Brooke S. BladesDepartment of Anthropology,University of Maryland,College Park, Maryland20742, U.S.A.E-mail: upblades@msn.com

Received 23 April 1998Revision received12 October 1998 andaccepted 7 February 1999

Keywords: Paleolithicarchaeology, southwesternFrance, Aurignacian, lithiceconomy.

Aurignacian lithic economy and earlymodern human mobility: new perspectivesfrom classic sites in the Vézère valley ofFrance

During the past decade the chronology and hominin attributions ofthe Aurignacian have been revised or called into question. Thesecontroversies have coincided with an increased appreciation for thesocial complexity of Aurignacian culture in the realms of organictechnologies and mobiliary and parietal manifestations of symbolicbehavior. Lithic raw material procurement and reduction intensityevidence from Aurignacian occupations at the Vézère Valley sites ofAbri Pataud, Le Facteur, and La Ferrassie may reflect complex groupmobility strategies. The lithic components under consideration werealways dominated by cherts available within a few kilometers radius.Assemblages associated with the early Aurignacian have elevatedproportions of cherts from distant sources. Lithic retouch dataindicate that some early Aurignacian assemblages reflect greaterextent and/or intensity of marginal retouch compared with the laterAurignacian. Lithic reduction data, however, reveal evidence ofgreater core reduction intensity during the later Aurignacian. Flexiblestrategies of residential mobility, possibly in response to changes inthe subsistence environment, may account for some of the variabilitybetween early and later Aurignacian assemblages. Similar shifts in rawmaterial procurement were evidently associated with the MiddlePaleolithic in southwestern France. However, Aurignacian popu-lations may have acquired most lithic materials by movement directlyto sources, while certain non-utilitarian materials were probablyobtained via some form of indirect social exchange. This suggestedcoexistence of direct and indirect procurement mechanisms serves todistinguish Aurignacian assemblages from earlier Middle Paleolithicdeposits and emphasizes that socially-directed intensification was oneof the fundamental elements of the suite of cultural changes referredto as the Middle–Upper Paleolithic transition.

? 1999 Academic Press

Journal of Human Evolution (1999) 37, 91–120Article No. jhev.1999.0303Available online at http://www.idealibrary.com on

Introduction

The Aurignacian is emerging as an increas-ingly complex and poorly understood cul-tural phenomenon. Indeed, the very natureof what is meant by the Aurignacian must bereevaluated in response to the identificationof Upper Paleolithic-like assemblages thatappear to extend back to the period between40 and 50 ka in eastern, southeastern, andcentral Europe (Valoch, 1976; Oliva, 1981,

0047–2484/99/070091+30$30.00/0

1984, 1993; Kozłowski, 1982, 1988;Svoboda, 1983, 1987, 1993; Allsworth-Jones, 1986; Soffer, 1989, 1991; Marks,1993), northern Spain (Bischoff et al., 1989;Cabrera Valdés & Bernaldo de Quirós,1991; Bernaldo de Quirós & CabreraValdés, 1993; Cabrera Valdés et al., 1997),and Belgium (Otte & Straus, 1995).

The dating of these earliest UpperPaleolithic assemblages and whether all ofthem may be considered Aurignacian or

? 1999 Academic Press

92 . .

Aurignacian-like in any meaningful sensehas given rise to various controversies.Conceptions of the origins of theChâtelperronian and whether it may beregarded as occurring prior to or contem-porary with the earliest Aurignacian remainsubjects of intense debate (D’Errico et al.,1998 and associated comments, particularlyMellars, 1998; Taborin, 1998; White,1998). Closely related to the issue ofchronology are differences regarding theextent to which manifestations of symbolicbehavior found in some Châtelperronianassemblages represent independent inven-tions by Neandertals or reflect some degreeof contact with Aurignacian groups.

Another controversy concerns whichhominin species or sub-species was respon-sible for ‘‘creating’’ the Aurignacian. Associ-ation of anatomically modern humans withthe Aurignacian is accepted by many paleo-anthropologists (Stringer et al., 1984;Binford, 1989; Gambier, 1989; Stringer,1989; Trinkaus, 1989; White, 1989a,b;Aitken et al., 1993; Klein, 1995). Othershave contended that the Aurignacian is infact a Neandertal industry. Specifically,Wolpoff (1981), Miracle & Crummett(1995), Karavanic (1995), Straus (1997),and Karavanic & Smith (1998) have arguedthat deposits in Vindija Cave in Croatiaestablish that Neandertals produced theAurignacian. Neandertal skeletal remainshave been recovered from a level at Vindijadating ca. 33 ka in association with a mixedMiddle and Upper Paleolithic artefacttypology that included a split-based andother organic points.

It is worth noting in this regard that othersites in central and southeastern Europeindicate different associations. Fossils suchas Stetten (Germany), Mladec (CzechRepublic), Velika Pecina (Croatia) and lessconfidently Hahnöfersand (Germany) aregenerally acknowledged to be modern inmorphology, although the significance ofapparent archaic features has been the

subject of considerable debate (Bräuer,1989; Mellars & Stringer, 1989; Stringer,1989; Wolpoff, 1989; Frayer et al., 1993;Lieberman, 1995; Frayer, 1997). Mladecand Velika Pecina are both considered todate prior to or about ca. 34 ka, whileStetten may be somewhat later (ca. 30 ka)(Brauer, 1989). These fossils are archaeo-logically associated with Aurignaciandeposits.

Hominin remains recovered fromAurignacian contexts in southwesternEurope are generally fragmentary ones.Garralda (1997:160) concluded that thefossils associated with the ‘‘ancientAurignacian’’ level 18 at El Castillo innorthern Spain possess archaic traits but aretaxonomically unresolved. As Rigaud (1989,1997) and Gambier (1989, 1997) empha-sized, the morphologies of the individualswho generated the earliest Aurignacian insouthwestern France are at presentunknown. The numerous remains fromCro-Magnon Cave are most frequently cor-related with the Aurignacian (Sonneville-Bordes, 1959; Movius, 1995; Lumley et al.,1984; Gambier, 1989), but the circum-stances of recovery during the nineteenthcentury will probably always cause questionsconcerning provenience. Chronology istherefore problematic unless the Cro-Magnon fossils themselves can be radio-metrically dated. Gambier (1989, 1997)reviewed various fossil associations with theAurignacian in France and noted thattwo sites in particular (Les Rois andIsturitz) yielded robust but, in her opinion,modern mandibular fragments. She thusargued that the early Aurignacian—i.e., inthe sense of Peyrony’s phase I to bediscussed—was generated by anatomicallymodern humans.

The meaning of the Aurignacian as acultural phenomenon is often entangledwithin the biological argument concerningthe broadly defined ‘‘independent con-tinuity’’ and ‘‘replacement’’ models for the

93

emergence of anatomically modern humans.The social complexity of the Aurignacian isincreasingly recognized in the appearance ofphysical manifestations of symbolic behavior(White, 1989a,b; Taborin, 1993a,b), thesophistication of techniques for produc-ing these symbolic manifestations as wellas a broader range of organic projectiletechnologies (Knecht, 1993), and the stun-ning parietal images in Chauvet Cave insoutheastern France that date ca. 32 ka(Clottes, 1996). The importance of open-airloci in Aurignacian settlement systemshas been emphasized by excavations bothwithin the Périgord and in neighboringregions (Chadelle, 1990; Le Brun-Ricalens,1990, 1993; Le Brun-Ricalens & Neveu,1991).

The present article focuses upon anotheraspect of Aurignacian complexity during theperiod 35–27 ka: the inference of variationsin group mobility strategies and settlementpatterns based upon comparisons of lithiceconomy within the Aurignacian. Datarelated to paleoenvironment, site geography,faunal remains, and lithic economy suggestearly modern humans adopted variablemobility strategies as a cultural meansof adjusting to changes in the structure ofsubsistence resources. Material evidence ofAurignacian social complexity is manifestedby the early Aurignacian within the studyarea of the lower Vézère Valley, and indeed,in some of the study assemblages, and cer-tainly argues for the additional considerationof intragroup and intergroup social relationsin the interpretation of settlement structure.The same archaeological assemblage maypreserve evidence of direct lithic procure-ment and acquisition of nonlithic materialsthrough indirect exchange. Such a duality inprocurement systems, with a concomitantinterplay between resource and socially-oriented strategies, forms the overridingframework for evaluating Aurignacianbehaviors in the portion of southwesternFrance under consideration.

Aurignacian systematics and sites

The classic bases for defining theAurignacian are well recognized: organicmaterial and lithic blade technologies, dis-tinctive heavy and elongated retouch tech-niques, characteristic artefacts such as thick‘‘Aurignacian’’ scrapers and inverselyretouched ‘‘Dufour’’ bladelets (Brooks,1982). Beads and pendants fashioned fromivory, steatite, or marine shells and lime-stone blocks bearing engraved lines orpainted designs represent radical departuresfrom the meager and usually questionablecollection of decorated objects in MiddlePaleolithic contexts (White, 1982, 1993).

The ordering of Aurignacian assemblagesin southwestern France has been directedfor the past half-century by the ‘‘phase’’structure proposed by Peyrony (1933, 1934)based upon his excavations at La Ferrassie.The importance of changing relative propor-tions of lithic tools between the phases wasrecognized by Sonneville-Bordes (1960,1980). The early Aurignacian (Peyrony’sphase I) is dominated by retouched bladesand endscrapers on blades, often with heavymarginal retouch. Split-based antler pointsare frequently encountered if soil matricespermitted preservation (Peyrony, 1934;Knecht, 1993). The later Aurignacian(Peyrony’s phase II) is traditionallydistinguished by an increase in burins and‘‘Aurignacian’’ endscrapers made on flakesand thicker blades, a decrease in the inten-sity of marginal retouch, and the appearanceof other antler point morphologies. Peyrony(1934) argued that subsequent manifes-tations (III, IV, and V) were present at LaFerrassie, but phases III and IV have notbeen isolated at other sites and are notwidely accepted. Phase V actually dates tomuch later in time.

The terms ‘‘early’’ and ‘‘later’’ have beenadopted herein to emphasize the importanceof focusing upon changing behaviors withinthe temporal continuum of the Aurignacian.

94 . .

AtlanticOcean

N

100 km

MassifCentral

Dordogne

Vézer

e´

CoastalPlain

100m

depth

0

Pyrénées

Figure 1. Southwestern France with major rivers and physiographic features indicated. The study area inthe lower Vézère Valley is denoted by a large dot. Redrawn from Bartholomew’s New International Atlas,1986 edition.

The alternating proportions of tool‘‘types’’ suggest a continuum within theAurignacian, with the arbitrary divisionbetween phases I and II occurring whenAurignacian scrapers and burins numericallysurpass blade endscrapers and retouchedblades. As will be discussed below, a con-tinuum is also indicated in the decreas-ing presence of particularly heavy or‘‘Aurignacian’’ marginal retouch.

The sites of Abri Pataud, Le Facteur, andLa Ferrassie are located in or near the villageof Les Eyzies in the Périgord of south-western France. The Périgord lies betweenthe younger Tertiary sediments of theAtlantic coastal plain to the west and theolder crystalline highlands of the MassifCentral to the east (Laville et al., 1980).Major geographic features defining thetopography are westward-flowing riversthat have cut downward into the underlyingCretaceous limestone (Figure 1). The land-

scape may be divided into three major land-forms: river valleys, adjacent cliff slopes andinterfluvial slopes, and plateaus betweenthe major river valleys (White, 1985). LeFacteur and Abri Pataud are located 50 mand 150 m respectively from the Vézère,while La Ferrassie is found in an interfluvialvalley 4 km north of the Vézère (Figure 2).

Climatic conditions between 35 and 27 kawere influenced by a very cold and dryglacial stadial during the early Aurignacian,which ameliorated during later Aurignacianoccupation into a warmer and moisterperiod associated with the Denekamp orArcy interstadial (Table 1). Laville et al.(1980) described general environmentalconditions during the latter part of oxygenisotope stage 3 as ranging from steppe ina cold, dry stadial to parkland during awarmer interstadial. A finer degree ofinterpretation, however, requires consider-ation of microclimatic variables such as solar

95

0

N

Abri Pataud

5 km

Le Facteur

La Ferrassie

Figure 2. The study area in the lower Vézère Valley. The interfluvial position of La Ferrassie contrasts withthe river valley locations of Abri Pataud and Le Facteur. Redrawn from Delluc & Delluc (1978; Figure 1).

Table 1 Paleoenvironmental summary

Level Sediments Pollen

Pataud 7 LH —Ferrassie J WM WMFerrassie K2 WM WMFacteur 19 WM WMFerrassie K3 WM WMFerrassie K4 LH CDPataud 8 LH —Facteur 21 LH CDPataud 11 WM —Pataud 12 CD —Ferrassie K5 CD CDFerrassie K6 CD CDPataud 13 CD —Pataud 14 CD —

C=cold, L=cool, W=warm, D=dry, H=humid,M=moist.

Sources: Laville, 1968, 1975; Leroi-Gourhan, 1968;Farrand, 1975, 1995; Laville et al., 1980; Laville &Tuffreau, 1984; Paquereau, 1984.

exposure and topographic settings. Openareas such as uplands and valley bottomswould have favored grassy steppe vegetationwith some arboreal elements, particularlyconiferous species on acidic plateau soils.

Sheltered valleys and south-facing slopes, onthe other hand, would have supportedthermophilous deciduous trees (Wilson,1975; White, 1985).

The three sites under examination areparticularly relevant since all contain earlyand late Aurignacian occupation levelsin stratigraphic succession. The earlyAurignacian level 21 represents the basalcultural component of Le Facteur and isseparated from the later Aurignacian level19 by a sterile deposit of éboulis or limestonefragments spalled from the shelter walls.No widely-accepted radiometric dates havebeen presented for these layers (Delporte,1968). The earliest occupation deposit atAbri Pataud was recovered from the ‘‘basal’’Aurignacian level 14. A sequence ofAurignacian occupations dating between35 ka and 29 ka and separated by éboulislayers was excavated, with rich faunaland/or artefact assemblages for the earlyAurignacian (levels 14–11), laterAurignacian (levels 8 and 7), and ‘‘evolved’’Aurignacian (level 6). Perigordian,Protomagdalenian, and Solutrean levels

96 . .

were also present (Movius, 1975, 1977;Bricker et al., 1995). Delporte’s reexami-nation and limited excavation of twointersecting stratigraphic sections at LaFerrassie revealed a complex sequenceof occupations: Mousterian (levelsM2–L3), Châtelperronian (L3–L2), basalAurignacian (L1–K7), Aurignacian I(K6–K5), Aurignacian II (K4–J), evolvedAurignacian (I3–F1), and Perigordian(D3–B3). Deposition dates for AurignacianI and II levels ranged from 35 ka to 27 ka(Delporte, 1984, 1991; Delibras, 1984).

Lithic economy

The inference of group residential mobilityfrom lithic raw material compositionprovides an important middle-range theor-etical paradigm linking static archaeologicalassemblages and dynamic past human socialbehavior. Systematic exploitation of specificdistant sources during the Upper Paleolithichas been considered indicative of antici-patory strategies rather than the moreopportunistic strategies of the Middle Paleo-lithic (Geneste, 1988, 1989). An increase inimported raw materials and the forms inwhich those materials appeared has beennoted on Aurignacian sites compared withMousterian and Châtelperronian occupa-tions in southwestern France (Turq, 1991).Studies of raw material economy in andnear the Périgord have suggested expandingzones of resource exploitation, reflected byprocurement distances for most lithics of5 km in the Mousterian to 15 km during theAurignacian and Perigordian (Geneste,1988). Lithics from increasingly distantsources may denote larger foraging rangesduring the Upper Paleolithic (Rigaud &Simek, 1990).

A dichotomy exists in the archaeologicalliterature concerning the interpretation oflithic raw material patterns, often posed asan opposition between some form of director indirect acquisition. Direct modes of

procurement are incidental to subsistencemobility or arise from dedicated extractionforays to a particular source. Indirectacquisition is attributed to exchange or tradewith neighboring social groups. Ethno-archaeological observations have beeninvoked to support the notion of lithic pro-curement as ‘‘embedded’’ in, and thus sub-sidiary to, subsistence pursuits (Binford,1979, 1980; Goodyear, 1979) or as con-ditioned by social structure and symbolicsignificance (Gould, 1980). The presence oflithics from distant sources on Paleolithicsites has been evaluated as evidence of vary-ing degrees of exchange between socialgroups (Wilmsen, 1970; Kozłowski, 1973,but see Kozłowski, 1991; Gamble, 1986;Oliva, 1993).

The interpretation of lithic raw materialcomposition within an archaeologicalassemblage is complicated by the problemof equifinality (Meltzer, 1989; Morrow &Jeffries, 1989), since procurement duringgroup movement or through social exchangemay produce similar patterns. It is par-ticularly important to realize, therefore,that mobility may provide simultaneousopportunities for direct source access andindirect procurement via social contact andexchange. Unless one conceives of sedentaryUpper Paleolithic groups exchanging all butthe most local lithic materials, group move-ment is ultimately responsible for at leastsome component of lithic raw material com-position. Further, lithics for utilitarian toolsmay reflect group mobility regardless of theextent to which other materials wereexchanged. Limited faunal seasonality datato be discussed below indicate that earlyUpper Paleolithic movement to and fromAbri Pataud and La Ferrassie occurred on aseasonal basis.

The reconstruction of Aurignacian lithiceconomy and more broadly of technologicalorganization is derived from three perspec-tives: lithic raw material source attributions,the technological stages in which each

97

material was introduced to a site, and theintensity of reduction for all lithic materials.These data collectively reflect the procure-ment areas that were encountered directly orindirectly by Aurignacians and economicstrategies for exploiting the lithic resourcesfrom those procurement areas, with impli-cations for organizational concerns such asgroup mobility and social interaction.

Lithic raw materials and technologicalsequences

Raw material analyses in the Périgordbenefit from three decades of intense geo-logic survey oriented to the identification oflithic raw material sources exploited duringthe Paleolithic. The result has been theidentification of nearly 1000 such sources,virtually all of which contain cherts locatedin primary outcrops and secondary alluvialand colluvial deposits (Bricker, 1975;Demars, 1982; Morala, 1984; Chadelle,1983; Larick, 1983; Geneste, 1985). Theupper Cretaceous limestone formationsdating to the Senonian epoch found inthe immediate vicinity of the rock sheltersites contain abundant chert deposits.‘‘Chalcedony’’ may be obtained 15–25 kmto the southwest. Other cryptocrystallinematerials considered distant relative to thesite loci are those from Bergerac to the westand the Isle Valley to the northwest(25–40 km), Fumel and Gavaudun to thesouth (35–45 km), and jasper sourcesprobably 30–40 km to the east. Quartzitesand materials of a noncryptocrystallinenature were present in very limitedquantities in the study assemblages.

The contrast between locally availableand distant materials is an important ele-ment in the interpretation of Aurignacianvariability. Most Paleolithic assemblages inchert-rich areas such as the Périgord aredominated by lithics obtained in the im-mediate vicinity. The distance at which alithic material may be considered ‘‘distant’’

relative to a site must be assessed withregard to the distribution of regionalresources and lithic economy manifestedin archaeological assemblages. Lithic pro-curement patterns encompassed greaterdistances in central Europe than in south-western France from the late MiddlePaleolithic through the Upper Paleolithic(Rensink et al., 1991; Féblot-Augustins,1997). The quantitative decrease inmaterials from sources 25 km and beyond,coupled with the introduction of thosematerials in a limited range of techno-logical forms, supports the argument thatAurignacian groups in the Vézère Valleywould have considered such sources to bedistant although not necessarily preferredones.

An analysis of reduction sequenceprovides an indication of the nature oftechnological activities undertaken duringdeposition of a particular assemblage. Thestructure employed to categorize the tech-nological remnants of lithic reduction andproduction reflects a combination ofapproaches advocated by Collins (1975),Geneste (1985), and particularly Chadelle(1983). The categories emphasize thetechnological ‘‘flow’’ from raw materialacquisition and core preparation throughblank production to tool refurbishment anddiscard:

O: raw material(untested or tested material blocks)

I: preparation/flake reduction(cortical and noncortical flakes)

II: blade production(cortical and noncortical blades)

III: preparation discard(crested blades, core tablettes)

IV: reduction remnants(cores and core fragments)

V: production discard(burins spalls, reutilized tools)

This technological reduction structurereflects the realization that a lithic piece at

98 . .

any stage—from untested block to broken ordiscarded tool—may be used expediently orshaped into a form that the modern analystperceives as a formal ‘‘tool.’’ The structuretherefore provides a means of evaluating theblank for a given retouched piece that isindependent of traditional typologies (Table2). Further, this structure indicates the tech-nological stage in which each material wasintroduced to a site during an occupationand, inferentially, the nature of the lithicproducts that were removed from a siteduring subsequent group movements. Thepresence of jasper burin spalls in LaFerrassie level K4 or unretouched Fumelblade fragments in Le Facteur level 19without the recovery of correspondingjasper burins or Fumel blade tools wouldinfer that those products were removedas Aurignacian groups moved to a new

Distribution by technological categories (percentages by weight)

Technological categoryPataud

14Ferr.K6

Facteur21

Ferr.K4

Facteur19

Debitage:O—raw material blocks — 5·3 8·2 — —I—flakes 47·7 52·5 50·1 35·1 75·1II—blades 6·2 6·8 13·8 3·2 5·2III—crested blades, tablettes 1·4 3·8 1·4 16·4 0·6IV—cores 5·0 25·0 25·6 38·0 12·2V—burin spalls — 0·1 0·1 0·5 0·4Other 39·7 6·5 0·9 6·8 6·5Weight (kg) * 24·6 8·4 24·3 6·5

Retouched pieces:†O—raw material blocks — — — — —I—flakes 19·0 39·7 26·1 40·7 61·0II—blades 65·9 40·5 66·2 43·3 28·6III—crested blades, tablettes — 9·2 4·5 10·2 2·2IV—cores — 8·4 3·2 3·8 2·5V—spalls, reutilized tools — 2·2 — 1·7 2·7Other 15·1 — — 0·3 3·0Weight (kg) * 4·8 2·6 7·8 5·3Total type tools (N) 171 231 114 443 175

Source for Abri Pataud level 14 data: Bondon, 1993.*Based on numerical count since weight not recorded; Other category for Abri

Pataud level 14 represents those pieces for which technological attribution was notapparent.

†Retouched pieces include all of the type tools from the Sonneville-Bordes andPerrot list of 92 tool types and varying numbers of slightly ‘‘retouched’’/utilized bladesand flakes.

Table 2

location, or that the products were notrecovered during the excavations. However,the presence of a core rejuvenation tabletteon Gavaudun chert in La Ferrassie level K6does not necessarily imply that core reduc-tion of that material occurred on site since athick tablette flake may have been trans-ported as an unretouched blank.

The dominance of locally availablematerials—Senonian gray and browncherts—in the debitage and retouched col-lections for all assemblages under study ismarked. Senonian materials accounted forthe vast majority of debitage flakes andblades, all of the untested blocks, all but oneof the cores, and most of the remnants ofcore preparation and rejuvenation such ascrested blades and tablettes. Senonianmaterials were present within each of thetechnological categories (Figure 3).

99

other

60

0block

Technological categories

Per

cen

tage

(w

eigh

t)

50

40

30

20

10

blade coreflake tablette tool

Figure 3. The distribution by weight of locally available raw materials among the various technologicalcategories for the early Aurignacian occupations of La Ferrassie level K6 and Le Facteur level 21 .The majority of the materials were present as flakes or cores.

Relative retouch percentages, lithic raw materials

Material kmPataud

14Ferr.K6

Facteur21

Ferr.K4

Facteur19

Senonian gray 0–5 10·6 16·4 20·1 26·6 27·6Senonian brown 0–5 9·7 18·8 11·3 13·0 16·1Quartzite 0–5 100·0 — — — 29·1Chalcedony 15–25 — 24·1 55·6 95·2 16·4Bergerac 25–40 90·9 57·7 91·6 99·7 96·0Jasper 30–40 100·0 81·6 74·2 58·6 44·1Fumel/Gavaudun 35–45 — 19·6 88·8 100·0 10·1Other ? 40·9 95·6 52·8 99·4 81·2

Source for Abri Pataud level 14 data: Bondon, 1993.

Table 3

Senonian cherts also dominated the toolcollections, but the percentages of theselocal materials that were actually convertedinto retouched tools were low compared tothe percentages of distant cherts so utilized(Table 3). Utilization of local materials con-sistently fell below 30% for gray Senonianand below 20% for brown Senonian. Con-siderable quantities of these cherts availablewithin a radius of a few kilometers weretransported to the shelters, but most of thatmaterial was expended in secondary corepreparation/flake reduction and blade pro-duction. The small number of untestedblocks and relatively limited quantities of

fully cortical flakes suggested that mostof the block testing and preliminarypreparation occurred at nearby sources.

Debitage amounts were consistently lowfor materials from distant sources, butinterassemblage variability was greater whenretouched pieces were considered. A markedincrease in retouched distant materialsoccurred in two of the three earlyAurignacian assemblages: Abri Pataud level14 and Le Facteur level 21. By contrast,both of the later Aurignacian assemblageshad low percentages of distant materials.

No cores on distant materials werepresent in the assemblages, although one

100 . .

other

60

0block

Technological categories

Per

cen

tage

(w

eigh

t)

50

40

30

20

10

blade coreflake tablette tool

Figure 4. The distribution by weight of distant materials among the various technological categoriesfor the early Aurignacian occupation of La Ferrassie level K6 and Le Facteur level 21 . No cores werepresent; the proportion of blades was higher in Le Facteur level 21.

chalcedony core was found in level K6 at LaFerrassie (Figure 4). Relatively few debitageflakes on cherts from distant sources wererecovered and those encountered wereusually small in size and noncortical. Un-retouched blade segments were rarelyencountered. These data indicate thatdistant materials were primarily but notexclusively imported as finished blade toolsand unretouched blade blanks that weresubsequently retouched on site.

These economic patterns—a dominanceof local materials with a relatively lowretouch percentage, compared with limitedquantities of generally retouched distantcherts—are essentially those noted byChadelle (1983) for the Perigordian levelVII at the Dordogne Valley site of LeFlageolet and represent the Upper Paleo-lithic equivalents to those defined for theMousterian by Geneste (1985, 1988).Further, the basic patterns did not changebetween the early and later Aurignacian.However, differences were indicated in theamounts of retouched distant materialspresent (Figure 5). The percentages of dis-tant materials among the retouched portionof the assemblages were greater during theearly portion of the Aurignacian.

Lithic reduction

Blade blanks were retouched more fre-quently than flakes and larger blanks werepreferentially selected for retouch, asillustrated for Le Facteur (Figures 6 & 7).Comparable trends were indicated at LaFerrassie. These data emphasize a point thatDibble et al. (1995) indicated has beenrepeatedly demonstrated in Middle Paleo-lithic studies: larger blanks were generallyselected for retouch (Geneste, 1985; Dibble,1988; Dibble & Holdaway, 1993; Meignen,1993; Dibble et al., 1995).

It has long been recognized that earlyAurignacian lithics were more heavilyretouched than those from the laterAurignacian (Peyrony, 1933; Sonneville-Bordes, 1960; Delporte, 1968; Delporteet al., 1977, 1983; Brooks, 1979). One formof particularly heavy retouch, known appro-priately as ‘‘Aurignacian,’’ was stronglylinked to temporal change, diminishingsteadily throughout the Aurignacian(Figure 8).

Retouch may be evaluated in terms of twocomponents, extent and intensity. Theextent of retouched blade edges, as evaluatedby the number of retouched edge segments,

101

0

25

Per

cen

tage

of

dist

ant

mat

eria

ls

20

10

Early Aur. Later Aur.

Pat 14Fer K6

Pey FPat 11

Fac 21Pat 8

Fer K4Fer K2

Pey HFac 19

Fer JPat 7

5

15

Figure 5. Percentages of tools made on distant materials for early and later Aurignacian assemblages,including Peyrony’s ‘‘Aurignacian I’’ level F and ‘‘Aurignacian II’’ level H from the earlier excavationsat La Ferrassie. (The Abri Pataud data are derived from Brooks, 1995 and Bondon, 1993.) Allearly Aurignacian assemblages, except Delporte’s level K6, contain relatively elevated percentages.Abbreviations: Fac, Le Facteur; Fer, La Ferrassie; Pat, Abri Pataud; Pey, Peyrony.

0

100

Bladelet/blade width (mm)

Per

cen

tage

by

nu

mbe

r

40

20

60

80

20–30

Facteur 21 Facteur 19

30+12–200–8 8–12

Figure 6. Percentages of retouched bladelets/blades within width categories from Le Facteur level 21 andlevel 19.

decreased considerably during the laterAurignacian in La Ferrassie level K4. Theintensity of retouch on each blade segment,as evaluated in categories of light, heavy,and ‘‘Aurignacian,’’ was pronounced in theearly Aurignacian Le Facteur level 21.Kolmogorov–Smirnov two-sample tests

(Table 4) suggest that these differences weresignificant ones.

The intensity of lithic reduction may beevaluated separately from the intensity oflithic retouch. A discussion of lithic reductionintensity is complicated by the variousmeans of evaluation. Dibble et al. (1995)

102 . .

0

100

Maximum flake diameter (cm)

Per

cen

tage

by

nu

mbe

r

40

20

60

8

Facteur 21 Facteur 19

1062 4

80

Figure 7. Percentages of retouched flakes within size categories for Le Facteur level 21 and 19.

0

50

Per

cen

tage

Pat 14

40

30

20

10

Fer K6Fac 21

Pat 11Fer K4

Pat 8Fac 19

Pat 7

Figure 8. Percentages of blade tools bearing particularly heavy, or ‘‘Aurignacian,’’ retouch. Theassemblages, with the possible exception of Abri Pataud level 7, are arranged in approximate chronologicalorder which suggests a continuous decrease in the presence of such heavy retouch. Abbreviations as inFigure 5.

observed that the intensity of core reductionexerted considerable influence upon variousrelationships within lithic assemblages,assuming considerations of technology andraw material variability remained constants.They cited numerous studies to argue thatas core reduction increases, the numberof blanks per core and extent of core

preparation also increase (Munday, 1977;Marks, 1988; Bar-Yosef, 1991; Montet-White, 1991), while average core size, flakesize, flake platform area, and cortex decrease(Newcomer, 1971; Stahle & Dunn, 1982;Henry, 1989; Marks et al., 1991).

It should be noted that the collection of‘‘blanks’’ consisted of intact and proximal

103

Table 4 Statistical comparisons (Kolmogorov–Smirnov two-sample) of retouch extent andintensity

Location D P

Extent of retouch:Ferrassie K6 and K4 0·594 <0·01Facteur 21 and 19 0·152 >0·05Ferrassie K6 and Facteur 21 0·157 >0·05Ferrassie K4 and Facteur 19 0·285 <0·01

Intensity of retouch:Ferrassie K6 and K4 0·136 >0·05Facteur 21 and 19 0·365 <0·01Ferrassie K6 and Facteur 21 0·205 <0·01Ferrassie K4 and Facteur 19 0·098 >0·05

flakes and blades, both unretouched andretouched. Henry (1989) suggested that adistinction may be drawn between potentialblanks and ‘‘debris’’ within the unretoucheddebitage from a given assemblage basedupon the minimum size threshold forretouched pieces within that assemblage.Following the suggestion of Henry, thesmallest flakes (less than 2 cm) and bladelets(less than 8 mm wide) were excluded fromconsideration as blanks since these sizeswere rarely retouched.

Certain measures should serve to indicatethe relative degree of core reduction, includ-ing blank to core ratio, core size, blank size,and amounts of remaining cortical covering.The extent of flake reduction/blade pro-duction may also be evaluated by cortexamounts, in addition to the relative percent-age of retouched blanks and the quantities ofscars reflecting previous blank removals.

The image that emerges is one of spatialand temporal variation, although certainlynot in all instances. The quantities of blanksper core were considerably greater at LeFacteur compared with La Ferrassie, due inpart to the limited number of cores recov-ered at the former site (Figure 9). However,the percentage of cortical covering on flakesand blades was also greater at Le Facteur,which suggests that a greater amount of

primary reduction occurred at this site.Temporal variability is reflected in thegreater number of blanks per core during thelater Aurignacian at Le Facteur, althoughthe ‘‘adjusted’’ ratio suggests comparabletotals at La Ferrassie. A decrease in theamount of cortical covering in the laterassemblages at both sites is consistent withevidence of increased reduction intensityduring the later Aurignacian.

Smaller core sizes are expected in assem-blages that are more intensively retouched.Metric data indicated that such was not thecase for intact cores from levels K6 (N=43)and K4 (N=47) at La Ferrassie (weight:t=0·534, df=69, P>0·50; length: t=0·007,df=69, P>0·90). (Le Facteur yielded toofew intact cores—four in level 21 and five inlevel 19—for meaningful analyses). Thecomparisons do suggest, however, that anydifferences in reduction intensity at LaFerrassie are not due to difficulties in rawmaterial procurement. This suggestion isreinforced by the comparable widths ofblade blanks (Table 5).

Changes in lithic technology or typologymay result in differences in blank sizes. Forexample, the increased production of thethick ‘‘Aurignacian’’ scrapers during thelater Aurignacian is certainly reflected inthe higher proportions of bladelets within LaFerrassie level K4. The elimination of thesmallest flake and bladelet size categoriesfrom consideration as blanks should aid incountering biases introduced by changingproportions of tool forms (Figure 10). Ananalysis of blank sizes indicated that bladewidths were comparable at both sites,although wider blades were selected forretouch in La Ferrassie level K4 comparedwith level K6 and Le Facteur level 19.Flake diameters were evidently significantlysmaller during the later Aurignacian at LaFerrassie, supporting indications of greaterreduction intensity (Figure 11). No differ-ences in flake sizes were apparent at LeFacteur.

104 . .

2540

55

Blank to core ratio

Cor

tex

perc

enta

ge

0

45

30

50

Fer K6

Fer K4

Fac 19

Fac 21

80 120 160 200

35

40

Figure 9. The blank to core ratios, based upon flakes larger than 2 cm in diameter and blades wider than8 mm, correlated with cortex percentages. Early Aurignacian assemblages (levels K6 and 21) have highercortex percentages than later ones at the same sites. Le Facteur assemblages are more cortical and haveconsiderably higher ratios than those at La Ferrassie. Abbreviations as in Figure 5.

Table 5 Statistical comparisons (Kolmogorov–Smirnov two-sample) of cortex and blank sizes

Location D P

Flake cortex:Ferrassie K6 and K4 0·128 <0·01Facteur 21 and 19 0·099 <0·01Ferrassie K6 and Facteur 21 0·098 <0·01Ferrassie K4 and Facteur 19 0·127 <0·01

Blade cortex:Ferrassie K6 and K4 0·075 >0·05Facteur 21 and 19 0·048 >0·05Ferrassie K6 and Facteur 21 0·135 >0·05Ferrassie K4 and Facteur 19 0·016 >0·05

Flake diameter (>2 cm):Ferrassie K6 and K4 0·122 <0·01Facteur 21 and 19 0·032 >0·05Ferrassie K6 and Facteur 21 0·080 >0·05Ferrassie K4 and Facteur 19 0·036 >0·05

Blade width (>8 mm):Ferrassie K6 and K4 0·112 >0·05Facteur 21 and 19 0·050 >0·05Ferrassie K6 and Facteur 21 0·083 >0·05Ferrassie K4 and Facteur 19 0·079 >0·05

Lithic economy and Aurignacianmobility strategies

Lithic reduction intensity reflects variabilitythat is temporal and spatial in nature. Vari-ous measures of core reduction suggest thatgreater intensity was associated with thelater Aurignacian assemblages. Further, theassemblages at Le Facteur would seem tohave been more intensively reduced thanthose at La Ferrassie. The reduction inten-sity data are intriguing since the extent andintensity of marginal retouch on blade toolswere greater earlier rather than later in theAurignacian.

The differences in flake sizes as measuresof reduction intensity at La Ferrassie aresupported by the quantity of scars fromprevious removals on dorsal surfaces of

intact and proximal flakes (Figure 12). Nosignificant differences were noticed amongthe blades from either site, but flakes inlevel K6 appeared to have significantlyfewer dorsal scars compared with thosefrom level K4 (Kolmogorov–Smirnov two-sample: D=0·273, P<0·01). Thus, flakeswere not only larger in the early Aurignacianat La Ferrassie but also bore fewer traces ofprevious removals, suggesting greater reduc-tion intensity during the later Aurignacian.

105

02

60

Average blank size (cm)

Cor

tex

perc

enta

ge

1

40

30

20

10

50

Fer K6Fac 19

Fer K4Fac 21 Fer K6

Fac 19

Fer K4

Fac 21

Blades Flakes

3 4 5

Figure 10. Average size for blanks correlated with cortical percentages for flakes and blades from LaFerrassie and Le Facteur. Variation in percentages of cortex is most pronounced for flakes and sizedifferences are marked for flakes at La Ferrassie. Abbreviations as in Figure 5.

010

60

Maximum flake diameter (cm)

Per

cen

tage

by

nu

mbe

r

2

40

30

20

10

50

4 6 8

Ferrassie K6 Ferrassie K4

Figure 11. The distributions of intact and proximal flake sizes for levels K6 and K4 at La Ferrassie.

Greater reduction intensity and higherproportions of local raw materials wereassociated with the later Aurignacian.Intensity of reduction may be related tochanging mobility patterns during thewarmer interstadial conditions that occurredlater in the Aurignacian. However, corereduction also mirrors differences in settle-ment pattern, site function, and raw

material access, while tool retouch intensi-ties reflect functional considerations, withor without the influence of mobility con-straints. Further consideration of rawmaterial distributions and other factors willassist in determining the extent to which rawmaterial economy reflects mobility patterns.

An increase of distant materials—particularly from Bergerac—has been

106 . .

04+

70

Quantity of dorsal scars

Per

cen

tage

0

40

30

20

10

60

50

1 2 3

Ferrassie K6 Ferrassie K4

Figure 12. Dorsal scar counts on intact and proximal flake tools from La Ferrassie level K6 (N=65) andlevel K4 (N=90). The scars on flakes in level K4 appear to be significantly greater in number.

Table 6 Blade tool metrics

FerrassieK6–K5

Facteur21

FerrassieK4

Width (mm)Local:

Mean 25·5 28·5 27·8S.D. 5·8 7·8 6·3N 96 28 152

Distant:Mean 26·3 30·3 27·1S.D. 6·2 4·4 3·6N 19 11 17

Thickness (mm)Local:

Mean 9·7 9·1 12·4S.D. 2·9 3·3 4·7N 96 34 164

Distant:Mean 10·3 9·3 12·0S.D. 3·2 1·6 4·4N 18 13 19

noted by Demars (1982, 1990a,b) in earlyAurignacian assemblages within thePérigord and the Brive Basin to the east.Data compiled by Turq (1991) from adozen early Aurignacian sites in the north-ern Périgord indicated similar increases inBergerac materials that were interpreted byDibble et al. (1995) as evidence of directprocurement.

The increased presence of distantmaterials during the early Aurignacian is amatter of particular interest. Demars hassuggested that the comparatively largeamounts of Bergerac chert—between 10 and30%—found in early Aurignacian assem-blages reflect the suitability of this materialfor the production of large and thick blades.However, metric data suggest that dimen-sions of blade blanks made on local anddistant materials were similar. Blades ondistant materials were generally wider andthicker but the differences were statisticallyrandom ones. Local and distant blades inthe early Aurignacian were slightly wider atLe Facteur level 21 than in La Ferrassielevel K6 (Table 6). A preference for widerblades on distant materials is indicated butlocal materials also supported blades of

comparable width and thickness, so sizeselection alone does not explain shiftingproportions of distant materials.

Higher percentages of type tools bynumerical count were made on blade blanksduring the early Aurignacian in La Ferrassielevel K6 (73%) and Le Fecteur level 21

107

0

25

Per

cen

tage

of

dist

ant

mat

eria

ls

Ferrassie K6

20

10

15

5

Facteur 21 Facteur 19Ferrassie K4

Flakes Blades

Figure 13. Percentages of blade and flake tools made on distant materials. The percentages arecomparable, with the possible exception of La Ferrassie level K4.

(82%) than in the later Aurignacian depositsfrom La Ferrassie level K4 (60%) and LeFacteur level 19 (50%). Distant materialswere imported primarily (but not exclu-sively) as blades, yet these materials werepresent in approximately the same percent-ages for blade and flake tools (Figure 13).Thus blades made on distant materialsare numerically more frequent but thedifferences are hardly dramatic enough toaccount for increased quantities of distantmaterials in the early Aurignacianassemblage of Le Facteur level 21.

Variable access to lithic resources throughtime may have influenced material propor-tions. Morala & Turq (1990) suggested acorrelation may exist in Lot-et-Garonnebetween the exploitation of locally availableraw material sources and climatic con-ditions. Material with unaltered cortex, con-sidered to be of good quality, may be foundin erosion deposits at the base of cliffs andwas accessible during colder periods. Thesesame cliff bases would have been coveredwith colluvium during warmer periods, soplateau deposits of material with alteredcortex and of variable quality were morefrequently exploited.

However, raw material availability wasprobably not a factor in the increased per-centages of distant materials. As mentionedabove, comparable core sizes at La Ferrassieand similar widths of blade blanks do notsuggest differences in the ‘‘quality’’ of rawmaterials available during the Aurignacian.Level K4 at La Ferrassie, which wasdeposited during slightly warmer conditions,reflects an increase in the natural qualityof locally available cores (i.e., a smallerpercentage with coarse inclusions thatnecessitated discard), suggesting a differentraw material trajectory in the Périgord thanthat noted in Lot-et-Garonne to the southby Morala & Turq (1990).

Since most sources agree that lithic rawmaterial acquisition was related at some levelto movement to sources, it is suggested thatfactors influencing group mobility may cor-relate with procurement of raw materials.The structure of the subsistence environ-ment is one of those factors. Studies of theassociations between faunal and lithicassemblages have proven to be fruitfulavenues of hunter-gatherer research formany geographic areas and time periods(for example, Clark, 1989; Jochim, 1989;

108 . .

Table 7 Large mammal (NISP) quantities

FaunaPataud

14Pataud

13Pataud

12Pataud

11Ferrassie

L1, K6–K4Ferrassie

K3–JPataud

7

Rangifer tarandus 1481 221 129 599 171 10 395Bos primigenius/Bison priscus 2 7 4 37 80 10Bos/Equus 3Equus caballus 1 37 359 7 5 60Equus hydruntinus 3 6Mammuthus primigenius 1 2Megaloceros giganteus? 1Sus scrofa 1 1 21 3Cervus elaphus 10 3 2 3 29 80Capreolus capreolus 4 7Capra ibex 1 1 2Rupicapra rupicapra 2 4

Ungulate N 1495 224 177 966 221 155 567Crocuta spelaea 1Ursus spelaea 1Canis lupus 4 7 1Vulpes vulpes 4 2 62 1Alopex/Vulpes 2 1

Total N 1499 224 183 1036 225 156 568

Sources: Bouchud, 1975: Table XXXIII; Delpech, 1984: Table 26-5.

Rolland & Dibble, 1990; Bamforth, 1991;Straus, 1991; Stiner & Kuhn, 1992;Lieberman & Shea, 1994). Goebel (1997)has noted a relationship between greaterfaunal specialization (i.e., fewer species withan emphasis upon reindeer) and a morediverse suite of lithic raw materials duringlate Upper Paleolithic (27–20 ka) occu-pations in Siberia. Straus (1991) suggestedthat an association may be drawn betweendiffering exploitations of local raw materialsources and subsistence shifts from huntingof reindeer and other gregarious, migratoryprey such as horse and bison during theMagdalenian to pursuit of less migratoryspecies such as red deer, roe deer, and boarduring the Azilian at the Abri Dufaure inextreme southwestern France.

Spatial variation in faunal remains isapparent between Abri Pataud and LaFerrassie (Table 7); large mammal remainsdid not survive for the most part at LeFacteur. The river valley locus of AbriPataud contained a series of occupation

levels, with reindeer (Rangifer tarandus)always dominating the number of individualspecimens (NISP) counts (Bouchud, 1975;Spiess, 1979). Minimum number countswere reported by Spiess (1979); forexample, level 14 yielded a minimum of16 reindeer, two red deer (Cervus elaphus),and one each of the following: bovid[aurochs (Bos primigenius) or bison (Bisonpriscus)], horse (Equus caballus), and ibex(Capra ibex). Skeletal elements werereported for reindeer from most levels atAbri Pataud (Bouchud, 1975). Cranialbones and elements of the axial skeletonwere poorly represented. The appendicularskeleton—particularly the meat andmarrow-rich lower limb portions (Enloe,1993)—was evidently transported to theshelter after disarticulation, since the low-utility joint bones were rarely encountered.

The interfluvial location of La Ferrassiewas probably closer to forested habitats,particularly during the later Aurignacian.Reindeer bones were numerically dominant

109

1 3

3.5

Faunal richness (number of species)

Fau

nal

eve

nn

ess

(Sim

pson

's in

dex)

3

2

2.5

1.5

Pat 12Fer L1-K4

Early Aurig. Later Aurig.

5 7 9

r = 0.770

P < 0.05

Pat 13 Pat 14

Pat 11

Fer K3-J

Pat 7

Figure 14. A correlation of the two measures of diversity—richness and evenness—for early and laterAurignacian assemblages at Abri Pataud and La Ferrassie. Although a relationship of apparent significanceis suggested, it should be remembered that these are related variables since richness is also reflected to acertain extent in the calculation of Simpson’s Index of Diversity. Abbreviations as in Figure 5.

during the early Aurignacian, which, for thepurposes of faunal analysis (Delpech, 1984),included level K4. Horse was never anumerically prominent element, but bovidbones were present. By contrast, the faunalassemblage from levels K3–J, with a domi-nance of bovids, the presence of red deer,wild boar (Sus scrofa), and roe deer(Capreolus capreolus), and the virtual dis-appearance of reindeer, bears a markedresemblance to those associated withMesolithic occupations in northern Europefollowing the last glaciation (P. Crabtree,personal communication, 1997).

A temporal pattern of increasing numbersof species (richness) and greater equality ofrepresentation among those species (even-ness) during the later Aurignacian is indi-cated (Figure 14). Further, an increasedpresence of those animals considered lessmobile and associated with more closed,forested habitats such as wild boar, red deer,and roe deer (Spiess, 1979; White, 1985;Gordon, 1988; Pike-Tay, 1991) argues for abroadening of the subsistence base arisingfrom environmental change.

This temporal pattern is reinforced uponconsideration of a regional perspective.Boyle (1990) indicated that reindeergenerally dominated faunal assemblagesassociated with the early Aurignacian insouthwestern France, particularly in theDordogne and Vézère Valleys. The reducedpercentage of reindeer and increased pres-ence of red deer, roe deer, and wild boarobserved at La Ferrassie are reflected atother sites in and beyond the Périgord dur-ing the later Aurignacian; Boyle interpretedthe augmented presence of these speciesas indicative of increasing environmentaldiversity.

It is suggested that changes in thefaunal assemblages and lithic raw materialproportions were related phenomena. Acomparison of lithic and faunal data revealsthat changing percentages of distant lithicmaterials may be correlated with diversity infaunal assemblages (Figure 15). The earlyAurignacian assemblages at Abri Pataudand, to a lesser extent, at La Ferrassie,reflected low faunal diversities and weredominated by reindeer. Early Aurignacian

110 . .

125

3.5

Distance material percentage

Fau

nal

eve

nn

ess

(Sim

pson

's in

dex)

0

3

2

2.5

1.5

Fer K3-J

Pat 7

Pey F

Pat 11

Pat 14

5 10 15 20

Early Aurig. Later Aurig.

Figure 15. Percentages of tools on distant materials and faunal diversity from Abri Pataud and LaFerrassie. (Data for Abri Pataud are derived from Brooks, 1979 and Bondon, 1993). It should be notedthat the material percentage from a blade endscraper sample excavated with Peyrony’s ‘‘Aurignacian I’’level F at La Ferrassie is substituted for that indicated in Delporte’s level K6. Abbreviations as in Figure 5.

groups procured a more limited range ofmobile fauna and consequently movedbeyond the local foraging area, obtainingand transporting materials from Bergeracand other areas. Distant material percent-ages are elevated at Abri Pataud and LeFacteur, and possibly at La Ferrassie.

During the later Aurignacian, reindeerbecame less dominant at Abri Pataud andvirtually disappear from the faunal assem-blage at La Ferrassie. Faunal diversityincreased at both sites, dramatically so at LaFerrassie. Distant lithic material percentageswere low at La Ferrassie, Le Facteur, and,based upon samples (Brooks, 1995), at AbriPataud. These later Aurignacian groupspursued a greater diversity of fauna withinan expanded faunal base, which permitted orrequired them to move for the most partwithin a smaller geographic area. The con-tinued procurement of lithics from distantsources is indicated, but groups evidentlyjourneyed to those distant areas less fre-quently or at least transported less materialfrom them.

The proposed association between sub-sistence strategies and raw material procure-

ment argues that groups pursuing a limitedrange of mobile fauna would have increasedquantities of distant lithic materials. Thecomparisons are based upon a limitednumber of assemblages, which for the mostpart conform to the expectations of aninverse relationship between faunal diversityand percentage of distant materials. Datarelating to La Ferrassie level K6 arevaried; artefacts recovered during Delporte’sexcavations indicated low percentagesof distant materials, while a sample oftools recovered by Peyrony from the‘‘Aurignacian I’’ levels reflected distantmaterial percentages similar to those fromthe early Aurignacian at Abri Pataud andLe Facteur. Further, data previously citedindicate that high proportions of materialfrom Bergerac (Demars, 1982, 1990a,b;Turq, 1991) and a dominance of reindeer(Boyle, 1990) were both characteristic of theearly Aurignacian throughout the Périgord,suggesting the association has a broaderregional basis.

The importance of topographic andseasonal variables in the comprehension ofAurignacian mobility must be recognized.

111

Delpech (1983) and White (1985) haveargued that the dominance of reindeer atAbri Pataud compared with the predomi-nance of bovids and red deer at LaFerrassie reflect both differences in geogra-phy between riverine and interfluvial set-tings and possibly season of occupation.Seasonal data indicated fall and winteroccupation throughout the Aurignacianand Perigordian sequence at Abri Pataud(Spiess, 1979), while evidence of springand summer exploitation of red deer dur-ing the Perigordian at La Ferrassie wasnoted by Pike-Tay (1991). These seasonalanalyses represent the strongest direct indi-cators of mobility during the early UpperPaleolithic in the Vézère Valley, althoughthe extent to which the riverine/interfluvialpattern may fully apply to the Aurignacianoccupation at La Ferrassie must awaitfurther study.

The geographic distribution of distantsources that were exploited by Aurignacianpopulations in the Vézère Valley providessome indication of the trajectory of move-ment. An east–west orientation along theriver valleys is suggested by the Bergeraccherts and the possible jasper sources, whilea north–south orientation across the inter-fluvial and upland terrains is reflected in thematerials from Fumel and Gavaudun. Theseaxes may correspond to foraging activitiesfocused upon animal herds migrating withinthe river valleys between the Massif Centralhighlands and the Atlantic coastal plainlowlands. Movement into interfluvial areasto exploit different habitats, possibly on aseasonal basis, is indicated as anothercomponent within the overall Aurignaciansettlement system.

An adjustment in mobility patterns inresponse to a changing subsistence structureshould not be interpreted as a behaviorassociated exclusively with anatomicallymodern humans or the Upper Paleolithic.Rolland & Dibble (1990) have observedthat aspects of Middle Paleolithic lithic

assemblage variability may be correlatedwith faunal exploitation and environmentalvariation in southwestern France. Theyargue that Charentian assemblages reflectintensive use of lithic resources arising fromreduced mobility in pursuit of ‘‘aggregatedresources’’ dominated by reindeer duringcolder oxygen-isotope stage 4 conditions.Milder environments during early oxygen-isotope stage 3 generated an expansion ofwoodland and mixed habitats and more‘‘dispersed resources’’ that included horseand bison. Archaeological assemblagesdeposited within these environments(Denticulate and some Typical Mousterian)reflect less intensive use of lithic materialsand are inferred to have been associatedwith increased local mobility. Rolland& Dibble thus emphasized an associ-ation between environment, subsistenceresources, and lithic utilizaton, althoughtheir expectations of reduced mobility inpursuit of migratory reindeer during colderclimates differs from those suggested by theresearch discussed in this paper.

Féblot-Augustins (1993) also argued thatmobility adjustments during the MiddlePaleolithic in the Aquitaine Basin ofsouthwestern France were influenced byenvironmental conditions. The presence of‘‘resident’’ faunal species such as roe deerand boar are common in assemblagesassociated with the early portion of theWeichselian glaciation. More mobile speciessuch as reindeer, as well as red deer and ibex,appear later. Increased raw material importsfrom distances beyond 60 km coincidewith the appearance in the archaeologicalassemblages of mobile fauna and,for Féblot-Augustins, reflect consequentchanges in group mobility.

A case for indirect procurement?

If lithic raw material procurement was influ-enced to a substantial degree by the behaviorand diversity of prey species, one would not

112 . .

expect to find archaeological assemblageswith diverse, nonmobile fauna and highpercentages of distant lithic materials. Thecoexistence of elevated distant materialpercentages and high faunal diversity maytherefore reflect the influence of techno-logical selectivity or various forms ofsocial exchange upon raw material procure-ment. The lack of such assemblagesamong those studied does not indicate thatindirect social exchange was absent duringthe Aurignacian, but simply that wecannot in this instance account for suchmechanisms in the procurement of lithicsfor tools.

The existence of socially-directed systemsmay indeed be indicated if we consider thepresence of other raw materials. The UpperPaleolithic in general and the Aurignacian inparticular manifested abundant evidence ofexotic material acquisition that implies theexistence of mobility on a greatly expandedscale or the development of exchange net-works. The existence of intensified socialbehavior and possible exchange networksduring the Aurignacian was proposed byWhite (1989a,b) based upon the use ofsteatite from the Massif Central to the eastand exotic marine shells, in addition tocarnivore teeth and mammoth ivory, as bodyornamentation during some of the earliestAurignacian occupations in the Périgord.These characteristics are dramatically pro-nounced at the site complex in the Vallonde Castel-Merle (Abri Blanchard, AbriCastanet, and Abri de la Souquette) locatedapproximately 4 km up the Vézère fromLe Facteur. Current excavations at theAbri Castanet have produced two radio-metric determinations on bone collagen ofca. 35 ka for the early Aurignacian level(R. White, personal communication,1998).

Taborin (1993a) noted that 15Aurignacian sites containing shells havebeen found in the Périgord; ten lie within aradius of a few kilometers in or near the

Vézère Valley. The richest shell assemblageswere accumulated during the earlyAurignacian, in association with split and‘‘simple’’ based points, numerous blades,and endscrapers. An Atlantic coastalspecies was present at the Abri Pataud inlevels 7, 14, and the éboulis depositbetween levels 13 and 14 (Dance, 1975).Atlantic species and fossil Miocene shellsare associated with the Peyrony assem-blages from La Ferrassie. A sea urchinfossil was recovered from Le Facteurlevel 21 during the Delporte excavations.Mediterranean shells are much rarer in theAurignacian Périgord, but do occur atBlanchard, Castanet, and La Combe(Taborin, 1993a).

These data provide some indication of thescale of Aurignacian regional contacts,although the scope of such ‘‘regionalism’’during the Upper Paleolithic clearly variedboth spatially and temporally throughoutEurope. Long-distance movements ofBergerac cherts out of the Périgord havebeen suggested in Charente to the west(Demars, 1990) and the Ariège to the south(Simonnet, 1982). These data are interest-ing since the former reinforce the shell evi-dence of movement to or contact with theAtlantic coastal plain while the latter maysuggest a trajectory oriented to the Pyrénéesor the Mediterranean.

White (1989c) contended that a centri-petal movement of materials into thePérigord occurred at times during UpperPaleolithic, which suggested variousmeans of procurement, including aggre-gation and dispersal, logistical directacquisition, and exchange. As indicatedabove, Taborin (1993a) commented thatthe richest shell collections in the Périgordwere associated with early Aurignacianassemblages. This observation mayreflect increased opportunities forexchange between mobile groups in thegenerally open environments prevailingduring the early Aurignacian.

113

SummaryA comparative analysis of lithic raw materialeconomy during Aurignacian occupation ofAbri Pataud, Le Facteur, and La Ferrassieindicates consistent differences in theutilization of locally available cherts com-pared with those obtained from sources atgreater distances from the sites. All stages oflithic reduction were reflected among thelocal materials, although emphasis wasplaced upon secondary core reduction andblade production. Distant materials weregenerally introduced as blade blanks andretouched blades.

The various trajectories of lithic procure-ment suggest movement along the majorriver valleys and between valleys acrossupland plateaus. Such movement mayreflect exploitation of these settings duringdifferent seasons. Indeed, indications of fall–winter occupations at riverine loci such asAbri Pataud and Le Flageolet and spring–summer habitations at the interfluvialshelter of La Ferrassie (Spiess, 1979; Pike-Tay, 1991) constitute the strongest sugges-tions of group mobility during the earlyUpper Paleolithic in the Périgord, althoughonly the data from Abri Pataud were derivedfrom Aurignacian deposits.

However, comparisons of faunal dataand evidence of variability within lithic econ-omies suggest that a more general patternmay be reflected within the Aurignacian.This pattern is specifically manifested in anassociation between the subsistence base,lithic raw materials, and—somewhat moretenuously—reduction intensity. During theearly Aurignacian, faunal profiles weredominated by reindeer and overall diversitywas low. The proportions of the tool assem-blage made on distant materials rangedbetween 15 and 25% at Le Facteur and AbriPataud, and possibly at La Ferrassie. LaterAurignacian assemblages reflect the pursuitof a diverse range of fauna, increasedemphasis upon locally available materials,and intensified lithic reduction. This pattern

suggests more frequent mobility beyond thelocal lithic resource area during the earlyAurignacian when faunal procurement wasfocused upon a mobile species. Reindeerremained dominant during the fall andwinter occupation at the Abri Pataud, butwere supplanted by bovids and red deer atLa Ferrassie during the later Aurignacian.The additional procurement of fauna associ-ated with more closed environments duringthe later Aurignacian may indicate a greateremphasis upon foraging within a smallergeographic range, which may be reflected inhigher proportions of locally-available lithicmaterials.

These observations emphasize the valueof actively articulating varied aspects of thearchaeological record, certainly includingbut not limited to faunal data, with theinterpretation of lithic economy. Further,important insights emerge pertaining toearly modern human variability in settle-ment patterns within annual cycles andbroader climatic periods. The suggestionthat patterns of settlement mobility wererelated to changes in the structure of subsist-ence resources may account for some ofthe typological variability that has beentraditionally employed in defining earlyand later Aurignacian assemblages.Blumenschine & Peters (1998) have recentlyargued for an ‘‘explicit ecological orien-tation’’ to the study of stone artefacts andanimal bones within an archaeological con-text. Although their specific interestsfocused upon spatial relationships and land-scape ecology, the argument certainly mayencompass a range of analytical perspectivesderived from consideration of the lithic andfaunal components of an archaeologicalassemblage.

The variations in Aurignacian mobilitypatterns proposed herein may not constitutea basis for suggesting differences in evol-utionary behavior compared with theMiddle Paleolithic since similar correspond-ences between lithic economy and faunal

114 . .

resources have been suggested forMousterian assemblages in southwesternFrance. A more complicated image of earlyUpper Paleolithic settlement emerges if con-sideration of the Aurignacian material realmis expanded to encompass the nonlithicmedia that reflect evidence of symbolicbehavior. An important difference betweenthe Middle Paleolithic and Upper Paleo-lithic within the portion of the Périgordunder examination was not necessarily thebasic mobility strategy in response tosubsistence resources or environmental con-ditions but the manner in which that mobil-ity was integrated into the broader socialstructure. The lithic materials reduced toform tools were obtained for the most part inthe immediate vicinity of the rock sheltersites. The variable presence of lithic ma-terials from greater distances (25–45 km)may be related to shifts in mobilityfrequency arising from variations in thesubsistence environment. Exotic materialssuch as shells and steatite—often utilizedin symbolic contexts related to bodyornamentation—were acquired throughdirect transport or indirect social exchangeover distances of several hundred kilo-meters. Direct and indirect means of pro-curement may thus simultaneously accountfor different aspects of the Aurignacianmaterial record. Mobility changes, evidentlyin response to differing conditions in thenatural environment, may have facilitated orbeen amplified by the intensification that isapparent in aspects of the Aurignacian socialenvironment.

Acknowledgements

I wish to express sincere appreciation for theassistance and advice that were providedduring the course of this study byJean-Pierre Chadelle, Pam Crabtree, HaroldDibble, Jean-Michelle Geneste, TerryHarrison, Heidi Knecht, André Morala,Jean-Philippe Rigaud, Alain Turq, and

especially Randy White and JacquesPelegrin. Any errors of fact or fancy aremy responsibility. Generous financial andlogistical support was provided by theNational Science Foundation (SBR-9311880), New York University, and theMaison Sugar in Paris. M Henri Delportevery kindly granted permission to studythe collections from La Ferrassie and LeFacteur at the Musée des AntiquitésNationals in Saint-Germain-en-Laye.Research at Saint-Germain was greatlyaided by Mme Marie-Hélène Marino andthe late Dominique Buisson. ProfesseurHenry de Lumley and Mme Marie Perpèregenerously granted permission to work withportions of the collections at the Muséede Préhistoire de l’Abri Pataud in LesEyzies, and this research was considerablyfacilitated by the assistance of Mme BrigitteDelluc. M Jean-Jacques Cleyet-Merlekindly granted permission to examine thePeyrony collections from La Ferrassie atthe Musée National de Préhistoire in LesEyzies.

References

Aitken, M. J., Stringer, C. & Mellars, P. (1993).The Origin of Modern Humans and the Impact ofChronometric Dating. Princeton: Princeton UniversityPress.

Allsworth-Jones, P. (1986). The Szeletian and theTransition from Middle to Upper Paleolithic in CentralEurope. Oxford: Clarendon Press.

Bamforth, D. (1991). Population dispersion andPaleoindian technology at the Allen site. In (A.Montet-White & S. Holen, Eds) Raw MaterialEconomies Among Prehistoric Hunter-Gatherers,pp. 357–397. Lawrence: University of KansasPublications in Anthropology No. 19.

Bar-Yosef, O. (1991). Raw material exploitation in theLevantine Epi-Paleolithic. In (A. Montet-White &S. Holen, Eds) Raw Material Economies Among Pre-historic Hunter-Gatherers, pp. 235–250. Lawrence:University of Kansas Publications in AnthropologyNo. 19.

Bernaldo de Quirós, F. & Cabrera Valdés, V. (1993).Early Upper Paleolithic industries of CantabrianSpain. In (H. Knecht, A. Pike-Tay & R. White,Eds) Before Lascaux: The Complex Record of theEarly Upper Paleolithic, pp. 57–69. Boca Raton: CRCPress.

115

Binford, L. (1979). Organization and formation pro-cesses: looking at curated technologies. J. Anthropol.Res. 25, 255–273.

Binford, L. (1980). Willow smoke and dogs’ tails:Hunter-gatherer settlement systems and archaeologi-cal site formation. Am. Antiq. 45, 4–20.

Binford, L. (1989). Isolating the transition to culturaladaptations: an organizational approach. In (E.Trinkaus, Ed.) The Emergence of Modern Humans:Biocultural Adaptations in the Later Pleistocene,pp. 18–41. Cambridge: Cambridge University Press.

Bischoff, J., Soler, N., Maroto, J. & Julia, R. (1989).Abrupt Mousterian/Aurignacian boundary atca. 40 ka BP: Accelerator radiocarbon dates fromL’Arbreda Cave (Catalunya, Spain). J. Archaeol. Sci.16, 563–576.

Blumenschine, R. & Peters, C. (1998). Archaeologicalpredictions for hominid land use in the paleo-Olduvai Basin, Tanzania, during lowermost Bed IItimes. J. hum. Evol. 34(6), 565–607.

Bondon, N. (1993). L’étude de l’industrie lithique aurig-nacienne (couche 14) de l’Abri Pataud (Les Eyzies-de-Tayac, Dordogne). Paris: Memoire de DEA, Museumd’Histoire Naturelle, Institut de PaléontologieHumaine.

Bouchud, J. (1968). L’abri du Facteur à Tursac(Dordogne). La faune et sa signification climatique.Gallia Préhistoire 11(1), 113–121.

Bouchud, J. (1975). Etude de la faune de l’abri Pataud.In (H. L. Movius, Jr, Ed.) Excavation of theAbri Pataud, Les Eyzies (Dordogne), pp. 69–153.Cambridge: Peabody Museum, Harvard University.

Boyle, K. (1990). Upper Paleolithic Faunas from South-West France: A Zoogeographic Perspective. Oxford:British Archaeological Reports International SeriesNo. 557.

Bräuer, G. (1989). The evolution of modern humans:A comparison of the African and non-Africanevidence. In (P. Mellars & C. Stringer, Eds) TheHuman Revolution: Behavioural and BiologicalPerspectives in the Origins of Modern Humans,pp. 123–154. Princeton: Princeton University Press.

Bricker, H. M. (1975). Provenience of flint used forthe manufacture of tools at the Abri Pataud, LesEyzies (Dordogne). In (H. L. Movius, Ed.) Ex-cavation of the Abri Pataud, Les Eyzies (Dordogne),pp. 194–197. Cambridge: Peabody Museum,Harvard University.

Bricker, H. M., Brooks, A., Clay, R. B. & David, N.(1995). Les fouilles de H. L. Movius Jr à l’abriPataud: généralités. In (H. M. Bricker, Ed.) LePaléolithique supérieur de l’abri Pataud (Dordogne): lesfouilles de H. L. Movius Jr, pp. 11–29. Paris: Editionsde la Maison des Sciences de l’Homme.

Brooks, A. (1979). The significance of variability inPaleolithic assemblages: An Aurignacian examplefrom southwestern France. Ph.D. Dissertation,Harvard University.

Brooks, A. (1982). Aurignacian assemblages from AbriPataud (Dordogne, France). In Aurignacien etPérigordien en Europe, pp. 92–104. Liège: Universitéde Liège.

Brooks, A. (1995). L’Aurignacien de l’abri Pataudniveaux 6 à 14. In (H. M. Bricker, Ed.) Le Paléo-lithique supérieur de l’abri Pataud (Dordogne): les fouillesde H. L. Movius Jr, pp. 167–222. Paris: Editions de laMaison des Sciences de l’Homme.

Cabrera Valdés, V. & Bernaldo de Quirós, F. (1991).L’Aurignacien de Castillo et en région cantabrique.Paper presented at the XIIth Congress, l’UnionInternationale des Sciences Préhistoriques etProtohistoriques, Bratislava.

Cabrera Valdés, V., Hoyos Gómez, M. & Bernaldo deQuirós Guidotti, F. (1997). The transition fromMiddle to Upper Paleolithic in the Cave of ElCastillo (Cantabria, Spain). In (G. Clark & C.Willermet, Eds) Conceptual Issues in Modern HumanOrigins Research, pp. 177–188. New York: Aldine deGruyter.

Chadelle, J.-P. (1983). Technologie et utilisation du silexau Périgordien supérieur: l’example de la couche VIII duFlageolet I. Toulouse: Ecole des Hautes Etudes enSciences Sociales.

Chadelle, J.-P. (1990). Le site de plein air de Corbiac-Vignoble à Bergerac (Dordogne): technologielithique et mode d’occupation. In (M. R. Séronie-Vivien & M. Lenoir, Eds) Le silex de sa genèse à l’outil,volume II, pp. 385–390. Paris: CNRS.

Clark, G. (1989). Romancing the stones: Biases, style,and lithics at La Riera. In (D. Henry & G. Odell,Eds) Alternative Approaches to Lithic Analysis, pp.27–50. Washington, DC: American AnthropologicalAssociation.

Clottes, J. (1996). Epilogue: Chauvet Cave today. In(J.-M. Chauvet, E. Deschamps & C. Hillaire, Eds)Dawn of Art: The Chauvet Cave, pp. 89–127.New York: Abrams.

Collins, M. (1975). Lithic technology as a means ofprocessual inference. In (E. H. Swanson, Ed.) LithicTechnology: Making and Using Stone Tools, pp. 15–34.Chicago: Aldine Press.

Dance, P. (1975). The molluscan fauna. In (H. L.Movius, Ed.) Excavation of the Abri Pataud, Les Eyzies(Dordogne), pp. 154–159. Cambridge: PeabodyMuseum, Harvard University.

D’Errico, F., Zilhão, J., Julien, M., Baffier, D. &Pelegrin, J. (1998). Neanderthal acculturation inwestern Europe? A critical review of the evidence andits interpretation. Curr. Anthrop. 39 Supplement,S1–S44.

Delibas, G. (1984). La datation par le carbone 14 desossements de la Ferrassie. In (H. Delporte, Ed.) Legrand abri de la Ferrassie, pp. 105–107. Paris: Institutde Paléontologie Humaine.

Delluc, B. & Delluc, G. (1978). Les manifestationsgraphiques aurignaciennes sur support rocheuxdes environs des Eyzies (Dordogne). Gallia Préhistoire21.

Delpech, F. (1983). Les faunes du Paléolithique supérieurdans le Sud-Ouest de la France. Paris: CNRS.

Delpech, F. (1984). La Ferrassie: Carnivores, artio-dactyles et périssodactyles. In (H. Delporte, Ed.) Legrand abri de La Ferrassie, pp. 61–89. Paris: Institutde Paléontologie Humaine.

116 . .

Delpech, F. (1993). The fauna of the Early UpperPaleolithic: Biostratigraphy of large mammals andcurrent problems in chronology. In (H. Knecht,A. Pike-Tay & R. White, Eds) Before Lascaux:The Complex Record of the Early Upper Paleolithic,pp. 71–84. Boca Raton: CRC Press.

Delporte, H. (1968). L’abri du Facteur à Tursac.Etude générale Gallia Préhistoire 11(1), 1–112.

Delporte, H. (1991). La séquence aurignacienne etpérigordienne sur la base des travaux récentsréalisés en Périgord. Bull. Soc. Préhist. Française 88,243–256.

Delporte, H. (Ed.) (1984). Le grand abri de la Ferrassie.Paris: Institut de Paléontologie Humaine.

Delporte, H., Djindjian, F. & Mazière, G. (1983).Etudes sur l’Aurignacien de la Ferrassie. InAurignacien et Périgordien en Europe, pp. 13–21. Liège:Université de Liège.

Delporte, H., Mazière, G. & Djindjian, F. (1977).L’Aurignacien de la Ferrassie. Observations prélimi-naires à la suite de fouilles récentes. Bull. Soc. Préhist.Française 74, 343–357.

Demars, P.-Y. (1982). L’utilisation du silex au Paléo-lithique supérieur. Choix, approvisionnement, circulation:l’example du Bassin de Brive. Paris: CNRS.

Demars, P.-Y. (1990a). Les interstratifications entreAurignacien et Châpelperronien à Roc-de-Combe etau Piage (Lot). Approvisionnement en matièrespremières et position chronologique. In (C. Farizy,Ed.) Paléolithique moyen récent et Paléolithique supérieurancien en Europe, pp. 235–239. Nemours: Mémoiresdu Musée de Préhistoire d’Ile de France No. 3.

Demars, P.-Y. (1990b). L’economie du silex àLaugerie-Haute (Dordogne). In (M. R. Séronie-Vivien & M. Lenoir, Eds) Le silex de sa genèse à l’outil,volume II, pp. 373–383. Paris: CNRS.

Dibble, H. (1988). Typological aspects of reductionand intensity of utilization of lithic resources in theFrench Mousterian. In (H. Dibble & A. Montet-White, Eds) Upper Pleistocene Prehistory of WesternEurasia, pp. 188–191. Philadelphia: UniversityMuseum Press.

Dibble, H. & Holdaway, S. (1993). The Middle Paleo-lithic of Warwasi Rockshelter. In (D. Olszewski &H. Dibble, Eds) The Paleolithic Prehistory of theZagros, pp. 75–99. Philadelphia: University MuseumPress.

Dibble, H., Roth, B. & Lenoir, M. (1995). The use ofraw materials at Combe-Capelle Bas. In (H. Dibble,B. Roth & M. Lenoir, Eds) The Middle PaleolithicSite of Combe-Capelle Bas (France), pp. 259–287.Philadelphia: University Museum Press.

Enloe, J. (1993). Subsistence organization in the EarlyUpper Paleolithic: Reindeer hunters of the Abri duFlageolet, couche V. In (H. Knecht, A. Pike-Tay &R. White, Eds) Before Lascaux: The Complex Record ofthe Early Upper Paleolithic, pp. 101–115. Boca Raton:CRC Press.

Farrand, W. (1975). Analysis of the Abri Pataudsediments. In (H. L. Movius, Ed.) Excavation ofthe Abri Pataud, Les Eyzies (Dordogne), pp. 27–68.Cambridge: Peabody Museum, Harvard University.

Farrand, W. (1995). Etude sédimentologique duremplissage de l’abri Pataud. In (H. M. Bricker, Ed.)Le Paléolithique supérieur de l’abri Pataud (Dordogne):les fouilles de H. L. Movius Jr, pp. 31–65. Paris:Editions de la Maison des Sciences de l’Homme.

Féblot-Augustins, J. (1993). Mobility strategies in thelate Middle Paleolithic of central Europe and westernEurope: elements of stability and variability. J.Anthropol. Archaeol. 12, 211–265.

Féblot-Augustins, J. (1997). Middle and Upper Paleo-lithic raw material transfers in western and centralEurope: assessing the pace of change. J. MiddleAtlantic Archaeol. 13, 57–90.

Frayer, D. (1997). Perspectives on Neanderthals asancestors. In (G. Clark & C. Willermet, Eds) Con-ceptual Issues in Modern Human Origins Research,pp. 220–234. New York: Aldine de Gruyter.

Frayer, D., Wolpoff, M., Thorne, A., Smith, F. &Pope, G. (1993). Theories of modern humanorigins: the palaeontological test. Am. Anthrop. 95,14–50.

Gambier, D. (1989). Fossil hominids from the EarlyUpper Paleolithic (Aurignacian) of France. In (P.Mellars & C. Stringer, Eds) The Human Revolution:Behavioral and Biological Perspectives in the Origin ofModern Humans, pp. 194–211. Princeton: PrincetonUniversity Press.

Gambier, D. (1997). Modern humans at the beginningof the Upper Paleolithic in France: anthropologicaldata and perspectives. In (G. Clark & C. Willermet,Eds) Conceptual Issues in Modern Human OriginsResearch, pp. 117–131. New York: Aldine deGruyter.

Gamble, C. (1986). The Palaeolithic Settlement ofEurope. Cambridge: Cambridge University Press.

Garralda, M. (1997). The human paleontology of theMiddle to Upper Paleolithic transition on the IberianPeninsula. In (G. Clark & C. Willermet, Eds) Con-ceptual Issues in Modern Human Origins Research,pp. 148–160. New York: Aldine de Gruyter.

Geneste, J.-M. (1985). Analyse lithique d’industriesmoustériennes du Périgord: une approche tech-nologique du comportement des groupes humainsau Paléolithique moyen. Thèse Sc., Université deBordeaux I.

Geneste, J.-M. (1988). Systèms d’approvisionnementen matières premières au Paléolithique moyen et auPaléolithique supérieur en Aquitaine. In (J. K.Kozłowski, Ed.) L’homme de Néandertal, tome 8: lamutation, pp. 61–70. Liège: Université de Liège.

Geneste, J.-M. (1989). Economie des ressourceslithiques dans le Moustérian du Sud-Ouest de laFrance. In (L. Freeman & M. Patou, Eds) L’hommede Néandertal, tome 6: la subsistance, pp. 75–97. Liège:Université de Liège.

Geneste, J.-M. (1990). Développement des systèmesde production lithique au cours du Paléolithiquemoyen en Aquitaine septentrionale. In (C. Farizy,Ed.) Paléolithique moyen récent et Paléolithiquesupérieur ancien en Europe, pp. 203–213. Nemours:Mémoires du Musée de Préhistoire d’Ile de FranceNo. 3.

117

Goebel, T. (1997). Pleistocene human settlement ofSiberia. Paper presented at the 6th annual meeting,Paleoanthropology Society, St Louis.

Goodyear, A. (1979). An hypothesis for the use ofcryptocrystalline raw materials among Paleoindiangroups of North America. Reprinted in (C. Ellis & J.Lothrop, Eds) Eastern Paleoindian Lithic Resource Use,pp. 1–9. Boulder: Westview Press, 1989.

Gordon, B. (1988). Of Men and Reindeer Herds inFrench Magdalenian Prehistory. Oxford: BritishArchaeological Reports International Series No.390.

Gould, R. (1980). Living Archaeology. Cambridge:Cambridge University Press.

Henry, D. (1989). Correlations between reductionstrategies and settlement patterns. In (D. Henry &G. Odell, Eds) Alternative Approaches to LithicAnalysis, pp. 139–155. Washington, DC: AmericanAnthropological Association.

Jochim, M. (1989). Optimization and stone toolstudies: Problems and potentials. In (R. Torrence,Ed.) Time, Energy, and Stone Tools, pp. 106–111.Cambridge: Cambridge University Press.

Karavanic, I. (1995). Upper Paleolithic occupationlevels and late-occurring Neandertal at Vindija Cave(Croatia) in the context of central Europe and theBalkans. J. Anthropol. Res. 51, 9–35.

Karavanic, I. & Smith, F. (1998). The Middle/UpperPaleolithic interface and the relationship ofNeanderthals and early modern humans in theHrvatsko Zagorje, Croatia. J. hum. Evol. 34(3),223–248.

Klein, R. (1995). Anatomy, behavior, and modernhuman origins. J. World Prehist. 9, 167–198.

Knecht, H. (1993). Splits and wedges: The techniquesand technology of early Aurignacian antler work-ing. In (H. Knecht, A. Pike-Tay & R. White,Eds) Before Lascaux: The Complex Record of theEarly Upper Paleolithic, pp. 137–162. Boca Raton:CRC Press.

Kozłowski, J. K. (1973). The origin of lithic rawmaterial used in the Paleolithic of the Carpathiancountries. Acta Archaeologica Carpathica 13, 5–19.

Kozłowski, J. K. (1988). Transition from the Middle tothe Early Upper Paleolithic in central Europe and theBalkans. In (J. Hoffecker & C. Wolf, Eds) The EarlyUpper Paleolithic: Evidence from Europe and the NearEast, pp. 193–235. Oxford: British ArchaeologicalReports International Series No. 437.

Kozłowski, J. K. (1991). Raw material procurement inthe Upper Paleolithic of central Europe. In (A.Montet-White & S. Holen, Eds) Raw MaterialEconomies Among Prehistoric Hunter-Gatherers,pp. 187–196. Lawrence: University of KansasPublications in Anthropology No. 19.

Kozłowski, J. K. (Ed.) (1982). Excavation in theBacho Kiro Cave (Bulgaria). Final report. Warszawa:Panstwowe Wydawnictwo Naukowe.

Larick, R. (1983). The circulation of Solutrean foliatepoint cherts: residential mobility in the Périgord.Ph.D. Dissertation, State University of New York atBinghampton.

Laville, H. (1968). L’abri du Facteur à Tursac(Dordogne). Etude sédimentologique du remplis-sage. Gallia Préhistoire 11(1), 133–145.

Laville, H. (1975). Climatologie et chronologie duPaléolithique en Périgord: étude sédimentologique dedépots en grottes et sous abris. Paris: Laboratoirede Paléontologie Humaine.

Laville, H. & Tuffreau, A. (1984). Les dépots dugrand abri de la Ferrassie: stratigraphie, significationclimatique et chronologie. In (H. Delporte, Ed.) Legrand abri de la Ferrassie, pp. 25–50. Paris: Institut dePaléontologie Humaine.

Laville, H., Rigaud, J.-P. & Sackett, J. (1980). RockShelters of the Périgord-Geological Stratigraphy andArchaeological Succession. New York: Academic Press.

Le Brun-Ricalens, F. (1990). L’occupationaurignacienne d’Hui (Beauville, Lot-et-Garonne).Présentation, problématique et premiers résultats.Bull. Soc. Préhist. Française 87(9), 275–282.

Le Brun-Ricalens, F. (1993). Reflexions préliminairessur le comportement litho-technologique etl’occupation du territoire du Pays des Serres: legisement de ‘‘Toulousète’’ à Beauville (Lot-et-Garonne), une occupation mousterienne etaurignacienne de plein air. Paléo 3, 127–153.

Le Brun-Ricalens, F. & Neveu, J. (1991). Le gisementde la Moulinière à Caussade (Tarn-et-Garonne): uneoccupation aurignacienne de plein air. Bull. Soc.Méridionale Spéléol. Préhist. 31, 33–49.

Leroi-Gourhan, A. (1968). L’abri du Facteur à Tursac(Dordogne). Analyse pollinique. Gallia Préhistoire11(1), 123–132.

Lieberman, D. (1995). Testing hypotheses aboutrecent human evolution from skulls: integrating mor-phology, function, development, and phylogeny.Curr. Anthrop. 36, 159–197.

Lieberman, D. & Shea, J. (1994). Behavioral differ-ences between archaic and modern humans inthe Levantine Mousterian. Am. Anthrop. 96(2),300–332.

Lumley, H. de, Couraud, C., Delluc, B., Delluc, G.,Delporte, H., Leroy-Prost, C., Lumley, M.-A., de,Perpère, M. & Vialou, D. (1984). Art et civilisationsdes chasseurs de la préhistoire, 34,000–8,000 ans av.J.-C. Paris: Laboratoire de Préhistoire du Musée del’Homme and Musée des Antiquités Nationales deSaint-Germain-en-Laye.

Marks, A. (1988). The curation of stone tools duringthe upper Pleistocene: a view from the central Negev,Israel. In (H. Dibble & A. Montet-White, Eds) UpperPleistocene Prehistory of Western Eurasia, pp. 87–94.Philadelphia: University Museum Press.

Marks, A. (1993). The Early Upper Paleolithic: theview from the Levant. In (H. Knecht, A. Pike-Tay &R. White, Eds) Before Lascaux: The Complex Record ofthe Early Upper Paleolithic, pp. 5–21. Boca Raton:CRC Press.

Marks, A., Shokler, J. & Zilhao, J. (1991). Rawmaterial usage in the Paleolithic. The effects oflocal availability on selection and economy. In (A.Montet-White & S. Holen, Eds) Raw MaterialEconomies Among Prehistoric Hunter-Gatherers,

118 . .

pp. 127–139. Lawrence: University of KansasPublications in Anthropology No. 19.

Meignen, L. (1993). L’abri des Canalettes. Un habitatmoustérian sur les Grands Causses (Nant, Aveyron).Paris: CNRS.

Mellars, P. (1998). Comment to Neanderthalacculturation in western Europe? A critical review ofthe evidence and its interpretation. Curr. Anthrop. 39Supplement, S25–S26.

Mellars, P. & Stringer, C. (1989). Introduction. In (P.Mellars & C. Stringer, Eds) The Human Revolution:Behavioral and Biological Perspectives in the Origins ofModern Humans, pp. 1–14. Princeton: PrincetonUniversity Press.

Meltzer, D. (1989). Was stone exchanged among east-ern North American Paleoindians? In (C. Ellis & J.Lothrop, Eds) Eastern Paleoindian Lithic Resource Use,pp. 11–39. Boulder: Westview Press.

Miracle, P. & Crummett, T. (1995). Just who werethese ‘‘Aurignacians’’? Vindija’s place in the centralEuropean early Upper Paleolithic. Paper presented atthe 60th annual meeting, Society for AmericanArchaeology, Minneapolis.

Montet-White, A. (1991). Lithic acquisition,settlements, and territory in the Epigravettian ofcentral Europe. In (A. Montet-White & S. Holen,Eds) Raw Material Economies Among PrehistoricHunter-Gatherers, pp. 205–220. Lawrence: Universityof Kansas Publications in Anthropology no. 19.

Morala, A. (1984). Périgordien et Aurignacien en Haut-Agenais: étude d’ensembles lithiques. Toulouse: Ecoledes Hautes Etudes en Sciences Sociales.

Morala, A. & Turq, A. (1990). Les stratégiesd’exploitation du milieu minéral, du Riss àl’Holocene, en Haut-Agenais (Sud-Ouest de laFrance). In (M. R. Séronie-Vivien & M. Lenoir, Eds)Le silex de sa genèse à l’outil, volume II, pp. 385–390.Paris: CNRS.

Morrow, C. & Jeffries, R. (1989). Trade or embeddedprocurement?: a test case from southern Illinois. In(R. Torrence, Ed.) Time, Energy, and Stone Tools,pp. 27–33. Cambridge: Cambridge University Press.

Movius, H. L., Jr (1977). Excavation of the Abri Pataud,Les Eyzies (Dordogne). Stratigraphy. Cambridge:Peabody Museum, Harvard University.

Movius, H. L., Jr (1995). Inventaire analytique dessites aurignaciens et périgordiens de Dordogne. In(H. M. Bricker, Ed.) Le Paléolithique supérieur de l’abriPataud (Dordogne): les fouilles de H. L. Movius, Jr,pp. 227–314. Paris: Editions de la Maison desSciences de l’Homme.

Movius, H. L., Jr (Ed.) (1975). Excavation of the AbriPataud, Les Eyzies (Dordogne). Cambridge: PeabodyMuseum, Harvard University.

Munday, F. (1977). Intersite variability in theMousterian occupation of the Avdat/Aqev area. In(A. Marks, Ed.) Prehistory and Paleoenvironments inthe Central Negev, Israel, Vol. 1. The Avdat/Aqev Area,Part 1, pp. 113–140. Dallas: Southern MethodistUniversity Press.

Newcomer, M. (1971). Some quantitative experimentsin handaxe manufacture. World Archaeol. 3, 85–94.

Oliva, M. (1981). Die Bohunicien- Station bei Podoliund ihre Stellung im beginnenden Jungpaläo-lithikum. C{asopis Moravského Muzea sc. soc. 66,7–45.

Oliva, M. (1984). Le Bohunicien, un nouveau groupeculturel en Moravie: quelques aspects psycho-technologiques du dévelopment des industriespaléolithiques. L’Anthropol. 88, 209–220.

Oliva, M. (1993). The Aurignacian in Moravia. In(H. Knecht, A. Pike-Tay & R. White, Eds) BeforeLascaux: The Complex Record of the Early UpperPaleolithic, pp. 37–55. Boca Raton: CRC Press.

Otte, M. & Straus, L. (1995). Le Trou Magrite. Liège:Université de Liège.

Paquereau, M.-M. (1984). Etude palynologique dugisement de la Ferrassie (Dordogne). In (H.Delporte, Ed.) Le grand abri de la Ferrassie,pp. 51–59. Paris: Institut de Paléontologie Humaine.

Pelegrin, J. (1986). Technologie lithique: une méthodeappliquée à l’étude de deux séries du Périgordienancien—Roc de Combe couche 8—la Côte niveauIII. Dissertation: Lettres, Université de Paris X.

Pelegrin, J. (1990). Observations technologiques surquelques séries du Châtelperronian et du MTA B duSud-Ouest de la France. In (C. Farizy, Ed.) Paléo-lithique moyen récent et Paléolithique supérieur ancien enEurope, pp. 195–201. Nemours: Mémoires du Muséede Préhistoire d’Ile de France No. 3.

Peyrony, D. (1933). Les industries aurignaciennes dansle bassin de la Vézère. Aurignacien et Périgordien.Bull. Soc. Préhist. Française 30, 543–559.

Peyrony, D. (1934). La Ferrassie—Moustérien,Périgordien, Aurignacien. Préhistoire 3, 1–92.

Pike-Tay, A. (1991). Red Deer Hunting in the UpperPaleolithic of Southwest France: A Study in Seasonality.Oxford: British Archaeological Reports InternationalSeries No. 569.

Rensink, E., Kolen, J. & Spieksma, A. (1991). Patternsof raw material distribution in the upper Pleistoceneof northwestern and central Europe. In (A. Montet-White & S. Holen, Eds) Raw Material EconomiesAmong Prehistoric Hunter-Gatherers, pp. 141–159.Lawrence: University of Kansas Publications inAnthropology No. 19.

Rigaud, J.-P. (1989). From the Middle to the UpperPaleolithic: Transition or convergence? In (E.Trinkaus, Ed.) The Emergence of Modern Humans:Biocultural Adaptations in the Later Pleistocene, pp.142–153. Cambridge: Cambridge University Press.

Rigaud, J.-P. (1997). Scenarios for the Middle toUpper Paleolithic transition: A Europeanperspective. In (G. Clark & C. Willermet, Eds)Conceptual Issues in Modern Human Origins Research,pp. 161–167. New York: Aldine de Gruyter.

Rigaud, J.-P. & Simek, J. (1990). The last pleniglacialin the south of France (24,000–14,000 years ago). In(O. Soffer & C. Gamble, Eds) The World at 18000BP, Vol. 1: High Latitudes, pp. 69–86. London:Unwin Hyman.

Rolland, N. & Dibble, H. (1990). A new synthesisof Middle Paleolithic variability. Am. Antiq. 55,480–499.

119

Simmonet, R. (1982). Grandes lames de silex dans lePaléolithique supérieur de Pyrénées Centrales, essaisur des documents marginaux. Soc. Préhist. l’Ariège37, 61–106.

Soffer, O. (1989). The Middle to Upper Palaeolithictransition on the Russian Plain. In (P. Mellars & C.Stringer, Eds) The Human Revolution: Behavioral andBiological Perspectives in the Origins of ModernHumans, pp. 714–742. Princeton: PrincetonUniversity Press.

Soffer, O. (1991). Lithics and lifeways: the diversity inraw material procurement and settlement systemson the Upper Paleolithic East European Plain. In(A. Montet-White & S. Holen, Eds) Raw MaterialEconomies Among Prehistoric Hunter-Gatherers,pp. 221–234. Lawrence: University of KansasPublications in Anthropology No. 19.

Sonneville-Bordes, D. de (1959). Position strati-graphique et chronologique relative des resteshumains du Paléolithique supérieur entre Loire etPyrénées. Ann. Paléontol. 45, 19–51.

Sonneville-Bordes, D. de (1960). Le Paléolithiquesupérieur en Périgord. Bordeaux: Delmas.

Sonneville-Bordes, D. de (1980). L’évolution desindustries aurignaciennes. In (L. Bánesz & J. K.Kozłowski, Eds) L’Aurignacien et le Gravettian(Périgordien) dans leur cadre écologique, pp. 255–273.Nitra: Institut Archéologique de l’AcadémieSlovaque des Sciences.

Spiess, A. (1979). Reindeer and Caribou Hunters: AnArchaeological Study. New York: Academic Press.

Stahle, D. & Dunn, J. (1982). An analysis and appli-cation of the size distribution of waste flakes from themanufacture of bifacial stone tools. World Archaeol.14, 84–97.

Stiner, M. & Kuhn, S. (1992). Subsistence, technology,and adaptive variation in Middle Paleolithic Italy.Am. Anthrop. 94, 306–399.

Straus, L. (1991). The role of raw materials in UpperPaleolithic and Mesolithic stone artifact assemblagevariability in southwest Europe. In (A. Montet-White& S. Holen, Eds) Raw Material Economies AmongPrehistoric Hunter-Gatherers, pp. 169–185. Lawrence:University of Kansas Publications in AnthropologyNo. 19.

Straus, L. (1997). The Iberian situation between40,000 and 30,000 B.P. in light of European modelsof migration and convergence. In (G. Clark & C.Willermet, Eds) Conceptual Issues in Modern HumanOrigins Research, pp. 235–252. New York: Aldine deGruyter.

Stringer, C. (1989). Documenting the origin of modernhumans. In (E. Trinkaus, Ed.) The Emergence ofModern Humans: Biocultural Adaptations in the LaterPleistocene, pp. 67–96. Cambridge: CambridgeUniversity Press.

Stringer, C., Hublin, J.-J. & Vandermeersch, B. (1984).The origin of anatomically modern humans inwestern Europe. In (F. Smith & F. Spencer, Eds)The Origins of Modern Humans: A World Survey ofthe Fossil Evidence, pp. 1–50. New York: Alan R.Liss.

Svoboda, J. (1983). Raw material sources in earlyUpper Paleolithic Moravia. The concept of lithicexploitation areas. L’Anthropol. 21, 147–158.

Svoboda, J. (1987). Stránská skála. Bohunicky typ vbrnenske kotlinie. Studie Arch. ustauv C.S.A.V.14(1).

Svoboda, J. (1993). The complex origin of the UpperPaleolithic in the Czech and Slovak Republics. In(H. Knecht, A. Pike-Tay & R. White, Eds) BeforeLascaux: The Complex Record of the Early UpperPaleolithic, pp. 23–36. Boca Raton: CRC Press.

Taborin, Y. (1993a). Shells of the French Aurignacianand Périgordian. In (H. Knecht, A. Pike-Tay & R.White, Eds) Before Lascaux: The Complex Record of theEarly Upper Paleolithic, pp. 211–227. Boca Raton:CRC Press.

Taborin, Y. (1993b). La parue en coquillage au Paléo-lithique. Gallia Préhistoire 29.

Taborin, Y. (1998). Comment on Neanderthalacculturation in western Europe? A critical review ofthe evidence and its interpretation. Curr. Anthrop. 39(Suppl.), S1–S22.

Trinkaus, E. (1989). The upper Pleistocene transition.In (E. Trinkaus, Ed.) The Emergence of ModernHumans: Biocultural Adaptations in the LaterPleistocene, pp. 42–66. Cambridge: CambridgeUniversity Press.

Turq, A. (1991). L’approvisionnement en matièrespremières lithiques au Moustérian et au début duPaléolithique supérieur dans le Nord-Est du BassinAquitain (France). Paper presented at ‘‘El origendel hombre moderno en el suroeste du europa’’conference, Madrid.

Valoch, K. (Ed.) (1976). Die altsteinzeitlicheFundstelle in Brno-Bohunice. Studie Arch. ustavuC.S.A.V. 4(1).

White, R. (1982). Rethinking the Middle/Upper Paleo-lithic transition. Curr. Anthrop. 23, 169–192.

White, R. (1985). Upper Paleolithic Land Use in thePérigord: A Topographic Approach to Subsistence andSettlement. Oxford: British Archaeological ReportsInternational Series No. 253.

White, R. (1989a). Production complexity and stand-ardization in early Aurignacian bead and pendantmanufacture: evolutionary implications. In (P.Mellars & C. Stinger, Eds) The Human Revolution:Behavioral and Biological Perspectives in the Origins ofModern Humans, pp. 366–390. Princeton: PrincetonUniversity Press.

White, R. (1989b). Toward a contextual understandingof the earliest body ornaments. In (E. Trinkaus,Ed.) The Emergence of Modern Humans: BioculturalAdaptations in the Later Pleistocene, pp. 211–231.Cambridge: Cambridge University Press.

White, R. (1989c). Husbandry and herd control in theUpper Paleolithic: a critical review of the evidence.Curr. Anthrop. 30(5), 609–632.

White, R. (1993). Technological and social dimensionsof ‘‘Aurignacian-age’’ body ornaments across Europe.In (H. Knecht, A. Pike-Tay & R. White, Eds) BeforeLascaux: The Complex Record of the Early Upper Paleo-lithic, pp. 277–299. Boca Raton: CRC Press.

120 . .

White, R. (1998). Comment on Neanderthal accultur-ation in western Europe? A critical review of theevidence and its interpretation. Curr. Anthrop. 39Supplement, S30–S32.

Wilmsen, E. (1970). Lithic Analysis and Cultural Infer-ence: A Paleo-Indian Case. Tucson: AnthropologicalPapers of the University of Arizona No. 16.

Wilson, J. (1975). The last glacial environment at theAbri Pataud: A possible comparison. In (H. L.Movius, Jr, Ed.) Excavation of the Abri Pataud,Les Eyzies (Dordogne), pp. 175–186. Cambridge:Peabody Museum, Harvard University.

Wolpoff, M. (1981). Allez Neanderthal. Nature 289,823.

Wolpoff, M. (1989). Multiregional evolution: Thefossil alternative to Eden. In (P. Mellars & C.Stringer, Eds) The Human Revolution: Behavioral andBiological Perspectives in the Origins of ModernHumans, pp. 62–108. Princeton: PrincetonUniversity Press.

Wolpoff, M., Smith, F. & Frayer, D. (1997).Neandertals are a race of Homo sapiens. Paper pre-sented at the 6th annual meeting, PaleoanthropologySociety, St Louis.