lable at ScienceDirect

Journal of Human Evolution 70 (2014) 16e35

Contents lists avai

Journal of Human Evolution

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

Intensification and sedentism in the terminal Pleistocene Natufiansequence of el-Wad Terrace (Israel)

Reuven Yeshurun a,b,*, Guy Bar-Oz a, Mina Weinstein-Evron a

a Zinman Institute of Archaeology, University of Haifa, Mount Carmel, Haifa 3498838, Israelb Program in Human Ecology and Archaeobiology, National Museum of Natural History, Smithsonian Institution, PO Box 37012, MRC 112, Washington,DC 20013-7012, United States

a r t i c l e i n f o

Article history:Received 10 August 2013Accepted 21 February 2014Available online 21 March 2014

Keywords:EpipaleolithicLevantZooarchaeologyBroad-spectrum revolutionMobilityContextual taphonomy

* Corresponding author.E-mail addresses: ryeshuru@research.haifa.ac.il (R.

haifa.ac.il (G. Bar-Oz), evron@research.haifa.ac.il (M. W

http://dx.doi.org/10.1016/j.jhevol.2014.02.0110047-2484/� 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Measuring subsistence intensification in the archaeofaunal record has provided strong evidence forsocioeconomic shifts related to sedentarization in the terminal Pleistocene Mediterranean Basin, but theprecise timing and scale of the intensification trend and its place in the evolution of settled societiesremain contentious. New archaeofaunal data from the key Natufian sequence of el-Wad Terrace (MountCarmel, Israel, ca. 15.0e11.7 ka [thousands of years ago]) is used here to clarify and contextualize pale-oeconomy and mobility trends in the latest Pleistocene Levant, representing the culmination of Epi-paleolithic subsistence strategies. Taphonomic variables serve as supplementary indicators of habitationfunction and occupation intensity along the sequence. At el-Wad, a very broad range of animals, mostlysmall to medium in size, were captured and consumed. Consumption leftovers were discarded inintensively occupied domestic spaces and suffered moderate attrition. The Early (ca. 15.0e13.7/13.0 ka)and Late (ca. 13.7/13.0e11.7 ka) Natufian phases display some differences in prey exploitation andtaphonomic markers of occupation intensity, corresponding with other archaeological signals. Wefurther set the intra-Natufian taxonomic and demographic trends in perspective by considering theearlier Epipaleolithic sequence of the same region, the Israeli coastal plain. Consequently, we show thatthe Early Natufian record constituted an important dietary shift related to greater occupation intensityand sedentarization, rather than a gradual development, and that the Late Natufian record appears to bemaintaining, if not amplifying, many of these novel signals. These conclusions are important for un-derstanding the mode and tempo of the transition to settled life in human evolution.

� 2014 Elsevier Ltd. All rights reserved.

Introduction

The process of settling down by hunter-gatherer groups in theterminal Pleistocene (ca. 20,000e11,700 years cal. BP [calibratedbefore present]) was an important milestone in human evolution,entailing a series of changes in mobility, economy and society.Sedentary or semi-sedentary groups were emerging in the Medi-terranean Basin in the millennia following the Last GlacialMaximum (LGM), displaying novel adaptations and the roots ofsocial complexity. Specifically in the Levant region, the culturalperiod bridging the LGM and the end of the Pleistocene, the Epi-paleolithic, has been a major subject of investigation concerningthe formation of complex foraging societies, which eventually laid

Yeshurun), guybar@research.einstein-Evron).

the foundations for the subsequent Neolithic Period (e.g., Kaufman,1992; Bar-Yosef and Meadow, 1995; Henry, 1995; Stiner and Kuhn,2006; Watkins, 2010; Belfer-Cohen and Goring-Morris, 2011;Maher et al., 2012a). Some of the earliest and most conspicuousmanifestations of the pre-agricultural shift to sedentary livingappear ca. 15,000e11,700 years cal. BP, in the Late EpipaleolithicNatufian culture of the Levant. These include stone structures andterraces, large cemeteries, diverse groundstone assemblages andhewn bedrock features, plus numerous personal adornments andart items, as well as remains of commensal animals. These traits aremuch better represented, quantitatively and qualitatively, relativeto the earlier Epipaleolithic record, indicating increased perma-nence of occupation and probably increasingly complex societiesand greater human impact on their surroundings (e.g., Garrod,1957; Henry, 1991; Tchernov, 1993a,b; Valla, 1995; Bar-Yosef,1998; Belfer-Cohen and Bar-Yosef, 2000; Munro, 2004, 2009;Byrd, 2005; Goring-Morris and Belfer-Cohen, 2008).

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 17

Changes in subsistence go hand in hand with sedentarization inthe Mediterranean Basin. While most pre-LGM groups habituallyhunted ungulates and paid less attention to small game species,thereby continuing earlier Paleolithic traditions of big-gamehunting as the main source of animal food, it is generallyaccepted that Epipaleolithic, and particularly Natufian groups,intensified their resource base, habitually exploiting juvenile, smallor non-terrestrial animals as important dietary components (Davis,1991, 2005; Tchernov,1993a,b,c; Stiner et al., 1999, 2000; Bar-El andTchernov, 2000; Stiner, 2001; Hockett and Haws, 2002; Bar-Oz,2004; Munro, 2004; Atici, 2009; Munro and Atici, 2009; Stutzet al., 2009; Bar-Yosef Mayer and Zohar, 2010; Stiner and Munro,2011; Starkovich, 2012; Zeder, 2012). The addition of a suite ofanimal taxa, many of which are small-bodied or non-terrestrial, tothe regular human diet in the millennia just preceding the onset offood production is commonly referred to as the Broad SpectrumRevolution (BSR) (Flannery, 1969).

Subsistence intensification and the BSR are important in thewider context of human evolution because of their bearing on site-occupation intensity in the Epipaleolithic. A strong link exists be-tween increasing sedentism and intensifying subsistence. It is oftenconceptualized using a behavioral ecology approach, where gametaxa may be ranked according to their dietary gains versus searchand handling costs. Foragers are likely to habitually procure lower-ranked game only when higher-ranked resources become lessavailable (Winterhalder and Smith, 2000; Stiner and Munro, 2002;Munro, 2004, 2009). Economic intensification means that lower-ranked prey is regularly included in the diet, presumably becauseencounter rateswith higher-rank prey decrease or because demandfor prey increases. During the Epipaleolithic, the ability to reside ina site for longer periods would have meant relying on such inten-sification, i.e., extracting more nutrients from the environment(Munro, 2009), all the more so if the number of people inhabiting asettlement was higher than before, or if they maintained smallerterritories (Rosenberg, 1998). In the Levantine context, low-rankedprey include small and fast-escaping mammals and birds, whichprovide small quantities of edible material for a high capture cost,as well as juvenile ungulates, which provide less meat and fat thanadults (Stiner, 2001; Munro, 2004). Hence, economic intensificationcan be related to increases in human population in a given territory(Stiner et al., 1999, 2000), but the precise timing and scale of thepost-LGM intensification process and its place in the evolution ofsedentary societies remain contentious (see below).

This study investigates sedentism and economic intensificationin the key Natufian base-camp of el-Wad Terrace (Mount Carmel,Israel). This is the classic Natufian sequence where the long-heldview of this culture, as a complex and sedentary society at thethreshold of farming, was first conceived (Garrod, 1932, 1957). Weaim to clarify, refine and contextualize trends in Natufian economyand sedentism by employing detailed zooarchaeological data fromour new excavations at the site. Taxonomic abundances, gazelleculling patterns and taphonomic indicators are used to evaluatehabitation type, the magnitude of intensification and site occupa-tion intensity. Our results are set in context by comparing theNatufian animal economy with the earlier, well-studied Epi-paleolithic sequence of the same region (Fig. 1A). Ultimately, weaim to shed light on Epipaleolithic socioeconomic developments bypinpointing precisely when and how sedentism-related intensifi-cation occurred within the long process of settling down in theterminal Pleistocene of Southwest Asia.

Intensification and sedentarization processes in the Epipaleolithic

The Natufian record of the late Epipaleolithic Levant figuresprominently in all discussions of pre-agricultural intensification

and sedentarization. Scholarly opinions differ regarding the Natu-fian phenomenon. Since the early days of research, the large EarlyNatufian (EN) hamlets displaying architecture, cemeteries, art andabundant groundstone items caused the Natufian to be viewed as amajor break from preceding Paleolithic cultures (Garrod, 1957;Valla, 1995; Bar-Yosef, 1998). Based on archaeological criteria suchas changes in architecture, mortuary practices and art, an impor-tant intra-Natufian difference was noted by some, in that the ENphase (ca. 15.0e13.7/13.0 ka [thousands of years ago]) wasconsidered the classic sedentary phase and the Late Natufian (LN)phase (ca. 13.7/13.0e11.7 ka) was interpreted as having a retreat togreater mobility (Garrod, 1957; Belfer-Cohen and Bar-Yosef, 2000;Bar-Yosef and Belfer-Cohen, 2002). Growing archaeological evi-dence in the last three decades has placed the Natufian culture incontext, demonstrating its Epipaleolithic roots (Kaufman, 1989,1992; Maher et al., 2012a,b). Recent data from Israel and Jordanindicate prolonged site habitation and the presence of defined hutsthat were rich in symbolic meaning, echoing the hallmark Natufianfeatures, as early as 23e20 ka, and therefore the EN was viewed asgradually evolving from the preceding Epipaleolithic cultures(Nadel et al., 2004; Maher et al., 2012b).

Interpretations of Natufian zooarchaeological data are crucialfor understanding the nature of Epipaleolithic sedentarization.However, the timing, mode and tempo of the terminal Pleistoceneshift to intensified economy and BSR are often contested or looselydefined. Regarding the well-studied Levantine record, severalscholars maintained that the Natufian is exceptional relative to thepreceding cultures in the high proportion of small mammals andsometimes birds and fish in conjunction with diminishing pro-portions of medium and large ungulates (e.g., Davis et al., 1988;Davis, 1991; Pichon, 1991; Bar-Oz, 2004). In contrast, it has beensuggested that diversification of Levantine animal economiesoccurred millennia before the terminal Pleistocene (Edwards,1989), although this analysis was countered on the basis of inap-propriate statistical methods, poor sampling quality and lack oftaphonomic consideration (Neeley and Clark, 1993; Bar-Oz, 2004).

Several fine-grained archaeofaunal analyses pertaining to Epi-paleolithic intensification have recently been published. Combinedwith multiple lines of archaeological evidence for increased sed-entism, the Natufian subsistence trends were interpreted as evi-dence for intensification and diversification of animal exploitationdue to greater permanence of site occupation (Bar-Oz, 2004;Munro, 2004, 2009; Davis, 2005; Stutz et al., 2009). The Natu-fians not only exploited small game in unprecedented proportions,they began in particular to exploit less cost-effective but resilientanimals such as lagomorphs and birds, hinting at novel capturetechniques, elevated pursuit costs and rising occupation intensity(Stiner et al., 1999, 2000; Stiner and Munro, 2002; Munro, 2004,2009). The exploitation of the main ungulate prey, the mountaingazelle (Gazella gazella), was geared towards culling of lower-yieldyounger individuals, which was taken as another sign of foragingintensification (Munro, 2004, 2009; see also; Davis, 1983, 2005;Bar-Oz, 2004). In some of these studies, it was possible to investi-gate intra-Natufian trends, which showed that more fast smallgame was exploited in the EN phase, while in the LN phase slowsmall game was more dominant (Munro, 2004, 2009; Stutz et al.,2009). Zooarchaeological evidence suggesting Natufian economicintensification was recently viewed as the product of gradualtrends throughout the Epipaleolithic, culminating in the Natufian(Munro, 2009; Stutz et al., 2009). Proponents of the ‘gradualtransition’ downplayed the BSR hallmark of the Natufian, claimingsimilarity to earlier Epipaleolithic subsistence (Maher et al., 2012a),or altogether ignoring economic factors when comparing pre-Natufian and Natufian adaptations (Richter et al., 2011; Maheret al., 2012b).

Figure 1. Provenance of the faunal samples: (A) Location map showing el-Wad and other sites mentioned in the text; (B) Plan of el-Wad in the late Early Natufian (LEN) phase, themain ‘architectural’ phase of the site (after Weinstein-Evron et al., 2013); (C) plan of the NE Terrace excavation e the Late Natufian layer; (D) plan of the NE Terrace excavation e lateEarly Natufian layer. Note the ‘domestic’ (Structure II) and ‘non-domestic’ (Locus 25) areas.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3518

The key sequence of el-Wad Terrace (EWT) is ideally positionedin time and space to shed light on the mode and tempo of terminalPleistocene socioeconomic shifts pertaining to economic and sed-entism. These new data, put in regional context, allow us tocontribute to two of the pressing issues in the study of Levantinesedentarization:

1. Did Early Natufian subsistence and mobility patterns constitutea significant break from preceding Epipaleolithic cultures?

2. Was the Late Natufian fundamentally different from the EarlyNatufian in mobility and subsistence?

Framework of analysis

Here we provide a detailed taphonomic and zooarchaeologicalanalysis of EWT, identifying the economically important animalsand their deposition and preservation patterns. We then put ourresults in wider diachronic perspective by considering the pre-

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 19

Natufian sites. In such large and complex sites it is quite possiblethat intra-site depositional mechanisms, type of habitation andtopographic factors blur the sought paleoeconomic signals throughtime (Munro and Grosman, 2010; Yeshurun et al., 2013a, submittedfor publication). Thus, we narrow the geographic, ecological and‘functional’ (e.g., site type) scope by focusing exclusively on a seriesof Epipaleolithic ‘base-camps’ on the Israeli coastal plain (Fig. 1A).This sequence contains the well-published archaeofaunas of NahalHadera V (NHV, Early Epipaleolithic, Kebaran Culture), Hefzibah 7e18 (HEF) and Neve David (NVD, Middle Epipaleolithic, GeometricKebaran Culture) and the new results from EWT, presented below(Late Epipaleolithic, Early and Late Natufian Culture). The coastalplain series forms a natural geographic cluster during the Epi-paleolithic (Bar-Oz et al., 1999; Bar-Oz and Dayan, 2002, 2003; Bar-Oz, 2004; Bar-Oz et al., 2004).

Stiner and colleagues (Stiner et al., 1999, 2000; Munro, 2004)argued that intensification needs to be measured by categorizinganimals according to their locomotion and resilience, rather thantraditional taxonomic groups. Therefore, the first avenue of inves-tigation examines trends in the taxonomic abundance (based onNumber of Identified Specimens, NISP) of small versus big game,larger (medium-size) ungulates versus small ungulates, and fastsmall game versus slow small game, following Stutz et al. (2009:their Table 4). Stutz et al. (2009) predicted, based on the behavioralecology approach explained above, that economic intensificationwould be reflected by: (1) elevated hunting of small game toaugment dwindling ungulates; (2) decrease in larger ungulates(primarily fallow deer and aurochs) relative to the more resilientsmall ungulates (gazelle); and (3) acquisition of fast-escape smallanimals (e.g., hare and partridge) rather than easily caught but lessresilient slowly escaping small animals (e.g., tortoise). These trendsmay reflect human predation pressure at different scales. As un-gulates have larger home ranges than most small game species,they provide a measure of hunting pressure on the regional scale,while small game response to predation may be more local-scale,around the sites themselves (Stiner and Munro, 2002; Munro,2009, 2012).

The second line of inquiry investigates culling patterns (age andsex profiles) of mountain gazelle, the most abundant ungulate preyin the Epipaleolithic of the Mediterranean southern Levant. It hasbeen suggested, based on the same set of considerations as above,that intensifying foragers inflicting predation pressure on gazelleherds will progressively broaden their diet by hunting more low-return individuals, such as juveniles and eventually fawns, whichprovide less meat and fat than adults (Munro, 2004, 2009).

The third line of inquiry, which constrains our paleoeconomicinferences, employs taphonomic measures of site occupation in-tensity to evaluate the habitation type of the different phases inEWT (Yeshurun et al., submitted for publication). These measuresinclude volumetric bone densities and frequencies of post-discarddamage such as weathering, animal gnawing, trampling and non-nutritional burning and fragmentation. Volumetric densities serveas rough correlates for the intensity of site use, assuming that, allelse being equal, higher quantities of refuse will be discarded whenmore people inhabit the site for a given time. The magnitude ofpost-discard destruction processes among the EN and LN sampleswill serve as an estimate of occupation intensity; a greater numberof repeated occupations or more prolonged occupations wouldhave a different taphonomic effect on the faunal remains than shortand sporadic occupations. The latter would theoretically displayless non-nutritional burning (Stiner and Munro, 2011), reducedrates of trampling and subsequent dry breakage (Haynes, 1983),and more weathering and animal damage as a result of decreasedbuildup of cultural refuse, leaving the bones exposed for a longerduration (Behrensmeyer, 1978; Kent, 1993; Magdwick and Mulville,

2012). Thesemultiple measures, though rough and indirect proxies,are tied to the type of habitation, which we strive to keep as con-stant as possible when comparing paleoeconomic indices throughtime.

The site and its setting

El-Wad Cave is part of the UNESCOWorld Heritage Site complexof Nahal Me’arot/Wadi el-Mughara that also includes the caves ofTabun, Jamal and Skhul (Garrod and Bate, 1937). The site, a largecave with an adjacent terrace containing a long and rich Early, Lateand Final Natufian sequence, is situated on the western face ofMount Carmel, Israel, where the mountain cliff meets the openexpanses of the Mediterranean coastal plain, 45 m above modernsea level, within the Mediterranean climatic zone of the Levant(Fig. 1A). The site was first investigated by Lambert in 1928(Weinstein-Evron, 2009), but became well-known as a result ofGarrod’s 1929e1933 excavation campaign (Garrod and Bate, 1937).Garrod’s finds from el-Wad were the foundation of her subsequentdefinition and interpretation of the Natufian culture as a transi-tional phase between foraging and fully agricultural life-styles(Garrod, 1932, 1957). The terrace was later revisited (Valla et al.,1986), as was the cave (Weinstein-Evron, 1998). The renewed andon-going excavation was initiated in 1994 and has been focused onthe north-eastern part of the terrace (Fig. 1B). An area of ca. 70 m2

was exposed and the attained thickness of Natufian sedimentsranges between ca. 0.5 and 1.5 m. Structures, burials and a highdensity of finds, specifically chipped lithic and groundstone tools,bone tools, bone and shell ornaments, ochre and a rich faunalassemblage were retrieved during the renewed excavation(Weinstein-Evron et al., 2007, 2012, 2013; Yeshurun et al., 2013b).

A composite stratigraphy of the site, based on a compilation ofdata from all excavations (Weinstein-Evron, 2009; Weinstein-Evron et al., 2013), suggests an ephemeral occupation at the baseof the Early Natufian (designated Early Early Natufian or EEN),followed by a prolific burial phase comprising almost 100 in-dividuals (Middle Early Natufian or MEN) and culminating with theLate Early Natufian (LEN), the ‘classic’ Early Natufian layer of thesite, with its varied architectural features (Figs. 1D and 2). Thisphase appears as a massive, >0.5 m thick accumulation of repeatedoccupations. Overlying this architectural phase are thick EarlyNatufian living levels with a few stone features but normally lack-ing structures. The Natufian sequence ends with a thinner LateNatufian layer devoid of architecture, but displaying several con-centrations of graves (Weinstein-Evron et al., 2007; Bachrach et al.,2013). The Early Natufian occupations (Garrod’s layer B2, our Unit2) are much thicker and richer than the Late and Final Natufianlayers throughout the site (Garrod’s layer B1; here, the upper part ofUnit 2 and base of Unit 1).

The renewed excavation has exposed an architectural complexin the LEN phase, composed of a 9-m long curvilinear wall (Wall I)encompassing a sequence of at least nine architectural sub-phases,each defined by a thin stony floor. In the area enclosed by Wall I,several partially preserved stone structures and stone-rich ‘livingfloors’ have been excavated, of which the best preserved is Struc-ture II (Fig. 1D). Several discrete types of Early Natufian architec-tural contexts were defined: inside Structure II, including stonyfloors and between-floor fills; outside Structure II, in levels corre-sponding to some or all of the dwelling floors; the area of Locus 67,a massive stone pile lying northwest of the Structure II area; andLocus 25, an amalgamation of stones outside the living compounddemarcated by Wall I (Fig. 1D). A contextual taphonomic analysisindicated meaningful differences in the tempo and mode of depo-sition of the faunal remains among these contexts, ranging fromrecurrent deposition of consumption refuse in domestic space in

Figure 2. (A) Section through the NE terrace excavation, corresponding to the NeO line in Fig. 1D. Note the stone architecture (Walls I and II) in the lowest attained levels. The insetdepicts the proportions of Helwan-retouched lunates, abruptly-retouched lunates and microburins in each phase (after Weinstein-Evron et al., 2012); (B and C) examples of the ENarchitecture in the renewed excavation.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3520

the first three contexts to an occasional toss area in the last one(Yeshurun et al., submitted for publication). Importantly, the studyconcluded that the Wall I living area at the heart of the excavationcontains discarded fauna in primary deposition in the context ofwell-preserved architecture, and that post-discard damage wasinflicted on the bones as a result of repeated human occupation anda fast rate of accumulation in the same spot. This depositionalcircumstance caused indirect burning, trampling and non-nutritional fragmentation, in tandem with low weathering andanimal gnawing. In contrast, the volumetric density of faunal re-mains, as well as burning, trampling and fragmentation, is mark-edly reduced outside of the Wall I compound (Locus 25) andweathering and gnawing appear in elevated proportions, all indi-cating a non-domestic locale (Yeshurun et al., 2013b, submitted forpublication). The distribution of groundstone items, bone tools andmicromammal remains generally show similar patterns of intra-site use, namely the discard of used, broken pestles and bonepoints and consumed rodents in the Wall I complex rather thanoutside it (Yeshurun, 2011; Rosenberg et al., 2012; Weissbrod et al.,2012).

Thirteen radiocarbon measurements on charcoals and ungulatebones yielded a calibrated age range of 14,660e14,030 years BP(�1s) for the architecture-bearing phases and a range of ca.15,000e13,000 years BP (�1s) for the entire Unit 2 accumulation(Eckmeier et al., 2012; Weinstein-Evron et al., 2012). The dating ofthe Late Natufian contexts is underway. The density of finds in theWall I compound is extremely high, but human remains arevirtually absent. The stone structures, numerous living floors,density and diversity of finds and the absence of burials indicatethat this part of the site was used primarily for habitation and dailyactivities in the later parts of the Early Natufian. The character of

the site in LN times is less clear; our ensuing comparison to bothkinds of EN depositions (dwelling contexts versus Locus 25) willhelp in tracing the nature of the LN settlement at this part of the siteand illuminate changes in site function and intensity of occupation.

Materials and methods

Sampling

The faunal assemblage retrieved by the renewed excavationis vast, and therefore had to be sampled with special attentionto the chronological and contextual proveniences. Diachronicarchaeofaunal variability was investigated by comparing two sub-assemblages: an EN sample from the lowest attained phases (theWall I compound; 9141 identified specimens) against an LN samplefrom the same spatial location (Unit 1e2 and Phase W-0 in Unit 2,squares OeP/6e7; see Fig.1C; 2420 identified specimens). In the ENlithic sample, >70% of all lunates are Helwan-retouched, micro-burins are very rare, and all radiocarbon dates are older than13,300 years cal BP, in line with other EN occurrences (Weinstein-Evron et al., 2012). In the LN sample, >64% of the lunates areabruptly retouched, and microburins are present in ratios of 1:2 to1:4 compared with the lunates. In these squares, located in themiddle of the excavation, mid-way between the cliff wall and thefall of the talus (Figs. 1 and 2), the deposits attain maximumthickness and are minimally disturbed. Importantly, in thesesquares no burials were cut into the layer, making in situ preser-vation more likely than in the east area of the excavation, whereeleven LN graves were dug, creating a significant potential formixture of sediments. Still, the EN sample is much larger than theLN sample by virtue of the much larger magnitude of the former

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 21

layer. The material in the phases vertically sandwiched betweenour two samples was left for future investigations, in order toensure that the EN versus LN samples were as clean and distinct aspossible. The good architectural preservation of the lower layersallows the comparison of the LN material with two kinds of ENdeposits: inside Structure II, interpreted as a domestic area withconsumption refuse in primary deposition, and Locus 25, inter-preted as an occasional toss-zone outside the living compound(Fig. 1D; Yeshurun et al., submitted for publication). The geogenicprocesses contributing to the formation of the strata appear to besimilar in both the EN and LN samples in this work. Both are hardlydifferentiated sedimentologically (Weinstein-Evron et al., 2007),suggesting a similar rate of sedimentation in the central part of theterrace excavation (i.e., the Structure II area in the EN sample, andthe studied LN sample). Care was taken not to sample the LNcemetery or other contexts conspicuously different in their’behavioral’ mode of deposition compared with the EN sample inthis study.

Faunal analysis procedures

All faunal remains in this study were either plotted and packedin the field or retrieved by wet sieving employing 5 mm sieves. The5 mm sieves were used to separate the zooarchaeological assem-blage described here from most of the micromammal remains,which were captured by the underlying 1 mm sieve and are un-dergoing detailed study (Weissbrod et al., 2005, 2012, 2013).Extremely few recognizable remains of animals larger thanmicromammals passed through the 5 mm sieve, in line withempirical studies comparing sieving methods (Lyman, 2008). Theanalysis included reptiles, birds larger than Passeriformes and allmammals except rodents, insectivores and bats. The NISP (definedas fragments whose precise location in the skeletal element, orportion thereof, can be determined and quantified, and can beassigned to species or size class) was as inclusive as possible (seeYeshurun, 2011 for details of skeletal-element identification andcounting). Intra- and inter-site taxonomic comparisons in this workare based on the NISP counts (Lyman, 2008). Due to the highfragmentation of the bone assemblage, the majority of identifiedspecimens were assigned to size class rather than species (Table 1).While the assemblage includes large numbers of unidentifiedfragments (a sample of which was counted and divided by bonetype, size and burning for the sake of bone-fracture analyses(Yeshurun, 2011)), all of the following procedures concern the NISP.Body-part profiles were assessed by calculating Minimum Numberof Element (MNE; Lyman, 1994) values for each bone portion.Standardizing the MNE by also taking bone side and age into ac-count produced theMinimumNumber of Individuals (MNI) figures.

All identified specimens except isolated teeth and squamatevertebrae were systematically examined for bone-surface

Table 1Animal groups used in this study.

Group Species in the Epipaleolithic assemblages

Small mammal Lepus capensis, Vulpes vulpes, Martes foina,Felis sp., Meles meles, Herpestes incheumon,Vormela peregusna, Canis lupus

Small ungulate Gazella gazella, Capreolus capreolusMedium ungulate Dama mesopotamica, Sus scrofaLarge ungulate Bos primigeniusTortoise Testudo graecaLizard/snake Ophisaurus apodusBird Alectoris chukar, Grus grus, Ardeidae, Falconiformes,

Strigiformes

The dominant species in each group appears in bold.

modifications, using a stereoscopic microscope (Olympus SZX7)with a high intensity oblique light source, at 8e56 magnification,following the procedure described in Blumenschine et al. (1996).We searched for cutmarks (Binford, 1981) and hammerstone per-cussion marks, including conchoidal notches (Bunn, 1981; Capaldoand Blumenschine, 1994; Pickering and Egeland, 2006) and per-cussion pits and striations (Blumenschine and Selvaggio, 1988;Blumenschine et al., 1996; Pickering and Egeland, 2006). Evidencefor bone-working, ubiquitous in many Natufian contexts, wasrecorded (Campana, 1989). We also looked for carnivore punctures,scoring and digestion marks (Binford, 1981), as well as rodent gnawmarks (Brain, 1981) and biochemical (root) marks (Domínguez-Rodrigo and Barba, 2006). Trampling striations (Behrensmeyeret al., 1986; Domínguez-Rodrigo et al., 2009; de Juana et al., 2010;Gaudzinski-Windheuser et al., 2010) and abrasion of bone edges(Shipman and Rose, 1988) were sought, and weathering was noted(Behrensmeyer, 1978). Burning presence and intensity wererecorded by bone color, generally following Stiner et al. (1995).Bones with external and internal faces were assigned two burningcodes according to their external and internal surfaces, in order todiscern when in the life history of the specimen burning took place(i.e., fleshed bone, defleshed bone, or cracked bone, following Cain,2005). The mode of bone fragmentation was assessed by recordingshaft fracture-plane typology, shaft circumference and fragmentlengths (Villa and Mahieu, 1991; Bar-Oz, 2004),

Age-at-death profiles of gazelle, the most prominent game an-imal in the assemblages, were reconstructed by tooth eruption andwear patterns and by bone fusion, following Munro et al. (2009).We recorded the occlusal wear stages of isolated lower M3, P4 anddP4 teeth and of M3 through P4/dP4 teeth in mandibles (where theM2 and M1 could be distinguished). The tooth eruption and weardata, and epiphyseal fusion data were assigned actual ages ac-cording to Munro et al.’s (2009) study of modern mountain gazellespecimens from Israel, and also according to Davis’ (1983) scheme,to ensure the compatibility with older studies. The proportion ofgazelle fawns was measured by the unfused proximal epiphysis ofthe phalanx I, which fuses at approximately five to eight months ofage (Munro, 2009).

Gazelle sexing was performed using applicable character traitsand by multiple measurements, as defined by Munro et al. (2011)for modern, known-sex mountain gazelles from Israel. The char-acter traits used here included horn cores, the caudal wing tip ofthe atlas and the cross-section of the pubis shaft. Measurementsincluded the relevant portions in the pelvis, atlas, axis, phalanx II,calcaneum, metapodials, tibia and radius (see Munro et al., 2011:their Table 7 for details of the measurements taken on each bone).Additionally, we measured the distal humerus (BT*HDH, followingDavis, 1981) to facilitate comparisons with older studies, where itwas widely used for reconstructing both sex ratios and body-size(e.g., Davis, 1981; Bar-Oz et al., 2004). Even though this measureis not as sexually dimorphic in gazelles as previously thought(Munro et al., 2011), it may still yield important information on sex-independent body-size patterns.

Statistical significance was tested at a ¼ 0.05 level (Shennan,1988), using PAST freeware (Hammer et al., 2001) and SPSSversion 19 software. We employed the c2 distribution forcomparing nominal variables, such as the relative abundance ofspecies or taphonomic variables among selected samples. Whencomparingmany variables, Adjusted Residuals (AR) were calculatedfor each cell and presented along with the composite c2, in order todiscern which cells most significantly differ from the expectedvalues. Adjusted Residual values are standard normal deviates,indicating the probability that a single-cell comparison is statisti-cally significant. Significant ARs have values equal to or greaterthan �2 (Everitt, 1977; see Grayson and Delpech, 2008 for a

Table 2Taxonomic composition of the EWT assemblages.

Early Natufian Late Natufian

NISP MNI NISP MNI

UngulatesGazella gazella 1068 43 300 11Capreolus capreolus 2 1 1 1Small ungulate 2873 697Dama mesopotamica 24 2 3 1Sus scrofa 42 2 6 1Medium ungulate 88 7Bos primigenius 5 1 7 2Large ungulate 6 1Fetus/neonate ungulate 3 1

Small mammalsLepus capensis 231 17 30 3Smal mammal-indet. (1) 212 32Vulpes vulpes 186 7 43 3Canis lupus 20 2Martes foina 19 2 6 2Felis sp. 25 3 3 1Meles meles 11 1 2 1Herpestes incheumon 2 1Vormela peregusna 1 1Small carnivore-indet. 112 27

BirdsAlectoris chuckar 43 10 5 1Grus grus 1 1Ardeidae 2 1Falconiformes-small 3 1Falconiformes-medium 2 1 1 1Falconiformes-large 3 1Strigiformes 6 1 1 1Raptor (indet.) 7Bird-medium 14 2Bird-large 1

ReptilesTestudo graeca 1134 38 303 11Ophisaurus apodus (2) 78 13 10 2Lizard (3) 481 94Snake (3) 2424 834Squamata (indet.) (4) 12 9 4 1Total 9141 2420

Notes: (1) May include remains of small carnivores; (2) All skull and jaw fragments;(3) All vertebrae; (4) All jaws.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3522

zooarchaeological application). We compared anatomical mea-surements and bone fragment lengths using Student’s t-test andanalysis of variance (ANOVA). For comparing taxonomic diversityand similarity we used the ShannoneWiener Index for Heteroge-neity (H), which considers both taxonomic richness and evenness(Krebs,1989). We used the Evenness Index e (Shannon H divided bythe natural log of the number of taxa or categories) for assessing thelevel of evenness of the values of a given variable across anatomicalelements, or the evenness of the relative frequencies of age classes(Faith and Gordon, 2007). The latter analysis was done to quanti-tatively discern even culling of age-classes from uneven culling, astrategy focusing on the exploitation of one or two age classes(Davis, 1983).

Results and preliminary discussion

Zooarchaeology and taphonomy of el-Wad Terrace

The Natufian of EWT is extremely rich in faunal remains (andother classes of finds), yielding 2100e2500 identified specimensper cubic meter of sediment.Taxonomic spectrum The sample is based on 9141 identified spec-imens in the EN layer, and 2420 identified specimens in the LN layer,representing a minimum number of 159 and 43 individual animals,respectively (Table 2). The MNI tallies are strongly correlated withthe NISP counts (Spearman’s rho: r ¼ 0.92, p < 0.001 and r ¼ 0.90,p < 0.001 for the EN and LN samples, respectively). This tight fitmeans that the proportional representation of animals of verydifferent build (e.g., ungulates versus tortoises) is not biased dueto our identification and counting procedure; for example, tortoiseshell counts did not inflate the frequency of this animal. The ENand LN samples are similar in their taxonomic spectrums. Bothassemblages are diverse and very uneven (Evenness Index:e ¼ 0.27 and 0.29, respectively). Ungulates are the most numerousgroup in NISP terms, followed closely by squamates (lizards andsnakes). Tortoises and small mammals are frequent, and somebirds are also found (Fig. 3).

The squamates are very well represented in the assemblages, bytheir skull pieces and especially vertebrae (Table 2). The onlysquamate that was identified to species in this study, based on thedentary bone, is the legless lizard (Ophisaurus apodus), a large lizardthat may reach 1.2 m in length and weighs 300e600 g (Amitai andBouskila, 2003). Ungulates make up 42e44% of the assemblageswith mountain gazelle (Gazella gazella) being by far the mostcommon ungulate species in both periods. Larger ungulates are rare(Table 2). The spur-thighed tortoise (Testudo graeca) is abundant inboth periods. It is represented by numerous carapace, plastron andlimb remains. Small mammals are well represented by the Capehare (Lepus capensis) and by four to seven species of small carni-vores, predominantly red fox (Vulpes vulpes) (Table 2). Birds areinfrequently represented in both periods (Table 2, Fig. 3).

Complementing the faunal spectrum of Natufian EWT, but notincluded in this study, are numerous micromammals (Weissobrodet al., 2005, 2013), and some fish and marine shellfish. Amongthe rodents, mole-rats (Spalax ehrenbergi) are very abundant, andtheir taphonomic, contextual and demographic traits suggest reg-ular consumption by humans (Weissbrod et al., 2012). Fish arepresent along the sequence. Desse (in Valla et al., 1986) identifiedfour families of Mediterranean fish (Sparidae, Serranidae, Mullidaeand Muglidae) from the 1980s soundings at EWT. The fish from therecent excavation are yet to be studied; thus far Sparidae have beententatively identified (I. Zohar, Personal communication). Thepresence of these taxa suggests exploitation of the Mediterraneancoastal plain and estuarine zones (Bar-Yosef Mayer and Zohar,2010). Large and rich assemblages of marine shells were

retrieved, including edible mollusks, such as the Mediterraneanlimpet Patella caerulea (Weinstein-Evron et al., 2007; Bar-YosefMayer and Zohar, 2010).

Vertebrate taphonomy Counts of bone-surfacemodifications in theEN and LN samples are grouped by the four mammal size classes(small mammal, small ungulate, medium ungulate, and largeungulate), the tortoises, birds and squamates (Table 3). Since thesquamates were not subjected to systematic microscopicinspection, only burning data are given for them. Bone-surfacemodification data indicate that the assemblages overwhelminglyconstitute butchered, consumed and discarded animal remains,which consequently underwent some pre- and post-burialdestruction processes. Cutmarks related to all butchery stages, andespecially dismemberment and filleting, attest to the exploitationof all ungulates, hare, fox and tortoises as food. A modest amountof carnivore ravaging affected the assemblages, probably by asmall-medium carnivore. Subsequently, some of the skeletalelements suffered moderate exposure related damage attested byweathering, physical abrasion, rodent gnawing, and trampling.

Burning (indicated by bone color) appears on 27% of NISP in theEN sample and on 24% of NISP in the LN sample. A sample of un-identified small ungulate specimens from the EN layer yielded anidentical proportion of burned bones, 27% (1553 of 5670 specimens

Figure 3. Taxonomic composition based on lumping species and size-classes intohigher taxonomic groups: (a) EN sample; (b) LN sample.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 23

>8 mm in maximum dimension). In spite of the fact that boneburning is abundant in EWT, burning intensity is not high. Toquantify bone burning, we used the Combustion Index (CI) definedby Costamagno et al. (2005). The index values range from 0 (noburned bones) to 1 (when all specimens are calcined), and thusmeasures both burning frequency and intensity. The CI of the entireassemblages is 0.10e0.11, reflecting moderately burned bones inboth periods. Looking at burning intensity by taxonomic group, thesquamates display the highest burning intensity (highest CI) duringboth periods, with asmuch as 7e8% calcined bone. Birds display theleast burning, while small mammals, small ungulates and tortoisesdisplay similar CI values (Table 4).

In the gazelle size class, the distribution of burning by skeletalelement is quite even and does not correlate with food utility.Numerous cortical and compact bones (limb shafts, toes, teeth,carpals/tarsals) display burning, and therefore preferential burningof fat-rich portions cannot be demonstrated (Table 5). Additionally,in both the EN and LN samples, most shaft fragments were burnedto a similar intensity on their exterior and interior surfaces,meaning that the bone was burned when already defleshed andfractured. Nevertheless, a significant difference exists between thetwo samples in that during the LN a considerable number of shaftsexhibited burning with greater intensity on the exterior (25% inthe LN versus only 10% in the EN; c2 ¼ 12.49, p ¼ 0.002; Table 5).Hence, it seems that the majority of ungulate gazelle bone burningin EWT, especially in the EN sample, was unintentional, in accor-dance with lighting of hearths on pre-existing refuse, therebyunintentionally inflicting secondary burning on bones buried un-derneath (Stiner et al., 1995; see also; Hardy-Smith and Edwards,2004). The averaging pattern of such an activity in the faunal as-semblageswould eventually lead to near-uniform burning intensity

and frequency on different skeletal parts, probably masking someof the effects of roasting body-parts or heating bones for marrowextraction.

The analysis of breakage patterns was conducted on mammallimb bones and phalanx I specimens, yielding largely similar resultsfor both periods (Table 3). Green, or fresh, fractures, attributed todeliberate fracturing by humans or carnivores to access the bonemarrow, appear on 42e44% of the small mammal and ungulatebones. This sizeable proportion is in line with a processed assem-blage that subsequently suffered additional cycle/s of breakagewhile the bones were already dry, resulting in intermediate and drybreaks amounting to more than half of the specimens (Villa andMahieu, 1991).Gazelle mortality profiles Mortality profiles were constructed forgazelles by three tooth eruption and wear series and by two bonefusionmethods (following Davis, 1983;Munro et al., 2009). The rawtooth-wear and bone fusion data are presented in SupplementaryOnline Material [SOM] Tables A.1 and A.2, and five differentmethods for presenting age data are compared in Fig. 4.

In the EN sample, the three tooth eruption and wear seriesindicate a low number of juvenile (<18 months) gazelles,amounting to 10e18% of ageable specimens. A more detailedmortality profile, covering the entire life-span of the animal, in-dicates the overwhelming preponderance of adult animals, 18e96months of age, alongwith a few fawns and juveniles (2e18months)and one very old animal (>96 months) in the EN sample (Fig. 4).However, the picture changes when bone fusion is examined,indicating a much larger proportion of juveniles (34e53%). Themost recently published and refined bone fusion method, basedupon a large sample of known-age gazelles from Israel, diverges inthe most marked way from the values obtained by tooth eruptionandwear (both followingMunro et al., 2009). The skeletal elementsadvocated by Munro et al. (2009) as being comparable to the dP4-P4/M3 replacement and the attainment of adult body size are theproximal humerus, proximal tibia and distal radius (MNE ¼ 17 inthe EN sample), while the older and widely used method of Davis(1983) employs the summation of fused and unfused distal radii,metapodials, tibiae, femora and the tuber calcis (MNE ¼ 104 in theEN sample).

The results of the LN sample show a much sounder agreementamong tooth eruption and wear and bone fusion methods, eventhough the samples are smaller than the EN sample (Fig. 4). All agepresentation methods indicate that a significant portion (31e50%)of the culled gazelles were juveniles. Notably, Davis’ (1983) andMunro et al.’s (2009) bone-fusion methods generate nearly iden-tical figures here (31% and 33% juveniles, respectively) and theproportion of juveniles is lower when presented by bone fusionthan by tooth eruption and wear, in contrast to the situation in theEN sample. The detailed dentition-based sequence points to cullingof just two age groups, in equal proportions: fawns (two to sevenmonths of age) and young adults (18e36 months).

The marked discrepancy between the dental eruption and wearand the bone fusion methods in the EN sample is difficult toreconcile. Juvenile epiphyses could be less well preserved becauseof their greater porosity, and epiphyses in general may be under-represented in relation to teeth in an assemblage that suffereddensity-mediated attrition and carnivore ravaging. This bias is ex-pected to reduce the number of juveniles in the bone fusionmethod comparedwith the tooth wearmethod, but the age profilesexhibit the reverse pattern, namely a much larger representation ofjuveniles by epiphyseal fusion.

Looking at the EN bone fusion results more closely (SOMTable A.2), it appears that early-fusing elements are usually foundfused and that the high juvenile ratio stems from the late-fusingelements. For example, the fusion state of the proximal phalanx I

Table 3Bone-surface modifications and bone fracture patterns for all taxonomic groups: (a) Early Natufian sample; (b) Late Natufian sample.

A Smallmammal

Smallungulate

Mediumungulate

Largeungulate

Tortoise Lizard and snake Birds Total

(a)NISP 819 3943 154 11 1134 2995 83 9139Burning n 212 1027 32 3 279 929 11 2493

% 25.9% 26.0% 20.8% 27.3% 24.6% 31.0% 13.3% 27.3%Green fracture n 20 328 10 1Dry fracture n 18 233 14 0Intermediate n 10 172 10 0Limb shaft circumference <50 22 900 39 1 0 962

>50 0 8 0 0 0 8100 56 42 4 0 30 132

Weathering (stage 3e5) n 6 70 8 2 4 1 91of 749 3775 136 10 1132 83 5885% 0.8% 1.9% 5.9% 20.0% 0.4% 1.2% 1.5%

Cutmarks n 21 239 14 1 6 3 284% 2.8% 6.3% 10.3% 10.0% 0.5% 3.6% 4.8%

Percussion marks n 0 108 3 0 111of 78 1616 71 1 1766% 0.0% 6.7% 4.2% 0.0% 6.3%

Working n 0 91 2 0 0 1 94% 0.0% 2.3% 1.3% 0.0% 0.0% 1.2% 1.0%

Gnawing (carnivore) n 17 212 8 1 14 0 252% 2.3% 5.6% 5.9% 10.0% 1.2% 0.0% 4.3%

Gnawing (rodent) n 8 63 3 0 4 0 78% 1.1% 1.7% 2.2% 0.0% 0.4% 0.0% 1.3%

Root-marks n 215 1238 55 5 185 17 1715% 28.7% 32.8% 40.4% 50.0% 16.3% 20.5% 29.1%

Trampling striations n 23 299 15 2 11 2 352% 3.1% 7.9% 11.0% 20.0% 1.0% 2.4% 6.0%

Abrasion n 27 169 2 0 13 5 216% 3.6% 4.5% 1.5% 0.0% 1.1% 6.0% 3.7%

(b)NISP 143 998 16 8 303 942 9 2419Burning n 31 267 2 1 57 217 0 575

% 21.7% 26.8% 12.5% 12.5% 18.8% 23.0% 0.0% 23.8%Green fracture n 2 83 0 0 85Dry fracture n 3 71 1 0 75Intermediate n 1 42 1 0 44Limb shaft circumference <50 2 190 2 0 0 194

>50 0 1 1 0 0 2100 7 11 0 0 4 22

Weathering (stage 3e5) n 0 48 1 0 0 0 49of 132 965 15 4 303 9 1428% 0.0% 5.0% 6.7% 0.0% 0.0% 0.0% 3.4%

Cutmarks n 5 39 2 1 0 0 47% 3.8% 4.0% 13.3% 25.0% 0.0% 0.0% 3.3%

Percussion marks n 0 13 0 0 13of 9 346 7 1 363% 0.0% 3.8% 0.0% 0.0% 3.6%

Working n 0 16 0 0 1 0 17% 0.0% 1.6% 0.0% 0.0% 0.3% 0.0% 0.7%

Gnawing (carnivore) n 2 61 0 1 4 0 68% 1.5% 6.3% 0.0% 25.0% 1.3% 0.0% 4.8%

Gnawing (rodent) n 3 14 0 0 2 0 19% 2.3% 1.5% 0.0% 0.0% 0.7% 0.0% 1.3%

Root-marks n 41 326 8 4 84 3 466% 31.1% 33.8% 53.3% 100.0% 27.7% 33.3% 32.6%

Trampling striations n 3 76 1 1 4 0 85% 2.3% 7.9% 6.7% 25.0% 1.3% 0.0% 6.0%

Abrasion n 7 39 1 6 1 54% 5.3% 4.0% 6.7% 0.0% 2.0% 11.1% 3.8%

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3524

(which fuses at five to eight months) indicates 13% fawns, closelyreplicating the lower proportion of juveniles obtained by thedental methods. Thus, it may be that near-adult specimens weremisidentified as young-adults in the tooth wear methods. More-over, the modern tooth eruption and wear, and bone fusion se-quences were derived from captive animals (for which knownages at death exist), and these could exhibit somewhat differentwear patterns due to different diets, and/or varying rates of growthand development due to their captive status. This suggestion is

obviously hard to test without more actualistic data. Presently, wedraw on the fact that the high juvenile ratio is caused by late-fusingelements to suggest that adults, as well as animals that nearlyreached adulthood and were osteologically immature, were pri-marily culled during the EN. The detailed tooth wear series point to‘even culling’ of various age classes (e ¼ 0.712), with specialemphasis on adults, during the EN. Conversely, during the LN fewerage classes were targeted (e ¼ 0.387) and more juveniles werecaptured.

Table 4NISP counts of bones burned to various degrees and the Combustion Index for eachtaxonomic group.

Unburned Brown Black Gray White TotalNISP

Combustionindex

Early NatufianSmall mammal 607 127 71 11 3 819 0.10Small ungulate 2913 607 328 62 40 3950 0.10Med. ungulate 122 22 7 2 1 154 0.07Large ungulate 8 2 1 0 0 11 0.09Tortoise 855 160 99 11 8 1133 0.09Lizardand snake

2058 408 275 187 59 2987 0.15

Bird 71 6 4 1 0 82 0.05Total 6634 1332 785 274 111 9136 0.11

Late NatufianSmall mammal 112 16 14 0 1 143 0.08Small ungulate 731 154 96 7 10 998 0.10Med. ungulate 14 2 0 0 0 16 0.03Large ungulate 7 0 0 0 1 8 0.13Tortoise 246 32 22 3 303 0.07Lizardand snake

725 72 72 56 17 942 0.12

Birds 9 0 0 0 0 9 0.00Total 1844 276 204 66 29 2419 0.10

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 25

Using the dP4eM3 series results, the LN sample is much morebiased towards juveniles younger than 18 months than the ENsample, a difference that is statistically significant (c2 ¼ 5.43,p ¼ 0.02). These figures can be cautiously interpreted using thedrive counts conducted monthly by Baharav (1974) betweenNovember 1971eNovember 1972 on living gazelle herds in RamotYissakhar (southeast Galilee, Israel; Baharav’s year-round averagesof the proportions of age classes and sexes are used here, becausethey are reasonably comparable with the zooarchaeological data).The modern counts indicated that the proportion of juveniles (�18months of age) in living herds is approximately 45% on a year-roundaverage. The low juvenile proportion in the EN sample divergessignificantly from the natural age structure (c2 ¼ 5.40, p ¼ 0.02),highlighting the special emphasis on adult culling in that period.Conversely, the LN proportion of juveniles is similar to a naturalherd structure (c2 ¼ 0.07, p ¼ 0.78; see also Bar-Oz et al., 2004).

Gazelle sexing The sexing results of the character traits show amarked male bias in both periods (Fig. 5; SOM Table A.3) stemmingmainly from horn cores. This element is unmistakably dimorphicbut its utility for reflecting the actual sex ratios of the huntedpopulation has been questioned, because male horn cores are

Table 5Characteristics of burned specimens in the small ungulate group.

Early Natufian Late Natufian

Evenness index for theanatomical distributionof burning

0.98 0.99

Correlation FUI*elementburning

r ¼ �0.29, p ¼ 0.274 r ¼ 0.03, p ¼ 0.502

Limb burning NISP NISPTwo sides equally burned 308 39Exterior more burned 37 15Interior more burned 25 6Burned bone end 166 36.6% 39 31.7%Burned bone shaft 208 30.0% 23 20.7%Comparison end*shaft c2 ¼ 5.35, p ¼ 0.024 c2 ¼ 3.61, p ¼ 0.057

Comparison of fragmentlengths

BurnedNISP

UnburnedNISP

BurnedNISP

Unburned NISP

N: 991 2523 257 627Mean: 19.662 23.972 17.222 19.742Var.: 131.06 313.44 47.853 109.82T �8.51 (p < 0.001) �4.19 (p < 0.001)

much more robust than female horn cores, possibly reflecting apreservational difference, and because of the possibility that malehorn cores had been used as raw material and therefore mighthave been preferentially cached in the habitation area (Munro,2001; Bar-Oz, 2004). While no evidence exists for horn corecaches in el-Wad, ca. 10% of these elements bear traces offabrication (Yeshurun, 2011) and some finished artifacts made ofhorn were discovered elsewhere at the site (Garrod and Bate,1937; Weinstein-Evron, 1998). On the other hand, contextualtaphonomic analysis of the EN layer showed that both bone toolproduction waste and unmodified butchery refuse are present inthe living surfaces investigated here, with little evidence ofremoval of selected skeletal elements, especially not small andnon-obstructing elements such as female horn cores, from thehabitation area (Yeshurun et al., submitted for publication).Nevertheless, the male-dominated picture is augmented by theonly other available character trait in the samples, the pubic shaft,displaying four male elements versus one female element in theEN sample (SOM Table A.3). Only one sexable (male) pubis shaftwas available in the LN sample.

An additional avenue is sexing by osteometrics, which necessi-tates an investigation of general body-size variability (see SOMTable A.4 for the raw gazelle measurement data). Two skeletal el-ements that provide a sufficient number (n > 5) of measurementsin both the EN and LN samples are the distal humerus (BT and HDH,following Davis, 1981) and the phalanx II (GL, Bd and Dd, followingMunro et al., 2011). The measurements on these elements are in-dependent of age and taphonomic processes, since only fused,unburned and unabraded elements were measured (von denDriesch, 1976). Additionally, no significant intra-individual sizedifference is evident for the phalanx II (Munro et al., 2011). All fivemeasurements either show significantly higher means in the ENsample, or statistically similar means, while no measurementprovided a higher mean value in the LN sample (SOM Table A.5).Plotting the phalanx II measurements (greatest length versus distaldepth) in both samples revealed that the EN gazelles are dividedinto two size groups, with the LN gazelles corresponding with thesmaller one (Fig. 6). Thus it seems that EN gazelles possessed alarger mean body-size compared with LN gazelles, and that a po-tential cause may be shifting sex ratios in favor of females in the LN.

The samples that were measured and sexed using the discrim-inate functions, following Munro et al. (2011) (altogether 42 spec-imens in the EN sample and 14 in the LN sample) yielded resultsthat were mostly larger than the female/male cutting point for themodern specimens in the EN sample, and mostly smaller than thecutting point in the LN sample (SOM Table A.4; cutting pointsfollow Munro et al. (2011)). Had the EWT gazelles been identical insize to modern Israeli gazelles, the assemblages would becomposed of 81% males in the EN sample and 43% males in the LNsample, a statistically significant difference (c2 ¼ 7.47, p < 0.01).However, it is possible that Natufian gazelles differed in sizecompared with present-day gazelles, which could cause the cuttingpoint to be too low and to misidentify large females as males.

A possible solution to separate body-size trends and sex ratioswould be to determine if EN and LN gazelle body-size differedsignificantly from the modern Israeli gazelles that were used todetermine the sexing functions. This comparison (Table 6) indicatesthat EN gazelles usually possess a similar body-size to modernmales and are significantly larger than modern females, while LNgazelles exhibit the reverse pattern, being significantly smaller thanmodern males but as large as modern females. Importantly, neitherthe EN nor LN gazelles display larger body-size compared withmodernmales, or smaller body-size thanmodern females (Table 6).This indicates that the range of gazelle body-size in the EN and LNwas un-skewed compared with present-day gazelles, lending some

Figure 4. Gazelle aging. Left: juvenile-adult ratios based on five bone-fusion or tooth-wear series. Bone fusion 1 refers to the elements fusing at 18 months, equivalent to the dP4-P4/M3 replacement (following Munro et al., 2009). Bone fusion 2 follows the traditional method by Davis (1983), which includes more elements. Sample size is indicated to the right ofeach bar; Right: detailed mortality profile using the dP4eM3 series, based on tooth NISP. The numbers in parentheses indicate age in months.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3526

support to the inter-period body-size difference in EWT as indeedstemming from shifting sex ratios.

According to phalanx II measurements, male gazelles dominatethe EN assemblage by ratio of approximately 4:1 while males andfemales are similarly represented in the LN, perhaps with femalesslightly dominating. The inter-period difference in sex ratios ishighly significant (EN versus LN: c2 ¼ 7.47, p < 0.001). Baharav’s(1974) drive counts indicated that in modern Israeli herds, fe-males outnumbered males by a ratio of 100:81 in a year-roundaverage (see also Baharav, 1983). Thus, the EN sex ratios differsignificantly from the sex ratios of living herds (c2 ¼ 17.89,p< 0.001), indicating specific targeting of males. On the other hand,the LN sample is similar to the living herd in this respect (c2 ¼ 0.02,p ¼ 0.90), meaning that LN hunters culled gazelles with no pref-erence to sex.

Figure 5. Gazelle sexing by character traits.

Taphonomic markers for occupation type: Early versus LateNatufian Several taphonomic markers were defined above forgauging site-occupation intensity through the Natufian sequence ofEWT. We compare the LN with two different EN accumulations(described above), displaying clearer contextual (architectural)patterns: a domestic faunal assemblage in primary depositionfollowing consumption activities on the spot versus Locus 25, anon-domestic area beyond the living compound, containingoccasionally-tossed consumption refuse (Yeshurun et al., submittedfor publication).

The comparison results (Table 7) show that the volumetricdensity of faunal remains by total bone mass is highest in the do-mestic EN and about half of that in Locus 25 and in the LN. Whenthe volumetric densities of NISP counts are taken into account, the

8

9

10

11

12

13

14

18 20 22 24 26 28

phal

anx

II Dd

(mm

)

phalanx II GL (mm)

Figure 6. Gazelle sexing by body-size: comparison of phalanx II measurements in theEN (black diamonds) and LN (gray squares) samples.

Table 6Summary of the comparisons of Early Natufian (EN) and Late Natufian (LN) gazellemeasurements with modern male and modern female gazelle measurements.

BThumerus

GLphalanx II

Bdphalanx II

Ddphalanx II

Dd tibia

EN comparedwith modern male

No diff. No diff. No diff. No diff. No diff.

EN comparedwith modern female

No diff. Larger Larger Larger Larger

LN comparedwith modern male

Smaller Smaller No diff. Smaller

LN comparedwith modern female

No diff. No diff. Larger No diff.

Raw measurements and t-test results are detailed in SOM Tables A.5 and A.6.Modern gazelle measurements were taken from Munro et al. (2011: their Table 3).

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 27

difference among the domestic EN and the LN is less dramatic,while Locus 25 still lags behind. The NISP-based counts of bonemodifications (e.g., burning, gnawing, etc.) are transformed intoAdjusted Residuals (AR) in order to assess which counts signifi-cantly diverge from the expected values. Some significant differ-ences are evident in the statistical comparison (Fig. 7; see SOMTable A.7 for details, AR values and composite c2 results): sub-arealweathering is significantly higher in the Locus 25 and LN samplescompared with the domestic EN, attesting to slower rate of sedi-ment build-up. On the other hand, carnivore gnawing is signifi-cantly underrepresented in the LN. The slow-accumulatingmaterial in Locus 25 displays both higher weathering and moreintense carnivore gnawing compared with the LN sample. Bone-burning measures are similarly represented in the LN comparedwith the domestic EN. In contrast, burned bones are significantlyunderrepresented in Locus 25. Finally, trampling is significantlyoverrepresented in the domestic EN sample and significantly un-derrepresented in the LN sample. Post-discard fractures (definedhere as the ‘dry’ and ‘intermediate’ fracture categories combined)are significantly underrepresented in Locus 25 but similarlydistributed in the domestic EN and in the LN samples (Fig. 7,Table 7).

To conclude, the taphonomic measures of occupation intensityposition the LN as intermediate between the EN domestic and non-domestic samples. Some important characteristics such as boneburning and breakage align the domestic EN and the LN, high-lighting the similarity in post-discard destruction processes andconsequently, the degree of occupation repetitiveness and type ofhabitation. This similarity, in turn, helps to keep site type as con-stant as possible and renders the inter-phase paleoeconomiccomparison more robust.

Table 7Comparison of taphonomic markers for occupation intensity among the EarlyNatufian (EN) and Late Natufian (LN) samples.

EN domestic area(inside Structure II)

EN occasionaltoss area (Locus 25)

LN

Bone mass (gr)/m3 15,267 9883.89 7708.57NISP count/m3 2716.54 1738.61 2180.95Weathering

(long bones)13 (2.5%) 4 (14.3%) 25 (8.8%)

Carnivore gnawing 106 (6.2%) 11 (10.7%) 68 (4.8%)Rodent gnawing 35 (2.1%) 4 (3.9%) 19 (1.3%)Trampling striations 136 (8.0%) 8 (7.8%) 85 (6.0%)Burning (total) 944 (23.5%) 35 (12.0%) 575 (23.8%)Indirect burning

(shafts)136 (84.0%) 2 (67.0%) 45 (75.0%)

Dry þ intermediatefractures

149 (58.0%) 16 (36.0%) 119 (58.0%)

Excavation volumeswere not available for all units, so the faunal remains from theseunits were excluded from calculations.

El-Wad in diachronic context

A broader perspective of intra-Natufian economic trends isachieved by comparing the entire Epipaleolithic sequence of theIsraeli coastal plain, including the Kebaran site of Nahal Hadera V(Bar-Oz and Dayan, 2002), the Geometric Kebaran sites of Hefzibah(Bar-Oz and Dayan, 2003) and Neve David (Bar-Oz et al., 1999), andthe Early and Late Natufian faunas of EWT presented above. All ofthese sites are positioned in a north-south strip along the Mediter-ranean coastal plain of Israel, ca. 40 km long, representing denseanthropogenic accumulations of vertebrate faunas and other cul-tural refuse. Their zooarchaeology and taphonomy were publishedin detail (Bar-Oz, 2004) using compatible methods to the onesemployed here.Taxonomic abundances in the Epipaleolithic coastal plain The ENand LN phases of EWT are generally similar in the presence andabundance of taxa, but some variations do occur. Rather thanfocusing on particular differences in relative abundances of certaintaxa, which are sometimes difficult to interpret and may stem fromdiffering sample sizes or other idiosyncrasies, the taxonomicabundances are evaluated here using Stutz et al.’s (2009) indices,described above. Our results indicate no gradual temporal trends(as interpreted by Stutz et al., 2009); rather, marked differencescharacterize the pre-Natufian and Natufian assemblages,regardless of chronological position of the sample within thisdivision (Fig. 8, Table 8). The notable differences amongst the pre-Natufian and Natufian strata, namely the increase in small gameand the dwindling numbers of larger ungulates, overshadow thesubtle differences that exist among the Early and Late Natufianphases of EWT. The fast small game index declines during theNatufian of EWT, while the slow small game index remainsconstant in the EN-LN comparison.Gazelle culling patterns in the Epipaleolithic coastal plain Detailedinvestigations of gazelle age and sex profiles at EWT provide apicture of shifting culling patterns through the Natufian sequence.Putting the EWT results in broader Epipaleolithic perspective, thepre-Natufian gazelle culling trends are summarized in Table 9. Theproportion of juveniles in the Kebaran and Geometric Kebaranfaunas is 24e39%, higher than the Early Natufian EWT (18%).However, only the Neve David assemblage significantly differs inthis respect from the natural herd age structure (c2 ¼ 9.73,p ¼ 0.001). Nahal Hadera V and Hefzibah 7e18 do not differ froma natural herd, as also Late Natufian EWT (c2 ¼ 1.07, p ¼ 0.30 andc2 ¼ 0.35, p ¼ 0.55, respectively). The evenness of age classes issimilar across all assemblages (e ¼ 0.671e0.712) except for themarkedly low value of the LN sample from EWT (Fig. 9).Additionally, the proportion of gazelle fawns (about six monthsold or less), measured by the proportion of unfused to fusedproximal phalanx I, is variable, reaching 18% at the most in the LNassemblage. Notably, fawns increase significantly from the earlierEpipaleolithic to the Natufian (Fig. 9; pooled pre-Natufian versuspooled Natufian: c2 ¼ 11.66, p < 0.001). Since the average year-round proportion of fawns in living herds is about one-seventh ofthe population (Baharav, 1974), the low pre-Natufian proportionof fawns significantly differs from living herds (c2 ¼ 10.92,p < 0.001) while the Early and Late Natufian assemblages samplefawns according to their living abundance (c2 ¼ 0.25, p ¼ 0.62).

Sex ratios, reconstructed by body-size in comparison to known-sex gazelle collections, again showed an inter-phase difference inEWT: male culling in the EN versus unselective culling of sexes inthe LN. Comparison with the pre-Natufian Epipaleolithic of thesame ecological region can help separate conflating factors thataffect body-size (Bar-Oz, 2004). Gazelle body-size trends wereevaluated using the humerus BT measurement, a relatively well-preserved portion that was measured in previous studies. Results

-6

-4

-2

0

2

4

6

Adju

sted

Res

idua

ls

Early Natufian - domes c

-6

-4

-2

0

2

4

6

Adju

sted

Res

idua

ls

Early Natufian - non-domes c

-6

-4

-2

0

2

4

6

Adju

sted

Res

idua

ls

Late Natufian

Figure 7. Comparison of taphonomic markers for occupation intensity among the EN(2 different contexts) and LN samples, using Adjusted Residuals (AR). Positive ARvalues denote overrepresented traits, while negative AR values denote underrepre-sented traits. Significant values (AR � �2) appear as black columns.

00.10.20.30.40.50.60.70.80.9

1

NHV NVD HEF EWT-EN EWT-LN

Total big game index Total small game index

00.10.20.30.40.50.60.70.80.9

1

NHV NVD HEF EWT-EN EWT-LN

Large big game index Medium big game index

00.10.20.30.40.50.60.70.80.9

1

NHV NVD HEF EWT-EN EWT-LN

Slow small game index Fast small game index

Figure 8. Index values for the Epipaleolithic sequence of the Israeli coastal plain(abbreviations as in Table 8). Black dashed lines separate the Natufian and pre-Natufiansamples.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3528

indicate that gazelle body-size underwent significant changesduring the Epipaleolithic (ANOVA F ¼ 5.31, p < 0.001). Tukey’spairwise comparisons indicate that this difference derives mostlyfrom the EN assemblage, which displays significantly larger mea-surements than the preceding Nahal Hadera V and Neve Davidgazelles (Fig. 10A; data from SOM Table A.8).

The trend of larger gazelle body-size in the EN was attributed toincreased proportion of males, but other potential explanationsshould be considered, such as cooler climate (e.g., Bergmann’s rule,see Davis, 1981) and human impact on gazelle populations, assuggested by Cope (1991) and later countered by Bar-Oz et al.(2004). If cooling was the reason for the increase in gazelle body-size, other mammals should also follow this trend (Ducos andHorwitz, 1997). The only mammal available for comparison wasthe hare (humerus BTand scapula glenoid fossa BGmeasurements).Although the samples are small (SOM Table A.8.), the hare does notexhibit a parallel trend of body-size increase (EN versus pre-Natufian hares: ANOVA F ¼ 2.87, p ¼ 0.07 and F ¼ 0.07, p ¼ 0.959for the humerus and scapula measurements, respectively)(Fig. 10B,C). Hence, the most plausible interpretation of the signif-icantly larger-sized gazelles in the EN is change in sex ratios.Apparently males and females were culled in similar proportionsalong the coastal plain sequence, except for Early Natufian EWT.

In sum, during most of the Epipaleolithic hunters exploited awide range of gazelle age classes in the Israeli coastal plain, withvarying degrees of reliance on juveniles, and in most cases did notseek to capture a particular sex. The Early Natufian stands out inthis series for its emphasis on male culling and its particularly lowproportion of juveniles, albeit with a seemingly contradictory trendof taking fawns in relation to their natural abundance (Fig. 9).

Discussion

Similar site function through time

The ancient function of excavated localities may affect paleo-economic inferences. Feasting activities or ephemeral use of partsof large and complex sites may introduce serious biases whenevaluating the long-term subsistence trends of Epipaleolithic

Table 8NISP relative abundance indices used tomeasure intensification of animal resources during the Epipaleolithic and index values for the Epipaleolithic series of the coastal plain,following Stutz et al. (2009: their Table 4).

Formula KEB GKEB GKEB EN LN

NHV NVD HEF EWT-EN EWT-LN

Total big game index NISPbg/(NISPbg þ NISPsg) 0.93 0.94 0.94 0.67 0.69Total small game index NISPsg/(NISPbg þ NISPsg) 0.07 0.06 0.06 0.33 0.31Large big game index NISPlbg/(NISPlbg þ NISPsbg) 0.01 0.01 0.03 0.00 0.01Medium big game index NISPmbg/(NISPmbg þ NISPsbg) 0.31 0.34 0.22 0.04 0.02Slow small game index NISPssg/(NISPssg þ NISPsbg) 0.03 0.01 0.01 0.22 0.23Fast small game index NISPfsg/(NISPfsg þ NISPsbg) 0.08 0.07 0.06 0.18 0.13

‘Big game’ refers to ungulates: small, medium and large. ‘Small game’ refers to all other animals included in this study except squamates, divided to fast types (hare, smallcarnivores and respective size class, game birds such as partridge) and slow types (tortoise). Note that in Stutz et al. (2009) original indices small carnivores were excluded.Abbreviations: KEB, Kebaran Culture; GKEB, Geometric Kebaran Culture; EN, Early Natufian; LN, Late Natufian; NHV, Nahal Hadera V; NVD, Neve-David; HEF, Hefzibah 7e18;EWT, el-Wad Terrace.

Table 9Summary of gazelle age and sex data for the two assemblages in the present study(in bold type), augmented by data from the Epipaleolithic sequence of the coastalplain (Nahal Hadera V, Hefzibah 7e18 and Neve David, all taken from Bar-Oz, 2004).

KEB GKEB GKEB EN LN

NHV NVD HEF EWT EWTProportion of juveniles (dental) 39% 25% 24% 18% 50%Evenness of age classes 0.671 0.683 0.710 0.712 0.387Proportion of fawns

(unfused phalanx I)7.3% 8.5% 0.5% 13.0% 18.0%

Proportion of males(by body size)

Under 50% 81% 43%

Abbreviations are as in Table 8.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 29

households (Yeshurun et al., 2013a, submitted for publication; seealso Asouti and Fuller, 2013). We applied taphonomic proxies ofsite-occupation intensity to investigate this issue and control forchanging site type. Our intra-Natufian analysis within the key siteof el-Wad Terrace compared domestic and non-domestic EN ac-cumulations with the more enigmatic LN phase. Since the post-discard damage patterns in the LN sample are intermediate be-tween the domestic and non-domestic EN samples, we suggest thatthe difference between the two Natufian chronological phases wasof degree, not of kind. The intensive post-discard damage fromrecurrent human activities and the considerable quantities offaunal remains that were found in the LN probably mean that EWTcontinued to serve as an important habitation site during the latterhalf of the Natufian. This conclusion is somewhat unexpected,

Figure 9. Inter-phase comparison of gazelle culling patterns in the coastal plain Epi-paleolithic: proportions of juveniles (<18 months old); proportion of fawns (<6months old); and the evenness of age classes (e). The proportion of males, based onbody-size considerations, is given for the sites in this study and approximated values(based on osteometrics) are given for the pre-Natufian sites (see discussion in Bar-Oz,2004). The dashed line separates the pre-Natufian and Natufian samples.

given the dissimilarities evident in the Natufian sequence of thesite: the dramatic decrease in the thickness of habitation sedi-ments, the disappearance of stone architecture, the dwindling ofbone ornaments and the higher accumulation rates of naturally-deposited micromammals (Yeshurun, 2011; Weinstein-Evronet al., 2012; Weissbrod et al., 2012). This discrepancy may beexplained when the entire Mt. Carmel settlement system isconsidered (see below).

The general depositional similarity through time in EWT isreinforced when compared with several conspicuously differenttypes of Natufian accumulations. The first example comes from thesame site, the EN fauna from Chamber III of el-Wad Cave (EWC;Weinstein-Evron, 1998; see Fig. 1B for location). The EWC assem-blage exhibits a medium big game index almost reaching pre-Natufian proportions and a very high fast small game index (Stutzet al., 2009), both very high compared with the EN terrace fauna.On the basis of the location of this excavation area in the dark partof the cave, the rarity of built features, lack of burials and compo-sition of the lithic assemblage, it was suggested that it representeda specialized ‘dumping area’ and not a domestic or a burial locality(Weinstein-Evron, 1998). Supporting evidence for the non-domestic use of this locality may be found in the scanty evidencefor in situ fire activity; the proportion of burned bone specimens isvery low (4% of NISP; Rabinovich, 1998). These characters arereminiscent of the ‘occasional-toss area’ of Locus 25 in the ENterrace. Moving to the LN, Raqefet Cave (Mount Carmel, ca. 10 kmeast of el-Wad) that was used primarily for burial and othercommunal activities (Nadel et al., 2012, 2013) yielded a faunalassemblage weakly affected by post-discard damage (notably, lowbone-burning intensity and frequency), as well as a relatively highvalue for the medium big game index, clustering with Locus 25 andthe EWC samples in some attributes (Yeshurun et al., 2013a).Similarly, the LN fauna in the burial cave of Hilazon Tachtit (Galilee,Israel) yielded a low proportion of burned bones, almost identicalto the EWC and Raqefet cases (Grosman and Munro, 2007; Munroand Grosman, 2010). These three non-domestic examples contrastwith the intensively-trampled and unintentionally-burned EN andLN faunas in the current study. The LN fauna sampled here clustersbetter with LN habitation sites such as Hayonim Terrace (Munro,2012) and ‘Eynan (Bridault et al., 2008), which also show exten-sive post-discard damage and, unlike Late Natufian EWT, preservearchitectural features.

By the same token, the pre-Natufian assemblages in the Israelicoastal plain Epipaleolithic series, namely, Nahal Hadera V, Hefzi-bah and Neve David, seem to represent spatially and verticallyextensive amalgamations of human occupations attesting tovarious domestic (and other) activities and preserving ‘livingfloors’, a hut-like feature (in NHV), spatially defined concentrationsof finds and diverse chipped lithics and groundstone assemblages

Figure 10. Comparison of measurements among the coastal plain Epipaleolithic sites: (A) gazelle humerus BT; (B) hare humerus Bd; and (C) hare scapula-glenoid fossa (GLP). TheGeometric Kebaran assemblages were pooled due to sample size limitations. See SOM Table A.8 for statistics.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3530

(Ronen et al., 1975; Saxon et al., 1978; Kaufman and Ronen, 1987;Kaufman, 1989, 1992; Barkai and Gopher, 2001; Bar-Oz, 2004;Bocquentin et al., 2011). The general similarity in the function ofhabitation through time in our case-study enables a more robustappreciation of Natufian economic trends, relatively independentof changes in site function.

Economic intensification through time

Economic trends in the Natufian sequence of EWT were put in abroader perspective, using the geographically distinctive Epi-paleolithic series of the coastal plain (Bar-Oz, 2004) and a behav-ioral ecology approach (Stiner andMunro, 2002; Munro, 2004). Theresults attest to regional-scale intensification of game resourceexploitation in the Natufian, relative to the pre-Natufian Epi-paleolithic of the same region. Starting with the EN and lastingthrough the LN, ungulates dwindle in numbers and small animalscomprise amuch larger portion of Natufian prey. The ungulates thatwere still routinely captured were overwhelmingly small ungulates(i.e., gazelles) as opposed to a significant input of larger ungulates(i.e., fallow deer) in the preceding Epipaleolithic samples. No dif-ference was observed in the EN versus LN phases of EWT in theserespects. On the more local (site) scale, fast small game is some-what more abundant in the EN phase, possibly signaling strongerhuman prey pressure in the camp’s catchment area (Stiner et al.,1999, 2000; Munro, 2004).

The observed prey trends were previously identified in thecoastal plain sequence (Bar-Oz, 2004) as well as in the Kinneret-Golan region (Davis et al., 1988; Davis, 2005), the Wadi Meged se-ries in the lower Galilee (Munro, 2004; Stiner, 2005) and theDamascus Basin (Napierala, 2011). However, only the Wadi Megedseries had previously included EN samples in the analysis, yieldingsignificant differences in small game procurement between the ENand LN (Munro, 2004). Other inter-regional syntheses (Munro,2009; Stutz et al., 2009) argued for a gradual decline of mediumand large big game and an increase of small game species over time,between the Early Epipaleolithic and the EN and LN of the Medi-terranean zone. While the present results generally support theseassertions, it is important to note that no ‘gradual’ trend isapparent. Rather, the shift in ungulate proportions relative to smallgame, and the near-disappearance of medium and large ungulates,occur abruptly in the EN and remain largely unchanged in the LN.Additionally, small game indices (fast versus slow) display the samemarked ‘step’ in the pre-Natufian e EN transition, exhibitingnegligible index values in the Kebaran and Geometric Kebaransamples as opposed to high values in either the EN or LN. Thestriking difference between the pre-Natufian and Natufian assem-blages overshadows any minor variations within these two time-

periods, including the modest decline in fast small game abun-dance in the LN.

The importance of small animals in human diet during the lateEpipaleolithic reaffirms the true broad spectrum nature of Natufiananimal economy. Such common exploitation of tortoises, hares,foxes and possibly squamates, which comprise around 50% of theidentified specimens, along withmole-rats, Mediterranean fish andedible mollusks (Yeshurun et al., 2009; Bar-Yosef Mayer and Zohar,2010; Weissbrod et al., 2012) was rarely observed in the pre-Natufian Epipaleolithic and Upper Paleolithic record of northernIsrael (e.g., Rabinovich, 2003; Bar-Oz, 2004; Stiner, 2005; Maromand Bar-Oz, 2008), let alone the more arid areas of the southernLevant (e.g., Martin et al., 2010). The lakeside settlement of Ohalo IIis currently the only example of pre-Natufian fishing (of freshwaterfish) and one of the rare cases of fowling (Simmons and Nadel,1998; Zohar, 2003). However, pre-Natufian sites that are notlocated in a lacustrine setting have not yielded any significant ev-idence of fishing, in contrast with the presence of Mediterraneanfish in several Natufian inland sites (Davis et al., 1994; Bar-YosefMayer and Zohar, 2010). Importantly, four families of Mediterra-nean fishes were identified at EWT (Valla et al., 1986), even thoughthe site was 8e12 km away from the sea shore. Birds are sporadi-cally present in the pre-Natufian Epipaleolithic, but can usually beconsidered to reflect specialized exploitation for raw materials, notregular and long-term dietary contributions (Martin et al., 2013).The extremely broad spectrum animal diet of the Natufian lendsstrong support to the view of Natufian economy as being intensi-fied. Natufian foragers chose to hunt significant amounts of lower-ranked game, probably in response to population packing in adefined territory (Munro, 2004). Taking a long-term perspectivebeyond the Epipaleolithic, ungulates had been the main target ofhunting in the Levant since the Middle Pleistocene and Natufian-like exploitation of many small game taxa was very uncommonup until this stage (Stiner, 2005; Yeshurun, 2013).

Plant exploitation must have played a very important role inEpipaleolithic subsistence, and intensified plant exploitation iscommonly mentioned as an Epipaleolithic hallmark. While spo-radic charcoal, pollen and phytolith data exist in the Natufian (Lev-Yadun and Weisntein-Evron, 1994; Weinstein-Evron, 1998; Portilloet al., 2010; Rosen, 2010; Power et al., 2014) macrobotanical evi-dence from the southern Levant Epipaleolithic is so scarce (Asoutiand Fuller, 2012: their Table 2) that quantitative analyses cannotrealistically be performed. Therefore, until new archaeobotanicaldata sets are available, the BSR must be measured by animalexploitation alone in this period.

Results from gazelle culling patterns augment the economic‘break’ seen between the EN and its forerunners. The probabletargeting of males during the EN is unique in the Epipaleolithic

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 31

series of the coastal plain, paralleling other Natufian assemblages(including LN ones; Garrard, 1980; Cope, 1991; Bar-Oz et al., 2004).It was suggested that targeting males was a (possibly seasonal)adaptation to ecologically sustainable hunting, attempting tominimize the effects of human predation on the non-migratorymountain gazelle herds in a territory (Cope, 1991).

The pre-Natufian assemblages exhibit age distributions that aresimilar to the year-round average of modern herds and cannotpoint to exploitation of specific age groups. An exception is theavoidance of fawns, presumably due to small meat and fat returns(Munro and Bar-Oz, 2005). The picture changes in the Early Natu-fian, where adult male gazelles were targeted, and fawns weretaken in proportions similar to their natural abundance, in linewiththe BSR hypothesis. The evenness of age groups in both the earlierEpipaleolithic and the EN is moderately high. It is possible that theage group evenness index reflects two different mechanisms in thepre-Natufian and EN assemblages, namely, culling of both youngand older gazelles in the former and averaging of exceedinglynumerous occupations in the latter, when gazelles were depositedduring several seasons (Davis, 1983). Such a scenario may explainthe seemingly contradictory evidence of adult-focused culling inconjunction with hunting some fawns in the EN. The picturechanges again in the LN, demonstrating an unselective culling ofsexes, juveniles and fawns. The low evenness of age groups heremay stem from the small sample in comparison to the other as-semblages and therefore enlarged samples will be needed to verifythe LN culling evenness trend.

Age profiles generally present a complex and inconsistent pic-ture throughout the Natufian assemblages. The proportion of ga-zelle fawns in the EN period ranges between 4 and 40%, with oldersub-adult proportions very variable (Munro, 2009; Edwards andMartin, 2013). High juvenile proportions of 30e50% characterizeseveral, but not all, LN assemblages (Davis, 1983, 2005; Munro,2012; Yeshurun et al., 2013a). Hence, ambiguous patterns areevident in the issue of Natufian gazelle culling, which, although oneof the most researched topics in Natufian archaeology, it is one ofthe least understood. Better interpretative frameworks are espe-cially needed if we are to understand the Natufian preferences forgazelles of certain age classes.

Decreasing mobility and increasing catchment exploitation

The importance of the Natufian economic transformation lies inits implications for inferring terminal Pleistocenemobility patterns.Small animal exploitation has been seen as an important marker ofprolonged site occupation and population growth in the Natufian,because of the need to feed a growing or more permanent popu-lation in a given territory (Tchernov, 1993b; Stiner and Munro,2002; Munro, 2004, 2009; Weissbrod et al., 2012). Thus, thebroad-spectrum economy and the increased culling of immaturegazelles were interpreted as a certain facet of the EN sedentariza-tion phenomenon, which has also been inferred from multiplearchaeological proxies. The notion of EN sedentism is stronglysupported by our findings, as the fauna presented here comes froma major hamlet, where stone architecture is plentiful and thedensity and diversity of finds are very high (Weinstein-Evron et al.,2013). Both the intensified animal economy and the archaeologicalnature of the site contrast sharply with the pre-Natufian Epi-paleolithic evidence from the same region (Bar-Oz, 2004). Thus,keeping the site function, site catchment and analytical methods asconstant as possible convincingly demonstrated the economicintensification marking the emergence of sedentary life in the EN,ca. 15e14 ka in the Israeli coastal plain.

In conjunctionwith decreasing mobility, one of the hallmarks ofthe Natufian economy, the regular capture of small mammals, may

reflect intensified site catchment exploitation in this period. Whilegazelle hunting was probably performed by shooting withmicrolith-tipped projectiles (Yeshurun and Yaroshevich, 2014),albeit with increasingly efficient designs (Yaroshevich et al., 2010),the regular hunting of small and agile terrestrial animals, primarilyhares and foxes, which are solitary and nocturnal, was likely per-formed by traps and snares (e.g., Winterhalder, 1980; Holliday andChurchill, 2006). Modern studies on poaching of Israeli wildlifeshowed that most trapping methods, such as simple cable snaring,would indiscriminately capture a broad spectrum of small andmedium mammal species (Yom-Tov, 2003). Thus, snaring wouldhave constituted an efficient and predictable hunting method forthe systematic execution of the broad-spectrum animal economy ina defined Natufian territory. Setting, maintaining and keeping trackof a system of traps and snares requires a long presence of humansin a given territory and excellent acquaintance with the landscape,the animals and their local patterns of foraging, and thus may berelated in the ethnographic record to sedentary or semi-sedentarysocieties (Holliday, 1998; Holliday and Churchill, 2006).

The inferred patterns of LN sedentism and site-catchmentexploitation do not differ much from EN patterns in this study.The LN period in the Mediterranean southern Levant was some-times viewed as one of cultural decline and a return to a moremobile way of life, comparedwith the Early Natufian (Garrod, 1957;Henry, 1991; Valla, 1995; Belfer-Cohen and Bar-Yosef, 2000; Bar-Yosef and Belfer-Cohen, 2002; Grosman, 2003; Munro, 2004).This was mainly due to the more impoverished and thinner natureof LN layers in the intensively-researched sites of el-Wad, HayonimCave and ‘Eynan, where architecture, elaborate burials and art ob-jects were considered to be less frequent. However, it has becomeincreasingly clear that stone-built dwellings and other architecturalfeatures are common at some Late/Final Natufian sites, as apparentfrom the new fieldwork at ‘Eynan (Valla et al., 2007), Huzuk Musa(Rosenberg et al., 2010) and Nahal Ein Gev II (L. Grosman, Personalcommunication), as well as other sites in the more arid regions(Henry, 1976; Goring-Morris, 1991; Conard et al., 2013; Richteret al., 2013). Elaborate LN funerary customs are now known inHayonim Terrace (Tchernov and Valla, 1997), Hilazon Tachtit Cave(Grosman et al., 2008;Munro and Grosman, 2010) and Raqefet Cave(Nadel et al., 2012, 2013; Yeshurun et al., 2013a). While at EWT thenature of the LN layer is indeed impoverished compared with theEN, it was taphonomically demonstrated here that the site was stillused for diverse domestic activities (as well as for burial, which isspatially and perhaps temporally segregated from the habitationsediments; Weinstein-Evron, 2009) and was intensively andrecurrently inhabited. While the taphonomic markers of site-occupation intensity indicate that the LN habitation were some-what less intensive compared with the EN, the economic indicatorspertaining to site-occupation intensity do not differ much from theEN phase, especially when the pre-Natufian sites are used asreference. Furthermore, just 5 km north of el-Wad there existed amajor LN base-camp at the site of Nahal Oren, displaying anarchitectural compound (a terrace wall and associated living sur-faces), as well as a spatially distinct cemetery (Stekelis and Yizraely,1963; Noy, 1991) and evidence of occupations throughout all of theLN sub-phases (Grosman et al., 2005), much like EN el-Wad. Thus,the major base-camp of the Carmel Coast may have simply shiftedfrom EN el-Wad to LN Nahal Oren, hardly suggesting any loss ofmobility or cultural complexity in the Mediterranean Levant(Weinstein-Evron, 2009). The presence of both base-camps andspecialized sites may suggest stable (or even growing) LN pop-ulations in the Mount Carmel-Galilee region ca. 13.7e11.7 ka, cor-responding with the evidence for geographic expansion ofsettlements and the emergence of novel technological and socialadaptations in the latter half of the Natufian (e.g., Bar-Yosef, 2002;

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3532

Valla et al., 2007; Grosman et al., 2008; Stutz et al., 2009;Weinstein-Evron, 2009; Rosen, 2010; Valla, 2012; Nadel et al.,2013).

Conclusions

We evaluated terminal Pleistocene subsistence change in theLevantine Epipaleolithic, when mobile foragers were settling downin the course of the earliest foraging-to-farming transition.Archaeologically, the salient features of the late EpipaleolithicNatufian Culture at key sites such as el-Wad (the regular appear-ance of stone architecture, cemeteries, groundstone and art) standout, suggesting remarkable socioeconomic changes. Early Natufianel-Wad Terrace yielded novel paleoeconomic signals comparedwith the pre-Natufian Epipaleolithic of the same region, accom-panied by the preponderance of stone-constructed dwellings, highdensity of finds, and heavy taphonomic damage from repeated andintensive occupations. The Late Natufian sample differs in gazelleculling patterns but nonetheless represents enduring economicintensification, coupled with relatively reduced, but still consider-able post-discard taphonomic damage from repeated occupations.The LN occupations may have been somewhat less intensive thanthe EN occupations in the particular case of el-Wad, but still leftbehind habitation deposits of an important hamlet. The use ofcontextual taphonomy, focusing on post-discard damage in Epi-paleolithic camps, carries a somewhat unexplored potential andshould be implemented in forthcoming studies to distinguish de-grees in the intensity of accumulation and repetition of habitations.

It is evident from the Epipaleolithic series of the Israeli coastalplain that the Early Natufian constituted an economic break,reflecting decreasing mobility and increasing exploitation of thesite catchment. While the roots of the Early Natufian can surely betraced back to the preceding cultures, the economic and socialtransformation it reflects should not be underestimated. Concur-rently, the EN-LN differences should not be overemphasized. TheNatufian economic transformation may be mentioned as a genuinestage in the evolution of human diet in the latest PleistoceneLevant, characterized by obtaining a broader spectrum of animalresources and reflecting the rise of sedentism in this period. TheNatufian strategy, commencing in the EN and maintained in the LN,was successfully employed for several millennia until it was grad-ually replaced by food production during the Pre-Pottery NeolithicPeriod.

Acknowledgments

This paper is based on R.Y.’s doctoral research at the Universityof Haifa, generously funded by the Graduate Studies Authority, theHecht Scholarship, the Wolf Foundation Scholarship and the Car-mel Research Center Grant. The manuscript was written during hisFulbright post-doctoral fellowship in the Smithsonian Institution.We thank D. Kaufman for his help throughout this research and forcommenting on a previous draft and A. Regev for graphic assis-tance. The paper greatly benefitted from the helpful commentsand suggestions by the editors and two anonymous reviewers. Therenewed excavation at el-Wad Terrace is sponsored by theWenner-Gren Foundation, the Care Foundation and the Faculty ofHumanities, University of Haifa. Thanks are also due to the DanDavid Foundation and to Sarah and Avie Arenson for their support.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jhevol.2014.02.011.

References

Amitai, P., Bouskila, A., 2003. Handbook of Amphibians and Reptiles in Israel. Keter,Jerusalem (Hebrew).

Asouti, E., Fuller, D.Q., 2012. From foraging to farming in the southern Levant: Thedevelopment of Epipalaeolithic and Pre-Pottery Neolithic plant managementstrategies. Veget. Hist. Archaeobot. 21, 149e162.

Asouti, E., Fuller, D.Q., 2013. A contextual approach to the emergence of agriculturein Southwest Asia: Reconstructing Early Neolithic plant-food production. Curr.Anthropol. 54, 299e345.

Atici, L., 2009. Specialization and diversification: Animal exploitation strategies inthe terminal Pleistocene Mediterranean Turkey. Before Farming 3, 136e152.

Bachrach, N., Hershkovitz, I., Kaufman, D., Weinstein-Evron, M., 2013. The lastNatufian inhabitants of el-Wad Terrace. In: Bar-Yosef, O., Valla, F.R. (Eds.),Natufian Foragers in the Levant: Terminal Pleistocene Social Changes inWestern Asia. International Monographs in Prehistory, Ann Arbor, pp. 107e117.

Baharav, D., 1974. Notes on the population structure and biomass of the mountaingazelle, Gazella gazella gazella. Israel J. Zool. 23, 39e44.

Baharav, D., 1983. Observation on the ecology of the mountain gazelle in the UpperGalilee, Israel. Mammalia 47, 59e69.

Bar-El, T., Tchernov, E., 2000. Lagomorph remains at prehistoric sites in Israel andsouthern Sinai. Paléorient 26, 93e109.

Bar-Oz, G., 2004. Epipaleolithic Subsistence Strategies in the Levant: A Zooarch-aeological Perspective. In: ASPR Monograph Series. Brill, Boston.

Bar-Oz, G., Dayan, T., 2002. “After 20 Years”: a taphonomic re-evaluation of NahalHadera V, an Epipalaeolithic site on the Israeli coastal plain. J. Archaeol. Sci. 29,145e156.

Bar-Oz, G., Dayan, T., 2003. Testing the use of multivariate inter-site taphonomiccomparisons: the faunal analysis of Heftziba in its Epipalaeolithic culturalcontext. J. Archaeol. Sci. 30, 885e900.

Bar-Oz, G., Dayan, T., Kaufman, D., 1999. The Epipalaeolithic faunal sequence inIsrael: a view from Neve-David. J. Archaeol. Sci. 26, 67e82.

Bar-Oz, G., Dayan, T., Kaufman, D., Weinstein-Evron, M., 2004. The Natufian econ-omy at el-Wad Terrace with special reference to gazelle exploitation patterns.J. Archaeol. Sci. 31, 217e231.

Bar-Yosef, O., 1998. The Natufian culture in the Levant, threshold to the origins ofagriculture. Evol. Anthropol. 6, 159e177.

Bar-Yosef, O., 2002. Natufian: A complex society of foragers. In: Fitzhugh, B., Habu, J.(Eds.), Beyond Foraging and Collecting: Evolutionary Change in Hunter-Gatherer Settlement Systems. Kluwer Academic/Plenum Publishers, NewYork, pp. 91e147.

Bar-Yosef, O., Belfer-Cohen, A., 2002. Facing environmental crisis: Societal andcultural changes at the transition from the Younger Dryas to the Holocene inthe Levant. In: Cappers, R.T.J., Bottema, S. (Eds.), The Dawn of Farming in theNear East, Studies in Early Near Eastern Production, Subsistence, and Envi-ronment. Ex oriente, Berlin, pp. 55e66.

Bar-Yosef, O., Meadow, R.H., 1995. The origins of agriculture in the Near East. In:Price, T.D., Gebauer, A.B. (Eds.), Last Hunters-First Farmers: New Perspectives onthe Prehistoric Transition to Agriculture. School of American Research Press,Santa Fe, pp. 39e94.

Bar-Yosef Mayer, D.E., Zohar, I., 2010. The role of aquatic resources in the NatufianCulture. Euras. Prehist. 7, 29e43.

Barkai, R., Gopher, A., 2001. Nahal Hadera V. Excavat. Surv. Israel 113, 35.Behrensmeyer, A.K., 1978. Taphonomic and ecological information from bone

weathering. Paleobiology 4, 150e162.Behrensmeyer, A.K., Gordon, K.D., Yanagi, G.T., 1986. Trampling as a cause of bone

surface damage and pseudo-cutmarks. Nature 319, 768e771.Belfer-Cohen, A., Bar-Yosef, O., 2000. Early sedentism in the Near East: A bumpy ride

to village life. In: Kuijt, I. (Ed.), Life in Neolithic Farming Communities: SocialOrganization, Identity and Differentiation. Plenum Publishers, New York,pp. 19e37.

Belfer-Cohen, A., Goring-Morris, A.N., 2011. Becoming farmers. Curr. Anthropol. 52(S4), S209eS220.

Binford, L.R., 1981. Bones: Ancient Men and Modern Myths. Academic Press, NewYork.

Blumenschine, R.G., Selvaggio, M.M., 1988. Percussion marks on bone surfaces as anew diagnostic of hominid behavior. Nature 333, 763e765.

Blumenschine, R.G., Marean, C.W., Capaldo, S.D., 1996. Blind tests of inter-analystcorrespondence and accuracy in the identification of cut marks, percussionmarks, and carnivore toothmarks on bone surfaces. J. Archaeol. Sci. 23, 493e507.

Bocquentin, F., Crevecoeur, I., Arensburg, B., Kaufman, D., Ronen, A., 2011. Leshommes du Kébarien géométrique de Neve David, Mont Carmel (Israël). Bull.Mém. Soc. Anthropol. Paris 23, 38e51.

Brain, C.K., 1981. The Hunters or the Hunted? An Introduction to African CaveTaphonomy. University of Chicago Press, Chicago.

Bridault, A., Rabinovich, R., Simmons, T., 2008. Human activities, site location andtaphonomic process: a relevant combination for understanding the fauna ofEynan (Ain Mallaha), level Ib (final Natufian, Israel). In: Vila, E., Gourichon, L.,Choyke, A., Buitenhuis, H. (Eds.), Proceedings of the 8th International Meetingof the ASWA. Maison de l’Orient et de la Méditerranée, Lyon, pp. 99e117.

Bunn, H.T., 1981. Archaeological evidence for meat-eating by Plio-Pleistocenehominids from Koobi Fora and Olduvai Gorge. Nature 291, 574e577.

Byrd, B.F., 2005. Reassessing the emergence of village life in the Near East.J. Archaeol. Res. 13, 231e290.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 33

Cain, C.R., 2005. Using burned animal bone to look at Middle Stone Age occupationand behavior. J. Archaeol. Sci. 32, 873e884.

Campana, D.V., 1989. Natufian and Protoneolithic Bone Tools: The Manufacture andUse of Bone Implements in the Zagros and the Levant. In: British ArchaeologicalReports International Series 494, Oxford.

Capaldo, S.D., Blumenschine, R.J., 1994. A quantitative diagnosis of notches made byhammerstone percussion and carnivore gnawing on bovid long bones. Am.Antiq. 59, 724e748.

Conard, N.J., Bretzke, K., Deckers, K., Kandel, A.W., Masri, M., Napierala, H., Riehl, S.,Stahlschmidt, M., 2013. Natufian Lifeways in the Eastern Foothills of the Anti-Lebanon Mountains. In: Bar-Yosef, O., Valla, F.R. (Eds.), Natufian Foragers inthe Levant: Terminal Pleistocene Social Changes in Western Asia. InternationalMonographs in Prehistory, Ann Arbor, pp. 1e16.

Cope, C., 1991. Gazelle hunting strategies in the Southern Levant. In: Bar-Yosef, O.,Valla, F.R. (Eds.), The Natufian Culture in the Levant, Archaeological Series 1.International Monographs in Prehistory, Ann Arbor, pp. 341e358.

Costamagno, L., Théry-Parisot, I., Brugal, J.-Ph., Guibert, R., 2005. Taphonomicconsequences of the use of bones as fuel. Experimental data and archaeologicalapplications. In: O’Connor, T. (Ed.), Biosphere to Lithosphere: New Studies inVertebrate Taphonomy. Oxbow Books, Oxford, pp. 51e62.

Davis, S.J.M., 1981. The effect of temperature change and domestication on the bodysize of late Pleistocene to Holocene mammals of Israel. Paleobiology 7, 101e114.

Davis, S.J.M., 1983. The age profiles of gazelles predated by ancient man in Israel:possible evidence for a shift from seasonality to sedentism in the Natufian.Paléorient 9, 55e62.

Davis, S.J.M., 1991. When and why did prehistoric people domesticate animals?Some evidence from Israel and Cyprus. In: Bar-Yosef, O., Valla, F.R. (Eds.), TheNatufian Culture in the Levant. International Monographs in Prehistory, AnnArbor, pp. 381e390.

Davis, S.J.M., 2005. Why domesticate food animals? Some zoo-archaeological evi-dence from the Levant. J. Archaeol. Sci. 32, 1408e1416.

Davis, S., Rabinovich, R., Goren-Inbar, N., 1988. Quaternary extinctions and popu-lation increase in Western Asia: the animal remains from Biq’at Quneitra.Paléorient 14, 95e105.

Davis, S.J.M., Lernau, O., Pichon, J., 1994. The animal remains: new light on the originof animal husbandry. In: Lechevalier, M., Ronen, A. (Eds.), Le Gisement deHatoula en Judée Occidentale, Israël. Association Paléorient, Paris, pp. 83e100.

de Juana, S., Galán, A.B., Domínguez-Rodrigo, M., 2010. Taphonomic identification ofcut marks made with lithic handaxes: an experimental study. J. Archaeol. Sci. 7,1841e1850.

Domínguez-Rodrigo, M., Barba, R., 2006. New estimates of tooth mark and per-cussion mark frequencies at the FLK Zinj site: the carnivore e hominid e

carnivore hypothesis falsified. J. Hum. Evol. 50, 170e194.Domínguez-Rodrigo, M., de Juana, S., Galán, A.B., Rodríguez, M., 2009. A new pro-

tocol to differentiate trampling marks from butchery cut marks. J. Archaeol. Sci.36, 2643e2654.

Ducos, P., Horwitz, L.K., 1997. The influence of climate on artiodactyls size during thelate Pleistocene-Early Holocene of the southern Levant. Paléorient 23, 229e247.

Eckmeier, E., Yeshurun, R., Weinstein-Evron, M., Mintz, E., Boaretto, E., 2012.Radiocarbon dating of the Early Natufian at el-Wad Terrace, Mount Carmel,Israel: methodology and materials characterization. Radiocarbon 54, 823e836.

Edwards, P.C., 1989. Revisiting the broad spectrum revolution: its role in the originsof Southwest Asian food production. Antiquity 63, 225e246.

Edwards, Y.H., Martin, L., 2013. Animal bones and archaeozoological analysis. In:Edwards, P.C. (Ed.), Wadi Hammeh 27, an Early Natufian Settlement at Pella inJordan. Brill, Leiden, pp. 321e352.

Everitt, B.S., 1977. The Analysis of Contingency Tables. Wiley and Sons, New York.Faith, J.T., Gordon, A.D., 2007. Skeletal element abundances in archaeofaunal as-

semblages: economic utility, sample size, and assessment of carcass transportstrategies. J. Archaeol. Sci. 34, 872e882.

Flannery, K.V., 1969. Origins and ecological effects of early domestication in Iran andthe Near East. In: Ucko, P.J., Dimbleby, G.W. (Eds.), The Domestication andExploitation of Plants and Animals. Atherton, Chicago, pp. 73e100.

Garrard, A.N., 1980. Man-animal-plant relationships during the Upper Pleistoceneand Early Holocene. Ph.D. Dissertation, University of Cambridge.

Garrod, D.A.E., 1932. New Mesolithic industry: the Natufian of Palestine. J.R.Anthropol. Inst. 62, 257e269.

Garrod, D.A.E., 1957. The Natufian culture: the life and economy of a Mesolithicpeople in the Near East. Proc. Brit. Acad. 43, 211e227.

Garrod, D.A.E., Bate, D.M.A., 1937. The Stone Age of Mount Carmel. In: Excavations atthe Wadi Mughara, vol. I. Clarendon Press, Oxford.

Gaudzinski-Windheuser, S., Kindler, L., Rabinovich, R., Goren-Inbar, N., 2010. Testingheterogeneity in faunal assemblages from archaeological sites. Tumbling andtrampling experiments at the Early-Middle Pleistocene site of Gesher BenotYa’aqov (Israel). J. Archaeol. Sci. 37, 3170e3190.

Goring-Morris, N., 1991. The Harifian of the Southern Levant. In: Bar-Yosef, O.,Valla, F.R. (Eds.), The Natufian Culture in the Levant. International Monographsin Prehistory, Ann Arbor, pp. 173e216.

Goring-Morris, A.N., Belfer-Cohen, A., 2008. A roof over one’s head: developmentsin Near Eastern residential architecture across the Epipalaeolithic-Neolithictransition. In: Bocquet-Appel, J.-P., Bar-Yosef, O. (Eds.), The Neolithic de-mographic transition and its consequences. Springer, Dordrecht, pp. 239e286.

Grayson, D.K., Delpech, F., 2008. The large mammals of Roc de Combe (Lot, France):the Châtelperronian and Aurignacian assemblages. J. Anthropol. Archaeol. 27,338e362.

Grosman, L., 2003. Preserving cultural traditions in a period of instability: The LateNatufian of the Hilly Mediterranean Zone. Curr. Anthropol. 44, 571e580.

Grosman, L., Munro, N.D., 2007. The sacred and the mundane: domestic activities ata Late Natufian burial site in the Levant. Before Farming 4, 1e14.

Grosman, L., Ashkenazi, H., Belfer-Cohen, A., 2005. The Natufian occupation ofNahal Oren, Mt. Carmel, Israel e the lithic evidence. Paléorient 31, 5e26.

Grosman, L., Munro, N.D., Belfer-Cohen, A., 2008. A 12,000-year-old Shaman burialfrom the southern Levant (Israel). Proc. Natl. Acad. Sci. 105, 17665e17669.

Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: Paleontological Statistics soft-ware package for education and data analysis. Palaeont. Electron. 4, 9.

Hardy-Smith, T., Edwards, P.C., 2004. The garbage crisis in prehistory:Artifact discard patterns at the Early Natufian site of Wadi Hammeh 27 andthe origins of household refuse disposal strategies. J. Anthropol. Archaeol. 23,253e289.

Haynes, G., 1983. Frequencies of spiral and green-bone fractures on ungulate limbbones in modern surface assemblages. Am. Antiq. 48, 102e114.

Henry, D.O., 1976. Rosh Zin: A Natufian settlement near Avdat. In: Marks, A.E. (Ed.),Prehistory and Palaeoenvironments in the Central Negev, Israel. SouthernMethodist University Press, Dallas, pp. 317e347.

Henry, D.O., 1991. Foraging, sedentism, and adaptive vigor in the Natufian:rethinking the linkages. In: Clark, G.A. (Ed.), Perspectives on the Past: Theo-retical Biases in Mediterranean Hunter-Gatherer Research. University of Penn-sylvania, Philadelphia, pp. 353e370.

Henry, D.O., 1995. Prehistoric Cultural Ecology and Evolution: Insights fromSouthern Jordan. Plenum Press, New York.

Hockett, B., Haws, J.A., 2002. Taphonomic and methodological perspectives ofleporid hunting during the Upper Paleolithic of the western MediterraneanBasin. J. Archaeol. Method Th. 9, 269e302.

Holliday, T.W., 1998. The ecological context of trapping among recent hunter-gatherers: Implications for subsistence in terminal Pleistocene Europe. Curr.Anthropol. 39, 711e720.

Holliday, T.W., Churchill, S.E., 2006. Mustelid hunting by recent foragers and thedetection of trapping in the European Paleolithic. In: Haws, J.A., Hockett, B.S.,Brugal, J.-P. (Eds.), Paleolithic Zooarchaeology in Practice, British ArchaeologicalReports International Series 1564, Oxford, pp. 45e56.

Kaufman, D., 1989. Observations on the Geometric Kebaran: A view from NeveDavid. In: Bar-Yosef, O., Vandermeersch, B. (Eds.), Investigations in SouthLevantine Prehistory, British Archaeological Reports International Series 497,Oxford, pp. 275e285.

Kaufman, D., 1992. Hunter-gatherers of the Levantine Epipaleolithic: The socio-ecological origins of sedentism. J. Mediterr. Archaeol. 5, 165e201.

Kaufman, D., Ronen, A., 1987. La sépulture kébarienne géométrique de Neve David,Haïfa, Israël. L’Anthropologie 91, 335e342.

Kent, S., 1993. Variability in faunal assemblages: The influence of hunting skill,sharing, dogs, and mode of cooking on faunal remains at a sedentary Kalaharicommunity. J. Anthropol. Archaeol. 12, 323e385.

Krebs, J.R., 1989. Ecological Methodology. Harper and Collins, New York.Lev-Yadun, S., Weisntein-Evron, M., 1994. Late Epipalaeolithic wood remains from

el-Wad Cave, Mount Carmel, Israel. New Phytologist 127, 391e396.Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge University Press, Cambridge.Lyman, R.L., 2008. Quantitative Paleozoology. Cambridge University Press,

Cambridge.Magdwick, R., Mulville, J., 2012. Investigating variation in the prevalence of

weathering in faunal assemblages in the UK: a multivariate statistical approach.Int. J. Osteoarchaeol. 22, 509e522.

Maher, L.A., Richter, T., Stock, J.T., 2012a. The pre-Natufian Epipaleolithic: Long-termbehavioral trends in the Levant. Evol. Anthropol. 21, 69e81.

Maher, L.A., Richter, T., Macdonald, D., Jones, M.D., Martin, L., Stock, J.T., 2012b.Twenty thousand-year-old huts at a hunter-gatherer settlement in easternJordan. PLoS One 7, e31447.

Marom, N., Bar-Oz, G., 2008. Measure for measure: a taphonomic reconsideration ofthe Kebaran site of Ein Gev 1, Israel. J. Arch. Sci. 35, 214e227.

Martin, L., Edwards, Y., Garrard, A., 2010. Hunting practices at an eastern JordanianEpipalaeolithic aggregation site: The case of Kharaneh IV. Levant 42, 107e135.

Martin, L.A., Edwards, Y., Garrard, A., 2013. Broad spectrum or specialised activity:Birds and tortoises in the Levantine Epipalaeolithicethe case of Wadi Jilat 22.Antiquity 87, 649e665.

Munro, N.D., 2001. A prelude to agriculture: Game use and occupation intensityduring the Natufian period in the Southern Levant. Ph.D. Dissertation, Univer-sity of Arizona.

Munro, N.D., 2004. Zooarchaeological measures of hunting pressure and occupationintensity in the Natufian: implications for agricultural origins. Curr. Anthropol.45, S5eS33.

Munro, N.D., 2009. Epipalaeolithic subsistence intensification in the southernLevant: the faunal evidence. In: Richards, M.J., Hublin, J.-J. (Eds.), Evolution ofHominid Diets: Integrating Approaches to the Study of Paleolithic Subsistence.Springer, Dordrecht, pp. 141e155.

Munro, N.D., 2012. The Natufian faunal assemblage from Hayonim Terrace. In:Valla, F.R. (Ed.), Les Fouilles de la Terrasse d’Hayonim, Israël 1980e1981 et1985e1989. De Boccard, Paris, pp. 321e348.

Munro, N.D., Atici, L., 2009. Human subsistence change in the Late PleistoceneMediterranean Basin: the status of research on faunal intensification, diversi-fication and specialization. Before Farming, 2009/1 article 1.

Munro, N.D., Bar-Oz, G., 2005. Gazelle bone fat processing in the Levantine Epi-palaeolithic. J. Archaeol. Sci. 32, 223e239.

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e3534

Munro, N.D., Grosman, L., 2010. Early evidence (ca. 12,000 B.P.) for feasting at aburial cave in Israel. Proc. Natl. Acad. Sci. 107, 15362e15366.

Munro, N.D., Bar-Oz, G., Stutz, A.J., 2009. Aging mountain gazelle (Gazella gazella):Refining methods of tooth eruption and wear and bone fusion. J. Archaeol. Sci.36, 752e763.

Munro, N.D., Bar-Oz, G., Hill, A.C., 2011. An exploration of character traits and linearmeasurements for sexing mountain gazelle (Gazella gazella) skeletons.J. Archaeol. Sci. 38, 1253e1265.

Nadel, D., Weiss, E., Simchoni, O., Tsatskin, A., Danin, A., Kislev, M., 2004. Stone Agehut in Israel yields world’s oldest evidence of bedding. Proc. Natl. Acad. Sci. 101,6821e6826.

Nadel, D., Lambert, A., Bosset, G., Bocquentin, F., Rosenberg, D., Yeshurun, R.,Weissbrod, L., Tsatskin, A., Bachrach, N., Bar-Matthews, M., Ayalon, A.,Zaidner, Y., Beeri, R., Grinberg, H., 2012. The 2010 and 2011 seasons of exca-vation at Raqefet Cave. J. Isr. Prehist. Soc. 42, 35e73.

Nadel, D., Danin, A., Power, R.C., Rosen, A.M., Bocquentin, F., Tsatskin, A.,Rosenberg, D., Yeshurun, R., Weissbrod, L., Rebollo, N., Barzilai, O., Boaretto, E.,2013. Earliest floral grave lining from 13,700e11,700-y-old Natufian burials atRaqefet Cave, Mt. Carmel, Israel. Proc. Natl. Acad. Sci. 110, 11774e11778.

Napierala, H., 2011. The Paleolithic background of early food producing societies inthe fertile crescent e faunal analysis. Ph.D. Dissertation, Universität Tübingen.

Neeley, M.P., Clark, G.A., 1993. The human food niche in the Levant over the past150,000 years. In: Peterkin, G.L., Bricker, H., Mellars, P. (Eds.), Hunting and Ani-mal Exploitation in the Later Palaeolithic and Mesolithic of Eurasia, Archaeo-logical Papers of the American Anthropological Association, vol. 4, pp. 221e240.

Noy, T., 1991. Art and decoration of the Natufian at Nahal Oren. In: Bar-Yosef, O.,Valla, F.R. (Eds.), The Natufian Culture in the Levant, Archaeological Series 1.International Monographs in Prehistory, Ann Arbor, pp. 557e568.

Pichon, J., 1991. Les oiseaux au Natoufien: avifaune et sédentarité. In: Bar-Yosef, O.,Valla, F. (Eds.), The Natufian Culture in the Levant. International Monographs inPrehistory, Ann Arbor, pp. 371e380.

Pickering, T.R., Egeland, C.P., 2006. Experimental patterns of hammerstone per-cussion damage on bones: implications for inferences of carcass processing byhumans. J. Archaeol. Sci. 33, 459e469.

Portillo, M., Rosen, A.M., Weinstein-Evron, M., 2010. Natufian plant uses at el-Wadterrace (Mount Carmel, Israel): The phytolith evidence. Euras. Prehist. 7, 99e112.

Power, R.C., Rosen, A.M., Nadel, D., 2014. The economic and ritual utilization ofplants at the Raqefet Cave Natufian site: The evidence from phytoliths.J. Anthropol. Archaeol. 33, 49e65.

Rabinovich, R., 1998. Taphonomical aspects of the recent excavations at El-Wad(appendix III). In: Weinstein-Evron, M. (Ed.), Early Natufian el-Wad revisited.ERAUL 77, Liege, pp. 199e224.

Rabinovich, R., 2003. The Levantine Upper Palaeolithic faunal record. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More Than Meets The Eye: Studies onUpper Palaeolithic Diversity in the Near East. Oxbow, Oxford, pp. 33e48.

Richter, T., Garrard, A.N., Allock, S., Maher, L.A., 2011. Interaction before agriculture:Exchanging material and sharing knowledge in the final Pleistocene Levant.Cambridge Archaeol. J. 21, 95e114.

Richter, T., Bode, L., House, M., Iversen, R., Arranz, A., Saehle, I., Thaarup, G.,Tvede, M.L., Yeomans, L., 2013. Excavations at the Late Epipalaeolithic site ofShubayqa 1: Preliminary report on the first season. Neo-Lithics 2012 (2), 3e14.

Ronen, A., Kaufman, D., Gophna, R., Bakler, N., Smith, P., Amiel, A., 1975. The Epi-palaeolithic site of Hefziba, central coastal plain of Israel. Quartär 26, 54e72.

Rosen, A.M., 2010. Natufian plant exploitation: Managing risk and stability in anenvironment of change. Euras. Prehist. 7, 117e131.

Rosenberg, M., 1998. Cheating at musical chairs: territoriality and sedentism in anevolutionary context. Curr. Anthropol. 39, 653e681.

Rosenberg, D., Kaufman, D., Yeshurun, R., Weinstein-Evron, M., 2012. The brokenrecord: The Natufian groundstone assemblage from el-Wad Terrace (MountCarmel, Israel) e attributes and their interpretation. Euras. Prehist. 9, 93e128.

Rosenberg, D., Yeshurun, R., Groman-Yaroslavski, I., Winter, H., Zertal, A., Brown-Goodman, R., Nadel, D., 2010. Huzuq Musa e a preliminary report on the testexcavation at a Final Epipalaeolithic/PPNA site in the Jordan Valley. Paléorient36, 189e204.

Saxon, E.C., Martin, G., Bar-Yosef, O., 1978. Nahal Hadera: An open-air site on theIsraeli littoral. Paléorient 4, 253e266.

Shennan, S., 1988. Quantifying Archaeology. Edinburgh University Press, Edinburgh.Shipman, P., Rose, J., 1988. Bone tools: an experimental approach. In: Olsen, S.L.

(Ed.), Scanning Electron Microscopy in Archaeology, British ArchaeologicalReports International Series, Oxford, vol. 291, pp. 303e335.

Simmons, T., Nadel, D., 1998. The avifauna of the Early Epipalaeolithic site of Ohalo II(19,400 years BP), Israel: species diversity, habitat and seasonality. Int. J.Osteoarchaeol. 8, 79e96.

Starkovich, B.M., 2012. Intensification of small game resources at Klissoura Cave 1(Peloponnese, Greece) from the Middle Paleolithic to Mesolithic. Quatern. Int.264, 17e31.

Stekelis, M., Yizraely, T., 1963. Excavations at Nahal Oren: Preliminary report. Isr.Explor. J. 13, 1e12.

Stiner, M.C., 2001. Thirty years on the “Broad Spectrum Revolution” and Paleolithicdemography. Proc. Natl. Acad. Sci. 98, 6993e6996.

Stiner, M.C., 2005. The Faunas of Hayonim Cave (Israel): A 200,000-Year Record ofPaleolithic Diet, Demography and Society. Peabody Museum of Archaeology andEthnology, Cambridge.

Stiner, M.C., Kuhn, S.L., 2006. Changes in the ‘connectedness’ and resilience ofPaleolithic societies in Mediterranean ecosystems. Hum. Ecol. 34, 693e712.

Stiner, M.C., Munro, N.D., 2002. Approached to prehistoric diet breadth, demog-raphy, and prey ranking systems in time and space. J. Archaeol. Method Th. 9,181e214.

Stiner, M.C., Munro, N.D., 2011. On the evolution of diet and landscape during theUpper Paleolithic through Mesolithic at Franchthi cave (Peloponnese, Greece).J. Hum. Evol. 60, 618e636.

Stiner, M.C., Kuhn, S.L., Weiner, S., Bar-Yosef, O., 1995. Differential burning, recrystal-ization, and fragmentation of archaeological bone. J. Archaeol. Sci. 22, 223e237.

Stiner, M.C., Munro, N.D., Surovell, T.A., Tchernov, E., Bar-Yosef, O., 1999. Palaeolithicgrowth pulses evidenced by small animal exploitation. Science 283, 190e194.

Stiner, M.C., Munro, N.D., Surovell, T.A., 2000. The tortoise and the hare: small-gameuse, the broad-spectrum revolution, and Paleolithic demography. Curr.Anthropol. 41, 39e73.

Stutz, A.J., Munro, N.D., Bar-Oz, G., 2009. Increasing the resolution of the BroadSpectrum Revolution in the Southern Levantine Epipaleolithic (19e12 ka).J. Hum. Evol. 56, 294e306.

Tchernov, E., 1993a. The impact of sedentism on animal exploitation in the southernLevant. In: Buitenhuis, H., Clason, A.T. (Eds.), Archaeozoology of the Near-East.Universal Book Services/Dr W. Backhuys, Leiden, pp. 10e26.

Tchernov, E., 1993b. Evolution of complexities, exploitation of the biosphere, andzooarchaeology. Archeozoologia 5, 9e42.

Tchernov, E., 1993c. Exploitation of birds during the Natufian and early Neolithic ofthe Southern Levant. Archaeofauna 2, 121e143.

Tchernov, E., Valla, F.R., 1997. Two new dogs, and other Natufian dogs, from theSouthern Levant. J. Archaeol. Sci. 24, 65e95.

Valla, F.R., 1995. The first settled societies e Natufian (12,500e10,200 BP). In:Levy, T. (Ed.), The Archaeology of Society in the Holy Land. Leicester UniversityPress, London, pp. 169e189.

Valla, F.R. (Ed.), 2012. Les Fouilles de la Terrasse d’Hayonim, Israël 1980e1981 et1985e1989. De Boccard, Paris.

Valla, F.R., Bar-Yosef, O., Smith, P., Tchernov, E., Desse, J., 1986. Un nouveau sondagesur la terrasse d’El Ouad, Israel. Paléorient 12, 21e38.

Valla, F.R., Khalaily, H., Valladas, H., Kaltnecker, E., Bocquentin, F., Cabellos, T., Bar-Yosef Mayer, D.E., 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.,Rosen, A.M., Porat,N., Belfer-Cohen, A., 2007. Les fouilles deAinMallha (Eynan) de2003 à 2005: Quatrième rapport préliminaire. J. Isr. Prehist. Soc. 37, 135e383.

Villa, P., Mahieu, E., 1991. Breakage patterns of human long bones. J. Hum. Evol. 21,27e48.

von den Driesch, A., 1976. A Guide to the Measurement of Animal Bones fromArchaeological Sites. In: Peabody Museum Bulletin 1, Cambridge.

Watkins, T., 2010. Changing people, changing environments: How hunter-gatherersbecame communities that changed the world. In: Finlayson, B., Warren, G.(Eds.), Landscapes in Transition. Oxbow, London, pp. 106e114.

Weinstein-Evron, M., 1998. Early Natufian el-Wad revisited. ERAUL 77, Liège.Weinstein-Evron, M., 2009. Archaeology in the Archives: Unveiling the Natufian

Culture of Mount Carmel. In: ASPR Monograph Series. Brill, Boston.Weinstein-Evron, M., Kaufman, D., Bachrach, N., Bar-Oz, G., Bar-Yosef Mayer, D.E.,

Chaim, S., Druck, D., Groman-Yaroslavski, I., Hershkovitz, I., Liber, N.,Rosenberg, D., Tsatskin, A., Weissbrod, L., 2007. After 70 years: new excavationsat the el-Wad Terrace, Mount Carmel, Israel. J. Isr. Prehist. Soc. 37, 37e134.

Weinstein-Evron, M., Yeshurun, R., Kaufman, D., Boaretto, E., Eckmeier, E., 2012.New 14C dates for the Early Natufian of el-Wad Terrace, Mount Carmel, Israel.Radiocarbon 54, 813e822.

Weinstein-Evron, M., Kaufman, D., Yeshurun, R., 2013. Spatial organization ofNatufian el-Wad through time: Combining the results of past and present ex-cavations. In: Bar-Yosef, O., Valla, F.R. (Eds.), Natufian Foragers in the Levant:Terminal Pleistocene Social Changes in Western Asia. International Monographsin Prehistory, Ann Arbor, pp. 88e106.

Weissbrod, L., Dayan, T., Kaufman, D., Weinstein-Evron, M., 2005. Micromammaltaphonomy of el-Wad Terrace, Mount Carmel, Israel: Distinguishing culturalfrom natural depositional agents in the late Natufian. J. Archaeol. Sci. 32, 1e17.

Weissbrod, L., Bar-Oz, G., Yeshurun, R., Weinstein-Evron, M., 2012. Beyond fast andslow: The mole rat Spalax ehrenbergi (Order Rodentia) as a test case for sub-sistence intensification of complex Natufian foragers in southwest Asia. Qua-tern. Int. 264, 4e16.

Weissbrod, L., Kaufman, D., Nadel, D., Yeshurun, R., Weinstein-Evron, M., 2013.Commensalism: was it truly a Natufian phenomenon? Recent contributionsfrom ethnoarchaeology and ecology. In: Bar-Yosef, O., Valla, F.R. (Eds.), NatufianForagers in the Levant: Terminal Pleistocene Social Changes in Western Asia.International Monographs in Prehistory, Ann Arbor, pp. 699e717.

Winterhalder, B.P., 1980. Canadian fur bearer cycles and Cree-Ojibwa hunting andtrapping practices. Am. Nat. 115, 870e879.

Winterhalder, B., Smith, E.A., 2000. Analyzing adaptive strategies: Human behav-ioral ecology at twenty-five. Evol. Anthropol. 9, 51e72.

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. J. Archaeol. Sci. 37, 368e388.

Yeshurun, R., 2011. The zooarchaeology of Natufian el-Wad Terrace and RaqefetCave, Mount Carmel: Taphonomic and diachronic intra- and inter-site analysis.Ph.D. Dissertation, University of Haifa.

Yeshurun, R., 2013. Middle Paleolithic prey-choice inferred from a natural pitfalltrap: Rantis Cave, Israel. In: Clark, J.L., Speth, J.D. (Eds.), Zooarchaeology and

R. Yeshurun et al. / Journal of Human Evolution 70 (2014) 16e35 35

Modern Human Origins: Human Hunting Behavior During the Later Pleisto-cene, Vertebrate Paleobiology and Paleoanthropology Series. Springer, Dor-drecht, pp. 45e58.

Yeshurun, R., Yaroshevich, A., 2014. Bone projectile injuries and Epipaleolithichunting:Newexperimental and archaeological results. J. Archaeol. Sci. 44, 61e68.

Yeshurun, R., Bar-Oz, G., Weinstein-Evron, M., 2009. The role of foxes in theNatufian economy: A view from Mount Carmel, Israel. Before Farming, 2009/1article 3.

Yeshurun, R., Bar-Oz, G., Nadel, D., 2013a. The social role of food in the Natufian cem-etery of Raqefet Cave, Mount Carmel, Israel. J. Anthropol. Archaeol. 32, 511e526.

Yeshurun, R., Bar-Oz, G., Kaufman, D., Weinstein-Evron, M., 2013b. Domestic refusemaintenance in the Natufian: Faunal evidence from el-Wad Terrace, MountCarmel. In: Bar-Yosef, O., Valla, F.R. (Eds.), Natufian Foragers in the Levant:

Terminal Pleistocene Social Changes in Western Asia. International Monographsin Prehistory, Ann Arbor, pp. 118e138.

Yeshurun, R., Bar-Oz, G., Kaufman, D., Weinstein-Evron, M. Purpose, Permanence,and Perception of 14,000-Year-Old architecture: Contextual taphonomy of foodrefuse (submitted for publication).

Yom-Tov, Y., 2003. Poaching of Israeli wildlife by guestworkers. Biol. Conserv.110,11e20.Zeder, M.A., 2012. The Broad Spectrum Revolution at 40: Resource diversity,

intensification, and an alternative to optimal foraging explanations.J. Anthropol. Archaeol. 31, 241e264.

Zohar, I., 2003. Fish exploitation at the Sea of Galilee (Israel) by early fisher-hunter-gatherers (23,000 B.P.): ecological, economical, and cultural implications. Ph.D.Dissertation, Tel Aviv University.