l

Preistoria Alpina v. 39 (2003) Museo Tridentino di Scienze Naturali Trento 2003 ISSN 0393-0157

La Combette (, , ): a Mousterian sequence in the mountain chain, between the plains of the and rivers

P IERRE-JEAN T EXIER, J EAN-PHILIP B RUGAL, E MMANUEL Dsscr.xux, C RISTINA L EMORINI, J OSE A NTONIO LOPEz S AEZ , ISABELLE THERY & Lucv WILSON.

ABSTRACT - Integration ofdata from various fields of investigation shows that the La Combette rock shelter was, at least during the Middle Palaeolithic , situated on the southern margin of a more northerly settlement area. The shelter appears to have been used several times as a refuge by small groups of hunters venturing into the southern reaches of a larger territory, so that they could exploit the resources offered by a mid-altitude mountain environment.

Key words: La Combette, Mousterian sequence, Durance and Calavon rivers. Parole chiave: La Combette, sequenza Musteriana, Durance e Calavon

Pierre-Jean Texier - CNRS, UMR 6130-CEPAM, 250 rue Albert Einstein, F-05 650 Valbonne. E-mail: [email protected] Jean-Philip Brugal- CNRS, UMR 6636 et GDR 1051-MMSH - BP 647, 5 rue du Chateau de I'Horloge, F­ 13094 Aix-en- . E-mail: [email protected] Emmanuel Desclaux - Laboratoire du Lazaret, 33bis Bd Franck Pilatte, F-06300 . Cristina Lemorini - Museo delle Origini, Dip. di Scienze Storiche,Archeologiche, Antropologiche dell' Antichita, Univ. "La Sapienza" di Roma, P.leAldo Moro 5, 1-001 85 Roma. E-mail : musori @rmcisadu.let.uniromal.it Jose Antonio Lopez Saez - Laboratorio de Arqueobotanica, Departamento de Prehistoria, Instituto de Historia, CSIC, Duque de Medinaceli 6,E-28014 Madrid. E-mail: [email protected] Isab elle Thery - CNRS, UMR 6l30-CEPAM, 250 rue Albert Einstein, F-05650 Valbonne. E-mail: thery@cepam .cnrs.fr Lucy Wilson - Department ofPhysical Sciences, University ofNew Brunswick in Saint John, P.O. Box 5050, CA-Saint John, NB, E2L 4L5. E-mail: [email protected]

1. INTRODUCTION ving a level ofresolution which has only rarely been possible in the past.

Located in the heart ofthe Luberon mountain chain (1 125 m), the site of La Combette (327 m) 1.1. Geologic and geographic outline contains a 7 metre thick stratigraphic sequence, with several layers ofextremely well-preserved Mouste­ By establishing direct connections between pos­ rian occupations. The shelter is oriented eastward, sible areas of raw material supply and the lithic arti­ has a surface area ofnot more than 50 sq.m., and is facts discovered at the site, an archaeopetrographic located in a narrow valley beside a spring. Multidi­ study has given us the outline ofthe territory covered sciplinary studies have resulted in palaeoenviron­ by the Mousterian groups who used the La Combette mental and palaeoethnologic reconstructions ofthe rock shelter. This territory features great geographic al use ofthe site within a well defined territory, achie - and geologic unity. It is naturally limited westwards .

•, F

78

07 ARDECHE

Vaison ­ la-Romaine I 11 ./ r • Les Arg i l ie rs ~ Baqnots­ COrn I Malaucene...... 1909m ) e Noyers­ sur-Ceze sur-Jabron t t I • • ~ o r a n ge 5Cl V 1826m ~ "( s eaumes' s Montagne de Lure ~ oe-versse /) B d I'A be CARPENTR.AS r:;/~o.rmo i ro/n u.J <5 au e I ser Saint-Etienne- • Bedarrides IS' _ • les-Orgues • IS' , ~~ I ~/ 5 e • Pernes-tes- ~ M t de V a UC\U 04 FOR CALQUIER Fomaines Berigoule 0 n s • L'lsle-sur- ) ALPES DE • Remouhns la- Bnquets - - i

Aramon • • Gordes CoJ\<§' ~ H A ~T E - P R O V E N C E

o~ 10 l~km

Echelle 1 : 600.000 (1cm = 6 km) Fig. 1 - Locations of the principal Mousterian stratified sites in the Vaucluse (France) .

by the Rhone corridor, southwards and eastwards by slopes, dissected by erosion through the brightly ochre­ the wide valley of the powerful watercourse that the coloured series ofthe Albian-Cenomanian. Durance used to be (before human alteration ofits flow These three geographic sectors are connected the Durance's discharge could be up to 3000 mvsec), geologically to the sub-alpine area , which consists of and finally northwards by the southern slopes ofMount calcareous series, including the Urgonian facies lime­ Ventoux and the Lure mountain, whose highest points stones which have played a major role in the morpho­ approach 1900 m (Fig. 1). After the chasm at Mirabe­ logic, tectonic and hydrologic development of the re­ au, and guided by the mid-Durance fracture zone, the gion (BRGM, 1986, 1998). The most southerly ofthe­ course of the Durance is oriented first of all NNE­ se, the Cretaceous Luberon chain, marks the boundary SSw, before turning NW-SE at Peyrolles and running between the sub-alpine area and Provence. It is orien­ i I for about 50 km along the southern side ofthe Lube­ ted E-W,about 50 km long, ranges from 5 to 10 km in I ron anticline. It reaches the Rhone south ofAvignon , width, and has an average altitude of 800 m. not long after passing through the Orgon Gap, which The only way through the Luberon, from south marks the western end ofthe Luberon chain. to north, is through the gorge of the Aiguebrun stre­ The landscape of this region is thus one of ex­ am. This passage was thus , even during the Middle treme contrasts.The dry calcareous mountains, ofcon­ Palaeolithic, an ideal location for ambushing animals, siderable height (Mt. Ventoux - 1909 m, Lure - 1876 especially those such as horse and deer, which are re­ m, Luberon - 1125 m), and the karstic remnants, of latively poorly adapted to steep , rocky, terrain. It was Barremian-Bedoulian age (the Vaucluse mountains, also an excellent location for finding mountain goats and the Albion and St. Christol plateaus, extending in their preferred environment. The La Combette rock from 600 to 1200 m), contrast, on the plains leading shelter is located near this passage. down to the main rivers, with gentler and more humid Several lines ofevidence indicate that the latest 79

Neandertal occupations of the shelter (levels A, B/C, tant reef-building episode, form the geologic under­ and D), were due to small groups ofhunters in search pinning ofthe region. ofmountain goats and horses. These hunters were tho­ Tectonic activity was responsible for a major roughly aware ofthe specific features ofthis mountai­ advance ofmarine waters from the south to the north, nous environment, and in particular they knew of its around 20 million years ago . During the Upper Mio­ lack of lithic raw materials. cene, 10 million years ago , the Luberon underwent a At the base ofthe sequence (levels E and FIG), process offolding and tilting, and then was thrust sou­ the faunal data are comparable, except that in this case , thwards. Around 8.5 million years ago, the sea retrea­ deer are the main species, since the climatic condi­ ted, making way for continental sedimentation. The tions were more favourable for their development (end sediments were rich in quartz grains, gravel and de­ of isotope stage 4). On the other hand, the sources of bris of organisms which followed the progression of raw materials used, the technical traditions, the use of the sea. They contributed to the formation of the mo­ fire, and the organisation of space within the shelter, lasse, which in particular covers all of the northern are very different than in the upper layers. side of the anticline in the area of Bonnieux. It is in this molasse layer that the rock shelters of La Com­ bette and of the Baume des Peyrards are found (de 1.2. The present climate LUMLEY, 1969). It took about three million years at the end of The local climate is as much affected by the the Miocene for the Luberon to take the shape that we corridors ofthe Rhone and the Durance, as by the high know today. The Aiguebrun, which existed even befo­ reliefbounding the region to the north and south. Con­ re the formation ofthe chain, following a fault, cut its sequently, it can show quite dramatic variations. Nowa­ way progressively downward through the ant icline as days the southern side ofthe Luberon benefits from a it rose, making a deep canyon. mild Mediterranean climate, while the plain ofApt has Around 5.5 million years ago, during the "Mes­ a humid Mediterranean climate, and the grassy sum­ sinian Crisis", the level ofthe Mediterranean was con­ mit ofthe Grand Luberon shows that it is subject to a siderably lowered, causing the downcutting of the colder, mountainous climate. The vegetation of the canyons ofboth the Rhone and the Durance, and fur­ summit zone is adapted to the rigourous conditions ther accentuating the gorge ofthe Aiguebrun. caused by the combined effect ofaltitude, wind, sum­ The beginning of the Quaternary is marked by mer drought and low winter temperatures. the intense erosion and sedimentation which accom­ In the valleys and combes, where springs and panied the advance and retreat ofthe Alpine glaciers. perennial rivers are rare, the occasional humid areas A whole system of fluvial terraces was created at the shelter a specific flora and fauna. The mistral wind same time as a thick layer of colluvium accumulated, comes from the north, starting at the entrance to the smoothing the contours ofthe steep hillsides (BRGM, Rhone corridor, and is cooled by passing over the hei­ 1986 , 1998). ghts ofMt. Ventoux and the Albion plateau. It is a very During the Palaeolithic, the mountainous ridge important component ofthe local climate, since it can ofthe Luberon, with its average height of800 m, would blow up to 260 kmlh on the summits. have been an important obstacle to travel, were it not There is a difference in insolation by a factor of for the Aiguebrun, which created a passage by exploi­ 10 between the most sun-exposed side and the shady ting a weak area and a fault in the centre ofthe chain. side ofthe hills, which results nowadays in a typically In addition, many other parameters oflife are also in­ Mediterranean vegetation on the southern slope of the fluenced, directly or indirectly, by the geological con­ Luberon, where green oak and Aleppo pine (Mediter­ text. These include for instance the relief (which af­ ranean stage) can be found. The northern slope is pre­ fects the thermal gradient and precipitation patterns); dominantly covered with a deciduous forest system the natural communication routes; the erosion ofshe1­ with white oak (hill stage). The long strip ofgrass near ters, caves and dolines; the size and quality ofthe sour­ Mourre Negre (1125 m) represents the mountain sta­ ces ofraw material or water; the nature ofthe vegeta­ ge ofthe Grand Luberon (MOLINIER, 1963; GlRED, 1978; tion cover (calciphile or siliciphile); and the animal BRGM, 1986; GUIDES GALLIMARD, 1996). species that are able to live there. The geological context thus had important con­ sequences for the Middle Palaeolithic inhabitants of 1.3. Short geological history ofthe Luberon the region. Their presence, on the other hand, had very little impact on the natural environment. They merely Thick deposits of limestone and mad dating used the hunting territories and the rich sources ofraw from the Upper Jurassic to the Lower Cretaceous, material, leaving behind traces offire , and the cutting marked during the Barremian-Bedoulian by an impor- edges oftheir unnaturally shaped artefacts. These, ac- ..... I

80

companied by the bones oftheir hunted or scavenged graphy, geomorphology, and geology, as well as on the prey, are the discreet but tangible and widespread evi­ circulation ofsilicic raw materials and the function of dence oftheir passage. the tools (TEXIERet al., 1996, 1998). The sequences ofthese three sites, which either have recently been or still are the subject ofinvestiga­ tion, cover a large chronological span, from the begin­ 2. MAJOR PALAEOLITHIC SITES IN THE RE­ ning ofisotope stage 7 to the middle of isotope stage GION 3 (BUISSON CATIL, ed., 1994). Being oriented eastward, the La Combette rock shelter is well protected from the cold northern winds. The shelter of La Combette was discovered by It is located in a narrow valley at the foot ofa molasse chance in 1971 by Brochier and Livache (BROCHIER, cliff, near some ofthe perennial springs that are scat­ 1976) who were looking for deposits contemporary tered through the region at the discordant contact with the final Magdalenian ofthe cave ofLa Combet­ between the Oligocene and the Cretaceous. te, which had been excavated starting in 1930 by Mo­ irenc and Vayson de Pradenne (MOIRENC& VAISONDE PRADENNE, 1934) and then, much later, by Paccard and Livache (PACCARD, 1976). The rock shelter is located 3. STRATIGRAPHY several dozen metres from the road bridge over the ravine, and several hundred meters downstream ofthe cave ofLa Combette (BROCHIER, 1976). The shelter is The shelter was filled up to its vault at the mo­ just a few kilometres from the Baume des Peyrards, ment of its discovery, with a stratigraphic sequence another famous Mousterian site in the Luberon, whi­ more than 7 m thick (Fig. 2) in which five principal ch was a major site in the development ofthe study of Mousterian layers have been detected (TEXIER et al., prehistory in (DELUMLEY, 1969; POR­ 1999). RAz,200l). Deposition occurred in three main stages that The La Combette rock shelter had to wait 15 may be interpreted as follows (starting at the base) : years after its discovery until it became the subject of a) Lower silty series. investigation, first of all by Andre Tavoso (TAVO­ - Deposition on the bedrock base ofa first sterile se­ so, 1986) and later by Pierre-Jean Texier. The excava­ dimentary layer with gravel , flakes and blocks of tions have just recently been completed. molasse. At about the same time, in 1987, the Canadian - Deposition ofa first silty layer (called the granular Serge Lebel undertook the first major scientific exca­ lower silty layer) containing the vestiges of one or vations at the Bau de l' Aubesier, on the territory of more Mousterian occupations (level F/G). The geo­ the commune of . With its 14 m ofPalaeoli­ logic study in progress will assess the processes of thic sediments and its rich archaeological levels, in­ deposition and the post-depositional evolution ofthe cluding H-l , which dates back to more than 200000 sediments. Ash and charcoal have given a greyish years, and the even older, lower level J-4, the Bau de colour to the sediments, which were originally yel­ l' Aubesier has become a stratigraphic reference for lowish but have been strongly affected by human calcareous sites in southern France (FERNANDEz et aI., actions. 1998; FERNANDEz, 2001; LEBEL, 2000 ; LEBEL et al., b) Sedimentary deposits oftorrential origin . 2001 ; BLACKWELL et al., 2001) . - Deposition from the north, in discordance with the The short, well dated sequence of the open-air erosion of the ravine , of coarse sediments (gravel, site of'Berigoule, investigated between 1988 and 1991, blocks and slabs ofmolasse ofvariable length, from is situated at about the beginning ofisotope stage 5. It a few decimetres to a few metres) of torrential ori­ produced a rich assemblage documenting repeated gin, forming a fan which spread into the southern centripetal Levallois debitage of a local raw material part ofthe shelter. (BRUGAL et al.,1989; JAUBERT et al.,199l; TEXIER & - Deposition of silty sediments correlated with an FRANCISCO-ORTEGA, 1995). important Mousterian occupation (level E). The The 7 m ofdeposits which have been excavated silt is sometimes full ofashe s and remarkably well in the La Combette rock shelter make this, too, an preserved charcoal. The human influence is clear I important reference site. Its importance lies in its par­ macroscopically (hearths, numerous charcoal pie­ 11 ticular geologic context, the Burdigalian molasse, in ces of considerable length in cm, bones and bur­ its data concerning Neandertal behaviour, and in the ned flints),'microscopically (presence of spores function ofthe site and its place in a territory that we of a fungus that prefers a carbon-rich environ­ can determine with unusual precision, based on geo- ment), and even physico-chemically (very high I

,I I ~ 81

, [ values ofmagnetic susceptibility ofthe sediment: 5. PALAEOENVIRONMENT i the oxidation due to the fire generated magnetic minerals). - Again from the north and in discordance with the Pollen analysis (Fig. 3a) has only been possible erosion ofthe ravine, a second deposit ofcoarse se­ for the lower part of the sequence, including the ar­ diments (gravel, blocks and slabs ofmolasse ofva­ chaeologicallayers FIG, E and D (LOPEZ SAEZ et al., riable length, from a few decimetres to a few me­ 1998). tres) oftorrential origin. The fan is even larger, and The APINAP ratio is always around 60/40. The preserves the second archaeological layer (E) in the pollen spectrum oflevel FIG suggests a fairly wooded northern halfofthe shelter; it spreads out and trun­ environment, dominated by Scotch pine, deciduous cates the archaeological layer in the southern part, oak, and juniper, while grassy areas are represented where the two fans merge . by the species Artemisia and Chenopodiaceae.The pre­ c) Upper silty series. sence ofbirch suggests that the climate was relatively There must have been partial or total infilling cold and humid. ofthe valley of La Combette during this sedimentary The spectrum of layer E shares the same spe­ cycle. It can be divided into three main episodes (star­ cies as FIG, with the addition ofchestnut, hazel, wil­ ting at the base): low and linden, which suggests that the climate was - Sub-horizontal stream-deposited silts, called "silt milder. The important human influence on this layer is with blocks", overlie the irregular upper surface of underscored by the abundance of ascospores of the the fan deposits, and form the base of the silty se­ carbonicolous type Chcetomium (type 7A), which in­ ries . In these sediments we discovered and excava­ dicate that the sediments are very rich in charcoal. ted over an area ofabout 50 sq.m . what we consider Human influence is also indicated by an increa­ to be a Mousterian living floor, sensu lato, (layer se by a factor of10 (Fig. 2) in the volumetric magnetic D), including several hearths. susceptibility of the sediment in the archaeological - The sequence continues with about 1 m of strati­ layers (k= 8 at 80x 10-5 SI; analysis by M. Dubar, Ce­ fied, sterile silts "exposing a rhythmic sequence of pam) . clayey silts, silty sands, sandy silts and silts, in whi­ The appearance offir, an increasing abundance ch greenish clay beds get thinner and progressively ofBetula, goosefoot and Artemisia, along with the di­ more widely spaced from the base to the top" (tran­ sappearance of Castanea and an abrupt decrease in slated from Txvoso, 1986). These sediments were probably deposited in a fairly humid period during which a gently flowing stream seems to have been the primary agent oftransport. Finely laminated stream-deposited silts giving way E progressively to definite loess deposits mark the uppermost 2 m ofthe La Combette sequence (Bro­ chier in TEXIER, 1999). These sediments contain two archaeological layers (B /C and A), which are distur­ bed, but whose assemblages are nonetheless plainly comparable to that oflayer D.

4. AGE OF DEPOSITS

Based on IRSL and TL measurements, the com­ plete excavated sequence may be situated between 45 and 70000 years ago (Wagner in TEXIER 1999; Valla­ das in TEXIER 1999). A large number ofmeasurements I: are currently in progress for a series of burned flints selected from the lithic assemblages oflevels E and FI I .. . h...... " ••" .." , j ~ .~ I""·..·,·'..lto. "" l_""'· G. The analysis ofthe thickness index oftooth enamel :.~ ;:;~" ; " ", ,,,Mo " I from micromammals ofthe species Arvicola (DESCLAUX t • Fig.2 - Sagittal section W-E in the Mousterian deposits of et al., 2000) also suggests a fairly old Upper Pleisto­ the La Combette rock shelter (Bonnieux, Vaucluse, cene age (stage 4?). France).

J 82

~ 70

~ 60

-- 50

--40 vaieurs en % - -3D

-- 20

10

;- 0

n-oes prelevements et coucne

taxons

Fig. 3a - Simplified pollen diagram ofLa Combette. Only a Fig. 3b - Preliminary anthracologic analysis of the few taxa are shown. The conventional order has been archaeological levels ofLa Combette (levels B/C to FIG). changed for the sake of readability.

the abundance of oak (APINAP: 45155) in the pollen available in the environment ofthe Bau de l'Aubesier spectrum of layer D (at the base ofthe upper silty se­ and La Combette (WILSON, 1996; TEXIER et al., 1998; ries) mark the return of a somewhat colder and more Wilson in TEXIER, 1999; TEXIER et al., in press) indica­ humid climate. tes that the potential sources are (presently) and were The charcoal is palaeoecologically significant (during the upper Pleistocene) abundant and diverse . because it results from the burning of fuel collected A recent study of raw material sources used exclusively as dead wood. The charcoals serves to cor­ during the Mesolithic (GUILBERT, 2000) has signifi­ roborate the data on the evolution of the vegetation cantly enriched the data available, extending it over a ofthis landscape, reflecting a deterioration in the cli­ larger geographic area. mate from the lower levels (FIG, E and D) towards the The potential sources can be classified in four top ofthe sequence (B/C). Pine is omnipresent in layers main groups: flint in primary position ofCretaceous, FIG, E and then D, where it begins to decrease sli­ Eocene or Oligocene age; and flint in secondary po­ ghtly, giving way to willow (Fig. 3b). Pine had almost sitions, in accumulations resulting from the weathe­ disappeared from the area of the shelter by the time ring andlor erosion and transportation ofolder rocks, that layer BIC was deposited, while willow had beco­ such as conglomerates or fluvial terraces. These can me the dominant taxon (Thery in TEXIER 1997; THERY date to the Cretaceous, Oligocene or even the Qua­ & TEXIER, in press). ternary (Fig. 4). Studies of the microfauna and of the malaco­ In the still preliminary research on the various fauna confirm that the overall climate during the for­ archaeological assemblages from La Combette (TEXIER mation of these deposits was colder than at present, et al., in press), we have been able to classify more and furthermore that the uppermost sediments of the sequence were deposited in an environment that was distinctly colder than that at the base of the sequence Local Tertiary Cretaceous Patinated (Brugal, Desclaux and Magnin in TEXIER 1999; De­ A 16,4% 27,1% 38,2% 18,4% sclaux in TEXIER, 2001). BIC 10,7% 39,2% 40,0% 10,1% D 6,3% 54,9% 32,7% 6,0% E 11,8% 46,9% 28,1% 13,2% 6. THE SILICIOUS RAW MATERIALS FIG 12,0% 43,0% 31,7% 13,4% TOTAL Col. 11,7% 44,0% 31,5% 12,9% Extensive research carried out in the Vaucluse Tab. 1 - % values and main geologic origins of identified and in neighbouring departments (particularly the Al­ material in the different archaeological layers of La pes-de-Haute Provence) regarding the raw materials Combette. 83

•

e-g

N A LEGENDE

Motass e, sables , mames et conqlomerats -Mioc ene Oalcaires . marnes et brecb ee -Eoc ene ~ Oliqo cene I n6-e2 I Mar nes , qres verts, sables ocreux et barioles * -Gargasien a Paleoqene / Calcaires, mames et marno -calc aires ~ Fail1es et chevauchements o 5 km -Jurassiqua super ieur a Bedoulien --=- -'--']

DAD P-J Texier, tracres les cartes geologiques BRGM Carpentras, Saett, Forcalquier, Reillanne au 1:50 000 et Pare Naturel Feg ional du Luberon au 1:100000 Fig. 4 - Principal areas of supply of siliceous materials used by the Mousterian groups that frequented la Combette rock shelter (Bonnieux, Vaucluse, France).

than 5000 pieces so far, and identify more than sixty petrographic types. 60 ,0% - For between 6% (D) and 18% (A) ofthe pieces, depending on the layer, we have no indication oftheir geologic origin, because ofthe alteration (patination) oftheir surfaces. The figures below are rough, inasmuch as they do not take into consideration all ofthe activities that could have occurred at the rock shelter that may have caused the over- or sub-representation of any source or area ofsupply (Tab. I, Fig. 5). The use ofTertiary flint (Eocene and Oligocene from the basin of Apt and from small basins within the Vaucluse mountains) is never below 33%, calcula­ ted excluding the patinated pieces. However, the Ter­ FIG tiary flints are the most likely to vary in their percen­ Fig. 5 - Layer D shows the most significant variability in tage of use, from 33% in layer A to 58% or more in raw material supply: anticipation of certain activities and layer D, a difference of around 25%. the needs associatedwith them resulted in the importation Use ofUrgonian flint from the north and north­ of large quantities of Tertiary (54.0%) and Cretaceous east of the department seems to have been more sta­ (32.4%) flints, withconcomitantlylessuse ofloeal materials ble, the percentage varying from 32 to 46% between (6.7%), whieh are of poor quality. layers, a difference of 14%. -

84

Local sources within the Luberon (such as a archaeological layers ofthe lower unit (E and FIG) that conglomerate ofpebbles with a greenish cortex, whi­ they have changed the colour of the sediment. The ch marks the base of the Miocene in contact with the bones and the lithic remains have also been greatly Urgonian limestone) represent on average about 13% affected by the action ofthis element (from 35 to 55% of the material; layer D is an exception to the rule, of the bones, and from 52 to 60% of the lithics), so with under 7%, which is 13.4% less than in layer A. much so that we suspect that the bones may have been A more in-depth analysis shows that even within used as an additional fuel for the fire. the geologic subdivisions described above, the precise The discovery of polygenic hearths, including sources used might have varied through time. Some of one built against the wall ofthe shelter using a block the types were present in all levels, although in diffe­ which fell from the roof, and on top ofwhat seems to rent proportions, while others were totally absent from be a stone pavement, suggests that the duration ofoc­ some layers. cupation of this layer was somewhat longer than we The accessibility ofthe different sources ofsup­ infer for the levels ofthe upper unit. ply must have varied through time, both during and Anthracologic analysis of the firewood used since the Middle Palaeolithic, depending on the deve­ during the different Mousterian occupations shows that . lopment and evolution ofthe periglacial scree deposi­ it was dead wood, greatly affected by micro-organi­ ts which accumulated in parallel with the infilling of sms. It would have been easy to ignite, but it must have the numerous caves in the region. burned out very quickly. The fact that this is the only combustible that has been found shows that its users preferred to use the fire to transfer heat through con­ duction (indirect cooking or thermal transfer to another 7. THE MAIN ARCHAEOLOGICAL LEVELS: material), rather than using it as a source oflight, or of D, EAND FIG radiant or convection heat (producing a flaming fire requires very dry firewood, rich in inflammable mate­ rials and therefore not weathered) (TEXIER et al., in The sequence of La Combette can be broadly press; THERY & TEXIER, in press). divided into two main units, with the summit of the second torrential episode marking the boundary betwe­ en the lower unit (layers FIG and E) and the upper unit 7.2. The upper unit (layers D, B/C and A). In order to understand how sediments have been 7.2.1. Layer D deposited, it is sedimentological practice to describe a The results obtained for this layer (so far only stratigraphy from the base to the top, thus respecting partially published) show that it is in many respects the chronological order of the deposits. This is what very close to the image ofan actual living floor. This we have done, so far. However, when we come to the interpretation is possible because ofthe integration of archaeological material, it becomes impossible to con­ data resulting from analyses in many different fields tinue to respect this rule, since the oldest layers, being of investigation, such as lithic technology, use-wear the most recently uncovered, are less well studied. Thus, studies, petroarchaeology, archaeozoology, anthraco­ I I at La Combette, one of the most recent Mousterian logy, and the spatial organisation within the shelter. levels, D, which is a reference layer for the site, has to The 50 sq.m of preserved surface area in this be compared with the older levels E and FIG, whose layer (Fig. 6) is just about the largest surface area that study has only just begun. could be used at La Combette. Furthermore, it is so well preserved that it has allowed us attempt a kind of high-definition archaeology (GowLETT, 1997) which is 7.1. The use offire in the main archaeological levels rarely even conceivable for a Middle Palaeolithic site. Thus we have been able to conclude that the len­ Although fire is present all through the sequen­ gth ofoccupation was brief, and we have determined II ce, there are major differences from one unit to another. the specific activities that took place there, in a short In the upper unit, fire was well circumscribed and has seasonal context. The occupation surface actually pro­ only slightly affected the archaeological remains. The tected by the shelter is relatively small (around 30 hearths, which are small and clearly marked by chan­ sq.m), and is delimited by the hearths. The distribu­ ges in the colour ofthe sediments, seem to have been tion ofthe remains shows that the activities were syn­ intentionally placed along the line marking the extent chronic and took place during a single period ofoccu­ ofthe rock shelter's overhanging roof(TEXIER et al., in pation. The remains are in a primary context, as the press) . depositional processes were favourable to the preser­ Charcoal and wood ash are so abundant in the vation ofthe structures. 85

11 12 13 14 15 16 17

s mcnnee ceccmtc snco et rapamnon au scttp jdes vesliges archaolog iques N w Vestiges osseux CJ Vesligesl ithiques

~ Surlaces ruberioose lfoye rs M '; ~ Galetsu tilises

D Blocs de calcairejurassique G Planes

L Parol ce I'abri B2J (molassebu rdigaliellne) .. . :; 0 ~ Lig n O de s u r p lo mb

K

I J

I',

Fig. 6 - Distribution ofarchaeolog ical material, hearts and large stones (Layer D).

The information that has come to us, through a remains, mainly rough knapping products, are oflocal limited taphonomic filter, gives us the image ofa tem­ origin. porary camp on the southern edge of a territory. The The numerous reduction sequences identified, small group ofNeandertals who stayed briefly in the based usually on only a small number of natural or shelter at that particular time engaged in sometimes complex activities like the tanning ofhides (Texier et al. 1996, Lemorini 2002) .These activities were strictly related to the result of their hunting expeditions. The fact that these activities were part ofa well organised system of exploitation of the territory, and that the Neandertals were fully aware ofthe distribution offood and mineral resources within that territory, is clearly shown by the fact that some oftheir actions show anti­ cipation ofthe constraints imposed by the environment. It is clear that the people ofLa Combette alrea­ dy knew ofthe existence ofthis shelter near a spring, and knew that the surrounding area was favourable for hunting, but poor in lithic resources. The techno-func­ tional study ofthe lithic assemblage shows that most of the pieces were already knapped, often retouched, and designed for very specific functions, and that they had been transported over distances ranging from 5 to 40 km to their place of use. Only 6,7% of the lithic 86

7.2.2. A grid for reading the archaeological data of layer D The concept of "operative chain" (LEROI-GOU­ RHAN, 1964) is very useful pour organiser les donnees archeologiques et les informations manquantes en tant qu 'elements et/ou phases des differentes chaines opera­ toires (PELEGRIN et al., 1988). Using this concept, the structures ofthe lithic and bone assemblages were pla­ ced in parallel in order to construct a grid (BRUGAL, 1995). The usefulness of this grid was tested with the archaeological data from the layer that seemed best su­ ited to it (TEXIER et al., 1998). This grid allows us to Fig. 8 - Small retouched tools used for wood-working, from propose a synthetic representation ofthe function ofthe Layers B/C and D of LaCombette rockshelter(traceological site at a particular moment in its use (layer D). Such an analyses: C.Lemorini, drawings: J.-G. Marcillaud). Note the interpretation ofthe role ofthe site is aided in particular great range of forms and types. by the excellent state of preservation of the remains, and their spatial integrity, which comes to us through a knapped retouched supports, seem to have been aimed reduced taphonomic filter. Another major role in this at maximizing size and usefulness. The knapped sup­ analysis is played by experimentation and functional ports were selected from both the initial stages and the analysis, which allow us to determine the nature and more advanced stages of reduction, unlike the situa­ sometimes even the processes used to transform the or­ tion that we observed for instance at the nearby lithic ganic materials which have not been preserved. knapping workshop and habitation site of Berigoule. The proposed model (Fig. 9) is divided into two Long, very curved, sidescraper edges were produced distinct zones . The upper part refers to what the avai­ on large, very flat flint nodules, on large cortical flakes, lable information can tell us about the behaviour of a on large flakes produced with hard (stone) hammers, given prehistoric group at the level of their territory. and even on large frost-cracked flakes. These consti­ The lower part refers to what we know about activities tute the basis ofa highly specialised tool kit designed carried out in the shelter. A combination ofhorizontal for preparing hides, which gives a very unusual cha­ reading and vertical reading allows us to compare the racter to the lithic assemblages of the upper levels of faunal and lithic data, while passing from a study of La Combette (Fig. 7). More diverse tools, of a more the nature ofthe different occupations to a study ofall normal size, were used for wood working (Fig. 8), while ofthe activities that could have been carried out within quarters of meat were cut up with the help of rough a specific area of exploitation (BRUGAL, 1995). cutting edges. This very specific organisation and use This change in scale is only possible when the of materials is well documented, and allows us to de­ transportation of resources, both lithic and biotic, is duce that there was also forethought with respect to taken into consideration. The parallel development of the acquisition of particular animal products and by­ the lithic and animal operative chains provides a sche­ products, whether for food (horse and mountain goat) matic representation of the adaptative relationships or for functional reasons. between the environment (physical and bioclimatic) A main camp (near the sources of raw mate ­ and the human groups who created the sites , whose rials), and some temporary camps (sometimes far from function was for the greatest part determined by the the main one), near which animals were killed and ecosystem. various activities depending on the result of the hunt By shedding some light on the principal causal were carried out, thus made up the basic socio-econo­ relationships linking the fauna and the tools, this ap­ mic organisation ofthe Mousterians at La Combette. proach provides stronger arguments in support of our The characteristics ofthe lithic and bone assem­ hypotheses about the function ofthe site at a particular blages of the two more recent archaeological layers moment in time. It also allows us to extrapolate these (B/C and A) suggest that the La Combette rock shelter ideas, with all necessary prudence, to other occupation was used as a temporary camp at least three times du­ layers presenting a similar structure, even though these ring the climatic period corresponding to the deposi­ are not as well preserved (TEXIER et al., 1998). tion ofthe silts and loess ofthe upper unit. As such it was an important part ofthe logistical organisation that groups of hunter-gatherers had developed in order to 7.3. The lower unit exploit a territory that they knew extremely well, in terms ofboth its potential and ofthe distribution ofits The oldest layers amongst those preserved in resources. the shelter (E and FIG) were affected by two periods 87

of deposition of sedirnents of torrential origin. They ces offire are not restricted to the area ofcombustion, have many sedimentological and archaeological fea­ but rather show that fire intensely affected the entire tures in common: excavated surface. The sediments are strongly influenced by hu­ The faunaI assemblages differ from those ofthe man action and in particular by the action offire, whi­ upper levels: horse (46% of the identifiable remains ch has left traces such as reddened sediments, ashes in E vs. 10% in F) and mountain goat dominate alter­ mixed with the silt, burned stones and charcoal. Many nately in E and FIG, while deer, which is scarce in remains ofmultiphase hearths have been detected, both layer D, is much more common in the lowermost le­ in layer FIG and in layer E. One of these, in layer E, vel. was positioned up against a block ofbreccia, and se­ The remains of a large bovine, probably of the ems to have been built on a bed ofstones piled against genus Bos, have also been found in these two layers. the wall ofthe shelter. A great deal ofbone and lithic Based on the faunal associations (macro- and micro­ material in different stages of incineration was found vertebrates), the malacology and the palaeobotany, the in this structure. The sediment is composed ofpowde­ oldest deposits at La Combette are evidence ofa fairly ry black or grey ashes, very rich in organic material, temperate climate and a forested environment, which but with no phytolites present now (ifthere ever were evolved towards more severe conditions and a more any) . There are no pieces of charcoal at the centre of open landscape as we approach the top ofthe sequen­ II 'I the structure, but they are particularly abundant and ce. I well preserved on its fringes (one 50 mm long piece There are many clear signs ofhuman actions on was collected from this layer) . the faunal material: removal ofmuscular mass (striae 56% ofthe flints and 35% ofthe bones recove­ of flesh removal), fractures (spiral fractures, percus­ red in layer E are burned, ifnot partly destroyed by the sion marks, splinters), and combustion (bones may action offire, while the respective figures for layer FI have been used accidentally or intentionally to feed G are 67% and 52% (TEXIERet al., in press). The tra- the fires).

ENVIRONMENT I CLIMATE Geomcrpnoroqv I Sea so n

Durcrops >

Acq uisition Specres oivers.ty

Regional area Raw ltldl..nals Selectl o~ (chOice) Prev r Carc asses ", (Territory ) 1----,-- Wf'tgnt iOuallty ­ _ Aoa rol11y l PhySlO1o(:J yrS ize - ' ,--- Domestic area (Site) , ' Typo lag'C81 I T,1xonomteal , Techn ll!ogJeal 5 rrucrur~ ... _ . ~ _ . • _ . ~ .. BIOIogiI;a I struclurE' , TnphonomJC

o Cm" ~e~v PM" " "~ ~ ' g ~ Mm"", ""'';'"'' ~ .~ '; Part s of seeteton e: I I re pres ,mlil:lion i cneraoees hamm,,~:'d ... ------... M~al l Marks I .!:! son US"VWid',i1l8IysIS Skln, SInews .. Abundance ~~::;'hed I eX(Je'lmema llon ~ Frequ onc y '1il 5 INR. NMI.UAM...} fCl o u c h e d ~ ... - - - M;lfrowlFraclul tl ~ ~ I. - - .... ,- ---JL- ...... • En~?D p;rterns

Datation: end of isotopic stage 4. beginning of Isotopic stage 3 (Tl) • - - - Sp

OPEN/COOL P,mwlsl 01Occup alionlsl ~ Length at occup'lhon(s} Topograph y, Water Spatial discontinu ity of the resources MODEL' ...... - Plateaus SOCtO·ECOLOGICAL FUNCTION(S) O FTHE SITE > 5"30 Km .. Gorges Clilfs / ' " Acquis ition l ow

Regional area ,.r Importod Se lection (choce) {Territory} .: --- MObHrty __ DIFFER ENTIAL _ ANTICIPATION ~~Ta rg c l ed l e s s ou rt; cS- Domestic area ," in a spcciscccetext (Site) ,' S I Low : Equus. Dapra " D~::~~r~:d . ------... sub.Ad, Ad " ~e~v m/ /:c'-,,,""',.,,-,,,""',,:-:,,'---"" ~g\ i 7=:i~~ ~~8ns- 1 I fragments ~ Adapted to a specific maintenance hard ~ Selected g: 1: task I Hammer Meal ; 11 soft --- - WOOD <'" SKIN !! ~~~~~fied ~ ~ ~S NRem. ,,400 hafting tHigh Ncoord.Remet tda & retouched ~ Bonellakes ~ ... ------Struclures andHe arlhs- - - - - .1 ~ Short Sheltered. S ri -River semt-ocmcc ac specialized groups ? LA COMBE TTE ROCK SHE LTER (Vauc luse, France) : Cr~ical ressource: Storage? T EMPORARY CAMP: A EAS ON ED EXPLOITATION OF SPECIF IC RESSOUR CES Ibex? Wood? •Skin " Meat ? I I I Fig.9 - Model ofthe operative chain applied to faunal and lithic data for a functional interpretation ofLa Combette Layer D (Brugal 1995; Texier et al. 1998)

--- 88

The lithic structures ofthe lower unit (FIG and The preliminary techno-functional analysis of E) are highly comparable to each other in terms ofte­ the lithic production sequences, and the study of the chnology, but differ distinctly from those of the silty traces left by intensive use offire, provide evidence of layers of the upper unit (0, BIC and A). Broadly spe­ very different technical and domestic behaviours. It aking, as we have seen, the main areas of supply re­ still remains to be shown whether this variability is mained the same all through the sequence, providing due simply to a difference in cultural tradition, or ifit siliceous raw materials in equivalent proportions. is due for instance to a different, and perhaps less cle­ However, a more detailed analysis suggests that the arly defined, use ofthe shelter, which may have been precise sources of supply might have varied. The li­ used for a longer period at a different time ofyear. thic assemblages ofthe lower unit are greatly affected Finally, the use-wear analysis demonstrated that by fire. However, we can see that they consist primari­ some of the tools were used for carving deer antler, ly ofdiscoid and centripetal recurrent Levallois reduc­ before we had found any remains of deer. This result tion sequences on blocks of more modest size which was then confirmed in the most striking way by the were in at least some cases exploited on site. Cores discovery of a fragment of deer antler bearing deep, and knapping debris are considerably more abundant semicircular incisions which had reached the inner than in the archaeological layers ofthe upper unit. spongiosa layer, thus preparing the piece to be broken On the other hand, retouched tools, mainly sim­ by bending. Although this is a manifest case of work ple or double sidescrapers which are small in size and on a deer antler by Neandertals, any similar discove­ low in resharpening potential, make up a much smal­ ries, particularly from this chrono-cultural period, are ler proportion ofthe assemblages here than in the up­ still very rare. per unit, and are often accompanied by knapped flakes (Fig . 10). This equipment was either imported already worked or produced on site. It consistently shows that CONCLUSION these people had a good knowledge ofthe mineral re­ sources in the local, neighbouring, and even distant environment (as much as 40 km from the site). Just as The role we believe this site to have played, all in the upper unit, the Mousterians of the lower unit through the Mousterian sequence ofLa Combette, re­ anticipated the local lack of siliceous raw materials flects the perennial nature ofthe territory used during south of the valley of the Cavalon. Thus the study of the Middle Palaeolithic: a territory whose limits were the archaeological layers has shown that the sources clearly defined by obvious geographic features. Du­ ofraw materials and their transport distances are very ring the entire sequence, the shelter was in a marginal similar in the two sedimentary units. position in the territory, serving as a southern camp Despite some encouraging preliminary results site for hunting expeditions of limited size and dura­ in the use-wear analysis, it has not so far been possible tion. to establish any evidence of anticipation with respect The anticipation ofthe objectives ofthese expe­ to specific domestic activities, such as was shown for ditions is more variable through time. The consisten­ instance for layer 0 (TEXIER et al., 1996). tly high proportion of allochthonous raw materials in the lithic units is evidence ofan outstanding knowled­ ge ofthe territory, ofits mineral resources, and ofthe­ ir locations. The variability in the technological behaviours, ~a ~&~ clearly detectable from one unit to the next, is due as

I ~ ~ .... -. much to the weight of tradition and technology as to the anticipation ofneeds related to very specific acti­ vities such as tanning hides (layer D). In one case (the lower unit, layers E and FIG), Levallois and discoid cores, partially worked on site, were used to provide small blanks to be retouched into a small number of sidescrapers without any particu­ larly marked characteristics. This interpretation is cor­ ~~~~ ~~ic~ ~ ~i~' ] roborated by the use-wear analysis. In the other case (upper unit, layers A, BIC, and D), the great majority ofthe tools were imported alrea­ Fig.IO - Examples (discoid block, Levallois flakes and sidescrapers) of the Mousterian lithic sequence from the dy worked and retouched into sidescrapers. The tools lower layers (E and FIG) in the La Combette rock shelter. represent around 30% of the potential or modified I1

89

supports. The low technological, but very marked pe­ looks more modern, and which was very near the tran­ trographic and especially functional specificity ofthe sition to the Upper Palaeolithic . largest ofthese, clearly documented by the functional analysis, is a clear indicator ofanticipation ofparticu­ lar, targeted, activities linked to the results ofhunting expeditions. ACKNOWLEDGEMENTS The La Combette rock shelter is a very special case among Middle Palaeolithic sites. Multidiscipli­ nary studies have been of great benefit in increasing This work stems from research developed within our understanding of the domestic and technological a program of investigation and excavation headed by behaviours of these groups of hunter-gatherers . The PJ. Texier,financially and logistically supported by the results give us a glimpse ofthe logistical organisation Ministry ofCulture and Communications (France), the of their socio-economic system. This approach also General Council ofthe Vaucluse, and the municipality gives us an impression of Neandertal society which of Bonnieux.

RESUME -L integration des donnees actuellement disponibles dans de multiples domaines d'investigation, permet d'ores et deja de disposer de nombreux arguments pour demontrer que l'abri de La Combette fut, un temps tout au moins durant le Paleolithique moyen, un element marginal d'un dispositif logistique situe plus au nord. Il apparait que I'abri a servi a plu­ sieurs reprises de refuge a de petits groupes de chasseurs s'aventurant aux confins meridionaux d'un vaste territoire pour y exploiter les ressources cynegetiques d'un environnement de moyenne montagne.

RIASSUNTO - Lintegrazione dei dati attualmente disponibili nei vari settori d'indagine, permette di disporre di numerose argomentazioni per dimostrare che il riparo di La Combette fu, alm eno durante il paleolitico medio, un punto marginale di un dispositivo logistico situato piu a Nord. Sembra infatti che il riparo abbia svolto in piu occasioni il ruolo di rifugio per i piccoli gruppi di cacciatori che si spingevano verso i confini meridionali di un vasto territorio per sfruttare le risorse cinegetiche di un ambiente di media montagna.

[I BIBLIOGRAPHY I

11 Arvicola Lacepede, 1799 (Mammalia, Rodentia). In : BLACKWELLB., SKI NNERA.R., BLICKSTEI NILB., LEBEL S. & i LEUNG H.Y.M., 2001- ESR Isochron dat ing analyses at France and Liguria (Italy) during the Middle and Upper I Bau de l'Aubesier, Provence, France: clues to U uptake Pleistocene., Act zool. crcov., 43, 1-2: 107-125. in fossil teeth. Geoarchaeology: An International FERNANDEZ P., FAURE M., GUERlN M.& LEBEL S., 1998 - Strategic Journal, vo1.16,7: 719-761. de chasse des neandertaliens du Bau de I'Aubesier ; choix BRGM, 1986 - Atlas des res sources du sous-sol et des eaux et opportunisme. In : Economie prehistorique : les souterraines du departement de Vaucluse, BRGM, comportements de subsistance au Paleolithique, Actes des : 40 . XVIl1e Rencontres Internationales d 'Archeologie et BR GM, 1998 - Decouverte geologique du Luberon, guide d'Histoire d 'Antibes-Juan-Les-Pins, octobre 1997. Sophia et carte geologique au 1/100.000. Antipolis, APDCA : 309-323. BROCHIER lE., 1976 - Abri de la Combette (Bonnieux, FERNANDEZ P., 2001 - Etude paleontologique et archeo­ Vaucluse).IXe Congres VISPp, Nice 13-18 septembre zoologique des niveaux d' occupations mousteriens au Bau 11 1976, Livret-Guide de I'excursion C2:113. de I' Aubesier (Monieux, Vaucluse): implications biochro­ BRUGAL I-P., 1995 - Archeologie et Zoologie pour un nologiques et palethnologiques. These, Univ. Claude ,! nouveau concept, la paleoethologie humaine, Prehistoire Bemard, 1. Anthropologie Mediterraneennes, 4: 17-26. GlRED B., 1978 - Inventaire ecologique et biostratigraphique BRUGAL I -P., JAUB ERT 1 & TEXIER P.-I, 1989 - Decouverte de la flore du departement de Vaucluse, Bulletin de la d 'un site mousterien de plein-air en Vaucluse (Berigoule, Societe d'Etude des Sciences Naturelles de Vaucluse, 47­ Murs) , Bulletin de la Societe Prehistorique Francaise, 48 eme annee. 86,3: 69-71. GOWLETT IAJ., 1997 - High definition archaeology: ideas BUISSON-CATIL J . (ed.), 1994 - Le Paleolithique moyen en and evaluation, WorldArchaeology, 29,2: 152-171. Vaucluse. A la rencontre des chasseurs neandertaliens GUIDES GALLIMARD, 1996 - Pare Naturel Regional : Luberon, en Provence nord-occidentale, Notices d 'Archeologie Eds. Nouveaux Loisirs. Vauclusienne, 3, . GUILB ERT R ., 2000 - Gestion des industries lithiques DESCLAUX E. , ABASSI M., MARQ uET r-c., CHALINE 1& VAN mesolithiques du sud-est de la France. These de

KOLFSHOTEN T., 2000 r- Distribution and evolution of l'Universite de I Pantheon-Sorbonne: 369. 90

JAUBERT I , TEXIER P.-J & BRUGAL I -P., 1991 - Berigoule TAVOSOA., 1986 - Bonnieux-La Combette (Vaucluse). Rapport (Murs, Vaucluse) . Rapport de fouill e programmee sur les fouilles effectuees du ler au 24 Juin 1986: 8. pluriannuelle 1989-1991 : 76. TEXIER P.-I, 1997 - Ab ri du pont de La Com bette, site LEBEL S., 2000 - Monieux, Bau de I'Aubesier. In : Bilan paleol ithique moyen aBonnieux (Vaucluse). Operati on Scientifique 1999, DRAC PACA, Service Regi onal de programmee pluriannuelle (1997-1999). Rapport I'Archeologie : 179 -180. interrnediaire 1997: 21 . LEBEL S., T RI NKAUS E., FAURE M. , FERNANDEZ P., GUERI NC., TEXIER P.-I, BRUGALI-P., LEMORINI C.& WILSONL., 199 8­ RICHTERD., M ERClERN ., VALLADAS H. & W AGNER G. A., Fonction d'un site du paleolithique moyen en marge d'un 2001 - Co mparative morphology and pal aeobiology of territoire: I'abri de La Combe tte (Bonnieux , Vaucluse), Middle Pleistocen e human remains from the Bau de In :Econom ie prehistorique: les comportements de I' Aubesier, Vaucluse-France, PNAS, vo!. 98, n 20 : 11097­ su b sistan ce a u paleolithique, X VIIIe Rencontres 11102. Internationales d 'Archeologie et d 'Histoire d 'Antibes, LEMORINI c.,2000 - Reconnaitre des tactiques d'expl oitation Editions APDCA, Ju an-I es-Pins: 325-348. du milieu au Paleolithique moyen. La contribution de TEXIER P.-I& FRANClSCO-ORTEGA I., 1995 -Main techno­ I'analyse fonctionnelle, BAR Int. Series 858: 142. logical and typological characteristi cs of the lithic L EROI -GOURHAN A. , 1964 - Le geste et la parole, t.l, as semblage from level I at Berigoul e (Murs-Vaucluse, Technique et language, Paris. France), The Definition and Interpretation ofLevallois LOPEZ SAEZ lA., TEXIER P.-I& BUI-THI-MAI M. , 1998 ­ Technology, Monographs in World Arc ha eo logy, 23, Paleoenvironnement durant le Pleistocene superieur en Prehisto ry Press: 21 3-248. Vaucluse: analyse palynologique des couches inferieures TEXIER P.-I, LEMORINI c., BRUGALI-P.& WILSON L., 1996 ­ de I'abri de La Combette (Bonnieux, Vaucluse, France), Une activite de proto-tannage des peaux dans I'habitat Trabajos de Prehistoria , 552: 151 -162. mousterien de La Combette (Bonni eux, Vaucluse , LUMLEY H . DE, 1969 - La Baume des Peyrards. In : Le France). Quaternaria Nova, actes de la Table Ronde Paleolithique inferieur et moyen du midi mediter­ Reduction processes (" Chaines operato ires") fo r the ran een dans son ca dre geologique, tome I, Ligurie­ European Mousterian, Rome 25-28 May 1995 : 189-211. Provence, CNRS vesupplement cl Gallia Prehistoire: TEXIERP-J, 2001 - L'abri de La Combette, site paleolithique 419-461. moyen a Bonnieux (Vaucluse), Rapport interrnedi aire: MOIRENCA. & VAISON DE P RA DENNESA., 1934 - La grotte de 25 . La Combe tte a Bonnieu x. Congres Prehistorique de TEXIER P-J et alii, 199 9 - L'abri de La Combe tte, site France, compte rendu de la 10e session: 427-434. paleolithique moyen aBonnieu x (Vaucluse), Rapport de MOLI NIER R. , 1963 - Vue d'ensemble sur la vege tation du synthese (1997-1 999): 84. Grand et du Petit Luberon (Vaucluse), Bulletin du TEXI ERP-J, BRUGALJ-P, LEMO RI NI C., THERY1.& WILSON L. Museum d 'Histoire Naturelle de Marseille, 1. XXIII: 5­ (in press) - Abri du pont de La Combette (Bo nnieux, 28 . Vaucl us e) : variabilite intrasite des comportements PACCARD M ., 1976 - Grotte de la Combette (Bonnie ux, neandertaliens. Actes du 126emes congres du CTHS, Vaucluse).IXe Congres UISPp, Nice 13-1 8 septembre Colloque Pre et Protohistoire: territoires, deplacements, 1976, Livret-Guide de I'excursion C2: 113. mobilite, echanges avril 2001, CTHS, Barbaza P ELEGRI N J. , K ARLI N C .& B OD U P., 19 88 - " Chaines & Jaubert ed .: 16. operatoires": un outil pour le preh isto rien , Technologie THERY1.& TEXIERP.-I (in press.) - L'utilisati on du bois mo rt Prehistorique, Ed. du CNRS, notes et monographies dan s le site mousterien de La Combett e (Vaucluse) : techniques n025: 55-62. apport d'une approche morphometrique de s charbons PORRAZG., 2001 - Etude techno-morphologique des pieces de bois, Bulletin de la Societe Prehistorique Francaise. amincies de s couch es 9 a 5 de la Baume des Peyrards WILSON L., 1996 - Mousteri an raw material strategies in a (Vaucluse). Mem oire de DEA, Universite de Pr ovence, reg iona l context in Southern France. Actes du XIJIeme Ce ntre d'Aix, UFR Civil isations et human ite s: 157. Congres UISPp, Forli, vo1.6, 1. 11 : 783-7 91.