Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/248725448

TowardstheReconstructionofMagdalenianArtisticTechniques:theContributionofMicroscopicAnalysisofMobiliaryArt

ARTICLEinCAMBRIDGEARCHAEOLOGICALJOURNAL·OCTOBER1999

DOI:10.1017/S0959774300015377

CITATIONS

4

READS

132

1AUTHOR:

CaroleFritz

FrenchNationalCentreforScientificResearch

44PUBLICATIONS69CITATIONS

SEEPROFILE

Availablefrom:CaroleFritz

Retrievedon:17November2015

Cambridge Archaeological Journalhttp://journals.cambridge.org/CAJ

Additional services for Cambridge Archaeological Journal:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

Towards the Reconstruction of Magdalenian Artistic Techniques: the Contribution of MicroscopicAnalysis of Mobiliary Art

Carole Fritz

Cambridge Archaeological Journal / Volume 9 / Issue 02 / October 1999, pp 189 - 208DOI: 10.1017/S0959774300015377, Published online: 14 October 2009

Link to this article: http://journals.cambridge.org/abstract_S0959774300015377

How to cite this article:Carole Fritz (1999). Towards the Reconstruction of Magdalenian Artistic Techniques: the Contribution of Microscopic Analysis of Mobiliary Art. CambridgeArchaeological Journal, 9, pp 189-208 doi:10.1017/S0959774300015377

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/CAJ, IP address: 123.183.211.9 on 11 Dec 2013

Cambridge Archaeological Journal 9:2 (1999), pp. 189-208

Towards the Reconstruction of Magdalenian ArtisticTechniques: the Contribution of Microscopic Analysis of

Mobiliary Art

Carole Fritz

On the assumption that Magdalenian mobiliary art is the bearer of certain technicaltraditions which express cognitive and cultural values, research is directed towards theidentification and appreciation of these values through the understanding of techniquesinvolved in bone engraving. Examination of a set of 90 objects from sites in southwestFrance sheds light on repeated artistic activity and graphic 'recipes' which do not appearto be characteristic of local or regional groups. In the sequence of actions as well as in theirfinal product, there is no activity specific to the Ariege as opposed to the Gironde.

Magdalenian artists have identical shared artistic procedures.

the past twenty years, the results obtained fromtechnological study of flint tool production have

shown that an understanding of entire phases ofPrehistoric daily life can be obtained through in-

ePth analysis of modes of production and the re-construction of chaines operatoires. Moreover, the^chlogical levels thus defined, and their practice,

us to identify genuine collective knowledgeto characterize a cultural group or a part of such

a group (Pigeot 1987; Pelegrin et al 1988).Studies of mobiliary art follow this same path,vi a mixture of methodology and concept,ve contributions from this line of research were

197*Seen i n t h e 1 9 7 0 s (Delporte & Mons 1973; 1975;76) but there has been a renewal of interest as a

esult of the introduction of new analytical tools1Q7 aS/ ^ r s t ' ^ e binocular microscope (Marshack^72,415; Cremades 1989) and then, during the 1990s,i QG USe °^ *^e scanning electron microscope (d'Erricoiy88; 1994; Fritz et al 1993).

The time has come to consider objects of mobil-.fr*\ a r t (irrespective of their appearance and aes-

etics) as technical products which bear traces ofe procedures involved in their manufacture. A tech-

lclUe can be defined as a group of methods wherebyn end is reached. It can be defined as 'actions and°ols organized as a chain with true syntax' (Leroi-

Gourhan 1965). Depending on its complexity, oneshould picture the finished product as a point ofconvergence of chaines operatoires of differing lengthand variety.

Adapting the approaches and vocabulary of theethnologist to prehistoric archaeology, lithic special-ists have taught us how to understand technique inorder to decipher the social and cognitive codeswhich it conceals (Cresswell 1983; Lemonnier 1983;Pigeot 1987; 1988; 1991; Pelegrin et al 1988). Initiallyone may be troubled by the use of terms such ashardness, solidity, suitability, and transformation todescribe objects 'which we consider with venerationto be in the sanctuary of art' (Seris 1994). Neverthe-less, we should not overlook the fact that techniqueis the basis of all transformation from natural tofinished product, from the realization of a simpleoutline to the representation of themes, figurative orotherwise.

Within the field of art, technique acts at alllevels of transformation as much in the fashioning ofa support as in the production of the decoration. Todescribe the formation of the support, one can detailthe tools and procedures involved in preparing thesurface, in forming the outline and in giving lifeto the form. In the representational phase the expres-sion is simultaneously semiotic, iconographic,

189

Carole Fritz

cognitive and cultural, for it formalizes, fixes andmaterializes the rules, styles, conventions, perspec-tive and models without which there would be noresulting image.

The distinction between the shaping of the sup-port and the production of the decoration is purelytheoretical, for the finished decoration will dependlargely on how it is created; 'the aesthetic value of animage is not simply plastic, it depends on its techni-cal value' (Seris 1994). One should no longer con-sider the stylistic approach to be essential but ratherto be equivalent to, and indissociable from, the analy-sis of procedures of manufacture.

Making the assumption that mobiliary art isthe bearer of part of a tradition which can, in itself,express cultural values and cognitive characteristics,we have directed our research towards the identifi-cation of these criteria by way of understandingMagdalenian bone engraving techniques.1 By observ-ing the superimposition of marks on bone, we shallbring to the fore cognitive schemes which governthe construction of form. We shall try better to deter-mine the thoughts of Magdalenian engravers, ob-serving in particular whether their technique varieswith the figurative or geometric nature of the sub-ject. Finally, we shall consider the cultural compo-nent of the technique, looking for possible local andregional models.

The specific choice of bone material studied hasbeen dictated by its morphological and physical qual-ity. For reasons of its relatively homogeneous struc-ture and its plasticity, bone quite clearly preservesmarks left by tools used to cut it and tells us, indi-rectly, about the actions carried out. Beyond thisthere is another, more decisive argument: hard ani-mal material is the preferred choice for Upper Pal-aeolithic mobiliary art.2 Engravings are the mostwidespread means of expression; engraved bone hasemerged as the richest source of information bothtechnical and iconographic.

Methodology

The reading and analysis of these engravings weresubjected to numerous methods and observationaltechniques. This applied both to the experimentalpieces and the original works of art. The stages wereprogressive, the scale of magnification varying frommacroscopy (binocular microscope) to microscopy(scanning electron microscopy, SEM). An in-depthknowledge of the material also figured amongst theanalytical techniques; through understanding thephysical and chemical structure (thread of bone fi-

bre, structural homogeneity or heterogeneity) of thematerial used (bone and flint), it is possible to specifythe mechanical phenomena employed and to under-stand the formation of certain marks located at thebase of an incision.

Observational techniques

A binocular microscope may be used in the firstplace to select archaeological pieces at a macroscopiclevel. These were examined in order to determinethe state of preservation of the bone itself (flaked orwell-preserved surface, cracks, glue or castingresidues and traces of colouring) and that of theengravings. As will be seen the pieces retained formicroscopic analysis should be in the best possiblestate of preservation; not only because they have towithstand an impression before being examined byscanning electron microscopy, but also because ex-amination of the bottom of the groove by SEM at alater stage requires optimal preservation for a goodreading of the incisions. We must also insist on astrict code of ethics: whenever the taking of an im-pression appeared likely to damage the original piecewe did not proceed.

In this study, the Jeol 840 was used as the pre-ferred SEM, allowing high magnification and givinga depth of field 300 times greater than that of aclassic optimal microscope. It also has a high resolu-tion (200A) and therefore achieves an image of higherquality than do other observational tools.

Use of the SEM involves several preparatorystages owing to the fact that the sample, during itspreparation, is bombarded with electrons (d'Errico1988; 1994; Fritz el ah 1993). In the case of non-con-ductive material such as bone it is necessary to en-sure contact with the material in order to eliminatethe electrical charge carried by the electron beam;otherwise the received image would be blurred. Itis therefore necessary to cover the sample withconductive material. In the present case, since itwas technically possible, replicas were gilded, i.e.a very thin film of gold was placed on the surface ofthe resin. Problems of electrical charge and the needto gild samples mean that we are obliged to takeimpressions.3 We therefore have to work with rep-licas.

Taking an impression

In order to take impressions, i.e. the moulds whichare used to make replicas, a dental silicon4 (mixtureof base and catalyst) was used. This has several ad-

190

Reconstruction of Magdalenian Artistic Techniques

vantages. Combining the base and catalyst forms asingle mixture. We can thus control the time in-volved, which is generally very short (2-5 minutes).This speed is particularly useful as it allows immedi-ate control of impression quality, and repetition asnecessary.

Application to the sample is done without caus-ing damage, since in mixing the base and catalyst itls possible to control the fluidity of the mixture,which sets rapidly and reliably. The bone surfaceremains in contact with the silicon for a short timeonly. Nothing is applied to the object in order toremove it from the mould as this is already incorpo-rated into the product. Numerous preliminary, at-tempts were made, especially on experimental greenbone. Even at microscopic level, no modification ofthe surface conditions could be perceived; these re-sults are satisfactory and comply with the necessaryrequirements for the preservation of original objectsoutlined above.

The dental silicon is of better quality than ordi-nary silicon and provides precision in the order ofnycrometres and therefore access to observation atnigh magnification (with a resolution of 100 — 5micrometres). This precision and the fact that it doesn ° damage to the archaeological material makes it avaluable tool.

The replicas were made from the impressionth a polyurethane resin (Hexcel 538), which has

advantage of not creating too many air bubblesing polymerization; such bubbles would be trou-

during high-powered magnification.

Experiments

xPeriments in known and controlled conditions in-volved producing cut-marks in order to observe and

yze the performance of the material and to iden-^arks left by flint on the bone (Leguay 1877;•'* Cremades 1989; d'Errico 1991; Fritz et ah 1993).These analyses make it possible to answer the

question: What is a mark? The answer should beXplained both macro- and microscopically.

Macroscopic analysis defines the mark by con-. ering the visible result on the surface of the mate-

The cut-mark is an incision, a hollow depressionProduced with the help of a tool. The latter should® adapted to the hardness of the support which it°uld, of necessity, damage in some way: 'it can be

° varied nature: flint, antler, plant material, fingeron clay)' (Feruglio 1993, 267). The cut is composed

the several elements which define it: the edges,e s and base. During macroscopic analysis these

characteristics are described and comprise the tech-nological comparative collection.

At the microscopic level the components whichform the incision, mentioned above, are no longerthe most important. They become the supports ofmicrotraces which record a succession of events, theresults of which are observed on the bone surfacefrom the moment when the tool first makes contactwith the surface to the moment when contact ceases.We are now able to consider the engraved mark as acombination of technical actions. As a result of theimage obtained with the SEM each of these dynamic'events' which occurred during engraving has beenobserved and described.

The experiments, and SEM observations of themarks obtained, provide varied information and de-fining criteria for each of the parameters which wehave controlled (direction of cut, number of timesthe tool was used) (d'Errico 1988; 1994). In order toexplain these results it has been necessary to createspecific terminology for the technological marks re-vealed within the incision.

In the following micrographs (Figs. 1 & 2) com-parable marks of different origin are compared: onthe left the experimental pieces, on the right thearchaeological. The former are in a fresh state andare easily readable. The latter, which are slightlymore eroded, require more careful observation.

The attcique de trait (ATT or point of impact)(Fig. la) occurs at the point where the tool makes aninitial cut on the material (d'Errico 1988; 1994; Fritzet ah 1993). In the case of multiple passes or incisionsseveral points of impact can be identified. In fact,successive ATTs are distinct and microstriationswhich correspond to a specific blow damage theprevious ones, leaving a small gap between them,causing a more or less visible break in the line. Bycounting the different ATTs, it is possible to deter-mine the number of times an implement has cut intothe same groove.

The point marking the end of the groove (labutee defin de trait (BFT)) (Fig. 1b) refers, on the otherhand, to the point at which the tool finishes itsprogress along the line, identified as a small ridge ofbone. All along its course the tool cuts into the boneleaving behind chips of variable size. The larger onesare removed from the surface and the smaller arepushed along in front of the tool, and end up form-ing the BFT. Such marks also allow one to countsuccessive tool movements in the same groove. Theirexact frequency often remains uncertain on archaeo-logical material. Nevertheless, experiments haveshown that it is very difficult to make a deep cut at a

191

Carole Fritz

Figure 1. a) Attaques de trait (ATT or point of impact): arrows indicate the number of impacts, b) Butees de fin detrait (BFT or point marking the end of the groove). Two BFTs can be clearly seen on the micrograph on the left andthree on that to the right, c) Stigmates de direction (marks of direction). This experimental photograph showsparticular marks left by the removed chips; the base of the line is 'flaked'. On the micrograph of the archaeological piecethe surface is more weathered.

192

Reconstruction of Magdalenian Artistic Techniques

single attempt; in order to achieve a readable cut (i.e.one visible to the naked eye), in most cases two orthree strokes are necessary. As a result, we knew inadvance that several strokes would have to be lookedfor in the same groove. It only remained to countthem precisely.

Apart from the ATTs and BFTs, it is possible toestimate the number of times a tool is used by meansof the micro-ridges formed on the edges of the lines.These 'ridges' look like a succession of ripples ofmaterial pushed to the side each time the tool isused. It is certain that not all movements are identifi-able, for one ridge can hide another; however, bycorrelating the number of ATTs, BFTs and ridges,the frequency of tool movements can be determinedwith minimal error.

Taking into consideration both the ATTs andBFTs, one can logically deduce the direction of thelines. Where earlier marks are absent (because of anobject breaking or owing to poor preservation), thestigmates de direction (marks of direction) (Fig. lc) tellus the direction of execution of the engraving, ormore exactly the direction of the last cut. When thegrooves are slightly misaligned it is possible to de-termine the direction of each cut.5 The directionalmarks, visible at the bottom of the groove or on theridges, are microreliefs with chips which have beenonly partially removed and which indicate the pointof resistance of the bone, the thickest part of theridge facing the direction of movement of the tool.

Morphologically similar to the above, stigmatesde changement de pression (marks resulting from chang-ing pressure) indicate, by the same characteristics,the direction of the line. A change of pressure ap-pears in two contrasting forms: reduction in pressure(of the hand on the tool), where a micro-accumulationof chippings is visible along the whole section of thegroove; or a small cut. These marks are most oftenlocated on the ATTs where one observes increasedpressure at the moment of burin impact followed bya reduction in pressure which corresponds to minorchanges in the execution of the line.

Examination of the dynamics of the line throughthe intermediary of the stigmates d'accident (acciden-tal marks) (Fig. 2a), provides a way of determiningwhether or not the engraver's hand was steady. Ac-cidents consist of repeated malformation in the courseof the line. They are indicated by the removal ofmaterial at the edge of the lines, by small perpen-dicular incisions and most often by scratches, simi-lar to small waves, which are close together and atright angles to the direction of the groove (Fig. 2b).In the majority of cases they are interpreted as the

result of a bad tool angle or its poor position in thehand. Nevertheless, the resistance of materials mustalso be born in mind and we note recurring scratcheswhen the tool makes perpendicular contact with bonefibres. Between the number and type of accidents onthe one hand and the dimensions of the graphic fieldon the other, there exists a correlation which we cansummarize as: the frequency of accidents increasesas space decreases. In other words, it is difficult,even for a skilled engraver, to produce a miniaturedrawing.

Taking into account the dimensions of the sup-port and its characteristics (material, direction of fi-bres, etc.) it must be admitted that concentration andrecurrence of these indices, on one surface, definesthe knowledge of the engraver, their mastery of thetool and, in the final analysis, their skill.

We now know how to obtain precise informa-tion concerning the hand of the engraver. Such in-formation can also be obtained for the tool byexamining code-banes (bar-codes). The 'bar code' isthe identity card of the agent impressed into thematerial. At the scale of its crystalline structure, weknow that a burin bevel is not 'smooth' but that ithas a specific inclusion of fossils and grains of quartzgreater than that of chalcedony cement (Walter 1993,15). When the tool cuts the surface it produces linearmicrostriations, parallel to the edge of the groove(Fig. 3). These are caused by the microfossils and thequartz lepispheres, observable at a scale between 50and 10 microns (depending on the preservation ofthe archaeological material). When they are visiblethey appear as a collection of parallel lines of vary-ing width and density (hence the term 'bar code').Each combination is specific to the state of the activeedge at any given moment. It is possible to superim-pose different bar codes and thus compare them. Intheory we should be able to determine the numberof tools used for an engraving and to establish atempting equivalence between:

[bar code A = tool A] [bar code B = tool B]

Unfortunately, the primary weakness of this propo-sition stems from the possibility of re-sharpening.When a burin is retouched, the quartz microfossilsand lepispheres are completely modified and ouranalysis would conclude that a different tool hasbeen used.

A second difficulty comes from the state of pres-ervation of the artefacts. Reading bar-codes onarchaeological material demands exceptionally well-preserved crevice depth. In these conditions such an

193

Carole Fritz

mu

Figure 2. tf) Accidental scratches, b) Experimental scratches in a vertical line. Archaeological scratches on reindeerantler, c) Experimental bar codes. Archaeological bar codes. On archaeological bar codes, the tool pressure is lessvisible owing to greater erosion.

194

Reconstruction of Magdalenian Artistic Techniques

analysis is possible only ona small number of pieces.In practice, bar-code obser-vation remains partial; nooverall view can be pro-posed (Fig. 2c).

Other details maycharacterize a tool. TheMorphology of certainmarks, for example theATT, may be regarded asspecific to an active agent.It is then possible to ob-serve the ATTs and BFTswhich belong to the samebevel. Others have re-marked that certain 'para-site traces' which are left°n the edge or at the endof a line indicate the sameactive agent on an objectwhich has multiple lines(d'Errico 1988; 1994). Wehave not actually observedsuch marks, no doubt be-cause they are specificallylinked (at least in part) tothe morphology of the par-ticular supports (such asf'at pebbles with convexedges) and to the gesturaldynamics which followfrom them.

Experimentation andanalysis have re-

sulted in reliable ways ofdeciphering layers of linesO r marks. Thanks to thehigh definition and depth°f field of the scanningelectron microscope, allthe superimposed markscan be observed at increas-l n g scales according totheir legibility. So far wehave experienced no fail-ure in this area; study oftile chronology of the marksand therefore the succes-sion of actions may beundertaken with greatconfidence.

1 2 3 4 5

Figure 3. Figure showing the origin of inicrostriations which are recognizable in thegroove. No. 1 is caused by a niicrofossil; 2, 3, 4 and 5 by quartz lepispheres. The termbar code is applied when they occur together.

LabastideEnlAe

Latigerie-Basse

• Monn

Figure 4. Geographic distribution of Magdalenian sites studied.

195

Carole Fritz

These results derive from the comparison ofexperimental and microscopic observations (White1982; d'Errico 1987; 1988; 1994; d'Errico & David1993; Fritz et al. 1993). Yet the question remains howwe can be certain that the observations on the ex-perimental pieces are identical to those observed onthe archaeological material.

We can reasonably infer that the physicalmechanisms used when a flint tool cuts bone havenot varied significantly over time. In other words, inthe physics of the raw material, the problems facingthe Magdalenian engraver are identical to those metwith in the course of experimentation. The basis forreproducing marks experimentally (on bone) is there-fore identical to that faced by Palaeolithic engravers,only the condition of surface preservation makingthe interpretation of marks more difficult. The na-ture both of the support and the tool can consider-ably affect the readability of the technological indicesand their morphology. In keeping a flint tool con-stant but changing the surface to be engraved (opt-ing for example for a rock), the criteria of technicalevaluation remain, perhaps, structurally close to theircounterparts on bone. Their physiognomy and theirclarity, however, would be affected by the lesserplasticity of the raw material. In addition, the meansof removal and of displacing the material are com-pletely different because the granulometry of theraw material plays a significant part in the preserva-tion of the marks.

As in other areas, notably microwear, we pro-ceeded by analogy. The structural similarity ofmaterials (prehistoric and modern) and the repro-ducibility of physical phenomena allow us to main-tain that the marks observed on the experimentalpieces are identical to those seen on archaeologicalmaterial. Experimentation helps to reproduce a se-lection of actions which generate models of micro-scopic structures (BFTs, ATTs, bar codes), thedynamic origin of which is known; as a second step,we thus observe the archaeological pieces and com-pare marks on the two categories of object. To returnto comparisons with microwear, the field of analysisshifts from visible microtraces on the tool to thoseleft on the engraved material.

Graphic elements anci construction of outlines

The sites considered in this study are representativeof the Magdalenian of the southwestern quarter ofFrance (Pyrenees, Aquitaine). Their geographic dis-tribution illustrates the limits of expansion of theFrench classic Magdalenian: from the Dordogne

(Laugerie Basse) to the Gironde (Abri Morin,Fontarnaud) to the Pyrenees Atlantiques (Arancou)and Ariege (Enlene, Mas-d'Azil). In total, ninety-oneobjects have been studied (ribs, bone blades, birdbones, pendant and antler harpoons and points).

In each representation the engraved mark isconsidered as a part of a graphic element (back,belly, collection of juxtaposed and parallel lines)which, in association and articulation with others,makes up forms, whether figurative or not. The piecesexamined have therefore been studied line by line,element by element, without making any a prioridistinctions between the figurative representationsbe they geometric or unspecified. This hierarchicalstructure of microscopic observation allows us toreconstruct the process of producing graphic elementsbased on their superimposition. Sixty-nine animalswere examined. This is much less than the total ofbone pieces examined: in the first place we ex-cluded animals which were too incomplete (cervidsrepresented by a section of antler) or unidentifi-able, and for which technical comparisons wouldhave been of little interest. On the other hand, 13geometric motifs were included, along with com-plete animal figures on which body segments can becompared.

Quadrupeds

We include in this category observations made on 38herbivores (horse, cervids, bison, aurochs), with theexception of frontal views of ibex heads and uniden-tifiable shapes (stylized birds?) discussed below. On

Figure 5. Chronological sequence of execution ofgraphical elements.

196

Reconstruction of Magdalenian Artistic Techniques

these representations, the engraving of contourstraces the outlines of shapes and draws the silhou-ette of the animal. These contours comprise vari-ously positioned graphical elements which formseparate components of the outline: the head (snout,muzzle and cheek), horns or antlers, breast, legs, lineof the belly (the angle of the sheath where it exists),the rump, the tail and the cervico-dorsal line.

Breaking down the graphical elements * intochronological sequences illustrates identical gesturalsequences (of the engraver's hands). The outline ofthe head was generally engraved first, then the backor the breast. In the absence of superimpositions it isimpossible to establish a distinction between theseelements. The fourth component to be executed isthe front leg, drawn after the breast (or alternatively:leg/breast/leg). The line of the belly comes in fifthplace (with its sexual attributes), then the hind limbsand rump, and finally the tail (Fig. 5).

This sequence of execution has been identified°n all the animal figures considered. Such consist-ency shows that Magdalenian artists constructed theiranimal silhouettes, whatever their size and orienta-tion, starting at the head and finishing at the rear.

One of the most complex cases is illustrated bythe bone slab of Arancou called the 'bicheauxpoissons',°n which we can see gestural sequences alternatingbetween the breast, the front feet and the line of thebelly. The Magdalenian artist first engraved thebreast, the front hooves and the belly, then succes-sively deepened the line of the belly, the foreleg andcompleted the final lines of the breast (Fig. 6). With-out distinguishing the alternating sequences, thechronology of the animal would be difficult to inter-pret. The final passes of the tool would have com-pleted the hoof/belly/breast construction.

It is the different layers of superimposed en-graving and, to a lesser extent, the depth of the linewhich help us to understand the 'alternating gesturalsequences'. They are not easily recognizable on allthe artefacts, but we have been able to detect many°^ them on the most complete animals without anypreconceived idea of the quality of execution. Aswell as a gestural progression from the front to therear of the animal there is a hierarchy in the con-struction of other figurative elements. As has al-ready been shown, the artist chooses to begin withlh head (less frequently the back). The head is

up of graphical elements which are oftencrucial for its identification: horn or antler, ears,snout, muzzle and cheek. On animals with cephalicappendages these extremities appear to be essential(pigs. 7 & 8).

The specific case of ibex heads in frontal view

The five ibexes in frontal view all come from theGrotte de la Vache (Fig. 9). They are found on ribs,bone fragments and bird bones. The heads in frontalview are drawn in an identical way: two divergingcurves indicate the horns, two smaller lateral curvesrepresent the ears. The heads of the animals weredepicted by two lines converging below. On somepieces 'dots' denote the coat. Comparison of the meth-ods of execution leads us to discern a single scheme:the horns are systematically drawn first, then theears and the lines of the face and finally, when theyare present, the dots of the coat. They show theexistence of a repetitive model of the individual ele-ments and therefore an archetypal mode of construc-tion and representation of ibexes in frontal view. Isthis mental scheme unique to the Magdalenian artistsof La Vache or is it shared by other groups? Sincethis sample is limited to the figures from one site it istoo soon to rule out the possibility of a commonschema for this type of stylization.

Indeterminate shapes from Fontarnaud (Gironde)and La Vache (Ariege)

Ten examples of unidentifiable figures are found ontwo pieces of bone from two sites: four on a piecefrom Fontarnaud and six on a piece from La Vache(Fig. 10). Strong similarities between these forms areimmediately noticeable: their small dimensions andthe fact that each one comprises a round head markedby a somewhat open angular line. The four speci-mens from Fontarnaud are decorated with a form ofbreast to which the head is attached. This breast isnot legible on the examples from La Vache, but someof them nevertheless possess a 'neck'. All these shapeshave an eye. At La Vache it is a simple vertical dash.At Fontarnaud on the other hand it is more detailed:we see a semi-circular line often with vertical dashes(eye-lashes?), thereby giving the impression of aclosed eyelid. The hypothesis of stylized birds isplausible (Breuil 1905; Marshack 1972).

At La Vache, pictures are engraved in a line,with one following another. Of comparable struc-ture, the quality of their execution deteriorates fromright to left; the first are clearly drawn, the last morerapidly sketched. The examples from Fontarnaudform a vertical frieze, with figures alternately facingtowards the right and left, all executed with care.

On both objects, these enigmatic creatures aredrawn in the same way; the artist engraves the top ofthe head, then the neck, an angular shape (beak?),

197

Carole Fritz

5 ••"•-•

• ' \ ' y

•' • *- *f• r. -•''" '•"; .>'

E

A

- • -

• .

D

. . ' • ' ! ! - ' • • . ' • >

<»•_ ••;_j-» .V_•-__

^ ;f- f* i--'V

B

* • • • • " ' .• " , . • •

' • • • • • - r >

c

Figure 6. Slab bone 'biche aux poissons' (Arancou, Pyrenees Atlantiques, Musee National de Prehistoire). Gesturalsequences in the execution of the breast, front legs and belli/ of the hind.

the lower part of the head and finally the eye (fol- Among possible explanations, the most compellinglowed by the dashes (eye-lashes?) on the examples would be the diffusion of a type of figure and itsfrom Fontarnaud). execution sequence between the Ariege and Gironde

To date, such stylized heads have only been regions. Such contact does nothing to contradict thefound at these two sites. Of very similar appearance, chronological attribution of artefacts from Upperthey indicate the same sequences of construction. Magdalenian levels. Nevertheless, these forms being

198

Reconstruction of Magdalenian Artistic Techniques

relatively non-complex, one can not ignore the pos-sibility of simple convergence.

Aquatic fauna

Aquatic fauna consists of the marine mammals andfish. In all, nine specimens have been recognized:four unidentifiable fishes and a cetacean fromArancou; a seal, an Atlantic salmon and two uniden-tifiable fishes from La Vache.

Despite their zoological diversity one noticesVery similar graphical construction. The aquatic ani-mal is perceived as a fusiform silhouette to whichspecies-specific attributes are added. The outlinesare drawn initially, beginning in eight out of thettine cases with the upper part of the body. In allcases, after the outlines, the artist 'traces' the finsar»d the tail, and the lateral lines (specific to theSe>lmonids). One notices during the execution of theseParts alternating gestural sequences; sometimes the

Figure 7. Fish-shaped bone slab from Lingerie Basse(Dordogne: 38189 1362 Musee de VHomme). (544) Theupper part of the outline of the muzzle is part of the front.In the lower part, repented lines are drawn before engravingthe mouth. The latter is linear, engraved with three strokeswhich are slightly out of line. (556) Mane follows after ear.

199

Carole Fritz

d v-\>; fi.

t.v * v f>:."«• v

5i:--wJ

i«

Figure 8. Fragment of rib bone decorated with three aligned bovids (Laugerie Basse, Dordogne: 38189 1722, Musec dcI'Homme). (588) On the left part of the micrograph the BFTs of the front and those completing the upper part of thesnout are clearly seen. The superimposed lines show a sequence: front/muzzle/mouth/chin. (582) The eye is drawn fromright to left with three strokes in each line. The eyelashes are drawn after the outline of the eye which is drawn after thefront.

200

Reconstruction of Magdalenian Artistic Techniques

1 C

Figure 9. Frontal view of ibex heads from La Vache (Ariege).

engraver takes the opportunity to place an attributeso as to deepen the existing outline, for example fin/outline/fin. This sequence can be observed on thecetacean on the Arancou pendant.

The rudiments of outlines and internal attributes

Whatever the animal, the Magdalenian artist seemsto follow a precise sequence in the chain of actionsand therefore in the construction of a figure. Theseactions can be interpreted as the reflection of rela-tively strict mental schemes governing shapes.

The outline defines the general shape. It is thiswhich separates the shape from the object on whichit is engraved. Recognition of the engraved subjectand the expressive force of the animal rests, to alarge degree, on the precision of its outline. Its posi-tioning is therefore of primary importance. In addi-tion to the outline, animals show details (coat, eyes,etc.) which we choose to call 'internal attributes'.

Internal attributes complete the animal outline,illustrating certain anatomical characteristics or ful-filling a symbolic role.

Anatomical characteristics can be divided intotwo categories:* easily identifiable6 segments such as eyes, mouth,

nostrils, teeth, and coat;• details revealing a familiarity with the animals and

their habitual behaviour which only in-depthobservation of the wild animal can provide; forexample, bony protuberances or muscular in-sertions visible under the hair, shades of coat,specific sexual differences, age- or season-re-lated variation.

Figure 10. Unidentifiable forms: above - Fontarnaud(Gironde), below - La Vache.

Symbolic signs are less common on animal bodies inmobiliary than in parietal art. Angular signs do ap-pear, however, notably the tapered triangular formsclearly identifiable on the flank or stomach of ani-mals and the long lines on bodies which appear likewounds (real or symbolic) inflicted on the animals.

Outlines and internal attributes

Outlines and internal attributes involve distinctmodes of execution which reinforce differences ingraphical status.

Outlines, essential to identification, are gener-ally deep grooves (between three and eight strokes

201

Carole Fritz

.-*»<-*.__._.'••

£

jr..

Figure 11. Pendant with cetacean and red deer head. (Arancou, Pyrenees Atlantiques: Musee National dePrehistoire). Alternating gestural sequences between the fin and the line of the stomach of the cetacean. The fin isinserted befrveen two i>ersions of the ventral line.

202

Reconstruction of Magdalenian Artistic Techniques

884774

884775884748

Figure 12. A horse and two aurochs from Abri Morin (Gironde: Musee d'Aquitaine). (Drawn by P. Laurent.)

°f the tool) thereby fixing the animal firmly on thematerial. At La Vache (Ariege), an Atlantic salmon°n a bird bone shows between three and five strokesfor the outline of the head; the chest of a reindeerfrom the Abri Morin on bird bone between four andsix strokes; the line of the belly of the hind fromArancou between five and seven (Fig. 6). This list ofexamples is not exhaustive and on most of the piecesexamined with the scanning electron microscope, aswell as others observed in the traditional way, sys-tematic deepening of the outlines of animal figuresmay be noticed. On three objects in our collection,however, this is not observed; these are pieces fromthe Abri Morin (Fig. 12). They belong to the FinalMagdalenian layer Al.

The characteristics of the internal attributes arethe complete reverse of the above. They are mostoften executed with two to four strokes. The mostconsistent elements are the coat, the distinguishingmarks and nostrils. There are differences in the eyesand mouth; it would seem that in some cases theseelements are considered to belong to the outline and,consequently, are executed as such (deeply marked).

One can cite as an example the rib fragment fromLaugerie Basse on which the central animal has a mouthengraved with five or six strokes, the legibility of whichis comparable to that of very deep outlines (betweensix and nine strokes according to element) (Fig. 7).

Internal attributes present an immutable andpredictable chronological sequence: they are alwaysadded after the outline, irrespective of the type ofanimal. It is possible to affirm a sequence of ele-ments: sensory organs (eye, mouth, nostrils) precedethe coat, and the distinguishing marks come later.Such a sequence is clearly legible on the reindeerheads from the Abri Morin. The chronological se-quence between the outline and internal attributesappears to be logical; it is difficult to imagine theengraver positioning the eye or nostril before theoutline of the animal. Nevertheless, one could imag-ine gestural sequences alternating between theoutline and internal attributes; for example, the po-sitioning of the head, then the details relating to it(eye, mouth) before following with the rest of thebody and the legs. Such a scenario has never beendocumented among the objects studied.

203

Carole Fritz

6099 lmm WD39

Figure 13. (a) Angular signs used as internal attributes on the body of a red deer, (b) Single geometric decoration on abird bone, (c) Angular signs used as an outline separate from animal figures, (d) Dots produced by pecking, causingaccumulation of material at the BFT. (e) Spindle-shaped detail: the outline is deepened and the dashes used as internalattributes, (f) Complex geometric assemblage combining elements considered as outline and those equivalent tointernal attributes.

204

Reconstruction of Magdalenian Artistic Techniques

Geometric decoration

Geometric marks have always been difficult to de-tect and understand, for they remain enigmatic, notonly because of their formal structure but also theirsignificance.

It is no less true that abstract designs are madeup of organized graphical elements, which can bedescribed with reference to geometric forms. Eachform comprises incisions which mark the outline ofthe shape and therefore, just like the figurative mo-tifs, isolate it from the base material.

Geometric designs are present at all the sites inour data base. They may consist of unrelated decora-tion but they are, most frequently, associated (di-rectly or indirectly) with animals. We have chosen toisolate purely geometric forms, i.e. those whose struc-ture has no figurative element. This choice is par-tially arbitrary. The principal types include linearrnarks, dashes, angular signs, arc shapes, broken lines,funates, dots, various impact marks and combina-tions and repetitions thereof.

Two modes of execution can be distinguished:• Most often these motifs are made up of a series of

lines (straight or curved) associated in variousways. They can accordingly occur in a largenumber of combinations: there are long lines,short lines, angular lines, crosses, broken lines,arc-shaped lines, lunates, triangles (Fig. 13).

• More rarely, geometric motifs are produced bypecking on the material, thereby giving a falseimpression of impact on the surface. We havediscerned two main types: dots and triangularmarks. They invoke a different technique. It isno longer a question of producing a groove inthe material; instead the engraver takes accountof the specific morphology of the impact pro-duced by a very short and powerful action. Notechnical example combines both of these types.The pecked motifs are found twice at Arancouand on the frontal view of the ibexes at La Vache.

Geometric themes constructed with the aid of in-cised lines benefit from the same graphical treat-ment as the animal figures. The lines present identicaltechnical characteristics (sections of marks, number°f strokes, depth, etc.). Nevertheless, we must dis-tinguish the geometric elements strictly associatedwith animals (body signs) from those which are ad-Jacent but beyond the outline.

The first, incorporated into the body of the ani-mal, form part of the scheme of construction of inter-nal attributes. They include linear marks, dashes andangular signs. On average, two or three cuts in each

mark can be counted. Indications of coat are alsofound on cervids at Arancou and La Vache, andeven bird plumage is indicated on the 'grasshopperbone' from Enlene.

Geometric elements placed outside the figuresare considered separate motifs. As in the case ofanimal outlines, the number of tool strokes variesfrom two to seven for each line (Fig. 13).

Overall, the geometric motifs which wereanalyzed are treated graphically in a way compara-ble to that of figurative decoration, with an equivalentdistinction between marks considered to be outlinesand those relating to internal characteristics.

Conclusion

Comparison of motifs produces an identical schemeof figure construction from the Pyrenees to theGironde. The animal is conceived and produced fromfront to back, beginning with the shape whichsketches the outline of the animal; then the internalattributes are gradually put in place: the anatomicalelements (eyes, nostril), the coat and finally the sym-bolic signs. In practice, this simple model of con-struction is refined by phases of deepening of marks,thereby giving rise to gestural sequences which al-ternate between the different segments of the shape.

Given that the animal figure is made up of anoutline plus internal attributes, differences in treat-ment between the two categories are noticeable: theoutline is more deeply incised in order to allow rapididentification; the internal attributes play a comple-mentary role as anatomical details, indication of vol-ume, signs, and so forth, and are executed moredelicately.

These categories are equally applicable to geo-metric motifs according to their position both on thesupport and on the animal: they are considered asinternal attributes if they are superimposed or inte-grated on the animals (symbolic signs), and as indi-vidual motifs when found on the periphery. Withinthe technical realm, no distinction between figura-tive and geometric motifs can be identified: they aretreated, and should therefore be considered, inequivalent terms.

Microscopic analysis of the animal and geomet-ric images from the sample analyzed demonstratesconceptual uniformity of the mode of execution. De-spite redoubled attention, we have never identifiedfeatures illustrating site-specific or regional particu-larities. From one site to another in this area theMagdalenian engravers possessed a single concep-tual scheme without any significant variation in space

205

Carole Fritz

or time. This constant formal concept raises ques-tions about the cultural or cognitive origin of themodes of execution. Were the gestural sequencesimposed by culture, and transmitted by learning? Orwere they natural, revealing the cognitive capacitiesof Homo sapiens sapiens?

This body of data includes products primarilyof the Middle and Upper Magdalenian; in additionthere are two engravings (a horse and an auroch)attributed to the Magdalenian-Azilian transition. Thetechnical sequences of these two latter pieces presentgestural sequences distinct from those identified onthe Magdalenian pieces. These comparisons suggestthat a profound change occurred between theMagdalenian and Azilian. Gestural discontinuitywould without doubt parallel other formal changesidentified in Azilian pieces (Roussot 1990; Guy 1993).From our point of view, the break between these twoperiods seems certainly to be cultural and illustratesone of the results of this work; the modes of execu-tion belong to the perception of form across a cul-tural filter than to a mental scheme particular of theUpper Palaeolithic. Consequently, the cognitive partof engraving widely predates the Magdalenian andremains within the capability of each Homo sapienssapiens to conceptualize and to inscribe figurative orabstract motifs on the material used.

To confirm this hypothesis, to make clear itsmethods and limits, we must extend the cross-cul-tural comparisons in order to establish essential docu-mentation bases before determining the specificmental processes.

The stability of mental schemes can be traced inorigin to the heart of Magdalenian symbols. To be-gin designing a figure with its head presupposesthat the artist associates emblematic or metaphoricvalue with this element. The head would indicatethe essential centre of the animal, as much physicalas spiritual. It would therefore be necessary to give itthe primary role. The scheme is reinforced by theposition of the antlers and horns, completed beforeall other segments. For the engraver-hunters, thesemay have been above all dangerous attributes, the'weapons' of the animal. They may alternatively havebeen simply the elements by which the animal wasimmediately identified.

In Magdalenian art, the intentional absence ofthe head seems to obey certain rules, since one findsheadless figures beside separate heads (protomes). Inone case the head is intentionally absent, in anotherit is the main feature of the image. Can one deducefrom this that the execution of figures is intimatelyallied to their symbolic value, to the perception which

engravers had of them, and through them, the ethnicgroup?

The recognition of a single procedure of execu-tion in the entire series of images implies controlledtransmission of method from the stage of learning.Isolating gestural norms suggests that the coding ofart is present at its source, from the first stages ofproduction. Since mobiliary art was the bearer of asemiological system, understood and mastered bythe group, one assumes that learning the graphicalcodes figured amongst the social constraints.

In Magdalenian mobiliary art, the presence ofsimilar themes, iconography or stylistic treatmenton objects found at neighbouring sites (the fawnspearthrowers from Mas d'Azil and Bedeilhac), tendsto confirm their local originality (Bahn 1982; Simonnet1989; Conkey 1992; Taborin 1992). Besides these ex-amples of confirmed particularism, one finds identi-cal pieces at distant sites which attest specific parallelsat the regional scale (sculptures from Isturitz andEnlene) or an even wider distribution ('contourdecoupes' of horse heads from Perigord, Ariege andAsturia) (Buisson et ah 1996).

The two trends, strong local identity and re-gional or long-distance diffusion, can be perceivedto be dichotomous. Part of the problem lies in thedefinition of the similarities or differences; severalparameters (stylistic, thematic, etc.) must be consid-ered. As a result of this study we can demonstratethat technique in the strict sense, that is to say pro-duction of incisions on-bone surfaces, does not fig-ure among the determining criteria. Across our database, it has been impossible to identify features spe-cific to one site or even one region; the corollaryapplies that no diffusion of techniques has been ob-served.

Without ignoring the need to increase the sizeof our sample, we note that other research has reachedcomparable results: 'one can see that all the anatomi-cal details (horse heads in 'contours decoupes') havebeen subjected to conventional graphic treatment,but that none of them is site specific. There is no"Isturitz-type", nor is there a "Mas d'Azil" variety,not even at a detailed level' (Buisson et ah 1996,338).

The above statement and the graphical coher-ence of this collection, in terms both of region andchronology, indicate that these procedures belong toa common Magdalenian repertoire. At this level ofanalysis, we have to reassess the identification ofhypothetical regional centres of artistic production.The existence of such centres remains only theoreti-cal. The originality of a site lies more in the creationof emblematic themes than in the manner of express-

206

Reconstruction of Magdalenian Artistic Techniques

ing them. Does this originality necessarily lead to anartistic diffusion? Magdalenian groups should havebeen sufficiently isolated to have developed certainparticularities, but still maintain sufficient contactwith others to remain within the current of sharedgraphic expression. We return therefore to the no-tion of aggregation sites (Bahn 1982; Conkey 1992)which furnishes a model of ethnic organization whichis compatible with the archaeological evidence andnotably allows us to explain the dual tendency (iso-lation and diffusion) mentioned above.

The examination of large Pyrenean sites andtheir mobiliary art favours a comprehensive assess-ment of the evidence. Should we not place the char-acteristics of a site in a wider context before discerningpotentially diffused elements? Behind the abundanceof evidence, the same general question applies: onwhat criteria is the definition of a cultural groupbased? As a result, at what level can it really bedistinguished from its neighbour? To answer thesequestions, several types of archaeological material(worked or used stone tools, personal ornaments,art) must be considered and the results compared. Itis in this perspective of multidisciplinary research,ever changing, that we have conceived (and hope tofollow up) this work.

Acknowledgements

should like to thank the Fondation Singer-Polignacits financial support and the Laboratoire de Re-

cherche des Musees de France for its logistical su-pervision as well as the curators of the followingmuseums for loan of artefacts: Musee de l'Homme,paris (Laugerie-Basse), Musee d'Aquitaine, Bordeaux(Fontarnaud and Abri Morin), Musee National dePrehistoire, Les Eyzies de Tayac (Arancou) and theMusee des Antiquites Nationales, Saint-Germain-en-Laye (La Vache).

Carole FritzUMR 5608 Toulouse (Priliistoire)

Laboratoire de Recherche des Musees de France6 rue des Pyramides

75041 Paris Cedex 01 .France

Notes

The term 'bone' is used here sensu lato, referring to allbone material, including cervid bone and antler.This abundance of mobiliary art on bone, antler orivory should be tempered by factors of preservation.Many other potential supports in wood, clay, copper,

etc., will have disappeared.3. It is of course out of the question to cover an archaeo-

logical artefact with gold.4. Brand: Coltene President Body Light.5. This observation can only be made with a S.E.M.6. Identification based on personal knowledge and ex-

perience (and its limitations).

References

Bahn, P., 1982. Inter-site and inter-regional links duringthe Upper Paleolithic: Pyrenean evidence. OxfordJournal of Archaeology 1(3), 247-68.

Breuil, H., 1905. La degenerescence des figures d'animauxen motifs ornementaux de l'epoque du renne, inComptes rendus des seances de I Academie des Inscrip-tions el Belles Lellres. Paris: Picard.

Buisson, Dv C. Fritz, D. Kandel, G. Pincon, G. Sauvet & G.Tosello, 1996. Analyse formelle des contours de-coupes de tetes de chevaux: implications archeo-logiques. 118eme Congres national des SocietesHisloriques el Scientifiques (Pau 1993), 327-40.

Conkey, M., 1992. Les sites d'agregation et la repartitionde l'art mobilier: ou y a t-il des sites d'agregationmagdaleniens?, in Actes du Colloques de Chancelade(1988)>.Taris: Editions du CNRS, 19-25.

Cremades, M., 1989. Contribution a l'etude de l'art mobilierdu Paleolithique superieur du Bassin aquitain: tech-niques de gravure sur os et materinux organiques.Unpublished Ph.D. theses, Universite de BordeauxI. 2 volumes.

Cresswell, R., 1983. Transfert de technique et chainesoperatoire, in Techniques et culture, vol. 2: juillet-decembre. Paris: Editions de la Maison des Sciencesde l'Homme, 145-63.

Delporte, H. & L. Mons, 1973. Notes de technologie et demorphologie de l'art paleolithique mobilier: bisongrave sur plaquette de limon de Bedeilhac (Ariege).Antiquites Nationales 5,20-32.

Delporte, H. & L. Mons, 1975. Omoplate decoree du Masd'Azil (Ariege). Antiquites Nationales 7,14-23.

Delporte, H. & L. Mons, 1976. Principes d'une etude surles supports osseux de l'art paleolithique mobilier,in Methodologie appliquee a I'industrie de I'osprehistorique. (Colloque international du CNRS 568,Abbaye de Senanque (Vaucluse), 9-12 juin 1976.)Paris: CNRS, 69-75.

D'Errico, F., 1987. Nouveaux indices et nouvelles tech-niques microscopiques pour la lecture de l'art gravemobilier. Comptes-Rendus de VAcademie des Sciencesde Paris 304,761.

D'Errico, F., 1988. Study of Upper Palaeolithic andEpipalaeolithic engraved pebbles, in Scanning Elec-tron Microscopy in Archaeology, ed. S.L. Olsen. (Brit-ish Archaeological Reports 452.) Oxford: BAR,169-84.

D'Errico, F., 1991. Etude technologique a base expdri-mentale des entailles sur matiere dure animale. Im-plications pour 1'identification de systeme de

207

Carole Fritz

notation, in 25 arts d'etudes technologiques en preliistoire.(XI rencontres Internationales d'archdologie et d'his-toire d'Antibes.) Juan-les-Pins: Editions APDCA, 83-97.

D'Errico, F., 1993a. Identification des traces de manipula-tion, suspension, polissage sur l'art mobilier en os,bois de cervides, ivoire, in Traces etfonction: lesgestesretrouves. (Colloque international de Liege 50(1).)Liege: Edition ERAUL, 177-87.

D'Errico, F., 1993b. La vie sociale de l'art mobilierpaleolithique. Manipulation, transport, suspensiondes objets en os, bois de cervide, ivoire. Oxford Jour-nal of Archaeology 12(2), 145-74.

D'Errico, F., 1994. L'art grave Azilien. De la technique a lasignification. (XXXIeme suppl. a Gallia Prehistoire,323.) Paris: Editions du CNRS.

D'Errico, F. & S. David, 1993. L'analyse technologique del'art mobilier. Le cas de l'Abri des Cabones (Ranchot,Jura). Gallia Prehistoire 35,76-139.

D'Errico, F., G. Giacobini & P.-F. Puech, 1984. Les repliquesen vernis des surfaces osseuses fac,onne"es: etudesexpe'rimentales. Bulletin de la Societe PrehistoriqueFrangaise 81(6), 169-70.

Feruglio, V., 1993. La gravure, in L'art parietal paleolithique,techniques et methodes d'etude. Paris: Editions du Comitedes Travaux Historiques et Scientifiques, 265-74.

Fritz, C , 1995. L'art mobilier paleolithique: du visible al'invisible. Cahier art et science 2, revue annuelle,Universite de Bordeaux I, 97-108.

Fritz, C , 1996. Procddes artistiques et societes mag-daleniennes: contribution de l'analyse microscopiquea l'Stude de la gravare sur matiere animale (os etbois de cervide). Unpublished doctoral thesis,l'Universite' de Paris I Pantheon-Sorbonne. 2 vol-umes.

Fritz, C , 1999. La gravure dans l'art mobilier magdalenien: dugesle a la representation. (Documents d'ArcheologieFranchise 75.) Paris: Maison des Sciences del'Homme.

Fritz, C. & R. Simonnet, 1996. Du geste a l'objet: les con-tours de"coupe"s de Labastide, re"sultats prdliminaires,in TECHNE, la science au service de l'art et des civilisa-tions. Paris: Editions du Laboratoire de Recherchedes Muse"es de France, Reunion des MuseesNationaux, 63-77.

Fritz, C , M. Menu, G. Tosello & P. Walter, 1993. La gra-vure sur os au magdale*nien: e"tude microscopiqued'une cote de la Vache (commune d'Alliat, Ariege).Bulletin de la Societe Prehistorique Frangaise 90(6), 4161-425.

Guy, E., 1993. Enquete stylistique sur l'expression figura-

tive epipaleolithique en France: de la forme au con-cept. PALEO 5, 333-73.

Legua'y, L., 1877. Procedes employes pour la gravure et lasculpture des os avec le silex a l'dpoque pre-historique. Bulletin de la Societe d'Anthropologie deParis 12,280-96.

Leguay, L., 1882. Sur la gravure des os au moyen du silex.A.F.A.S. He session, la Rochelle 1882,677-9.

Lemonnier, P., 1983. L'etude des systemes techniques, uneurgence en technologie culturelle, in Technique etculture, vol. 2: juillet-decembre. Paris: Editions de laMaison des Sciences de l'Homme, 11-34.

Leroi-Gourhan, A., 1965. Prehistoire de l'art occidental. Paris:Mazeno, 498.

Marshack, A., 1972. Les racines de la civilisation. Paris: Plon.Pelegrin, J., C. Karlin & P. Bodu, 1988. Chalnes operatoire:

un outil pour le prehistorien. Technologie prehistorique,notes et monographies 25,55-62.

Pigeot, N., 1987. Magdaleniens d'Etiolles, economie de debitageet organisation sociale (Unite d'habitation U). (XXVemesuppl. a Gallia Prehistoire.) Paris: Editions du CNRS,168.

Pigeot, N., 1988. Apprendre a debiter des lames: un casd'education technique chez des magdaleniensd'Etiolles. Technologie prehistorique, notes et mono-graphies 25,27-36.

Pigeot, N., 1991. Entre nature et culture, valeur heuristiquede la technologie lithique par des approchessystemique et cognitives. These d'habilitation del'Universite de Paris I, Pantheon, Sorbonne, 200.

Roussot, A., 1990. Art mobilier et parietal du Pdrigord etde la Gironde: comparaison stylistique, in L'art desobjets au Paleolithique, vol. I: L'art mobiliet et soncontexte, 189-205. (Actes du colloque de Fois-le Masd'Azil, 1987.)

Sdris, J.P., 1994. La technique. Paris: Presses Universitairesde France.

Simonnet, R., 1981. Cartes des gites a silex des Pre-Pyr<§-nees. XXIe Congres S.P.F., Cahors-Montauban, 19791,308-23.

Taborin, Y., 1992. Les espaces d'acheminement de certainscoquillages magdaldnien, in Le Peuplement mag-dalenien: Actes du Colloque de Chancelade (1988). Paris:Editions du CNRS, 417-29.

Walter, P., 1993. Etude du comportement du fluor lors desinteractions silice-solutions aqueuses: Applicationsarcheologiques. Unpublished thesis, l'UniversitePaul Sabatier de Toulouse (Sciences).

White, R., 1982. The manipulation of burins in incisionand notation. Canadian Journal of Anthropology 2(2),129-35.

208