The lithic tool arsenal of a Mousterian hunter

lable at ScienceDirect

Quaternary International xxx (2013) 1e19

Contents lists avai

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

The lithic tool arsenal of a Mousterian hunter

Gonen Sharon a,*, Maya Oron b

aMultidisciplinary Studies, Tel Hai College, Upper Galilee 12208, Israelb Institute of Archaeology, Hebrew University of Jerusalem, Israel

a r t i c l e i n f o

Article history:Available online xxx

* Corresponding author. Tel Hai College, PrehistoGalilee 12208, Israel.

E-mail addresses: [email protected] (G. Sha(M. Oron).

1040-6182/$ e see front matter � 2013 Elsevier Ltd ahttp://dx.doi.org/10.1016/j.quaint.2013.10.024

Please cite this article in press as: Sharon, Gdx.doi.org/10.1016/j.quaint.2013.10.024

a b s t r a c t

The lithic assemblage excavated from the Mousterian site of Nahal Mahanayeem Outlet (NMO) enablesus to reconstruct a brief moment in the life of the Middle Palaeolithic hunter. The site, located on theeastern bank of the Jordan River at its outflow south from the Hula Valley, is a short-term, task specifichunting location at the shore of the Paleo-Hula Lake. Dated by OSL to ca. 65,000 years ago, the site hasyielded a small assemblage of flint artifacts alongside exceptionally well-preserved animal bones andbotanical remains. While only some 1000 artifacts have been counted, the lithic assemblage has thehighest percentage of tools ever recorded in a Levantine Mousterian site. The primary lithic groupsrepresented are pointed elements (over 10% of the entire assemblage) and cutting tools (over 5%). Othertool types typical of Mousterian sites, such as scrapers, are either absent or represented in very smallnumbers. The uniqueness of the assemblage is further highlighted by refitted sequences that, whencombined with technological observations, suggest a non-Levallois, “blade-core like” reduction sequence.The NMO assemblage represents the tool kit used for the hunting and butchering of large mammals by agroup of Levantine Mousterian hunters. It enables us to explore what tool types were selected forslaughtering and carcass processing, which tools were brought to the site and which were produced onsite, what tools were left behind, and much more. The site was inhabited for a very short period,providing an opportunity to study hunting practices and human life ways in a resolution rarely possiblefor Late Pleistocene sites.

� 2013 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction

Most of our knowledge of the life ways and behavior of pre-historic people during the Middle Palaeolithic comes from largecave sites with long sequences of sedimentary accumulation and awealth of finds, primarily of stone tools and bones. This is partic-ularly true for the Middle Palaeolithic era in the Levant. Here,almost all that we know originated from the excavation of longsequences of the famous cave sites (e.g. Tabun Cave e Garrod andBate, 1937; Amud Cave e Suzuki and Takai, 1970; Abu Sif andKebara Caves e Neuville, 1951; Kebara Cave e Bar-Yosef et al., 2007,1992; Hayonim Cave e Stiner, 2005; Meignen, 2011; Qafzeh Cave e

Hovers, 2009 and many others). The layers of the cave sites are, inmany cases, littered with hundreds of thousands of stone tools andanimal bones. Due to time averaging, the cave archaeological ho-rizons represent hard-to-estimate time spans, in which every cubic

ry Lab, East Campus, Upper

ron), [email protected]

nd INQUA. All rights reserved.

., Oron, M., The lithic tool ars

centimeter may represent a long occupation and an unknownnumber of different activities. While producing themost significantinformation on a large scale, the Levantine cave sites are usually notthe best localities for high resolution data from which everydayactivity patterns can be discerned. This type of data generallyoriginates from open-air, short occupation sites. Open-air sites inEurope are now yielding data regarding site function, lithic tech-nology and emerging models of mobility patterns for EuropeanNeanderthals (see recent examples in Moncel and Rivals, 2011;Delagnes and Rendu, 2011). Apparently, even open-air huntingand butchering sites (or layers within sites) are not easy to distin-guish within longer duration, multi-activity sites (but see papers inCarbonell, 2012). Accordingly, many of the models suggested todescribe site pattern and mobility activity, such as the foraging andlogistical models of Binford (1980), are of a theoretical nature basedupon ethnographic observation or non-empiric considerationrather than upon archaeological data (Binford, 1980; Kuhn, 1992,1995). Furthermore, their development to describe the behaviorof modern humans and their applicability to early human behavioris questionable. Nevertheless, they provide definitions and guide-lines for the current discussion. Specifically, we think that thedefinition provided by Binford (1980) for a task-specific location is

enal of a Mousterian hunter, Quaternary International (2013), http://

mailto:[email protected]

mailto:[email protected]

www.sciencedirect.com/science/journal/10406182

http://www.elsevier.com/locate/quaint

http://dx.doi.org/10.1016/j.quaint.2013.10.024



Fig. 1. Map of the Upper Jordan Valley and location of the NMO site.

G. Sharon, M. Oron / Quaternary International xxx (2013) 1e192

themost suitable termdescribing the nature of the NMO site.Whenstudying the assemblages from the few Levantine open-air sitesexcavated to date, a complex picture emerges (Hovers, 2009). Thesite of Quneitra, for example (Goren-Inbar, 1990; Oron and Goren-Inbar, in this issue), seems to represent a more long term occupa-tion judging from its large number of stone tools and bones, and itslithic typological composition may suggest different activities. It isclear that at Quneitra, stone tool knappingwas a significant activity.While it may be possible to assign all of these activities to theprocessing of animal carcasses, the presence of numerous scrapersin the assemblage may indicate that activities in addition to “slicingthe meat” took place at the site (see Moncel et al., 2009; Delagnesand Rendu, 2011).

The lithic assemblage from the Mousterian site of NahalMahanayeem Outlet (NMO) is different. It is small, comprising lessthan one thousand artifacts. It is unique in its composition withdominancy of pointed elements and cutting tools. In this paper, wepresent an analysis of the lithic assemblage unearthed during thefirst five excavation seasons at NMO. We use the unique nature ofthis assemblage to define and describe the “tool kit” of the group of

Please cite this article in press as: Sharon, G., Oron, M., The lithic tool arsdx.doi.org/10.1016/j.quaint.2013.10.024

Mousterian hunters who inhabited the site. The excavation’s highresolution made possible by the short-term and task-specific na-ture of the archaeological horizons at NMO, combined with a smallnumber of highly significant refitted knapping sequences, enablesus to discuss questions of raw material acquisition strategies, toolmobility and selection, discard patterns and tool use and lithictechnology, which rarely can be asked in larger, richer MiddlePalaeolithic sites.

2. The site of NMO

The Mousterian site of Nahal Mahanayeem Outlet at the bank ofthe Jordan River was discovered during a drainage operation in1999 (Sharon et al., 2002a, 2002b, 2010). The site is located on theeast bank of the Jordan River, opposite the outlet of the now arti-ficial channel of the Mahanayeem Stream into the Jordan north ofthe Benot Ya’aqov Bridge (Fig. 1). The piles of sediments dug byheavy machinery from the channel of the river were found to belittered with flint tools and animal bones in excellent preservationcondition, indicating the presence of archaeological horizons


Fig. 2. Location of excavation areas, excavated squares and general schematic stratigraphic composition of the NMO geology and archaeology.

G. Sharon, M. Oron / Quaternary International xxx (2013) 1e19 3

worthy of excavation. A team from Tel Hai College and the HebrewUniversity of Jerusalem began excavating the site in 2007. Pre-liminary results of the first two excavation seasons have been dis-cussed elsewhere (Sharon et al., 2010). The site is OSL dated to c.65,000 years before present (Kalb et al., 2013). In this paper wefocus on the unique flint assemblage excavated at NMO with thegoal of describing its special nature and implications for the


understanding of Middle Paleolithic (MP) lithic technology andhuman behavior.

2.1. Stratigraphy

The site of NMO is located on the primary geological faultforming the northern section of the Dead Sea Rift, an area subject to



constant tectonic activity as well as intensive volcanism (Belitzky,2002; Schattner and Weinberger, 2008; Spiro et al., 2011). Theresulting geology is highly complex on a regional scale (see detaileddiscussion in Kalbe et al., 2013). However, the local archaeologicalstratigraphy recorded at the site, based on the excavation areas aswell as on geological trenches and river bank sections, wasobserved to be quite straightforward (Fig. 2). At the base of thearchaeological stratigraphic sequence lies a layer of basalt bouldersand cobbles of unknown thickness recorded as Layer 5. Themorphology of this conglomerate is of a hill sloping gently towardsthe north and tilted to the east (Fig. 3). To the west the basalt formsa steep, low cliff to which the artifacts and bones of the archaeo-logical horizon are directly attached, indicating the formation ofthis morphology prior to human presence at the site (Fig. 3). Thismorphology resembles a bar or cutting of mini-streams into thebasaltic boulder fans of the nearby Golan Heights stream outletsinto the Rift Valley. Despite our current knowledge, we are still farfrom fully understanding the processes that formed this basaltaccumulation (for further discussion and details see Kalbe et al.,2013). The basalt of Layer 5 is archaeologically sterile; it is the lowerbed-rock of the archaeological sequence of NMO. The reconstruc-tion of the excavated surface suggests that the NMO inhabitantsfound this “basalt hill” as part of the landscape at the time ofoccupation.

The primary archaeological bearing layer, Layer 4 (Fig. 2), is alayer of dark fine silt, lying directly on top of the basalt floor. Thearchaeological remains, primarily bones and flint tools, are scat-tered in distinct horizons within the layer’s dark mud. A minimumof two and probably three such horizons were identified in theLayer 4 sequence (Fig. 4). The lowest horizon, that lying in directcontact with the basalt floor, is rich in artifacts and bones (horizon1). At its thickest point, horizon 1 reaches a depth of 40 cm. Asecond archaeological horizon (horizon 2) was observed in themudabove horizon 1. Horizon 2 includes a few basalt cobbles, bones of acow in poor preservation state (when compared to the preservationof the bones in the lower horizon) and a few flint artifacts. Duringthe 2012 season, a possible distinct, third horizon was excavated,

Fig. 3. “The basalt hill” of Area D, La


lying some 50 cm above horizon 2 and yielding stone tools, poorlypreserved bones and a few wood fragments. It can be suggested,therefore, that these upper horizons represent additional, shortoccupation events that may be related to the presence of the basalthill of Layer 5. Based upon the findings, we suggest that the sitewasa lake shore-accumulation environment at which short-term hu-man occupation activities were performed. The remains of suchactivities were quickly covered by rising water levels that accu-mulated fine, silty mud that sealed the lowest archaeological ho-rizon. Later visits to the same spot (possibly land marked by thebasalt hill) resulted in the formation of new, higher level horizonsthat were sealed in turn by the continually rising water. The lower,northern section of the basalt hill of Layer 5 was probably alreadycovered with mud prior to the occupation of horizon 1, which ex-plains the thickness of this horizon. Furthermore, it is suggested bythe orientation of the bones and flint tools, some of which werefound standing vertically in the mud, that they were discardeddirectly into the soft mud at the foot of the basalt hill of Layer 5 (Fig4). Additional excavation is needed before certainty can be reachedwith regard to this stratigraphic interpretation, yet the accumu-lating evidence seems to support it further (Fig. 4).

The vast majority of bones found in the Layer 4 horizons arethose of giant cows. Other animal bones found in much smallernumbers in these horizons include wild boar, deer, gazelle andothers (Sharon et al., 2010). Analysis of the fauna is ongoing. Whilethe overwhelming presence of cow bones has led to the interpre-tation of NMO as a site for the processing and probably hunting ofcows, the presence of additional species suggests a somewhat morecomplex behavior than the slaughtering of a single animal species.

On top of archaeological Layer 4 is a sequence of silty mud andclays that accumulated during the Late Pleistocene. This sequencecomprises later archaeological entities including the UpperPalaeolithic assemblage excavated from Area A at the site (Sharonet al., 2010) and even Byzantine coins and lead weights for fishingnets found higher up in the sequence. As stated above, thegeological and geomorphological history of this sequence of mudand clays is complicated and will not be discussed in detail here

yer 5, the bed-rock of the site.


Fig. 4. A west-east section of Area D along the 159 grid line (Fig. 2) showing horizontal distribution of flint artifacts and animal bones (context integrity 1 and 2 only). Note thethickness of the archaeological horizon 1 of lower Layer 4, the approx. standing of the Basalt Layer 5 and the presence of an additional archaeological horizon “floating” in the mudabove Layer 4. Note that the “hill” topography of Layer 5 and that the finds located at the lower part of the section are not embedded within this layer but laying on its slopingmargins.

Table 1Tools by excavation areas.

Excavation area Flakes and flake tools Cores and non-flakes

N % N %

Surface 144 10.5 9 19.1Area A (UP?) 201 14.7Area B 4 0.3 1 2.1Area C 142 10.3 3 6.4Area D 730 53.2 24 51.1Area E 63 4.6 5 10.6Area F 11 0.8 1 2.1Sections and Trenches 77 5.6 4 8.5


(see Kalbe et al., 2013). The archaeological horizons within Layer 4were sealed by this mud and were probably never subject to sig-nificant post-depositional movement.

The NMO stratigraphy and nature of archaeological accumula-tion indicate a short-term occupation of the Layer 4 horizons. Thedata and observations supporting this conclusion (some of whichwill be presented in detail below) include the small number of flintartifacts in the assemblage, the techno-typological nature of theassemblage, the fair number of refitted artifacts and their locationand, finally, the excellent preservation of flints, bones and botanicalremains indicating rapid coverage of the horizon with waterloggedmud. Analysis of the stone tools of the site, the best indicator forhuman behavior, provides a unique opportunity to study a site thatwas probably occupied for a very short time and, judging from theassemblage, was task specific.

Total 1372 100 47 100.0

2.2. Context of artifacts

Excavation of the NMOMousterian layers was focused on Area Cand D (Fig. 2). Other excavation areas, river bank sections andgeological test trenches also yielded artifacts, but some of theseoriginated in different layers or are of stratigraphically uncertaincontext. The NMO database, after 5 excavation seasons from 2007to 2011, currently holds 1342 flakes and 47 cores and core tools(Table 1). Out of these, the group of 201 artifacts excavated fromArea A (Fig. 2 and Table 1) belong to the Upper Palaeolithic horizonof the site (Sharon et al., 2010). An additional group of artifactscomprising flint implements excavated from Area E are clearlyMousterian. However, the artifacts originated in an occurrence thatappears to be very different from the primary Area D Layer 4assemblage andwill be described elsewhere. Hence, out of the 1342artifacts recorded, only 670 flakes and 24 cores and non-flake ar-tifacts comprise the integral part of the NMO Layer 4 assemblage.This group includes the following: a) the flint artifacts excavated insitu from Areas C and D that together form the primary excavationarea at the site; b) the small number of flakes from Area B (Fig. 2and see Sharon et al., 2010), as the flakes are now known to haveoriginated from the margin of the same archaeological Layer 4; c)artifacts that originated from the 3 sections of the Jordan Riverexposed and studied during the 2007 season with the goal of


clarifying the site’s stratigraphy now known as Area D (Sharonet al., 2010); and d) artifacts from Area F and Trench IV, now thatLayer 4 has been verified as their stratigraphic context (Fig. 2).

While we are satisfied with the context of the artifacts excavatedin situ from the excavation areas, the flint artifact assemblage fromNMO comprises additional artifacts whose stratigraphic context isless clear. The vicinity of the NMO site has been heavily disturbed byover one hundred and fifty years of drainage operations. It is alsoclear from the excavation that after the accumulation of the site,channels of ancient streams cut into its layers causing disturbance tothe upper parts of the sequence (Fig. 2). While these ancient dis-turbances are, in most cases, easily defined and recognizable due tothe very different nature of their sediments, special care has beentaken to ensure that artifacts taken from disturbed contexts are notmixed with those whose archaeological integrity is verifiable.Therefore, when analyzing the flint artifacts from NMO we appliedfour “context integrity” categories to their classification: 1. in situ e

finds excavated from the archaeological layer; 2. integrity in questione finds that probably originated in a verifiable archaeologicalcontext, however, their integrity cannot be determined with cer-tainty; 3. mixed context e finds that can be attributed to an archae-ological layer, e.g. artifacts dugout by tractor shovel from a knownlocality; and 4. surface e finds from disturbed contexts. The classifi-cation of the artifacts into these categories is noted when relevant.When the lithics are sorted according to their origin in excavationareas (Table 1), the small assemblage size is even more notable, as


Fig. 5. Distribution map of flint flakes and fauna density of Area D, all flakes and bones through the 2011 season. Context integrity 1 and 2 only.


many of the artifacts originated from either non in situ or non-Mousterian contexts. For the purpose of the current paper we didnot distinguish between the different archaeological horizons ofLayer 4. Such separationwill be established in the futurewhenmorespatial data is available. As can be seen from Fig. 4, the greatmajorityof the stone tools can be attributed to the lower horizon (horizon 1).Someof the tools describedherewillmost likely need to be shifted toother horizons, reducing even further the number of artifactsbelonging to horizon 1 and emphasizing the observations presentedhere regarding the unique nature and size of the assemblage.

3. Excavation and lithic analysis methodology

The excavation is being recorded using a 1 m2 grid system.During excavation, all artifacts (as well as bones, wood and otherfinds) are recorded using a Leica total station device, resulting inhighly accurate location data. This data is then processed using theArcMap GIS program of ESRI (see: http://www.esri.com/software).All sediments are wet-sieved (2 mm mesh) and sorted in the lab,and the sorted artifacts are analyzed together with the excavatedassemblage. The analysis of the artifacts is based upon the meth-odology developed for the site of Quneitra (Goren-Inbar, 1990) andsubsequently modified for other sites (Hovers, 1998, 2009). Themethod is a combination of morphological, technological andtypological observations recorded for each of the artifacts. Specificmodifications to the method were made to best describe the NMOassemblages and are presented below in the data section.

4. NMO lithic assemblage

According to the “context integrity” categories presented above,the observations presented in this section are derived only from ar-tifacts excavated in themain excavationAreasD andC (Table 1). Theseartifacts are from a verifiable archaeological context, classified aseither category 1 (in situ) or category 2 (context in question). Thisreduces the number of artifacts under study here to only 694 (670flakes and 24 cores and non-flakes). Nevertheless, this small numberof artifacts represents the entire excavated assemblage of Layer 4 in


the site. The advantage of the small number is thatwe canobtain a fullandcomprehensivepictureof the tool kitusedby the site’s inhabitantsin a resolution very hard to achieve for other, flint-rich assemblages.

4.1. Artifact density

The primary excavation areas at NMO are Areas C and D (Fig. 2).The total excavated surface within these areas is 54 m2. Thus, theaverage density of the lithic artifacts for Areas C and D is approxi-mately 11 artifacts per m2 (with a range of 1e66 lithic items perm2). This density is very low in comparison to other MP sites in theLevant. This is particularly true for the large cave sites in whichdensity can reach tens of thousands of artifacts per square meter(e.g. Bar-Yosef et al., 2007; Yeshurun et al., 2007; Hovers, 2009). Insome caves, such as Hayonim Cave, density is much lower(Meignen, 2011). Nevertheless, the NMO artifact density also is lowwhen compared to other MP open-air sites of the Levant. Forexample, in Quneitra the average density is 83 (Area A) and 100(Area B) lithic finds per m2 (Oron and Goren-Inbar, in this volume).

Calculating artifact density for the entire excavation area is some-what misleading. The artifacts are not evenly scattered on the exca-vated surface. The spatial distribution of the artifacts in Area D ispresented in Fig. 5. Clear concentrations of artifacts can be observed inArea D. Detailed discussion of the spatial distribution of the lithic ar-tifacts and their correlationwith other types of evidence such as faunaland botanic finds will be presented elsewhere. For the purpose of thecurrent discussion, it is suggested that these high density areas are theresult of human agency and may reflect activity zones (see forexample, Alperson-Afil and Hovers, 2005; Alperson-Afil et al., 2009).The concentration of artifacts indicates that some sections of the sitearemoredenselyoccupiedwith stone tools. This evidence is presentedhere as support for the anthropological nature of the accumulation.

4.2. Preservation state

The flint artifacts from Area D at NMO are exceptionally well-preserved. As many as 83.4% of the flakes were recorded as “fresh”,meaning here “as fresh as flint tools can get”. In no other site havewe


http://www.esri.com/software


seenflakes in abetter preservation state (Table2). Anadditional 10.2%of the flakes were recorded as “slightly abraded”, together yielding93.6% of the flakes in good preservation condition. This observationrules out any possibility that the flint artifacts were subject to anymajor, or even minor transportation prior to or post-deposition.

Table 2Preservation state Layer 4 flakes.

Preservation N %

Fresh 571 85.2Slightly abraded 62 9.3Abraded 23 3.4Rolled 14 2.1Total 670 100.0

4.3. Artifact typology e the Bordesian typological system

Table 3 presents the frequencies of typological types within theassemblage according to the “traditional” Bordesian typologicalsystem (Bordes, 1961), which has been modified for Levantine MPassemblages (Goren-Inbar, 1990; Hovers, 1998, 2009). Thefollowing primary observations regarding the NMO assemblage canbe drawn from Table 3:

Table 3Assemblage typology.

Category Typology N %

Levallois Typical Levallois flake 2 0.3Atypical Levallois flake 4 0.6

Points Levallois point 16 2.4Retouched Levallois point 1 0.1Pseudo-Levallois point 3 0.4Mousterian point 1 0.1Tanged point 1 0.1Total points 28 4.2

Scrapers Limace 1 0.1Single straight side scraper 1 0.1Single Convex side-scraper 7 1.0Double convex side-scraper 1 0.1Double concave side-scraper 1 0.1Convergent straight scraper 1 0.1Side-scraper on ventral face 2 0.3Alternate retouch side-scraper 1 0.1H D scraper 1 0.1Total scrapers 16 2.4Typical end-scraper 1 0.1Atypical end scraper 5 0.7Typical burin 2 0.3Atypical burin 1 0.1Typical borer 2 0.3Atypical borer 1 0.1Notch 23 3.4Denticulate 9 1.3Total other tools 44 6.6Atypical backed knife 6 0.9Naturally backed knife 33 4.9Knives total 39 5.8Retouched flake 103 15.4Total tools 230 34.3

Waste Angular fragment 3 0.4Core waste 6 0.9Flake 394 58.8Blade 37 5.5Total waste 440 65.7Total flakes 670 100.0

Cores and core tools Levallois core for flakes 2 8.3Nucleus divers 9 37.5Nucleus inform 4 16.7Percuteur 1 4.2Angular fragment 2 8.3Core on flake 6 25.0

Total 24 100.0Total assemblage 694


a. Exceptionally high percentage of tools: The percentage oftools at NMO is 34.3% of the assemblage. This percentage seemsto be the highest for any excavated Mousterian assemblagerecorded to date in the Levant. Typically, the percentage of toolswithin a Levantine Mousterian cave site assemblage will rangearound 5%. At open-air sites the percentage of tools is usuallyhigher than for caves. However, even for an open-air assemblagethe NMO tool frequency is high (Goren-Inbar, 1990: Table 32;Hovers, 2009: Table 8.4).

b. Typological composition: The assemblage is also unique for thetool types that are under-represented. Scrapers of all types formonly 2.1% of the flake and flake tool assemblage. Other tool typessuch as end scrapers, burins and borers (Upper Paleolithic types)form less than 1% when grouped together. Even notches anddenticulates, normally a substantial percentage of Mousterian as-semblages, appear in very low frequencies at NMO (Hovers, 2009).

c. Levallois knapping method: Technologically, the use of theLevallois knapping method is low, although it is clearly present.The Levallois index is only 0.04 (Hovers, 2009: Table 8.3). Two ofthe cores were defined as Levallois cores for flakes, yet they aresmall and atypical. Only 32 flint artifacts, measuring 4.3% of theflakes in the assemblage, show definitive evidence of havingbeen knapped using Levallois technology. An additional 76 ar-tifacts (10% of the flakes) were classified as “Maybe Levallois”,meaning that they show some attributes of Levallois technologybut cannot be classified as such with full certainty.

d. Low frequency of knapping waste: The waste includes cores,core waste, and core trimming elements. Some of the flakesexcavated at NMO are on-site knapping waste, as evident fromthe results of the refitting. However, stone tool knapping couldnot have been a significant activity at NMO, and the knappingmethod used on-site shows minimal core preparation, resultingin a minimal amount of core waste.

The picture emerging from the typological composition of thesite’s assemblage is of a small assemblage with a very high per-centage of tools and low percentage of knapping waste. The tools inthe assemblage are somewhat different from other sites, as scrapersand notches are represented in small numbers at NMO, whileknives and points are more frequent. Similar to many sites, thelargest category of tools is “retouched flakes”, comprising 15.4% ofthe entire flake assemblage.

The NMO assemblage has additional technological and typo-logical aspects for which the Bordesian typology, developed pri-marily to describe European sites, seems inadequate. The detailedinformation that follows is intended to provide a clearer descrip-tion of the assemblage and its unique character.

5. Technological and morphological characteristics of theNMO assemblage

During the analysis of the NMO flint assemblage, two elementsemerged as characteristic of the assemblage: pointed elements andlong cutting-edge elements. These two elements do not fit easilyinto the Bordesian typological system which, while open to inter-pretation, is very rigid in its definitions. For example, the label“point” refers to Levallois points and Mousterian points and in-cludes only a few additional types such as pseudo-Levallois andtanged points. The result of this rigidity is that many artifactspointed in morphology and very similar in shape to “real” pointsare classified as either waste flakes or blades in this system or, atbest, as “retouched flakes” since they fail to possess all necessaryattributes (Table 3 and Fig. 6).

During our analysis of the NMO lithic assemblage, the distinc-tion of “pointed” and “cutting” elements arose from the repeated


Fig. 6. Pointed elements from NMO e Non-Levallois.


morphology of artifacts observed during analysis. We describebelow the characteristic morphology and technology that definethese groups in an attempt to present a repeatable procedure thatcan be applied to other MP assemblages.

5.1. Pointed elements

The pointed elements group is comprised of flakes and bladeshaving pointed morphology ending in a pointed tip. The elementswere grouped based on their morphology and on the technologicalattributes defined below. A total of 75 pointed elements wereidentified in the NMO Area D assemblage. They comprise 11.2% ofthe entire assemblage (Table 4). Two technological approaches canbe observed in the pointed elements, Levallois and non-Levallois, asdescribed below:

Table 4Pointed elements technological group.

N % % Within pointed elements

Levallois point 21 3.1 28.0Non-Levallois point 39 5.8 52.0Possibly point 15 2.2 20.0Total points 75 11.2 100.0Non pointed elements 595 88.8Total 670 100.0

Points created using Levallois technological concept (term afterBoëda, 1995): These include primary Levallois points as well as afew additional point types such as Mousterian points and pseudo-Levallois points (in very small numbers; n¼ 3). The NMO points aresimilar to Levallois points recorded in many other Levantine MP


sites. It should be noted, however, that the typical “Tabun B” broad-base short points (Copeland, 1975; Hovers, 2009) are practicallyabsent from NMO. The NMO Levallois pointed elements showcarefully prepared, faceted striking platforms and comprise 2.4%(n ¼ 19) of the flake assemblage (Fig. 7: a, e and f). Levallois pro-duced elements make up, therefore, up to 25% of the pointed ele-ments. For comparison, see Table 8.3 in Hovers (2009) showing thevery low percentage of Levallois points in other MP sites of theLevant.

Points created using non-Levallois knapping method: In additionto the Levallois points, a distinctive group of pointed elementsexists at NMO that could not be assigned to the Levallois knappingmethod (Boëda, 1995). These artifacts show the following typicalmorphological features (Figs. 6e8):

1. They are pointed, showing a basic, triangular blank morphology,the distal edge being the narrowest part.

2. Many of them are of elongated, with blade proportions (Fig. 9).No difference was observed in metric proportions betweenLevallois and non-Levallois pointed elements.

3. The striking platforms are, in most cases, either plain or dihedraland are typically thick (best examples are Fig. 6h and i). It is clearthat for non-Levallois pointed elements plain and dihedralstriking platforms are dominant. This observation is somewhatcircular as one of the definitions of a Levallois element is afacetted striking platform; yet, it is evident that most of thepointed elements in the NMO assemblage originated from non-Levallois cores.

4. On many of the striking platforms, evidence of platform prep-aration appears that is typically attributed to the production ofblades (Fig. 10). These “blade core” preparation morphologies


Fig. 7. Levallois (a, b, e, f), Tanged (c) and Mousterian (d) points from NMO.


include the presence of many small scars on the proximal, distaledge of the tool (angle de chasse e Inizan et al., 1999), some-times evidence of “step scar” morphology resulting from thecreation of many small scars in the same place; hinged endingscars, also typical of blade production; and, lastly, clear evidenceof abrasion of the striking platform prior to removal of the flake(Fig. 10).

5. The scar pattern on the dorsal face of these pointed elements isnot convergent as on the typical Levallois points. It is a unidi-rectional/unipolar, parallel pattern. In other words, the blankswere produced from a single platform core, designed to produceprimary, elongated elements based on a parallel scar pattern.

The morphology of these pointed elements is dictated by theirplace within the removal sequence of the core (Fig. 11). Some ofthem follow a single, guiding arris (“ridge”, terminology fromInizan et al., 1999) separating two scars. Others follow two sucharrises. The single arris points are elongated, narrow and, in manycases, very fine and aesthetic (Fig. 6). The double arris points arewider and, in some cases, end in a tip that is not strictly pointed orwith the pointed tip diverging from the middle axis of the point(Fig. 8).

In addition, a very small group of pointed blades can be distin-guished from among the NMO pointed elements. This is a group ofvery well-made, long and pointed blades, most resulting fromsingle arris removals. They are very similar in proportion to each


other and, in most cases, were found unbroken and in mint con-dition (see Fig. 6h and i for best examples). This group is too small innumber from which to draw further conclusions; nonetheless, itmay be suggested that the particular shape and proportion of thepoints led to their selection by the NMO inhabitants to bring to thesite (or to leave behind).

Another important feature of the NMO pointed elements is thatthey are infrequently retouched. At Quneitra, for example, manypointed elements were retouched and hence classified as conver-gent or dejete scrapers (Goren-Inbar, 1990). Such tools are practi-cally absent from NMO.

Because they are classified as either flake or blade waste, wehave no way of knowing the percentage of pointed, un-retouchedelements in other MP assemblages. One of the goals of this paperis to draw the attention of researchers analyzing Mousterian as-semblages to this group of artifacts, since it appears that they wereselected specifically by the NMO inhabitants.

Additional support for the classification of these artifacts aspoints comes from their high frequency of broken tips. Table 5shows the frequency of breakage patterns for the NMO pointedelements. While 28 (32.2%) of the pointed elements were foundcomplete, as many as 20 (23%) are distally broken and an additional26 (29.9%) show distal damage, indicating minimal tip shatter. Thepresence of impact fracture was identified for the NMO points(Yaroshevich et al., 2010; Yaroshevich, A., personal communication)but quantified conclusions await further analysis.


Table 5Breakage pattern of NMO pointed elements.

Breakagelocation

Levallois point Non-Levallois point Maybe point Total

N % N % N % N %

Complete 8 10.7 14 18.7 22 29.3Distal 3 4.0 7 9.3 6 8.0 16 21.3Proximal 1 1.3 2 2.7 2 2.7 5 6.7Lateral & distal 1 1.3 1 1.3Proximal &

distal1 1.3 3 4.0 1 1.3 5 6.7

Fragment 1 1.3 1 1.3Distal end use

damage1 1.3 1 1.3 1 1.3 3 4.0

Distal minimaldamage

8 9.3 12 14.7 5 5.3 22 29.3

Total 21 28.0 40 53.3 14 18.7 75 100

Table 6NMO cutting tools techno-typological groups.

N % % Within cutting tools

Cutting tool 25 3.7 39.1ABK 6 0.9 9.4NBK 33 4.9 51.6Total cutting tools 64 9.6 100.0Non Cutting tools 606 90.4Total 670 100.0


The pointed elements at NMO are numerous. They form morethan 10% of the entire assemblage. The number of points in mostpublished assemblages rarely reaches over 5% of the tools. At NMOthey are probably the most significant aspect of the assemblage. Inaddition, while the NMO pointed elements were produced bydifferent reduction sequences, the pointed elements resulting fromthese sequences are similar in morphology and size. The NMOhunters selected their target tools from a variety of possible blanksand brought to the site (or, more precisely e left behind) a quitehomogenous group of tools (Fig. 9). In addition, technological ob-servations and refitting results enable us to claim with a high de-gree of certainty that the better-made pointed elements foundwere imported into the site and not produced on-site.

5.2. Cutting implements

In addition to the pointed elements, a second group of imple-ments was observed at NMO, defined as flakes and blades with along, uninterrupted, frequently straight, sharp and un-retouchededge (Figs. 12 and 13). We use the term “cutting tool” with nofunctional bearing. The definition is strictly based on themorphology of the tools and the presence of a long, sharp edge.Some of these implements are classified as knives according totypological criteria (naturally backed knives and atypical backed

Fig. 8. “Two arris” po


knives) while others are grouped here due to the presence of anedge following the criteria described above (Table 6). It is suggestedthat these implements could all have been used as knives for slicingmeat during carcass processing at the site. The group of cutting andslicing elements at NMO includes:

Naturally backed knivese (NBK; Bordes, 1961; n¼ 34e53% of thecutting tools). These are frequently elongated, narrow blades andflakes that have one lateral edge covered with cortex while theother side is a sharp, untouched edge. Few of the naturally backedknives at NMO would be considered primary blades in the termi-nology of Shimelmitz et al. (2011) due to the angle between theback and the ventral face that is smaller than the maximum limit of60� (Fig. 12: a, b, c, f, g).

Atypical backed knives e (ABK; Bordes, 1961; n ¼ 6e11% of thecutting tools). These are blades and flakes that have an abrupt backat one lateral edge that is not shaped by retouch. This attribute isthe reason for their classification as atypical (Fig. 13). No typicalknives (back shaped by retouch) were found at NMO.

Blades and flakes with a long, un-retouched sharp edgee (n¼ 23e36% of the cutting tools). This is a group of artifacts classified ascutting tools due to the presence of a sharp, un-retouched edge,longer than 5 cm. Since this is not a typological category, some ofthese cutting tools actually belong to other tool types according totheir retouch (e.g. notches, end-scrapers or retouched flakes), butthese retouches do not appear on the cutting edge.

The mean length of the cutting edge of all three categories is89 mm, with the third group (blades and flakes) showing thehighest values with a mean edge length of 107 mm. This is due, inpart, to the fact that some have no back and the cutting edge is

ints from NMO.


Fig. 9. Length vs. width of NMO pointed elements.

Fig. 11. NMO technology of pointed element production.


spread along both lateral edges. Away to illustrate the length of thisedge is by testing the ratio of edge to artifact circumference. Whencalculating this ratio for the entire group of cutting tools the meanis 46.5% length to circumference. When calculating separately foreach sub-group, the ratio for the naturally backed knives andatypical backed knives is 40% and 41% respectively, while the valuefor the blades and flakes is 57%. The reason for this difference is thepresence or absence of a backed lateral edge.

Altogether, 64 flakes and blades are classified as cutting toolsdue to the presence of a long, sharp edge (>5 cm). These make up9.6% of the entire assemblage (Table 6). As in the case of the pointedelements, these artifacts are not retouched. In addition, of partic-ular interest is the scarcity of scrapers, end scrapers and other toolsin the assemblage. At all MP sites one finds a group of artifacts suchas flakes and blades possessing a long, sharp edge suitable for meatprocessing. The NMO assemblage is unique due to the rarity ofother types such as scrapers and the very high percentage of im-plements suitable for cutting. This observation is emphasizedfurther by the fact that the great majority of the artifacts were notproduced on-site. The artifacts that were produced on site (evidentfrom refitting) are also elongated flakes or blades with cutting toolproportions. It is evident that the inhabitants of NMO chose thistype of morphology for the tools they imported into the site sincemany other shapes and types are not represented. Analysis of theNMO lithic assemblage demonstrates that the group of artifacts

Fig. 10. Striking platforms of NMO pointed elements sh


with a long cutting edge is very prominent. Use wear studies willundoubtedly contribute to the refinement of this observation (e.g.Moncel et al., 2009). Nonetheless, it is clear that the NMO huntersselected elongated, sharp flakes and blades as part of their toolarsenal for the task executed at the site.

6. Flint refitting and technological implications

The refitting effort at NMO is still in its early stages but alreadyhas proven to be very rewarding. To date, as many as 30 artifactswere joined together into eight reduction sequences. This is areasonably high rate for a Levantine MP site at which most of theartifacts were imported as tools, and where some of the flint wassubject to the heavy patination typical of waterlogged environ-ments of the Jordan River bank sites (Sharon and Goren-Inbar,1999). The 8 sequences contain 2e10 refitted flakes. Some of thesequences consist of flakes joined together while others consist ofcores to which flakes could be reattached. One of the latter is madeup of a large, irregular core to which 3 flakes could be refitted. The

owing “blade core preparation” including abrasion.


Fig. 12. Drawing of cutting elements from NMO.


longest and most significant sequence comprises a combination of10 flakes and blades representing a number of stages of a singlereduction sequence, from which a great deal of technological in-formation can be extracted (Fig.14). The refitting effort for the NMOlithic assemblage is ongoing and will be described in detail in thefuture. Here we focus on the technological information retrievedfrom the long sequence (Sequence 1) and its application to un-derstanding the lithic technology and economy of the Mousterianhunters at the site.

The long sequence (Sequence 1, Fig. 14) consists of 10 flakes andblades with the core still missing. The blank for the core was acortical nodule of medium quality flint (a knapper would describe


the flint as dry, grainy and not very homogenous) that is quitecommon to the site. The flint has different colors and patterns thathelped in the refitting effort (Fig. 14).

Observed technology: the presence of cortical, primary flakesindicates that the first step applied by the knapper was decortica-tion of the nodule. This was done on site, suggesting that the nodulewas brought unprepared to the site and all reduction stages wereconducted in situ. This observation is supported by the presence ofan additional nodule with only two removals that was probablytested for its rawmaterial quality. This additional nodule was foundin square J159 in close proximity to a wild boar mandible. It seemsthat together with some knapped implements, the NMO knappers


Fig. 13. Cutting elements from NMO.


brought a few untouched or tested-only nodules into the site toserve as raw material for on-site tool production.

The knappingmethod applied as reconstructed from the refittedsequence indicates the sequential removal of elongated flakes, insome cases blades, from a single platform. The refitted flake strikingplatforms are, in most cases, thick and plain. No evidence exists forstriking platform preparation by means of faceting that is typical ofthe Levallois method. Some evidence for abrasion of the platformprior to flake removal can be observed in the form of small scars,but the scars are not as prominent as those on some well-madepointed elements at the site (Fig. 10). The flakes and blades wereremoved sequentially following along the outer rim of a singleplatform in a practice typical of blade core method (�Skrdla, 2003;Shimelmitz et al., 2011). The products of the reduction sequenceare as follows:

� Two small pointed elements (Fig. 14) roughly shaped. The ele-ments resulted from an early stage of the core reduction andmay represent a natural morphology derived from the reductiontechnology in which flakes follow the scars of the previousflakes. However, they fit well within the group of pointed ele-ments presented above.

� Three cutting elements with a long cutting edge, one of which isan atypical backed knife.

� Two massive, elongated and thick denticulates.


� Two cortical fragments.� Small waste flake.

Other refitted sequences at NMO also include massive, elon-gated flakes shaped in the same non-meticulous manner. The ef-ficiency of the sequence is demonstrated by the production of manyusable artifacts. Almost all products of the core could be used, whileapplying minimal time and effort in core preparation and building.Sequence 1 is clearly not a Levallois sequence. It can be described asa blade core method, but it differs from the Amudian system by thethickness of the platforms and the resulting products (Shimelmitzet al., 2011). It also differs quite clearly from later, much moredelicate and systematic Upper Palaeolithic Early Ahmarianmethods (Davidzon and Goring-Morris, 2003). It would be inap-propriate to speculate or generalize about the stone technology ofthe NMO inhabitants on the basis of a single core. Such speculationwould be even more inappropriate given that the core is, in ourinterpretation, the result of ad hoc, on-site, rapid knapping whosegoal was to produce cutting tools for meat processing. As such, thereduction sequence of this core is different from the full sequenceof careful manufacturing evident from other tools at the site.Rather, it is an example of a practice applied by the knappers on-site, most likely in response to the need for processing a greatquantity of meat. For this objective, medium-sized nodules werebrought into the site and worked in situ when other, much better


Fig. 14. Refitted “sequence 1”.


prepared tools were either lacking or inadequate for the task athand.

The location of the refitted artifacts in the site is illustrated inFig. 15. Detailed discussion of the implications of this distribution isbeyond the scope of the current paper, however, it can be seen thatartifacts were refitted from a large area of the excavated squares.They were excavated at relatively large distances, a fact that can beexplained by the nature of their use at the site. Tools were knappedin a specific place and used (and discarded) at a distance of a fewmeters from their detachment locality. The presence of refittedartifacts throughout the area of archaeological horizon 1 of Layer 4indicates that this layer was accumulated during a single event.

7. Blade preparation

A special technological feature of the NMO pointed elementsand blades is the preparation of the striking platform by means ofabrasions on the angle de chasse, the ridge between the butt of theflake and the upper proximal surface (terminology after Inizanet al., 1999). These abrasions, attributed to the removal of over-hangs from cores (in particular to facilitate blade removal), wereobserved on 81 striking platforms in the NMO assemblage. Of these,38 were observed on artifacts classified as pointed elements. Theabrasions generally appear in the form of numerous micro-scarsand sometimes in the form of short, hinge-ending, burin spell-like scars (Fig. 10).


This phenomenon, typical of blade production, has rarely beendescribed in the context of Mousterian lithic technology. Moreover,it was attributed as a technological marker dividing Upper Palae-olithic blade technology from the earlier Mousterian technology(Goring-Morris and Belfer-Cohen, 2003). Clearly, the NMO platformpreparation is unlike the UP indirect blade core preparationobserved with the same abrasions. At NMO, the striking platformsare thick and the flakes and blades were most probably removedusing hard hammer technique. The platform preparation is alsovery different from the typical Levallois preparation that the NMOknappers applied for the production of their fine tools. Similarabrasion marks were observed on artifacts from other MP sitesincluding Amud (A. Buler, personal communication) and even onthe finely-produced Abu-Sif points from the site of Abu-Sif (GSpersonal observation). It is suggested that more evidence of suchabrasions will be reported when awareness of this phenomenonincreases.

8. Raw material economy of flint

Good quality, sufficiently-sized flint is unavailable in the basalticterrain forming the immediate vicinity of NMO. The streamsrunning into the Hula Valley from the north and from the westoccasionally carry flint nodules of suitable size and quality that mayhave collected in the river beds a few kilometers from the NMO site.However, the muddy lakeside environment accumulating on the


Fig. 15. Distribution map of refitted elements from NMO. Lines represent different sequences.


basalt boulders of Layer 5 offered no immediately available rawmaterial to the NMO knappers. It is clear that some of the tools,predominantly points, were brought as finished tools. Due to theunavailability of flint, it is also suggested that the NMO tool makersbrought to the site flint nodules as a source of raw material for on-site knapping. The small amount of flint waste at the site indicatesthat only a few such nodules were brought to the site. Lastly, asdescribed above, the refitting reveals much about the strategy andtechnology applied to the imported raw material.

Basalt as raw material e Basalt is very common in the vicinity ofthe site in the shape of cobbles and pebbles as well as large boul-ders. Layer 5, the “floor” of the site, is a layer of basalt cobbles andboulders. The quality of basalt as rawmaterial for tool production isreasonable, and the same rawmaterial was used extensively by theAcheulian knappers at Gesher Benot Ya’aqov (Sharon, 2008). Basalt


is a significant raw material at the Mousterian site of Quneitra,located in similar basaltic terrain of the Golan Heights, where up to10% of the assemblage is shaped on this raw material (Goren-Inbar,1990). Given these facts, it is interesting to note that with theexception of a few small, sporadic flakes and possible hammer-stones, no basalt artifacts were identified within the NMO assem-blage. This observation supports the suggestion that the tools at thesite represent a task-specific tool kit, of which basalt tools were nota part at NMO. The presence of basalt tools at the site of Quneitra(Oron and Goren-Inbar, in this volume) indicates complex activitypatterns for MP hominid hunting sites. Many of the basalt artifactsat Quneitra were classified as massive scrapers (Goren-Inbar, 1990).At the site of Fa’ara II in the Northern Negev, massive scrapers werealso identified, but there they were produced out of limestone(Gilead, 1980, 1988). This makes the absence of basalt artifacts and



massive scrapers from the NMO assemblage even more significantand suggests a very task-specific nature for the assemblage.

9. Discussion

The lithic assemblage excavated at NMO is unique amongMousterian assemblages in the Levant. It is a small assemblage, yetit is very characteristic in its typological and technological attri-butes. The archaeological horizons of Layer 4 unearthed at NMOaccumulated during a very short time prehistorically, possiblymeasured in days, as a result of the butchering and most likelyhunting of large game on the banks of the Paleo-Hula Lake. Fromobservations including the spatial distribution of the artifacts, theabsence of weathering and the refitting data, the in situ nature ofthe assemblage is evident (Figs. 5 and 15). Both archaeological andsedimentological data indicate that the finds of this horizon did notmove from their location of discard (Kalbe et al., 2013). Some of theflint tools and bones were found in an upraised position within thefine sediment, suggesting amuddy environment of accumulation inwhich some of the artifacts and bones were immediately coveredby, or even sank into, soft mud.

The archaeological horizon’s short duration is reconstructedfrom three primary factors: the sedimentology of the layers; thesmall number of lithic artifacts; and the mint condition of the lithicartifacts, bones and, in particular, the botanical remains. The NMOarchaeological horizons observed within Layer 4 were accumulatedin very fine-grain silt of large thickness (probably > 2 m). The siltwas accumulating in a lake shore environment and no evidence forlamination is evident within this sediment, suggesting rapidaccumulation in the rising water level of the lake. A possible causefor the accumulation may be a single flood bringing a large amountof sorted, fine material to the southern edge of the lake. Such finemud accumulation can be seen today at the southern margin ofLake Kinneret, at the southern shore, far from the entrance of thelake’s primary tributaries.

The NMO lithic assemblage is small, containing only a fewhundred artifacts. This number is much smaller than that of anyother excavated site in the Levant, including open-air sites. Forexample, at the site of Quneitra, an excavated area of ca. 125 m2

yielded a total of almost 13,000 artifacts, while the much smallerexcavation at Fa’ara II unearthed over 3700 artifacts (Goren-Inbar,1990: Table 32). The small assemblage size is strong evidenceindicating that the site was occupied for only a very short timeperiod. Sites at which there were repeated visits over a long timeperiod have a much larger number of artifacts. The NMO horizonsexposed in Layer 4 fit within Binford’s (1980) definition of a “tasklocation”: “A location is a place where extractive tasks are exclu-sively carried out. Since foragers generally do not stockpile foods orother raw materials, such locations are generally “low bulk” pro-curement sites. That is to say, only limited quantities are procuredthere during any one episode, and therefore the site is occupiedfor only a very short period of time. In addition, since bulkprocurement is rare, the use, exhaustion, and abandonment of toolsis at a very low rate” (Binford, 1980, 9). Binford further quotesHayden: “As a rule, they are spatially segregated from base campsand are occupied for short durations (usually only a matter of hoursat the most) by task-specific groups. the lithic tools employed aregenerally very distinctive and the assemblages highly differentiatedin terms of proportional frequencies compared to base campassemblages. (Binford, 1980, 9)”

Short occupation duration is further evident from the mintcondition of the lithic artifacts (see Table 2), the well-preservedbones and the presence of a large amount of wood and otherbotanical remains. Such preservation could not have been observedfor artifacts exposed to the destructive effect of the harsh Hula


Valley sun for a long time (years or even months). These are theprimary observations that enable us to conclude that the bones,botanical remains and lithic assemblage are evidence of a short-term occupation. The lithic assemblage includes only those toolsthat were used (or, more accurately, left behind) by the site’s in-habitants for the specific task they executed at the site, namely theslaughtering, processing and most likely hunting of big game, pri-marily very large cows.

9.1. Lithic technology and flint economy

The assemblage is comprised primarily of pointed elements andcutting tools. Analysis of the artifacts reveals the use of two pri-mary, identifiable lithic core technologies. The pointed elements, inparticular, were knapped by the Levallois core method and a non-Levallois blade core method. The artifacts produced using theLevallois core method were not knapped on-site, as evident fromthe nearly complete absence of cores and primary waste products.The majority of the items classified as Levallois products at NMOare points; Levallois flakes are rare and very few of them wereretouched. It is important to note that the Levallois points at NMOcannot be classified as typical “Tabun B” short, broad-base pointssince many are elongated in proportion.

The second core method applied by the NMO knappers for theproduction of their tools, both pointed elements and cutting toolsalike, is a non-Levallois e blade volumetric concept. This method isevident from the presence of thick, plain platforms that show clearmarks of abrasion and typical blade-core preparation as well as thescar pattern observed on many of the NMO artifacts (Figs. 6 and 8).

Finally, a third, ad hoc but very efficient core method was usedfor on-site knapping. This core method is evident from the refittedsequences at the site as well as from the presence of chunks of rawmaterial, some tested for raw material quality. This method wasused primarily for the production of elongated, thick cutting toolsproduced from cortical nodules brought into the site as raw ma-terial. These nodules were brought for immediate use as dictated bythe needs of the site’s inhabitants. Hence, two distinct groups canbe defined within the assemblage:

� Carefully produced artifacts that were imported into the site intheir finished form, including a high percentage of pointed el-ements and cutting tools. These artifacts were made on highquality flint and reveal a high level of preparation and dexterity;

� On-site knapped artifacts primarily including rough, thick arti-facts produced using a very efficient but not highly systematicmethod.

Such complex behavior involving sophisticated raw materialutilization strategies was documented for MP sites in Europe (e.g.Delagnes and Rendu, 2011 and references therein), but rarely hassuch resolution been achieved for the few open-air sites of theLevantine MP.

9.2. Typology and assemblage composition

The NMO assemblage is characterized by a high percentage oftools to waste. Goren-Inbar summarized the data for excavatedLevantine Mousterian sites (Goren-Inbar, 1990: Table 32). Thehighest percentage of tools, up to 26% of the assemblage, wasrecorded at the site of Quneitra (excluding the tiny, probably se-lective collection from Adlun and Naame in Lebanon). AfterQuneitra, the only assemblage comprising over 20% tools is that ofVadi Hasa 621. In all other sites the values are much smaller,reaching only a few percent in some sites. More recently, Hovers(2009: Table 8.4) presented data from additional sites. Most of



the sites recorded have a tool percentage below 10% and, in manycases, below 5%. The highest values are from a few layers in QafzehCave, with Layer XIV being the highest at 16.77%. It should be notedthat the layers of the large cave sites range dramatically in theirpercentage of tools (between 1.8 and 16.8 in Qafzeh). At NMO, theBordesian typology has revealed a tool percentage of well over 30%of the assemblage. This is high by any standard and the highestvalue we are aware of in any excavated site in the Levant.

There are also very few cores, primary cortical flakes and coretrimming elements (CTE) in the assemblage. This fact, togetherwith the high tool to waste percentage, demonstrates that themajority of the tools at the site were imported into the site in theirdiscard shape and that knapping was not one of the primary ac-tivities at the site.

A very illustrative aspect of the NMO assemblage is the scarcity ofmany tool types typical of Levantine Mousterian assemblages. Moststriking, scrapers of all types form a small percentage of the assem-blage: a total of 16 scrapers out of 230 tools or out of 670 flakes. Thesame holds true for the other tool types (Table 3). This is one of thestrongest pieces of evidence demonstrating that the NMO assem-blage resulted from highly selective tool behavior. Only very specifictool types were brought into or produced on site, namely pointedelements and cutting tools. All other tool types, for which the in-habitantsmost likely hadno use for the tasks executed at the site, areeither absent or present in very low frequencies. It should beemphasized that the different morphologies, tool types and tech-nologies observed at NMO are found in other, contemporary Mous-terian assemblages. For example,many of themwere observed in thelithic assemblage excavated at the Amud Cave (A. Buler, personalcommunication). It is their frequency, however, and the absence ofdifferent tool types that makes the NMO assemblage so unique.

Moreover, many of the NMO artifacts have elongated pro-portions, further indication of preference for a specific tool type.This is particularly truewhen examining the better produced, finelyshaped tools at the site. The term elongated refers here to an arti-fact whose length is more than 1.5 times its width (for blades thelength to width ratio is 2:1). The assemblage is not laminar as awhole (Meignen, 2011) but when the pointed elements and cuttingtools are measured as a group they tend to be elongated (Fig. 9).This is in contrast to MP assemblages classified as Tabun B Type, inwhich the pointed elements are broad in morphology (Hovers,2009).

Two techno-morphological groups of artifacts emerge from theanalysis of the lithic assemblage: pointed elements and cuttingtools. Together, the artifacts classified into these two groupscomprise 21% of the entire assemblage. The evidence from NMOallows us to suggest that, when preparing for a hunting expedition,Levantine Mousterian hunters equipped themselves with thefollowing:

1. Pointed elements, possibly serving as projectile tips (Shea, 1988,2006; Villa et al., 2009; Yaroshevich et al., 2010), as evident fromthe damage fractures at the tip of many of the NMO points. Themorphology of many of the points, in particular the finely madeones, suggests that the NMO hunters had a preference forelongated, narrow, pointed elements.

2. Cutting tools, flakes and blades, some of which are very finelymade in the shape of well-balanced blades, possibly used formeat cutting and carcass processing.

3. A few nodules of un-flaked or tested-only flint serving as rawmaterial for ad-hoc, yet very efficient, on-site knapping of roughcutting tools and additional tools as needed.

This is the picture emerging from the lithic artifacts the NMOinhabitants left behind them. Parts of it remain unclear.While some


of the artifacts seem to have been discarded due to breakage, othersseem to be in perfectly usable condition. Discard behavior surelyinvolves patterns beyond the scope of our modern understanding.We are also still lacking any evidence for hafting. Wood has beenpreserved at NMO; hence, the absence of shafts unearthed to dateseems to indicate that the lithic artifacts were discarded withouttheir shafts. It was noted (Keeley, 1982; Villa and Lenoir, 2009) thatthe preparation of a wooden spear shaft is much more time andenergy consuming than the shaping of a stone tip. This may explainwhy wooden shafts have not yet been found at NMO.

A variety of site types are described for Neanderthal subsistencestrategies and mobility in the well-documented open-air sites ofWestern France (see Delagnes and Rendu, 2011 and referencestherein). These include sites that were revisited numerous times forhunting thousands of bison, and sites where animals were fullyprocessed and consumed (evident from the presence of a highpercentage of humanly modified animal bones as well as manyburnt bones and flints). Additional site types have been suggestedthat distinguish between “kill sites”, at which the carcasses werebriefly prepared, and “butchering sites” to which the carcasseswere then transported for further slaughtering in order to preparethe richest body parts (in terms of food utility index) for trans-portation to the base camp. Our observations to date suggest thatthe NMO site is a kill/butchering location (Binford, 1980terminology).

The NMO lithic assemblage is typologically and technologicallydifferent from other known MP lithic assemblages. Much of thediscussion regarding cultural variability in the Levantine Mouste-rian is based upon typological and technological characteristics ofthese other assemblages (see Hovers, 2009 for overview). A keyquestion arises as to whether the unique tool collection from NMOrepresents a new, distinguishable cultural phase within theLevantine Mousterian. If this is true, then there will be resultingimplications regarding the chronology of the assemblage. It wasnoted long ago that the entire lithic variability of the LevantineMousterian does not follow any of the European lithic classifica-tions and this trend is further supported by analysis of the lithicassemblages of the Levantine open-air sites (see Hovers, 2009 foroverview and references). Chronologically, the elongated nature ofthe NMO artifacts suggests that the assemblage should be attrib-uted to the Early “Tabun D” stage of the Levantine Mousterian. Thiscontradicts the dates obtained for the site placing it within the finalstages of the Levantine MP. The NMO lithic assemblage does notherald a new cultural phase in the Levantine MP, nor should it beattributed to a defined stage. The typological composition of theassemblage is a direct outcome of the primary activity carried out atthe site: processing the meat of large game.

10. Conclusion

The lithic assemblage excavated from Layer 4 at the Mousteriansite of Nahal Mahanayeem Outlet, OSL dated to ca. 60 ka, indicatesthat the site was a short-term, task specific location (Binford, 1980)at which large game was slaughtered and processed. The shortduration of occupation is demonstrated by the small size of thesample and its typological composition, the presence of refittedlithic sequences, the excellent preservation of archaeological re-mains that excludes the probability of long exposure to atmo-spheric conditions prior to burial and the sedimentological natureof the archaeological horizons. Typologically, the assemblage isunique for the tools it comprises and, no less important, for the tooltypes absent from the assemblage. The NMO assemblage has thehighest percentage of tools-to-waste recorded to date in the Levant.These unique conditions, very rare in the Levantine Mousterian,enabled us to explore and describe the tool kit used by the



Mousterian hunters for the specific activity of meat processing (andprobably hunting) in high resolution. Such resolution is unattain-able in the long sequences of the large Levantine cave sites.

The tool kit of the Levantine Mousterian hunter included pri-marily pointed elements and cutting tools. These tool types formover 20% of the entire assemblage of the site. Additional tool typesthat are normally a very pronounced part of the Levantine Mous-terian, in particular scrapers of all types, are practically absent fromthe NMO assemblage. Technological study and the presence ofrefitted sequences have enabled us to describe the tool economy andmobility of the NMO inhabitants. At least three different lithicmethods were applied for the production of the points and blades atthe site, including Levallois method, Non-Levallois blade coremethod and an efficient, ad hoc knapping sequence designed for fastproduction of rough, elongated cutting tools. The NMO tool usersimported into the site (and left behind) finely made, elongatedpoints and blades that show careful selectivity. They also broughtinto the site un-retouched flint nodules that served as raw materialfor the on-site, ad hoc knapping. Both are evidence of advancedplanning of raw material economy by the NMO inhabitants. Theypredicted the need for a large quantity of cutting edge during theslaughtering process of large game, mainly giant cows (up to12,000 kg each), whichwas the primary activity at horizon 1 of NMOLayer 4. The site’s environment is a lake shore and its geology sug-gests that it was a high, dry spot surrounded by water on at least 3sides (Kalb et al., 2013), a good location for meat processing. Slicingand removing large quantities of flesh, evident from the presence ofcut marks on the bones at the site, requires a lot of cutting edge. Thisexplains the frequency of cutting tools at NMO.

Hunting activity is suggested from the presence of many pointedelements, many of which show damage fractures of their tip.Moreover, it is difficult to imagine that somany animals were founddead and scavenged during the short time of the site’s occupation.At the current stage of research we are unable to say that bothhunting and meat processing took place at the site. It could be thatthe meat was brought from a nearby hunting location, although thesize and quantity of bones at the site suggests that a great weightwould then have been carried to the site. Alternatively, the siterepresents the location where both hunting and meat processingtook place. It is clear, however, that other activities such as hideprocessing or flint knapping were not primary activities carried outat the site. This may suggest that these other activities were notpart of the immediate hunting practice and took place in differentlocalities, possibly at the larger, long-term duration cave sites.

The site of NMO is interpreted as a short-term hunting campwhere the meat of large game was processed. The unique nature ofthe lithic assemblage excavated from NMO does not represent alithic traditional or cultural deviation within the Levantine Mous-terian. Rather, it is a functional selection of specific morpho-typological artifact groups out of the inventory available to thesite’s inhabitants. It was dictated by the function of and the activitythat took place at the site, providing a glimpse into what theLevantine Mousterian hunters selected as their tool kit when goingout on a hunting expedition. The assemblage reflects sophisticatedbehavior, preplanning, and knowledge of the environmentmasteredby Upper Jordan Valley MP hunters during the Late Pleistocene.

Acknowledgments

Excavation and research at NMO is supported primarily by theIsrael Science Foundation (Grant No. 645/09, titled: Behavior,Subsistence Strategies and Paleo-environmental Background ofMiddle Paleolithic Hominids in the Northern Dead Sea Rift). Inaddition, the project is supported by the following grantingagencies: The Leakey Foundation, The Wenner-Gren Foundation,


National Geographic and the Irene Levi-Sala CARE ArchaeologicalFoundation. Tel Hai College supports the NMO excavation projectwith internal grants and by providing research facilities, including aprehistory laboratory. Artifacts drawing by A. Sumner & M. Oron.Many people contributed to this paper, discussing ideas and mak-ing important suggestions. In particular, we would like to thank A.Buler for sharing his ideas and knowledge. Most of all wewould liketo thank the NMO international excavation team for many hoursspent knee-deep in the mud with the Jordan River sun over theirheads. It was fun after all.

References

Alperson-Afil, N., Hovers, E., 2005. Differential use of space in the Neandertal site ofAmud Cave, Israel. Eurasian Prehistory 3, 3e22.

Alperson-Afil, N., Sharon, G., Kislev, M., Melamed, Y., Zohar, I., Ashkenazi, S.,Rabinovich, R., Biton, R., Werker, E., Hartman, G., Feibel, C., Goren-Inbar, N.,2009. Spatial organization of hominin activities at Gesher Benot Ya’aqov, Israel.Science 326, 1677e1680.

Bar-Yosef, O., Vandermeersch, B., Arensburg, B., Belfer-Cohen, A., Goldberg, P.,Laville, H., Meignen, L., Rak, Y., Speth, J.D., Tchernov, E., Tillier, A.-M., Weiner, S.,1992. The excavations in Kebara Cave. Current Anthropology 33, 497e550.

Bar-Yosef, O., Meignen, L., Albert, R.M., 2007. Kebara Cave, Mt. Carmel, Israel: theMiddle and Upper Paleolithic Archaeology: Part 1. Peabody Museum ofArchaeology and Ethnology, Harvard University.

Belitzky, S., 2002. The structure and morphotectonics of the Gesher Benot Ya’aqovArea, Northern Dead Sea Rift, Israel. Quaternary Research 58, 372e380.

Binford, L.R., 1980. Willow smoke and dogs’ tails: hunter-gatherer settlement sys-tems and archaeological site formation. American Antiquity 45, 4e20.

Boëda, E., 1995. Levallois: a volumetric construction, methods, a technique. In:Dibbel, H.L., Bar-Yosef, O. (Eds.), The Definition and Interpretation of LevalloisTechnology. Prehistory Press, Madison Wisconsin, p. 52.

Bordes, F., 1961. Typologie du Paleolithique Ancien et Moyen. CNRS, Paris.Carbonell, E., 2012. High Resolution Archaeology and Neanderthal Behavior: Time

and Space in Level J of Abric Romaní (Capellades, Spain). In: Delson, Eric.,Sargis, Eric J. (Eds.), Vertebrate Paleobiology and Paleoanthropology Series.Springer, New York.

Copeland, L., 1975. The Middle and Upper Paleolithic of Lebanon and Syria in thelight of Recent Research. In: Wendorf, F., Marks, A.E. (Eds.), Problems in Pre-history: North Africa and the Levant. SMU Press, Dallas, pp. 317e350.

Davidzon, A., Goring-Morris, N., 2003. Sealed in stone: the Upper Palaeolithic EarlyAhmarian Knapping Method in the Light of Refitting Studies at Nahal NizzanaXIII, Western Negev, Israel. Mitekufat Haeven e Journal of the Israel PrehistoricSociety 33, 75e205.

Delagnes, A., Rendu, W., 2011. Shifts in Neandertal mobility, technology and sub-sistence strategies in western France. Journal of Archaeological Science 38,1771e1783.

Garrod, D.A.E., Bate, D.M.A., 1937. The Stone Age of Mount Carmel. In: Excavations inthe Wady el-Mughara, vol. 1. Clarendon Press, Oxford, UK.

Gilead, I., 1980. A Middle Paleolithic open-air site near Tell Far’ah, Western Negev:preliminary report. Israel Exploration Journal 30, 52e62.

Gilead, I., 1988. Le site moustérien de Fara II (Neguev septentrional, Israel) et leremontage de son industrie. L’Anthropologie 92, 797e807.

Goren-Inbar, N., 1990. Quneitra: a Mousterian Site on the Golan Heights. The He-brew University of Jerusalem, Jerusalem.

Goring-Morris,A.N., Belfer-Cohen,A., 2003.More thanMeets theEye:StudiesonUpperPalaeolithic Diversity in the Near East. Oxbow; David Brown Book Co., Oxford.

Hovers, E., 1998. The lithic assemblages of Amud Cave: implications for under-standing the end of the Mousterian in the Levant. In: Akazawa, T., Aoki, K., Bar-Yosef, O. (Eds.), Neandertals and Modern Humans in Western Asia. Plenum,New York, pp. 143e163.

Hovers, E., 2009. TheLithicAssemblagesofQafzehCave.OxfordUniversity Press,Oxford.Inizan, M.-L., Reduron-Ballinger, M., Roche, H., Tixier, J., 1999. Technology and Ter-

minology of Knapping Stone (J. Féblot-Augustins, Trans.). In: Préhistoire de laPierre Taillée, vol. 5. CREP, Nanterre.

Keeley, L.H., 1982. Hafting and retooling: effects on the archaeological record.American Antiquity 47, 798e809.

Kuhn, S.L., 1992. On planning and curated technologies in the Middle Palaeolithic.Journal of Anthropological Research 21, 43e66.

Kuhn, S.L., 1995. Mousterian Lithic Technology e an Ecological Perspective.Princeton University Press, Princeton N.J.

Meignen, L., 2011. Contribution of Hayonim Cave assemblage to the understandingof the so-called “Early Levantine Mousterian”. In: Tensorer, J.-M.L., Jagher, R.,Otte, M. (Eds.), The Lower and Middle Palaeolithic in the Middle East andNeighboring Regions, vol. 126 ERAUL, Liége, pp. 85e100.

Moncel, M.-H., Rivals, F., 2011. On the question of short-term Neanderthal Site oc-cupations e Payre, France (MIS 8e7), and Taubach/Weimar, Germany (MIS 5).Journal of Anthropological Research 67, 47e75.

Moncel, M.-H., Chacón, M.G., Coudenneau, A., Fernandes, P., 2009. Points andconvergent tools in the European Early Middle Palaeolithic site of Payre (SE,France). Journal of Archaeological Science 36, 1892e1909.


http://refhub.elsevier.com/S1040-6182(13)00796-9/sref1




















































































Neuville, R., 1951. Le Paléolithique et le Mesolithique du desert du Judee. Masson etCie, Paris.

Schattner, U., Weinberger, R., 2008. A mid-Pleistocene deformation transition in theHula basin, northern Israel: implications for the tectonic evolution of the DeadSea Fault. Geochemistry Geophysics Geosystems 9 (7), Q07009.

Sharon, G., 2008. The impact of raw material on Acheulian large flake production.Journal of Archaeological Science 35, 1329e1344.

Sharon, G., Goren-Inbar, N., 1999. Soft Percussor use at the Gesher Benot Ya’aqovAcheulian Site? Mitekufat Haeven e Journal of the Israel Prehistoric Society 28,55e79.

Sharon, G., Feibel, C.S., Belitzky, S., Marder, O., Khalaily, H., Rabinovich, R., 2002a.1999 Jordan River Drainage Project Damages Gesher Benot Ya’aqov: a pre-liminary study of the archaeological and geological implications. In: Gal, Z.(Ed.), Eretz Zafon e Studies in Galilean Archaeology. Israel Antiquities Authority,Jerusalem, pp. 1e19.

Sharon, G., Marder, O., Boaretto, E., 2002b. A note on 14C dates from the Epi-Paleolithic Site at Gesher Benot Ya’aqov. Mitekufat Haeven e Journal of theIsrael Prehistoric Society 32, 5e15.

Sharon, G., Grosman, L., Fluck, H., Melamed, Y., Rak, Y., Rabinovich, R., Oron, M.,2010. The first two excavation seasons at NMO: a Mousterian site at the bank ofthe Jordan River. Eurasian Prehistory 7, 135e157.

Shea, J.J., 1988. Spear points from the Middle Paleolithic of the Levant. Journal ofField Archaeology 15, 441e450.

Shea, J.J., 2006. The origins of lithic projectile point technology: evidence fromAfrica, the Levant, and Europe. Journal of Archaeological Science 33, 823e846.


Shimelmitz, R., Barkai, R., Gopher, A., 2011. Systematic blade production at late LowerPaleolithic (400e200kyr)QesemCave, Israel. JournalofHumanEvolution61,458e479.

Spiro, B., Ashkenazi, S., Starinsky, A., Katz, A., 2011. Strontium isotopes in Mela-nopsis sp. as indicators of variation in hydrology and climate in the UpperJordan Valley during the EarlyeMiddle Pleistocene, and wider implications.Journal of Human Evolution 60, 407e416.

Stiner, M., 2005. The Faunas of Hayonim Cave, Israel. Peabody Museum ofArchaeology and Ethnology. Harvard University, Cambridge, Massachusetts.

Suzuki,H., Takai, F.,1970.TheAmudManandHisCaveSite. TheUniversityof Tokyo, Tokyo.Villa, P., Lenoir, M., 2009. Hunting and hunting weapons of the lower and middle

Paleolithic of Europe. In: Hublin, J.-J., Richards, M.P. (Eds.), The Evolution ofHominin Diets: Integrating Approaches to the Study of Palaeolithic Subsistence.Springer, New York, pp. 59e85.

Villa, P., Boscato, P., Ranaldo, F., Ronchitelli, A., 2009. Stone tools for the hunt: pointswith impact scars from a Middle Paleolithic site in southern Italy. Journal ofArchaeological Science 36, 850e859.

Yaroshevich, A., Kaufman, D., Nuzhnyy, D., Bar-Yosef, O., Weinstein-Evron, M., 2010.Design and performance of microlith implemented projectiles during theMiddle and the Late Epipaleolithic of the Levant: experimental and archaeo-logical evidence. Journal of Archaeological Science 37, 368e388.

Yeshurun, R., Bar-Oz, G., Weinstein-Evron, M., 2007. Modern Hunting Behavior inthe Early Middle Paleolithic: Faunal Remains from Misliya Cave, Mount Carmel,Israel. Elsevier, pp. 656e677.

�Skrdla, P., 2003. Comparison of Boker Tachtit and Stránská skála MP/UP transitionalindustries. Journal of the Israel Prehistoric Society 33, 37e73.








































































The lithic tool arsenal of a Mousterian hunter

Documents

Transcript of The lithic tool arsenal of a Mousterian hunter