Short human occupations in the Middle Palaeolithic level i of the Abric Romaní rock-shelter (Capellades, Barcelona, Spain)

Journal of Human Evolution 48 (2005) 157e174

Short human occupations in the Middle Palaeolithiclevel i of the Abric Romanı rock-shelter

(Capellades, Barcelona, Spain)

J. Vallverdua,b,*, E. Alluea,b, J.L. Bischoff c, I. Caceresa, E. Carbonella,A. Cebriaa, D. Garcıa-Antona, R. Hugueta,b, N. Ibaneza, K. Martıneza,

I. Pastoa, J. Rosella, P. Saladiea,d, M. Vaqueroa,*

aArea de Prehistoria, Departament d’Historia i Geografia, Universitat Rovira i Virgili,

Placa Imperial Tarraco, 1, 43.005 Tarragona, SpainbFundacion Atapuerca

cUS Geological Survey, 345, Middlefield Road, Menlo Park, CA 94025, United StatesdFundacion Duques de Soria

Received 26 March 2004; accepted 6 October 2004

Abstract

This paper presents a multidisciplinary study on the size of the occupied surfaces, provisioning strategies andbehaviour planning at the Romani rock-shelter, using the Middle Palaeolithic record of the level i. This level is datedaround 46.000 BP through U/Th ages. A behavioural interpretation is proposed, which emphasises the activities and the

systemic value of the archaeological artefacts and structures.Occupation patterns are identified on the basis of the accumulations formed by human activities. These

archaeological accumulations, consisting of artefacts and hearths, are easily defined visually as spatial units. The

relationships between these accumulations, established by means of refitted remains, indicate that differences can beestablished between: 1) small and medium-sized occupation surfaces; 2) restricted and diversified provisioningstrategies. This variability suggests that different modes of occupation are represented in the same archaeological level.

The human activities reveal the generalization of fire technology. In almost all sizes of the occupation surfaces, theexploitation of vegetal resources near the Abric Romani marks the threshold of the restricted provisioning strategy.Limited use and fragmented knapping activities are recorded in the lithic assemblage. Faunal remains show differential

transport. The exploitation of lithic, faunal and vegetal resources characterizes the diversified provisioning strategy.The small occupation surfaces and restricted provisioning strategies suggest short settlements in the Abric Romani.

This shorter occupation model complements the longer diversified provisioning strategy recorded in both small andmedium-sized occupied surfaces. The selection of precise elements for transport and the possible deferred consumption

* Corresponding authors. Tel.: C34 977 250640.

E-mail addresses: [email protected] (J. Vallverdu), [email protected] (M. Vaquero).

0047-2484/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jhevol.2004.10.004

158 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

in the diversified provision strategy suggest an individual supply. In this respect, Neanderthal occupations in theRomani rock-shelter show a direct relation to: 1) hunting strategic resources; 2) high, linear mobility.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: Middle Palaeolithic; Neanderthals; rock-shelter occupation patterns; Abric Romanı

Introduction

Behavioural strategies are currently consideredto be one of the most essential components ininterpreting the archaeological record of prehis-toric hunter-gatherers. This is especially clear inthe case of the European Middle Palaeolithic, sincebehaviour has become central to the debate on thebiological and cultural characterisation of Nean-derthals and their role in the evolutionarysequence of human populations. Of the variousdomains that define human behaviour, settlementpatterns are particularly important due to theireffects on the variability of lithic and faunalassemblages, a long-standing issue in MiddlePalaeolithic research.

Research into settlement patterns, then, focuseson two interrelated issues: on the one hand, themodes and types of occupation and, on the other,mobility strategies. Both issues can be approachedusing theoretical models derived from ethnoarch-aeological studies, which provide the sole availablereferent to define the variability spectrum in thearchaeological record. Nevertheless, these modelsgive rise to a series of significant problems and,although most archaeologists are aware of them,their implications are rarely accepted. The mainproblem derives from the different time scales thatcharacterise the archaeological and ethnoarchaeo-logical records. Archaeological units e generallyestablished using stratigraphic criteria e are notcorrelated with the time scale observed in theethnographic context. On the contrary, they arethe product of cultural and natural formationprocesses that have taken place over differentperiods of time. Unlike the occupational contem-poraneity of the material remains documented ina settlement of current hunter-gatherers, archaeo-logical levels are palimpsests that normally includeremains of different occupational events, which

cannot be distinguished only by stratigraphiccriteria.

Although at present archaeologists tend torecognise that the archaeological units theyanalyse are palimpsests, their reconstruction ofsettlement patterns often does not take intoaccount this temporal framework. As a matter offact, many interpretations seem to assume that allthe remains recovered in an archaeological levelare contemporaneous or, at least, correspond tothe same occupation type. The aim of this paper isto identify the occupation modes represented inlevel i of the Abric Romanı site (Capellades,Spain). We assume from the start that severaloccupational events may have contributed to itsformation. We also assume that these events couldhave been different, as far as factors like theduration of the occupation or group size areconcerned. Our approach to the archaeologicalrecord is, therefore, to define spatial units that arevertically and horizontally well characterised in-side the stratigraphic level. Establishing thetemporal relationships between these units is oneof the main aims of this research. Their synchro-nicity will not be assumed, but justified by theconnections found between them. This procedurewill make it possible to show the variability ofoccupational models, which would go unnoticed ifthe archaeological level were approached asa homogeneous unit.

In this paper we adopt a multidisciplinaryapproach to the so-called mousterian variability(Binford, 1973; Bordes, 1981). We aim to bringtogether different theoretical and methodologicalperspectives, such as lithic technology, zooarch-aeology, taphonomy and spatial analysis. Thearchaeological interpretation is based on a behav-ioural approach, which takes into account theprocessual nature of interpretation, and thesystemic value of the material remains and

159J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

the archaeological structures. This paper dealswith the organisational strategies of the occupa-tional events on level i of the Abric Romanı.

Antecedents and geo-ecological framework

of Abric Romanı and level i

The Abric Romanı is a sedimentary depositlocated in a cliff known as the Cinglera delCapello, at the village of Capellades (50 km westof Barcelona, 280 m above sea level) (Fig. 1). TheCinglera is made up of carbonated quaternarysediments of lacustrine and spring origin. Thecomplex neotectonic history of the Capellades areaand the fault of the river Anoia have contributedto the cliff formation. At the beginning of the 20thcentury, Amador Romanı carried out intensearchaeological activity at the Cinglera del Capello,and several pleistocene and holocene sites werefound and documented (Bartroli et al., 1995). Thedates of these deposits have been established byarchaeological seriation and radiometric dating atdifferent sites. The quantity and quality of U/Th

dates obtained at the Abric Romanı site areconsiderable (Bischoff et al., 1988; Bischoff andFitzpatrick, 1991; Carbonell et al., 1994), and thestratigraphic succession has been dated by U/Thbetween 70 and 40 ka BP. This time span extendsfrom the middle of OIS 3 back to OIS 5(Dansgaard et al., 1993). In addition, humanoccupations have contributed to the sedimentarycomposition with materials that are suitable forradiocarbon dating. This research has made itpossible to verify and calibrate the U/Th and C14

dates in key biocultural periods, like the Middle/Upper Palaeolithic boundary at the Abric Romanı(Bischoff et al., 1994), or the Upper Pleistocene/Holocene boundary at the Abric Agut (Vaqueroet al., 2002)

The vegetation at the cornices of the Cingleradel Capello is outstanding because of the presenceof species encrusted by carbonates. There is anabundance of different mosses, but the presence ofbryophyte pseudomorphs (Dydymodon, Eucladium,Ctenidium) and the fern Adiantum capillus-veneris(Carbonell, 1992) in fossil travertines is par-ticularly significant. This is an association that is

EUROPE

0 - 400 m

> 400 m

C, Capellades

B, Barcelona

Regional framewok

0 60 km

N

Anoia riverAnoia

Penedès

cB

41°33'

32'

31'

1°41'42'0 5Km

8

11

10

9

6

5

4

2

3

1

AbricRomaní

7

Fig. 1. Geographic and geologic location of the Abric Romanı (Capellades, Anoia, Barcelona) in the NE of the Iberian Peninsula.

Legend of the geological framework of the Capellades area (IGME, 1973, 1975). 1, Plutonic Rock. 2, Palæozoic. 3, Mesozoic. 4,

Coenozoic. 5, Quaternary travertine. 6, Quaternary undifferentiated. 7, anticline. 8, syncline. 9, thrust fault. 10, reverse fault (?). 11,

fault.

160 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

characteristic of dripping environments, which arecalled Adiantetea communities, typical in travertinedeposits (Folch, 1981).

Twelve archaeological levels (from A to M)have been excavated so far at the Abric Romanısequence. Except for level A (Early Upper Palae-olithic), all the other levels correspond to MiddlePalaeolithic occupations. Amador Romanı mostlyexcavated the uppermost archaeological layers atthe beginning of the 20th century and they provideonly very fragmentary information on humanstrategies. It is from level H downwards (Carbonell,1992) that the archaeological fieldwork carriedout since 1983 has exposed a continuous surfacesuitable for a spatial reading of the archaeologicalrecord. Level i, mainly excavated between 1991and 1992, is in this context. The chronology oflevel i has been established by U/Th at ca.46 ka BP. The U/Th dates correspond tothe travertine layers overlying and underlyingthis level. We have four results for the underlyingtravertine: 48.6G 2.3; 46.9G 2.6; 46.5G 1.5 and46.3G 2.4. The overlying travertine has beendated at 45.1G 3.1 ka BP. This layer has recentlybeen dated at 46.6G 1.7 ka BP (Bischoff, per.com. 1.07.02).

According to the pollen analysis, level i islocated in the coldest phase of the Abric Romanıstratigraphy (Burjachs and Julia, 1994). This pollenzone is characterised by low percentages ofarboreal pollen. The dominance of Asteraceae,Poaceae, and Artemisia suggests steppe vegetation.

The archaeological excavation of level i exposeda large surface of over 195 m2. Nevertheless, insome areas ancient excavations had already re-moved the level. These areas were basically Pits 1and 2, excavated by Amador Romanı (Bartroliet al., 1995). During the fifties, Dr. Eduard Ripollcleaned Pit 1 and excavated two L-shaped trenchesthat also affected level i. In Romani’s and Ripoll’sstratigraphies, level i appears as Bed 11.

Level i is located on a discontinuity surface,which is documented in all the available sections ofthe trial excavations in the archaeological inter-ventions mentioned earlier. This surface wasformed by gours or pools situated along a steppedsurface. Two transversal travertine bars separatingthree relatively flat surfaces formed these steps.

The two upper surfaces contained metre-size goursclose to the travertine bars and joined by smallpaleo-channels. In the lowermost surface, thegours were centimetre-size. In the sedimentarydeposits filling the gours, level i overlay a series oflow energy facies (laminated sands). In fact, thearchaeological record is mostly included in a layerof crystalline gravels from the shelter’s wall. Thesurfaces regularised by the travertine bars, wherethe gours are situated, consisted of travertineaccretions and tuffaceous bioconstructions.

The discontinuity surface where level i wasdocumented was fossilised by a tabular stratum.This stratum was also overlaid throughout theshelter by a discontinuity surface with metre-sizeand centimetre-size gours in travertine accretiondeposits. Archaeological level H is located overthis stratum.

Sedimentary processes of subaerial exposure esuch as stromatolithic accretion and cryoclastismeare common in the lithological composition ofarchaeological level i. It should be emphasisedthat the sedimentary successions at the top andthe bottom of level i do not contain archaeolog-ical remains. Excavation over a large surfacemade it possible to characterise the variouslithological contexts where the archaeologicalremains are located: crystalline gravels, laminatedsands of gour fillings and travertine deposits.

The structures and archaeological record of thewell-preserved human occupations correspond toa phase of dry sedimentation. At least, this phase isdrier than the one corresponding to the successionof gour facies and travertine accretion, whichindicates wetter sedimentary processes with a pre-dominance of dripping from the shelter’s ceiling(Giralt and Julia, 1996). The detailed archaeostra-tigraphy indicates that these wetter facies aredistributed before and after the formation of thearchaeological record from human activities.

Materials and methods

The interpretation of the archaeological recordfrom level i is the outcome of a multidisciplinarystudy recently published in a monographic book(Carbonell, 2002). Below we synthesise the results

161J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

provided by the different analytic domains: lithictechnology, use-wear analysis of artifacts, rawmaterial provisioning, zooarchaeological and taph-onomic studies of faunal remains, identificationand characterization of vegetal remains (basicallycharcoal and wood pseudomorphs). Finally, wedescribe and analyse the evident structures(Meignen, 1994). The most common archaeologi-cal structures are hearths. Other archaeologicalstructures documented in the rock-shelter are theproduct of travertine blocks moving inside theshelter. Other natural features, such as largeblocks, or the stalagmite of the Coveta Ripoll,may have been used in occupation patterns. CovetaRipoll is a small dissolution cavity close to the wallof the rock-shelter delimited by a speleothemdevelopment from the wall and the ground.

The lithic assemblage

The number of lithic remains recovered in level iis relatively low (about 500 remains larger than1 cm), especially if we take into account that theamount of surface excavated was extensive. Flint isthe dominant raw material (60%), although thevariability is considerable compared to otherarchaeological levels of the Abric Romanı succes-sion (Carbonell et al., 1996a; Carbonell et al.,1996b). The percentage of quartz and limestonewas also significant (39%). At other archaeologicallevels, especially the uppermost ones, the domi-nance of flint is more pronounced. In this respect,level i resembles the immediately underlyingarchaeological level Ja, characterised by a signifi-cant use of quartz and limestone, both of whichshould be considered as strictly local, since theyare available and plentiful in the vicinity of thesite. The geological formations bearing a significantquantity of flint are further away, although only5 km from the shelter. In addition, there are largeoutcrops of good quality flint within a 15 kmradius.

The technological processes are characterisedby a flexible reduction strategy, aimed at max-imising the number of flakes per core. The knap-ping methods should be placed in a context definedby the absence of hierarchical cores and a lowdegree of predetermination. These characteristics

are common in the framework of discoid methods.Nevertheless, the small number of cores found inlevel i prevents the reduction strategies from beingreconstructed in detail. The number of retouchedartifacts is also very low and almost all of them aredenticulates and notches, as is the case with theother Middle Palaeolithic levels at the AbricRomanı. The technological processes indicatelow intensity knapping activities, and use-wearanalyses also show a limited use of artifacts. Thereduction sequences are temporally and spatiallyvery fragmented. There is some relationshipbetween the evidence of woodworking and re-duction sequences where most of the chaıneoperatoire is present at the site. On the otherhand, the artifacts used in butchering are itemstransported from reduction sequences carried outelsewhere. In general, the retouched artifacts seemto correspond to cores knapped outside theshelter. Transport is therefore a key factor in theprovisioning and use of raw materials.

The spatial distribution of lithic remains showssome well-defined and delimited accumulations,corresponding to the location of core reductionsequences. The size and rawmaterial of the artifactsin these accumulations are clearly homogeneous,which, in turn, indicates a small number ofknapping episodes. Moreover, the various accumu-lations are not connected by refits. The absence ofartifact transport between the main knapping areassuggests that they were not contemporaneous andmay belong to different occupation events.

The faunal remains

The faunal remains are made up of equid,cervid and bovid and they clearly show evidence ofhuman activity directed towards obtaining animalnutrients. The analysis of cut marks, bonefractures, and signals of burning on bones in-dicated a primary access pattern to these herbi-vores, mostly prime-age adults. The prey was notbrought into the camp as complete carcasses;cranial and limb parts were selected and trans-ported to the rock-shelter. The axial elements ofcervid were also transported occasionally. Largelongitudinal cut-marks on bones also suggest thatstrips of meat were obtained (Bunn, 1983).

162 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

The technique of fracturing bone to obtain themarrow consists of separating the diaphysis andthe epiphysis. It is important to note thatepiphyses were systematically destroyed, sincetheir representation in the record is almost null.The areas of bone breakage are spatially welldelimited. One of these areas is located beside theCoveta Ripoll. Another is located in the squaresI-J/53-54, close to a travertine block.

The impact of other taphonomic agents is low.For instance, carnivore marks have only beenrecognised in 1.5% of the bone remains. Of thenatural agents, water stands out as it affects 9.9%of the bones, even though most of the roundingand smoothing affects only one side (the oneexposed to the surface). This, in turn, also suggestsweak water streams. However, we do not rule outthe possibility that, in certain areas of the rock-shelter, the record was displaced. Finally, theincidence of trampling and weathering in therecord is particularly low (1.3% and 0.5%respectively).

The archaeobotanic record

The archaeobotanic record of Level i is diverseand peculiar, since we have been able to recoverdifferent categories of macro and micro remains:wood imprints (WI), charred woods, charcoalsand phytoliths.

A total of 452 charcoals were recovered fromthe hearths. The main taxon identified was Pinussilvestris/nigra salzmannii. This taxon includes twospecies, scots pine and black pine, that currentlygrow at over 800 m above sea level and which havebeen recognised in about 60% of the charcoals.Similarly, as in the rest of the sequence, theprevalent percentage of conifers denotes theexistence of a forest growing near the site.Furthermore, pinewood was preferred as fuelbecause of its abundance in the immediateenvironment (Allue, 2002).

Phytolith analysis was performed on both theregular sediments from level i and the sedimentsfrom the hearths. Gymnosperms, and in particulartracheids, are the most important silicate phyto-liths. Phytoliths from gramineae are less commonand cells from the Festucoideae subfamily have

been recognized. There are many oxalate crystalsin most of the samples and prismatic shapes ofPinus sp., although some of the parenchymatoustissue shapes may belong to Pinus sylvestris.

Over 100 wood imprints (WI) are documentedin level i. Their average length is 30 cm and theiraverage width 7 cm. The most outstanding im-prints from level i are a trunk (measuring 3 x 0.5 x0.5 meters) (Fig. 2) and two large accumulations.One of them is located beside a hearth and theother one on top of it, in a position that looks likea tripod that has fallen on top of a hearth (Fig. 3)(Carbonell and Castro-Curiel, 1995). There is alsoanother group consisting of a dense concentrationof small imprints that we interpret as the remainsof woodworking activities. The ratio of burned tounburned remains in these imprints is 1:4 (burned/unburned). In many imprints, only the distalsegment is burned.

Fig. 2. Imprint on a tree trunk from level i.

163J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Hearths

Sixteen hearths have been recorded on level i.They are flat, without preparations and blocks,and correspond to individual episodes of combus-tion. Hearths and archaeological artefacts arespatially associated. The accumulations of archae-ological material in and around hearths show anoccupational pattern characterised by the absenceof human secondary deposits. The micromorpho-logical analysis of hearths also shows the absenceof secondary deposits, and an intensity of com-bustion lower than 300 degrees (Carbonell, 2002).

Anthropic and naturally located travertine blocks

There are many travertine blocks from theshelter’s roof and wall. The spatial distribution ofthese blocks shows two types of accumulations.One line of blocks, parallel to the wall of theshelter, was formed as a result of the roofcollapsing. Other groups of travertine blocks arelocated outside this line. These blocks clearly havea stratigraphic relation with the archaeologicalassemblage and hearths, which suggests displace-ment during human occupations. In the fieldwork,these associations of blocks are known as InternalOrganisation of Travertine (IOT).

Spatial distribution: observed accumulations

The problems involved in characterising thearchaeological assemblage of level i can be summed

Fig. 3. Singular tripod of wooded imprints from level i.

up with the concept of ‘‘living floor’’ (Bordes,1975). We propose an empirical approach based onthe concept of accumulation: an area where thearchaeological assemblage formed by naturaland anthropic processes can be easily delimitedby means of visual observation (Villa, 1988)(Fig. 4).

Archaeological remains are spatially clusteredin well-delimited accumulations that are generallyonly a few centimetres thick. They can be easilydefined as spatial analytical units, especially whenthey are related to hearths (Vaquero and Pasto,2001). Each accumulation is referred to withArabic numbers (Table 1 and Fig. 5). Our archaeo-stratigraphic research, based on the vertical plot ofthe remains larger than 2 cm and the depth of thearchaeological structures such as hearths andtravertine blocks (Canals, 1993), indicated thepresence of stratified and non-stratified accumu-lations (Carbonell, 2002). The stratified accumu-lations have a clear diachronous archaeologicalrecord, whereas the non-stratified accumulationsonly show one horizon of remains. Stratifiedaccumulations are subdivided using a letter (Table1). Despite the high resolution of the sedimentaryprocesses and their chronostratigraphy (Bischoffet al., 1988), and because of the lack of corre-spondence between these processes and the timespan of the human occupations, the stratified andnon-stratified accumulations should be consideredas occupational palimpsests or a long-term ex-posed surface, in the sense of a geological palimp-sest (Riba, 1997).

Accumulations and natural formation processesArchaeological materials from the bottom of

level i are located in accumulation 12c, whichconsists of bone remains recovered in gourdeposits (L-K/45-46 squares). In this accumulation3 MNI of cervid without stone tools have beenidentified. Taphonomic analysis of these red deerbones show that this assemblage is redeposited(Caceres, 1995). A bone refit between accumu-lations 12c and 11c has been documented. Otherarchaeological materials inside gours have beenrecognized in accumulation 13b.

The archaeostratigraphy of level i is betterdefined in accumulation 11 (Carbonell, 2002).

164 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Fig. 4. Accumulations observed on level i. Legend. A, travertine cliff of the shelter. B, gour. C, travertine block. D, hearth. E, wood

imprint. F, boundary of diffuse accumulation. G, ancient archaeological surveys.

During the fieldwork, we recognised two phases ofcultural formation (11b and 11c), with two hearths(III and IV) separated by a sterile layer. We alsoidentified a third accumulation (11a) by means ofan archaeostratigraphical analysis. This accumu-lation is made up of a senile equid tooth, a jawfragment of bovid and a wood imprint. Thestratigraphical relations between these archaeolog-ical materials enable us to characterise the mainnatural formation processes of level i. From thebase to the top these are as follows:

The senile equid tooth from this accumulation11a can be related to another isolated toothlocated in square M49 (accumulation 13b). This

skeletal element was the key in determining theonly MNI of senile equid. It was found in the baseof a succession of gour facies.

The bovid mandible from accumulation 11asuggests a relationship with other bovid elementsfrom accumulation 14c near hearth IX, whichdetermine the only MNI of bovid in this level.

The natural processes that formed the woodimprints from accumulation 11a show a relationwith the wood imprints from accumulations 6, 14band 12a.

Accumulation 11a is formed of archaeologicalremains that show a spatial relation with otheraccumulations attributed to different sedimentary

FAUNA REMAINS

RTZITE

2 3 1 2 3 4 5 6

- - E-sE-C C-AX H - - -

- - E C-AP H 1E A -

- - E C H 1E A E?

- - E C-AP L - - -

- - sE-C AP L - - -

- - E C H - - -

- - E C L - - -

- - E-C C-AP L - - -

- - C AP L 1E A -

- - E C H - - I

- - E-C-B C-AP H 1E A E?

I L C C-AP H 1E sA -

- - E-C C-AP L - - E

- - E-C AX-AP - 1C sA -

- - E-sE-C C-AX-AP L 3C 2A-1sA -

- - C C-AX-AP H 1E A E

I L E-C C-AP L 1E S -

- - E C - - - E?

- - E-C C-AX-AP L - - -

- - E-C C-AX-AP H 1E A -

I L E-C-B C-AX-AP L 1B - E?

A: absence. P: presence. Lithic remains legend. UW A:

concentrated (O10 remains m�2); D, dispersed (2-10

); L, large (O5.5 cm).

ts: C, cranial; AX, axial. AP, apendicular. 3, number of

, subadult; S, senile. 6, faunal refittings: E, external; I,

165

J.Vallverd

uet

al./JournalofHumanEvolution

48(2005)157e

174

Table 1

Synthetic table of archaeological components observed and analysed in level i accumulations

LITHICS REMAINS

FLINT QUARTZ LIMESTONE QUA

H IOT WI UW A 1 2 3 1 2 3 1 2 3 1

ACCUMULATIONS

NON-BEDDED

1 I A A M I C S - - - - - - -

2 A A A - - - - - - - - D M-L -

3 II A P W - - - - - - E I M-L -

4 VIII P P - - I M-L - I M-L E I M-L -

5 X P A - I I S - D S-M I D M-L -

6 XI A P - - D S-M - I S I-E D S-M-L -

7 A P P W - I M - I S I D M-L -

8 XIII P A M I C S-M I C S - I M -

9 XV A A - - D M-L - D S - I M -

10 XVI P A M - D S-M - D S-M - D M-L -

BEDDED

11a A A A M - - - - I S - I M-L -

11b III P P W I C M-L - I S I D M-L -

11c IV P A - I D M-L - - - - - - -

12a VII A A - - D M-L - D S-M - - - -

12b V P P W E I M-L - C S-M - - - -

12c A A A - - - - - - - - - - -

13a VI P P - - - - - I S - I M-L i

13b A A A - - - - - - - - - - -

14a A A A - - - - - - - - I M-L -

14b XIV P P W-M I-E C S-M-L - I M-L I-E C S-M-L -

14c IX A A - - - - - I S-M - I M-L -

General legend. H: hearth, with a roman number. IOT: Internal Organization of Travertines. WI, wood imprints.

usewear analyses; W, wood; M, meat. 1, lithic refitting; I, internal; E, external. 2, density of lithic remains: C,

remains m�2); I, isolated (1 remains m�2). 3, format of lithic remains: S, small (!3 cm); M, medium (3-5.5 cm

Faunal remains legend. 1, Specific determination: E, equid; sE, senile equid; C, cervid; B, bovid. 2, Skeletal elemen

burnt remains: H, high; L, low. 4, location of INM and number: E, equid, C, cervid, B, bovid. 5, Age: A, adult; sA

internal.

166 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Fig. 5. Roman number for hearths and Arabic number for accumulations for the spatial archaeological record on level i.

layers. The bovid remains from accumulation 14care on the top of gour deposits. The stratigraphicposition of the equid tooth, however, is 20 cmbelow hearth VI, in accumulation 13b. Thus, thebovid remains are later than the dental remains ofthe senile Equus. These archaeostratigraphic rela-tions show that accumulation 11a is polygeneticwith redeposited and displaced remains.

We also identified disturbed materials in accu-mulation 14a. They are separated from the loweraccumulation 14b by a layer of silty gour sedi-ments. These remains show orthogonal orientationamongst them, which suggests displacement bystream action (Isaac, 1989).

In conclusion,we have identified twomoments ofnatural formation processes of the archaeological

assemblages. The first one is in the base of level i, inaccumulation 12c, and can be characterised by theredeposition of archaeological remains. Thesearchaeological materials are inside gours with littlehorizontal displacement. The second stratigraphiccontext is at the top of the level, where naturalformation processes like re-sedimentation madeaccumulations 11a and 14a. These accumulationsalso consist of bone and vegetal remains (charcoalsand wood imprints). It is also worth pointing outthat they have a low density of archaeologicalmaterials.

Finally, lithic refits are not located in theseaccumulations, which are formed by natural pro-cesses. Some bone refits, however, have beenidentified, such as the ones in 11a and 13b.

167J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Accumulations and anthropic formation processesThe evidence of accumulations formed by

natural processes allows us to consider theexistence of accumulations through anthropicprocesses. Cultural accumulations can be charac-terised by the same archaeostratigraphic andlithological context with a stratigraphic connectionbetween artifacts and archaeological structuressuch as hearths and travertine blocs.

On the basis of the presence of refits, we havedistinguished two kinds of accumulations: onesthat are related, characterised by connectionsbetween different accumulations, and ones thatare unrelated, characterised by the absence ofconnections with other clusters, although somerefits have been identified within them.

Some accumulations cannot be properly char-acterised, because of the ancient excavations.Therefore, we do not take into account theaccumulations formed by the archaeological re-cord around hearths VII, XII and XV. Theseaccumulations make it very difficult to makea comparative characterisation.

Non-related anthropic accumulationsAccumulation number 1 contains hearth I and

a dense concentration of flint. In this accumulationsome faunal remains from equid, small equid, andcervid have also been identified. The use wear ofa flint flake from this accumulation shows that ithad been used in meat processing.

The archaeological record from accumulation11b is very specific. There are cervid remains, withno axial parts. It also contains hearth III and anInternal Organization of Travertine (IOT). Theflint and limestone remains, connected by refits,form a dense lithic concentration over the hearth.Some of these lithic remains present a thermalimpact. The functional analysis of one of the flintflakes from this accumulation shows traces ofwoodwork. Some internal refits have been foundon the periphery of this accumulation, close toaccumulation 5.

Accumulation 11c appears around hearth IVand contains a high diversity of archaeologicalremains. The lithic assemblage consists of flint andquartz. The faunal remains correspond to cervidsand equids.

Accumulation 13a consists of the archaeologi-cal remains located around hearth VI, and it is notclearly delimited with respect to accumulation 14caround hearth IX. These two accumulationspresented a high diversity of faunal remains ofequids, cervids and bovids. Moreover, the lithicassemblage is characterised by the absence of flintand the presence of a quartzite refitting. There areno refits between accumulations 13a and 14.

Accumulation 8 includes hearth XIII, an In-ternal Organization of Travertine, and lithic andfaunal remains. It should be pointed out that thereare few bone remains, and that these belong to anequid cranium. The lithic assemblage around thehearth consists of quartz, flint and limestone.

Accumulation 5, with hearth X, has very fewlithic and bone remains. Three cervid tibiae anda small equid tibia make up the bone record of thisaccumulation. The lithic assemblage consists ofquartz, flint and limestone. This accumulation hasrefittings near the travertine blocks in squaresM-N58-59, between accumulations 5 and 7.

Accumulation 7 has no hearth, and has aninternal limestone refitting. This accumulation ispartly limited by the above mentioned travertineblocks. One of the lithic remains has traces ofwoodworking.

Related anthropic accumulationsLithic and bone refitting between accumula-

tions suggest that there is a connection between thenon-stratified accumulations 3-4-6 and 10 and thestratified ones 12b-14b. These related accumula-tions have a high degree of material diversity. Themain raw materials are flint and limestone, fol-lowed by quartz. The faunal remains are domi-nated by equids.

Accumulations 3-4 and 6 are not clearly charac-terized because of the old archaeological exca-vations. Accumulation 3, with hearth II, hasdispersed lithic and bone remains, including anequid molar with a singular modification (Fig. 6).This same modification has also been documentedin another equid molar found at accumulation 14b,near hearth XIV. Accumulation 3 contains largewooded imprints (WI) that belong to a tree trunk.Accumulation 4 has more bone remains and fewlithic remains, but there is a lithic refitting between

168 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

accumulations 3 and 4. Accumulation 6 showsa high degree of burnt bones and near hearth XIthere is also a group of large wood imprints.

The related accumulation 14b has the greatestdensity of archaeological material. These archaeo-logical remains, around hearth XIV, are located infront of the stalagmite that delimits the entrance tothe Coveta Ripoll. This accumulation has variousinternal and external refittings of flint and lime-stone that show connections with accumulations12b, 10 and 6. It is also characterised by small lithicand bone remains, as well as small WI fragments.In this accumulation, there are a large number ofequid cranial fragments, and also a few axial cervidbones. It also contains many burned bone remains.Additionally, the functional analysis of a flint flakeshowed that it had been used on wood and meat.

In accumulation 10, the faunal assemblageconsists of refitted equid dental and mandibleremains. The use-wear analyses indicate that a flintflake had been used in meat processing.

Accumulation 12b includes hearth V and itsarchaeological record consists of a few flint andlimestone remains, and abundant cervid and equidfaunal remains. There are also many dispersed WI,although there is a group of three lying togetherover hearth V, which suggests the analogy witha tripod mentioned earlier (Carbonell and Castro-Curiel, 1995).

Fig. 6. Two horse molars with deep incisions on their surface

from accumulations 3 and 14b. These horse teeth are from the

same individual. They were 8 meters apart. The marks on their

surface are currently being studied.

Discussion

Level i consists of accumulations formed bynatural and cultural processes. The record of theaccumulations formed by anthropic processes islocated basically around archaeological structures,such as hearths and Internal Organization ofTravertines. The record of the cultural accumu-lations shows two different sedimentary momentsfor human occupation, which had already beendocumented during the fieldwork and verified by thearchaeostratigraphic and lithic study (Carbonell,2002).

The lithic and faunal refitting made it possibleto distinguish between related and unrelatedaccumulations. The lithic and faunal compositionof some accumulations is often highly specific.This is due to sedimentary processes limiting theformation of developed occupational palimpsests,in the framework of a centennial-millennial timespan. Also, the dripping sedimentary processes erelated to the above mentioned wetter facies emay have limited the suitability of the rockshelter for human occupation (Vallverdu et al., inpress).

The main activities identified in these accumu-lations are the following (Table 2):

1. Knapping activities, as well as the consump-tion and abandonment of faunal remains.These activities have been identified at thenon-related accumulations 1, 2, and 8 and 13aand 14c and the related accumulations 2, 12b,4, 6, 14b and 10.

2. Butchering activities. These activities havebeen identified at accumulations 11b, 14b and10. In the first, there is evidence of thebutchering of a cervid that shows no axialskeleton. Butchering activities at the relatedaccumulations 14b and 10 e near the CovetaRipoll e have been linked to the abandonmentof equid cranial elements.

3. Animal biomass conservation. These activitiesmay have taken place at accumulation 12b.This accumulation contains the tripod onhearth V (Carbonell and Castro-Curiel, 1995).The interpretation of the meat conservationactivities is based on the large longitudinal

169J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Table 2

Main anthropic accumulations, estimated surface (in m2) and human activities

SURFACE

(m2)

ACCUMULATION ARCHEOLOGICAL

MATERIAL

ACTIVITIES

6 7 Wood imprint. Vegetal exploitation and partial consumption.

Flint, quartz, limestone. Lithic transport and exploitation.

Equid. Animal transport and consumption.

9 1 Hearth I. Vegetal exploitation and consumption.

Flint. Lithic transport and exploitation.

Equid, cervid. Animal transport and consumption.

9 9 Hearth XV. Vegetal exploitation and consumption.

Flint, quartz, limestone. Lithic transport and exploitation.

Cervid. Animal transport and consumption.

12 5 Hearth X. Vegetal exploitation and consumption.

Flint, quartz, limestone. Lithic transport and exploitation.

Cervid, equid. Animal transport and consumption.

12 8 Hearth XIII. Vegetal exploitation and consumption.

Flint, quartz, limestone. Lithic transport and exploitation.

Equid. Animal transport and consumption.

16 11C Hearth IV. Vegetal exploitation and consumption.

Flint, quartz. Lithic transport and exploitation.

Cervid, equid. Animal transport and consumption.

16 13AC 14C Hearth VI, IX. Vegetal exploitation and consumption.

Quartz, limestone, quartzite. Lithic transport and exploitation.

Cervid, equid, bovid. Animal transport and consumption.

20 11B Hearth III. Vegetal exploitation and consumption.

Flint, quartz, limestone, Lithic exploitation and supply.

Quartzite. Cervid. Animal butchering and consumption.

140 3C 4C 6C 10C

2BC 14B

Hearth II, V, VIII, XI, XIV, XVI Vegetal exploitation and partial consumption.

Wood imprints Lithic exploitation and supply.

Flint, quartz, limestone. Animal butchering.

Equid, cervid, bovid. Immediate and differed consumption.

cut marks mentioned above, which show thatmeat strips may have been cut.

4. Exploitation and consumption of vegetalresources. These activities are recorded inalmost all the accumulations on level i. Thereare also other areas with vegetal remainsunrelated to fire, such as the small WIfragments at accumulation 14b. Traces ofwoodworking were also identified throughtraceological analyses at accumulations 12b,11b, 7 and 3.

Occupational models at level i

The size of the accumulations formed byanthropic activities and the analysis of the

different categories of the record have made itpossible to identify two occupation models at theshelter in level i. The first has a small surfacearea while the second is medium sized. Bothmodels are based on the distinction between therelated and unrelated accumulations (Figs. 7and 8).

Occupation model of the small surfaceThe small model can be characterised by the

occupation of a surface between 9 and 20 squaremeters (Table 2). The superimposition of lithic andfaunal remains around or over the hearthscharacterises this model of occupation (Vaqueroand Pasto, 2001). The lithic activities identified inthe technological study are the introduction of

170 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Fig. 7. Spatial location for the hearths and internal refitted archaeological material from the small model.

large flakes and retouched artefacts, partiallyreduced cores and limestone blocks.

This model shows two types of animal resourceprovisioning: in-site processing and consumption,or transport and consumption of animal partsprocessed outside the site.

Occupation model on the medium-sized surfaceThis occupation model occupies 120 square

meters (Table 2). The various materials aredistributed differently on the surface. The faunalremains are concentrated towards the exterior ofthe area, at accumulations 4 and 12a. The lithicremains tend to cluster at the accumulations nearto the wall, where the small faunal and lithicremains overlap, just as they do in the small

occupation model. There are knapping sequences,the introduction of limestone cobbles, and largeflakes and retouched objects. Faunal remains showthat animal biomass was exploited, processed,conserved and consumed at the site.

The fragmentation of the medium-sized occu-pation models, with a central domestic area anda specialized peripheral area, has been described byethnoarchaeologists (Yellen, 1977; Binford, 1987).No secondary accumulation related to intentionalrefuse has been described. The overlapping ofstructures (hearths and IOT) and archaeologicalremains, as well as the clearly delimited accumu-lations, suggest the absence of this type of deposits.

The identification of a domestic zone andspecialised areas, in the medium-sized occupation

171J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Fig. 8. Spatial location of hearths and external and internal refitted archaeological material from accumulations in the medium-sized

model.

model, indicates an occupation pattern that wasmore common during the Upper Paleolithic(Simek, 1987). So, on level i the occupation modelvaries according to the size of the occupied surface.

Exploitation models at the Abric Romanıand territorial mobility

The provisioning strategies identified showa threshold of minimal activity, related to theexploitation of the vegetal materials from thesurroundings of the site for fuel. This indicatesthat fire technology was widespread during thisperiod (Binford, 1988) and is present in almost all

the accumulations of human formation processes(Arteaga et al., 2001). The combustion activitiesaround the hearths show a specific combustion ofpinewood (Allue, 2002).

The restricted provisioning strategy is charac-terised by highly fragmented fauna and lithicoperatory sequences, and suggests the existenceof a previous provisioning site. The Abric Romaniwas basically a place of passage (Bibouac). Therestricted strategy of lithic and faunal resourcesshows that the immediate surroundings were notexploited (Table 3).

The diversified strategy documents the exploita-tion of lithic and faunal resources from the shelter’s

172 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Table 3

Main human activities according to the kind of resources and the occupation models

MODELS VEGETAL RESOURCES ANIMAL RESOURCES LITHIC RESOURCES

Small sized extension

and restricted provision

Combustion and configuration Final consumption Introduction of tools and final

knapping sequences

Small sized extension

and diversified provision

Combustion and configuration Zone of processing and final

consumption

Final knapping sequences and

introduction raw material

Middle sized extension

and diversified provision

Combustion and configuration Specialized zone of

processing and final

consumption

Introduction of tools, Final and

initial knapping sequences and

introduction of raw material

immediate environment. This diversified exploita-tion is recorded through the input and output oflithic elements, the primary or secondary pro-visioning strategy and a possible deferred treatmentof animal and lithic resources. The selection of

precise elements for transport and consumption atthe site has been designed as a strategy ofprovisioning individuals (Khun, 1995).

The exploitation models on level i show thatthe groups were highly mobile. The diversified

Fig. 9. Density of burned fauna per square metre from level i.

173J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

provisioning strategy probably complements therestricted one. This is based on accumulation 11b(Table 3), which is small and has a diversifiedexploitation. In these accumulations large quanti-ties of burned bones coincide with diversifiedprovisioning strategies (Fig. 9).

It is also possible to suggest the absence ofprevious occupation sites characterised by a di-versified exploitation and a deferred consumptionof resources. In this way, regardless of thedemography or size of the occupied surface, themobility of the group is directly related tothe exploitation of hunting strategic resources.The linear mobility between diverse geographicalareas and low segregation suggest the planning oflong movements with the abandonment of easilyreplaceable objects and the transport of surplus.

Acknowledgements

The Generalitat de Catalunya, Diputacio deBarcelona, Ajuntament de Capellades and theRomanya-Valls Company provided financial sup-port for the research and fieldwork in the AbricRomani. The Fundacion Atapuerca supports theresearch of E. A., R. H. and J. V.; P. S. has a grantprovided by the Fundacion Duques de Soria.

Bibliography

Allue, E., 2002. Dinamica de la vegetacion y explotacion del

combustible lenosos durante el Pleistoceno Superior y el

Holoceno del Noreste de la Penınsula Iberica a partir del

analisis antracologico. Ph.D. Dissertation, Universitat

Rovira i Virgili.

Arteaga, I., Allue, E., Pasto, I., Vallverdu, J., Carbonell, E.,

2001. Els fogars del paleolıtic mitja de l’Abric Romanı

(Capellades, Anoia). Cypsela 13, 11e27.

Bartroli, R., Cebria, R., Muro, I., Riu, E. y Vaquero, M., 1995.

A frec de ciencia. L’Atles d’Amador Romani i Guerra.

Ajuntament de Capellades, Capellades. 232 pp.

Binford, 1973. Interassemblage Variability -the Mousterian

and the functional argument. In: Renfrew, C. (Ed.), The

Explanation of Culture Change: Models in Prehistory.

Duckworth, London, pp. 225e254.

Binford, L.R., 1988. En busca del pasado. Descifrando el

registro arqueologico. Crıtica, Barcelona.

Binford, L.R., 1987. Researching ambiguity: frames of refer-

ence and site structure. In: Kent, S. (Ed.), Method and

Theory for Activity Area Research. Columbia University

Press, New York, pp. 449e512.

Bischoff, J.L., Fitzpatrick, J.A., 1991. U-series dating of impure

carbonates: An isochron technique using total-sample

dissolution. Geochimica et Cosmochimica Acta 55,

543e554.

Bischoff, J.L., Julia, R., Mora, R., 1988. Uranium-series dating

of the Mousterian Occupation at Abric Romani, Spain.

Nature 332, 68e70.

Bischoff, J.L., Ludwing, K., Garcia, J.F., Carbonell, E.,

Vaquero, M., Stafford, T.W., 1994. Dating of the basal

Aurignacian Sandwich at Abric Romani (Catalunya, Spain)

by Radiocarbon and Uranium-series. Journal of Archaeo-

logical Science 21, 541e551.Bordes, F., 1975. Sur la notion de sol d’habitat en prehistoire

paleolithique. Bulletin de la Societe Prehistorique Francaise

72, 139e144.

Bordes, F., 1981. Vingt-cinq ans apres: le complexe mousterien

revisite. Bulletin de la Societe Prehistorique Francaise 78,

77e87.

Bunn, H.T., 1983. Comparative analyses of modern bone

assemblages from a San Hunter-Gatherer camp in the

Kalahari desert, Botswana, and from a Spotted Hyena den

near Nairobi Kenia. In: Clutton-Brock, J., Grigson, C.

(Eds.), Animals and Archaeology. 1, Hunters and their

Prey. pp. 143e148. Oxford: British Archaeologica Reports

International Series 163.

Burjachs, F., Julia, R., 1994. Abrupt climatic changes during the

last glaciation based on pollen analysis of the Abric Romani,

Catalonia, Spain. Quaternary Research 42, 308e315.

Caceres, I., 1995. Estudio Tafonomico de los Procesos de

Formacion del n.(i) del Abric Romanı (Capellades, Barce-

lona). La influencia de la actividad antropica. M.A.

Dissertation, Universitat Rovira i Virgili.

Canals, A., 1993. Methode et techniques archeostratigraphi-

ques pour l’etude des gisements archeologiques en sediment

homogene: application au complexe CIII de la Grotte de

Lazaret, Nice (Alpes Maritimes). Ph.D. Dissertation,

Museum d’Histoire Naturelle.

Carbonell, E., 1992. Abric Romanı, nivell H: un model

d’estrategia ocupacional al Plistoce Superior mediterrani.

Estrat 5, pp. 159e308.

Carbonell, E., 2002. Abric Romanı nivell i. Models d’ocupacio

de curta durar de fa 46.000 anys a la Cinglera del Capello

(Capellades, Anoia, Barcelona). Universitat Rovira i Virgili

Press, Tarragona.

Carbonell, E., Castro-Curiel, Z., 1995. Wood pseudomorphs

from level i at Abric Romani (Barcelona, Spain). Journal of

Field Archaeology 22, 376e384.

Carbonell, E., Cebria, A., Allue, E., Caceres, I., Castro, Z.,

Diaz, R., Esteban, M., Olle, A., Pasto, I., Rodriguez, X.P.,

Rosell, J., Sala, R., Vallverdu, J., Vaquero, M., Verges,

J.M., 1996a. Behavioural and organizational complexity in

the Middle Palaeolithic from The Abric Romani. In:

Carbonell, E., Vaquero, M. (Eds.), The Last Neandertals.

The First Anatomically Modern Humans. Universitat

Rovira i Virgili Press, Tarragona, pp. 385e434.

174 J. Vallverdu et al. / Journal of Human Evolution 48 (2005) 157e174

Carbonell, E., Cebria, A., Rosell, J., Sala, R., Vaquero, M.,

1996b. Els conjunts III i IV de l’Abric Romanı de

Capellades (campanes 1988-1994). Una sequencia d’ocupa-

cions del paleolıtic mitja. Tribuna d’Arqueologia 1994-1995,

27e38.

Carbonell, E., Giralt, S., Vaquero, M., 1994. Abric Romanı

(Capellades, Barcelone, Espagne): une importante sequence

anthropisee du Pleistocene Superieur. Bulletin de la Societe

Prehistorique Francaise 91 (1), 47e55.

Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D.,

Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffen-

sen, J.P., Svelnbjornsdottir, A.E., Jouzel, J., Bond, G., 1993.

Evidence for general instability of past climate from 250-kyr

core record. Nature 364, 218e220.Folch, R., 1981. La vegetacio dels paısos catalans. Ketres,

Barcelona.

Giralt, S., Julia, R., 1996. The sedimentary record of the

Middle-Upper Palaeolithic transition in the Capellades area

(NE Spain). In: Carbonell, E., Vaquero, M. (Eds.), The Last

Neandertals. The First Anatomically Modern Humans.

Universitat Rovira i Virgili Press, Tarragona, pp. 365e376.

Isaac, G., 1989. Towards the interpretation of occupation

debris: some experiments and observations. In: Isaac, B.

(Ed.), The Archaeology of Human Origins. Cambridge

University Press, Cambridge, pp. 191e205.Khun, S.L., 1995. Mousterian Lithic Technology, An Ecolog-

ical Perspective. Princenton University Press, Princenton.

Meignen, L., 1994. L’analyse de l’organisation spatiale dans les

sites du paleolithique moyen: structures evidentes, structures

latentes. Prehistoire et Anthropologie Mediterraneennes 3,

7e23.

Riba, O. (Dir.), 1997. Diccionari de Geologia. Enciclopedia

Catalana. Barcelona.

Simek, J.F., 1987. Spatial order and behavioural change in the

French Palaeolithic. Antiquity 61, 25e40.

Vallverdu, J. Vaquero, M., Courty, M.A., Carbonell, E., in

press. Procesos sedimentarios rapidos y ocupaciones hu-

manas poco frecuentes. La seccion del Pleistoceno Superior

de la coveta nord del Abric Romanı (Capellades, comarca

de l’Anoia, Barcelona). IV Reunion Nacional de Geo-

arqueologia. Almazan. Spain.

Vaquero, M., Esteban, M., Allue, E., Carbonell, E., Bischoff,

J.L., 2002. Middle Palaeolithic Refugium, or Archaeological

Misconception? A New U-series and radiocarbon chronol-

ogy of Abric Agut (Capellades, Spain). Journal of

Archaeological Science 29, 953e958.

Vaquero, M., Pasto, I., 2001. The definition spatial units in

Middle Paleolithic sites: the hearth-related assemblages.

Journal of Archaeological Science 28, 1209e1220.Villa, P., 1988. A propos d’analyse spatiale et de taphonomie.

Les nouvelles de l’archeologie 31 (Printemps 1988), 55e59.

Yellen, J.E., 1977. Archaeological Approaches to the Present:

Models for Reconstructing the Past. Academic Press, New

York.