Makers of the early Aurignacian of Europe - Cameron M

Makers of the Early Aurignacian of EuropeSTEVEN E. CHURCHILL1* AND FRED H. SMITH2

1Department of Biological Anthropology and Anatomy, Duke University,Durham, North Carolina 277082Department of Anthropology, Northern Illinois University, DeKalb,Illinois 60115

KEY WORDS Neandertals; origin of modern humans; Middle-to-Upper Paleolithic transition

ABSTRACT Despite intensive study and a number of remarkable discover-ies in the last two decades of the 20th century, our understanding of the culturaland biological processes that resulted in the emergence of the Upper Paleolithicand the establishment of modern humans in Interpleniglacial Europe remainsfar from complete. There is active debate concerning the timing and location ofthe origins of the Aurignacian, the nature of the origins of Initial Upper Paleo-lithic industries (whether by autochthonous development or through accultura-tion by Aurignacian peoples), the timing of the appearance of early modernhumans and the disappearance of the Neandertals, and the relationship ofarcheologically defined cultures to these different types of hominids. Frustratingour attempts to address these latter two questions is a general paucity oftaxonomically diagnostic human fossil material from early Upper Paleolithiccontexts. We undertake here a review of the human fossil record of Interpleni-glacial Europe, and its archeological and chronological context, to clarify to theextent possible the nature of the relationship between hominid groups and theearliest Upper Paleolithic artifact industries, particularly the early Aurignacian.Although substantial difficulties involved in interpreting the fossil, archeologi-cal, and geochronological records of this time period prohibit making any defin-itive statements, a number of observations are suggested by the current data: 1)the Middle Paleolithic of Europe appears to have been made exclusively byNeandertals; 2) Initial Upper Paleolithic industries (with the exception of theBachokirian) appear to have their roots in the late Middle Paleolithic industriesof their respective regions; 3) all of the human fossils yet recovered from InitialUpper Paleolithic (except the Bachokirian) contexts for which any diagnosticmorphology is present have their greatest morphological affinities with Nean-dertals and not early modern humans; 4) modern humans were almost certainlyestablished in Europe by ca. 32 ky BP, with a strong possibility that they werethere by ca. 36 ky BP. Claims for an appearance before 36 ky BP cannot besubstantiated with currently available evidence; 5) the hypothesis that modernhumans are uniquely associated with the Aurignacian cannot yet be refuted.Aurignacian-associated human fossils (including those from the Bachokirian) forwhich any diagnostic morphology is present have their greatest affinities withearly modern Europeans and not Neandertals; and 6) Neandertals and modernhumans coexisted in Europe for at least 2,000–4,000 years, and perhaps for8,000–10,000 years or longer. The overall picture is one of an extended period ofcultural contact, involving some degree of genetic exchange, between Neander-tals and early modern Europeans. Yrbk Phys Anthropol 43:61–115, 2000.© 2000 Wiley-Liss, Inc.

*Correspondence to: Steven E. Churchill, Department of Bio-logical Anthropology & Anatomy, Box 90383, Duke University,Durham, NC 27708-0383 USA, Phone: 919-660-7388, Fax: 919-660-7348, [email protected]

YEARBOOK OF PHYSICAL ANTHROPOLOGY 43:61–115 (2000)

© 2000 WILEY-LISS, INC.

TABLE OF CONTENTS

Introduction ........................................................................................................................... 62Interpleniglacial Chronology and Environments ............................................................... 65The End of the Mousterian in Europe ................................................................................ 71

The makers of the Mousterian ......................................................................................... 72Cariguela ........................................................................................................................ 73Veternica ........................................................................................................................ 73Svaty Prokop .................................................................................................................. 73Starosel’e ........................................................................................................................ 74

The Initial Upper Paleolithic of Europe ............................................................................. 74Chatelperronian ................................................................................................................ 76Uluzzian ............................................................................................................................. 77Szeletian ............................................................................................................................ 78The makers of the IUP ..................................................................................................... 79

Human Fossils From the Earliest Aurgignacian ............................................................... 79El Castillo ....................................................................................................................... 81Bacho Kiro ...................................................................................................................... 82Hahnofersand ................................................................................................................. 87Vogelherd ....................................................................................................................... 88Riparo Bombrini ............................................................................................................ 91La Ferrassie ................................................................................................................... 92Vindija Cave ................................................................................................................... 92Mladec ............................................................................................................................ 95Zlaty kun ........................................................................................................................ 98Fossellone ....................................................................................................................... 99Kelsterbach .................................................................................................................... 99Kent’s Cavern .............................................................................................................. 100Cioclovina ..................................................................................................................... 101Podbaba ........................................................................................................................ 101Camargo ....................................................................................................................... 101

The Makers of the Earliest Aurignacian and the Timing of the Appearance of ModernHumans in Europe ................................................................................................... 102

Acknowledgments ............................................................................................................... 109Literature Cited .................................................................................................................. 109

At the end of the Mousterian phase of paleolithic cul-ture, the Neandertal inhabitants of Europe wereabruptly replaced by people of the completely modernhuman type. There is reason to suppose that this newpopulation, the Aurignacians, having developed theirdistinctive culture elsewhere, probably in Asia, mi-grated into Europe and, with their superior social orga-nization, quickly displaced Mousterian man and occu-pied his territory. (Le Gros Clark, 1966, p. 116–117.)

INTRODUCTION

The latter part of the Interpleniglacial,marked by the approximately 10,000-year-long Hengelo/Denekamp temperate periodbetween 39–29 ky BP, was a dynamic timein European prehistory. It was during thisrelatively warm and wet interval of oxygenisotope stage 3, in which sediments in cavesand rock shelters were as often washing

away as accumulating (Cabrera Valdesand Bischoff, 1989), that the Mousterianculture and Neandertals disappeared, Ini-tial Upper Paleolithic (IUP) cultures pro-liferated and then disappeared, and theAurignacian and early modern humansbecame established across Europe. Therecontinue to be many uncertainties sur-rounding the emergence of the UpperPaleolithic, particularly the Aurignacian,and early modern humans in Europe. Re-gardless, accumulating evidence points tomarked complexity in the biocultural dy-namics of this period, which renders therather simple perspective on these phe-nomena reflected in the quotation aboveincreasingly less robust an explanation.

62 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 43, 2000

The complexity of the biological and cul-tural processes associated with the emer-gence of the Upper Paleolithic and modernhumans in Europe is well illustrated by twoimportant discoveries in the last year of the20th century. First, the direct dating of Ne-andertal fossils, found in association withAurignacian-like tools at Vindija Cave(Croatia), to ca. 28–29 ky BP (Smith et al.,1999), has revealed a new intricacy to thepattern of Neandertal extinction, showingas it did that populations of these hominidssurvived in Central Europe (as well as inthe Iberian peninsula and, less certainly, inthe Caucasus Mountains) long after the ap-pearance of modern humans on the conti-nent.1 The meaning and reliability of thestratigraphic association of these late Nean-dertals with Aurignacian-like tools is notentirely certain (see Zilhao and d’Errico,1999), but this possible association does callinto question the widespread assumption ofan Aurignacian-modern human correlation,and also raises doubts about the sagacity ofassuming that the contemporaneity of theChatelperronian and Aurignacian in West-ern Europe demonstrates coexistence of Ne-andertals and early modern humans in thetime range of 40–28 ky BP. Second, thediscovery of a Gravettian-aged juvenileskeleton from Lagar Velho (Portugal) (Du-arte et al., 1999) has added fuel to the al-ready heated debate about the extent of thegenetic contribution Neandertals made tomodern populations. This child has been ar-gued to evince a mosaic of Neandertal andearly modern human traits—traits thathave been interpreted as demonstratingsubstantial genetic input from Neandertalsto the early modern European gene pool(Duarte et al., 1999). This interpretationstands in contradistinction to results of re-

cent mitochondrial DNA analyses, first ongenetic material from the Feldhofer Nean-dertal (Krings et al., 1997, 1999), and thenon a Mousterian-associated infant fromMezmaiskaya Cave (Ovchinnikov et al.,2000), suggesting little if any genetic contri-bution from Neandertals to the modern Eu-ropean gene pool. Neither of these claimshas gone unchallenged (see Tattersall andSchwartz, 1999 on Lagar Velho, and Nord-burg, 1998 on the mtDNA results), and thepercentage of Neandertal genes in modernEuropeans remains a matter of debate. Ifthe morphology of the Lagar Velho childdoes reflect a degree of Neandertal ancestry,it invokes an image of significant exchangeof material culture and mates betweenneighboring populations of humans.2

A number of other developments in recentyears add to the picture of complexity dur-ing late Interpleniglacial Europe. These in-clude, but are not limited to: the discovery ofNeandertals associated with Initial UpperPaleolithic assemblages at La Roche a Pier-rot (St. Cesaire: Leveque and Vandermeer-sch, 1980) and the Grotte du Renne (Arcy-sur-Cure: Hublin et al., 1996) (the lattercase including objects of personal adorn-ment), the radiocarbon dating of the basalAurignacian to ca. 38.5 ky BP in northernSpain (Bischoff et al., 1989, 1994; CabreraValdes and Bischoff, 1989), and the discov-ery of Neandertals and Mousterian technol-ogy with dates showing they persisted in therefugia of southern and western Iberia untilas recently as perhaps 28 ky BP (Vega To-scano, 1990; Straus et al., 1993; Hublin etal., 1995), and possibly in the northernCaucasus until about the same time(Golovanova et al., 1999; Ovchinnikov et al.,

1The dating of the Vindija G1 hominids was done concurrentlywith the direct dating of the important hominid frontal bonefrom Velika Pecina. Originally thought to represent one of theearliest modern humans in Europe based on its association withan Aurignacian component dated to greater than 34 ky BP(Smith, 1976; 1982), Velika Pecina now appears to represent amiddle Holocene specimen that was intrusive into the Aurigna-cian layer (Smith et al., 1999). Thus while the direct dating ofthe late surviving Vindija hominids suggests a long period oftemporal overlap of Neandertals and early modern humans, thedating of the Velika Pecina frontal removes from considerationone of the formerly strongest lines of evidence for an earlyoccupation of Europe by modern humans.

2But populations of what? Opinions differ about the taxonomiclevel at which Neandertals and modern humans are distin-guished, i.e., whether the two groups should be seen as conspe-cifics (distinguished as subspecies) or as separate species. Sincethe application of a strict biological species distinction is bothpremature given the current state of evidence, and presumesthat which we seek to know (specifically, the degree of inter-breeding), we feel the only defensible approaches are to 1) em-ploy an evolutionary species concept that sees the two groups assister species that potentially interbred along hybrid zones, or 2)to consider them as subspecies. This question revolves aroundissues of taxonomy and systematics that are beyond the scope ofthis review. Therefore, we choose here to simply refer through-out the paper to Neandertals as “Neandertals” and early modernEuropeans as “modern humans,” “early modern humans,” “ana-tomically modern humans,” or early modern Europeans.

MAKERS OF THE AURIGNACIAN 63Churchill and Smith]

2000). All of these developments suggest theMiddle-to-Upper Paleolithic “transition”was a multifaceted, regionally variable phe-nomenon. Out of this realization, five inter-related questions have emerged that aredriving much of the research and debateabout the biocultural dynamics of this pe-riod. These questions, or sets of questions,are:

1) Whence the Aurignacian? Did it arise di-rectly from some regional variant of theMousterian (e.g., as Cabrera Valdes etal., 1997 have suggested for the earlyAurignacian of Cantabria), or perhapsfrom some regionally distinct IUP cul-ture (such as the “Aurignacoid” Ba-chokirian: see Kozłowski et al., 1982)? Ifso, why did the bone and lithic artifacttypes of the Aurignacian appear so rap-idly and so widely across Europe (Har-rold, 2000)? Or was the origin of theAurignacian exogenous to Europe, oc-curring perhaps in the Near East andarriving in Europe in the hands of in-migrating populations (Mellars, 1996)?If the latter, the source area has yet tobe discovered (despite the fact thatLevantine assemblages that could becharacterized as Upper Paleolithic coulddate to as old as 50 ky BP (see Mellars,1996), specific Near Eastern analogs ofthe Aurignacian emerge only after theappearance of the Aurignacian in Eu-rope; Marks, 1993). The center of originsof the Aurignacian remains elusive, andthis raises uncertainty over whether thebirth of the Aurignacian was a single,regionally circumscribed event, or theend product of a pan-European trend intechnology (favoring blade production)that resulted in the convergent develop-ment of the Aurignacian or Aurignacian-like industries in different regions (Oliva,1993; Cabrera Valdes and Bernaldo deQuiros, 1996; Straus, 1997; Karavanicand Smith, 1998). This latter perspectiveraises the question whether the Aurigna-cian was a single ethnic or cultural en-tity, either across space or through time(see Mellars, 1996; Miracle, 1998).

2) How were hominid taxa and lithic cul-tures related? The well-documented as-

sociation of modern humans with Mous-terian artifacts in the Near East andNorth Africa (McCown and Keith, 1939;Vandermeersch, 1981; Vermeersch et al.,1998), combined with the clear associa-tion of Neandertals with IUP (e.g., at St.Cesaire and Arcy-sur-Cure: Leveque andVandermeersch, 1980; Hublin et al.,1996) and Aurignacian-like (at Vindija:Karavanic, 1995; Karavanic and Smith,1998) assemblages in Europe, calls intoquestion any generalizations aboutwhich hominids were making which in-dustries. This question is critical, how-ever, to our understanding of the dynam-ics of the emergence of the UpperPaleolithic in Europe, since it lies at thebase of our models of the interaction be-tween Neandertals and early modern hu-mans. As pointed out by Harrold (1989,2000), Straus (1997), and others, the pos-sibility that the European Mousterianwas, until its very end, the product ofNeandertal cultural behavior, and thatthe Upper Paleolithic, at least from Au-rignacian times onward, was made byanatomically modern humans, raisesthe additional possibility that the cul-tural transition from Middle-to-UpperPaleolithic reflects differences in thecultural capacities of these two types ofhominid.

3) What were the biological and culturalprocesses that led to the emergence andspread of the Upper Paleolithic, as seenin both the regional fluorescence of IUPindustries and in the emergence of a pan-European Aurignacian? Do regionallydistinct IUP industries represent autoch-thonous development from local Moust-erian antecedents (e.g., d’Errico et al.,1998; Zilhao and d’Errico, 1999), or ac-culturated derivations of the Moust-erian, possibly having come aboutthrough cultural diffusion from contactwith Aurignacian-bearing peoples (seeHarrold, 1989; Mellars, 1996)?

4) What was the timing of the disappear-ance of Neandertals and the appearanceof modern humans in different regions(and vis-a-vis the questions above, how isthis related to the appearance of the Up-per Paleolithic and the disappearance of


the Middle Paleolithic in different re-gions)? Recent dating of late Mousterian(Vega Toscano, 1990; Straus et al., 1993)and Neandertal (Hublin et al., 1995;Smith et al., 1999; Ovchinnikov et al.,2000) remains suggests a considerabletime of Neandertal/modern human co-oc-cupation of Europe, spanning perhaps10,000 years. Confirming this possibilityis made difficult, however, by uncertain-ties in radiometric dating (see Mellars,1999 vs. Zilhao and d’Errico, 1999) andby a paucity of diagnostic human re-mains in good association with early Up-per Paleolithic assemblages in Europe.

5) What were the adaptive and behavioralcharacteristics of late Neandertals andearly modern Europeans, and what wasthe nature of the interaction betweenthem? Do typological and technologicaldifferences between artifact assemblagesreflect important differences in adaptivemodalities, and might these differencesthen provide insight into the bioculturaldynamics surrounding the developmentof the Upper Paleolithic and the demiseof the Neandertals? Furthermore, whatwas the nature of settlement patterningand niche partitioning that could allowtwo adaptively similar groups of humansto coexist in Europe for thousands ofyears (Mellars, 1998)?

The answers to these questions are inex-tricably and frustratingly bound up in oneanother, making the resolution of any par-ticular question difficult. Added to this dif-ficulty are ambiguities in the fossil, archeo-logical, and geochronological records thathave contributed additional subjectivity tothe interpretation of the prehistory of Inter-pleniglacial Europe. In this paper we reviewearly (pre-30 ky BP) Aurignacian, as well asInitial Upper Paleolithic, sites that haveproduced human remains, and we reviewthe evidence pertaining to the taxonomicdiagnosis of those remains, to clarify thecurrent state of knowledge concerning hom-inid-cultural associations. We do so becauseimplicit in much of the discussion of thedynamics of modern human origins in thisregion is the assumption that Neandertalsproduced the Middle Paleolithic and Initial

Upper Paleolithic industries, and that ana-tomically modern humans produced the Au-rignacian and subsequent Upper Paleolithicindustries (see Stringer et al., 1984; Gam-bier, 1989; Stringer, 1992; Hahn, 1993;d’Errico et al., 1998; Miracle, 1998).

INTERPLENIGLACIAL CHRONOLOGYAND ENVIRONMENTS

An understanding of the biocultural dy-namics of the Middle-to-Upper Paleolithictransition requires a firm geochronologicalframework, yet difficulties with the reliabil-ity and interpretation of various datingmethods add still more ambiguity to thatalready surrounding the classification ofscrappy, undiagnostic fossils and small, ty-pologically mixed artifact assemblages (seeZilhao and d’Errico, 1999). Radiocarbon dat-ing continues to play the dominant role indating the emergence of the Upper Paleo-lithic and the advent of modern humans.The development of accelerator mass spec-trometry (AMS) in the late 1970s greatlyimproved the utility of 14C dating (by allow-ing use of smaller samples, thereby mini-mizing the effects of contamination, and byextending the datable range of organic ma-terial an additional 10,000 years),3 but ac-celerator dating still suffers the same prob-lems of contamination and calibration as itsolder sibling, conventional radiocarbon(Gowlett, 1987). Because atmospheric levelsof 14C vary in response to fluctuations ingeomagnetic fields, solar activity, and theearth’s surficial geochemistry, radiocarbonyears do not directly correspond to calendaryears (Kitagawa and van der Plicht, 1998).Only the recent end of the radiocarbon scale(back to ca. 13 ky BP) has been calibrated bydendrochronology and data from glacialvarves (references in Kitagawa and van derPlicht, 1998). Recent work comparing U-se-ries and 14C dates in corals (Bard et al.,1990) and speleothems (Vogel and Kronfeld,1997), and dating macrofossils in lake sedi-mentation varves (Kitagawa and van der

3AMS relies on direct spectrometric counting of 14C atoms,whereas conventional radiocarbon measures radioactive decayrates to estimate 14C content. AMS can therefore detect asmaller 14C fraction, allowing for an extension of the dateablerange of radiocarbon time.


TABLE 1. Representative absolute dates for the final Mousterian, earliest Initial Upper Paleolithic, and earliestAurignacian for various regions in Europe1

Site/layer Method Date Source

Eastern EuropeFinal MousterianAkhstyr (Russia) U-Th 35.0 6 2.0 Lioubine, 1993Betovo (Russia) 14C 36.5 6 0.1 Soffer, 1989Mezmaiskaya (Russia) AMS 29.2 6 1.0 Ovchinnikov et al., 2000

40.7 6 1.6.45.0

Malaıa Voroncov 14C 35.7 6 0.5 Lioubine, 1993Molodova V (Russia)/ashy band 14C .35.6 Soffer, 1989StreletskayaKostenki 12 (Russia)/1a 14C 32.7 6 0.7 Soffer, 1989SpitsynskayaKostenki 17 (Russia)/2 14C 32.2 1 2.0–1.6 Soffer, 1989

32.8 6 0.336.4 1 1.7–1.4

Central EuropeFinal MousterianCrvena Stijera (Croatia?)/12 14C 40.8 6 0.9 Vogel and Waterbolk, 1972Erd (Hungary)/d 14C 35.3 6 0.9 Vogel and Waterbolk, 1972

39.4 6 0.8Gura Cheii (Romania) 14C 29.7 1 1.7–1.4 Carciumaru, 1989Ohaba Ponor (Romania)/IIIa 14C 39.2 1 4.5–2.9 Allsworth-Jones, 1990a

.41.0Pestera Cioarei (Romania)/II (top) 14C 37.8 6 1.0 Allsworth-Jones, 1990aRipiceni-Izvor (Romania)/IV–V 14C 40.2 1 1.1–1.0 Allsworth-Jones, 1990aTata (Hungary) 14C 33.3 6 0.9 Vogel and Waterbolk, 1972Tokod (Hungary) 14C 36.2 Gabori-Csank, 1970BachokirianBacho Kiro (Bulgaria)/11 14C .43.0 Mook, 1982

AMS 33.8 6 0.9 Hedges et al., 199434.8 6 1.237.7 6 1.538.5 6 1.7

Istallosko (Hungary)/lower 14C 39.7 6 0.9 Vogel and Waterbolk, 197244.3 6 1.9

Temnata (Bulgaria)/4b AMS 36.9 6 1.3 Ginter et al., 199638.2 6 1.538.3 6 1.838.8 6 1.739.1 6 1.8

TL 45 6 7 Ginter et al., 199646 6 8

SzeletianBohunice (Czech Republic) 14C 40.2 6 1.2 Svoboda, 1990

41.4 1 1.4–1.242.9 1 1.7–1.4

Stranska Skala (Czech Republic)/IIIa 14C 41.3 1 3.1–2.2 Svoboda, 1990Szeleta (Hungary)/C3 14C 43.0 6 1.1 Allsworth-Jones, 1990bAurignacianKrems-Hundssteig (Austria) 14C 35.2 6 2.0 Allsworth-Jones, 1990bMitoc Malu Galben (Romania)/III 14C 31.9 6 0.8 Allsworth-Jones, 1990aPesko (Hungary) 14C 34.6 6 0.6 Allsworth-Jones, 1990b

35.2 6 0.7 Vogel and Waterbolk, 1972Samuilica Cave (Bulgaria) 14C 42.8 6 1.3 Vogel and Waterbolk, 1972Willendorf C (Austria)/1 14C 41.7 1 3.7 Allsworth-Jones, 1990bWillendorf C (Austria)/2 14C 39.5 1 1.5–1.2 Allsworth-Jones, 1990bNorthwestern EuropeFinal MousterianCoygan Cave (Great Britain) 14C 38.7 1 2.7–2.0 Aldhouse-Green and Pettitt, 1998Creswell Crags (Great Britain) 14C 37.2 6 1.3 Aldhouse-Green and Pettitt, 1998

.42.7 6 1.6Hyaena Den (Great Britain) 14C 34.9 6 1.5 Aldhouse-Green and Pettitt, 1998

40.4 6 1.6Scladina (Belgium)/1A 14C 38.7 6 1.5 Bastin et al., 1986Leaf-point early Upper PaleolithicBench Tunnel Cavern (Great Britain) 14C 27.2–34.5 Hedges et al., 1989Couvin (Belgium) 14C ;45.0 Otte, 1990(?) Picken’s Hole (Great Britain) 14C 27.5 6 2.6 ApSimon, 1986

34.4 1 2.6–1.9(Continued)


TABLE 1. (Continued)


AurignacianGeißenklosterle (Germany)/IIa AMS 33.2 6 0.5 Hahn, 1995, 1996

33.7 6 1.136.8 6 1.0

Geißenklosterle (Germany)/IIIa AMS 33.1 6 0.7 Hahn, 199633.5 6 0.637.8 6 1.140.2 6 1.6 Zilhao and d’Errico, 1999

TL 40.2 6 1.5Trou Magrite (Belgium)/3 AMS 41.0 6 1.7 Otte and Straus, 1995Vogelherd (Germany)/V 14C 31.9 6 1.1 Muller-Beck, 1983

30.2 6 1.3Italian PeninsulaFinal MousterianBuca del Iena (Italy) U-Th ,40.0 Pitti and Tozzi, 1971Grotta Breuil (Italy)/3 ESR 36.6 6 2.7 Schwarcz et al., 1991UluzzianCavallo (Italy)/EII–I 14C .31.0 Bietti, 1997Grotta di Castelcivita (Italy)/rpi 14C 33.0–32.0 Bietti, 1997AurignacianGrotta di Fumane (Italy)/A2 AMS 31.6 6 0.4 Bietti, 1997

36.8 1 1.2–1.4Grotta di Paina (Italy) AMS 37.9 6 0.8 Bietti, 1997

38.6 1 1.4–1.8France and Northeastern SpainFinal MousterianL’Arbreda (Spain)/B1 AMS 34.1 6 0.8 Bischoff et al., 1989

39.4 6 1.441.4 6 1.6

Camiac (France) 14C 35.1 1 2.0–1.5 Delibrias and Evin, 1980Les Cottes I (France) 14C 37.6 6 0.7 Vogel and Waterbolk, 1967Cueva Millan (Spain)/Ia 14C 37.6 6 0.7 Moure Romanillo and Garcia Soto, 1983Els Ermitons (Spain) 14C 36.4 6 1.8 Harrold, 1989La Quina (France) 14C 35.3 6 0.5 Vogel and Waterbolk, 1967

34.1 6 0.7Arcy (France)/XII 14C 34.6 6 0.9 Vogel and Waterbolk, 1967La Rochette (France)/7 14C 36.0 6 0.5 Vogel and Waterbolk, 1967Romanı (Spain) U-series 39–43 Harrold, 1989ChatelperronianLe Moustier (France)/K TL 42.6 6 3.2 Valladas et al., 1986St. Cesaire (France)/EJOP Sup (8) TL 36.3 6 2.7 Mercier et al., 1991Arcy (France)/X AMS 33.8 6 0.7 Zilhao and d’Errico, 1999Arcy (France)/IX AMS 45.1 6 2.8 Zilhao and d’Errico, 1999Combe Sauniere (France)/X AMS 33.0 6 0.9 Zilhao and d’Errico, 1999

38.1 6 1.0Grotte XVI (France)/B AMS 35.0 6 1.2 Zilhao and d’Errico, 1999

38.1 6 1.7.39.8

Roc de Combe (France)/X AMS 31.0 6 0.8 Zilhao and d’Errico, 199938.0 6 2.0

AurignacianCaminade (France)/G AMS 37.2 6 1.5 Zilhao and d’Errico, 1999Caminade (France)/F AMS 35.4 6 1.1 Zilhao and d’Errico, 1999Castanet (France)/Inferieur AMS 35.2 6 1.1 Zilhao and d’Errico, 1999Castillo (Spain)/18c AMS 39.8 6 1.4 Cabrera Valdes and Bischoff, 1989; Cabrera

Valdes and Bernaldo de Quiros, 199640.0 6 2.140.7 6 1.541.1 6 1.742.2 6 2.1

Castillo (Spain)/18b2 AMS 37.1 6 2.2 Cabrera Valdes and Bischoff, 1989; CabreraValdes and Bernaldo de Quiros, 199637.7 6 1.8

38.5 6 1.340.7 6 1.6

Castillo (Spain)/18b1 AMS 38.5 6 1.8 Cabrera Valdes and Bischoff, 1989Combe Sauniere (France)/VII AMS 34.0 6 0.9 Zilhao and d’Errico, 1999Isturitz (France)/U27 4d AMS 36.5 6 0.6 Zilhao and d’Errico, 1999Isturitz (France)/V1 26 AMS 34.6 6 0.6 Zilhao and d’Errico, 1999

(Continued)


Plicht, 1998), are for the first time allowingpreliminary extension of the calibrated 14Crange back to about 38 ky BP. While muchremains to be resolved (see van Andel, 1998;van der Plicht, 1999), these studies showthat in the time range from 30–40 ky BP(the practical end of the 14C range), 14C con-sistently (but not uniformly) produces dates2,000–4,000 years too recent. Other datingmethods applicable to the later Pleisto-cene—thermoluminescence (TL), electronspin resonance (ESR), and uranium series(U-series)—are independent of the earth’s14C history, and thus are expected to pro-duce dates more concordant with calendartime. Some prehistorians recommend “cor-recting” TL, ESR, and U-series dates bysubtraction of age-specific constants to con-vert them to radiocarbon years (see Zilhaoand d’Errico, 1999), but uncertainties in theform of the calibration curve (see van derPlicht, 1999) render the use of such correc-tions premature. All dates discussed in thispaper are in radiocarbon years unless oth-erwise indicated.

Given that dating the events of the Inter-pleniglacial requires operating at the prac-tical limits of radiocarbon, and given un-evenness in the radiocarbon calibrationcurves, reported 14C determinations shouldbe treated with caution. While we can gen-erally assume that radiocarbon dates pro-vide a reasonably accurate probable meanage of deposition within a known error

range, it is important to bear in mind thatthe older the sample, the greater the possi-bility and the impact of contamination, andthat radiocarbon produces dates on a differ-ent scale than other methods, and thus caremust be taken in the interpretation of dat-ing results (see Zilhao and d’Errico, 1999).

Paleoenvironmental indicators within anarcheological horizon provide additional in-formation that may help to place a site intime. Consideration of the overall pattern ofradiocarbon dates across Europe (Table 1)suggests that both IUP and Aurignaciancultures arose around the time of the WurmII/III interstadial (the Hengelo or Podrahemtemperate period: Fig. 1), during the inter-val between ca. 39–37 ky BP. Kozłowski(1996) notes that, given the underestima-tion bias of radiocarbon dating, horizonsdated to the Hengelo may actually havebeen deposited during the preceding Moer-shoofd temperate period (between ca. 45–50ky BP). Discrepancies between radiocarbondates from land and marine cores have beennoted (Rossignol-Strick, 1995), making cor-relation of the climatic and chronologicalrecords difficult. However, since 14C datesfrom pollen sequences (e.g., Shotton, 1977)are subject to the same calibration error asall 14C dates, the pollen zone chronology (asin Fig. 1) should generally correspond toradiocarbon years.

The Interpleniglacial (or Middle Wurm),even during full stadial episodes, was

TABLE 1. (Continued)


L’Arbreda (Spain)/B1 AMS 37.7 6 1.0 Bischoff et al., 198937.7 6 1.038.7 6 1.239.9 6 1.3

Reclau Viver (Spain)/TIII-27 AMS 40.0 6 1.4 Straus, 1997Romanı (Spain)/2 AMS 37.1 6 1.0 Bischoff et al., 1994

U-series .42.6 6 1.1 Bischoff et al., 1994Portugal/southern IberiaFinal MousterianColumbeira 14C 28.9 6 1.0 Antunes et al., 1989Figueira Brava 14C/U-

series31–30 Antunes, 1990

Gruta do Caldeirao AMS 27.6 6 0.6 Zilhao, 1993Pedreira das Salemas 14C 29.9 6 1.0 Antunes et al., 1989Zafarraya AMS/U-

series32–29 Hublin et al., 1995

AurignacianGato Preto (Portugal) TL 38.1 6 3.9 Marks et al., 1994La Vina (Spain)/XIII inf 14C 36.5 6 0.8 Zilhao and d’Errico, 19991 All dates in ky BP.


Fig. 1. Approximate correlations between intersta-dial and glacial stadial periods in Central and WesternEurope during the early and mid-Wurm. (Wurm subdi-visions are based on Alpine river terrace deposits. TheWurm glacial is also known as the Weichsel in NorthernEurope and the Devensian in the British Isles.) Ages ofinterstadials are based on 14C determinations fromNorthern European pollen sequences (largely but notexactly following Shotton, 1977) and should be taken as

approximate. Precipitation and temperature curvesfrom regional pollen sequences vary in the timing of theonset and end of temperate periods, making correla-tions across regions difficult. The rainfall and tempera-ture curves depicted here are derived from the pollensequence at Les Eschets, southwestern France (datafrom Guiot et al., 1989). The present-day base (dashed)line represents rainfall of 800 mm and mean annualtemperature of 11°C.


milder in climate than the glacial cycles be-fore and after (the Lower and Upper Pleni-glacial periods, respectively). Climatic fluc-tuations during this period were generallymild but frequent, due to intermittent acti-vation of the north Atlantic thermohalinecirculation (see van Andel and Tzedakis,1996a). Although mild, fluctuations werelikely rapid, occurring in some cases on theorder of a few decades or less (as reflected inoxygen isotope ratios in the GreenlandSummit ice core: Dansgaard et al., 1993). Inaddition, the Interpleniglacial was punctu-ated by three major temperate oscillations.The first of these was the Moershoofd (in thenorthwestern European pollen scheme, orthe Loopstedt or Heraklitsa in the centralEuropean scheme), between ca. 50–43 kyBP (date based on Shotton, 1977). As re-flected in the pollen sequences from France(Les Echets and La Grande Pile), the Moer-shoofd saw rainfall in excess of modern pre-cipiations (by as much as 100 mm) but meanannual temperatures about 4°C below mod-ern norms (Guiot et al., 1989). During thesubsequent stadials of Wurm II, IIIa, andIIIb, mean annual temperatures may havebeen 10°C or more below modern Europeanvalues, with precipitation half or less of thatreceived by Europe today (Guiot et al.,1989). The stadials saw the spread of step-pic grasslands in Southern Europe (with theretention of wooded refugia in sheltered val-ley systems), arctic steppe and tundra above45° N latitude, and possible ice advances inNorthern Europe (van Andel and Tzedakis,1996a). Perhaps the most severe of theseoccurred during Wurm IIIa, with mean an-nual temperatures about 12°C below mod-ern norms (Guiot et al., 1989), and withwinter temperatures of 220°C not unusual(Mellars, 1998). Even in the south of Eu-rope, faunal assemblages dated to this in-terval are heavily dominated by reindeer,while those of the interstadials are typicallyrepresentative of a more temperate wood-land fauna (red deer, horse, aurochs, bison:references in Mellars, 1998).

The second temperate oscillation of theInterpleniglacial was the Hengelo intersta-dial (the Wurm II/III interstadial, alsoknown as Les Cottes in Western Europe, orPodrahem or Kalabaki in Central Europe),

between ca. 40–37 ky BP. Interstadial con-ditions fostered soil formation in Centraland Northwest European loess deposits,and thus the interstadials are commonlyreferred to stratigraphically identified soils.The Hengelo has thus variously been called“Soil I” or “Les Vaux soil” in loess regions(e.g., Svoboda, 1988). The final interstadialof the Interpleniglacial occurred roughly be-tween 32–29 ky BP, and is known in West-ern European schemes as the Denekamp,Arcy, Kesselt, or Les Eyzies (although someconsider the Arcy to be a separate, short-lived, and earlier temperate oscillation),and in Central Europe as the Krinides, Still-fried B, or Soil II period. Despite the moretemperate conditions, during Hengelo andDenekamp times temperatures were 4–5°Ccolder than today, and precipitation wasroughly 20–25% below modern Europeanvalues (Guiot et al., 1989). During thewarmer interstadials, much of Europe wascovered with open (almost parkland) mixedpine and deciduous forest, with decreasingabundance of hardwoods moving north (vanAndel and Tzedakis, 1996a). Between 45°–55° N latitude, much of Europe was charac-terized by evergreen woodlands (open conif-erous forest), while above 55° N, shrubtundra dominated (van Andel and Tzedakis,1996b).

Although it seems certain that climaticinstability played a role in the bioculturaltransitions of the Interpleniglacial, whatthat role might have been is unclear. It haslong been recognized that the stadial cyclesof the Ice Age had important consequencesfor human settlement and demography inEurope (e.g., Gamble, 1986; Roebroeks etal., 1992; van Andel, 1998; Mellars, 1998),following a general (although not rigid: seeRoebroeks et al., 1992) pattern of northerlypopulation advances during temperate in-tervals and southerly retreats when the cli-mate deteriorated. Recently, Mellars (1998)proposed that the warmer and wetter con-ditions of the interstadials, especially thatof the Hengelo temperate oscillation, mayhave allowed the initial colonization of Eu-rope by anatomically modern humans de-spite the presence of resident populations ofNeandertals. The warmer periods may havedone so by both expanding the range of tem-


perate environments in Southern Europe towhich modern humans (presumably occupy-ing portions of the Levant and northern Af-rica) were already well adapted, and also bydisrupting the distribution of indigenousNeandertal populations, thus creating va-cant zones into which modern humans couldexpand without facing demographic compe-tition (Mellars, 1998; Hoffecker, 1999). Thesubsequent cold interval of Wurm IIIa mayhave stressed the adaptive capacities ofboth groups (as modern humans faced thenew adaptive challenges of stadial Europe,and Neandertals faced stadial conditions inthe context of new competitive pressuresfrom modern populations), and may havebeen an important factor in the demise ofthe Neandertals (Mellars, 1998). The tem-poral patterning of IUP and Aurignaciansites in southwestern Europe broadly sup-ports this model (Mellars, 1998), but only ifone makes the usual assumptions abouthominid-cultural associations (i.e., Middleand Initial Upper Paleolithic assemblagesequate with Neandertals, Aurignacian andlater Upper Paleolithic assemblages equatewith modern humans).

The warmer and wetter environmentsduring the above-named temperate oscilla-tions appear to have accelerated erosion ofsediments in caves and rockshelters. Forexample, at the abri La Roche a Pierrot (St.Cesaire), a 2,000-year hiatus separates theuppermost Mousterian from the lowermostChatelperronian (Mercier and Valladas,1996). Cabrera-Valdes and Bischoff (1989)contend that a depositional hiatus begin-ning 40–37 ky BP and ending perhapsaround 34 ky BP (thus likely correspondingto the Hengelo temperate oscillation) hasremoved critical early Upper Paleolithic lev-els from many Western European sites. Theresulting Mousterian/Aurignacian discon-formity may then lend a false appearance ofan abrupt cultural replacement. While therole these temperate oscillations played inthe human biocultural transition is not en-tirely certain, the increased erosion ratesthat they brought have no doubt contrib-uted to the scarcity of diagnostic human andearliest Upper Paleolithic artifactual re-mains from this time period.

THE END OF THE MOUSTERIAN INEUROPE

Terminal dates for the Mousterian areuncertain, and no clear geographic patternto its disappearance has yet been estab-lished (Table 1). Until recently, the young-est dates reported from Russia suggested aMousterian terminus there at around 36.5ky BP (Soffer, 1989), although a fewyounger dates have been reported (e.g., aU-Thorium date of 35.0 6 2.0 ky BP on level3 at Akhstyr Cave, and a 14C determinationof 35.7 6 0.5 ky BP from level 3 of MalaıaVoroncov: Lioubine, 1993). The bulk of dateson Russian Early Upper Paleolithic assem-blages tend to fall around 33–32 ky BP (Sof-fer, 1989), which may indicate a deposi-tional hiatus that has obscured either theend of the Mousterian or the beginning ofthe Upper Paleolithic. However, recent ac-celerator mass spectrometric (AMS) datingof presumably Neandertal subadult re-mains, associated with Mousterian tools atMezmaiskaya cave in the northern Cauca-sus, indicates an age of 29.2 6 1.0 ky BP(Ovchinnikov et al., 2000). Similar dates (ca.30 ky BP) have been reported for late Mid-dle Paleolithic assemblages in Crimea(Chabai and Marks, 1998). This suggests alater terminus for the Mousterian, as wellas a 3,000–4,000-year temporal overlap ofMousterian and early Upper Paleolithic inEastern Europe. However, previous 14C de-terminations from the Mousterian level atMezmaiskaya indicate ages of 45.0 and40.7 6 1.6 ky BP (Golovanova et al., 1999),raising questions about the true age of theMousterian at the site, and highlighting theneed for further dating of the terminal Mid-dle Paleolithic in Eastern Europe.

Dates for final Mousterian layers at Cen-tral European sites tend to fall between36–41 ky BP, with most of them clusteringaround 40–39 ky BP. However, youngerdates have been reported from two Hungar-ian sites (Erd layer d and Tata; 35.3 6 0.9and 33.3 6 0.9 ky BP, respectively: Vogeland Waterbolk, 1972) and from the Roma-nian site of Gura Cheii Cave (29.7 6 1.7 kyBP: Carciumaru, 1989). In France andnortheastern Spain there are a number ofradiocarbon determinations that indicate


the Mousterian lasted until as recently as35–34 ky BP, and it is now well establishedthat the Mousterian persisted until ca. 28ky BP in Iberia (Table 1).

The makers of the Mousterian

Despite the unquestionable association ofearly modern humans with Mousterian as-semblages in the Levant (McCown andKeith, 1939; Vandermeersch, 1981) andNorth Africa (Vermeersch et al., 1998),there is as yet no evidence to refute the ideathat Neandertals were the sole producers ofthe Mousterian in Europe. Neandertal re-mains have been recovered from terminalMousterian levels on the Iberian Peninsula(at Zafarraya: Hublin et al., 1995; and Cari-guela: Garcıa Sanchez, 1960; de Lumley andGarcıa Sanchez, 1971; and possibly Figueira

Brava: Antunes, 1990), in France (at LeMoustier: Bordes, 1959; Valladas et al.,1986; and Hortus: Piveteau et al., 1963), inBelgium (at Couvin: Ulrix-Closset et al.,1988), in Italy (at Grotta Breuil: Manzi andPassarello, 1995), and in Romania (atOhaba Ponor: Allsworth-Jones, 1990a) (Fig.2). Thus far no diagnostic modern humanremains have been found in good associa-tion with a Mousterian assemblage in Eu-rope. Claims for such associations, however,have been made for Cariguela Cave (Anda-lusia, Spain), Veternica (Croatia), andStarosel’e (Ukraine). Craniofacial and post-cranial remains from the Czech site of SvatyProkop, possibly associated with Moust-erian-like lithic artifacts and fauna sugges-tive of Wurm II, may also fit with this group.Because such associations would have im-

Fig. 2. Locations of sites discussed in the text: 1,Figueira Brava (Portugal); 2, Zafarraya (Spain), Cari-guela (Spain); 4, El Castillo and Camargo (Spain); 5,Paviland (England); 6, Kent’s Cavern (England); 7, St.Cesaire (France); 8, Combe Capelle and La Ferrassie(France); 9, Le Moustier and Font de Gaume (France);10, Cro-Magnon (France); 11, Hortus (France); 12, Arcy-sur-Cure (France); 13, Couvin (Belgium); 14, RiparoBombrini (Italy); 15, Kelsterbach (Germany); 16, Vo-

gelherd (Germany); 17, Hahnofersand (Germany); 18,Fossellone and Grotta Breuil (Italy); 19, Zlaty kun(Czech Republic); 20, Svaty Prokop and Podbaba (CzechRepublic); 21, Veternica (Croatia); 22, Vindija (Croatia);23, Mladec (Czech Republic); 24, Dzerava Skala (CzechRepublic); 25, Cavallo (Italy); 26, Mariaremete (Hunga-ry); 27, Istallosko (Hungary); 28, Cioclovina (Romania);29, Ohaba Ponor (Romania); 30, Bacho Kiro (Bulgaria);and 31, Mezmaiskaya (Russia).


portant implications for our understandingof the biocultural dynamics of this time pe-riod, we briefly review the nature of theevidence from each site below.

Cariguela. Excavated initially by J.-C.Spahni in 1954–1955, Cariguela (or Cari-huela) has a long Mousterian sequence (pos-sibly extending from Wurm IV [OIS 2] to thelast or Eem interglacial [OIS 5e]: see Strauset al., 1993), capped by a Neolithic upper-most level (Garralda, 1997). In addition tofinding remains attributable to Neandertalsin Mousterian levels 6, 7, and 9 (now UnitsV, VI, and VIII: Vega Toscano et al., 1988),Spahni also found the remains of gracileanatomically modern humans in levels 2and 3 (Garcıa Sanchez, 1960; de Lumley andGarcıa Sanchez, 1971). These levels he at-tributed to a Mousterian with Upper Paleo-lithic (Aurignacian) influence. The remainsinclude a small parietal fragment (Pınar 4)from strata 2 (now Unit III), a right-sidehemimandible with M1–3 (Pınar 5) also fromlevel 2 (Unit III), and a partial tibia (Pınar6) from the top of level 3 (now Unit IVa)(Straus et al., 1993; Garralda, 1997). Re-study of the artifact assemblages and stra-tigraphy of Cariguela failed to document anUpper Paleolithic component (Almagro etal., 1970; Vega Toscano et al., 1988), andthere appears to have been an occupationalhiatus at the site between the Middle Paleo-lithic and the Neolithic. The tools fromSpahni’s strata 2 and 3 are typologicallyand technologically Mousterian (with no in-dication of Upper Paleolithic tool types) thatmight have been secondarily deposited inlevels 2 and 3 as colluvium from the cavemouth (Vega Toscano et al., 1988). The hu-man remains are gracile even by Upper Pa-leolithic standards, and are similar in mor-phology, dental pathology, and postmortemmodification to Neolithic human remainsfrom Cariguela level 1 and other local sites(Garcıa Sanchez, 1960; Garralda, 1997).The association of gracile modern humansand Mousterian lithics in the upper portionof Cariguela Cave is most likely the result ofsecondary mixing (either colluvial deposi-tion of Mousterian artifacts from the cavemouth onto a Neolithic-aged cave surface, orintrusive Neolithic burial into Mousterian

levels) or poor stratigraphic control on thepart of Spahni (see Straus et al., 1993). Un-til the human remains can be directly dated,the evidence at Cariguela is too tenuous tosupport a serious claim of a modern human-Mousterian association in Europe.

Veternica. A similar example is providedby the case of Veternica Cave, located justoutside Zagreb in Croatia. A long strati-graphic sequence, extending from the Mous-terian through the Roman period, was exca-vated here in the 1950s (Malez, 1956). Afragmentary calvarium (VTR 830/55) wasinterpreted as deriving from a Riss-WurmMousterian stratum and was widely cited asevidence of modern humans associated withthe Mousterian (e.g., Skerlj, 1958; Coon,1962). However, the state of preservationand morphology of this specimen conformedexactly to a series of later (mostly Meso-lithic) skulls from the site, and the strati-graphic location of the specimen stronglysuggests an intrusive burial (Smith, 1977).As with Cariguela, Veternica does not pro-vide a compelling case for modern human-Mousterian association.

Svaty Prokop. An adult occipital frag-ment and femoral head were recovered froma travertine/loess breccia in a fissure at thebase of Svaty Prokop Cave in 1887. Thebreccia also yielded a questionable bonetool, lithic artifacts with affinities to theMousterian, and a fauna suggestive of earlyInterpleniglacial (Wurm II) times (Vlcek,1951, 1996). According to Vlcek (1951,1996), the occipital morphology of the SvatyProkop specimen compares favorably withZlaty Kun, suggesting it derives from anearly modern human. Brauer and Broeg(1998) also mention a mandible that pre-sumably belongs to this individual. Stillpartly encased in matrix, the mandible ex-hibits a mental eminence and trigone, aswell as other features, that align it withmodern human mandibular form. Unfortu-nately, the typological diagnosis of the arti-fact assemblage is uncertain, and the tapho-nomy of the breccia formation has not beenstudied, making the true age of this speci-men highly uncertain. If the Wurm II attri-bution from the fauna is correct, this speci-


men would rate as one of the earliestmodern humans in Europe, a possibilitythat should be further explored with directdating.

Starosel’e. In 1953, A. Formozov exca-vated a child’s skeleton from below a MiddlePaleolithic level at the site of Starosel’e(Formozov, 1958). The taxonomic status ofthe specimen is somewhat uncertain, withmany accepting the presence of both Nean-dertal and modern human features in theskeleton, and accordingly there has been adiversity of views on the significance of thisspecimen to modern human origins (see re-view in Marks et al., 1997). The skeleton,representing a child between the ages of1.5–3 years at death, exhibits the moderntraits of a high frontal angle, pronouncedchin, and marked canine fossa, but appearsmore primitive in its cranial vault thicknessand details of mandibular alveolar morphol-ogy (Alexeeva, 1997). At the time of discov-ery, Formozov arranged a commission ofprehistorians to examine the circumstancesof the burial and verify its antiquity (Alexe-eva, 1997; Marks et al., 1997). Close exam-ination revealed no evidence of an intrusivepit in the sediments, and most (but not all)of the committee members concluded thatthe burial was of Mousterian age (Alexeeva,1997). Attempts to directly date the speci-men have produced equivocal results, butfluorine analysis suggested that the speci-men was younger than the associated fauna(Marks et al., 1997). Subsequent work atStarosel’e has greatly clarified the picture.During excavations in 1993 and 1994,Marks et al. (1997) recovered two additionalhuman burials (an adult and a child) in thesame area (within 3–6 m) as the originalfind. The adult skeleton was at the samestratigraphic level (below the uppermostMousterian level) and was oriented (head tothe west) and positioned (lying supine andextended, left hand on pelvis, face to thesouth) identically to the 1953 burial. Theadult burial was found within a clearlyidentifiable burial pit, one that containedMousterian artifacts that had been second-arily deposited in the grave fill. The recentlydiscovered child’s burial, like the one discov-ered in 1953, had its head towards the west

and was lying on its back (although the bodyappears to have been semiflexed ratherthan extended). This child’s burial wasstratigraphically higher than the other two,but like the original discovery it lacked anidentifiable burial pit. The orientation andposition of the burials, along with the pres-ence of an historic Muslim graveyard nearthe site, led Marks et al. (1997) to arguethat all three skeletons—including theStarosel’e child—represent late-medievalMuslim burials.

All four of these cases (Cariguela, Veter-nica, Svaty Prokop, and Starosel’e) concernsituations in which the stratigraphic con-text and/or the archeological association ofthe fossils is highly questionable. While fu-ture efforts at direct dating may support theclaimed antiquity of some of these speci-mens, we suspect that specimens with mod-ern morphology are not likely to be foundwith the Mousterian, at least not in excessof ca. 35 ky BP. It is certainly true that noexisting case provides compelling evidencefor a modern human-Mousterian connectionin Europe.

THE INITIAL UPPER PALEOLITHIC OFEUROPE

During Wurm II/III (Hengelo) and IIIatimes, there existed throughout Europe in-dustries that combined Middle and UpperPaleolithic tool types and modes of pro-duction (technology). Often referred to as“transitional,” these industries are gener-ally considered as belonging to the UpperPaleolithic because of the occurrence of Up-per Paleolithic blade technologies and toolforms.4 These Initial Upper Paleolithic in-

4It is reasonable to ask if IUP assemblages might better beconsidered as belonging to the final Middle Paleolithic, contain-ing as they do elements of Middle Paleolithic technology and tooltypes. A number of issues make IUP typology especially thorny.If one employs a typological construct based on the frequencies offormal tools, then the blade-rich assemblages of the IUP shouldmost reasonably be considered as Upper Paleolithic. However,the past few decades have seen a growing discomfort with classicBordesian typology (including the Upper Paleolithic typologyestablished by De Sonneville-Bordes and Perrot: see referencesin Reynolds, 1990), yet no new theoretical perspective hasemerged to replace it (Harrold, 2000). What has emerged in thepost-Bordesian era has been an increased emphasis on technol-ogy over typology in the interpretation of lithic assemblages,most notably illustrated by the recent fluoresence of studies ofthe chaine operatoire (e.g., Bietti and Grimaldi, 1996). Thischange in emphasis has not necessarily clarified the position ofthe IUP, however. Blades, the sine qua non of the Upper Paleo-


dustries (to borrow the term applied byKuhn et al. (1999) to comparable assem-blages from Asia Minor) may be contempo-raneous with other Upper Paleolithic (Au-rignacian) industries in the same region,but generally do not show temporal overlapwith the local Mousterian. Across Easternand Central Europe, IUP assemblages aretypically laminar (but with blades com-monly produced by Levallois reduction),have both Middle Paleolithic (e.g., sidescrapers, notched pieces, denticulates) andUpper Paleolithic type elements (e.g., endscrapers, burins), and contain bifaciallyworked foliates (“leaf points”). Most of thevarious IUP cultures of Central and EasternEurope that are characterized by leaf pointshave at one time or another been seen asregional variants of the Szeletian, a culturedefined at Szeleta Cave in the Bukk Moun-tains of Hungary (Allsworth-Jones, 1990a).Regional distinctions can be identified, how-ever, and the possibility exists as well thatthe occurrence of leaf points in different re-gions may be the result, at least in somecases, of convergence (Allsworth-Jones,1990a). Regionally defined leaf-point culturesinclude the Brynzeny, the Gordineshty, andthe Kostenki-Streletsian (Streletskaya) of theRussian Plain (Anikovich, 1992), the Altmuh-lian of southern Germany, the Jerzmanowi-cian of eastern Germany and Poland, the Bo-hunician of the Czech Republic (in which leafpoints are relatively rare, leading to sugges-tions that this industry should be consideredas a development separate from the Szeletian:Svoboda, 1988, 1990), and the Jankovichian(although this might best be seen as a MiddlePaleolithic industry with leaf points: Gabori-Csank, 1990) and the Szeletian sensu strictoof Hungary (see Allsworth-Jones, 1990a).Leaf points also occur in the early Upper Pa-leolithic of Belgium and Great Britain (at the

sites of Kent’s Cavern, Spy, and Paviland),suggesting either a pan-European (at least tothe east and north of the Alps) development ofthe Szeletian sensu lato or convergence inpoint morphology in different parts of Europe.

The IUP in France and northern Spain isrepresented by the Chatelperronian (alsoknown as the Castelperronian or LowerPerigordian). While the Szeletian appearsto have its greatest affinities with the Mico-quian5 of Central Europe (since leaf pointsoccur in the Micoquian: Allsworth-Jones,1990a), the Chatelperronian looks as if itsroots are in the Mousterian of AcheulianTradition type B (MATb) of southwest Eu-rope (Bordes, 1972). This association isbased on a relatively high occurrence ofUpper Paleolithic elements in the MATb,including burins, end-scrapers, and espe-cially, backed knives (such as the Chatelp-erron knife) (Harrold, 1989). In addition tothese Upper Paleolithic type tools, the Chat-elperronian shares with other IUP indus-tries in retaining relatively high frequenciesof Middle Paleolithic elements (side scrap-ers, denticulates, and notched pieces). Un-like most Mousterian assemblages, how-ever, Chatelperronian assemblages can berich in bone artifacts, including small andlozenge-shaped bone points, awls, ba-guettes, and beads and pendants (seed’Errico et al., 1998). The Chatelperronianappears to interstratify with the Aurigna-cian at three sites (le Roc de Combe and lePiage (Bordes and Labrot, 1967) in France,and El Pendo in Spain (Gonzalez Echega-ray, 1980)), although this claim has recentlybeen challenged by Zilhao and d’Errico(1999) (see below).

In Italy the Mousterian gives way, atsome sites, to an Upper Paleolithic culturecalled the Uluzzian which is characterizedby crescent-shaped backed pieces, or lu-

lithic, can be produced in a variety of ways, and early forms ofprismatic core reduction methods (as evidenced by the presenceof crested blades), generally considered characteristic of theUpper Paleolithic, can be found in relatively old (even MiddlePleistocene) Mousterian assemblages (Bar-Yosef and Kuhn,1999). Likewise, Middle Paleolithic style Levallois blade corescan be found in later Upper Paleolithic assemblages (Bordes,1947; Newcomer, 1975). Given that there seems to be no perfectdiagnostic criteria for classifying “transitional” industries, or forsharply demarking Middle from Upper Paleolithic for that mat-ter, we adhere to tradition in treating these industries as UpperPaleolithic (e.g., Harrold, 1989; Farizy, 1990).

5When the European Middle Paleolithic is contrasted with thewhole of the Upper Paleolithic, it appears in comparison to lacksubstantial variation across space and through time. However,the Middle Paleolithic was not a monolithic entity, but instead alithic culture with considerable regional and temporal complex-ity. In addition to the well-known facies of the Mousterian (Bor-des, 1954, 1968), there were regional variants, such as the “Vas-conian” Mousterian of Spain and the Pyrenees, as well as non-Mousterian variants of the Middle Paleolithic, such as the“Micoquian” of Central Europe (see Harrold, 1989; Svoboda etal., 1996).


nates, as well as end scrapers, side scrapersand burins, and bone points or sagaies(Palma di Cesnola, 1993). First identified atsites on the Bay of Uluzzo in the south ofItaly (on the Ionian coast of Apulia, the“bootheel” of the peninsula), the Uluzzianoccurs as far north as Tuscany (Palma diCesnola, 1993), but is not known from cen-tral Italy (Zampetti and Mussi, 1988). Thisindustry has also been argued to be a re-gional variant of the Chatelperronian, inwhich backed crescents take the place ofbacked knives (Gioia, 1988).

Despite good archeological representationof IUP industries, the makers of these as-semblages are still poorly known. Humanremains have been recovered from IUP lev-els at only seven sites (although a few hom-inid-bearing sites with Aurignacian-typetools might be added to this list: see below).These human remains (reviewed below)have tended to be isolated teeth and veryfragmentary bones, frustrating attempts tomake taxonomic attributions.

Chatelperronian

The best known of the IUP fossils are theNeandertal remains from Chatelperronianlevels at La Roche a Pierrot (Saint Cesaire:Leveque and Vandermeersch, 1980; Van-dermeersch, 1984) and the Grotte du Renne(Arcy-sur-Cure: Leroi-Gourhan, 1958; Hub-lin et al., 1996). The former comprises apartial skeleton of an adult male, recoveredfrom an apparent burial in the higher of twoChatelperronian levels at the site (EJOPsuperieur). Six burnt flints from this levelwere dated by thermoluminescence, result-ing in a mean calendar age of 36.3 6 2.7 kyBP (Mercier et al., 1991). The St. Cesaireskeleton is by far the most complete andinformative specimen associated with theIUP in Europe. In craniofacial and postcra-nial morphology, St. Cesaire 1 clearly con-forms to a Neandertal pattern (Levequeand Vandermeersch, 1980; Vandermeersch,1984; Stringer et al., 1984; Trinkaus et al.,1998, 1999), albeit with some details ofanatomy that are closer to modern humanmorphology (such as reduced midfacialprognathism, greater craniofacial gracility,reduced nasal piriform aperture width, andgreat width and projection of the humeral

deltoid tuberosity: Trinkaus et al., 1999;Wolpoff, 1999). Geometric analysis of long-bone cross sections suggests that in someaspects of behavior, most notably locomotorlevels, St. Cesaire was more similar to EarlyUpper Paleolithic modern humans thanMiddle Paleolithic Neandertals (Trinkauset al., 1998, 1999).

The remains from Arcy-sur-Cure includenine isolated teeth and a fragmentary tem-poral bone from an infant. On the basis ofsize and morphology, especially the expres-sion of taurodontism in the recovered mo-lars, Leroi-Gourhan (1958) attributed theremains to Neandertals. Given the uncer-tainty associated with classifying fossil ma-terial based on isolated teeth, and the pre-vailing idea (until the discovery of St.Cesaire) that the Chatelperronian wasmade by modern humans (see discussion ofCombe Capelle, below), Leroi-Gourhan’s at-tribution was for a long time consideredvery tentative. Indeed, Brose and Wolpoff(1971) even saw the dental material fromArcy-sur-Cure as indicative of a high degreeof Neandertal affinities in the earliest mod-ern humans in France. High-resolutioncomputed tomography of the temporal bonein 1996 revealed a morphology of the bonylabyrinth of the inner ear characteristic ofNeandertals (Hublin et al., 1996). While notall Neandertals are characterized by anidentical bony labyrinth morphology (Poncede Leon and Zollikofer, 1999), the patternnoted in the Arcy specimen appears to so-lidify its Neandertal affinities. The tempo-ral bone, derived from layer Xb, has beendated by 14C to 33.8 6 0.7 ky BP (Hedges etal., 1994). Given the roughly 10% age un-derestimation by 14C in this time range, thedates from St. Cesaire and Arcy-sur-Cureare highly concordant. It bears reiterating—given the questions that surround the cul-tural capacities of Neandertals—that Chat-elperronian levels at Arcy-sur-Cure andelsewhere have produced numerous itemsof personal adornment (see Zilhao andd’Errico, 1999).

To these remains we can add an isolatedtooth from Chatelperronian levels at Fontde Gaume, from the Dordogne of southwest-ern France (Gambier et al., 1990). The toothis an incompletely formed crown of a perma-


nent right C1, and based on the degree ofcrown formation it probably derived from achild of 2–4 years (Gambier et al., 1990). Aswith many Neandertal and modern humanmandibular canines, the crown has a largemedian crest, separated from the marginalcrests by distinct grooves. The region of thelingual cingulum is not fully formed, thusthe morphology of this feature cannot beevaluated. The specimen lacks the commonNeandertal characteristics of an expandeddistal border and a distolingual tubercle,but as noted by Gambier et al. (1990), theexpression of these features is variablewithin Neandertals. According to Gambieret al. (1990), this isolated tooth crown doesnot provide enough information to securelyattribute it to (or exclude it from) eitherNeandertals or early modern humans.

Another possibly Chatelperronian-associ-ated human fossil, this time a nearly com-plete adult male, was recovered by OttoHauser from the site of Combe Capelle (Dor-dogne, France) in 1909 (Klaatsch andHauser, 1910). The skeleton (destroyed dur-ing the Allied bombing of Berlin in WorldWar II) was that of an anatomically modernhuman, leading to the widespread miscon-ception for the better part of the 20th cen-tury that the Upper Paleolithic, in all of itsforms, was equated with modern humans.According to Gambier (1989), the excava-tions at Combe Capelle were conductedmainly by workmen and lacked strati-graphic control, and the stratigraphy of thesite and the provenience of the skeleton re-main open to interpretation. Gambier(1989) also notes that, morphologically, theCombe Capelle specimen would fit better ina sample of more recent Europeans than itwould with Upper Paleolithic modern hu-mans, and she suggests that the skeletonmay have been an intrusive burial frommore recent times. Given the questions sur-rounding the archeological provenience andmorphological affinities of the skeleton, andseeing as the fossil is no longer available fordirect dating or further study, the CombeCapelle skeleton cannot be considered evi-dence of a modern human-Chatelperronianconnection.

Uluzzian

Two human milk teeth were found in1965 in Uluzzian level E at Grotta del Cav-allo (Palma di Cesnola and Messeri, 1967),the type site if not the eponymous site forthe Uluzzian (Palma di Cesnola, 1993). Re-covered near the base of level E (EIII orpossibly EII), the remains include a left dm1

and, from 15–20 cm higher in the level, aright dm2. Level EIII contains an “early Ul-uzzian” (Palma di Cesnola, 1993), and islikely greater than 31 ky BP based on asingle 14C determination from the upperpart of level E (EII-I) (Bietti, 1997). Theearliest Uluzzian at the Campanian site ofGrotta di Castelcivita yielded radiocarbondates clustering around 33–32 ky BP (Bi-etti, 1997), likely the same approximate ageas the material in the base of Cavallolevel E.

The Cavallo dm1 is large-crowned, mea-suring 11.0 mm mesiodistally (MD) by 7.5mm buccolingually (BL). To Messeri andPalma di Cesnola (1976), the overall dimen-sions and morphology of this tooth com-pared favorably to European early modernhuman samples. On the basis of the pub-lished crown dimensions, the Cavallo dm1

appears to us to be more closely alignedwith Neandertals. The mesiodistal diameterof the tooth is large, well above the meanvalues of both Neandertals and early mod-ern humans (Neandertals, 8.6 6 1.1, n 5 9;early modern humans, 7.3 6 0.6, n 5 10:data from Frayer, 1978), but closer to that ofthe Neandertals. The buccolingual dimen-sion is small relative to both groups (Nean-dertals, 8.7 6 1.0, n 5 9; early modern hu-mans, 8.7 6 0.6, n 5 9: data from Frayer,1978). These dimensions produce a com-puted crown area (MD * BL) of 82.5, morethan two standard deviations above themean early Upper Paleolithic mean (63.3 68.2, n 5 9) but within one standard devia-tion of the mean for a sample of Neandertals(75.1 6 10.0, n 5 9: all data from Frayer,1978). The tooth is mesiodistally large rela-tive to its buccolingual dimension, which isopposite the condition usually seen in bothNeandertals and early modern humans(although this morphology can be seen inthe presumably Neandertal dm1 from


Gibraltar). According to Palma di Cesnolaand Messeri (1967), the tooth has a ratherlarge pulp chamber, which may suggest tau-rodontism, but which would also not be un-usual for an early modern human deciduousmolar (see below). In short, there is nothingin this tooth to rule out the possibility thatit derived from a Neandertal, or a modernhuman for that matter, although we feelthat its dimensions more closely align itwith the former.

The Cavallo dm2, from stratigraphicallyhigher in level E, is similar to Neandertalspecimens in size, cusp morphology, andtaurodontism (Messeri and Palma diCesnola, 1976). Measuring 10.0 mm (MD)by 11.0 mm (BL), the Cavallo dm2 has acrown that is large relative to, but withinthe range of, early modern human values.Both its MD and BL diameters are close tobut above mean values for small samples ofNeandertals and early modern humans(MD, 9.8 6 0.8 for 13 Neandertals, 9.6 6 0.8for 5 early modern humans; BL, 10.3 6 0.8for 13 Neandertals, 10.5 6 0.5 for 10 earlymodern humans: data from Frayer, 1978),and accordingly its computed crown area issomewhat above that of both groups (Cav-allo, 110 mm; Neandertals, 101.5 6 14.8,n 5 13; early modern humans, 103.3 6 13.4,n 5 4: data from Frayer, 1978). Thus, met-rically, the Cavallo dm2 could fit with eithersample, and only the observed taurodontismand occlusal morphology (see Palma diCesnola and Messeri, 1967, p. 258) suggestattribution to Neandertals. Taurodontism isgenerally characteristic of Neandertal de-ciduous molars (although it is often obfus-cated by root resorption) and is not evidentin any of the early modern human juveniles(n 5 16) radiographed by Skinner and Sper-ber (1982). Some of these early modern hu-man specimens (e.g., Laugerie-Basse 4 andBadger Hole) have somewhat enlarged pulpcavities but do not approach the conditionseen in Devil’s Tower (Gibraltar) and LaQuina 18. Thus the presence of taurodon-tism in this specimen would appear to linkit with the Neandertals. Palma di Cesnolaand Messeri (1967) were also struck by theextreme occlusal wear in this specimen(which made difficult their observation ofcusp morphology), a feature they felt re-

flected a tough vegetal diet. Unfortunatelylittle is known about wear patterns in juve-nile Neandertal and early modern humanteeth, and thus group assignment based onocclusal wear is clearly not justified.

Szeletian

Two sites have produced human fossils inassociation with Szeletian (or Szeletian-like) assemblages. In the Upper RemeteCave at Mariaremete (Hungary), a singleculture-bearing horizon produced a tinysample of lithics (n 5 13) and a single ma-rine shell thought to have been collected byhumans (Allsworth-Jones, 1990a). The lith-ics include Mousterian-type tools (sidescrapers and flakes) made by Levallois re-duction, and five bifacially worked imple-ments, considered by Gabori-Csank (1983)to be bifaces or bifacial scrapers. The overallcharacter of this small assemblage appearsto be Jankovichian (Szeletian sensu lato: seeAllsworth-Jones, 1990a), and according toGabori-Csank (1983, p. 284), “a modifiedform of the Micoquian.” These bifacialpieces were thought to be leaf points byAllsworth-Jones (1990a), and the assem-blage from Mariaremete could just as wellbe pre-Szeletian, namely a late Micoquianor Moustero-Levalloisian with leaf points(Allsworth-Jones, 1990a; Kozłowski, 1996).Also in this archeological level (layer 4)were found three heavily worn teeth, a rightI1, I2, and C1, all thought to belong to asingle Neandertal individual (Allsworth-Jones, 1990a). Unfortunately, no morpho-logically detailed defense of this attributionis given.

The Slovakian site of Dzerava Skala (orPalffy) produced the germ of a left M2 froma cryoturbated sediment containing bothAurignacian and Szeletian type artifacts(massive-base bone points and leaf points,respectively). Although soliflucted, the sed-iments showed some cultural superposition,with the leaf points being more abundant inthe lower portion (Prosek, 1951). The molarwas described by the site’s excavator, J. Hil-lebrand, as having a well-developed foveaanterior and as being generally similar tothe Krapina Neandertal mandibular molars(Hillebrand, 1914). Recent reassessments oflate Pleistocene fossil human remains from


the former Czechoslovakia (Svoboda et al.,1996) provide no further information on theDzerava Skala tooth. Thus the conclusionthat this specimen could be either Nean-dertal or early modern human (Smith,1984) is still valid. The same interpreta-tion would also apply for the Remete teeth(see above).

The Makers of the IUP

Based on the limited fossil evidence avail-able to us now, we can say only that 1) theMousterian and other European Middle Pa-leolithic industries (such as the Micoquianat Kulna) appear to have been made exclu-sively by Neandertals, 2) IUP industries,with the exception of the Bachokirian(which is treated with the Aurignacian be-low, and can likely best be considered aproto-Aurignacian or early archaic Aurigna-cian: Kołowski, 1996), appear to have theirroots in the late Middle Paleolithic indus-tries of their respective regions, 3) Neander-tals were making some of these IUP indus-tries (at least the Chatelperronian inFrance), and 4) all of the human fossils yetrecovered from IUP contexts for which anyindication of diagnostic morphology ispresent have their greatest morphologicalaffinities with Neandertals and not earlyUpper Paleolithic modern humans. In short,there is nothing in the archeological or pa-leontological records to refute the claim thatthe various manifestations of the non-Au-rignacian earliest Upper Paleolithic repre-sent in situ cultural evolution from MiddlePaleolithic (usually Mousterian) anteced-ents, and at the hands of Neandertals (cf.Kozłowski, 1996; Allsworth-Jones, 1990a).However, it must be remembered that, ex-cept for the Chatelperronian, the fragmen-tary and often undiagnostic nature of fossilhuman associations with IUP componentsdoes not allow this conclusion to be made ina definitive fashion.

HUMAN FOSSILS FROM THE EARLIESTAURIGNACIAN

Both the timing of the appearance of theAurignacian and its typological character-istics are issues of current debate. Prehis-torians have relied heavily on the classicsequence of five Aurignacian stages delin-

eated by Peyrony (1933), based largely onthe sequence of assemblages in southwest-ern France. The earliest of Peyrony’s stages,Aurignacian I, was characterized by bladeswith scalar retouch (Aurignacian blades),end-scrapers on Aurignacian blades, stran-gled blades, carinated and nosed end-scrap-ers fashioned on thick flakes or chunks,burins (although rare) and awls, batons decommandements, and split base pointsmade from bone (Bordes, 1968). The split-base bone point has played the role of fossiledirecteur for the early Aurignacian, a detailthat has complicated efforts to classify as-semblages outside of France. Often thesenon-French manifestations are lithically notlike the Aurignacian but contain split-baseand other bone points (e.g., Montet-White,1996). In France, the Aurignacian I gaveway to the Aurignacian II, in whose assem-blages lozenge-shaped points replace split-base points, Dufour bladelets are common,and the busked burin assumes the role oftype-fossil. De Sonneville Bordes (1960)added an earlier, “Aurignacian 0,” stage toPeyrony’s scheme. This came from the rec-ognition that what Peyrony had called“Perigordian II,” which lacked split-basebone points and Aurignacian blades andwhich commonly had Dufour bladelets, wasAurignacian in its other attributes (Har-rold, 1989). Aurignacian 0 (also variouslycalled the Correzian, the Archaic Aurigna-cian, Basal Aurignacian, Aurignacian An-cien, or proto-Aurignacian) has only rarelybeen found outside of France, and the defin-ing characteristics of the early Aurignacianremain a matter of continuing debate (seeHarrold, 1989; Montet-White, 1996; Mira-cle, 1998).

The use of type-fossils for classifying ar-tifact assemblages is clearly laden withproblems: bone points, for example, do notappear in all early Aurignacian assem-blages, and do appear in some later Moust-erian assemblages (e.g., at Große Grotte(Riek, 1934; Albrecht et al., 1972) and Sal-zgitter-Lebenstedt (Gaudzinski, 1999), bothin Germany: see also Montet-White, 1996);Aurignacian assemblages may also occa-sionally contain type-fossils from other in-dustries, such as Chatelperron knives orleaf points. Bordes (1968) clearly recognized


this problem and advocated that relativefrequencies of retouched tool types, ratherthan fossiles directeurs, should be used tocompare assemblages (see Miracle, 1998).But even this approach is fraught with prob-lems of intra- and intersite variability,small sample sizes, and sampling error thatmake discrimination of assemblages diffi-cult. Finally, it is not entirely clear that theconcept of unified lithic industries, repre-senting cultural commonality, can be ap-plied in a pan-European context. As ob-served by Straus (1995, p. 4):

. . . there is little purpose in repeating the traditional,normative characterizations of the classic industries,for it has been shown in recent years that all of them arehighly variable internally, intra- and interregionally,synchronically and diachronically, among and withinindividual sites. Much of this variability can be ex-plained in terms of sampling factors, differences in ac-tivities or site functions, artifact disposal modes, anddifferences in raw materials. The stylistic or ethnicaspects of Upper Paleolithic industries continue to behotly debated, but it is difficult to imagine that ethni-cally bounded units existed in the same sense that mod-ern ”cultures“ have existed over thousands of years andkilometers.

In short, archeologists are confrontedwith many of the same problems in trying toclassify artifact assemblages as paleontolo-gists are in classifying fossils: uncertaintyabout the validity of the typologic or taxo-nomic units, disagreement and doubtsabout the appropriate diagnostic criteria,uncertainty about how to interpret sites orspecimens with mixed diagnostic features,and small, fragmentary, and generally inad-equate samples for analysis. Added to thesedifficulties are ambiguities in dating, partic-ularly with older excavations in whichstratigraphic complexity was not fully ap-preciated and in which contamination ofsamples for conventional 14C determina-tions was not controlled. This situation hasmade research on the biological and culturalinteractions between Neandertals and earlymodern humans, and between Mousterianand Upper Paleolithic peoples, especiallyhard to decipher. Symptomatic of this diffi-culty is the current debate about the rela-tive chronological and stratigraphic posi-tions of IUP and Aurignacian industries.The Chatelperronian, for example, has con-ventionally been thought to interstratify

with the Aurignacian at some sites (le Rocde Combe, le Piage, and El Pendo: Bordesand Labrot, 1967; Gonzalez Echegaray,1980), and this observation has formed thecornerstone of Chatelperronian/Aurigna-cian acculturation scenarios. It has been ar-gued, however, that when due considerationis given to taphonomic factors and problemswith radiocarbon contamination and cali-bration, the remaining well-dated sites re-veal first an emergence of IUP cultures invarious regions, followed in time by the ap-pearance of the Aurignacian (d’Errico et al.,1998; Zilhao and d’Errico, 1999). The reex-amination by Zilhao and d’Errico (1999, p.43) of the archeological record of the Inter-pleniglacial led them to claim, “(n)owhere inEurope does the true Aurignacian—with itscharacteristic bone or ivory points and stonebarbs—date to before ca. 36,500 radiocar-bon years ago.” Zilhao and d’Errico’s per-spective has not gone unchallenged (e.g.,Harrold, 2000; Richter et al., 2000; and com-ments in d’Errico et al., 1998), and this re-mains an issue of heated debate.

Below, we examine the sites that haveproduced human fossils in association withthe earliest Aurignacian. These includesome sites in which fossils have been foundwith diagnostically Aurignacian assem-blages, others in which human remainshave been found in association with Aurig-nacian type-fossils but in which the appro-priate classification of the entire lithic andbone industry is unclear, and some withhuman remains with no archeological con-text but with direct dates that would sug-gest an Interpleniglacial age. There is anabundant, albeit generally fragmentary,fossil record from the later Aurignacian ofWestern Europe. A number of sites haveyielded human remains in association withPeyrony’s Aurignacian 1, such as LaCrouzade, Isturitz, and Les Rois in France(Gambier, 1989), Fontana Nuova in Italy(Chilardi et al., 1996), and Istallosko inHungary (Vertes and de Vries, 1952; Malan,1954). In France, the Aurignacian 1 has 14Cdates clustering around 31.5–30.5 ky BP(Delibrias and Fontugne, 1990). At ca. 30 kyBP in age (Movius, 1969), the modern hu-man remains found with an evolved Aurig-nacian assemblage (De Sonneville Bordes,


1959) at Cro-Magnon can also be placedwith this group. In every case that a confi-dent taxonomic diagnosis can be made,these later Aurignacian remains representmodern humans (see Gambier, 1989, 1993,1997). There is a temptation to view theseearly modern humans as the sole propri-etors of the Aurignacian, and to deputizethe Cro-Magnon to fill the early Aurigna-cian fossil void. In truth the makers of theearliest Aurignacian are not well-known,despite a growing (but still diagnosticallydifficult) fossil record. We have thus limitedour review to human fossils associated withthe earliest Aurignacian, and present thefossil record in roughly chronological orderand without regard to geography. In generalwe devote somewhat greater attention tothe more diagnostically difficult material(simply because it is more difficult to diag-nose and therefore warrants greater atten-tion to detail).

El Castillo (Santander, Spain)

The cave of El Castillo, with 20 m of strat-ified cultural deposits spanning virtuallythe entire duration of the European Paleo-lithic, is one of the cornerstones of Spanishprehistory. Excavation was undertaken in1910 by Hugo Obermaier (in collaborationwith other noted prehistorians of the day,including the Abbe Breuil and Paul Wern-ert), and by 1912 Aurignacian levels in thecave had been reached (Cabrera Valdes,1984). Obermaier (1924) identified threecultural levels containing Aurignacian arti-facts: layer o, a late Aurignacian containinggravers and typical Gravette points, under-lain by a sterile clay layer n; layer m, an-other late Aurignacian level, underlain by anearly sterile clay layer i; layer h, withObermaier’s “Aurignacian delta,” underlainby a sterile stalagmitic layer g (actually asterile silt: Butzer, 1981), and below that, alayer (f) containing a typical Mousterian as-semblage. Analysis of cultural materialfrom stratum h (now designated level 18)excavated by Obermaier and later (begin-ning in 1980) by Victoria Cabrera Valdesreveals an assemblage characteristic of thebasal Aurignacian of Cantabrian Spain (Ca-brera Valdes, 1984). Artifacts recoveredfrom level 18 include split-base bone points

(more common in the lower portion of thelayer), Aurignacian blades, nosed and cari-nated end scrapers, and occasional dihedralburins (Cabrera Valdes, 1984). A singleChatelperronian point was also recoveredfrom level 18 (Cabrera Valdes and Bischoff,1989). Initial AMS radiocarbon determina-tions on three small pieces of charcoal fromlevel 18 yielded dates ranging from 40.0 62.1 to 37.7 6 1.8 ky BP (Cabrera Valdes andBischoff, 1989). Seven additional determi-nations, performed by two different labora-tories, have reinforced the original dates,and suggest that the base of level 18 (18c)was deposited during the beginning of theHengelo Interstadial, at approximately 40ky BP, while the top of the level (18b1) ac-cumulated around 38.5 ky BP (CabreraValdes and Bernaldo de Quiros, 1996).

In the course of the early excavations atEl Castillo, Obermaier recovered humanfossil material from the basal Aurignacianlevel 18. The remains derive from at leasttwo individuals, an adult (El Castillo B) anda child of 3–5 years (El Castillo C) (Gar-ralda, 1989, 1997). Detailed descriptions ofthe human remains were never published,and the fossils are now lost. Brief accountsof the Castillo human material have beenput forth by Garralda (1989) and Garraldaet al. (1992), based on an unpublished 1933description of the material by Henri Vallois.The adult remains included a slightly wornright mandibular second molar and threesmall yet robust cranial fragments. Thechild was represented by a partial mandible(Fig. 3), preserved from the region of thesymphysis to the right-side unerupted firstpermanent molar and containing the firstand second deciduous molars (and the firstadult molar in its crypt), as well as sevensmall cranial fragments (Garralda, 1989;Garralda et al., 1992).

As to the morphologic affinities of theCastillo Aurignacian people, the dental di-mensions (Table 2) are largely uninforma-tive. The El Castillo B adult mandibularsecond molar is relatively large by both Ne-andertal and early modern human stan-dards, but could fit comfortably within therange of variation of either group in terms ofboth size and shape (Table 2). Both of the ElCastillo C deciduous molars have mesiodis-


tal (MD) diameters equal to or greater thanthe mean values for both comparativegroups, but exhibit small buccolingual (BL)diameters (Table 2). The El Castillo C dm1is mesiodistally long relative to a sample oftwo early modern humans, but falls right onthe Neandertal mean (also from a smallsample) for this dimension. However, thedm1 is more than three standard deviationsbelow the mean of the Neandertal sample inBL diameter, yet is close to the mean de-rived from two early modern humans. Thisresults in a crown surface area intermediatebetween the two samples (but some 6.7standard deviations below the Neandertalmean) and a crown index well below the

mean of two modern humans (and fartherstill [1.5 standard deviations] below the Ne-andertal mean). The same pattern is seen inthe El Castillo C dm2, with the exceptionbeing that its crown area is very close to theearly modern human sample mean (thistime based on six specimens). The crownindex in the El Castillo C dm2 is 2.5 and 1.9standard deviations below the early modernand Neandertal means, respectively. Gar-ralda (1997) noted that the symphysis of theEl Castillo C mandible is robust and lacksan accentuated chin, noting (p. 158), “Thehuman remains from El Castillo confirm thegeneral robusticity of early Upper Paleo-lithic humans and also show some archaicfeatures, but it is impossible to classifythem as ‘evolved’ Neandertals or ‘archaic’modern humans.” Based on the limitedamount of information available on theseremains, we agree that an unequivocal as-sessment of their affinities is not possible.

Bacho Kiro (Balkan Mountains, Bulgaria)

The fragmentary human remains fromBacho Kiro Cave, Bulgaria, enter into virtu-ally every consideration of the Middle-to-Upper Paleolithic and Neandertal/modernhuman transitions. Associated with an“Aurignacoid” lithic and bone industry(Kozłowski et al., 1982), the oldest of theseremains derive their importance first fromradiometric dates that may grant them sta-tus as the earliest known makers of theUpper Paleolithic in Europe (but see below),and second from typological and taxonomicambiguities in the lithic and hominid fossilsamples, respectively, that allow consider-able latitude in their interpretation. Theseremains have been variously seen as repre-senting the earliest modern human makersof the Aurignacian, as modern humans as-sociated with an IUP industry, or even aspossible Neandertals with an IUP or evenAurignacian association (see Wolpoff, 1996;Miracle, 1998).

A total of eight human fossils was recov-ered from four stratigraphic levels at BachoKiro (Glen and Kaczanowski, 1982). Sevenof these fossils can reasonably be placed intothree temporal groups (the eighth specimen,a deciduous mesial incisor crown, is not well

Fig. 3. El Castillo C mandibular fragment in occlu-sal (top) and inferior (bottom) views. Redrawn fromGarralda et al. (1992).


provenienced and is not considered furtherhere) (Table 3).

The first, and stratigraphically oldest,group contains a single fossil: a fragment ofthe left side of a mandibular corpus contain-ing a first deciduous molar. This specimenwas recovered from level IV at the base oflayer 11 (Ginter and Kozłowski, 1982b),along with artifacts classified as Ba-chokirian (an “Aurignacoid” early Upper Pa-leolithic industry of Central Europe: see be-low). Level 11 occurs between 354–380 cmbelow datum (Ginter and Kozłowski,1982a), with subunit IV present at a depthof 375–380 cm. Charcoal from the top oflayer 11 (at a depth of 356–357 cm) wasdated by conventional radiocarbon to be.43 ky BP (Mook, 1982). Ginter andKozłowski (1982b) place layer 11 in a periodof warming and humidification based onsedimentological analysis, a greater abun-dance of Pitymys subterraneus relative to

the more dry-habitat Microtus arvalis, andan increased abundance of fish and themarsh-loving European mole, Talpa euro-paea. If the radiocarbon date for the upperreaches of layer 11 is correct, then thiswarm period may correspond to the Moer-shoofd (Heraklitsa) temperate oscillation(Ginter and Kozłowski, 1982b), making theBachokirian the earliest dated Upper Paleo-lithic in Europe. However, bone, charcoal,and dental samples from layer 11 submittedto the Oxford accelerator laboratory in 1990produced AMS radiocarbon dates rangingfrom 38.5 6 1.7 to 33.8 6 0.9 ky BP (Hedgeset al., 1994). Two of these dates (38.5 6 1.7ky BP on bone and 37.7 61.5 ky BP oncharcoal) are not significantly different fromone another, but are more than two stan-dard deviations older than the two youngestdates obtained from layer 11 (34.8 61.2 kyBP on a tooth, and 33.8 6 0.9 ky BP onbone). These younger dates suggest that the

TABLE 2. Dimensions of teeth from El Castillo, level 18, relative to Neandertals and early modern humans1

Neandertals El Castillo Early moderns

M2Mesiodistal length 11.8 6 0.7 (26) 12.0 11.3 6 1.0 (22)Buccolingual breadth 11.3 6 0.7 (26) 11.5 10.8 6 0.8 (22)Computed crown area 132.9 6 14.3 (26) 138.0 122.6 6 19.0 (20)Crown index 95.7 6 5.6 (26) 95.8 96.9 6 5.0 (20)

dm1Mesiodistal length 9.0 6 0.5 (7) 9.0 8.6 (2)Buccolingual breadth 7.7 6 0.2 (7) 7.0 7.1 (2)Computed crown area 69.1 6 0.9 (7) 63.0 60.3 (2)Crown index 85.8 6 5.4 (7) 77.8 82.4 (2)

dm2Mesiodistal length 10.6 6 0.6 (14) 11.0 10.6 6 0.6 (6)Buccolingual breadth 9.7 6 0.5 (14) 9.0 9.4 6 0.6 (6)Computed crown area 102.7 6 9.7 (14) 99.0 99.1 6 11.2 (6)Crown index 91.4 6 4.9 (14) 81.8 88.8 6 2.8 (6)

1 Dimensions of El Castillo teeth from Garralda (1997); Neandertal and early modern human data from Frayer (1978). Computedcrown area 5 mesiodistal length p buccolingual breadth. Crown index 5 100 p buccolingual breadth/mesiodistal length.

TABLE 3. Stratigraphic context of Bacho Kiro human fossils1

Bacho Kirospecimen no. Element Level Carbon 14 date Cultural association

559 Right mandibular corpuswith dm2 and M1

6a/7 29.15 6 0.95 ky BP Aurignacian

1702 RP4 6a/7 29.15 6 0.95 ky BP Aurignacian1704 RC1 6a/7 29.15 6 0.95 ky BP Aurignacian2641 RI2 7 None Aurignacian3575 Right parietal fragment 7 None Aurignacian2823 RI1 7/6b 32.15 Aurignacian?

32.7 6 0.3 ky BP1124 Left mandibular corpus

with dm1

11-IV .43.0 Bachokirian (“proto-Aurignacian”)

W-1 di1 ? None ?1 From Kozłowski (1982).


warmer and wetter conditions reflected inthe layer 11 sediments correspond to theHengelo (Podrahem) interstadial ratherthan the Heraklitsa.

As pointed out by Mellars (in Hedges etal., 1994), the range of dates from layer 11indicates either stratigraphic misplacementof samples, serious contamination effects, ora long duration for the accumulation of thelayer. A similar situation occurs at TemnataCave, also in the Bulgarian Balkan Moun-tains. Here artifacts attributed to the Ba-chokirian are distributed throughout a sin-gle level (layer 4), with radiocarbon datessuggesting a 5,000–9,000 year period of ac-cumulation of the sediments (Ginter et al.,1996). Recent 14C AMS determinationsplace the occupation at the bottom of Tem-nata layer 4 between 39.1 6 1.8 and 38.3 61.8 ky BP, while the top of the layer dates tobetween 33.0 6 0.9 and 31.9 6 1.6 ky BP(Ginter et al., 1996). If the Bachokirian oc-cupations of Bacho Kiro and Temnata werepenecontemporaneous, the dates from Tem-nata would support a Hengelo/Podrahemdate for Bacho Kiro layer 11. The emergenceof the Bachokirian, then, may correspond tothe emergence of the Upper Paleolithic gen-erally across all of Europe.

The single human specimen from BachoKiro layer 11 does not provide many usefulfeatures for taxonomic classification. Thedeciduous first molar in this mandibularfragment was described by Glen and Kacza-nowski (1982) as having a relatively wornocclusal surface and a large mesio-buccaltubercle. The crown dimensions are inter-mediate between small samples of Neander-tals and early modern Europeans (Table 4).The buccolingual diameter of the crown isrelatively large, and in this feature and inits relatively high crown index (produced bythe large BL diameter), the specimen iscloser to the Neandertal mean. Indeed, Wol-

poff (1996) noted that the layer 11 molar hasexactly the same dimensions as the Nean-dertal dm1 from Taubach. However, smallsample sizes prohibit a reliable assessmentof the range of variation in the comparativesamples, and it is unlikely that the dimen-sions of the Bacho Kiro tooth are signifi-cantly different than those of early UpperPaleolithic modern humans. Radiography(Glen and Kaczanowski, 1982, p. 75) re-vealed a nontaurodont root to the dm1,which might suggest modern human affini-ties. As previously noted, the Neandertaldeciduous molars reported by Skinner andSperber (1982) are characterized by taur-odontism, as is the Uluzzian-associatedCavallo dm2.

Kozłowski et al. (1982) place the lithicassemblage in layer 11 into an “Aurignacoi-dal tradition,” but as “Bachokirian” ratherthan true Aurignacian. Assemblages thathave been classified as Bachokirian areknown from only a handful of other sites,including Istallosko in Hungary, and Tem-nata Cave, Pest, and V. Levski in Bulgaria(Hahn, 1993). The lithic elements of the Ba-chokirian are typologically and technologi-cally Aurignacian, emphasizing blade pro-duction from cores with typical UPpreparation (single- and opposite-platformcores), retouched blades, carinated end-scrapers, and with Aurignacian retouch onsome retouched blades (Delporte and Djind-jian, 1979; Kozłowski et al., 1982). Perfo-rated bear and fox teeth were recoveredfrom layer 11, suggesting personal adorn-ment as is generally characteristic of earlyUpper Paleolithic cultures. These featureshave led many researchers to consider thelayer 11 assemblage as a very early or“proto-Aurignacian” (e.g., Mellars in Hedgeset al., 1994), archaic Aurignacian (Kozłowski,1996), or Early Balkan Aurignacian (Ginteret al., 1996). The main distinction between

TABLE 4. Dimensions of dm1 from Bacho Kiro, level 11, relative to Neandertals and early modern humans1

Neandertals Bacho Kiro Early moderns

Mesiodistal length 9.0 6 0.5 (7) 8.8 8.6 (2)Buccolingual breadth 7.7 6 0.2 (7) 7.5 7.1 (2)Computed crown area 69.1 6 0.9 (7) 66.0 60.3 (2)Crown index 85.8 6 5.4 (7) 85.2 82.4 (2)1 Dimensions of Bacho Kiro 1124 from Glen and Kaczanowski (1982); Neandertal and early modern human data from Frayer (1978).Computed crown area 5 mesiodistal length p buccolingual breadth. Crown index 5 100 p buccolingual breadth/mesiodistal length.


it and more classic Aurignacian assem-blages lies in the absence of bone points inthe Bachokirian (which is also characteris-tic of the basal Aurignacian [Aurignacian 0]in France). It is also interesting that, unlikeother IUP industries, the Bachokirian rep-resents a break with the local Mousterian inhaving a fully developed Upper Paleolithicmethod of blank production (with no dis-cernible tradition of the Levallois tech-nique) and a low count of typologically Mid-dle Paleolithic elements (e.g., side-scrapers,points), suggesting it is allochthonous(Kozłowski, 1996), possibly with ties to theearly Aurignacian (Baradostian) of theZagros Mountains (Olszewski and Dibble,1994). The Bachokirian also differs fromCentral European IUP cultures, but shareswith the Aurignacian, in its lack of leafpoints and backed pieces (Kozłowski et al.,1982).

The second group of human fossils fromBacho Kiro includes a right central perma-nent mandibular incisor from the level 7/6binterface, and a fragment of a right parietaland a right lateral permanent mandibularincisor from level 7. The top of level 6b (be-low the contact with level 7)6 has been ra-diocarbon dated to 32.7 6 0.3 ky BP (Mook,1982), and according to Glen and Kacza-nowski (1982), the human I1 from the inter-face of 7/6b was radiocarbon dated to 32,150years BP (errors and other details are notgiven by the authors). The level 7/6b inter-face yielded 76 lithic and four bone artifacts(Kozłowski et al., 1982). The lithics includedend-scrapers, burins, retouched blades, re-touched flakes, and side scrapers, and theosseous artifacts included a bone point witha triangular cross section. While clearly Up-per Paleolithic in character, the level 7/6bassemblage is too small to allow for a confi-dent assessment of its typological attributes(Kozłowski et al., 1982). The middle of level7 contained a larger lithic sample (654pieces) but only slightly more bone tools(11), including the base of a bone point, ovalin section (Kozłowski et al., 1982). The lithicand osseous assemblages in this level are

generally similar to those of level 7/6b, andKozłowski et al. (1982) classified this assem-blages as typical (Balkan) Aurignacian.Sedimentological and faunal data suggestdry but relatively warm conditions as level 7began to form, but with increasingly coolertemperatures throughout level 7 times (Gin-ter and Kozłowski, 1982b). The radiocarbondeterminations and the climatic indicatorssuggest that the base of level 7 was depos-ited during the early part of the Denekamptemperate period.

As with the material from level 11, thehuman remains from level 7 do not providemuch diagnostic information. The parietalfragment is thick (5 mm), but within therange of variation for recent humans (2–5mm: Glen and Kaczanowski, 1982). The twoadult mandibular incisors from levels 7/6band 7 have their greatest size and shapesimilarity with the early Upper Paleolithicsample (Table 5). Both of these incisors havereduced BL diameters relative to Neander-tals, resulting in incisors that have crownindices that are very close to the modernhuman sample means but more than onestandard deviation away from Neandertalmeans.

The third group of human fossils fromBacho Kiro derives from the contact betweenlevels 7 and 6a. Artifacts from this depthinclude nosed and carinated end-scrapers,Dufour bladelets, and bone points with cir-cular cross sections, leading Kozłowski et al.(1982) and Ginter and Kozłowski (1982b) toconsider them representative of the typicalAurignacian. A single 14C determinationfrom the 6a/7 contact gave a date of 29.2 61.0 ky BP (Mook, 1982), and climatic indi-cators suggest relatively warm and humidconditions (Ginter and Kozłowski, 1982b),which may indicate deposition during theDenekamp interstadial.

The human remains from the 6a/7 contactinclude a fragment of right-side mandibularcorpus containing the second deciduous mo-lar and the first permanent molar, a rightmaxillary permanent fourth premolar, anda right maxillary permanent canine. Again,the available fossil evidence is fragmentaryand includes juvenile specimens, makingtaxonomic diagnosis difficult. The ambigu-ity inherent in taxonomically diagnosing

6The strata at Bacho Kiro are numbered, from highest tolowest, as follows: 1, 2, 3, 4, 3a, 5, 4a, 4b, 6a, 7, 6b, 8, 6c, 9, 10,11, 11a, 12, 13, 13h, and 14.


the Bacho Kiro remains is apparent whenone compares the size and shape of the fourteeth from this level to those from samplesof available Neandertal and early Upper Pa-leolithic (Aurignacian and Gravettian) mod-ern human teeth (Table 5). Glen and Kac-zanowski (1982) remarked merely that theBacho Kiro dentition falls metrically be-tween Neandertal and early modern humanvalues, albeit somewhat closer to theformer. They further note that the adultcanine and fourth premolar are metricallymost similar to the small sample of Nean-dertals they included in their comparativedata. In terms of mesiodistal (MD) and buc-colingual (BL) crown diameters, computedcrown areas (MD * BL), and crown indices(100 * BL/MD), the Bacho Kiro M1 and C1

are indistinguishable from either the Nean-dertal or early modern human samples. TheBacho Kiro P4 is large relative to the meansof both comparative samples, in large partdue to its great MD diameter. The relativelygreat MD length of this premolar gives it acrown index that is well below the meanindex values of both comparative groups

(1.3 standard deviations below the modernhuman mean, 1.5 standard deviations belowthe Neandertal mean). Thus the initial as-sessment by Glen and Kaczanowski (1982)of the C1 and P4 similarities to Neandertalteeth does not hold when larger fossil sam-ples are taken into consideration. Althoughvalues are given for the crown dimensions ofthe deciduous second molar, heavy occlusalwear precludes an accurate assessment ofthe size and shape of this tooth.

Unfortunately, the Aurignacian-associ-ated teeth from Bacho Kiro preserve fewdiagnostic characteristics that would behelpful in determining their affinity. Most ofthe discrete traits that differentiate Nean-dertal from modern human teeth (see Table13 in Bermudez de Castro et al., 1999) arenot of help in this case, since they involveeither teeth not represented in the BachoKiro sample, or require adjacent pairs ofteeth for examination of size sequences, orconsist of crown features that do not un-equivocally separate the groups. For exam-ple, the Bacho Kiro M1 (from the 6a/7 con-tact) has a small but clearly developed

TABLE 5. Dimensions of teeth from the Aurignacian levels of Bacho Kiro, relative to Neandertals and earlymodern humans1

Neandertals Bacho Kiro Early moderns

I1Mesiodistal length 6.0 6 0.5 (10) 5.7 5.8 6 0.7 (12)Buccolingual breadth 7.4 6 0.4 (14) 6.3 6.4 6 0.4 (13)Crown index 126.6 6 11.4 (8) 110.5 111.9 6 12.0 (12)

I2Mesiodistal length 6.8 6 0.6 (18) 6.7 6.4 6 0.7 (15)Buccolingual breadth 8.0 6 0.5 (20) 7.3 7.0 6 0.6 (17)Crown index 120.3 6 9.8 (16) 109.0 108.8 6 11.9 (15)

dm2Mesiodistal length 10.6 6 0.6 (14) (10.0) 10.6 6 0.6 (6)Buccolingual breadth 9.7 6 0.5 (14) (9.5) 9.4 6 0.6 (6)Computed crown area 102.7 6 9.7 (14) (95.0) 99.1 6 11.2 (6)Crown index 91.4 6 4.9 (14) (95.0) 88.8 6 2.8 (6)

M1Mesiodistal length 11.7 6 1.0 (31) 12.3 11.6 6 0.9 (27)Buccolingual breadth 11.1 6 0.7 (32) 11.5 11.0 6 0.6 (28)Computed crown area 129.6 6 15.8 (31) 141.5 127.1 6 14.8 (27)Crown index 95.7 6 9.1 (31) 93.5 94.9 6 6.1 (27)

C9Mesiodistal length 8.2 6 0.6 (17) 8.5 8.1 6 0.6 (11)Buccolingual breadth 9.4 6 0.6 (18) 9.3 9.0 6 1.0 (11)Computed crown area 76.5 6 8.0 (17) 79.1 73.3 6 13.1 (11)Crown index 115.2 6 12.6 (17) 109.4 112.0 6 7.3 (11)

P4

Mesiodistal length 7.0 6 0.6 (16) 7.9 7.1 6 0.5 (13)Buccolingual breadth 10.4 6 0.8 (18) 10.2 9.7 6 0.7 (15)Computed crown area 72.8 6 9.7 (16) 80.6 69.6 6 9.7 (13)Crown index 150.9 6 14.3 (16) 129.1 137.4 6 6.2 (13)

1 Dimensions of Bacho Kiro teeth from Glen and Kaczanowski (1982); Neandertal and early modern human data from Frayer (1978).Computed crown area 5 mesiodistal length p buccolingual breadth. Crown index 5 100 p buccolingual breadth/mesiodistal length.


hypoconulid. According to Bermudez deCastro et al. (1999, p. 557), in Neandertalsthe hypoconulid is reduced in size and/orincidentally absent in M1 and M2, while inmodern humans it is “frequently absent inM2 and less so in M1.” Thus without traitfrequencies based on large samples, it isimpossible to evaluate the taxonomic rele-vance of this character. Taurodontism of themolar roots is one feature that could poten-tially be of benefit, but regrettably no radio-graphs of the Bacho Kiro M1 have yet beenpublished.

Our overall evaluation of the Bacho KiroAurignacian adult dental sample indicatesthat these teeth are largely ambiguous withrespect to taxonomic characteristics, butthat those few aspects of size, shape, andcrown morphology that do distinguish themfrom one or the other of the comparativesamples tend to align them with modernhumans.

Hahnofersand (Elbe Valley, Germany)

An isolated frontal bone, the Hahnofer-sand specimen reflects a robust modern hu-man in its vault and brow ridge architecture(Brauer, 1980, 1981; Smith, 1984). As withother early modern Europeans, the Hahnof-ersand frontal (Fig. 4) has well-developedsuperciliary arches with marked thinninglaterally (Smith, 1984), and with separationof the superciliary arch and the supraorbitaltrigone (Brauer, 1980). The Hahnofersandfossil appears to have a rather low frontalangle (60° 6 5°: Brauer, 1980), but this an-gle is difficult to measure accurately on thisspecimen. Hahnofersand also shares withlater Upper Paleolithic crania from Pader-born and Oberkassel the presence of a smallmidsagittal ridge (Smith, 1984). Overall,Hahnofersand is modern in morphology, butwith some primitive or even Neandertal-reminiscent features (low frontal angle,prominent brow ridges, and large overalldimensions: Smith, 1984). The strength ofthese features is reflected in morphometricanalyses that group Hahnofersand eitherwith Neandertals or in a position interme-diate between Neandertals and early mod-ern human crania (Brauer, 1980, 1981).

Discovered in a sand deposit of the ElbeRiver near Hamburg in1973, the Hahnofer-

sand frontal appears to have been second-arily deposited and has no archeologicalcontext. Both amino-acid racemization(AAR) and conventional radiocarbon datingof residual bone collagen give consistent ab-solute ages for Hahnofersand at ca. 36 kyBP (AAR, 36 ky BP; 14C, 36.3 6 0.6 ky BP;Brauer, 1980). If these dates can be ac-cepted, this makes the Hahnofersand fron-tal the earliest taxonomically diagnosablespecimen of a modern human yet recoveredin Europe. Hahnofersand obviously does nothelp to clarify the issue of who made whichlithic industry, but it does tell us, if the AAR

Fig. 4. Hahnofersand frontal. a: Anterior. b: lateral.


and 14C dates are correct, that modern hu-mans were installed in North-Central Eu-rope by end-Hengelo times. Given the lackof a clear geological or archaeological con-text for this specimen, along with the recentsuccesses of AMS dating, the time has prob-ably come to redate this fossil.

Vogelherd (Swabian Jura, Germany)

Vogelherd Cave, near the town of Stetten,preserves 2–4 m of cultural deposits, ex-tending from the early Mousterian to theNeolithic. After excavations in 1931, GustavRiek identified what he believed to be threeAurignacian phases at Vogelherd Cave,which he designated lower, middle, and up-per (Riek, 1934). Riek’s “lower Aurignacian”was contained in horizon VI, and was rep-resented by only a few stone and two bonetools. These bone tools are large bone points,which are also found in association with lateMousterian assemblages in some areas(Montet-White, 1996). This level also lackedmammoth ivory animal and anthropomor-phic figurines like those recovered from theAurignacian in higher levels. On reexami-nation of the material from this component,Muller-Beck (1983, p. 250–251) attributedthe tools to the late Mousterian. Riek’s“middle Aurignacian” from horizon V haslikewise been redesignated as early Aurig-nacian (Muller-Beck, 1983). The typicallyAurignacian tools and carvings derive fromthis layer, and from the late Aurignacian ofhorizon IV (Hahn, 1983).

Human skeletal elements from at least 3adult individuals were recovered from levelsV and IV (Gieseler, 1937, 1940, 1971;Czarnetzki, 1980, 1983; Smith, 1984). Thefirst specimen, Vogelherd 1 (also known asStetten 1), consists of a modern human cra-nium (lacking most of the face; Fig. 5) and apartial mandible (Fig. 6), found 0.15 m eastof the cranium and 0.33 m deeper. Bothwere found in a “Brandschicht” or burnedlayer in horizon V in front of the southwestentrance (Riek, 1932, 1934). Two lumbarvertebrae, of identical color to the skull,were found 2–3 m east of the cranium in thesame layer and were also attributed to Vo-gelherd 1 (Riek, 1932). Czarnetzki (1980)also mentions the presence of an unde-scribed second metacarpal, presumably

from horizon V, which may belong to thisindividual. There is some confusion in theliterature concerning where in horizon Vthe Vogelherd 1 remains were found. Gie-seler (1937, p. 42) reported that the craniumand burned layer were located between the“middle and upper Aurignacian,” whichwould place it at the top of horizon V. How-ever, Riek (1934, p. 40–41), the cave’s exca-vator, clearly describes and illustrates Vo-gelherd 1 and the burned layer to lie at thebase of horizon V. Riek’s placement of thespecimen, based on his personal observationof the cave’s stratigraphy, is likely to besounder than Gieseler’s, but either waythere is little doubt that the specimen de-rived from the early Aurignacian compo-nent of Vogelherd.

Fig. 5. Vogelherd 1 cranium. a: Anterior. b: Lateral.


The second individual, Vogelherd 2 (Stet-ten 2) (Fig. 7), was discovered just inside thesouth entrance, separated by the entirelength of the cave from Vogelherd 1. Accord-ing to Riek (1934), it was found at the top ofthe Upper Aurignacian layer (now level IV),and there were noncultural layers depositedabove it in this area of the cave (a layer offine splintered yellow-white limestone).While Riek (1934) pointed out that therewas no mixing of this limestone layer withthe Aurignacian layer in this area, Gieseler(1937) believed that Vogelherd 2 might havebeen deposited by later Magdalenian peo-ples.

The third individual, Vogelherd 3 (Stetten3), consists of a robust right humerus (Fig.8), lacking its proximal end. According toGieseler (1937, p. 43), this specimen derivedfrom near the middle of the cave at the baseof the “middle Aurignacian” (now early Au-rignacian, horizon V). Furthermore, Gie-seler (1937) stated that the specimen wasexcavated partly from this level and partlyfrom the underlying sterile layer comprisedof large, coarse pieces of limestone rubble(“grobstuckiger Kalkschutt”). According toGieseler (1937, p. 43), “Riek assumes that,according to the profile, this humerus isprobably somewhat older than the Stetten Iskull.”

Radiocarbon dates from horizons IV(30.73 6 0.75 ky BP) and V (30.16 6 1.34and 31.9 6 1.1 ky BP) indicate a late Inter-pleniglacial age for the Vogelherd Aurigna-cian (Muller-Beck, 1983). The 14C determi-nations from Vogelherd horizon V provide aminimum age for the humerus, however,since it was recovered from the base of thishorizon. Some finer chronological resolutioncan be had by considering dates from theAurignacian levels at the German site ofGeißenklosterle. Here a German “proto-Au-rignacian” occurs in horizon IIIa, with a typ-ical Aurignacian component in IIa–IId(Hahn, 1983, 1996; but see Zilhao andd’Errico, 1999, concerning the possibilitythat Hahn’s proto-Aurignacian is an IUPassemblage with cryoturbational mixingfrom Aurignacian levels). Climatic recon-

Fig. 6. Vogelherd 1 mandible, lateral view.

Fig. 7. Vogelherd 2 cranium. a: Anterior. b: Lateral.


struction suggests that Vogelherd horizonIV and Geißenklosterle horizons IIa and IIbcorrespond to cold-dry periods (likely WurmIIIa), while Vogelherd horizon V and Gei-ßenklosterle horizons IId and III correspondto cold-humid periods. The humid climaticconditions suggest a Wurm II/III intersta-dial age (Hengelo temperate oscillation),which would be consistent with dates from

the proto-Aurignacian levels at Geißenk-losterle. AMS radiocarbon dates from thebase of Geißenklosterle (horizon III) rangefrom 40.2 6 1.6 to 37.3 6 1.8 ky BP, with acentral tendency of 38.4 6 0.9 (Richter etal., 2000). Richter et al. (2000) also obtainedsix TL dates on burned flints from level IIIa,producing an average age of 40.2 6 1.5 kyBP. Given the systematic underestimationof 14C dates in this time range, the TL datesare concordant with the AMS determina-tions for the level. The taphonomic situationat Geißenklosterle is complex, as is the ty-pological diagnosis of the artifact assem-blage (see Zilhao and d’Errico, 1999). Obvi-ously we need better chronological datafrom Vogelherd if we wish to place the hu-man remains from that site in time. How-ever, if Geißenklosterle level III correspondsto Vogelherd V (as suggested by climaticindicators and artifacts), then the Vo-gelherd humerus (found at the very base ofhorizon V) and the Vogelherd 1 cranium(also from the base of horizon V according toRiek, 1934) would date to the Wurm II/IIIinterstadial or early Wurm IIIa times.

From a morphological standpoint, bothVogelherd crania unequivocally representmodern humans (Gieseler, 1937), albeitwith some archaic features (Smith, 1984;Frayer et al., 1993). Both crania have high-domed vaults with steep frontal angles andexhibit a level of occipital bunning (slightlymore pronounced in Vogelherd 1; Figs. 5, 7)characteristic of most early Upper Paleo-lithic crania. Both specimens also preservethe lateral portions of the supraorbitalridges, which are only moderately project-ing and fall comfortably within the range ofother early modern humans in both thick-ness and projection (Smith, 1984, p. 154). Inrear view, both Vogelherd 1 and 2 have ver-tical sides and exhibit vault contours thatgenerally resemble modern humans. How-ever, both crania have their maximumbreadths either low on the parietals (Vo-gelherd 2) or at the level of the supramas-toid crest (Vogelherd 1). In these featuresthe Vogelherd crania are similar to the maleMladec 5 and 6 crania (Smith, 1984; Frayer,1986). The Vogelherd 1 mandible (Fig. 6)has a well-developed chin with a distinctmental trigone. The mandible also exhibits

Fig. 8. Vogelherd 3 right humerus, anterior view.


a retromolar space and a horizontal-ovalconfiguration of the mandibular foramen,both of which are more characteristic of Ne-andertals than early modern humans(Smith 1984; Frayer et al., 1993).

The Vogelherd 3 humerus is robust andhas large, rugose muscle markings. Gieseler(1937, 1940) stated that these features sug-gest Vogelherd 3 might represent a Nean-dertal and that its stratigraphic positionsupported that possibility. However, basedon an analysis of humeral epiphyseal anddiaphyseal shape and strength measures ofVogelherd 3 relative to humeri of Neander-tals and early modern humans, Churchilland Smith (2000) concluded that the speci-men derived from an anatomically modernhuman. The Vogelherd humerus shareswith other early modern human humeri(and differs from Neandertals) in having awide deltoid tuberosity with three crests, anexpanded (“nonstenotic”) diaphyseal crosssection, and a mediolaterally narrow olecra-non fossa bounded by relatively thickdistodorsal pillars (Churchill and Smith,2000). The most striking feature of the Vo-gelherd 3 humerus is its large, rugose, andprojecting deltoid tuberosity. The overallstoutness and rugose muscle scarring of thisspecimen were what led to initial claimsthat it may represent a Neandertal, andundoubtedly the enlarged, rugged deltoidtuberosity must have contributed to this im-pression. Ironically, it is the morphology ofthe deltoid tuberosity that most clearly dis-tinguishes the Vogelherd 3 humerus fromthat of Neandertals. In its absolute and rel-ative width, and in its crest configurations,the deltoid tuberosity of Vogelherd 3 is mostsimilar to that of early modern humansfrom later Aurignacian and Gravettian con-texts.

The Vogelherd humerus evidently derivedfrom a robust, muscular individual, yetthere is nothing that clearly indicates theretention of Neandertal features. In this re-spect the Vogelherd humerus is similar tothe postcranial material recovered from Au-rignacian contexts at Mladec Cave (CzechRepublic) (Smith et al., 1989b). Like Vo-gelherd 3, some aspects of the Mladec post-cranial material evince skeletal hypertro-phy (most notably, the size of the joint

surfaces: Wolpoff, 1992), but the overallmorphology is unequivocally more modernthan archaic. This includes nonstenoticlong-bone diaphyses with reduced corticalthickness, and the clear absence of typicalNeandertal features (such as narrow hu-meral deltoid tuberosities, medially ori-ented radial tuberosities, and absence offemoral pilasters) (Smith et al., 1989b). Atboth Mladec and Vogelherd, possibly Nean-dertal-reminiscent features are morereadily identifiable in the cranial than post-cranial material (see Smith et al., 1989b;Frayer, 1992; Wolpoff, 1992; Frayer et al.,1993).

Riparo Bombrini (Liguria, Italy)

Excavations in 1976 resulted in the recov-ery of an isolated human tooth—a decidu-ous left lateral mandibular incisor (Li2)—inlevel III of Riparo Bombrini. Situated atop alevel (IV) containing tools characteristic of alate denticulate Mousterian (Vicino, 1986),level III appears to represent the base of anearly Aurignacian sequence (extending up-wards through level I) characterized by Du-four bladelets, bone points, decorated bones,perforated shells, and lumps of red ochre(Vicino, 1986). Gioia (1990) puts the Aurig-nacian at Riparo Bombrini between 33.0–31.5 ky BP, in the dry/cold interval of WurmIIIa. However, a very similar early Aurigna-cian assemblage with Dufour bladelets atnearby Riparo Mochi dates between 34.5–33.0 ky BP (Gioia, 1990). If the two assem-blages are contemporaneous, the humanmilk tooth may date to the period of climaticinstability at the beginning of Wurm IIIa.

According to Formicola (1989), the toothis small in both mesiodistal and bucco-lingual dimensions compared to both Nean-dertals (n 5 2) and a combined sample ofMesolithic and late Upper Paleolithic spec-imens. He also notes that the tooth does notappear to have the Neandertal characteris-tic of a pointed incisal edge, and judgingfrom drawings of the specimen (Formicola,1989, p. 288), the preserved root does notexhibit mesiodistal broadening and bucco-lingual flattening as one would expect in aNeandertal. This incisor likely derives froma modern human.


It should be noted that new dating andstratigraphic reanalysis of possibly Aurig-nacian-aged human remains from Grimaldi(at Grotte des Enfants, Barma Grande, andBaousso da Torre) revealed these remains tomost likely derive from Gravettian or laterperiods (Mussi, 1986; Bisson et al., 1996).

La Ferrassie (Dordogne, France)

A single human left I1 from level E9 of LaFerrassie was discovered in the collectionsof the Musee National de Prehistoire desEyzies (Gambier et al., 1990). The artifactsfrom this level were considered to representan early (“ancient”) Aurignacian by De Son-neville Bordes (1960), and this makes theLa Ferrassie tooth the only specimen to de-rive from the basal Aurignacian in France.The level is also thought by Leroyer (1988),on palynological and sedimentologicalgrounds, to have been deposited during thecold episode of Wurm IIIa, just prior to theDenekamp temperate oscillation, ca. 34–32ky BP.

According to Gambier et al. (1990), thetooth is not shovel-shaped (although strongmarginal ridges can be seen in the accom-panying photograph), and morphologicallyand metrically it falls closest to samples offossil modern humans. They are careful tonote, however, that the tooth could also fitcomfortably within a sample of Neander-tals, and that a definitive classification ofthe specimen is not possible.

Vindija Cave (Hrvatsko Zagorje, Croatia)

During the course of excavations between1974–1986, Mirko Malez and collaboratorsrecovered more than 100 hominid fossilsfrom three stratigraphic levels at VindijaCave (Malez et al., 1980; Smith and Ahern,1994) (Fig. 9). Level D, the stratigraphicallyhighest level with hominid remains, hasproduced 45 anatomically modern speci-mens (Malez et al., 1980; Smith et al., 1985)in association with artifacts typical of thefinal Gravettian (Karavanic, 1995), andlikely deposited at the end of Wurm III(Malez and Rukavina, 1979). Level G3, thestratigraphically lowest level to producehominid remains, yielded 48 specimens inassociation with Mousterian artifacts(Malez et al., 1980; Smith et al., 1985). This

layer is provisionally dated to the end of thelower Wurm stadial (end Wurm II: Wolpoffet al., 1981; Karavanic and Smith, 1998).The human remains from G3 are diagnosti-cally Neandertal, but exhibit a mosaic ofNeandertal and modern human features,indicating either a parallel evolution ofmodern human morphological traits in lateNeandertals or significant flow of modernhuman genes into the Hrvatsko Zagorje Ne-andertal population (see Wolpoff et al.,1981; Smith, 1994 and references therein).

Between Vindija’s layers D and G3 lie theearliest Upper Paleolithic strata: G1, Fd/d,and Fd. G1 is a thin but distinctive leveldating toward the end of the Interplenigla-cial (Rukavina, 1983) that has producedonly a small sample of typologically diag-

Fig. 9. Generalized stratigraphic section of VindijaCave, redrawn with modifications from Karavanic(1995). The left-hand scale represents meters below thesurface. The right-hand column indicates radiometricdates (in uncalibrated radiocarbon years) for the G3, G1,Fd/d, Fd, and E levels (from bottom to top). Dates fromKaravanic (1995) and Smith et al. (1999).


nostic lithic pieces (15 out of 62 total lithicsrecovered). Six of these are typologically Up-per Paleolithic (including two end-scraperson flakes and one on an Aurignacian blade,a straight dihedral burin, a blade with twocontinuously retouched edges, and a leaf-shaped bifacial piece), while nine are moretypical of the Mousterian (five side-scrapersand four denticulates) (Karavanic, 1995;Karavanic and Smith, 1998). This layer hasalso produced a split-base bone point char-acteristic of the Aurignacian, and two mas-sive-base bone points, as well as fragmentsof others (Karavanic, 1995). Massive-basebone points are found in early Upper Paleo-lithic, and also probably late Mousterian(Montet-White, 1996), components through-out Central Europe. The interface betweenthe G and F layers and the lower part of theF layer have produced a larger lithic sam-ple, but one that is again characterized by acombination of Upper Paleolithic-type tools(including some type-fossils of the Aurigna-cian: Aurignacian blades, and keeled andnosed end-scrapers) and Middle Paleolithicelements (notched pieces, denticulates, andside-scrapers) (Karavanic, 1995; Miracle,1998). Radiocarbon dating of cave bearbones at the base of the F layer (level Fd/d)indicates an age of 26.7 6 0.9 ky BP, whilea sample of charcoal found near the inter-face of the two lowest strata in the F level(Fd/Fd/d) produced a radiocarbon date of27.0 6 0.6 ky BP (see Karavanic, 1995).

The typological attributes of the G1 andlower-F (Fd/d and Fd) assemblages, compris-ing as they do both Middle and Upper Pa-leolithic elements, are a matter of ongoingdebate. The occurrence of Aurignacian-typetools (including the fossile directeur of theAurignacian I—the split-base bone point;Bordes, 1968) certainly lends an Aurigna-cian flavor to the assemblages, and any di-agnosis that relied on index fossils wouldundoubtedly classify them as early Aurigna-cian (see Karavanic, 1995). However, Aurig-nacian type-fossils, including split-basebone points, occur with regularity in IUPassemblages (such as in the Szeletian atSzeleta Cave [level 4] and Dzerava Skala[levels 5–11]; Miracle, 1998), as well in as-semblages that are lithically undiagnostic(Miracle, 1998). Clearly the utility of these

artifacts as fossiles directeurs for the Aurig-nacian of East-Central Europe has not beenestablished (Karavanic and Smith, 1998).When frequencies of retouched lithics areexamined, the assemblage from VindijaG1–Fd does not compare favorably to thosefrom other European Aurignacian levels(Miracle, 1998). At present it is uncertainwhether the Vindija G1 assemblage repre-sents an early Aurignacian comparable tothat found elsewhere in Europe (althoughthis possibility appears increasingly doubt-ful), a Central European regional variant ofthe Aurignacian (“Olschewian,” as in Kara-vanic and Smith, 1998), a Szeletian compo-nent (Miracle, 1998), or a late Mousterianthat incorporates Aurignacian tools ob-tained through contact with (possibly) con-temporaneous Upper Paleolithic peoples liv-ing at nearby sites like Velika Pecina.

The F complex, specifically levels Fd andFd/d, is likely a separate time componentfrom level G1. This is supported by the dis-tinctly different color and lithology of thesediments, with the F complex comprisingsandy sediment with abundant stone rub-ble, while G1 is a distinctive reddish clay(Malez and Rukavina, 1979). Furthermore,the chronometric dates for G1 and the Fcomplex (see below) suggest a temporal sep-aration, and the lithics in the F complex aremore clearly and consistently Upper Paleo-lithic in character (Karavanic, 1995).

Six fragmentary human fossils have beenrecovered from the G1 layer at Vindija.These remains include the right ramus andposterior corpus (sans teeth) of a mandible(Fig. 10), a fragment of a left parietal, a leftzygomatic (Fig. 11), a left frontal fragment(preserving the medial portion of the su-praorbital torus; Fig. 12), and two isolatedright maxillary teeth (I1 and C1) (Wolpoff etal., 1981; Smith and Ahern, 1994). A num-ber of morphological features indicate thatthese remains can be confidently classifiedas Neandertal. The diagnostic Neandertalfeatures in the G1 fossils include a retromo-lar space, horizontal-oval mandibular fora-men and a medial pterygoid tubercle on themandibular fragment, a prominent Bre-schet’s sulcus on the parietal fragment,marked shoveling and large size of the max-illary incisor, a columnar frontal process


and multiple zygomaticofacial foramina onthe zygomatic, and a large frontal sinus re-stricted within a true supraorbital torus inthe frontal fragment (Wolpoff et al., 1981;Smith and Ahern, 1994).

Five human fossils were recovered fromthe base of the F complex: three isolatedpermanent teeth from level Fd (a right I2, aright I2, and a left C1), and two parietalfragments that articulate along the sagittalsuture, from Fd/Fd/d (Wolpoff et al., 1981;

Smith et al., 1985; Smith and Ahern, 1994).The maxillary incisor is markedly shoveled(the primitive condition for the genus Homo,which occurs in some modern humans aswell), and both incisors are similar to thoseof Neandertals in being buccolingually thickrelative to the crown mesiodistal diameter(see Table 6 in Wolpoff et al., 1981). How-ever, these dimensions lie within or justslightly above the ranges for early UpperPaleolithic specimens in Smith (1984, p.150–151). Furthermore, the most diagnos-tic specimen (the conjoining right and leftparietal fragments) exhibits a more modern,gabled contour to the cranial vault, andthere is an indication of lambdoidal flatten-ing comparable to that seen in Central Eu-ropean early modern human crania (Smithet al., 1985). Although the small size andfragmentary nature of this sample precludestaunch interpretations of the nature andaffinities of the F complex human sample,the most reasonable working hypothesis isthat these people were fundamentally mod-ern, with some Neandertal reminiscent fea-tures.

The significance of the association of Ne-andertal fossils with an Aurignacian-likeassemblage in Vindija G1 remains uncer-tain. Imprecise stratigraphic control duringexcavation, along with evidence of cryotur-bational and bioturbational disturbance of

Fig. 10. Vindija 207 mandible in occlusal (top) andmedial (bottom) views. Note the retromolar space be-tween the alveolus for M3 (B) and the anterior border ofthe ramus, the horizontal-oval mandibular foramen,and the medial position of the intersection between themandibular notch (incisura) and the condyle (A). Draw-ing by Kim Reed, reprinted from Karavanic and Smith(1998).

Fig. 11. Vindija 307 (right) and modern human (left)left zygoma in lateral view. Note the columnar frontalprocess in the Vindija specimen.

Fig. 12. Vindija 308 left frontal fragment, anteriorview.


some of the sediments, have plagued effortsto interpret the G1 artifact assemblage andfossil humans (see Karavanic, 1995; Mira-cle, 1998; Zilhao and d’Errico, 1999). Frostheave or sediment disturbance by denningcave bears could have mixed lithic and os-seous materials in G1 from lower Moust-erian (G3 and G2) and higher Aurignacian(F) horizons (Kozłowski, 1996; Montet-White, 1996; Karavanic and Smith, 1998;Zilhao and d’Errico, 1999). Inexact excava-tion records make establishing the prove-nience of some of the G1 artifacts difficult,and some of them do appear to derive fromthe cryoturbated portion of the sediments(Karavanic and Smith, 1998). Karavanicand Smith (1998) noted, however, that theG1 mandibular fragment and the split-basebone point derive from a portion of the cavein which sediments are undisturbed, andthat the Upper Paleolithic lithic and osse-ous tool types lack modifications character-istic of postdepositional movement (such asnibbling and edge rounding; see also Mira-cle, 1998 for further support for a reliableassociation of Neandertal remains with Up-per Paleolithic tools).

Attempts to radiometrically date the G1materials have clarified but not resolved theissue. Karavanic et al. (1998) attempted todirectly test the contemporanity of thehominids and artifacts in G1 through gam-ma-ray spectrometry of the mandibularfragment and the split-base bone point. Re-sults indicated ages of 51.0 6 8.0 and 46.0 67.0 ky BP (U-Th and U-Pa, respectively) forthe mandible, and 45.0 6 6.0 (U-Th) and30.0 6 5.0 ky BP (U-Pa) for the bone point(Karavanic et al., 1998). The determinationson the mandible are considerably older thana previous AMS date on a fragment of cavebear (Ursus spelaeus) bone of 33.0 6 0.4 kyBP (Karavanic, 1995), which may be seen assupport of the claim of stratigraphic mixingin G1, except that the inconsistent datesobtained on the bone point and the rela-tively high error ranges may also reflect theunreliability of gamma-ray dating. More re-cently, Smith et al. (1999) attempted to datetwo artifacts (including the split-base bonepoint) and two hominid fossils (the mandib-ular and parietal fragments) from G1 byAMS radiocarbon. Unfortunately, both bone

points lacked sufficient collagen and/or suf-fered from contamination, and failed to pro-duce reliable dates. The mandibular frag-ment produced an uncalibrated date of29.1 6 0.4 ky BP, while the parietal frag-ment was dated to 28.0 6 0.4 ky BP (Smithet al., 1999). While these two dates are notstatistically significantly different from oneanother, they are significantly younger thanthe date obtained on the cave bear bone, andindicate a minimum span of 3,000 years inthe formation of layer G1.

The young dates derived from the G1hominid material deserve discussion. In ad-dition to demonstrating a Neandertal pres-ence in Central Europe until well into theUpper Pleniglacial (several thousands ofyears after unequivocally modern humanshad appeared in Europe), the young agesalso make unlikely the possibility that theywere secondarily deposited in G1 from lowerlevels by cryo- or bioturbation. Unfortu-nately, the only date in existence for VindijaG3 is a questionable amino-acid racemiza-tion (Isoleucine) determination of 42.2 6 4.3ky BP (Smith et al., 1985). Dates for the Fdand Fd/d levels fall at the very end of theAurignacian range and overlap with therange for the younger Gravettian in thispart of Europe (Karavanic, 1995).

Mladec (Moravian karst, CzechRepublic)

The Mladec Caves have produced perhapsthe largest, most important, and best-stud-ied assemblage of early modern humanskeletal material associated with the Aurig-nacian. The morphological attributes andtaxonomic affinities of the cranial material,and to a lesser extent postcranial material,have been well described (Szombathy, 1925;Smith, 1982, 1984; Jelınek, 1983; Frayer,1986; see especially a new and detailedanalysis by Frayer et al., nd), and we pro-vide only a brief summary of that work here,choosing instead to focus on the archeologi-cal and geochronological context of the ma-terial.

The Mladec Caves (also known by theGerman name Lautsch) comprise a maincave system (variously known as Furst-Jo-hanns-Hohle or Bacova dıra) and a smallerside cave (known as the Quarry Cave or


simply the Side Cave) 50 m to the west ofthe main site (Smith, 1997). The main cavewas excavated under the direction of JosefSzombathy in 1881 and 1882, and later byJan Knies (in 1903), Jan Smycka (in 1912),and Johann Furst (in 1922), while theQuarry Cave was excavated in 1904 by JanKnies (Smith, 1997). These excavations pro-duced 101 identifiable human fossil speci-mens, including two largely complete youngadult female crania and two calottes withpossibly associated mandibles representingadult males (sadly, 59 of these specimenswere destroyed, along with the entire inven-tory of human remains from Predmostı andnumerous specimens from Dolnı Vestonice,by a fire in Mikulov Castle in the closingdays of World War II). There is widespreadagreement that these remains represent apopulation of early modern humans, al-though disagreement exists as to their de-gree of Neandertal affinity (see, e.g., Frayer,1986; Frayer et al., nd, vs. Brauer andBroeg, 1998). As with other early modernhuman crania from this region, the craniafrom Mladec (Figs. 13, 14) are characterizedby marked development of the brow ridges(in the male crania), long cranial vaults,lambdoidal flattening, occipital bunning,and robust nuchal areas (Smith, 1984). Sev-

eral of the specimens have been further ar-gued to exhibit traits considered by some tobe uniquely derived for Neandertals, includ-ing an elliptical suprainiac fossa, extensivelambdoidal flattening, and a short posterioroccipital face in Mladec 6, a Neandertal-likecranial vault form in lateral view coupledwith a marked occipitomastoid crest, smallmastoids, and midfacial prognathism inMladec 5, and a groove on the internal sur-face of the inferior nasal margin and a me-dial projection on the lateral internal wall ofthe piriform aperture in Mladec 8 (Frayer etal., nd).

Among early modern human remainsfrom Europe, the occipital morphology of theMladec males is the closest to the Neander-tal condition. Both Mladec 5 and 6 exhibitoccipital bunning that extends more later-ally than is typical for most early modernEuropeans. The fact that variation in the

Fig. 13. Mladec 1 female cranium in oblique view.

Fig. 14. Mladec 5 male cranium. a: Anterior. b:Lateral.


nature of occipital bunning is continuous inlate Pleistocene Europeans suggests thatthe factors underlying bunning are notqualitatively different in early modern Eu-ropeans compared to Neandertals, and thisunderscores a Neandertal contribution tothe early modern European gene pool. Onthe other hand, the bunning in Mladec andother early moderns occurs in cranial vaultsthat are quite different in overall shapefrom Neandertals (see also Lieberman et al.,2000). For example, although both Mladec 5and 6 have broad cranial bases, both speci-mens lack the characteristic oval (“enbombe”) shape characteristic of Neandertalsin rear view. Indeed, the parietal bosses inboth Mladec specimens are located high onthe sides of the cranial vault. This placesthe bunning in the context of a rather dif-ferent cranial shape in Mladec and otherearly modern Europeans compared to Nean-dertals.

If the morphology of the entire Mladecsample is considered, a number of impor-tant points emerge. The adult female cra-nia (Mladec 1 and 2) exhibit evidence ofbunning and upper midfacial prognathismbut otherwise have vault and facial formsthat are not typical of Neandertals. Theseinclude their lateral and posterior vaultcontours, the presence of canine fossaeand angled inferior zygomaticoalveolarmargins, mastoid morphology, and theanatomy of their supraorbital regions.Furthermore, the mandibles pictured bySzombathy (1925), which were destroyedat Mikulov, appear modern in form, espe-cially at the symphysis; and the postcra-nial remains, although not as extensivelystudied as the crania, are variable in sizeand robustness but are fundamentallymodern in anatomical form (Smith et al.,1989b). Thus, when the entire sample isconsidered, the Neandertal reminiscentmorphology at Mladec is primarily foundin what might be called “details” of anat-omy rather than in the fundamentalmorphological gestalt. Such Neandertalsimilarities are certainly more evidentin the male crania, as was recently de-tailed by Frayer et al. (nd), but even thesespecimens are clearly distinguishable

from Neandertals in overall morphologicalform.

In both the main and side caves, thehuman remains, artifacts, and faunal re-mains were collected from a reddish-brown clayey sediment that representsthe fan of a talus cone formed by infillingthrough a chimney. As with the site ofZlaty kun (below), there are no hominidoccupation levels in the Mladec Caves,and the remains in the caves were proba-bly deposited through chimneys in thecaves’ roofs, similar to the better-docu-mented situation at Zlaty kun (Svoboda,2000). However, a small sample of arti-facts have been recovered from the sedi-ments (including . 22 bone points and 24perforated animal teeth, but only a smallcollection of lithics; Frayer et al., nd). Inthe sediments of both caves, the bonepoints are flat with broad bases (the so-called Mladec-type point), an artifact thattends to occur in early Aurignacian assem-blages in Central Europe (Frayer et al.,nd), although it is not restricted to suchcomponents (Montet-White, 1996). Split-base bone points, another common occur-rence in the Moravian Aurignacian, areabsent in the Mladec deposits. Bone awls(some 15 fragments) showing characteris-tic Aurignacian patterns of boneworkingwere also recovered at Mladec (Frayer etal., nd). On the basis of the stratigraphy ofthe main cave, Szombathy (1925) thoughtthe sediments to have accumulated in ashort time. The homogeneity of the arti-fact assemblages at each site further sug-gests rapid accumulation, while the nearidentity of the artifacts and fauna fromthe main and side caves suggests that thetwo infillings were penecontemporaneous(Frayer et al., nd). Thus, according toFrayer et al. (nd), Mladec can reasonablybe considered a single component site (butsee Svoboda, 2000). The relatively abun-dant Mladec points and other bone toolssuggest an early or middle Aurignacianassociation for the Mladec early modern hu-mans, and the associated fauna indicate rel-atively temperate (perhaps Denekamp) con-ditions (Frayer et al., nd).


Zlaty kun (Bohemian karst, CzechRepublic)

Koneprus Cave, situated on Zlaty kunHill a half-kilometer from the village ofKoneprusy, was discovered during blastingoperations in a limestone quarry in 1950(Vlcek, 1996; Svoboda, 2000). The cave pro-duced a partial skeleton of a robust yet mor-phologically modern individual (an adult fe-male according to Vlcek, 1996),7 as well asartifacts of an Upper Paleolithic character,distributed within a large talus cone in themain chamber. The deposits in Koneprusyaccumulated as debris fell through a chim-ney in the cave during early Upper Pleni-glacial times (early Wurm IIIb) (Prosek etal., 1952; Vlcek, 1957; Svoboda, 2000). Someof the human material (the facial skeleton)exhibits gnawing from carnivores (Vlcek,1996). As there is no evidence that humansever occupied the cave, the associations be-tween the hominid fossils and cultural ma-terials remains tentative, although Vlcek(1996) notes that some of the artifacts werefound concentrated in a restricted area nearthe skeleton, and may represent the con-tents of a satchel in the person’s possessionat the time of death. Artifacts recoveredfrom Koneprusy have been described as rep-resenting an “Upper Paleolithic industry ofMousterian character” (Vlcek, 1967, p. 268),as possibly having affinities with the localSzeletian (Prosek et al., 1952), or as Aurig-nacian on the basis of the presence of afragmentary flat Mladec bone point (Frid-rich and Sklenar, 1976), which are charac-teristic of the early Aurignacian of South-Central Europe (Jelınek, 1978). Again, wehasten to point out that the use of type-fossils for classifying artifact assemblages isfraught with problems (see Miracle, 1998),and in any case the hominid remains andartifacts are not in clear association at Zlatykun. Efforts are currently underway (by P.Pettitt and E. Trinkaus) to derive directdates on this important specimen.

While the geological age and cultural as-sociations of the Zlaty kun hominid are notentirely certain, the morphological affinitiesof this specimen are quite clear. The skele-ton preserves most of the neurocranium(lacking most of the basicranium; Fig. 15),including the supraorbital tori, both zygo-matics, the right maxilla with C1–M2, themandible with right C1–M2 and left I2–M2,five fragmentary vertebrae, and three ribfragments (Smith, 1982). The Zlaty kun cra-nium is similar to other early modern Euro-peans (and different from Neandertals) inhaving well-developed occipital bunning, arobust supraorbital region divided into a su-perciliary arch and supraorbital trigone, an“en maison” vault contour in coronal profile,

7The posterior cranial vault and right zygomatic of Zlaty kunwere initially thought to represent one individual (Zlaty kun 1),while the frontal, mandible, and isolated postcranial elementswere thought to represent a second individual (Zlaty kun 2). Thecranial elements were later found to refit, and all of the materialis now considered to represent a single individual.

Fig. 15. Zlaty kun cranium. a: Anterior. b: Lateral.


zygoma without columnar frontal processes,and a maxilla with a narrow nasal apertureand a clear although weakly expressed ca-nine fossa. The specimen also preserves arobust mandible with a distinct mental em-inence and mental trigone, a moderately re-treating symphysis (symphyseal angle 581°; average of 6 early Upper Paleolithicspecimens 5 76.8° 6 6.5°; 11 Neandertals,98.5° 6 4.8°; Smith 1984, p. 161), and noretromolar space (Smith, 1982; Brauer andBroeg, 1998). The preserved teeth are small,even relative to early Upper Paleolithicspecimens (Smith, 1982). Although thespecimen does exhibit a weakly developedsuprainiac fossa on the occipital, the mod-ern morphology of Zlaty kun is indisputable.

Fossellone (Latium, Italy)

Excavations in 1953 by A.C. Blanc andothers resulted in the recovery of humanmaterial from both Mousterian and earlyUpper Paleolithic levels at Fossellone, oneof the many caves of Monte Circeo (Mallegniand Segre-Naldini, 1992; Mallegni, 1992). Afragment of the symphyseal region of amandible, and three isolated teeth (left P4,M1, and M2), likely all deriving from a singleindividual (Fossellone 3) about 9 years ofage at death, were found in close associationat the top of an undated level containinghyena teeth and coprolites and a small num-ber of Mousterian tools. The remains wereattributed by Mallegni (1992) to Neander-tals on the basis of 1) similarities in themorphology of the lingual surface of thesymphysis (namely, the presence of a me-dian sagittal crest bounded by shallow de-pressions) to that of Guattari 2, 2) a sulcal(closed sulcus) and cusp (with incipient for-mation of an entoconid along with the pro-toconid and metaconid) pattern on the P4similar to that seen in Ehringsdorf and LeMoustier, 3) the presence of archaic featuresincluding cingula, fovea anterior, tubercle,and sulcal patterns, and 4) M2 . M1 inmesiodistal and buccolingual diameters andcrown area. Mallegni (1992) notes, however,that the Fossellone 3 teeth are smaller indiameters and areas (and often more thanone standard deviation below the mean)than those of other Neandertals, and in factthe sizes of these teeth generally fall within

one standard deviation of the means for Up-per Paleolithic modern human teeth pre-sented by Mallegni and Segre-Naldini(1992).

Of more relevance to this review are twofossils that were recovered from Upper Pa-leolithic contexts. Fossellone 1 is a rightmaxillary fragment with the M1 and M2 inplace. This specimen was recovered from anAurignacian level (level E) that representeda late stage of the early Aurignacian to LaPlace (1964, 1966) or Aurignacian 1 to Zam-petti and Mussi (1988). Based on faunal at-tributes, this layer accumulated during thecold-dry interval of Wurm IIIa about 33.0–31.5 ky BP (Zampetti and Mussi, 1988;Gioia, 1990).

Fossellone 1 preserves the entire P4 anddistal P3 alveolar sockets, with a concavityof the buccal wall between them indicating adeep canine fossa (Mallegni and Segre-Nal-dini, 1992). The M2 exhibits a small cingu-lum, and in size and shape the molars aresimilar to those of other Italian Upper Pa-leolithic modern humans (Mallegni andSegre-Naldini, 1992).

Fossellone 2 is a partial left scapula, pre-serving the glenoid fossa, most of the spineand acromial process, most of the coracoidprocess, and the proximal third of the axil-lary border. This specimen was recoveredfrom an infilled erosion channel containingboth Mousterian and Aurignacian tools, andthus the chronostratigraphic position of thefossil is uncertain. The overall morphologyof this specimen is modern: the glenoid fossais relatively wide, and the axillary border isbisulcate and dorsoventrally thin (Mallegniand Segre-Naldini, 1992). Since no diagnos-tic Gravettian tools were found in the chan-nel fill, the erosion is thought to have oc-curred during a marine transgression at theend of the early Aurignacian occupation ofthe site, perhaps reflecting temperate con-ditions of the Denekamp interstadial orearly Wurm IIIb. Accordingly the scapulamost likely represents a modern human as-sociated with the Aurignacian at the site.

Kelsterbach (Frankfurt, Germany)

From a gravel pit near Frankfurt, Kels-terbach is a gracile calvarium (lacking muchof its base; Fig. 16) of a modern human


female. Although the stratigraphic positionof the specimen was noted in graphs anddrawings by workmen (Protsch von Zieten,1988), the age of the deposits has not beenfirmly established. The detailed recoverynotes (from 1952), along with adherent ma-trix in the left auditory meatus and exami-nation of the still-undisturbed stratigraphicsection, allowed Protsch von Zieten (1988)and Protsch and Semmel (1978) to place thefind in the lower part of a gravel bed at the“Obere Niederterrasse” or t(6) terrace of theriver Main (Protsch von Zieten, 1988). Ami-no-acid racemization and radiocarbon dat-ing of bone collagen have both been applieddirectly to the calvarium, with resulting ageestimates of 32–31 ky BP (AAR, 32 ky BP;14C, 31.2 6 1.6 ky BP; Protsch and Semmel,1978), and if these ages are correct thiswould place the specimen in the period of

the later Aurignacian in this area. The spec-imen is remarkable, given its purported age,for its gracility and lack of archaic features:the brow ridges are weakly developed, thefrontal angle is high, and there is only aslight indication of occipital bunning(Brauer, 1980; Smith, 1984). As there is noarcheological context for the specimen,Kelsterbach would seem another excellentcandidate for AMS dating.

Kent’s Cavern (Devonshire, England)

A right maxillary fragment (Kent’s Cav-ern 4), preserving the canine, fourth premo-lar, and first molar (all heavily worn), wasdiscovered stratigraphically below Aurigna-cian artifacts during excavations by A.H.Ogilvie in 1927 in trench C in the cave’svestibule (Oakley et al., 1971; Hedges et al.,1989). The “Aurignacoid” artifacts (Garrod,1926) consist of several blades struck fromopposed-platform cores (Hedges et al.,1989), which appear to represent the UpperPaleolithic but which are insufficient forcultural diagnosis. The hominid specimenwas attributed to Homo sapiens sapiens byKeith (1927), and has since been direct-dated by AMS radiocarbon to 30.9 6 0.9 kyBP (Hedges et al., 1989). Layer A2, fromwhich both the hominid maxilla and theearly Upper Paleolithic tools derive (Keith,1927; Campbell and Sampson, 1971), wassubsequently recognized as a debris flow(see Aldhouse-Green and Pettitt, 1998),which complicates the chronostratigraphicpicture somewhat.

A presumed Aurignacian-associated hu-man skeleton was also recovered from theBritish site of Goat’s Hole (Paviland). Theochre-stained partial skeleton of an adultmale was recovered from this site in 1823 bythe Rev. William Buckland (Aldhouse-Green and Pettitt, 1998). The presence ofbusked burins and nosed and carinatedscrapers in the deposits at Paviland, andthe apparent association of ivory braceletsand rods with the ochre-stained interment,led to the suggestion that the skeleton wasof Aurignacian age, likely corresponding tothe Aurignacian II of western continentalEurope (Sollas, 1913; Jacobi, 1980; see alsoAldhouse-Green and Pettitt, 1998). Subse-quent direct dating of the Paviland 1 skele-

Fig. 16. Kelsterbach cranium. a: Anterior. b:Lateral.


ton (Hedges et al., 1989) resulted in a dateof 26.350 6 0.55 ky BP, which is late rela-tive to the Aurignacian on the continent(Mellars and Bricker, 1986), suggesting ei-ther an extended duration of the Aurigna-cian on the British Isles or an incorrect cul-tural diagnosis of the associated artifacts.

Cioclovina (Transalvanian Alps,Romania)

The calvarium of an anatomically modernhuman (Fig. 17), morphologically similar toother Central European early Upper Paleo-lithic modern humans, was recovered withthree Aurignacian artifacts at Cioclovina(Necrasov and Cristescu, 1965). The speci-men likely represents a male (Smith, 1984;contra Necrasov and Cristescu, 1965), and

is similar to other early modern crania in itsexpression of occipital bunning and in browridge morphology. Without a larger culturalcomponent or absolute dates, the Cioclovinahominid contributes little to our under-standing of the nature of the Neandertal/modern human transition, but may well bean early representative of the latter group.

Podbaba (Prague, Czech Republic)

A partial calvarium with large, projectingsupraorbital tori and a low frontal squamawas discovered at Podbaba in 1883. Thespecimen is generally thought to have de-rived from the Aurignacian at the site, butthe exact stratigraphic provenience is un-known (Obermaier, 1905; Matiegka, 1924).Originally thought to represent a Neander-tal based on its large brow ridges and lowfrontal (Fric, 1885), Podbaba has since beenshown to be morphologically similar to othercrania of early modern humans from South-Central Europe (Matiegka, 1924; Vlcek,1956; Smith, 1984). Unfortunately, the Pod-baba cranium was destroyed during castingin 1921. Given its uncertain context and theimpossibility of directly dating the speci-men, the relevance of this specimen tobiocultural dynamics of the Middle/UpperPaleolithic transition will remain uncertain.

Camargo (Santander, Spain)

During excavations in 1908 at CamargoCave, a near neighbor of El Castillo (bothsites are in the valley of the river Pas), Fa-ther Lorenzo Sierra discovered a fragmen-tary human calotte (Fig. 18) from a levelbearing Upper Paleolithic artifacts. Thelithic assemblage was attributed to the Au-rignacian by Obermaier (1924),8 but no for-mal typological analysis was conducted onthe Camargo artifacts before their destruc-tion during the Spanish Civil War (Gar-ralda, 1997). The cranial vault, also de-stroyed during the war, is thought torepresent a female. The illustration of thisspecimen provided by Obermaier (1924,

8Prior to the reorganization by Peyrony (1933) of the earlyUpper Paleolithic, Gravettian assemblages were often includedas part of the Aurignacian (Harrold, 2000); thus, the possibilityexists that the Camargo assemblage was Gravettian in charac-ter.

Fig. 17. Cioclovina (cast). a: Anterior. b: Lateral.


Plate 17) shows a gracile, dolichocephalicvault with a high frontal angle and an in-cipient occipital bun. The region of the rightsupraorbital torus appears to be partiallypreserved and looks to be gracile and non-projecting, consistent with the female sexattribution. The specimen was consideredanatomically modern by Saller (1926), andthere is nothing in the published illustra-tions to argue against this attribution. How-ever, given the lack of a formal typologicalanalysis of the associated artifacts, incom-plete analysis of the human material, andlack of chronostratigraphic context for thespecimen, the Camargo cranial vault re-mains of little utility in determining theaffinities of the makers of the earliest Au-rignacian in Spain.

THE MAKERS OF THE EARLIESTAURIGNACIAN AND THE TIMING OF

THE APPEARANCE OF MODERNHUMANS IN EUROPE

The fossil and archaeological records ofthe early Aurignacian are summarized inTable 6, along with our subjective assess-ments of the degree of confidence (low, mod-erate-to-good, or high) in the taxonomic andtypological classifications, cultural associa-tions, and chronological placements, basedon the available evidence. Arranging thedata in this manner provides a foundationfor addressing the question of the timing ofthe appearance of modern humans and theUpper Paleolithic in Europe (Fig. 19).

The oft-cited conventional radiocarbondate of .43 ky BP for the Bachokirian inlayer 11 of Bacho Kiro Cave has not beensupported by further dating. More rigorousAMS dating of this layer suggests that itwas deposited over a 5,000-year interval be-tween ca. 39–34 ky BP. The appearance ofthe Bachokirian is roughly coincident, then,with the appearance of the early Aurigna-cian elsewhere, as well as with the Szeletian(sensu lato) and Chatelperronian (Table 1).The Bachokirian is probably best seen aspart of a larger pan-European emergence ofthe Upper Paleolithic during Hengelo tem-perate conditions, rather than as an espe-cially early forebear of the Upper Paleo-lithic. The single human fossil from BachoKiro’s layer 11 is frustratingly undiagnosticwith respect to taxonomy. Given the earlydate of the Bachokirian, the taxonomic af-finities of this specimen are of extreme in-terest. At least one diagnostic feature, theanatomy of the dm1 pulp cavity, suggestsmodern human affinities; but this singlefeature is not sufficient for a staunch claimthat this is a modern human. Such certaintyis unfortunately beyond our grasp atpresent.

The only other fossil-bearing sites thatmay date to the Hengelo (El Castillo andHahnofersand) present similar problems ofinterpretation. The El Castillo materialfalls within the ranges of variation of bothNeandertals and early modern Europeans,making taxonomic diagnosis unrealizable(especially given the impossibility of further

Fig. 18. Camargo cranium in lateral (top) and supe-rior (bottom) views. Drawn from photographs in Ober-maier (1924) by Dania Ermentrout.


TABLE 6. Summary of possible early Aurignacian fossil sites

Assemblage designation1Taxonomic designation

of human remains1Confidence in

archaeological association2Absolute date or

geologic age1

El Castillo/level 18 Basal Aurignacian Uncertain High ca. 40–38.5Bacho Kiro/layer 11 Bachokirian/proto-

AurignacianModern human High ca. 39–34 ky

BPBacho Kiro/7/6b and 7/6a Aurignacian Modern human High ca. 33–29 ky BPHahnofersand None Modern human NA ca. 36 ky BPVogelherd/level V Early Aurignacian Modern human High >32 ky BPRiparo Bombrini Early Aurignacian Modern human High 34.5–31.5La Ferrassie/Level E9 Aurignacian 0 Modern human High 34–32Vindija/Level G1 IUP with Aurignacian elements Neandertal Moderate-low ca. 33–28 ky

BPVindija/Level F Aurignacian Uncertain High ca. 27 ky BPMladec Early/middle Aurignacian Modern human Moderate-high Wurm IIIaZlaty kun Aurignacian Modern human Moderate-high Wurm IIIa or

IIIbFossellone Early Aurignacian Modern human High 33–31.5Kelsterbach None Modern human NA ca. 32–31 ky BPKent’s Cavern “Aurignacoid” British early

Upper PaleolithicModern human Moderate-low ca. 31 ky BP

Cioclovina Aurignacian Modern human Moderate NonePodbaba Aurignacian Modern human Poor NoneCamargo Aurignacian Modern human Moderate None1 Our assessment of the degree of confidence associated with the assemblage designation, taxonomic designation, and absolute date of each site is indicated as follows: bold text, high confidence;plain text, moderate-to-good confidence; italicized text, low confidence. All dates in ky BP.2 Confidence in association of human remains with early Aurignacian component of site.

Fig. 19. Temporal distribution of archeological levels with human remains and Aurignacian assem-blages or Aurignacian-like tools. Ages are in (uncalibrated) radiocarbon years; the Hengelo andDenekamp interstadials are indicated by shading on the Wurm chronology. Dashed lines indicate agreater than usual uncertainty in dating. The undated sites of Cioclovina, Podbaba, and Camargo (seetext) are not included.


analysis of the specimens). The Hahnofer-sand frontal has a basically modern mor-phology but lacks an archaeological context.Given the lack of context and the recentdemonstration that the morphologicallysimilar specimen from Velika Pecina doesnot date to .34 ky BP (Smith et al., 1999),unequivocal acceptance of the 36 ky BP agefor Hahnofersand will require redating ofthe specimen with the more precise tech-niques now available (e.g., AMS radiocar-bon dating).

The unquestionably modern human fos-sils from the lower portion of Vogelherdlevel V may rate as the earliest well-prove-nienced and confidently classified modernhuman fossils in Europe. At present, weknow that the Vogelherd 1 and 3 specimensare likely older than 32 ky BP, but if we arecorrect that paleoenvironmental indicatorsdesignate probable contemporanity of Vo-gelherd layer V and Geißenklosterle levelIII, the Vogelherd skeletal material maydate to the end of the Hengelo interstadialor to early Wurm IIIa times. Also in thisend-Hengelo time range may go the isolatedteeth from Riparo Bombrini and La Fer-rassie Level E9. The teeth are of question-able modernity (in both cases they appear tobe more similar to early moderns than toNeandertals, but secure attributions cannotbe made on the basis of single teeth) andquestionable age, but their associationswith early Aurignacian assemblages, alongwith the associated paleoclimatic indica-tors, suggest an early Wurm IIIa date.

In sediments dating to the end of theWurm IIIa stadial and beginning of theDenekamp interstadial, diagnostically mod-ern human fossils become more frequent,with many of them (e.g., Mladec, Zlaty kun,Kent’s Cavern) in association with Aurigna-cian type-fossils. To this group can likely beadded the undated specimens from Fossel-lone. Finally, the culturally unaffiliatedmodern human cranium from Kelsterbachmay derive from this period if its provisionaldate of 32–31 ky BP can be substantiated.

Based on the current state of the fossiland archeological records, we can drawsome tentative conclusions. First, we canconfidently say that modern humans wereestablished in Europe by mid-late Wurm

IIIa times (by ca. 32 ky BP). There is astrong suggestion that modern humans hadsettled in Europe by end-Hengelo times (ca.36 ky BP), but this must remain only asuggestion until further dating of the spec-imens from Hahnofersand and Vogelherd isaccomplished. Claims for an appearance ofmodern humans in Europe before 36 ky BP,including claims based on dated archeolog-ical complexes lacking fossil human re-mains, cannot be substantiated with cur-rently available fossil evidence.

Second, the hypothesis that modern hu-mans are uniquely associated with the Au-rignacian cannot yet be refuted. As Table 6demonstrates, all reasonably diagnostic hu-man remains confidently associated withthe Aurignacian represent modern humans.The single current best argument againstthis hypothesis is the case of the Vindija G1Neandertals, but even if the G1 artifact as-semblage is not artificially mixed, its typo-logical attributes are uncertain (Karavanicand Smith, 1998). Thus, Vindija may standin parallel with St. Cesaire and Arcy-sur-Cure as a testament to the Upper Paleo-lithic cultural abilities of Neandertals, butnot necessarily to their production of theAurignacian sensu stricto. However, it is im-portant to bear in mind that none of theconfidently Aurignacian associated modernhuman skeletal remains have been shownto predate ca. 32–33 ky BP (although thosefrom Vogelherd V may). Given the uncer-tainty surrounding the origin of the Aurig-nacian, we should not be confident that theearliest Aurignacian must have been madeby modern humans until we have the appro-priate fossil associations.

Third, whatever the ecological and cul-tural dynamics promoting the emergence ofthe Upper Paleolithic in Europe, Neander-tals were a part of those dynamics. On thewestern (Iberian) and eastern (Russian)fringes of their range, the last Neandertalsseem to have held to a Mousterian way oflife to the bitter end. But between theseextremes Neandertals appear to have beenfull participants in the evolving Upper Pa-leolithic. Cultural materials from Neander-tal-bearing levels at St. Cesaire, Arcy-sur-Cure, and Vindija, which include bone toolsand items of personal adornment, suggest


that Neandertals had cultural capacities ona par with those of early modern humans.Similarities in the behavioral repertoires ofChatelperronian Neandertals and Aurigna-cian and Gravettian modern humans arefurther suggested by biomechanical analy-ses of the St. Cesaire long-bone diaphyses(Trinkaus et al., 1998, 1999). Further evi-dence that Neandertal behavioral capacitieswere comparable to those of modern hu-mans derives from recent suggestions thatNeandertals were effective and efficienthunters, apparently taking all of their di-etary protein from meat (Bocherens et al.,1999; Richards et al., 2000). While the mostappropriate typological designation of theVindija G1 assemblage remains uncertain,the very difficulty we face in making thisdesignation underscores the similarity inthe cultural behavior of the Vindija G1 Ne-andertals and Aurignacian modern hu-mans.

Fourth, by conservative estimates, Nean-dertals and modern humans coexisted inEurope for 2,000–4,000 years. If modernhumans did arrive in Europe during Hen-gelo times, this coexistence may have beenmore on the order of 8,000–10,000 years.The geographic location of fossil sites pro-vides an idea of the minimum proximity ofthese groups of humans. Thus during lateWurm IIIa times, for example, we can saythat the Neandertals of the HrvatskoZagorje of Croatia (at Vindija) likely hadmodern human neighbors in Moravia (Mla-dec) and Bohemia (Zlaty kun), about 400 kmto the north, and within 600 km to thenorthwest (in southern Germany, at Vo-gelherd), and possibly west (in Liguria, ifthe Riparo Bombrini incisor represents amodern human). The isolated incisor fromLa Ferrassie is roughly contemporaneouswith the Chatelperronian Neandertal re-mains from St. Cesaire and Arcy-sur-Cure.If this specimen represents a modern hu-man (unfortunately this is a call that cannotbe made with any confidence on present ev-idence), it would signal very close proximitybetween these groups. It is tempting to usethe close geographic association of IUP andAurignacian sites (and the possible inter-stratification of these assemblages withinsome sites) to argue for close proximity be-

tween Neandertal and modern human pop-ulations in Interpleniglacial times. To do so,however, requires the assumption that spe-cific lithic cultures equate with specifictypes of hominids, the circularity of which(in the context of this review) has not es-caped us. Given that the evidence reviewedabove does not allow for the rejection of aNeandertal-IUP and modern human-Aurig-nacian association, all we can say at presentis that these groups had a considerable pe-riod of co-occupation of Europe, and that theassociated archeological record suggestsclose contact between them.

It is difficult to imagine a co-occupation ofEurope of 2,000–4,000 years or longer with-out a substantial amount of cultural ex-change between human groups, and withouta substantial degree of adaptive parity(which is not to say adaptive equality—dif-ferences in adaptive strategies may wellhave existed between groups; Churchill etal., 1996). Despite paleogenetic evidence tothe contrary (Krings et al., 1997, 1999;Ovchinnikov et al., 2000; but see Nordburg,1998), consideration of certain morphologi-cal details evident in the last Neandertalsand the earliest modern humans in Europesuggests that this coexistence also entaileda significant degree of genetic exchange aswell (Trinkaus and Smith, 1985; Smith etal., 1989a; Smith and Trinkaus, 1991; Du-arte et al., 1999). It is important to note thatthe evidence for morphological, and presum-ably genetic, continuity between EuropeanNeandertals and early modern humans isfound in anatomical details and not funda-mental aspects of morphological form. LateNeandertals from Vindija (G1 and G3),Kulna, Barakaevskaıa, Grotte du Renne, St.Cesaire, and Zafarraya are unequivocallyidentifiable as Neandertals, yet many ofthese specimens have morphological detailsthat approach the modern human condition(Smith 1984, 1994; Smith and Trinkaus,1991; Wolpoff, 1999). Although there is al-ways the possibility that these details re-flect parallelisms, the most logical explana-tion in our opinion is low-level gene flowfrom modern populations that were begin-ning to colonize Europe during Hengelotimes (or from populations on the peripher-ies of Europe prior to the period of active


colonization). The overall morphologicalform of early modern Europeans such asthose from Mladec, Vogelherd, Zlaty kun, orHahnofersand is clearly distinctive fromthat of Neandertals. Furthermore, the basicsimilarity of early modern European mor-phology to that of the Skhul/Qafzeh samplefrom the Near East is evident (Vandermeer-sch, 1981; Tillier, 1999; Holliday, 2000).Given that fossils from Skhul/Qafzeh ante-date any early modern European, it seemslogical to view the former as having had asignificant role in the origin of modern hu-mans in Europe. However, there are strongindications that Neandertals were not shutout of this process. Early modern Europeanshave a series of anatomical details (e.g.,similar frequencies and patterns of mandib-ular foramen form, retromolar spaces, su-prainiac fossae, and occipital bunning) thatappear to derive from Neandertal contribu-tion to their gene pools (Smith, 1984;Frayer, 1992; Frayer et al., 1993; Wolpoff,1999).

The phylogenetic significance of Neander-tal-reminiscent features in early modernEuropeans remains unclear (e.g., Caspari,1991; Lieberman et al., 2000 on occipitalbunning). Some of these features may beshared primitive traits, or the product ofparallel evolution, and thus of little use inassessing the phylogenetic relationships be-tween these groups. However, we see in themorphology of these groups a situationanalogous to that of the emergence of IUPcultures in Europe. In a query that appliesequally to all the various IUP industriesthat proliferated in the early Interplenigla-cial, Harrold (2000, p. 70) asks, “If the Chat-elperronian developed autochthonously,why did it do so after a long period of Mous-terian adaptive stability and just before thelocal appearance of the Aurignacian?” Like-wise we can ask of the fossil evidence: 1)Why, after at least 200,000 years of morpho-logical divergence from the presumed mod-ern human lineage in Africa, did Neander-tals begin to converge on details ofmorphology with modern humans, at pre-cisely the time that modern humans seem tohave been expanding their range into Eu-rope? This “convergence” is evident in spec-imens from western France (Vandermeer-

sch, 1984) to the northern Caucasus(Lioubine, 1998). 2) Why, in the absence ofgenetic exchange, would early modern hu-mans, only upon entering Europe, indepen-dently develop Neandertal-reminiscent fea-tures—features that are largely absent intheir presumed ancestors from the NearEast?

In our opinion, the best explanation forthis situation is the “assimilation” of Nean-dertals into the modern populations thatlikely entered Europe from the east andsouth, introducing some Neandertal ele-ments into the early modern European genepool. The opportunity for such biological in-teraction is certainly enhanced by the grow-ing evidence for a substantial period of over-lap between these populations in Europe.Assimilation rather than in situ evolution issupported by the fact that, even at the endof this overlap, Neandertal and early mod-ern populations remained fundamentallydistinct from each other, only evincing indi-cations of biological integration in a rela-tively few specific morphological features.

Virtually all current perspectives on mod-ern human origins are willing to accept thepossibility of genetic exchange between Ne-andertals and early modern Europeans.Leading proponents of a “Recent African Or-igin” for all modern humans do not rule out,at least theoretically, some Neandertal con-tribution to early modern gene pools; but itis clear from their writings that such contri-butions were at best minimal, and perhapsbordered on the insignificant (Brauer, 1992;Stringer and Brauer, 1994; Brauer andBroeg, 1998). The “Multiregional EvolutionModel,” on the other hand, is fundamentallybased on the existence of significant, butvariable, patterns of genetic exchange be-tween regional lineages of humans through-out the Middle and Late Pleistocene (Frayeret al., 1993; Wolpoff, 1999). In a recent dis-cussion of the emergence of modern mor-phology in Europe, Frayer et al. (nd) con-cluded that the majority contribution to thismorphology did not derive from Neander-tals and thus entered Europe via gene flow.However, they did not view this gene flow asthe catalyst for the emergence of early mod-ern European morphology, nor did they ac-cept the idea that modern morphology has a


specific region of origin (Frayer et al., 1993,p. 42). In contrast, our view of this processconsiders gene flow from outside Europe tohave been the catalyst for the appearance ofmodern Europeans. We interpret the “tran-sitional” morphology of some late EuropeanNeandertals as likely related to low levels ofgene flow by demic diffusion into Europejust prior to ca. 35 ky BP. However, theemergence of definitively modern humans(around or just after this date) is most logi-cally due to a systematic increase in geneflow (cf. Trinkaus and Smith, 1985) in theform of population movements into Europeat this time. At the same time, our interpre-tation of the morphology of early modernEuropeans convinces us that Neandertalsmust have been far from insignificant in themorphological and genetic makeup of thispopulation. In other words, we see evidenceof a significant “assimilation” of Neander-tals into a fundamentally modern, immi-grant population.

The origin of modern human morphology,based on the current state of the fossil andgeochronological records, appears to havebeen centered in Africa in the late MiddlePleistocene (Beaumont et al., 1978; Singerand Wymer, 1982; Rightmire and Deacon,1991). There are also indications from thearcheological record that some aspects ofmodern human behavior (or perhaps moreprecisely, Upper Paleolithic behavior: nota-bly symbolic expression, bone working, andthe production of sophisticated compositeextractive technology) also arose in sub-Sa-haran Africa during Middle-to-Late Pleisto-cene times (Wendt, 1976; Brooks et al.,1995; Yellen et al., 1995; Deacon and Wurz,1996; Henshilwood and Sealy, 1997). Whilethere is a natural tendency to view theemergence of modern form and modern be-havior as causally related (see Churchill,1997), hominid-cultural associations out-side of Africa (McCown and Keith, 1939;Leveque and Vandermeersch, 1980; Hublinet al., 1996) urge caution in uncritically ac-cepting this notion. Much of the classic re-search in European prehistory was donewithin a paradigm that saw modern hu-mans as having had a greater capacity forcultural behavior, and that envisioned aclear distinction between the Middle Paleo-

lithic of the Neandertals and the Upper Pa-leolithic of modern people (amply reflectedin the quote by Le Gros Clark at the begin-ning of this review), and much recent efforthas been devoted to untangling the hardevidence from the preconceptions that sur-round it.

In this vein we are careful to point outthat although movement of populationsfrom the south probably had a catalytic ef-fect on the biological emergence of modernEuropeans, there is certainly no evidencethat this occurred in concert with the originof the Aurignacian. The oft-repeated impli-cation that the appearance and spread ofboth in Europe must have occurred togetheris not demonstrated by the available evi-dence. In fact, the origin of the Aurignacianremains somewhat of a mystery and maywell be an internal event (perhaps even aseries of independent events) in Europe.

The emergence and fluorescence of theUpper Paleolithic across Europe were nodoubt a response to the climatic volatilityand cultural dynamics of the Interplenigla-cial. Certainly an important part of thatcultural dynamic was the interaction be-tween morphologically different, and possi-bly behaviorally different, groups of peoplewho shared the region. Scholarly under-standing of the nature of that interactionhas improved substanially in recent years.Although we certainly do not yet have allthe answers to the questions posed at thebeginning of this review, there are somevery clear patterns emerging (as presentedin the preceding paragraphs). Furthermore,it is important to bear in mind that we arebeginning to delve into very detailed aspectsof population and cultural dynamics in LatePleistocene Europe. This level of analyticaldetail cannot be applied to earlier periods inthe human evolutionary record, nor even toother geographic regions in the Late Pleis-tocene. Thus, it is not surprizing that manyof these questions continue to be the focus ofhealthy scientific debate. Despite this, it isbecoming increasingly apparent that simpleconceptions of cultural superiority and sup-plantation (the “colonial metaphor:” Graves,1991) are inadequate for understanding theevents of the Middle-to-Upper Paleolithic


transition and the demise of the Neander-tals on the European continent.

ACKNOWLEDGMENTS

This paper has benefited from discussionswith, and comments from, Michael Black,Nick Conard, Vincenzo Formicola, DavidFrayer, Laura Gruss, Dan Lieberman, ChrisRuff, Lawrence Straus, Erik Trinkaus, andtwo anonymous reviewers. For help withthe figures, we thank Dania Ermentrout,Michael Black, and Laura Gruss.

LITERATURE CITED

Albrecht G, Hahn J, Torke WG. 1972. Merkmalsanalysevon Geschoßspitzen des mittleren Jungpleistzans inMittel- und Osteuropa. Tubingen: Archaeologica Ve-natoria.

Aldhouse-Green S, Pettitt P. 1998. Paviland Cave: con-textualizing the “Red Lady”. Antiquity 72:756–772.

Alexeeva TI. 1997. More on Starosele. Curr Anthropol38:647–650.

Allsworth-Jones P. 1990a. The Szeletian and the strati-graphic succession in Central Europe and adjacentareas: main trends, recent results, and problems forresolution. In: Mellars P, editor. The emergence ofmodern humans: an archaeological perspective.Ithica, NY: Cornell University Press. p 160–242.

Allsworth-Jones P. 1990b. Les industries a pointe folia-cees d’Europe centrale: questions de definitions etrelations avec les autres techno-complexes. In: FarizyC, editor. Paleolithique moyen recent et Paleolithiquesuperieur ancien en Europe. Nemours: Musee de Pre-histoire d’Ile-de-France. Memoires 3. p 79–95.

Almagro M, Fryxell R, Irwin HT, Serna M. 1970.Avances a la investigacion arqueologica, geocrono-logica y ecologica de la Cueva de La Carihuela (Pinar,Granada). Trab Prehist 27:45–60.

Anikovich M. 1992. Early Upper Paleolithic industriesof eastern Europe. J World Prehist 6:205–245.

Antunes M. 1990. O homen da gruta da figueira brava.Mem Acad Cie Lisboa 31:487–536.

Antunes M, Peixoto J, Cardoso J, Monge A. 1989. Pa-leolitico medio e superior em Portugal: datas 14C,estado actual dos conhecimentos, sintese e discussao.Cie Terra 10:127–138.

ApSimon AP. 1986. Picken’s Hole, Compton Bishop,Somerset; early Devensian bear and wolf den, andMiddle Devensian hyaena den and Palaeolithic site.In: Collcutt SN, editor. The Palaeolithic of Britan andits nearest neighbours: recent trends. Sheffield: Uni-versity of Sheffield. p 55–56.

Bard E, Hamelin B, Fairbanks RG, Zindler A. 1990.Calibration of the C-14 timescale over the past 30,000years using mass spectrometric U-Th ages from Bar-bados corals. Nature 345:405–410.

Bar-Yosef O, Kuhn SL. 1999. The big deal about blades:laminar technologies and human evolution. Am An-thropol 101:322–338.

Bastin B, Cordy J-M, Gewelt M, Otte M. 1986. Fluctu-ations climatiques enregistrees depuis 125 000 ansdans les couches de remplissage de la grotte Scladina(Province de Namur, Belgique). Bull Assoc FranceEtude Quaternaire 25–26:168–177.

Beaumont PB, de Villiers H, Vogel JC. 1978. Modernman in sub-Saharan Africa prior to 49 000 years BP:

a review and evaluation with particular reference toBorder Cave. S Afr J Sci 74:409–419.

Bermudez de Castro JM, Rosas A, Nicolas ME. 1999.Dental remains from Atapuerca-TD6 (Gran Dolinasite, Burgos, Spain). J Hum Evol 37:523–566.

Bietti A. 1997. The transition to anatomically modernhumans: the case of peninsular Italy. In: Clark GA,Willermet CM, editors. Conceptual issues in modernhuman origins research. New York: Aldine deGruyter. p 132–147.

Bietti A, Grimaldi S, editors. 1996. Reduction processes(“chaines operatoires”) in the European Mousterian.Q Nova 6.

Bischoff JL, Soler N, Maroto J, Julia R. 1989. AbruptMousterian/Aurignacian boundary at c. 40 ka bp: ac-celerator 14C dates from L’Arbreda Cave (Catalunya,Spain). J Archaeol Sci 16:563–576.

Bischoff JL, Ludwig K, Francisco Garcia J, Carbonell E,Vaquero M, Stafford TWJ, Jull AJT. 1994. Dating ofthe basal Aurignacian sandwich at Abric Romanı(Catalunya, Spain) by radiocarbon and uranium-se-ries. J Archaeol Sci 21:541–551.

Bisson MS, Tisnerat N, White R. 1996. Radiocarbondates from the Upper Paleolithic of the BarmaGrande. Curr Anthropol 37:156–162.

Bocherens H, Billiou D, Mariotti A, Patou-Mathis M,Otte M, Bonjean D, Toussaint M. 1999. Palaeoenvi-ronmental and palaeodietary implications of isotopicbiogeochemistry of last interglacial Neanderthal andmammal bones in Scladina Cave (Belgium). J Ar-chaeol Sci 26:599–607.

Bordes F. 1947. Etude comparative des differentes tech-niques de taille du silex et des roches dures. Anthro-pologie (Paris) 51:1–29.

Bordes F. 1954. Notules de typologie paleolithique III:pointes mousteriennes, racloirs convergents et de-jetes, limaces. Bull Soc Prehist Franc 51:336–339.

Bordes F. 1959. Le contexte stratigraphique des hom-mes du Moustier et de Spy. Anthropologie (Paris)63:154–157.

Bordes F. 1968. The Old Stone Age. New York: McGrawHill.

Bordes F. 1972. Du Paleolithique moyen au Paleo-lithique superieur: continuite ou discontinuite? In:Bordes F, editor. The Origin of Homo sapiens. Paris:UNESCO. p 211–218.

Bordes F, Labrot J. 1967. La stratigraphie du gisementde Roc de Combe et ses implications. Bull Soc PrehistFranc 64:15–28.

Brauer G. 1980. Die morphologischen Affinitaten desjungpleistozanen Stirbeines aus dem Elbmundungs-gebiet bei Hahnofersand. Z Morphol Anthropol 71:1–42.

Brauer G. 1981. New evidence on the transitional pe-riod between Neanderthal and modern man. J HumEvol 10:467–474.

Brauer G. 1992. Africa’s place in the evolution of Homosapiens. In: Brauer G, Smith FH, editors. Continuityor replacement: controversies in Homo sapiens evolu-tion. Rotterdam: A.A. Balkema. p 83–98.

Brauer G, Broeg H. 1998. On the degree of Neandertal-modern continuity in the earliest Upper Palaeolithiccrania from the Czech Republic: evidence from non-metrical features. In: Omoto K, Tobias PV, editors.The origins and past of modern humans—towardsreconciliation. Singapore: World Scientific. p106–125.

Brooks AS, Helgren DM, Cramer JS, Franklin A,Hornyak W, Keating JM, Klein RG, Rink WJ,Schwarcz H, Leith Smith JN, Stewart K, Todd NE,Verniers J, Yellen JE. 1995. Dating and context of


three Middle Stone Age sites with bone points in theUpper Semliki Valley, Zaire. Science 268:548–553.

Brose DS, Wolpoff MH. 1971. Early Upper Palaeolithicman and late Middle Palaeolithic tools. Am Anthropol73:1156–1194.

Butzer KW. 1981. Cave sediments, Upper Pleistocenestratigraphy and Mousterian facies in CantabrianSpain. J Archaeol Sci 8:133–183.

Cabrera Valdes V. 1984. El yacimiento de la cueva de“El Castillo” (Puente Viesgo, Santander). Madrid:Bibliotheca Praehistorica Espana, volume XXII. In-stituto Espanol de Prehistoria.

Cabrera Valdes V, Bernaldo de Quiros F. 1996. Theorigin of the Upper Paleolithic: a Cantabrian perspec-tive. In: Carbonell E, Vaquero M, editors. The lastNeandertals, the first anatomically modern humans:a tale about the human diversity. Cultural changeand human evolution: the crisis at 40 KA BP. Tarra-gona: Universitat Rovira i Virgili. p 251–265.

Cabrera Valdes V, Bischoff J. 1989. Accelerator 14Cdates for early Upper Paleolithic (basal Aurignacian)at El Castillo Cave (Spain). J Archaeol Sci 16:577–584.

Cabrera Valdes V, Hoyos Gomez M, Bernaldo de QuirosF. 1997. The transition from the Middle to the UpperPaleolithic in the cave of El Castillo (Cantabria,Spain). In: Clark GA, Willermet CM, editors. Concep-tual issues in modern human origins research. NewYork: Aldine de Gruyter. p 177–188.

Campbell JB, Sampson CG. 1971. A new analysis ofKent’s Cavern, Devonshire, England. Eugene, OR:University of Oregon Press.

Carciumaru M. 1989. Contexte stratigraphique, paleo-climatique et geochronologique des civilisations duPaleolithique moyen et superieur en Roumanie. An-thropologie 93:99–122.

Caspari R. 1991. The evolution of the posterior cranialvault in the Central European Upper Pleistocene.Ph.D. Dissertation. Ann Arbor: University of Michi-gan.

Chabai VP, Marks AE. 1998. Preliminary synthesis:Middle Paleolithic assemblage variability in westernCrimea. In: Marks AE, Chabai VP, editors. The Mid-dle Paleolithic of western Crimea—volume 1. Liege:Etudes et Recherches Archeologiques de l’Universitede Liege.

Chilardi S, Frayer DW, Gioia P, Macchiarelli R, MussiM. 1996. Fontana Nuova di Ragusa (Sicily, Italy):southernmost Aurignacian site in Europe. Antiquity70:553–563.

Churchill SE. 1997. Morphological evolution, behaviorchange, and the origins of modern humans. In: ClarkGA, Willermet CM, editors. Conceptual issues inmodern human origins research. New York: Aldine deGruyter. p 202–219.

Churchill SE, Smith FH. 2000. A modern human hu-merus from the early Aurignacian of Vogelherdhohle(Stetten, Germany). Am J Phys Anthropol 112:251–273.

Churchill SE, Weaver AH, Niewoehner WA. 1996. LatePleistocene human technological and subsistence be-havior: functional interpretations of upper limb mor-phology. In: Bietti A, Grimaldi S, editors. Reductionprocesses (“chaines operatoires”) in the EuropeanMousterian. Q Nova 6:18–51.

Coon CS. 1962. The origin of races. New York: Knopf.Czarnetzki A. 1980. Pathological changes in the mor-

phology of the young Paleolithic skeletons from Stet-ten (southwest Germany). J Hum Evol 9:15–17.

Czarnetzki A. 1983. Zur Entwicklung des Menschen inSudwestdeutschland. In: Muller-Beck H, editor. Ur-

geschichte in Baden-Wurttemberg. Stuttgart: KonradTheiss Verlag. p 217–240.

Dansgaard W, Johnsen SJ, Clausen HB, Dahl-JensenD, Gundestrup NS, Hammer CU, Hvidberg CS, Stef-fensen JP, Sveinbjornsdottir AE, Jouzel J, Bond G.1993. Evidence for general instability of past climatefrom a 250-kyr ice-core record. Nature 364:218–220.

Deacon HJ, Wurz S. 1996. Klasies River main site, cave2: a Howiesons Poort occurrence. In: Pwiti G, Soper R,editors. Aspects of African archaeology: papers fromthe 10th Congress of the Pan African Association forPrehistory and Related Studies. Harare: Universityof Zimbabwe. p 213–218.

Delibrias G, Evin J. 1980. Sommaire des datations 14Cconcernant la prehistorique charentais. Bull Mem SocHist Archeol Charente 1971:189–202.

Delibrias G, Fontugne M. 1990. Datations des gise-ments de l’Aurignacien et du Mousterien en France.In: Farizy C, editor. Paleolithique moyen recent etPaleolithique superieur ancien en Europe. Nemours:Musee de Prehistoire d’Ile-de-France. Memoires 3. p39–42.

Delporte H, Djindjian F. 1979. Note a propos del’outillage aurignacian de la couche 11 de Bacho Kiro.In: Kozłowski J, editor. Middle and Upper Paleolithicin Balkans. Prace Archeol 28:101–103.

De Lumley H, Garcıa Sanchez M. 1971. L’enfant nean-dertalien de Cariguela a Pinar (Andalousie). Anthro-pologie (Paris) 75:29–55.

D’Errico F, Zilhao J, Julien M, Baffier D, Pelegrin J.1998. Neanderthal acculturation in western Europe?A critical review of the evidence and its interpreta-tion. Curr Anthropol [Suppl] 39:1–44.

De Sonneville Bordes D. 1959. Position-stratigraphiqueet chronologique relative des restes humains du Pa-leolithique superieur entre Loire et Pyrenees. AnnalPaleontol 45:19–51.

De Sonneville Bordes D. 1960. Le Paleolithique su-perieur en Perigord. Bordeaux: Delmas.

Duarte C, Maurıcio J, Pettitt PB, Souto P, Trinkaus E,van der Plicht H, Zilhao J. 1999. The early UpperPaleolithic human skeleton from the Abrigo do LagarVelho (Portugal) and modern human emergence inIberia. Proc Natl Acad Sci USA 96:7604–7609.

Farizy C. 1990. Du Mousterien au Chatelperronien aArcy-sur-Cure: un etat de la question. In: Farizy C,editor. Paleolithique moyen recent et Paleolithiquesuperieur ancien en Europe. Nemours: Musee de Pre-histoire d’Ile-de-France. Memoires 3. p 281–289.

Formicola V. 1989. Early Aurignacian deciduous incisorfrom Riparo Bombrini at Balzi Rossi (Grimaldi, Italy).Riv Antropol 67:287–292.

Formozov AA. 1958. Peshchernaya stoyanka Starosel’ei ee mesto v paleolite. Moscow: Materials and Inves-tigations of the Archaeology of the USSR, no. 71.

Frayer DW. 1978. Evolution of the dentition in UpperPaleolithic and Mesolithic Europe. Lawrence: Univer-sity of Kansas Publications in Anthropology 10.

Frayer DW. 1986. Cranial variation at Mladec and therelationship between Mousterian and Upper Paleo-lithic hominids. In: Novotny VV, Mizerova A, editors.Fossil man. New facts, new ideas. Papers in honor ofJan Jelınek’s life anniversary. Brno: Anthropos. p243–256.

Frayer DW. 1992. The persistence of Neanderthal fea-tures in post-Neanderthal Europeans. In: Brauer G,Smith F, editors. Continuity or replacement: contro-versies in Homo sapiens evolution. Rotterdam: A.A.Balkema. p 179–188.

Frayer DW, Wolpoff MH, Thorne AG, Smith FH, PopeGG. 1993. Theories of modern human origins: thepalaeontological test. Am Anthropol 95:14–50.


Frayer DW, Jelınek J, Oliva M, Wolpoff MH. nd. Aurig-nacian male crania from the Mladec Caves, Moravia,Czech Republic. Am J Phys Anthropol (in press).

Fric A. 1885. Uber einen Menschenschadel aus demLoss von Podbaba bei Prag. Sitzber koenig Bohm Gesvon Wissensch Prag, (Jahrgang 1884) 17:152.

Fridrich J, Sklenar K. 1976. Die palaolithische undmesolithische Hohlenbesiedlung des BohmischenKarstes. Prague: National Museum.

Gabori-Csank V. 1970. C-14 dates of the HungarianPaleolithic. Acta Archaeol Acad Sci Hung 22:1–11.

Gabori-Csank V. 1983. La grotte remete “Felso” (su-perieure) et le “Szeletien de Transdanubie”. Acta Ar-chaeol Acad Sci Hung 31:239–248.

Gabori-Csank V. 1990. Le Jankovichien en Hongrie del’Ouest. In: Farizy C, editor. Paleolithique moyen re-cent et Paleolithique superieur ancien en Europe.Nemours: Musee de Prehistoire d’Ile-de-France.Memoires 3. p 97–102.

Gambier D. 1989. Fossil hominids from the early UpperPalaeolithic (Aurignacian) of France. In: Mellars P,Stringer C, editors. The human revolution: behav-ioural and biological perspectives in the origins ofmodern humans. Princeton: Princeton UniversityPress. p 194–211.

Gambier D. 1993. Les hommes modernes du debut dePaleolithic superieur en France. Bilan des donneesanthropologiques et perspectives. In: Cabrera ValdesV, editor. El origen del hombre moderno en el sur-oeste de Europa. Madrid: Universidad Nacional deEducation a Distancıa. p 409–430.

Gambier D. 1997. Modern humans at the beginning ofthe Upper Paleolithic in France: anthropological dataand perspectives. In: Clark GA, Willermet CM, edi-tors. Conceptual issues in modern human origins re-search. New York: Aldine de Gruyter. p 117–131.

Gambier D, Houet F, Tillier AM. 1990. Dents de Font deGaume (Chatelperronien et Aurignacien) et de LaFerrassie (Aurignacien ancien) en Dordogne. Paleo2:143–152.

Gamble C. 1986. The Paleolithic settlement of Europe.Cambridge: Cambridge University Press.

Garcıa Sanchez M. 1960. Restos humanos del Paleo-lıtico medio y superior y del Neo-Eneolıtico de Pinar(Granada). Trab Inst Bernardino Sahagun AntropolEtnol 15:17–72.

Garralda MD. 1989. Upper Paleolithic human remainsfrom El Castillo Cave (Santander, Spain). In: Giaco-bini G, editor. Hominidae: Proceedings of the 2ndInternational Congress of Human Paleontology.Turin: Jaca Book. p 479–482.

Garralda MD. 1997. The human paleontology of theMiddle to Upper Paleolithic transition on the Iberianpeninsula. In: Clark GA, Willermet CM, editors. Con-ceptual issues in modern human origins research.New York: Aldine de Gruyter. p 148–160.

Garralda MD, Tillier AM, Vandermeersch B, CabreraV, Gambier D. 1992. Restes humains de l’Aurignacienarchaıque de la Cueva de El Castillo (Santander, Es-pagne). Anthropologie (Brno) 30:159–164.

Garrod DAE. 1926. The Upper Palaeolithic age in Brit-ain. Oxford: Oxford University Press.

Gaudzinski S. 1999. Middle Palaeolithic bone tools fromthe open-air site Salzgitter-Lebenstedt (Germany). JArchaeol Sci 26:125–141.

Gieseler W. 1937. Bericht uber die jungpalaolithischenSkeletreste von Stetten ob Lontal bei Ulm. Verh GesPhys Anthropol 8:41–48.

Gieseler W. 1940. Die urgeschichtlinchen Menschen-funde aus dem Lonetal und ihre Bedeutung fur diedeutsche Urgeshichte. Jahresbande Wiss Akad Tu-bingen NCD Dozentenbundes 1:102–127.

Gieseler W. 1971. Germany. In: Oakley KP, CampbellBG, Molleson TI, editors. Catalog of fossil hominidspart 1: Europe. London: British Museum (NaturalHistory). p 189–215.

Ginter B, Kozłowski JK. 1982a. Excavation and thestratigraphy of the cave. In: Kozłowski JK, editor.Excavations in the Bacho Kiro Cave (Bulgaria): finalreport. Warsaw: Panstwowe Wydawnictwo Naukowe.p 7–11.

Ginter B, Kozłowski JK. 1982b. Conclusions. In:Kozłowski JK, editor. Excavations in the Bacho KiroCave (Bulgaria): final report. Warsaw: PanstwoweWydawnictwo Naukowe. p 169–172.

Ginter B, Kozłowski JK, Laville H, Sirakov N, HedgesREM. 1996. Transition in the Balkans: news from theTemnata Cave, Bulgaria. In: Carbonell E, Vaquero M,editors. The last Neandertals, the first anatomicallymodern humans: a tale about the human diversity.Cultural change and human evolution: the crisis at 40KA BP. Tarragona: Universitat Rovira i Virgili. p169–200.

Gioia P. 1988. Problems related to the origins of ItalianUpper Palaeolithic: Uluzzian and Aurignacian. In:Kozłowski JK, editor. La mutation. Liege: Etudes etRecherches Archeologiques de l’Universite de Liege. p71–101.

Gioia P. 1990. An aspect of the transition between Mid-dle and Upper Palaeolithic in Italy: The Ulluzzian. In:Farizy C, editor. Paleolithique moyen recent et Paleo-lithique superieur ancien en Europe. Nemours:Musee de Prehistoire d’Ile-de-France. Memoires 3. p241–250.

Glen E, Kaczanowski K. 1982. Human remains. In:Kozłowski JK, editor. Excavations in the Bacho KiroCave (Bulgaria): final report. Warsaw: PanstwoweWydawnictwo Naukowe. p 75–79.

Golovanova LV, Hoffecker JF, Kharitonov VM, Ro-manova GP. 1999. Mezmaiskaya Cave: A Neander-thal occupation in the northern Caucasus. Curr An-thropol 40:77–86.

Gonzalez Echegaray J. 1980. El yacimiento de la Cuevade “El Pendo”. Madrid: Bibliotheca Praehistorica His-pana.

Gowlett JAJ. 1987. The archaeology of radiocarbon ac-celerator dating. J World Prehist 1:127–170.

Graves P. 1991. New models and metaphors for theNeanderthal debate. Curr Anthropol 32:513–541.

Guiot J, Pons A, de Beaulieu L, Reille M. 1989. A140,000-year continental climate reconstruction fromtwo European pollen records. Nature 338:309–313.

Hahn J. 1983. Eiszeitliche Jager zwischen 35000 und15000 Jahren vor heute. In: Muller-Beck H, editor.Urgeschichte in Baden-Wurttemberg. Stuttgart: Kon-rad Theiss Verlag. p 273–330.

Hahn J. 1993. L’origine du Paleolithique superieur enEurope centrale: les datations C14. In: CabreraValdes V, editor. El orıgen del hombre moderno en elsuroeste de Europe. Madrid: Universidad Nacional deEducacion a Distancia. p 61–80.

Hahn J. 1995. Neue Beschleuniger-C-14 Daten zumJungpalaolithikum in Sudwestdeutschland. Eiszeit-alter Gegenwart 45:86–92.

Hahn J. 1996. Le Paleolithique superieur en Allemagnemeridionale. In: Otte M, editor. Le Paleolithique su-perieur europeen: bilan quinquennal, 1991–1996.Liege: Etudes et Recherches Archeologiques del’Universite de Liege. p 181–186.

Harrold FB. 1989. Mousterian, Chatelperronian andearly Aurignacian in Western Europe: continuity ordiscontinuity? In: Mellars P, Stringer C, editors. Thehuman revolution: behavioural and biological per-


spectives in the origins of modern humans. Princeton:Princeton University Press. p 677–713.

Harrold FB. 2000. The Chatelperronian in historicalcontext. J Anthropol Res 56:59–75.

Hedges REM, Housley RA, Law IA, Bronk CR. 1989.Radiocarbon dates from the Oxford AMS system: ar-chaeometry datelist 9. Archaeometry 31:207–234.

Hedges REM, Housley RA, Bronk Ramsey C, vanKlinken GJ. 1994. Radiocarbon dates from the OxfordAMS system: archaeometry datelist 18. Archaeom-etry 36:337–374.

Henshilwood CS, Sealy JC. 1997. Bone artifacts fromthe Middle Stone Age at Blombos Cave, southernCape, South Africa. Curr Anthropol 38:890–895.

Hillebrand J. 1914. Az 1913 evi barlangkutatasaim er-emenyei. Barlangkutates 2:115–124.

Hoffecker JF. 1999. Neanderthals and modern humansin Eastern Europe. Evol Anthropol 8:129–141.

Holliday TW. 2000. Evolution at the crossroads: modernhuman emergence in Western Asia. Am Anthropol102:54–68.

Hublin JJ, Barroso Ruiz C, Medina Lara P, FontugneM, Reyss J-L. 1995. The Mousterian site of Zafarraya(Andalucia, Spain): dating and implications on thePaleolithic peopling process of Western Europe. C RAcad Sci [IIa] 321:931–937.

Hublin JJ, Spoor F, Braun M, Zonneveld F, Condemi S.1996. A late Neanderthal associated with UpperPalaeolithic artefacts. Nature 381:224–226.

Jacobi RM. 1980. The Upper Palaeolithic of Britain withspecial reference to Wales. In: Taylor JA, editor. Cul-ture and environment in prehistoric Wales: Oxford.British archaeological reports. p 15–100.

Jelınek J. 1978. Earliest Homo sapiens sapiens fromCentral Europe (Mladec, Czechoslovakia). Paper pre-sented at the Xth Int Cong Anthropol Ethnol Sci,Delhi, India.

Jelınek J. 1983. The Mladec finds and their evolution-ary importance. Anthropologie (Brno) 21:57–64.

Karavanic I. 1995. Upper Paleolithic occupation levelsand 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/Upper Paleo-lithic interface and the relationship of Neanderthalsand early modern humans in the Hrvatsko Zagorje,Croatia. J Hum Evol 34:223–248.

Karavanic I, Paunovic M, Yokoama Y, Falgueres C.1998. Neandertaliens et Paleolithique superieur dansla Grotte de Vindija, Croatie: controverses autour dela couche G1. Anthropologie 102:131–141.

Keith A. 1927. Report on a fragment of a human jawfound at a depth of (101⁄2 ft) 3.2 m. in the cave earth ofthe vestibule of Kent’s Cavern. Trans Proc TorquayNat Hist Soc 5:1–2.

Kitagawa H, van der Plicht J. 1998. Atmospheric radio-carbon calibration to 45,000 yr. B.P.: late glacial fluc-tuations and cosmogenic isotope production. Science279:1187–1190.

Klaatsch H, Hauser O. 1910. Homo Aurignacensis Hau-seri. Prahist Z 1:273–338.

Kozłowski JK. 1982. Excavation in the Bacho Kiro Cave(Bulgaria): final report. Warsaw: PanstwoweWydawnictwo Naukowe.

Kozłowski JK. 1996. Cultural context of the last Nean-dertals and early modern humans in Central-EasternEurope. In: Bar-Yosef O, Cavalli-Sforza LL, MarchRJ, Piperno M, editors. The Lower and MiddlePalaeolithic. Forlı: International Union of Prehistoricand Protohistoric Science. p 205–218.

Kozłowski JK, Dagnan-Ginter A, Gatsov I, Sirakova S.1982. Upper Paleolithic assemblages. In: Kozłowski

JK, editor. Excavation in the Bacho Kiro Cave(Bulgaria): final report. Warsaw: Panstwowe Wydaw-nictwo Naukowe. p 119–167.

Krings M, Stone A, Schmitz RW, Krainitzki H, Stonek-ing M, Paabo S. 1997. Neandertal DNA sequencesand the origin of modern humans. Cell 90:19–30.

Krings M, Geisert H, Schmitz RW, Krainitzki H, PaaboS. 1999. DNA sequence of the mitochondrial hyper-variable region II from the Neandertal type specimen.Proc Natl Acad Sci USA 96:5581–5585.

Kuhn SL, Stiner MC, Gulec E. 1999. Initial UpperPalaeolithic in south-central Turkey and its regionalcontext: a preliminary report. Antiquity 73:505–517.

La Place G. 1964. Les subdivisions du leptolithiqueitalien. Etude de typologie analytique. Bull PaletnItal LXXIII Fossellone 33–34.

La Place G. 1966. Recherches sur l’origine et l’evolutiondes complexes leptolithiques. Paris: Bonard Editions.

Le Gros Clark WE. 1966. History of the primates, 5thed. Chicago: Pheonix.

Leroi-Gourhan A. 1958. Etude des vestiges humainsfossiles provenant des grottes d’Arcy-sur-Cure. AnnPaleontol 44:87–148.

Leroyer C. 1988. Des occupations Castelperroniennes etAurignaciennes dans leur cadre chrono-climatique.In: Kozłowski JK, editor. La mutation. Liege: Etudeset Recherches Archeologiques de l’Universite deLiege. p 103–108.

Leveque F, Vandermeersch B. 1980. Decouverte derestes humains dans un niveau castelperronien aSaint-Cesaire (Charente-Maritime). C R Acad SciParis 291:187–189.

Lieberman DE, Pearson OM, Mowbray KM. 2000. Ba-sicranial influence on overall cranial shape. J HumEvol 38:291–315.

Lioubine VP. 1993. La chronostratigraphie du Paleo-lithique du Caucase. Anthropologie 97:291–298.

Lioubine VP. 1998. La grotte Mousterienne Barakae-vskaı (Nord Caucase). Anthropologie 102:67–90.

Malan M. 1954. Zahnkeim aus der zweiten AurignacienSchicht der Hohle von Istallosko. Acta Archeol Hung5:145–148.

Malez M. 1956. Die Hohle Veternica, eine neue pala-olithische Fundstelle in Kroatien. Bull Sci Yougosl3:11–12.

Malez M, Rukavina D. 1979. Polozaj naslaga spilje Vin-dije u sustavu clanjenja kvartara sireg podrucja Alpi.Rad Jugosl Akad Znanosti Umjetnosti 383:187–218.

Malez M, Smith F, Radovcic J, Rukavina D. 1980. Up-per Pleistocene hominids from Vindija, Croatia, Yu-goslavia. Curr Anthropol 21:365–367.

Mallegni F. 1992. Quelques restes humains immatures,des niveaux Mousteriens de la Grotte du Fossellone(Monte Circeo, Italie): Fossellone 3 (olim Circeo IV).Bull Mem Soc Anthropol Paris 4:21–32.

Mallegni F, Segre-Naldini E. 1992. A human maxilla(Fossellone 1) and scapula (Fossellone 2) recovered inthe Pleistocene layers of the Fossellone cave, Mt. Cir-ceo, Italy. Q Nova 2:211–225.

Manzi G, Passarello P. 1995. At the archaic/modernboundary of the genus Homo: the Neandertals fromGrotta Breuil. Curr Anthropol 36:355–366.

Marks AE. 1993. The early Upper Paleolithic: the viewfrom the Levant. In: Knecht H, Pike-Tay A, White R,editors. Before Lascaux: the complex record of theearly Upper Paleolithic. Boca Raton: CRC Press. p5–21.

Marks AE, Bicho N, Zilhao J, Ferring CR. 1994. UpperPleistocene prehistory in Portuguese Estremadura:results of preliminary research. J Field Archaeol 21:53–68.


Marks AE, Demidenko YE, Monigal K, Usik VI, FerringCR, Burke A, Rink J, McKinney C. 1997. Staroseleand the Starosele child: new excavations, new results.Curr Anthropol 38:112–123.

Matiegka J. 1924. Lebka podbabska. Anthropologie(Prague) 2:1–16.

McCown TD, Keith A. 1939. The Stone Age of MountCarmel II: The fossil human remains from the Leval-loiso-Mousterian. Oxford: Clarendon Press.

Mellars PA. 1996. Models for the dispersal of anatomi-cally modern populations across Europe: theoreticaland archaeological perspectives. In: Bar-Yosef O,Cavalli-Sforza LL, March RJ, Piperno M, editors. TheLower and Middle Paleolithic. Forlı: Int. Union Pre-hist. Protohist. Sci. p 225–235.

Mellars PA. 1998. The impact of climatic changes on thedemography of late Neandertal and early anatomi-cally modern populations in Europe. In: Akazawa T,Aoki K, Bar-Yosef O, editors. Neandertals and mod-ern humans in Western Asia. New York: PlenumPress. p 493–507.

Mellars PA. 1999. “Acculturation”, “co-existence” andthe end of the Neanderthals. J Hum Evol 36:13.

Mellars PA, Bricker HM. 1986. Radiocarbon acceleratordating in the earlier Upper Palaeolithic. In: GowlettJAJ, Hedges REM, editors. Archaeological resultsfrom accelerator dating. Oxford: Oxford UniversityCommittee for Archaeology. p 73–80.

Mercier N, Valladas H. 1996. GIF laboratory TL datesfor some Middle Palaeolithic sites. In: Bar-Yosef O,Cavalli-Sforza LL, March RJ, Piperno M, editors. TheLower and Middle Palaeolithic. Forlı: InternationalUnion of Prehistoric and Protohistoric Science. p 117–125.

Mercier N, Valladas H, Joron J-L, Reyss J-L, LevequeF, Vandermeersch B. 1991. Thermoluminescence dat-ing of the late Neanderthal remains from Saint-Ce-saire. Nature 351:737–739.

Messeri P, Palma di Cesnola A. 1976. Contemporaneitadi paleantropi e fanerantropi sulle coste dell’Italiameridionale. Zephyrus 26–27:7–30.

Miracle PT. 1998. The spread of modernity in Paleo-lithic Europe. In: Omoto K, Tobias PV, editors. Theorigins and past of modern humans—towards recon-ciliation. Singapore: World Scientific. p 171–187.

Montet-White A. 1996. Le Paleolithique en ancienneYougoslavie. Grenoble: Jerome Million.

Mook WG. 1982. Radiocarbon dating. In: Kozłowski JK,editor. Excavation in the Bacho Kiro Cave (Bulgaria):final report. Warsaw: Panstwowe WydawnictwoNaukowe. p 168.

Moure Romanillo JA, Garcia Soto E. 1983. Radiocarbondating of the Mousterian at Cueva Millan (Hor-tiguela, Burgos, Spain). Curr Anthropol 24:232–233.

Movius HL. 1969. The abri of Cro Magnon, Les Eyzies(Dordogne) and the probable age of the containedburials on the basis of the nearby Abri Pataud. AnEstud Atlantic 15:323–344.

Muller-Beck H. 1983. Sammlerinnen und Jager von denAufangen bis vor 35000 Jahren. In: Muller-Beck H,editor. Urgeschichte in Baden-Wurttemberg. Stutt-gart: Konrad Theiss Verlag. p 241–272.

Mussi M. 1986. On the chronology of the burials foundin the Grimaldi Caves. Antropol Contemp 9:95–104.

Necrasov O, Cristescu M. 1965. Donnees anthro-pologiques sur les populations de l’age de la pierre enRoumanie. Homo 16:129–161.

Newcomer MH. 1975. Punch technique and UpperPalaeolithic blades. In: Swanson EH, editor. LithicTechnology: making and using stone tools. TheHague: Mouton. p 97–102.

Nordburg M. 1998. On the probability of Neanderthalancestry. Am J Hum Genet 63:1237–1240.

Oakley KP, Campbell BG, Molleson TI. 1971. Catalogueof fossil hominids. Part II: Europe. London: BritishMuseum (Natural History).

Obermaier H. 1905. Les restes humains quaternairesdans l’Europe centrale. Anthropologie 16:385–410.

Obermaier H. 1924. Fossil man in Spain. New Haven:Yale University Press.

Oliva M. 1993. The Aurignacian in Moravia. In: KnechtH, Pike-Tay A, White R, editors. Before Lascaux: thecomplex record of the early Upper Paleolithic. BocaRaton: CRC Press. p 37–55.

Olszewski D, Dibble H. 1994. The Zagros Aurignacian.Curr Anthropol 35:68–75.

Otte M. 1990. From the Middle to the Upper Palaeo-lithic: the nature of the transition. In: Mellars P,editor. The emergence of modern humans: an archae-ological perspective. Ithica, NY: Cornell UniversityPress. p 438–456.

Otte M, Straus LG. 1995. Conclusions et resume. In:Otte M, Straus LG, editors. Le trou magrite: fouilles1991–1992. Liege: Etudes et Recherches Arche-ologiques de l’Universite de Liege. p 229–239.

Ovchinnikov IV, Gotherstrom A, Romanova GP, Khari-tonov VM, Liden K, Goodwin W. 2000. Molecularanalysis of Neanderthal DNA from the northern Cau-casus. Nature 404:490–493.

Palma di Cesnola A. 1993. Il Paleolitico superiore inItalia. Florence: Garlatti e Razzai.

Palma di Cesnola A, Messeri P. 1967. Quatre dentshumaines paleolithiques trouvees dans des cavernesde l’Italie meridionale. Anthropologie 71:249–262.

Peyrony D. 1933. Les industries “aurignaciennes” dansle bassin de la Vezere. Bull Soc Prehist Franc 30:543–559.

Pitti C, Tozzi C. 1971. La Grotta del Capriolo e la Bucadella Jena presso Mommio (Camaiore): sedimenti,fauna, industria litica. Riv Sci Preist 26:213–258.

Piveteau J, de Lumley H, de Lumley MA. 1963. Decou-verte de restes neandertaliens dans la grotte del’Hortus (Valflaunes, Herault). C R Hebd Seanc AcadSci Paris 256:40–44.

Ponce de Leon MS, Zollikofer CPE. 1999. New evidencefrom Le Moustier 1: computer-assisted reconstructionand morphometry of the skull. Anat Rec 254:474–489.

Prosek F. 1951. Vyzkum jeskyne Dzerave skaly MalychKarpatech. Archeol Rozhledy 3:293–298.

Prosek F, Starka V, Hrdlicka L, Hokr Z, Lozek V,Dohnal Z. 1952. Vyzkum jeskyne Zlateho Kone uKoneprus. Ceskoslovensk Kras 5:161–179.

Protsch R, Semmel A. 1978. Zur Chronologie des Kels-terbach-Hominiden. Eiszeitalter Gegenwart 28:200–210.

Protsch von Zieten RRR. 1988. Absolute dating by ra-diocarbon- and amino-acid-dating of latest Homo sa-piens neanderthalensis and earliest Homo sapienssapiens in Europe. In: Otte M, editor. L’homme deNeandertal, volume 1, la chronologie. Liege: Etudeset Recherches Archeologiques de l’Universite deLiege. p 81–115.

Reynolds TEG. 1990. The Middle-Upper Palaeolithictransition in southwestern France: interpreting thelithic evidence. In: Mellars P, editor. The emergenceof modern humans: an archaeological perspective.Ithaca, NY: Cornell University Press. p 262–275.

Richards MP, Pettitt PB, Trinkaus E, Smith FH,Paunovic M, Karavanic I. 2000. Neanderthal diet atVindija and Neanderthal predation: the evidencefrom stable isotopes. Proc Natl Acad Sci USA 97:7663–7666.


Richter D, Waiblinger J, Rink WJ, Wagner GA. 2000.Thermoluminescence, electron spin resonance and14C-dating of the late Middle and early UpperPalaeolithic site of Geißenklosterle Cave in southernGermany. J Archaeol Sci 27:71–89.

Riek G. 1932. Palaolithischen Station mit Tierplastikenund menschlichen Skelettresten bei Stetten ob Lon-tal. Germania 16:1–8.

Riek G. 1934. Die Eiszeitjagerstation am Vogelherd.Band 1: Die Kulturen. Leipzig: Curt Kabitzsch Ver-lag.

Rightmire GP, Deacon HJ. 1991. Comparative studiesof Late Pleistocene human remains from KlasiesRiver Mouth, South Africa. J Hum Evol 20:131–156.

Roebroeks W, Conard NJ, van Kolfschoten T. 1992.Dense forests, cold steppes, and the Palaeolithic set-tlement of northern Europe. Curr Anthropol 33:551–586.

Rossignol-Strick M. 1995. Sea-land correlation of pollenrecords in the eastern Mediterranean for the glacial-interglacial transition: biostratigraphy versus radio-metric time-scale. Q Sci Rev 14:893–915.

Rukavina D. 1983. O stratigrafiji gornjeg pleistocena sosvrtom na topla razdoblja i njihov odraz u nasl-agama na podrucju Jugoslavije. Rad Jugosl AkadZnanosti Umjetnosti 404:199–221.

Saller K. 1926. Die Rassen der juengeren Steinzeit inden Mittelmeerlandern; I: Die Iberische Halbinsel.Bull Assoc Catalan Antropol Etnol Prehist 4:1–36.

Schwarcz HP, Bietti A, Buhay WH, Stiner MC, Grun R,Segre AG. 1991. On the reexamination of GrottaGuattari: uranium-series and electron-spin-reso-nance dates. Curr Anthropol 32:313–316.

Shotton FW. 1977. The Devensian stage: its develop-ment, limits and substages. Philos Trans R Soc Lond[Biol] 280:107–118.

Singer R, Wymer JJ. 1982. The Middle Stone Age atKlasies River Mouth in South Africa. Chicago: Uni-versity of Chicago Press.

Skerlj B. 1958. Were Neanderthalers the only inhabit-ants of Krapina? Bull Sci Yougosl 4:44.

Skinner MF, Sperber GH. 1982. Atlas of radiographs ofearly man. New York: A.R. Liss.

Smith FH. 1976. A fossil hominid frontal from VelikaPecina (Croatia) and a consideration of Upper Pleis-tocene hominids from Yugoslavia. Am J Phys An-thropol 44:127–134.

Smith FH. 1977. On the application of morphological“dating” to the hominid fossil record. J Anthropol Res33:302–316.

Smith FH. 1982. Upper Pleistocene hominid evolutionin South-Central Europe: a review of the evidenceand analysis of trends. Curr Anthropol 6:667–703.

Smith FH. 1984. Fossil hominids from the Upper Pleis-tocene of Central Europe and the origin of modernEuropeans. In: Smith FH, Spencer F, editors. Theorigins of modern humans: a world survey of the fossilevidence. New York: Alan R. Liss. p 137–209.

Smith FH. 1994. Samples, species, and speculations inthe study of modern human origins. In: Nitecki MH,Nitecki DV, editors. Origins of anatomically modernhumans. New York: Plenum Press. p 227–-249.

Smith FH. 1997. Mladec. In: Spencer F, editor. Historyof physical anthropology. An encyclopedia. New York:Garland. p 659–660.

Smith FH, Ahern JC. 1994. Brief communication: addi-tional cranial remains from Vindija Cave, Croatia.Am J Phys Anthropol 93:275–280.

Smith FH, Trinkaus E. 1991. Les origines de l’hommemoderne en Europe centrale: un cas de continuite. In:Hublin J-J, Tillier A-M, editors. Aux origines d’Homo

sapiens. Paris: Presses Universitaires de France. p251–290.

Smith FH, Boyd DC, Malez M. 1985. Additional UpperPleistocene human remains from Vindija Cave,Croatia, Yugoslavia. Am J Phys Anthropol 68:375–383.

Smith FH, Falsetti AB, Donnelly SM. 1989a. Modernhuman origins. Yrbk Phys Anthropol 32:35–68.

Smith FH, Falsetti AB, Liston MA. 1989b. Morphomet-ric analysis of the Mladec postcranial remains. Am JPhys Anthropol 78:305 [abstract].

Smith FH, Trinkaus E, Pettitt PB, Karavanic I,Paunovic M. 1999. Direct radiocarbon dates for Vin-dija G1 and Velika Pecina late Pleistocene hominidremains. Proc Natl Acad Sci USA 96:12281–12286.

Soffer O. 1989. The Middle to Upper Palaeolithic tran-sition on the Russian Plain. In: Mellars P, Stringer C,editors. The human revolution: behavioural and bio-logical perspectives in the origins of modern humans.Princeton: Princeton University Press. p 714–742.

Sollas WJ. 1913. Paviland Cave: an Aurignacian stationin Wales. J R Anthropol Inst 43:325–374.

Straus LG. 1995. The Upper Paleolithic of Europe: anoverview. Evol Anthropol 4:4–16.

Straus LG. 1997. The Iberian situation between 40,000and 30,000 B.P. in light of European models of migra-tion and convergence. In: Clark GA, Willermet CM,editors. Conceptual issues in modern human originsresearch. New York: Aldine de Gruyter. p 235–252.

Straus LG, Bischoff JL, Carbonell E. 1993. A review ofthe Middle to Upper Paleolithic transition in Iberia.Prehist Eur 3:11–27.

Stringer CB. 1992. Replacement, continuity and theorigin of Homo sapiens. In: Brauer G, Smith FH,editors. Continuity or replacement: controversies inthe evolution of Homo sapiens. Rotterdam: Balkema.p 9–24.

Stringer CB, Brauer G. 1994. Methods, misreading, andbias. Am Anthropol 96:416–424.

Stringer CB, Hublin JJ, Vandermeersch B. 1984. Theorigin of anatomically modern humans in WesternEurope. In: Smith FH, Spencer F, editors. The originsof modern humans: a world survey of the fossil evi-dence. New York: Alan R. Liss. p 51–135.

Svoboda J. 1988. Early Upper Paleolithic industries inMoravia: a review of recent evidence. In: KozłowskiJK, editor. La mutation. Liege: Etudes et RecherchesArcheologiques de l’Universite de Liege. p 169–192.

Svoboda J. 1990. The Bohunician. In: Kozłowski JK,editor. Les feuilles de pierre. Liege: Etudes et Recher-ches Archeologiques de l’Universite de Liege. p 199–211.

Svoboda J. 2000. The depositional context of the EarlyUpper Paleolithic human fossils from the Koneprusy(Zlaty kun) and Mladec Caves, Czech Republic. JHum Evol 38:523–536.

Svoboda J, Lozek V, Vlcek E. 1996. Hunters betweenEast and West: the Paleolithic of Moravia. New York:Plenum Press.

Szombathy J. 1925. Die diluvialen Menschenreste ausder Furst-Johanns-Hohle bei Lautsch in Mahren.Eiszeit 2:1–34, 73–95.

Tattersall I, Schwartz JH. 1999. Hominids and hybrids:the place of Neanderthals in human evolution. ProcNatl Acad Sci USA 96:7117–7119.

Tillier A-M. 1999. Les enfants mousteriens de Qafzeh.Paris: Editions du CNRS.

Trinkaus E, Smith FH. 1985. The fate of the Neander-tals. In: Delson E, editor. Ancestors: the hard evi-dence. New York: Alan R. Liss. p 325–333.

Trinkaus E, Ruff CB, Churchill SE, Vandermeersch B.1998. Locomotion and body proportions of the Saint-


Cesaire 1 Chatelperronian Neandertal. Proc NatlAcad Sci USA 95:5836–5840.

Trinkaus E, Churchill SE, Ruff CB, Vandermeersch B.1999. Long bone shaft robusticity and body propor-tions of the Saint-Cesaire 1 Chatelperronian Nean-dertal. J Archaeol Sci 26:753–773.

Ulrix-Closset M, Otte M, Cattelain P. 1988. Le “trou del’abime” a Couvin. In: Kozłowski JK, editor. L’hommede Neandertal, volume 8, la mutation. Liege: Etudeset Recherches Archeologiques de l’Universite de Liege.

Valladas H, Geneste JM, Joron JL, Chadelle JP. 1986.Thermoluminescence dating of Le Moustier (Dor-dogne, France). Nature 322:452–454.

Van Andel TH. 1998. Middle and Upper Palaeolithicenvironments and the calibration of 14C dates beyond10,000 BP. Antiquity 72:26–33.

Van Andel TH, Tzedakis C. 1996a. European Palaeo-lithic landscapes: 140,000–30,000 years ago—a sum-mary. In: Bar-Yosef O, Cavalli-Sforza LL, March RJ,Piperno M, editors. The Lower and Middle Palaeo-lithic. Forlı: International Union of Prehistoric andProtohistoric Science. p 191–204.

Van Andel TH, Tzedakis C. 1996b. Palaeolithic land-scapes of Europe and environs, 150,000–25,000 yearsago: an overview. Q Sci Rev 15:481–500.

Vandermeersch B. 1981. Les hommes fossiles de Qafzeh(Israel). Paris: Editions du CNRS.

Vandermeersch B. 1984. A propos de la decouverte dusquelette Neandertalien de Saint-Cesaire. Bull MemSoc Anthropol Paris 14:191–196.

Van der Plicht J. 1999. Radiocarbon calibration for theMiddle/Upper Palaeolithic: a comment. Antiquity 73:119–123.

Vega Toscano LG. 1990. La fin du Paleolithique moyenau sud de l’Espagne: ses implications dans le contextede la peninsule Iberique. In: Farizy C, editor. Paleo-lithique moyen recent et Paleolithique superieur an-cien en Europe. Nemours: Musee de Prehistoire del’Ile de France. Memoires 3. p 169–176.

Vega Toscano LG, Hoyos M, Ruiz-Bustos A, Laville H.1988. La sequence de la grotte de la Carihuela (Pinar,Grenade): chronostratigraphie et paleoecologie duPleistocene superieur du sud de la Peninsule Ibe-rique. In: Otte M, editor. L’homme de Neandertal,volume 2, l’environnement. Liege: Etudes et Recher-ches Archeologiques de l’Universite de Liege. p 169–180.

Vermeersch PM, Paulissen E, Stokes S, Charlier C,Peer Pv, Stringer C, Lindsay W. 1998. A MiddlePalaeolithic burial of a modern human at TaramsaHill, Egypt. Antiquity 72:475–484.

Vertes L, de Vries H. 1952. Az Istalloskoi barlang au-rignaci II. Kulturajanak radiokarbon kormeghataro-zasa. Archaeol Ert 2:195.

Vicino G. 1986. Lo scavo paleolitico al Riparo Bombrini(Balzi Rossi di Grimaldi, Ventimiglia). Riv IngaunaIntemelia 39:1–10.

Vlcek E. 1951. Pleistocennı clovek z jeskyne sv.Prokopa. Anthropozoikum 1:213–226.

Vlcek E. 1956. Kalva pleistocennıho cloveka z Podbaby(Praha XLX). Anthropozoikum 5:191–217.

Vlcek E. 1957. Pleistocennı clovek z jeskyne na Zlatemkoni u Koneprus. Anthropozoikum 6:283–311.

Vlcek E. 1967. Der jungpleistozane Menschenfund ausSvitavka in Mahren. Anthropos 19:262–270.

Vlcek E. 1996. Patterns of human evolution. In: Svo-boda J, Lozek V, Vlcek E, editors. Hunters betweenEast and West: the Paleolithic of Moravia. New York:Plenum Press. p 37–74.

Vogel JC, Kronfeld J. 1997. Calibration of radiocarbondates for the late Pleistocene using U/Th dates onstalagmites. Radiocarbon 39:27–32.

Vogel JC, Waterbolk HT. 1967. Groningen radiocarbondates VII. Radiocarbon 9:107–155.

Vogel JC, Waterbolk HT. 1972. Groningen radiocarbondates X. Radiocarbon 14:6–110.

Wendt WE. 1976. “Art mobilier” from the Apollo 11cave, South West Africa: Africa’s oldest dated worksof art. S Afr Archaeol Bull 31:5–11.

Wolpoff MH. 1992. Theories of modern human origins.In: Brauer G, Smith FH, editors. Continuity or re-placement: controversies in Homo sapiens evolution.Rotterdam: A.A. Balkema. p 25–63.

Wolpoff MH. 1996. Neandertals of the Upper Paleo-lithic. In: Carbonell E, Vaquero M, editors. The lastNeandertals, the first anatomically modern humans:a tale about the human diversity. Cultural changeand human evolution: the crisis at 40 KA BP. Tarra-gona: Universitat Rovira i Virgili. p 51–76.

Wolpoff MH. 1999. Paleoanthropology, 2nd ed. Boston:McGraw-Hill.

Wolpoff MH, Smith FH, Malez M, Radovcic J, RukavinaD. 1981. Upper Pleistocene hominid remains fromVindija Cave, Croatia, Yugoslavia. Am J Phys An-thropol 54:499–545.

Yellen JE, Brooks AS, Cornelissen E, Mehlman MJ,Stewart K. 1995. A Middle Stone Age worked boneindustry from Katanda, Upper Semliki Valley, Zaire.Science 268:553–556.

Zampetti D, Mussi M. 1988. Du Paleolithique moyen auPaleolithique superieur dans le Latium. In:Kozłowski JK, editor. La mutation. Liege: Etudes etRecherches Archeologiques de l’Universite de Liege. p273–288.

Zilhao J. 1993. Le passage du Paleolithique moyen auPaleolithique superieur dans le Portugal. In: CabreraValdes V, editor. El origen del hombre moderno en elsuroeste de Europa. Madrid: Universidad Nacional deEducacion a Distancıa. p 127–146.

Zilhao J, d’Errico F. 1999. The chronology and tapho-nomy of the earliest Aurignacian and its implicationsfor the understanding of Neandertal extinction. JWorld Prehist 13:1–68.


Documents

Makers of the early Aurignacian of Europe - Cameron M