Pr&iistoire Europeenne, volume 13,1999, p. 33 a 66.

THE CHAINES OPERATOIRES OF LEVALLOIS SITE PRONYATYN,WESTERN UKRAINE

Vadim STEPANCHUK1, Olexander SYTNYK2

ABSTRACT :

The Middle Palaeolithic Levallois site of Pronyatyn geologically dated to Ammersfort is situated in south-west ofVolyno-Podolian upland at the right bank of the river Seret, the left tributary of Middle Dniester, 6 km north from Ternoptlcity, West Ukraine. The site was discovered and investigated between 1977 and 1985. Assemblage includes c. 6.000 lithics.The technological analysis allowed to elucidate highly complex system of different core reduction strategies, which wereapplied at the site. The overwhelming majority ofknapped products were obtained through the exploitation of flat cores,although the idea of volume was also known and time to time was applied. Three specific forms of pre-cores can bedistinguished, each of which adds specificity to modes of further reduction. Careful curation of properties of removals bydifferent methods of preparation of striking platform zone (trimming, facetting) and of maintenance of flaking surfaceconvexity (different debordants) was conducted during each stage of core reduction. There is evidence of a row of recurrent

technology are consist in frequent application of trimming technique during striking platform preparation and inacquaintance with volumetric mode of raw materials exploitation, resulted, among other, in particular type of pre-cores.

Key-words : Eastern Europe, West Ukraine, Middle Palaeolithic, Levalkis, Lithic Technology

Proposed paper is devoted to thestudy of stone knapping technology of theWest Ukrainian Middle Palaeolithic stratifiedsite Pronyatyn studied between the years 1977- 1985. The questions of detailed dataconcerned history of investigation, localisationof the site, available natural science data,typology of lithics are not the subject of thisarticle and were published elsewhere(SYTNYK 1978,1985,1994).

BRIEF DESCRIPTION OF THE SITE

bearing sediments were uncovered andstudied during this period and provides c.6.000 lithics, not frequent fauna remains, tracesof hearths and other human activity relatedrecords associated with single occupationepisode (SYTNYK 1978,1985,1994).

Stratigraphy was studied by A.Bohucki, Lviv State University, who arguedAmmersfort age of culture bearing sediments.Fauna species were determined by N. Belan,Kiev Zoological Institute. The list containswoolly rhinoceros, ancient bison, reindeer,mammoth.

The Middle Palaeolithic site ofPronyatyn is situated in south-west of Volyno-Podolian upland at the 30-40 m high rightbank of the river Seret, the left tributary ofMiddle Dniester, 6 km north from Ternopilcity, West Ukraine. The site was discoveredand investigated by O. Sytnyk in 1977-1981,1984 and 1985. 520 square meters of culture

•a, 12,1 Institute of Archaeology, Geroyev StalingradUA-254655, Kyiv-210, Ukraine.2 Institute of Ukraine Science, Kozelnycka, 4, UA-290026, Lviv-26, Ukraine.

STONE INDUSTRY

Local high quality black and greycretaceous flints of Turonian deposits wereused as raw material for knapping. The nearestknown outcrop of flints is about 6 km northfrom the site. Raw materials constitute eitherirregular crude nodules with massiveprotuberances or regular cobble like nodules.

Physical state of flint artifacts is rathergood. There are no traces of lustrage, erosion

33

Vadlm STEPANCHUK, Olexander SYTNYK

etc. Some pieces show signs of frost fractures.Almost all lithics are covered by light bluepatina.

According to calculations published in1994, the culture bearing layer has produced91 complete and 157 fragmented cores, 434Levallois blanks without secondary working,584 blades, 2.743 flakes, 295 tools, 2.106 debris(SYTNYK 1994). The assemblage was re-examined recently with the purpose to restoreas detail as possible the Pronyatyn chaineoperatoire and the series c. 1.000technologically meaningful pieces weresubjected to special analysis.

ANALOGIES

Technological and typological featuresof the assemblage under discussion allow toregard it as Levallois. The most close analogiescan be seen in the materials of multilayeredDniester Middle Palaeolithic sites Molodova Iand V (CHERNYSH 1982, 1987), as well asKetrosy (ANISYUTKIN 1981) and Korolevo,Hid layer (GLADILEN 1989). Basically similarindustry was recently discovered in theCrimea (KOLOSOV 1994, KOLOSOV &STEPANCHUK 1997).

TECHNOLOGICAL APPROACHES

Raw materials reduction sequence : generalconsiderations

A raw materials piece through itsexploitation as core usually passes severaldistinct - and generally cyclical - stages,namely the stage of primary selection [= rawmaterial stage]; the stage of preparation of acore [= pre-core stage], the stage of utilisationof a core [= core stage ], the stage of re-preparation of a core [= analogy of pre-corestage], the stage of abandonment of exhaustedcore [= post-core stage]. Basing on the chaineoperatoire approach (GENESTE 1985, 1988;BOEDA et. al., 1990) following conventionalstages or phases could be traced in thematerials of the site of Pronyatyn. ^

PHASE 0 "testing of raw materials". Thestage of primary selection is characterised bypresence of cortical flakes and unworkedpieces of raw materials. Assemblage underdiscussion provides virtually no specimens of

unworked pieces of raw materials, hence,initial phase of raw materials utilisation issuggested to be performed somewhere out ofthe site, probably, immediately on outcrops.The initialisation of raw materials processingdirectly at the site cannot be absolutelyexcluded but seems being conducted here in avery limited volumes.

PHASE I "initial preparation of workingsurface and working platform of core" [the phaseof "pre-core"]. Cortical and partly corticalflakes and debris are referred to this phase.They are rather frequent and constitute about50 % of the total number of flakes and debris.Even this stage produces comparativelynumerous bladey flakes and rough blades(graph 1). Worthy to note that c. 60 % of flakesmorphologically are very similar; typicalspecimens are presented at fig. 1:2-5; 6:4.

Parallel dorsal pattern is evidentlypredominate among other (table 1; graph 9).Special attention to careful curation ofproperties of removals is documented bycomparatively high indices of facetting strictand trimming (graph 2-4 and table 2). Specificevidence witness for the prevalence of hardhammer.

PHASE II "reduction of prepared workingsurface and striking platform of core" isrepresented by cores at different stages ofexhausting and by numerous flake products(fig. 2; 3; 4) comparable in number with wasteproducts of the preceding phase. Thisapproximate quantitative parity probablypoints to large initial dimensions of processedpieces of raw materials. Reduction was aimedto below described "ideal blank", but in realitythe number of desirable elongated flakes andblades is surprisingly fall in comparison withphase I (graph 1). Dorsal pattern statistics alsochanges. Centripetal items are especiallycharacteristical for this phase (table 1; graph 9),although artifacts with parallel scars at dorsalsurface are still numerous, as well. The level ofcuration of the properties of removals isbecome higher as it seen from schifts ofrelevant indices (graph 2-4 and table 2). Hardhammer is still prevailed during this phase.

PHASE III "re-preparation of workingsurfaces and striking platforms". As it followsfrom definition, this phase resulted in specificdebordant products and hardly recognisedtiny wastes. There are different types ofdebordant flakes, i.e. "classical" generally

34

THE CHAINES OPERATOIRES OF LEVALLOIS SITE PRONYATYN, WEST UKRAINE

elongated items (fig. 3 : 1; 5 : 5), shortdebordant or "debordant a dos limite" (fig. 5 :4). To the most characteristical feature theprevalence of products with perpendiculardorsal pattern can be referred (table 1). Thereare certain sufficient schifts in proportions ofproducts (graph 1). A hard hammerstone stillprevailed at this phase, as well.

PHASE IV "retouching of flakes andresharpening of already retouched pieces". Theassemblage of Pronyatyn produces virtuallyno typical examples of such wastes. It can notbe so surprisingly if we considered the generalcharacter of secondary working characterisedby light retouch regularly resulted in tinychips that are partly absent in collection andpartly has no clearly recognisable typicalfeatures. As it witnessed by available evidence,retouching was carried out by flint retouchersand stone hammerstones.

Types of initial raw materials

Several major types of initial pieces ofraw materials designated for knapping can bedistinguished, i.e. : natural untouched flintnodules, natural fragment of nodules, artificialfragments of nodules, natural flakes,decorticage flakes, ordinary massive flakes.Certain less serial objects, such as fragments ofcores, destroyed hammerstones etc were alsoused occasionally.

Available series of cores in certaincases allow us to recognise the initial type ofraw materials. Of course, the proposeddefinitions are not pretended to be absolutelydoubtless, nevertheless it more or lessobjectively illustrate the composition of initialcore blanks (table 3). Besides, several flakeshas clearly distinguishable traces, which allowus to suggest knapping of at least two differentnodules initially utilised as hammerstones (fig.4:4).

The cores: general regularities

The series of cores of the site ofPronyatyn is generally consists of non-volumetric artifacts oriented on utilisation offlat surface (fig. 8; 9; 10 : 1,2). Only few corescan be regarded as volumetric ones (fig. 10 :3,4). Besides, the overwhelming majority of socalled "flat" cores has only one workingsurface (cf. table 4; 5). At the same time, thediscussed series is often presented byexhausted artifacts, which can be defined as

post-cores. Among other, the latter statementwell argued by the sharp discordance betweenthe size of available flake products of theassemblage and dimensions of scars onworking surfaces of core being presented inour disposition. There is virtually no evidenceof pre-cores and pieces of raw materials in thevery beginning of its exploitation.

Analysis of systems of removals onworking surfaces of cores allow us to recogniseseveral different schemes of knapping aimedto the obtainment of blank and documented bymore or less representative series of cores.There are following items, i.e. : centripetal,parallel uni- and bi-directional, convergent,perpendicular, semi-crossed, and crossed, alloriented to recurrent production. There alsoare frequent examples of application of thescheme oriented to obtainment of preferentialor single removal in each distinct pre-core/corecycle of core reduction (table 6). Three groupsof cores are obviously dominate, i.e.centripetal, parallel and preferential. Othergroups are less frequent and - in principle - aregravitate toward the two main abovementioned ones, i.e. centripetal and parallel.

Table 6 provides certain needed dataconcerned the main parameters and tendenciestraced in the series of Pronyatyn cores. Thereare no crucial difference between dimensions,prevailed type of initial blank of comparedgroups of cores. More perceptible schifts aretraced through such indices as intensity ofcuration of working platforms and workingsurfaces, proportions of common type ofobtained flakes, productivity of cores. In thisrespect, the series of Pronyatyn cores providespotentially significant evidence, but its furtherdiscussion needs in broader comparable data,originated from at least one more regionalMiddle Palaeolithic site referred to the sametechnological facie.

Rather interesting data is obtained dueto reconstruction of the order of detachments offlakes. Generally speaking, there are two typesof records, i.e. a) working surfaces yieldedrandom, irregular, turning order of removals,and b) working surfaces provided successiveregular order of obtainment of flakey blanks(table 7). The data provided by centripetalartifacts and kindred groups of crossed andsemi-crossed items are especially complicateand various. There are certain - as it seemed -regularities, but once again, this kind of data istoo unusual and needs in accumulation of new

35

Vadim STEPANCHUK, Olexander SYTNYK

correspondent records for comparison andfurther discussion (cf.: table 8).

Modes ofcuration of working surface

The profile of working surfaces ofartifacts designated to further knapping wascurate by: a) usually elongated removals withspecific "debordant" type backs which wereoriented alongside the edge of core (fig. 3 :1; 5: 5); b) similar items but with cortical backs(fig. 1 : 1; 6 : 2,5; 7 : 1); c) comparatively shortand massive - especially in its basal part -flakes, removed from the edge of core towardits centre (fig. 3 : 4). The latter variation isrepresented sometimes by the series of smallflakes/ chips and constitute specific coretrimming (fig. 8 : 2,3) resulted in micro-debordant wastes. This, if to say, coretrimming is cardinally differ by itstechnological objectives from the method oftrimming of fringe zones of striking platformsof cores which is frequently traced on the areasbordering to butts of flakes. The first kind ofcore treatment concentrates on the generalproblem of shaping of the profile of theworking surface. Meantime, the second kind oftreatment evidently was served as good toolfor careful curation of properties of eachdistinct detached flake.

There also are several examples ofspecific practice of lowering of profile ofworking surface of core by series of blows forwhich the areas disposed immediately on thesurface of working plane were served asstriking platforms (fig. 7 : 2,4). Ratherexceptional examples of application of thistrick can be explained most likely in terms ofrarely aroused technological situation whenthis method was needed.

Modes ofcuration of striking platforms

The volumes of prepared - in that orthose way - striking platforms are ratherenormously high at the each stage ofutilisation of core and ranged between 50 - 87% from the number of determinable flakeproducts (table 2). Such a level of specialpreparation is sufficient and more advancedthan among many other sensu lato comparableMiddle Palaeolithic assemblages. It is worthyto note that if the meanings of indices of wellrecognisable facetted butts - being somewherec. 20 - 30 % - are more or less similar tocorrespondent indices provided by the otherEast European Middle Palaeolithic Levallois

related assemblages, then the quota oftrimmed flakes is evidently andextraordinarily higher and, hence, can beregarded as an characteristical feature of theindustry under discussion. Must be stressed,that trimmed products reach up to 49.2 (!)among flakes originated during exploitationphase II (table 2; graph 2-4; FIG 1 :2,5; 3 :4,6; 6: 2), whilst the technologically similar industryof Alyoshin Grot from the Crimea provides atthis phase only 23.5 % of trimmed products,i.e. even lower rate than the quota atPronyatyn phase I (= 28.2).

Technological reasons of retouching ofstriking platforms or trimming of the areasbordered with the fringe of striking platformsare satisfactory described (cf. :NEKHOROSHEV 1993; GIRIA &NEKHOROSHEV 1993; GIRIA 1997) andexplained in terms of intentional efforts aimedto as careful as possible curation of propertiesof detached removals. Must be stressed, thatthe trimming normally associated with plainstriking platform is usually understand asmore typical for the Upper Palaeolithictechnology; however, the assemblage ofPronyatyn add a little to the benefit of thisnotion. Some other artifacts might be formallyreferred as atypical avivages fig. 7:5 but moreplausibly can be explained coming from thespecificity of one of the Pronyatyn chaineoperatoire that will be shown later.

As exceptional might be regardedthree or four artifacts with evidence of roughabrasion of the edge between fringe zones ofstriking platform and working surface (fig. 2 :1; 4 : 2). Two more flakes provide amazinglycrude and careless treatment of this zone byseries of blows oriented approximately alongthe bissectriss of the angle between the planesof striking platform and working surface.

All in all, following modes of curationof striking platforms were applied, namely :prevailed trimming, well represented facetting(fine and rough), and very weak and doubtfulevidence of abrasion (?) and picketage (?).

Parameters of "ideal" blank

The concept of "ideal" blank can befruitfully applied to the analysis of assemblageunder discussion. This concept foresees therestoration of appearance of blanks whichwere most frequently used by ancienttoolmakers for the manufacture of secondary

36

THE CHAINES OPERATOIRES OF LEVALLOIS SITE PRONYATYN, WEST UKRAINE

working pieces, i.e. so called "tools".Generalised parameters of the most commontool blank a priori are supposed to be close, ifnot identical, to the parameters oftechnological aim of the process of modifyingof raw materials. This technological aim mightbe elucidated in the case of successivecomparative analysis of both series ofretouched pieces and unworked products.Such analysis foresees comparison through arow of different indices, e.g. raw materials,dimensions, proportion, shape, type of dorsalpattern, profile etc (STEPANCHUK 1995;KOLOSOV & STEPANCHUK 1997).Evidently, certain of these parameters directlypoint on properties of the most desirableblank, whilst the other indicate specialtechnological modes which were used forobtainment of blanks. As a result, theaccumulation of data on restored tendencies ofselection of flakes for next retouching or usewill provide needed information and thegeneral characteristic of "ideal blank" can bedrawn. The so called Index of Selection iswidely used and very helpful in this case (theway of calculation see KOLOSOV &STEPANCHUK 1997).

The data at hands allow us to statethat the most frequently used flake-blank ofPronyatyn assemblage can be defined as alarge, more than 6 cm, ovalaire or rectangular inshape blade or bladey flake, not cortical, usuallywith centripetal or perpendicular dorsal patternand finely prepared butt, either facetted or trimmed.Special calculations also point to conclusionabout predominant selection of blanks whichare straight in profile, symmetrical and notmassive in crossed-section. Graphs 5-11illustrate certain parameters of flake products,the dynamic of these parameters throughdifferent stages of raw materials reduction andtendencies of selection of flakes for futureretouching or use in respect of that or thoseparameter.

Industry of Pronyatyn was orientedexclusively to obtainment of blanks throughcore reduction. There are no any evidencewitnessed for familiarising of Pronyatyntoolmakers with the technique of bifaciallyprepared blank.

SUPPORTING INSTRUMENTS AND THEMOST RELIABLE MODE OF KNAPPING

Hard versus soft hammerstone technique

There are various data leaded to theassumptions about the intensive use of hardhammerstones. Firstly, it is very frequenttraces of cones of blows easily recognised onthe surface of flake product butts. Thanks tothe properties of the local fine grained flints,this kind of traces is well distinguishable andexpressive (fig. 1 : 1; 3 : 6; 4 : 4; 6 : 1). Onequarter to one third of examined flakesprovide this kind of evidence. It is interestingto note that only one tenth of cores yielded thesimilar traces. The frequency of advancedbulbs, the profile of ventral surface on certainflakes, and correspondent data provided bycores, - all suggests the knapping wasgenerally carried out by hard hammerstones.This assumption, finally, is strengthened byreal presence of such artifacts in the collectionof the assemblage under discussion.

These latter artifacts are presented bysingle complete and several fragments of flinthammerstones. Not destroyed artifactconstitutes regular ovoid nodule 74 X 66 X 53mm with typical damage concentrated withinarea 30 X 18 mm near the conical end of theobject. Fragmented hammerstones arepresented by irregular pieces yieldedcomparable surface damages. Worthy to notethe presence of numerous scratches on thecortex of the unbroken hammerstone. Thesescratches are not underlie or overlapped withdamage of hammerstone type and obviouslyhas different reason. This reason is still unclearas well as purpose of similar scratches tracedon cortex of certain cores and flakey products(fig. 6:4).

Conical fracture initiations (GIRIA1997) are crucially dominate among flakeseries of the Pronyatyn. Nevertheless, therealso are several examples of not conicalinitiation. Besides, the presence of lipscoincided with conical initiation, althoughvery weak, also was evaluated as sign of softhammer technique. Only these evidences arepointed to the date to the possibility of use ofsoft hammerstones. Anyway, this evidence isweak and seriously inferior to hard hammerrecords (table 2) and general conclusion mightbe stated about the predomination of hardhammerstone technique of knapping during

37

Vadim STEPANCHUK, Olexander SYTNYK

all the three first main stages of raw materialsprocessing.

Retouching implements

Assemblage of the site of Pronyatynyielded several retouchers. Flakes wereusually served as blanks for this kind of tools.Our observations under similar implementsoriginated from a row of Crimean MiddlePalaeolithic sites unanimously witness thattypical "retoucher type" damage regularlycoincided with secondarily worked orintensively used flake products. Pronyatynseries is not out of this rule, as well, as soon as10 of 12 known retouchers are made onmodified pieces. This kind of regularity mightbe regarded as additional data to the benefit ofsituational mode of manufacture ofsecondarily worked pieces. The last twoPronyatyn specimens were prepared by coreand primary flake (fig. 3 :6; 10:2).

Mode of knapping

As the most reliable and widely usedthe mode of free knapping in hands usingindirect percussion can be suggested.Herewith, there are certain evidence indicatedthe limited application of anvil duringflintknapping activity. This mode can besuggested by the presence of specific tracesvisible on the projected areas of back sides ofnine flat cores (fig. 9 : 3) and associated withstage of pre-cOre preparation. These traces areof "hammerstone type", although lesspronounced, and likely resulted due to contactwith hard substance.

RESTORATION AND DESCRIPTION OFREDUCTION SEQUENCE

In the most general form the rawmaterials reduction sequence recognised at thesite of Pronyatyn can be described as follows.

Stage of initialisation of core

Preferable initial pieces of rawmaterials destined for further knapping, as itwas shown above, constitute the nodules oflocal cretaceous flint, as well as massiveprimary - natural or artificial - flakes. The mostpreferable shape of crude piece of rawmaterials is ovoid flat-surface pebble-like flintnodule. If there were protuberances, they were

removed beforehand and this action hadproduced primary flakes in its turn constitutepre-forms for further knapping (fig. 11 : B,O.Then the plain striking platform wasprepared by single blow (or series of blows)either removed the conical extremity of nodule(fig. 11 : B) or splintered the latter alongside itslength (fig. 11 : C). Thus, there are two majortypes of core blanks, i.e. nodule, either slightlymodified or not, and primary flake, eitherartificial or natural. Hard hammer technique iscrucially prevailed.

Stage of preparation of core

The next steps of dealing with alreadyinitialised piece of raw materials might bedifferent.

Variety 1. Initialised nodule might bealmost completely decorticated by a series ofsuccessive blows used already existing plainstriking platform (fig. 12 : A). This action hadgenerally produced elongated flakes andbladey flakes in different extent covered bycortex (fig. 12 : A). Terminations of scars ofsuch decorticage flakes are well recognised onmany of cores presented in assemblage (fig. 9 :1-3), as well as on butt areas of flakes obtainedduring subsequent reduction (fig. 6 :1,6; 7: 5).Just during this stage of decorticating the anvilcould be used. Hard hammer technique ispredominate. Properties of detachments werecurate mainly by trimming. In fact, processedpiece of raw materials to the end of this stageconstitutes real pre-core and looks like crudeprismatic core (fig. 12 : A).

Variety 2. Initialised nodule might benot decorticated through the full or almost fullperimeter of plain striking platform butmanufacture of pre-core might be restricted tothe preparation of opposed striking platformsand, sometimes, additional preparation offlaking surface (fig. 12 : B). In this case waste-products constitute certain number ofelongated primary flakes similar to onesappeared during the above described mode ofcuration of raw materials, and specificelongate debordant flakes with natural "backs"(fig.l:l;6:2A7:l;12:B).

Variety 3. Massive primary flakeseither natural or artificial were prepared forforthcoming detachment of blanks bypreparation of striking platform (fig. 8 :1,2; 12: C) and, sometimes, by special core trimming(fig. 8 : 2,3; 12 : C) that modify the profile of

38

THE CHAINES OPERATOIRES OF LEVALLOIS SITE PRONYATYN, WEST UKRAINE

working surface (in this case the ventralsurface of flake).

Thus, as it likely followed by theanalysis of technologically meaningful flakesand other available artifacts, there were threekinds of pre-cores, which further exploitationalso has certain differences.

Stages of utilisation and re-preparation of a core

Variety 1. The first type of pre-core, ifto say, "prismatic type", allows to utilise bothvolume and surface. Let us to mark them,correspondingly, as l.A) and l.B). As data athand witnessed, Pronyatyn toolmakers weremostly preferred the second mode (l.B).

1. A) Volumetric mode of utilisation ofraw materials is presented in assemblage byseveral fragments of crude prismatic cores (fig.10 : 3,4). These records are characterised byplain or prepared striking platform andparallel sided bladey scars of removals. Thereis the specimen of bi-directionally orientedparallel scars that assumes presence of coreswith opposite striking platforms. Curation ofproperties of detached products wasmaintained mainly by trimming. As the mainstriking platform, the plain surface, alreadyachieved during the pre-core stage of rawmaterials managing was used. Hard hammertechnique, probably, was prevailed. Productsof this mode obligatory look as regular bladeyflake or blades and are presented inassemblage of the site (fig. 4 : 5,6; 14 : A).Several flakes can be regarded as by-productof re-juvination of striking platforms (cf. fig. 5 :3).

1. B) Non-volumetric mode, in thecontrary, was concentrate on utilisation - as aworking surface - of that one, which served involumetric mode as striking platform (fig. 13 :A). As it is argued by available flat cores andstatistics on dorsal pattern of flake products,there were different, that we call, schemes ofutilisation of working or flaking surface.Among these schemes are centripetal, paralleluni- and bi-directional, convergent,perpendicular, semi-crossed, crossed andoriented to obtainment of sole end product(single removal scheme). Were or not theabove mentioned recurrent schemes speciallyelaborated and maintained throughout the fullcycle of a single raw materials piece reductionit is not clear yet. In any case, to the date thereis no solid grounds to regard all these schemes

only as examples of different ways ofpreparation of convexity of working surfacefor completing the last blow aimed toobtainment of sole blank. To the contrary, weinclined to suggest independent existence of arow of different and distinct schemes ofutilisation of flaking surface. Keeping in mind,as it was shown above, the high level of specialpreparation of striking platforms and frequentrepreparation of working surfaces (resulted incomparatively frequent debordant flakesgrtavitated toward classical type (cf. fig. 5 :5; 6: 1,6) all these schemes are potentially able toproduce high quality regular elongated blanks,which could even formally be described asLevallois ones. Hard hammer technique stillprevailed during the treatment.

Variety 2. The second type of pre-corealso allows, in principle, the utilisation of bothvolume and surface.

2. A) As soon as we have at handsonly exhausted forms of volumetric cores,there is no convincible data witness to thebenefit of real existence of such cores. By theother hand, potentially prognosed flakeswhich could be obtained during 2.A) varietyexploitation are available in collection, i.e.regular elongated products with parallel/subparallel dorsal pattern and several types ofbladey products with cortex (fig. 4 :5,6). At thesame time these products are indistinguishablefrom by- and end-products of volumetric corevariety l.A) reduction or even from principallydifferent stage of raw materials processing(e.g. decorticage of pre-core variety 1).

2. B) This variety of cores, is in manyways similar to variety 1. B) save for morelarge and elongate proportion, and theapplication of the same schemes of knappingcould be suggested (fig. 13 : B). However, thestudy of series of flakes of phase HI (curationof working surface) reveals presence of specificand numerous debordant blades and bladeyflakes (fig. 1:1; 6 :2,5) which allow to supposethe prevalence of parallel and perpendicular(orthogonal) scheme. "Single blow" schemewas also applied, as it argued by presence oflarge blanks with signs of Levallois typepreparation (fig. 2; 3).

Variety 3. The third type of pre-coresallows only the utilisation of surface (fig.14: B).

39

Vadim STEPANCHUK, Olexander SYTNYK

To judge by available cores, there wereapplied following schemes, i.e. "single",parallel, perpendicular. In spite of rathernumerous cores, flakey end-products of theirutilisation are rather rare in collection at hands(fig. 4 : 3). Hard hammerstone technique waspreferable. This variety of raw materialsutilisation resemble the so called Kombewatechnique but cannot be regarded as typicalexample of this latter.

Stage of abandonment of exhausted core

Practically all available cores of theassemblage under discussion constituteexhausted artifacts, immediately related to thepost-core stage. Once again seems to bestrange and hardly explained the presence ofdiminutive cores, negatives of the lastremovals on which are far smaller thanaverage dimensions of flakes with retouch ortraces of use.

Nothing can be stated about theappearance of volumetric post-cores, as soonas they are represented only by fragmentedartifacts.

About one third of flat, or non-volumetric, cores represents near end stage ofutilisation of cores of variety l.B) (fig. 9 : 1-3).The latter part of small sized and thin coresrepresents end stage of exhausting of cores ofvarieties l.B, 2.8, and, probably 3 (fig. 10:1,2).

Two sided flat cores, also presentingthe stage of abandoned nucleus, constitutesomething special, inasmuch as it is not cleareither they represent the cores which from thevery beginning were possessed two oppositeworking surfaces, or became two sided only atthe stage close to the stage of post-core. At thesame time, there is evidence pointed tomanufacture of flat cores with two oppositeworking surfaces prepared on pre-cores ofvariety 1 (fig. 5 :5).

DISCUSSION

Assemblage under discussion yieldedcomplicate records in respect of applied corereduction strategies and we have tried to showthe variability of used strategies through thestressing out the presence of different schemesof knapping. Special analysis in frames of socalled "ideal blank concept" reveal the

Levallois appearance of desirable blankselected for further retouching or use. Must bestressed the obvious orientation of Pronyatyntechnology toward obtainment of blades andbladey flakes. General regularities of appliedtechnology clearly show powerful tendencytoward careful preparation of strikingplatforms and flaking surfaces and thesefeatures can again be regarded as signs ofLevallois technology. To the number ofspecific peculiarities of the assemblage underdiscussion following issues are refereed,namely : specific way of preparation of pre-core variety 1 forms (so-called "prismatictype"); co-existence of both volumetric(although comparatively rare) and notvolumetric ways of raw materials reduction;extremely high level of trimming technique,that alongside with facetting technique allowto curate properties of detached flakes andusually considered as Upper Palaeolithicinnovation; high variability of strategiesallowed to obtain desirable bladey products.

The certain tendency to use massiveprimary flakes as core blanks, can also beregarded as rather specifically. Ventral surfacewhich was used as flaking one had neededminimal additional preparation, but, instead,the productivity of such core was cruciallyrestricted due to its comparatively smallproportions. We do not regarded such cores asevidence of Kombewa-related technique, andrefereed it - in depend of dorsal pattern - asrecurrent or "single" removal techniqueevidence.

To the date there is no possibility ofexact reconstruction of each distinct recurrentreduction sequence, we mean different kindsof parallel and crossed knapping, thenconvergent, perpendicular etc. This task needsin accumulation of new factual data and,consequently, special refitting program. At thesame time, due to presence of characteristiccores and specific by- and end-products,following variations of recurrent corereduction strategies could be distinguished, i.e.centripetal, parallel uni- and bi-directional,convergent, perpendicular, semi-crossed andcrossed. Each variation embrace stages ofspecial preparation of striking platform(s) andworking surface. As it was already stressed,this high variability calls several questions andthe most important among them is whethereach of these schemes was steadily maintainedduring the reduction of a single raw materials,or they were highly interchanged and/or

40

THE CHAINES OPERATOIRES OF LEVALLOIS SITE PRONYATYN, WEST UKRAINE

simply supported "single" removal strategy.Anyway, Pronyatyn records provide such adoubtless kind of data, which to the date canbe evaluated as evidence of intensiveapplication of a row of recurrent knappingmethods, usually regarded as Levallois ones(BOEDA et.al. 1990; CHABAY & SITLIVY1994).

Currently unanimously regarded asLevallois, the method of "single" or"preferential" removal was also known toPronyatyn toolmakers, and this records arewell illustrated by representative series ofcores and flakey products.

Alongside with recurrent and "single"removal techniques based upon exploitation ofsurface, there is data of utilisation of volume.Complete forms of cores or post-cores are notavailable, but fragments of cores at hands, aswell as flake products, both suggest theexistence of crude "prismatic" nuclei,presenting consequent stage of pre-corevariety 1 utilisation. Flakes and fragment ofcores show intensive traces of trimmingtechnique and signs of hard hammer.Acquaintance of Pronyatyn toolmakers withthis mode of raw materials reduction and,nevertheless, their concentration on "surfacemode" of exploitation of flat cores, remarkablyclear show their orientation toward specificdesirable blank - rather wide and slightlyelongated blade/ bladey flake - which differsufficiently from regular end-productsresulted in the case of volumetric reduction.

Thus, the analysis of Pronyatynassemblage allowed to elucidate highlycomplex system of different core reductionstrategies, which were applied at the site. Theoverwhelming majority of knapped productswere obtained through the exploitation of flatcores, although the idea of volume was alsoknown and time to time was applied. Carefulcuration of properties of removals by differentmethods of preparation of striking platformzone and of maintenance of flaking surfaceconvexity was conducted during each stage ofcore reduction. Specific forms of pre-core stagetreatment can be distinguished. There isevidence of a row of recurrent methods,frequently regarded as Levallois. ClassicalLevallois technique oriented to obtainment ofa single removal through pre-core-core cycle isalso represented.

ACKNOWLEDGEMENTS

The presented work was prepared inframes and with financial support of currentproject INTAS-96-079. We are grateful to T. M.Kovalchuk, Ternopil Museum, for kindpromotion in treatment of materials of the siteof Pronyatyn.

REFERENCES

ANISYUTKIN N.K., 1981,Archaeological study of the Mousteriansite of Ketrosy In : N. D. Praslov (ed.)Ketrosy. Mousterian site in the MiddleDniester. Moscow, "Nauka". (In Russian).

BOEDA E., GENESTE J.-M. et MEIGNEN L.,1990,

Identification de chaines operatoireslithiques du Paleolithique ancien et moyen.Paleo n° 2 :43-80.

CHABAY V. & SITLIVY V., 1994,The periodization of core reductionstrategies of the Ancient, Lower andMiddle Palaeolithic PrehistoireEuropeenne, vol. 5:9-88.

CHERNYSH A.P., 1982,Multilayered Palaeolithic site Molodova IIn : G. I. Goretsky & I. K. Ivanova (eds)Molodova I. Unique Mousterian settlementin the Middle Dniester. Moscow, "Nauka".(In Russian).

CHERNYSH A.P.., 1987,The standard multilayered site ofMolodova V. Archaeology. In : I. K.Ivanova & S. M. Tzeitlin (eds) Themultilayered Paleolithic Site Molodova V.Moscow, "Nauka". (In Russian).

GENESTE J.-M., 1985,Analyse lithique d"industries mousterienesdu Perigord : une approche technologiquedu comportement des groupes humainesau Paleolithique moyen. These doctoral,Universite de Bordeaux.

GIRIA E.Yu., 1997,The technological analysis of stoneindustries. S-Petersbourg, "AkademPrint".(In Russian).

41

Vadim STEPANCHUK, Olexander SYTNYK

GIRIA E.Yu. & NEKHOROSHEV P.E., 1993,Some technological criteria of the stone ageperiodisation. Rossijskaya Arkheologiya n°4 :5-24. (In Russian).

GLADILINV.N.,1989,The problems of the Early Palaeolithic.These doctorat. Novosibirsk. (In Russian).

KOLOSOV Yu.G., 1994,Alyoshin Grot - new Mousterian site inCrimea (preliminary report).Arkheologicheski Al'manakh, Donetsk 4 :75-80. (In Russian).

KOLOSOV Yu. & STEPANCHUK V., 1997,New type of Middle Palaeolithic Industryin Eastern Crimea. Archeologicke RozhledyXLIX: 124-145.

NEKHOROSHEV P.E., 1993,On the methods of study of the LowerPalaeolithic stone flaking technique andtechnology. Rossijskaya Arkheologiya n° 3= 100-119. (In Russian).

STEPANCHUK V.N., 1995,The dependence of technology andtypology of Middle Palaeolithic industriesfrom the properties of used stone rawmaterials (An example of East Crimeansites Prolom I, Prolom II, and Sary-Kaya).Scientific Archives of Institute ofArchaeology of NUAS, Kyiv.dn Russian).

SYTNYK O.S., 1978,Engraved picture on bone from theMousterian site of Pronyatyn nearTernopol In : R. S. Vasilievski (ed.) Plastikai risunki drevnikh kul'tur. Novosibirsk: 39-46. (In Russian).

SYTNYK O.S., 1985,Mousterian site of Pronyatyn nearTernopil. Archeologiya n° 50. (InUkrainian).

SYTNYK O.S., 1994,Mousterian site of Pronyatyn and its placein the Palaeolithic of Eastern Europe.Arkheologicheski Al'manakh, Donetsk 3 :101-20. (In Russian).

42

TABLE 1. Pronyatyn. Principal types of dorsal pattern of flakes. Bold figures point to themost typical products for a given stage of core reduction.

DORSALPATTERNparallel*perpendicularcentripetalcorticalTOTAL

PHASE I

167661620269

62.1%24.5%5.9%7.5%

100%

PHASE II

1416453

258

54.7%24.8%20.5%

100%

PHASE III

22283

53

41.5%52.8%5.7%

100%

*;including parallel, subparallel, convergent, divergent uni- and bi-directional pcs

TABLE 2. Pronyatyn. The dynamic of certain technical indices trough the mainstages of core reduction sequence. Data achieved by flake productsanalysis. Bold figure points to the highest meaning of given index.

phase Iphase IIphase III

the type ofhammerstone

hard

25.527.733.3

soft

6.68.85.2

the main kinds ofstriking platform

preparationfacetted

22.7

37.731.6

trimmed

28.249.226.3

the main schemes ofknapping

parallel

62.154.741.5

perpen-dicular

24.524.852.8

centri-petal5.9

20.55.7

43

TABLE 3. Pronyatyn. Types of initial blanks of cores

CORES: THE TYPE OF BLANKprimary cortical flakemassive flakenatural flakenatural ? artificial ? fragment of noduleuntouched ? nodulefragment of core

N9348111

TABLE 4. Pronyatyn. Flat cores with two working surfaces.Combination of schemes of knapping of alternatesurfaces.

Flat cores with two alternate working surfaces: the types ofcombination of schemes of knapping

Combinationparallel - centripetalparallel - parallelperpendicular -convergentsingle -crossedsingle - parallelsemi-crossed - convergentsemi-crossed - centripetalcentripetal - centripetal

TOTAL

N of cores32111111

11

N of surfaces64222222

22

44

TABLE 5. Pronyatyn. The main varieties of cores: utilisation of volume

T H E M A I N V A R I E T I E S O F P R O N Y A T Y N C O R E SA. u t i l i s a t i o n of s u r f a c e : "flat cores'1

A.1A.2

One Working SurfaceTwo Working Surfaces

N6111

%84.715.3

Total: 72 100B. u t i l i s a t i o n of v o l u m e : "volumetric cores"

B.I One Working Surface 5 100

72 pcs= 93.5 %

5 pcs= 6.5 %

TABLE 7. Pronyatyn. The main kinds of flat cores in respect of usedorder of detachments

TYPE OFSCHEME

centripetalparallel uni-directionalparallel bi-directionalconvergentperpendicularsemi-crossedcrossedTOTAL

irregular

166

5

231-

33

59.3 %85.7 %

71.4 %

50%60%33%

058.9 %

regular

111

2

222323

40.7 %14.3 %

28.6 %

50%40%

66.7 %100 %31.1 %

indeterminable

64

-

2---

12

45

TABLE 6. Pronyatyn. Flat cores. The main technical data. F = flake; BF = bladeyflake; B = blade.

SCHEME

centripetalparallel uni-directional

singleparallel bi-directionalconvergentperpendi-

cularsemi-crossed

crossed

TOTAL

Nofsurfac

es

3313

127

66

33

83

AV.DIMEN-

SION

56X52X18

55X53X26

60X49X1864X49X26

66X45X2455X51X18

78X58X1274X64X24

64X53X21

INITIAL BLANK

flake

34.640

7016.7

25

80

--

40.7

core

-

-

10-

--

--

1.7

nodule

65.460

1083.3

7520

10010057.6

WORKING PLATFORMS

un-prepa

red

12.9

61.5

-28.5

33.3-

66.733.3

61.5

roughlyprepared

45.223.1

-42.9

33.3

66.7

--

23.1

finelyprepared

41.915.4

10028.5

33.333.3

33.3

66.715.4

WORKINGSURFACE

trimming

66.761.5

58.3100

83.366.7

10066.789.2

debordant

-

7.7

41.7-

16.733.3

--

10.1

NOF

REMOVALS

8-9

3-4

35-6

64

8106

THE TYPE OFREMOVALS

F

78.846.2

91.785.7

10066.7

33.333.3

73.5

F/BF

21.223.1

8.314.3

-33.3

66.766.7

21.7

BF

-

23

--

--

--

3.6

BF/B

-

7.7

--

--

--

1.2

MAXIMAL AV.DIMENSIONS

OF REMOVALS

63X3553X37

60X4351X54

40X4853X38

40X3540X33

50X40

TABLE 8. Pronyatyn. The order of detachments on working surfaces of flat cores ofdifferent schemes of knapping. Figures and small letters mark the order of individualremovals; capital letters point to the order of series of removals.

crossed

semi-crossed

centripetal

parallel uni-directional

convergent

perpendi-cular

parallel bi-directional

*AD-» 4-C

tB

4>A/DB-» «-C

XT8»V7K6

!•» «-5lH K42 *3

1 2 3 4 ...XXXXX

1 2 3 4 ...

A**-B

*BfA

*AB-» «-D

*C

XAB/C-* 4-B/C

6«*7K85-» *•!431 R2

t3

...432 1^XXXg

*AD-» *-B

*r

...SI ...*...M «•! 4-...-H

...rr •!>..K +1

~5" *..**

...+1M *...+2

'* V£2?lt.. ^"^

R +•...+» t.-.+K-

-Ji*11 .̂̂3-» «•»+731 Rn

fl "

lS|*a+Ka+

!•* «•»+231 R2"+3

bf» *C+Kc+!>*-» «-c

b^l Ra+

*•» «•*

4-•» «•

***•» «•

91 R

xxxxx

^4,4,4,^

4-4-

4-t

Graph 1. Pronyatyn. Proportions of flakes appeared at different stages of core reduction.

Pronyatyn. Proportions of flakes appeared at differentstages of core reduction (%).

54.5•flakes

D bladey flakes

CD blades

phase I phase II phase IIIN=269 N=233 N=73

GeneralTendency

Graph 2. Pronyatyn. Some indices reflecting ways and degree of core striking platforms preparation.Series of "decorticage" flakes.

Pronyatyn. Some indices reflecting ways and degree of core striking platformspreparation. Series of "decorticage" flakes (N=231).

100

unprepared unprepared dihedral butts polihedral butts facetted butts punctifbrrn andcortical butts plain butts N=25 N=22 N=68 lineal butts

N=11 N=89 N=16

48

Graph 3. Pronyatyn. Some indices reflecting ways and degree of core striking platforms preparation.Series of "blank" flakes.

100

75

50

25

0

Pronyatyn. Some indices reflecting ways and degree of core striking platformspreparation. Series of "blank" flakes (N=260).

Duntrimmed

•trimmed

91,6

unprepared unprepared dihedral butts polihedral butts facetted butts punctiform andcortical butts plain butts N=23 N=57 N=95 lineal butts

N=0 N=74 N=11

Graph 4. Pronyatyn. Some indices reflecting ways and degree of core striking platforms preparation.Series of "curation" flakes.

Pronyatyn. Some indices reflecting ways and degree of core striking platformspreparation. Series of "curation" flakes (N=57).

94.7

unprepared unprepared dihedral polihedral facetted punctiformcortical plain butts butts N=4 butts N=3 butts N= 19 and lineal

butts N=3 N=25 butts N=3

49

Graph 5. Pronyatyn. Distribution of flake products of phase II according to its general shape.

Pronyatyn. Distribution of flake products of phase II according to its generalshape. Flake tools/ flakes N = 65/148.

rectangular ovalaire(115 pcs) (50 pcs)

triangular circular trapezoide(36 pcs) (7 pcs) (5 pcs)

Graph 6. Pronyatyn. Indices of selection of flake products of phase II for retouching accordingly to itsshape.

Eft

40 -

Qfl

onf\j •

10 -

0 -

Pronyatyn. Indices of selection of flake products of phase II for retouchingaccording to its shape (%). Flake tools/ flakes N=65/148.

42.8

34 32.2

22.2

circular ovalaire rectangular triangular(3 flake (17 flake (37 flake (8 flaketools; 4 tools; 33 tools; 78 tools; 36flakes) flakes) flakes) flakes)

0

trapezoide( 0 flaketools; 5flakes)

50

Graph 7. Pronyatyn. Indices of selection of flake products for use or retouching accordingly to itsplace in core reduction sequence.

Pronyatyn. Indices of selection of flake products for use or retouchingaccording to its place in core reduction sequence (%). Flake tools/flake

(FT/F) N= 138/437

35 T30-25-20-15-10-

0-

22.3

: ^"

phase I :decorticage

60 FT/ 209 F

31.8

.-\

phase II:blanks74 FT/ 1 59 F

\5.5*_ . n*k U

1 ' i 1 1

phase III: phase IV:repreparation retouching4 FT/ 69 F 0 FT/ 0 F

Graph 8. Pronyatyn. Indices of selection of flakes for next retouching accordingly to their proportionsand place in core reduction sequence.

Pronyatyn. Indices of selection of flakes for next retouching accordingto their proportions and place in core reduction sequence (%).

LJ flakes

• bladey flakes

• blades

phase)decorticage

N=269

phase IIblanks N=233

(A)

phase IIIrepreparation

N=73

GeneralTendency

(B)

51

Graph 9. Pronyatyn. Indices of selection of flakes for next retouching accordingly to their dorspatterns and place in core reduction sequence.

Pronyatyn. Indices of selection of flakes for next retouching according totheir dorsal patterns and place in core reduction sequence (%).

B parallel

• perpendicular

D centripetal

D cortical

phase I phase II

N=269 N=258

(A)

Graph 10. Pronyatyn. Indices of selection of "flake-blanks" according to its dimensions anproportions.

Pronyatyn. Indices of selection of "flake-blanks" according to itsdimensions and proportions. N=227.

bladeyflakes N=69

blades N=47

flakes N=120

4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12cm cm cm cm cm cm cm cm

N=48 N=56 N=53 N=41 N=18 N=7 N=2 N=2

Graph 11. Pronyatyn. Indices of selection of "flake-blanks" according to its dimensions : general view.

Pronyatyn. Indices of selection of "flake-blanks" according its dimensions:general view. N=227.

66.7

4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12cm cm cm cm cm cm cm cm

N=48 N=56 N=53 N=41 N=18 N=7 N=2 N=2

53

Fig. 1. Pronyatyn. Debordant and decorticage products: 1. Refitted debordants with cortical backs. Bi-polar parallel dorsal pattern, rough and fine facetting of butts. 2. Decorticage bladey flake. Paralleldorsal pattern, trimmed plain butt. 3. Decorticage bladey flake. Centripetal dorsal pattern, plain linearbutt. 4. Decorticage bladey flake. Parallel dorsal pattern, facetted butt. 5. Decorticage flake. Sub-parallel dorsal pattern, trimmed plain butt.

54

Fig. 2. Pronyatyn. Blank products : 1. Levallois bladey flake. Sub-parallel/ convergent dorsal pattern,facetted and abraded butt. 2. Levallois flake. Sub-perpendicular dorsal pattern, trimmed and facettedbutt. Trimming associates with roughly facetted area. 3. Levallois flake. Perpendicular dorsal pattern,facetted butt. Regularly changed direction of removals. 4. Levallois flake. Centripetal dorsal pattern,roughly facetted trimmed butt suggested utilisation of pre-core of variety 1. 5. Levallois pointedbladey flake. Bi-polar parallel dorsal pattern, roughly facetted butt.

55

Fig. 3. Pronyatyn. Blank and debordant products : 1. Levallois bladey flake. Bi-polar parallel dorsalpattern, roughly facetted butt. 2. Levallois flake. Centripetal dorsal pattern, facetted butt. 3. Backedknife on Levallois blade. Sub-parallel/ convergent dorsal pattern, facetted butt. 4. Levallois pointedflake. Convergent dorsal pattern, trimmed dihedral butt. Debordant (?) versus blank product ofutilisation of volumetric core (?). 5. Bladey flake. Sub-parallel/ convergent dorsal pattern, facetted andtrimmed butt. Trimming associates with roughly facetted area. 6. Fragmented side-scraper onLevallois blank. Retoucher type damage on surface of bulb of percussion. Sub-perpendicular dorsalpattern, roughly facetted trimmed butt, suggested utilisation of pre-core of variety 1.

56

Fig. 4. Pronyatyn. Blank and decorticage products : 1. Levallois flake. Convergent dorsal pattern,facetted butt. 2. Levallois flake. Parallel dorsal pattern, roughly facetted trimmed and abraded butt. 3.Levallois bladey flake of Kombewa type. Convergent dorsal pattern, facetted butt. 4. Decorticagebladey flake. Parallel dorsal pattern, plain butt. Proximal ventral surface show evidence of previousutilisation of nodule as hammerstone. 5. Decorticage/ blank blade. Parallel dorsal pattern, facettedbutt. Product of volumetric core utilisation. 6. Decorticage/ blank blade. Parallel dorsal pattern, plainbutt. Product of volumetric core utilisation (?).

57

Fig. 5. Pronyatyn. Debordant products : 1. Bladey flake. Sub-perpendicular dorsal pattern, slightlytrimmed plain butt. 2. Flake. Semi-crossed dorsal pattern, roughly facetted slightly trimmed but. 3.Bladey flake. Plain dorsal pattern, cortical lineal butt. Product of initial phase of preparadon ofworking surface on pre-core variety 1 (?) versus re-animadon of volumetric core striking platformarea (?). 4. Flake. Orthogonal dorsal pattern, punctiform butt. 5. Blade. Parallel dorsal pattern, facettedand trimmed butt. Trimming associates with plain area of butt. Scars on back suggest morecomplicate case of udlisation of pre-core of variety 1 as it shown at FIG 13 : A and 14 : A. 6.Fragmented blade. Orthogonal dorsal pattern.

58

- 2

Fig. 6. Pronyatyn. Debordant and decorticage products : 1. Bladey flake. Sub-parallel dorsal pattern,roughly prepared trimmed butt. Scars on back suggest utilisation of pre-core of variety 1. 2. Bladeyflake. Bi-polar parallel dorsal pattern, trimmed plain butt. Cortical back suggests utilisation of pre-core of variety 2. 3. Flake. Sub-perpendicular dorsal pattern, facetted butt. Product of re-preparationof exhausted Levallois core of "preferential" type (?). 4. Decorticage flake with secondarily workedbutt area. Regular groups of scratches are disposed on dorsal surface. 5. Blade. Bi-polar parallel dorsalpattern, facetted butt. Cortical back suggests utilisation of pre-core of variety 2. 6. Bladey flake.Perpendicular dorsal pattern, roughly prepared trimmed butt and opposite platform. Scars on backsuggest utilisation of pre-core of variety 1.

59

Fig. 7. Pronyatyn. Decorticage, debordant and blank products : 1. Debordant (?) flake. Convergentdorsal pattern, cortical trimmed (?) butt. 2. Decorticage bladey flake. Sub-parallel dorsal pattern,dihedral butt. Evidence of specific method of curation of properties of flaking surface (area ofinitialisation of technological flakes is marked by circle). 3. Blank flake. Centripetal dorsal pattern,roughly prepared butt. 4. Unsuccessful blank flake. Semi-crossed dorsal pattern, facetted butt.Evidence of specific method of curation of properties of flaking surface (area of initialisation oftechnological flakes is marked by circle). 5. Flake. Centripetal dorsal pattern, roughly prepared butt.Avivage-like product of utilisation of pre-core of variety 1.

60

Fig. 8. Pronyatyn. Flat cores: 1. "Single" removal or "preferential" Levallois core on primary flake. Onefacetted striking platform. 2. Recurrent core of perpendicular scheme on primary flake, (a) points toremains of original ventral surface of flake. "Core trimming" technique evidence, three roughlyfacetted striking platforms. 3. Recurrent core of perpendicular scheme on fragment(?) of nodule. "Coretrimming" technique evidence, three striking platforms, two of which are roughly facetted and thelatter is naturally plain.

61

Fig. 9. Pronyatyn. Flat cores : 1. Recurrent core of centripetal scheme on pre-core of variety 1.Perimeter striking platform, no special preparation; almost exhausted form. 2. Recurrent core ofparallel scheme on pre-core of variety 1. Short debordants and "core trimming" technique evidence;almost exhausted form. 3. "Single" removal core of centripetal scheme of preparation of flakingsurface on pre-core of variety 1. Perimeter striking platform, rough preparation; almost exhaustedform. "Anvil type" damage on opposite side of the core.

62

Fig. 10. Pronyatyn. Flat and volumetric cores: 1. Recurrent flat core of parallel scheme of knapping onflake (?). Perimeter striking platform, rough and fine facetting of striking platforms, short debordant/"core trimming" technique evidence; absolutely exhausted form. 2. Two-sided flat core of recurrenttype; convergent and centripetal schemes. Rough and fine facetting of striking platforms, shortdebordant/ "core trimming" technique evidence; absolutely exhausted form. 3. Fragment ofvolumetric core of prismatic type. Roughly prepared striking platform. 4. Fragment of volumetric coreof prismatic type. Plain striking platform; facetting and trimming evidence.

63

Fig. 11. Pronyatyn chaine operatoire. A. Flint nodule. B, C. Different ways of initialisation of core onnodule.

64

B

Fig. 12. Pronyatyn chaine operatoire. A. Preparation of pre-core of variety 1. B. Preparation of pre-coreof variety 2. C. Preparation of pre-core of variety 3.

65

Bt, ft,

Fig. 13. Pronyatyn chaine operatoire. A. Utilisation of pre-core of variety 1 through not volumetriccentripetal mode of knapping. B. Utilisation of pre-core of variety 2 through not volumetric bipolar,centripetal, and classical Levallois modes of knapping.

B

Fig. 14. Pronyatyn chaine operatoire. A. Utilisation of pre-core of variety 1 through volumetric modeof knapping. B. Utilisation of pre-core of variety 3 through not volumetric Kombewa-like mode ofknapping.