fortheir assistance and the production of Figs. 3e5. Special thanks aredue to Michael Dvorachek for the SEM analysis and the productionof Fig. 2 and Shimon Ilani, both of the Geological Survey of Israel, forhelpful discussions. Ella Werker defined the botanical remains asfibers and Gabi Laron photographed Fig. 1. Anna Belfer-Cohen andFrançois Valla provided invaluable comments and corrections. SueGorodetsky edited the manuscript with her usual professionalismand dedication. The authors wish to thank the five anonymousreviewers for their important comments and suggestions.

References

Anderson-Gerfaud, P., 1983. A consideration of the use of certain backed and“lustred” stone tools from Late Mesolithic and Natufian levels of Abu-Hureyraand Mureybet (Syria). In: Cauvin, M.-C. (Ed.), Traces d’Utilisation sur les OutilsNéolithiques du Proche Orient. Maison de L’Orient, Lyon, pp. 77e105.

Anderson-Gerfaud, P., 1988. Using prehistoric stone tools to harvest cultivated wildcereals: preliminary observations of traces and impact. In: Beyries, S. (Ed.),Industries Lithiques Tracéologie et Technologie. Aspects archéologiques, vol. 1.BAR International Series, Oxford, pp. 176e196.

Bar-Yosef, O., 1987. Direct and indirect evidence for hafting in the Epi-Palaeolithicand Neolithic of the Southern Levant. In: Stordeur, D. (Ed.), La Main et l’Outil:Manches et Émmanchements Préhistoriques. Maison de L’Orient, Lyon, pp.155e164.

Bar-Yosef, O., 1992. The Neolithic period. In: Ben Tor, A. (Ed.), The Archaeology ofAncient Israel. The Open University, Israel.

Bar-Yosef, O., 2002. The Upper Paleolithic revolution. Annual Review of Anthro-pology 31, 363e393.

Bar-Yosef, O., Goring-Morris, A.N., 1977. Geometric Kebaran occurrences. In: Bar-Yosef, O., Phillips, J.L. (Eds.), Prehistoric Investigations in Gebel Maghara,Northern Sinai. The Hebrew University, Jerusalem, pp. 115e148.

Bar-Yosef, O., Gopher, A., 1997. An Early Neolithic Village in the Jordan Valley e PartI: The Archaeology of Netiv Hagdud. Peabody Museum of Archaeology andEthnology Harvard University, Cambridge.

Basedow, H., 1929. The Australian Aboriginal. F.W. Preece and Sons, Adelaide.Belfer-Cohen, A., Goring-Morris, A.N., 2002. Why microliths? Microlithization in the

Levant. In: Elston, R.G., Kuhn, S.L. (Eds.), Thinking Small: Global Perspectives onMicrolithic Technologies. Archeology Papers of the American AnthropologicalAssociation, Arlington, Virginia, pp. 57e68.

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e2016 2015

Author's personal copy

Beyries, S., 1983. Fonction et mode d’utilisation d’une série de lames ocrées Cap-sienne. In: Cauvin, M.-C. (Ed.), Traces d’Utilisation sur les Outils Néolithiques duProche Orient. Maison de L’Orient, Lyon, pp. 135e142.

Boëda, E., Connan, J., Muhesen, S., 1998. Bitumen as hafting material on MiddlePaleolithic artifacts from El-Kowm basin, Syria. In: Akazawa, T., Aoki, K., BarYosef, O. (Eds.), Neandertals and Modern Humans in Western Asia. PlenumPress, New York, pp. 181e204.

Boëda, E., Bonilauri, S., Connan, J., Jarvie, D., Mercier, N., Tobey, M., Valladas, H., AlSakhel, H., Muhensen, S., 2008. Middle Palaeolithic bitumen use at Umm el Tlelaround 70 000 BP. Antiquity 82, 853e861.

Büller, H., 1983. Methodological problems in the microwear analysis of toolsselected from the Natufian sites of El-Wad and Ain-Mallaha. In: Cauvin, M.-C.(Ed.), Traces d’Utilisation sur les Outils Néolithiques du Proche Orient. Maisonde L’Orient, Lyon, pp. 107e133.

Cauvin, M.-C., 1973. Problémes d’émmanchement des faucilles du Proche-Orient:les documents de Tell Assouad (Djezireh, Syrie). Paléorient 1, 101e106.

Cauvin, M.-C., 1983. Les faucilles préhistoriques du Proche-Orient données mor-phologiques et fonctionnelles. Paléorient 9, 63e79.

Chu, V., Regev, L., Weiner, S., Boaretto, E., 2008. Differentiating between anthro-pogenic calcite in plaster, ash and natural calcite using infrared spectroscopy:implications in archaeology. Journal of Archaeological Science 35, 905e911.

Clark, J.D., 1959. The Prehistory of Southern Africa. Penguin Books, London.Copeland, L., Verhoeven, M., 1996. Bitumen-coated sickle-blade elements at Tell

Sabi Abyad II, Northern Syria. In: Kozlowski, S.K., Gebel, H.G.K. (Eds.), NeolithicChipped Stone Industries of the Fertile Crescent and their Contemporaries inAdjacent Regions. Ex Oriente, Berlin, pp. 327e330.

Curwen, E.C., 1941. Some food-gathering implements e a study in Mesolithictradition. Antiquity XV, 320e336.

Edwards, P.C., 1987. Late Pleistocene Occupation in Wadi al-Hammeh, Jordan Valley.University of Sydney, Sydney.

Edwards, P.C., 1991. Wadi Hammeh 27: an early Natufian site at Pella, Jordan. In: BarYosef, O., Valla, F.R. (Eds.), The Natufian Culture in the Levant. InternationalMonographs in Prehistory, U.S.A., pp. 123e148.

Ewers, J.C., 1958. The Blackfeet: Raiders of the Northwestern Plains. University ofOklahoma Press, Norman.

Friedman, E., Goren-Inbar, N., Rossenfeld, A., Marder, O., Burian, F., 1995. Haftingduring Mousterian times e further indications. Journal of the Israel PrehistoricSociety 26, 8e31.

Garfinkel, Y., 1989. The pre-pottery Neolithic A site of Gesher. Journal of the IsraelPrehistoric Society 22, 145.

Garfinkel, Y., 1990a. Excavations at Gesher e a pre-pottery a neolithic site anda middle bronze Age IIA cemetery. Qadmoniot 23, 26e31 (in Hebrew).

Garfinkel, Y., 1990b. Gesher, un nouveau site “Néolithique Précéramique a” dans laMoyenne Vallée du Jourdain, Israël. L’Anthropologie 94.

Garfinkel, Y., Belitzky, S., 2005. Late Pleistocene and Holocene tectonic deformationat the Gesher site, Kinnarot Valley, Dead Sea Rift. Israel Journal of Earth Sciences54, 133e143.

Garfinkel, Y., Dag, D. (Eds.), 2006. Gesher: A Pre-Pottery Neolithic A Site in theCentral Jordan Valley, Israel, A Final Report. Ex-Oriente, Berlin.

Garrod, D.A.E., Bate, D.M.A., 1937. The Stone Age of Mount Carmel. Clarendon Press,Oxford.

Gifford, E.W., 1965. The Coast Yuki. Sacramento Anthropological Society, Sacra-mento State College, Sacramento.

Goren, Y., Goring-Morris, A.N., Segal, I., 2001. The technology of skull modeling inthe Pre-Pottery Neolithic B (PPNB): regional variability, the relation of tech-nology and iconography and their archaeological implications. Journal ofArchaeological Science 28, 671e690.

Goren, Y., Goring-Morris, A.N., 2008. Early pyrotechnology in the Near East:experimental lime-plaster production at the Pre-Pottery Neolithic B site of KfarHaHoresh, Israel. Geoarchaeology 23, 779e798.

Goring-Morris, A.N., 1999. Saflulim: a Late Natufian base camp in the Central Negevhighlands, Israel. Palestine Exploration Quarterly 131, 36e64.

Gourdin, W.H., Kingery, W.D., 1975. The beginnings of pyrotechnology: Neolithicand Egyptian lime plaster. Journal of Field Archaeology 2, 133e150.

Grosman, L., Munro, N.D., Belfer-Cohen, A., 2008. A 12,000-year-old shaman burialfrom the southern Levant (Israel). Proceedings of the National Academy ofSciences 105, 17665e17669.

Helmer, D., 1983. Les faucilles et les gestes de la moisson. In: Cauvin, M.-C. (Ed.),Traces d’Utilisation sur les Outils Néolithiques du Proche-Orient. Maison deL’Orient, Lyon, pp. 189e198.

Helwig, K., 1998. The Characterisation of Iron Earth Pigments Using InfraredSpectroscopy, IRUG2 at V&A Postprints. Victoria Albert Museum, pp. 83e91.

Henry, D.O., Garrard, A.N., 1988. Tor Hamar: an Epipaleolithic rockshelter insouthern Jordan. Palestine Exploration Quarterly 120, 1e25.

Henry, D.O., Shen, C., 1995. The Madamaghan Sites. In: Henry, D.O. (Ed.), PrehistoricCultural Ecology and Evolution: Insights from Southern Jordan. Plenum Press,New York, pp. 295e318.

Horwitz, L.K., Garfinkel, Y., 1991. Animal remains from the Site of Gesher, CentralJordan Valley. Journal of the Israel Prehistoric Society 24, 64e76.

Ighilahriz, F., 1996. L’ocre dans l’industrie lithique de l’Ibéromaurusien d’AfalouBou-Rhummel (Algérie). L’Anthropologie 100, 77e87.

Inizan, M.L., 1976. Outils lithiques Caspiens ocrés. L’Anthropologie 80, 39e63.Keeley, L., 1982. Hafting and retooling: effects on the archaeological record.

American Antiquity 47, 798e809.Kingery, W.D., Vandiver, P.B., Prickett, M., 1988. The beginnings of pyrotechnology,

part II: production and use of lime and gypsum plaster in the Pre-PotteryNeolithic Near East. Journal of Field Archaeology 15, 219e244.

Krige, E.J., 1965. The Social System of the Zulu. Shuter & Shooter, Pietermaritzburg.Lombard, M., 2005. Evidence of hunting and hafting during the Middle Stone Age at

Sibidu Cave, KwaZulu-Natal, South Africa: a multianalytical approach. Journal ofHuman Evolution 48, 279e300.

Lombard, M., 2007. The gripping nature of ochre: The association of ochre withHowiesons Poort adhesives and Later Stone Age mastics from South Africa.Journal of Human Evolution 53, 406e419.

Lombard, M., 2008. Finding resolution for the Howiesons Poort through themicroscope: micro-residue analysis of segments from Sibudu Cave, SouthAfrica. Journal of Archaeological Science 35, 26e41.

Marks, A.,1983. The sites of Boker Tachtit and Boker: brief introduction. In:Marks, A.E.(Ed.), Prehistory and Palaeoenvironments in the Central Negev, Israel. The Avdat/Aqev Area, vol. III. Southern Methodist University Press, Dallas.

Mazza, P.P.A., Martini, F., Sala, B., Magi, M., Colombini, M.P., Giachi, G., Landucci, F.,Cristina, L., Modugno, F., Ribechini, E., 2006. A new Palaeolithic discovery: tar-hafted stone tools in a European Mid-Pleistocene bone-bearing bed. Journal ofArchaeological Science 33, 1310e1318.

Neuville, R., 1951. Le Paléolithique et le Mésolithique du Désert de Judée. Masson etCie, Paris.

Nissenbaum, A., 1997. 8,000 year old collagen from Nahal Hemar Cave. Archaeologyand Science Bulletin, Weizmann Institute of Science 5, 5e11.

Pomiès, M.P., Morin, G., Vignaud, C., 1998. XRD study of goethiteehematite trans-formation: application to identification of heated prehistoric pigments. Euro-pean Journal of Solid State and Inorganic Chemistry 35, 9e25.

Suárez, M.M., 1968. The Warao: Natives of the Orinoco Delta. Deparemento deAntropología, Instituto Venezolano de Investigaciones Científicas, Caracas.

Turville-Petre, F., 1932. Excavations in the Mugharet el-Kebara. Journal of the RoyalAnthropological Institute 62, 271e276.

Valla, F.R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquetin, F., Cabellos, T., Bar-Yosef Mayer, D., Le Dosseur, G., Regev, L., Chu, V., Weiner, S., Boaretto, E.,Samuelian, N., Valentin, B., Delerue, S., Poupeau, G., Bridault, A., Rabinovich, R.,Simmons, T., Zohar, I., Ashkenazi, S., Delgado Huertas, A., Spiro, B., Mienis, H.K.,Rosen, A.M., Porat, N., Belfer-Cohen, A., 2007. Les fouilles de Ain Mallaha(Eynan) de 2003 à 2005: Quatrième rapport préliminaire. Journal of the IsraelPrehistoric Society 37, 132e383.

Wadley, L., 2005. Putting ochre to the test: replication studies of adhesives that mayhave been used for hafting tools in the Middle Stone Age. Journal of HumanEvolution 49, 587e601.

Wadley, L., Williamson, B., Lombard, M., 2004. Ochre in hafting in Middle Stone Agesouthern Africa: a practical role. Antiquity 78, 661e675.

Wadley, L., Hodgskiss, T., Grant, M., 2009. Implications for complex cognition fromthe hafting of tools with compound adhesives in the Middle Stone Age, SouthAfrica. Proceedings of the National Academy of Sciences 106.

Weinstein-Evron, M., Ilani, S., 1994. Provenance of ochre in the Natufian layers of El-Wad Cave, Mount Carmel, Israel. Journal of Archaeological Science 21, 461e467.

Wreschner, E., 1983. Studies in Prehistoric Ochre Technology. The Hebrew Univer-sity, Jerusalem.

D. Shaham et al. / Journal of Archaeological Science 37 (2010) 2010e20162016