Refining Our Understanding of Howiesons Poort Lithic Technology: The Evidence from Grey Rocky Layer in Sibudu Cave (KwaZulu-Natal, South Africa)


previous item | next item

Citation
Refining Our Understanding of Howiesons Poort Lithic Technology: The Evidence from Grey Rocky Layer in Sibudu Cave (KwaZulu-Natal, South Africa)

Material Information

Title:
Refining Our Understanding of Howiesons Poort Lithic Technology: The Evidence from Grey Rocky Layer in Sibudu Cave (KwaZulu-Natal, South Africa)
Series Title:
PLOS One
Creator:
de la Pena, Paloma
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Howiesons Poort ( local )
Sibudu Cave ( local )
Grey Rocky ( local )
Genre:
serial ( sobekcm )

Notes

Abstract:
The detailed technological analysis of the youngest Howiesons Poort occupation in Sibudu Cave, layer Grey Rocky, has shown the importance of blade production (with different knapping methods involved), but also of flaking methods in coarse grained rock types. Moreover, new strategies of bifacial production and microlithism were important. Grey Rocky lithic technology shows a really versatile example of reduction strategies that were highly influenced by the characteristics of the rock types. This lithic assemblage is another example of the technological variability linked to the Howiesons Poort technocomplex. The reasons for this variability are still difficult to elucidate. Discrepancies between sites might be for different reasons: diachronic variations, functional variations, organizational variations or maybe different regional variations within what has been recognized traditionally and typologically as Howiesons Poort. The technological comparison of the Grey Rocky assemblage with assemblages from other Howiesons Poort sites demonstrates that there are common technological trends during the late Pleistocene, but they still need to be properly circumscribed chronologically. On the one hand, Howiesons Poort characteristics such as the bifacial production in quartz are reminiscent of production in some Still Bay or pre-Still Bay industries and the flake production or the prismatic blade production described here could be a point in common with pre-Still Bay and post-Howiesons Poort industries. On the other hand, the detailed analysis of the Grey Rocky lithics reinforces the particular character of this Howiesons Poort technocomplex, yet it also shows clear technological links with other Middle Stone Age assemblages.
Original Version:
PLOS One, Vol. 10, no. 12 (2015-12-03).

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
This item is licensed with the Creative Commons Attribution License. This license lets others distribute, remix, tweak, and build upon this work, even commercially, as long as they credit the author for the original creation.
Resource Identifier:
K26-00054 ( USFLDC: LOCAL DOI )
k26.54 ( USFLDC: LOCAL Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

RESEARCHARTICLERefiningOurUnderstandingofHowiesons PoortLithicTechnology:TheEvidencefrom GreyRockyLayerinSibuduCave(KwaZuluNatal,SouthAfrica)PalomadelaPe ñ a *EvolutionaryStudiesInstitute,UniversityoftheWitwatersrand,Johannesburg,SouthAfrica * paloma.delapenya@gmail.comAbstractThedetailedtechnologicalanalysisoftheyoungestHowiesonsPoortoccupationinSibudu Cave,layerGreyRocky,hasshowntheimportanceofbladeproduction(withdifferentknappingmethodsinvolved),butalsoofflakingmethodsincoarsegrainedrocktypes.Moreover, newstrategiesofbifacialproductionandmicrolithismwereimportant.GreyRockylithic technologyshowsareallyversatileexampleofreductionstrategiesthatwerehighlyinfluencedbythecharacteristicsoftherocktypes.Thislithicassemblageisanotherexampleof thetechnologicalvariabilitylinkedtotheHowiesonsPoorttechnocomplex.Thereasonsfor thisvariabilityarestilldifficulttoelucidate.Discrepanciesbetweensitesmightbefordifferentreasons:diachronicvariations,functionalvariations,organizationalvariationsormaybe differentregionalvariationswithinwhathasbeenrecognizedtraditionallyandtypologically asHowiesonsPoort.ThetechnologicalcomparisonoftheGreyRockyassemblagewith assemblagesfromotherHowiesonsPoortsitesdemonstratesthattherearecommontechnologicaltrendsduringthelatePleistocene,buttheystillneedtobeproperlycircumscribed chronologically.Ontheonehand,HowiesonsPoortcharacteristicssuchasthebifacialproductioninquartzarereminiscentofproductioninsomeStillBayorpre-StillBayindustries andtheflakeproductionortheprismaticbladeproductiondescribedherecouldbeapoint incommonwithpre-StillBayandpost-HowiesonsPoortindustries.Ontheotherhand,the detailedanalysisoftheGreyRockylithicsreinforcestheparticularcharacterofthisHowiesonsPoorttechnocomplex,yetitalsoshowscleartechnologicallinkswithotherMiddle StoneAgeassemblages.Introduction HowiesonsPoortinRetrospectiveHowiesonsPoort(HP)isoneofthebetterknownindustriesofthesouthernAfricanMiddle StoneAgesequence.However,althoughitiswellknown,itisnotwell-understood.Depending PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 1/52 OPENACCESS Citation: delaPeñaP(2015)RefiningOur UnderstandingofHowiesonsPoortLithicTechnology: TheEvidencefromGreyRockyLayerinSibuduCave (KwaZulu-Natal,SouthAfrica).PLoSONE10(12): e0143451.doi:10.1371/journal.pone.0143451 Editor: NunoBicho,UniversidadedoAlgarve, PORTUGAL Received: June19,2015 Accepted: November4,2015 Published: December3,2015 Copyright: ©2015PalomadelaPeña.Thisisan openaccessarticledistributedunderthetermsofthe CreativeCommonsAttributionLicense ,whichpermits unrestricteduse,distribution,andreproductioninany medium,providedtheoriginalauthorandsourceare credited. DataAvailabilityStatement: Allrelevantdataare withinthepaperanditsSupportingInformationfiles. Funding: Fundedbyapostdoctoralscholarshipof theDST-NRFCentreofExcellenceinPaleosciences, EvolutionaryStudiesInstitute(Universityofthe Witwatersrand). CompetingInterests: Theauthorhasdeclaredthat nocompetinginterestsexist.

PAGE 2

ontheresearcherconcerned,ithasbeennamedanindustry[ 1 – 3 ],ahorizonmarker[ 4 ],a MiddleStoneAge(MSA)entity[ 5 ]andrecentlyatechnocomplex[ 6 ]oratechno-tradition[ 7 ], followingthemainstreamtechnologicalapproach.Thetwolasttermsarepresentlythemost commononesintheMSAliterature.Theterm ‘ technocomplex ’ wasformulatedbyClarke[ 8 ] referringtoasetofcultureshavinganumberofcommonfeatures.InWesternEuropeanstudies,thistermisalsousedbythepractitionersoftheG.Laplacetypology.However,nowadays ‘ technocomplex ’ isusedinPalaeolithicstudiesinanothersense,asaresultoftherelatively recenttrendintechnologicalanalysis.Thusatechnocomplexisseenassynonymouswithlongstandingtechnicaltraditions.Itisstrikingthatwhilethemethodologyforanalysingtheindustrieshasradicallychanged-fromatypologicalapproachtoatechnologicalapproach-the namesusedtoidentifytheassemblageremainthesameandtypologystillseemsthemainway toguidetherecognitionofassemblages.Thisepistemologicaltrendhasalsobeenrecognisedin UpperPaleolithictechnocomplexesinEurope,whichwereformerlyidentifiedascultures,then astypologicaltraditions,andmorerecentlyastechnocomplexes.Thebestexampleismaybe theGravettiananditsmultiple facies [ 9 ].HPhasfollowedasimilarhistoriographicalevolution, fromaculturaltypologicalconceptiontoatechnocomplex. Thispaperfollowsatechnologicalapproachandexploresspecificallythelithic-technologicaldefinitionofHP,puttingthestressontheknappingmethodsandthemanagementofdifferentrocktypesasa proxy forunderstandingtheorganizationalandeconomicalbehaviourfor theassemblageanalysed.InthispaperIwillrelatetheknappingreductionstrategiesfoundat themostrecentHP ’ slayeratSibuduwithothercontextsalsoidentifiedasHP(eitherinatypologicalorinatechnologicalmanner)inordertotestwhethertheterm ‘ technocomplex ’ (understoodasalongstandingtechnicaltradition)canbeappliedfordifferentassemblagesidentified asHP. Sinceitsfirstinception,HPhasbeenconsideredasanunusualindustrybecauseofitsLater StoneAge/UpperPalaeolithic-likeimplementsand,lateron,thefocusonthestudyofHP industrieshashighlightedagreatvarietyofmaterialculturethatisnowknowntoaccompany theHPlithicrepertoire,suchasengravedochre,bonetechnologyandostricheggshellengravings[ 10 – 13 ]. InthefirsthalfoftheXXthCenturyresearcherssuchasGoodwin,usinganevolutionary perspective,proposedthattheadvancedmaterialculturedemonstratesthatHPcorrespondsto thefinalMSAortransitionbetweentheMSAandLSA(asimilarhypothesishaslasteduntil recently,see[ 14 ]). IntheThirdPanAfricanCongressonPrehistory(1955)theHP(togetherwiththeMagosian)wasincludedinthesocalled ‘ SecondIntermediateperiod ’ fortheMSA,asatransitional industrybetweentheMSAandtheLSA.Thebasisofthiswasmainlytypological,onthepresenceofbackedand/ortruncatedpiecesmanufacturatedonfine-grainedmaterials[ 15 ].Afterwards,theMagosianturnedouttobeafallaciouslithictradition(becausethesitewhereitwas definedwasamixedassemblage)whereasthenameandconceptHPremainedintact. LateronSingerandWymer[ 16 ]usedpopulationreplacementasanexplanation,thatis,the makersof ‘ traditional ’ MSAtoolswerereplacedbyanewpopulationcreatingHPtools,and thattheoriginalsouthernAfricaninhabitantsreturnedtotheirhomelandafterthedemiseof theHP.Besidesthisinterpretation,itisimportanttohighlightthattheSingerandWymer ’ s excavationsatKlasiesRiverMouthconfirmedthatHPwasanindustryinthemiddleofthe MSAsequenceandnotalateMSAindustry,astypologicalproposalssuggested.Thiswaslater confirmedinsequencessuchasRoseCottage[ 17 ],Umhlatuzana[ 18 ],BorderCave[ 19 ]and Apollo11[ 20 ]. Deacon[4 , 21 ]proposedthatlifewaysimpliedbymaterialcultureassociatedwiththeHP couldbecomparedtotheSanethnographicrecord.Theenvironmentalstressgeneratedby RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 2/52

PAGE 3

climaticdeteriorationatthebeginningoftheHPstimulatedthecreationofnewsocialadaptations;andbackedpieceswouldhavebeenawaytomarksocialidentitybetweengroups,with socialinterchangesystemsuchasthe hxaro gift-givingpartnershipsystemoftheKalahariSan describedbyWiessner[ 22 ].ForDeaconthissocialtrendwouldalsohavebeenasignof ‘ modernbehaviour ’ . Wadley[ 23 ]:90 – 91comparedtheHPwiththeWilton,suggestingthatbothcouldberesult ofsimilartechnologicalresponsestosimilardemographic,socialandeconomicconditions.She alsopointedouthowtheStoneAgehadfluctuatingformsofsocialrelationsandtheWilton andHPmayrepresentrelated,yetdifferentforms. Indeed,HPhasbeenoneofthemainarchaeologicalentitiesenablingaclaimthat ‘ modern human ’ behaviourwaspresentintheMSA,andthatbehaviourintheMSAwasnotthatdifferentfrombehaviouratthebeginningoftheLSAortheso-called ‘ UpperPalaeolithicrevolution ’ [ 24 ]. Nonetheless,thearchaeologicalcharacteristicsassociatedwithHPandconceivedas ‘ advanced ’ ,suchasthebladetechnology[ 25 ],symbolicbehaviour[ 26 ],etc.havebeensometimespresentedasephemerale.g.[ 27 ].Inotherwords,behaviourintheHPwasviewedasan anomalywithinthegeneraltrendofdevelopmentwithintheMSA. Finally,amongthedifferentpropositions,itmustbehighlightedthatharshenvironmental conditionswerealsousedinordertoexplainHPdevelopment[ 21 , 26 , 28 – 30 ].TechnologyandHowiesonsPoortTherealityisthat,regardlessofeffortstodefinetheHPtechnologically,itisstillusuallyrecognizedtypologically.ThisisaproblemwhichitshareswiththeSB,whichisalsomainlyidentifiedbytypology.ThisseemsparadoxicalasHPandSBarebyfarthemoststudiedandcited archaeologicalentitiesintheSouthAfricanMSA. FromatypologicalpointofviewThackeray[ 31 ]gaveanaccuratedefinitionofsouthern AfricanHPassemblages: ‘ southernAfricanMSAstoneartifactassemblagescharacterizedby variousbackedand/ortruncatedpiecessuchastrapezoidsandsegments(crescents) , oftenconsideredlargerthanthosefoundinLaterStoneAgeassemblages( ... ) , whichareoftenmadeon fine-grainedrawmaterialssuchassilcreteandchert . TheseartifactsarefoundinadditiontotypicalMiddleStoneAgeflake-blades(elongatedflakes)andflake-bladesections , anegligibleproportionofwhichisretouchedtoformunifacialandbifacialpoints , denticulates , orscrapers ’ [ 31 ]: 390. Inthelastfifteenyears,typologicalcriteriahavebeensupplementedbyHPtechnological definitionsindifferentregionsofSouthAfrica.Furthermore,therehaslatelybeenanotable efforttodefinetheentiretechnologyoftheHP.Nevertheless,thetechnologicaldefinitionof theHPrequiresalotmoreresearch,notwithstandingthattheworkofWurz[ 25 ]ontheKlasies Riverassemblagebegantochangethedefinitionofthisindustryfromatypologicaltowardsa technologicalone.Wurz[ 25 ]proposedthatHPbladesoriginatedfromarecurrentbladeproductionsystemusingasofthammer.Shealsoproposedthatthisbladeproductionwasthe basisforcreatingtheblanksforthebackedtoolsinKlasiesRiverHPlayers(inotherHPassemblages,however,flakesarethemostusedblanksinordertoproducebackedimplements,e.g. forDiepkloof,seeMackay[ 32 ]).Indeed,variantsofHPhavebeenidentifiedatsitessuchas Umhlatuzana,RoseCottage,Sibudu,DiepkloofandKlipdrift[ 6 , 18 , 25 , 33 – 39 ]. ArelevantexampleforthetechnologyofHPisRoseCottage.Inthissitethebladeproductionandtheformaltoolshavebeenprofuselydescribed[ 17 , 33 , 38 ].IntheHPlayersofRose Cottageknappingstrategiesseemreallyconstrainedbytherocktypechosen:almostexclusively opalinecryptocrystallinenodules.Sorianoetal.[ 33 ]comparetheknappingstrategieswith RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 3/52

PAGE 4

opalinenodulestoproducebladeletsfromtheHPofRoseCottagewiththebladeproductionin theChatelperronianinFrance,whichrepresentsabreakwithMiddlePaleolithicbladeproduction.Theyalsoemphasizetheimportanceof ‘ marginalpercussion ’ amongthetechniquesof knappingassociatedwiththeHP.Moreover,intheirstudytheydistinguishbetweena ‘‘ classic ” HP,whichcorrespondstothebaseandthemiddleoftheHPsequence.Intheselayers(layers EMDandMAS),bladeproductioniscarefullycarriedoutwithverymarginalpercussion.The secondphase,orfinalHP(layersETHandSUZ),ismarkedbyadegradationinthequalityof bladeproduction,observableinthelowerdegreeofregularityinproductsandalesselaborate platformpreparation.Furthermore,inthissecondphase,marginalpercussiondisappears. OneofmostrecentstudiesofHPtechnologyistheoneofDiepkloof.InthissitetheHP industrywasdefinedasaregionallyspecifictechnocomplexthat,unlikeotherHPindustries, couldbesubdividedintophases(Early,IntermediateandLate)[ 6 ].Moreover,intheirstudy theyproposethatthereisinterstratificationofanon-HPlithictechnologybetweentheEarly andtheIntermediateHP,thesocalledMSAtype ‘ Jack ’ ,eventhoughthislayercontainslarge backedpieceswhich,traditionallyhavebeenthehallmarkfortypologicalrecognitionofHP assemblages.Porrazetal.[ 6 ]seesimilarities,intermsoftechnologicalstrategies,amongthe threephasesofHP(suchasthe ‘ HPcorereduction ’ ).NeverthelessintheLatePhasetheypoint outanincreaseofirregularbladescoupledwithanincreaseinflakeproduction.Inother words,betweentheEarly,IntermediateandLateHPthattheydefinetheyobservethesame chaîneopératoires .Therefore,thevariationsthattheypointoutbetweenthesethreephasesare mostlyrelatedtotheformaltoolrepertoire,withmainly piècesesquillées ,truncatedpiecesand bifacialpiecesintheEarlyHP;strangulatedpiecesandnotchedpiecesintheIntermediateHP and,finally,mainlybackedpiecesintheLateHP.Theystatethatthesignificanceofthesethree distinctHPphasescannotbefunctionalandthatthecomparisonswithotherHPsitessuggest thatthesethreephasesrepresentatechnologicaltrendwithinHP.Furthermore,aftertheir analysistheyproposethatinmostofsouthernAfricansitestheindustriesidentifiedasHP probablycorrespondtotheIntermediateandLatephasefromDiepkloofthattheyname ‘ classic ’ HP(aterminologypreviouslyusedbySorianoetal.[ 33 ],butwithaslightlydifferentdefinition, vid . supra ). AtKlipdrifttheconclusionofthetechnologicalanalysiswasthatthreephaseswithinthe ‘ HPcomplex ’ couldbedistinguished.Thelowermostphase(layersPCA,PBE)ischaracterized bysilcreteasthemainrocktypeand,typologically,bythepresenceofnotchedpieces,strangulatedpiecesandhighlystandardizedtruncatedblades.Theintermediatephase(PBC,PBA/ PBB)ischaracterizedbyanincreaseinthequartzexploitationandbytheincreaseofbacked piecesamongtheretouchtoolkit.Thefinalmostrecentphase(PAYlayer)ischaracterizedby theuseofquartziteandbyanincreaseofthesizeofthebladeproduction,togetherwithan increaseofflakeknappingmethods(suchas Levallois ).Henshilwoodandcolleaguespropose thatthisphasecouldbeinterpretedasa ‘ transitionallayer ’ towardspost-HP.Thecommon trendinthesethreephasesisthattheknappingreductionsequenceisalmostentirelydevoted totheproductionofblades[ 39 ]. IntworecentstudiesbydelaPeñaandWadley[ 36 , 37 ],oneoftheHPlayersofSibuduwas studiedindepthfromatechnologicalpointofview:GreySand(whichhasanageestimateof 63.8±2.5ka,[ 40 ]).Inthesetwostudies,newknappingmethodswererecognisedintheHP, suchasawell-developedprismatictechnologyforbigbladeproduction(fordoleriteandhornfels),varietiesofcoreonflakesforbladeletproduction(fordoleriteandhornfels),andstrategiesofmicrolithismforquartzreduction,includinganextensiveuseofprismaticcoresand, subsequently,bipolarknappinginordertoproducemicroliths(bladeletsandsmallflakes). Moreover,recenttechnologicalstudiesatDiepkloofandSibuduhavehighlightedbifacial reductionsequenceswithinHPassemblages[ 6 , 41 ]. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 4/52

PAGE 5

TherecenttechnologicalanalysisofasampleofSibudu,RoseCottageandKlasiesRiver assemblagehavehighlightedseveraltechnologicaltrendsfortheHP[ 35 ].Theyassignall theseHPassemblagestotheso-called ‘ classicHP ’ definedatDiepkloof(thismeanstheIntermediateandlateHPphasesofDiepkloofsite).ForSibududifferenttechnologicaltrendsare pointedout,suchassoftstonehammerpercussi onfortheflakingtechniquesandanaxial haftingforbackedimplements.Furthermore,thisstudyhighlightsthesimilaritiesinblade knappingmethodsbetweenKlasiesandSibu du,thisisproposedfromthemorphologyofthe bladesandthesamefrequenciesinthesetwoa ssemblages.ThisrecentstudyconsidersHPas an ‘ industry ’ ,identifiedwitha ‘ culturalentity ’ ,anditscoherencewithinSouthAfrica,in termsoftechnologicalandtypologicaltrends.ThisisstressedincontrapositiontoSB assemblages. Asaresultofthesenewtechnologicalanalyses,variationinHPindustriesisbeginningtobe recognised.Inotherwords,itnolongerseemstobethehomogeneouslithicphenomenonpromotedbytheformertypologicalapproach. ItseemsthatoneofthemainchallengesforconstructingabetterdefinitionofHPisfirstto establishthemeaningofvariabilityinthecontextofthisindustry,therebyenablingthedistinctionbetweendiachronic/chronological,functional,organizationalandregionaltechnological variation. Togetherwithlithictechnologystudies,ochreandresidueanalyses,macrotraceusewear analyses,andexperimentalapproacheshavepointedtoevidenceforvariedhaftingstrategies forcompositetools,usingbackedHPimplements,firstinSibuduCave[ 42 – 45 ],andlateronin Diepkloof[ 46 ].Suchtechnologicalstrategies(whicharerelatedtolithictechnologyandwhich complementthem)might,inthefuture,giveusdifferentperspectivesontechnologicalvariationwithinHPassemblages.TheChronostratigraphicalPositionofHowiesonsPoortTheHPchronologyiscurrentlythesubjectofintensescientificdiscussion.Jacobsetal.[ 3 ]proposedashorttimeframeforHPandSBindustriesbasedontheevidencefromninesouthern Africansites.InthatstudyJacobsetal.[ 3 ],usingsingle-grainopticalluminescencedating (OSL),proposedthatHPandSBwereshort-lived(5000yearsorless)industriesseparated,at thesitesthatweredated,byabout7000years.TheyproposedthatHPstartedaround64.8ka andended59.5ka,withdurationinthetimeframeofabout5.3ka[ 3 ].Withinthisshort-lived model,HPbelongstoMarineIsotopeStage(MIS)4.However,Triboloetal.[ 47 ]havechallengedthemodelinrecentworkatDiepkloof,basedonTLandOSLdatingof39burntlithics and5sedimentsamples.TheresultsaresignificantlyolderthanthosereportedbyJacobsetal. [ 3 ].Theyobtainedachronologyof105±10kaand109±10kafortheEarlyHP;77±8ka, 85±9ka,83±8ka65±8kafortheIntermediateHPand52±5kafortheLateHP.Therefore, thetimespanproposedforthethreetechnologicalphasesrelatedtoHPinDiepkloofis between109and52kaago[ 47 ].Thus,inthisnewmodel,HPfirstappearsinMIS5anddisappearedinMIS3.Feathers[ 48 ]reanalysedsamplesfromHPlayersofDiepkloofandKathuPan 6andhasproducedslightlyolderorequivalentdatestothoseprovidedbyTriboloetal.[ 47 ]. Moreover,JacobsandRoberts[ 49 ]havereanalysedandreassessedtheirsedimentsamples fromDiepklooftoaddresssomeofthecriticismsoftheirDiepkloofchronology[ 3 ].Inthisnew researchtheyobtainedagesthatarerobustandconsistentwiththeiroriginalchronology,but theycannotsatisfactorilyexplainwhytheTLandOSLagesprovidedbyTriboloetal.[ 47 ]are different.Theydopointoutthattheirsedimentsampleswerenotcollectedfromthesamepart oftheDiepkloofexcavationsassamplescollectedbyTriboloandFeathers.Inconclusion, regardingtheHPchronology,thedebateremainsopen. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 5/52

PAGE 6

ThecontroversyaroundHPchronologypresentsamajorproblemfortworeasons:onthe onehand,fordiscussionsoftechnologicalvariabilitythroughtimeand,ontheotherhand,for thecorrelationofHPtechnologywithclimaticmodels.Moreover,theuncertainchronology makesitdifficulttounderstandwhatHPrepresents.Thereisabigdifferencebetweena techno-traditionthatisa ‘ shortlived ’ culturalphenomenon(5ka)asopposedtoalithictraditionlasting45-50ka.ObjectiveInthispaperIpresentanindepthtechnologicalstudyoftheyoungestHPlayerinSibudu Cave:GreyRocky(GR).Thisisarichlayerwithoptimalorganicpreservation[ 50 ].Previous lithicanalysesofpartofWadley ’ scollectionandGRhavebeenalsopublished[ 14 , 35 , 51 ]. Inapreliminarystudyitwasnoticedthatflakereductionstrategiesmightbepresent, togetherwithothertypesofbladeknappingmethods.Moreover,bifacialreductiononquartz hasbeendocumentedandexplainedinapreviouswork[ 41 ].Inotherwords,thislayerisparticularlyinterestingbecause,togetherwiththetypicaltechno-typologicalelementsassociated withHP(backedtools,profusebladeproduction),itwasalreadyknownthatotherspecific technologicaltraitswerepresent.Therefore,themainaimofdescribingGRtechnologicallyis toexploreHPlithicvariationasitoccursinthiseasternpartofSouthernAfrica.Sibudu remainsthemainsequenceforthispartofthecontinent,notleastbecauseofits chronostratigraphy. Themainobjectivesofthispaperarethree:first,togiveadetailedtechnologicaldescription oftheyoungestHPoccupationinSibuduCave,makinguseofallthelithicsexcavatedfromsix squaremetresduringWadley ’ scampaigns(notethattherecentpaperofSorianoetal.[ 35 ] onlytookintoaccountasampleofthislayer).Secondly,Ishallestablishastrongmethodologicalprotocol(qualitativeattributeanalysisandbasicstatisticaltests)thatinthefuturewillbe usedtotackletechnologicallyotherMSAlithicassemblagesinSibuduCaveandotherHP sequences.Thirdly,toplacetheresultsofthistechnologicalstudyinabroaderdiscussion aroundHPandtheMSAvariability.SibuduRockShelterandtheLayerStudiedSibuduislocatedapproximately40kmnorthofDurban,andabout15kminlandoftheIndian Ocean,onasteepcliffoverlookingtheuThongathiRiver(29.522627°S,31.085895°E)( Fig1 ). Theshelteris55mlongand18minbreadthandhasalongoccupationsequencewithseveral layersandfeaturescorrespondingtothepre-SB,SB,HP,post-HP,lateMSA,finalMSAand IronAge[ 52 ]. HPoccupationsreportedherecomefromsixsquaremetres(squaresB4,B5,B6,C4,C5and C6)ofWadley ’ sexcavationsinthedeepsounding. ThelayersassociatedwiththeHPare(fromthebasetothetop):PinkishGreySand(PGS), GreySand(GS,GS2andGS3),DarkReddishGrey(DRG)andGreyRocky(GRandGR2) ( Fig1 )[ 52 ]. Thestratigraphyisclear,combustionfeaturesarediscernible[ 50 , 53 ]( Fig2 ),andmicromorphologyimpliesthatmostSibudulayershavestratigraphicintegrity[ 54 ],althoughrockfall betweentheoldestHPlayer,PGS,andtheunderlyingSBlayer,ReddishGreySand(RGS),has causedsomedisturbance. ThelayerthatIdiscusshereisGRwhichislight,brownish-greysilt.InB6thelayerGRisin betweenarockfall.GR2isanartificialspittodividethelayer.Duringtheexcavationseveral featureswithinGRweredistinguished( Table1 )[ 50 , 53 ];ascanbeseenthereareabundant RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 6/52

PAGE 7

hearths( Fig2 ).ThechronologyofGRlayeris61.7±2Kaobtainedfromsinglegrainoptically stimulatedluminescenceonsedimentfromGR2[ 52 ].TheRockTypesKnappedattheSiteAsshownin Fig3 themainrocktypesandmineralsknappedintheGRlayeratSibuduare: dolerite,hornfels,sandstone,quartz,quartziteandcryptocrystallinematerial(inorderofpercentagerepresentationinthelayer). TheuThongathiRiverbelowSibuduisasourceofweatheredandriver-rolleddoleriteand quartziteaswellasofsmallquartzpebbles.TheroundedcortexonanumberofSibuducores andflakesimpliesawaterborneoriginforsomedoleritebroughttothesite,butcoresandflakes madefromtabulardoleritepiecesarealsopresent.AbundantigneousdoleritenearSibudu derivesfromintrusiveJurassicvolcanism,mostlyassills,althoughatruedoleritedykelies Fig1.LocationofSibuduCave. Ontopleft,locationofSibuduCave(29.522627S,31.085895E).Onthebottomleft,PlanofSibuduCave.Thisschematic mapwasmadeonthebasisofatopographicmapofSouthernAfrica,source:MapsattheCIA(publicdomain): https://www.cia.gov/library/publications/theworld-factbook/index.html .Onthebottomlefttheexcavationgridisrepresentedwiththesquaremetersforthisanalysishighlightedingrey.Onthebottom right,stratigraphyoftheNorthwallofSibuduCave(StratigraphycourtesyofLynWadley).TheHPlayersarehighlightedingrey. doi:10.1371/journal.pone.0143451.g001 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 7/52

PAGE 8

closetotherockshelterandthisseemslikelytohavebeenthesourceofmuchofthedoleriteat thesite.Doleritesillsintheareaincludefine-grainedonesliketheMhlasinisillandcoarsegrainedonesliketheVerulamsill[ 55 ].Petrographicanalysisofathinsectionoffine-grained doleriteshowedthatthemineralsinclude45%clinopyroxene,44.5%plagioclaseandsmallpercentagesofquartz,limoniteandgoethite[ 56 ].Unfortunately,doleriteischemicallysimilar acrosslargeregionsofSouthAfricaandcannotbedistinguished,althoughtheEffinghamsills inKwaZulu-Natalhavehighersilicacontents(inexcessof63wt%)thanmostoftheothers. Wheredoleriteintrudesintoshale,therearebandsofmetamorphichornfels.Differingtemperaturesoccurinthezoneofthermalmetamorphismwhereadoleriteintrusionoccurs.Consequently,therearedifferentgradesofbothhornfelsanddolerite.Hornfelsisdifficulttofind intheSibuduarea,butonesourceisnearVerulam,about15kmfromSibudu.Thishornfelsis poorlymetamorphosed,andoccursasthinslabs,onlyafewcentimetresthick.XRFelemental analysisofapieceofhornfelsfromSibududemonstrateditshighsilica(63.8%),lowmagnesium(1.4%)andcalcium(0.7%)content(relativetothatofdolerite)whichconfirmsthatitis metamorphosedshalefromacontactzonewithadoleriteintrusion[ 55 ].Asurveyalongthe Fig2.Hearth2inGRlayer,squareC4. PicturefromL.Wadley Â’ sexcavation.PhotographcourtesyofLyn Wadley. doi:10.1371/journal.pone.0143451.g002 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 8/52

PAGE 9

uThongathiRiverbyHelenKempsonin2011showedthathornfelsnodulesarenotpresent there[ 56 ].Moreover,itmustbetakenintoaccountthatthelandscapemighthavechangeda lotsincetheHP,andavailableoutcropsinthepastmightnotbevisiblenow. Sandstoneisthe ‘ local ’ rocktypeatthesite,astheSibudurockshelterissituatedinasandstonecliffwithintheMariannhillFormationoftheNatalGroup[ 55 ].Thesandstoneknapped atthesitelooksidenticaltotheonethatisformingtherockshelter.Itisquiteacoarsegrained subtype.Thegranularstructureofthisrockmakesitasuitablematerialtoknap. CrystallinequartzappearsasriverpebblesandaspebblesinsideconglomeratesintheVerulamarea[ 55 ].Somebigquartzpiecescouldcomefromriverpebblesbutthisdoesnotseemthe caseforsomeotherbigpieces.Probablyquartzdykeswereusedasasourceforthisrawmaterialinthepast,judgingbythemacroscopiccharacteristicsofsomecortexareasofthesepieces. Withinthecrystallinequartzassemblagetwomaincategoriescanbedistinguishedandboth wereexploitedintheHPlayersofSibudu[ 37 ]:veinquartz(milkyorxenomorph)andcrystal quartz(alsocalledhyalineorautomorphquartz)[ 57 , 58 ].AspointedoutbydeLombera-Hermida[ 58 ]:102crystalline ‘ quartzformationprocessesmustbetakenintoaccountinorderto establishagoodpetrologicalclassificationandcharacterization ’ ,andthesecharacteristicshave importantimplicationsforthemechanicalpropertiesofdifferentquartzvarieties.Xenomorph quartzdisplaysgreatervariabilitybecauseofdifferentchemicalandphysicalcausesofitsformationprocess[ 58 ]:102.Alltheseconditionsgeneratedifferenttypesofcrystallinequartzand, therefore,differenttypesofmechanicalproperties.MartínezandLlana[ 59 ]madeamorphostructuralclassificationtakingintoaccountgrainsandplanes(flawsorcrystallinesurfaces)of crystallinequartz.Fourmorphostructuralgroupsweredistinguished:NN:nograin,noplane; Table1.ListwiththedifferentstratigraphicfeaturesrecognisedaspartofGRandGR2inSibudu, fromWadley ’ sexcavation. GreyRockyunderHearth1 GreyRockyunderHearth2 GreyRockyunderrock Hearth1inGreyRocky Hearth2inGreyRocky Hearth3inGreyRocky Hearth4inGreyRocky Hearth5inGreyRocky HearthAinGreyRocky HearthBinGreyRocky HearthCinGreyRocky HearthDinGreyRocky HearthEinGreyRocky Hearth1inGreyRocky2 Hearth2inGreyRocky2(blackbase) Hearth3inGreyRocky2 HearthAinGreyRocky2 HearthBinGreyRocky2 HearthCinGreyRocky2 HearthDinGreyRocky2 WhiteashbelowGreyRocky2 WhiteashunderGreyRocky2 YellowAshinHeartEinGreyRocky2 doi:10.1371/journal.pone.0143451.t001 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 9/52

PAGE 10

NS:Nograin,plane;SN;grainy,noplane;SS:grainy,plane.FollowingthisclassificationSibudu Â’ sveinquartzisNSinlayerGR.Inaddition,aspointedoutinpreviouspublications,hyaline orcrystalquartzwasalsoknapped[ 37 , 60 ]. Fig3.Percentagesofrocktypes. Top:percentageofrocktypesinSibudu Â’ slayerGRamongstpiecesover 2cminsquareB4.Bottom:Percentageofrocktypesforcompleteflakesinthe6m2analysed(see S1File ).As canbeseenthesetwosamplesareverysimilarandthemainrocktypesknappedinGRlayeratSibuduare: dolerite,hornfels,sandstone,quartz,quartziteandcryptocrystallinematerial(inorderofpercentage representationinthelayer). doi:10.1371/journal.pone.0143451.g003 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 10/52

PAGE 11

Cryptocrystallineormicrocrystallinequartzalsoappearswithinconglomeratesinthisarea [ 55 ]anditisoneoftheminoritymineralsknappedinthislayer.Cryptocrystallinemineralsin GRarereallyfinegrainedandtheyhaveverywelldevelopedconchoidalfracturepatterns. Quartziteappearsasriver-rolledsmallpebblesanditisusuallymedium/finegrained.MethodologyTheexcavationofSibuduwasconductedwithapermitobtainedfromAmafaiKwaZuluNatali,theHeritageAgencybasedinPietermaritzburg,SouthAfrica.Noethicsclearanceor permitisrequiredtostudythelithicartefactsfromSibudu. Thisstudyemployedthe chaîneopératoire approach,withwhichanassemblageisviewedas theoutcomeofculturalchoicesbyahumangrouporseveralhumangroupsthatoccupiedthis rocksheltermorethan60000yearsago. The chaîneopératoire approachassumesthattechnologicalchoicesareorderedandcanbe classifiedintodifferentstages[ 60 – 63 ],fromtheselectionoftherocktypes(acquisition)tothe discardorrecyclingafteruse[ 62 ].Havingsaidthis,Iacknowledgethatthismethodologyisa systemicapproach.Inotherwords,oneisforcedtoconsidertheassemblageoflithicblanksas awhole,despitethefactthattheywerelikelyproducedbydifferenthumangroupsatdifferent times.Therefore,thesystemicnatureofthe chaîneopératoire cansometimesprovideaskewed perspective.Inotherwords,analysingallthelithicsfromGRthroughthe chaîneopératoire approachmeansthatImustassumethattheyaretechnologicallyrelatedandcanbeconnected; andthattheyalwayssharesomesortoftechnologicalaffiliation.Therearelimitationstothis approach,becausetodemonstratetheassumptionthatallthelithicsarerelatedandpartofthe samesystem,refittinganalysesshouldbeundertaken.This,however,wouldbeextremelydifficultfortheSibudumaterialsbecauseofthesimilarcolourandothermacroscopiccharacteristicsofthelithicassemblage ’ srawmaterialtypes,andbecausethesizeofWadley ’ sexcavation (6m2)isonlyabout2%oftheroughly300m2oftheSibududeposit.Furthermore,asdifferent scholarshavepointedout[ 64 – 66 ],the chaîneopératoire approachcanbehighlysubjective.In ordertomakethisstudyasobjectiveaspossible,andinordertoallowtheGRassemblagetobe comparedwithotherSibuduandMSAassemblages,Iincorporatedvariousquantitative parametersinanattempttolimitsubjectivitythatmayarisefromapurelyqualitativeandsystemicapproach.Furthermore,aftertheattributeanalysis,Iapplieddifferentbasicstatisticsto supportthequalitativearguments. ThesamplestudiedfromGRwasthefollowing:allcores(n=120),corerelatedby-products (n=63)andretouchedpieces(n=244)fromtheWadley ’ sexcavations.Inordertounderstand lithicproductionbetterIalsoanalysedallcompleteflakes(n=1091)(see S1File ).Finally,all lithicartefacts(includingchips)wereexaminedinordertoidentifyandanalyseanyitemsof potentialtechnologicalimportance.Thelithicassemblagewasdividedintofourbroadanalyticalcategories:(1)cores,(2)blankswithoutretouch,(3)retouchedblanksand(4)chips.The chipsincludepieceswithawiderangeofmorphologiesthataresmallerthan10mmforquartz, sincepreviousstudiesoftheGSlayershowedthatthiscut-offwasthemostappropriateforthis material[ 36 , 37 ],andsmallerthan20mmfortheotherrawmaterialtypes.Moreover,Iexaminedallthelithicremains,includingchipsinapreliminaryanalysisofthestudyofGR.Inthat analysisIseparatedalltheretouchedpiecesevenwhentheywereunder10mmandwasableto identifymorphotypessuchasbifacialpieces,smallnotchesandbackedpieces,especiallywith respecttoquartzpieces. Coresandcore-relatedby-productscanprovideimportantqualitativeinformationabout knappingmethods,although,aspreviousresearchershavealsopointedout[ 64 ]theycanprovidealimitedviewofthereductionsequence.Indeed,coresusuallyreflectonlythelatterstages RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 11/52

PAGE 12

ofreductionanddiscard.Anunderstandingoftheknappingprocessresultsfromaquantitative andqualitativestudyofthemajorityoftheblanks,especiallythosefromflakeproduction[ 67 ]. ThespecificlithicattributesrecordedaredescribedinTables 2 – 5 andincomplementary Figs 4 and 5 .Mostofthequalitativevariablesforthehornfelsanddoleritecorestudycome fromthePelegrinmethodforChatelperroniancores[ 63 ].Thestudyofcoresonflakesfollows previoustechnologicalandexperimentalwork,suchas[ 68 , 69 ].ThestudyofSibudu ’ squartz coresusestheattributespresentedinpreviousquartzstudies,suchas[ 70 – 71 ]andmyown experimentationcriteria[ 72 ],inordertodistinguishbetweenfreehandandbipolarknapping. Table2.Variablesrecordedforquartzcores. Typeofblank Presenceofcortex(0,1 – 10,10 – 40,40 – 60,60 – 90,90 – 99,100%). Length,breadth,thickness(maximummeasurementsforallthree)andweight Numberofstrikingplatforms Orientationofstrikingplatforms Lengthandbreadthofstrikingplatforms(incaseofseveral,thelarger) Typeofpreparationofthestrikingplatform(simple ake,faceted,etc.) Lengthandbreadthofknappingsurface Scarpatternofthelastnegativesontheknappingsurfaceofthecore Presenceofconchoidalnegatives(yes/no) Presenceof ssurationinoverhangorstrikingplatform(yes/no) Presenceofbluntnessinoverhangorstrikingplatform(yes/no) Freehandorbipolarcore Typeofquartzcorefollowing Fig4 doi:10.1371/journal.pone.0143451.t002 Table3.Variablesrecordedforcores(non-quartzrocktypes,suchashornfels,dolerite,sandstone, etc.). Typeofblank Presenceofcortex(0,1 – 10,10 – 40,40 – 60,60 – 90,90 – 99,100%) Length,breadth,thickness(maximalmeasuresforthethreeofthem)andweight Volumetricshape(prismatic,conical,globular) Lateraltrimming(crestorsemicrest)(yes/no) Numberoflateraltrimmings Othertypeoftrimming Numberofstrikingplatforms Orientationofthestrikingplatforms Preparationofthestrikingplatforms(simple ake,faceted,etc.) Shapeofthestrikingplatforms Shapeoftheexploitationsurface Lengthandbreadthofthestrikingplatform(maximummeasurements) Lengthandbreadthoftheexploitationsurface(maximummeasurements) Curvatureoftheexploitationsurface(yes/no) Lengthandbreadthoflastnegative(nottakingintoaccountaccidents) Anglebetweenstrikingplatformandexploitationsurfaceforprismaticcore,oranglebetweenfacesfor discoidalandLevalloiscoresinthe4subdivisions. Typeofcore:prismatic,discoidal,Levallois,HowiesonsPoortcore,Coreon ake(bipolarcore, Kostienki core,burin-likecore,endscrapercore).Thesegeneralcategorieswerepreviouslyusedandexplainedin [ 36 ]. doi:10.1371/journal.pone.0143451.t003 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 12/52

PAGE 13

Furthermore,inthisstudyIusedthecategoriesproposedforHPcoresintwoprevioustechnologicalpapersonSibudu ’ sHPlayerGS:onefocusedonquartz[ 37 ]andanotherexamined knappingmethodsfromcoremorphology[ 36 ].Suchconsistencyshouldfacilitatefuturecomparisons.Itmustbeclarifiedthattheterm ‘ coreonflake ’ refersheremainlyto Kostienki cores, burin-likecoresandend-scrapercores.Withincategory2(blankswithoutretouch)I distinguish: (2.1.)Flakes.Inthestudydatabase,itwasrecordedifaflakewasinfactwhatmaybetechnologicallyidentifiedasablade itemswhoselengthisgreaterthantwiceitswidthandwithparalleldorsalridgesandtrapezoidalortriangularincross-section; (2.2.)Core-relatedby-products.Thiscategoryincludesallblanksthatshowanycharacteristic ofcoremorphology(andthereforeknappingmethods).Theycouldbeproducedpurposefullyoraccidentally.Inotherlithicstudiesthiscategoryisreferredtoas ‘ coretrimmingelements ’ [ 73 ]andinpreviouspublicationstheyarereferredtoas ‘ trimmingormaintenance by-products ’ [ 36 ].Withinthiscategory,otherspecificpiecesareidentified(2.2.1.)ascrests orsemicrestsfortheinitiationofacoreorforcoremaintenanceand(2.2.2.)falsecrestsor semicrests.Withrespecttothelatter,theyarenotmaintenanceby-products,butrather resultfromachangeinthedirectionofblankremovalwiththecrestrepresentingaprevious overhang,see[ 36 ]:31,and Fig4 withinit,andGRexamplesbelow; (2.2.3.)Cleaningflakes.Thesearepiecesthatmayresultfromcorrectingerrorsorrepairing accidentsthataffectedthestrikingplatform; Table5.Variablesrecordedforretouchedblanks. Alltheattributespreviouslylistedforblankswithoutretouchin Table4 . Locationofretouch Delineationoftheedge/sretouched(rectilinear,curved,denticulated) Regularityoftheretouch(yes/no). Continuity(yes/no) Typeofmorphotypeincluding:Notch,end-scraper,side-scraper,strangulatedpiece,denticulate,burin, marginalretouch ake/blade,retouch ake/blade,borer,bifacialpieceandbackedpiece(includingallthe categoriesillustratedin Fig5 ).AllthesemorphotypeswerealreadyrecognisedintheGSstudyofSibudu [ 36 ] doi:10.1371/journal.pone.0143451.t005 Table4.Variablesrecordedforallunretouchedblanks. Typeofblank( ake,bipolarblank,burinspall) Length,breadth,thickness(maximummeasurementsforthethreeofthem) Presenceofcortex(0,1 – 10,10 – 40,40 – 60,60 – 90,90 – 99,100%). Typeofplatform:plain,cortical,dihedral,faceted,punctiform,linear.Removed,broken/crushed. Lengthandbreadthoftheplatform Anglebetweenplatformanddorsalfaceofthe ake Numberofscarsonthedorsalface Dorsalscarpattern:unidirectional,unidirectionalconvergent,bidirectional,crossed,subcentripetal, centripetalandunknown. Shapeoftheblank/lateraledge:Converging,expanding,ovoid,circular,indeterminate. Crosssection:triangular,right-triangular,trapezoidal,lenticular,domed,indeterminate. Accidents(yes/no) Typeofaccident(overshoot,hinge/step) doi:10.1371/journal.pone.0143451.t004 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 13/52

PAGE 14

(2.2.4.)Flakesexhibitingknappingaccidents; (2.2.5.)Flakesexhibitingcoremorphology. Thestatisticalanalyseshavetwomainobjectives:(1)todescribebetterthestudiedassemblageand,(2),toidentifyandexplorerelationshipsbetweenthedifferentrecordedquantitative andqualitativevariables.Thenullhypothesiswasthatallrocktypeswereknappedinthesame manner.Inotherwords,alltypesofrockswerereducedfollowingidenticalknappingmethods; andthereforeanidenticalsizedistributionofblankscanbefoundforeachoneoftherock types.Inordertoexaminethevariabilityofthedifferentrecordedvariables,anumberof descriptivemeanswasemployedsuchashistograms,dispersiondiagrams,box-plots,univariatestatistics,etc.FortherelationshipsamongthedifferentquantitativeandqualitativevariablesrecordedintheattributeanalysisIperformeddifferentbasicstatisticalanalyses.ShapiroWilknormalitytestswereperformedtoidentifythosequantitativevariablesthathadnormally distributedvalues.Additionally,mixtureanalysesforthenon-normaldistributionsofthe quantitativedatawereimplemented,inordertofindpotentialgroupsofquantitativedatawith normaldistributions[ 74 ].AsexplainedbySotoSebastián[ 74 ]:158mixtureanalysisisan agglomerative,hierarchicalandunivariatestatisticaltest.Itisamethodofmaximumlikelihood estimationtorecognizeparameters(suchasmeansandstandarddeviation)oftwoormore univariatenormaldistributionsgroupedinasinglesample.Forexample,inordertodiscover whether,withinablankcategorysuchasblades,differentknappingmethodswereapplied,differentnormalgroupsofblankswouldbeexpected(seeanapplicationofthismethodologyin lithicanalysisinRiosetal.[ 75 ]andSotoSebastián[ 74 ].Besides,t-tests(fornormaldistributions)andUMann-Whitney(fornon-parametricdistributions)werecompletedforcomparisonsofquantitativesetsofdata.Finally,chi-squaretestswereappliedtosearchfor relationshipsamongstqualitativedata(forexample,betweencortexdistributionandtypeof blankorbetweenshapeoftheblankandthedorsalscarordinations).Alltheseanalyseswere performedwiththeopen-accesssoftwarepackagePAST( http://folk.uio.no/ohammer/past/ ). Fig4.Freehandandbipolarquartzcoresclassifiedinthisanalysis. Left:typesoffreehandcores.F1 conicalcore;F2,F3,F4,F5,F6andF7varioustypesofprismaticcores(withdifferentdirectionsofremovals); F8centripetalcore;F9multifacialcore.Right:differenttypesofbipolarcoresrecognisedinthisstudy,with differentdirectionsofremovals.B1.Unidirectional.B2.Bidirectional.B3istheresultofrotation.B4istypical offractureaccidentsduringbipolarknapping. doi:10.1371/journal.pone.0143451.g004 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 14/52

PAGE 15

ResultsoftheGreyRockyLithicIndustryTechnologicalAnalysis TheCoresWhattypeofknappingmethodscanbeinferredfromthetechnologicalanalysisofthe cores?. Thefirststrikingcharacteristicofthecoresisthatthosemadefromhornfelsanddoleritearequitesimilar.Avarietyofdifferentcorestypes,suchascoreonflakes,Howiesons Poortcores,bladeletcores,centripetalcores,etc.weremadefromthesematerials,andtheyare verydifferentfromcoresmadeonotherrocktypessuchasquartzorquartzite,forwhichthe majorityofcoresisbipolar(Tables 6 and 7 ;and Fig6 ).Therefore,thepercentagesofcoretypes andtheirqualitativecharacteristicsimplythatthesedifferentbroadclassesofrawmaterial weremanagedandreduceddifferently. Anothergeneralcharacteristicisthatthepercentagesofcoresbyrocktypesdifferfrom thoseofrocktypesrepresentedintherestofthelithicassemblagerecoveredfromGR.Hornfels Fig5.Differenttypesofbackedpiecesconsideredinthisstudy. Thistypologicalclassificationtakesintoaccountthelateralizationoftheretouchandits curvatureorlackthereof.Thesingletrunctationtypeiswhat,inthesouthernAfricanliterature,isreferredtoasan ‘ obliquebackedpoint ’ doi:10.1371/journal.pone.0143451.g005 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 15/52

PAGE 16

coresaremoreabundantthanthosemadefromdolerite,andquartzmorethansandstone( cf . Fig3 , Table6 and Fig6 ).Onemustkeepinmind,though,thatthesepatternsmaybedueto samplingsinceWadley ’ sexcavationswerefromonlyasmallportionofthesite ’ sarchaeological deposits(6m2).Mostofthehornfelsanddoleritecoresarecoreonflakes,bipolarcores(which usuallyaremadealsoonflakes)andtheso-calledHowiesonsPoortCores[ 25 , 34 ]( Table6 ). Withinthecategorycoreonflakestherearedifferentsub-typessuchas Kostienki coresor burin-likecores( Fig7 ).Thesesub-typesarenotanoveltyintheHPbecausetheywerealready recognizedintheolderGSlayerofSibudu[ 36 ].Thesecoresonflakesandthebipolarcores weremostlikelygearedtowardstheproductionofbladeletblanks. TheHowiesonsPoortCores(HPcores)haveasimilarpatterntothosedescribedinWurz [ 25 ],Villaetal.[ 34 ]andintheSibuduGSanalysis[ 36 ].ThestrikingplatformofaHPcoreis preparedbycreatingalargetruncation(orfacetedsurface)andthelateralconvexitiesareproducedviasemicrestremovals.Sometimesthedistalpartoftheknappingsurfacehassmall removalstogivethecoremorphologyanarrowend.Theotherfaceofthecore(thenon-knappingsurface)isusuallypreparedwithcentripetalremovals.TheseHPcoreswerealsoprobably intendedforblade/bladeletproduction.AsmentionedindelaPeñaandWadley ’ s[ 36 ] Table6.GreyRockycoretypesbyrawmaterialtype. * Seequartzsubtypescoresin Table7 . CORESGRHornfels (N) Hornfels (%) Dolerite (N) Dolerite (%) Quartz (N) Quartz (%) Quarzite (N) Quarzite (%) Sandstone (N) Cryptocristalline (N) Centripetalcore 49.527.400111.101 Multifacialcore 12.413.70000.000 Bladeletcore(from asmallnodule) 49.5414.8717.500.000 Prismaticblade core 12.413.70000.000 CoreonFlake (burin-like,endscraper-like, Kostienki ) 819.0622.200111.100 Bipolarcore 921.4414.82972.5666.710 HowiesonsPoort Core 819.0311.10000.000 Indeterminate 716.7622.2410111.100 TOTAL42 100 27 100.0 40 100 9 100 11 doi:10.1371/journal.pone.0143451.t006 Table7.Typeofcoresinquartz(thetypes,suchasB1andB2)areillustratedin Fig4 . Ind=indeterminate. Type N B1 2 B2 22 B3 2 B4 3 F10 3 F2 2 F3 1 F9 1 Ind 4 TOTAL 40 doi:10.1371/journal.pone.0143451.t007 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 16/52

PAGE 17

discussionofSibudu ’ sGSlayer,somecoresonflakesfollowthisHPcore-morphologicalpatternandthesetypesarerecognizedinthePaleolithicliteratureas NahrIbrahim or Kostienki cores[ 76 – 80 ](see Fig7 #1). Bladeprismaticproductionisrevealedbyonedoleritecore( Fig8 )andonehornfelscore. Thedoleriteexampleexhibitsachangeofknappingdirectionduringthereductionsequence, whichisacharacteristicthatisevidentelsewhereinthehighpercentageoffalsesemi-crestsin thismaterialtype,aswellaswithhornfels( vid . infra ). Inaddition,therearemultifacialcoresandcoreswithcentripetalscarpatterns( Table6 ). However,noneoftheseexamplesfallswithinthe Levallois technologicaldefinition[ 81 – 83 ].In otherwords,theyhaveacentripetalscarpattern,buttheydonothavetwohierarchicalsurfaces characteristicof Levallois cores.Importantly,thesecoresdemonstratethattherewasaninterest intheproductionofflakesfromhornfelsanddolerite. Thepercentagesofcoresonhornfelsanddolerite(withanotablerepresentationofcoreson flakes)suggestthattherewasagreatemphasisonbladeletandsmallflakeproduction.This would,however,beabiasedconclusionbecausethestudyofthenon-retouchedblanks( vid . infra )revealsthattheknappingstrategieswereverybroad,anddidnotonlyfocusonsmall blanks. Quartzcoresaredividedintotwomaincategories:freehandbladeletprismaticcoresand bipolarcores( Fig4 ).Thedifferentsubtypesconsideredinthisstudy( Fig4 )arepresentedin Table7 .Itseemsthatthefreehandbladeletcoreswererecycledintobipolarcores,aswasthe caseforquartzmanagementintheGSlayer[ 37 ].Indeed,thepercentagerepresentationoffreehandandbipolarcoresinGRissimilartothatintheotherHPlayers(GSandPGS)see[ 84 ] and Fig4 withinit.Thedistinctionbetweenfreehandandbipolarcoresisclearfromaqualitativepointofview(seeinthisregard Fig9 ).Bipolarcoreshaveabundantfissurationatthe Fig6.GreyRockycoretypesbyrawmaterialtype. doi:10.1371/journal.pone.0143451.g006 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 17/52

PAGE 18

overhang,andbluntnessofthestrikingplatform,whereasconchoidalscarsarevirtuallyabsent [ 37 , 84 ].Onthecontrary,freehandquartzcoreshavetheoppositetendency(see Fig10 forthe percentageofthesequalitativecharacteristics).Therecyclingoffreehandcoresintobipolar coresisalsosupportedbytherepresentationofcortexinthesetwocategories,see[ 84 ]and Fig 8 withinit.Ihavenotusedstatisticalanalysistodistinguishbetweenthesetwotypesofcores becauseofthesmallnumberoffreehandquartzcores(n=9)inGR.Inthecaseofquartz,freehandprismaticbladeletcoresandbipolarcoresexplainmostofthequartznon-retouched blanks.However,asdemonstratedindelaPeñaetal.[ 41 ],thereisalsoquartzbifacialpoint productioninGR,andthesepointsweremadeonflakes.However,therearenodiscoidalor centripetalcoresforobtainingblanksforthequartzpoints. Thereisanexhaustedcentripetalcore( Fig11 #1)madefromariverpebbleofCCS.There arealsosomebipolarquartzitecores( Fig11 #2,3and4)andonecentripetalquartzitecore. Thesandstonecoresdonotgiveusmanycluesaboutthemanagementofthisrocktypebecause Fig7.DoleriteandhornfelscoresfromGRlayer. 1Coreonflake( Kostienki ),hornfels.2and3HPcore,hornfels.4.HPcore,dolerite.5.Coreonflake (burincore),dolerite.6.Coreonflake(burincore),hornfels.7.Bladeletcoreonsmallnodule,hornfels. doi:10.1371/journal.pone.0143451.g007 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 18/52

PAGE 19

thereisonlyonesandstonebladeprismaticcoreanditisdifficulttoreadowingtothecharacteristicsofthisrocktype. Whatthetypometricanalysisofthecoresshows. Thetypometryofthecoresshowsthat thereisacleardifferenceinlength,breadthandthicknessforthreeoftherocktypes:dolerite, hornfelsandquartz( Fig12 ).Thisdifferenceisespeciallyacuteinthecaseofquartz,wherethe majorityofpiecesislessthan2cminlengthbecauseofthefinalbipolarreductionofthese cores,whichsupportsatrulymicrolithicstrategyofrecyclingfortheproductionofsmall blanks. AnANOVAanalysisverifiedthisintuitiveperceptionforthecomparisonoflength,breadth andthicknessofthesethreerocktypes[p(same)forLength:1.215E-13;Breadth5.467E-15; Thickness:5.005E-6].Moreover,theproofof ‘ honestlysignificantdifferenceofTukey ’ also gaveusnoteworthydifferences,particularlyforlength,wherethedifferencesweresignificant betweeneachofthethreerocktypes.TheCoreRelatedBy-ProductsThecorerelatedby-productsthatIseparateinthisstudyareonlyinhornfelsanddolerite(Figs 13 and 14 ).FortherestoftherocktypesIdidnotfindanyblankthatcouldbeincludedwithin mydefinitionofcore-relatedby-productsasgiveninthe Methodology section.Ascanbeseen in Table8 ,mostofthecorerelatedby-productsinhornfelsanddoleritefallwithintwobroad categories.Ontheonehand,manyofthemwereproducedtostartorcorrectbladeprismatic cores(crestandsemicrests)( Fig13 ).Ontheotherhand,thereareseveralblankswhichdenote Fig8.PrismaticbladecoreexampleinGR. A.Prismaticbladecore(#61)(dolerite)showingachangeofdirectionintheknapping.1.Firststrikingplatform andexploitationsurface.Afterwardsthecorewasrotatedinordertocontinuethebladeproduction.2.Secondandthirdstrikingplatform(opposed) and knappingsurface.B.Threedoleritebladeblanksthatcouldcorrespondtoaprismaticcoreasexample#61.2.Isanovershoot. doi:10.1371/journal.pone.0143451.g008 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 19/52

PAGE 20

achangeofdirectioninthereductionsequenceofbladeprismaticcores;thisiswhatIhave decidedtocall ‘ falsecrestsandsemicrests ’ (see Methodology section)( Fig14 ).Theseby-productsofthebladeprismaticreductionsequencemightsignifythatbigprismaticbladecores wereheavilyreducedbyachangeofknappingdirectionduringtheproductionofblades(as shownbythedoleritebigbladeprismaticcoreof Fig8 ). Fig9.FreehandandbipolarquartzcoresinSibudu,layerGR. Scale1cm. doi:10.1371/journal.pone.0143451.g009 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 20/52

PAGE 21

TheBlankswithoutRetouchDothenon-retouchedblanksshowevidenceforotherknappingmethods?. Thecompleteblanksaremainlydolerite,hornfelsandsandstoneflakes(see S1File ).Thesearethethree mainrocktypesinthislayer( Fig3 ). Duringtheanalysistypicalby-productsof Levallois productionhavebeenobservedfordolerite.Thissituationisquitedifferentfromwhatwasseenduringthecoreanalysis.Someofthe Levallois flakesinthisassemblagefollowthecanonicaltypologicaldescriptionofthesetypeof blanks:symmetricalflakeswithcentripetalscarordinations(morethanthreenegatives),longer thatwider,andwithrectangularorquadrangularshapes[ 85 ].However,theplatformsarenot alwaysfaceted,andplainanddihedralplatformsareabundant.Fordoleritetherearealsosome flakeby-productsthatcouldcorrespondtoadiscoidalknappingmethod,buttheycouldalso comefromapreparatoryphaseofa Levallois core( Fig15 ).Asotherresearchershavepointed out[ 86 ],itisnotaneasytasktodetect Levallois reductiononlyfromtheblankproduction. Furthermore,makingadistinctionbetweendiscoidaland Levallois basedonblankproduction isparticularlyproblematic.Indeed,thetopichasbeenintenselydebatedinPalaeolithicliteratureforsomeyears[ 87 ].Nonetheless,whatitisinterestingisthatitisclearthatfordoleritethe productionofflakeswasintentionalandwasnotonlyasubsidiaryofbladeproduction.The typeofdoleriteblanksillustratedin Fig15 #1to4,seemclearlyrelatedtoa Levallois flake reductionsequence.Inaddition,someofthedoleriteflakesoriginatefromthepreparationof bladecores. Inhornfels,onthecontrary,Ididnotfindanytypical Levallois flakes.Inaddition,mostof thecompletedblankswithoutretouchseemassociatedwithblade/bladeletproduction.However,itseemsthattherewasalsodeliberateproductionofhornfelsflakes,asisalreadydemonstratedfromthecentripetalandmultifacialcores. Moreover,fordoleriteitisclearthatbigprismaticbladeswithunidirectionalorbidirectionalscarpatternsandtriangularortrapezoidalsections( Fig8B )probablycomefromprismaticcoresliketheoneshownin Fig8A .Thesametypeofbigprismaticbladealsooccurson hornfels.Moreover,thecoresandthecore-relatedby-productsconfirmthistypeofreduction Fig10.Percentageofconchoidalnegatives,fissurationandbluntnessoffreehandandbipolarcoresinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g010 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 21/52

PAGE 22

bythepresenceofmanyfalsesemicrests( Fig14 ).Therefore,theinformationfromtheblanks withoutretouchsupportsthatfromthecoresandthecorerelatedby-products.However,the abundanceofthesebigprismaticbladesseemsveryimportantjudgingfromthenumberof Fig11.ExamplesofCCSandquartzitecoresinSibudu,layerGR. 1.Centripetalcore( Levallois -like)onCCS.2,3and4.Bipolarquartzitecores.Scale1 cm. doi:10.1371/journal.pone.0143451.g011 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 22/52

PAGE 23

them,whereasinthecoresampletheywereweaklyrepresented(onlyoneprismaticbladecore forhornfelsandonefordolerite). Inregardtotheblade/bladeletproductionforhornfelsanddolerite,itmustbestressedthat probablythethinbladeswithslightcurvaturesinthedistalpartoftheirprofileswereproduced fromHPcoresandnotfromprismaticcores(butexperimentsshouldbeperformedinorderto confirmthissuggestion). Sandstoneisthethirdrocktyperepresentedintheflakestudyanditalsoreflectsasignificantcentripetalflakeproductionandabigbladeproduction. Fig12.Box-plotofthelength,breadthandthicknessofdolerite,hornfelsandquartzcores. Besides doleritelengththereis ‘ LengthCR ’ ,whichisthebox-plotofdoleritecrestandsemicrestandfalsesemicrest, asitcangivealsoanideaofthelengthofthecores.Ascanbeseen,themeasurementsareclearlydifferent amongrocktypes,onlybreadthofdoleriteandhornfelsshowsomeoverlap. doi:10.1371/journal.pone.0143451.g012 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 23/52

PAGE 24

Fig13.Corerelatedby-productsexamplesinSibudu,layerGR. 1.Overshoothornfelsflakeshowingthe morphologyofaprismaticcorewithtwoopposedstrikingplatforms.2and3.Semicrestinhornfels.4. Semicrestindoleritefromabigprismaticbladecore.Scale3cm. doi:10.1371/journal.pone.0143451.g013 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 24/52

PAGE 25

Allthepreviousremarksweremadefromtheidentificationoftypicalflakesandfromthe shapeoftheblanks(see S1File ).Therefore,itcouldbearguedthatthestudywasmadesubjectively.Inordertogiveamoreobjectiveperspectiveonthestrategiesofknappingitwouldbe appropriatetorecordthefrequenciesofvariousqualitativeattributes.Moreover,thistypeof approachcangiveusabetterideaoftherepresentationofthedifferentknappingmethodspreviouslydocumentedqualitativelyforeachoftherocktypes. Itisinterestingtoseethescarpatterndistributionforthedifferentrocktypes.Ascanbe seenin Fig16 ,doleritehasahighpercentageofcentripetalandsubcentripetalscarpatterns (~36%),whichmakessensewith Levallois ordiscoidalstrategies.Theunidirectionalscarpatternisalsoimportant(~37%),anditmainlycorrespondswiththebladeknappingmethods thatcharacterizesomeoftheblanksandthecores(prismaticbladeproduction,HPcoresand coreonflakes).Incontrast,thepercentageofunidirectionalscarsonhornfelsandsandstoneis higherthanfordolerite,whichmightindicatethatforthesetworocktypesblade/bladelet knappingmethodsweremorecommon(~50/53%)thanflakingmethods. Fig14.Corerelatedby-products-FalsesemicrestexamplesinSibudu,layerGr. Theseblanksaretheresultfromachangeinthedirectionofblank removalwiththecrestrepresentingapreviousoverhang.Onthetopleftschematicdrawingexplainingtheproductionofthesetypesofblanks.1,2,3a nd4 doleriteexamples.1,2and3arecomingfrombigbladeprismaticcoresastheoneshownin Fig8 .5and6.Hornfelsexamples.Scale1cm. doi:10.1371/journal.pone.0143451.g014 Table8.Numbersandpercentagesofcorerelatedby-productsinGRfordoleriteandhornfels. Corerelatedby-productsinGRDolerite%Hornfels% Crest 24.5500 Semicrest 715.9210.53 Cleaning akes 511.400 Falsesemicrest 2045.51368.42 Flakeshowingcoremorphology 715.9210.53 Knappingaccidents 36.82210.53 TOTAL4410019100 doi:10.1371/journal.pone.0143451.t008 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 25/52

PAGE 26

Thissamelineofreasoningcanbeappliedifwelookattheshapeoftheblanks.Ontheone hand,ascanbeseenin Fig17 ,doleritehasahigherpercentageofexpandingshapeblanksthan theotherrocktypes.Ontheotherhand,hornfelsandsandstonehaveahigherpercentageof Fig15.DifferentexamplesofdoleriteflakesinSibudu,layerGR. 1to4seemsrelatedtoa Levallois reductionsequence,whereaspiecessuchas6or11couldberelatedtoadiscoidalknappingmethodora preparatoryphaseofa Levallois core.Scale3cm. doi:10.1371/journal.pone.0143451.g015 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 26/52

PAGE 27

parallel-edgedblanksthandolerite;thiswouldalsosupporttheideathatflakingmethods(such as Levalloi s,discoidalormultifacial)werefavouredmoreoftenfordoleritethanforhornfels andsandstoneinthislayer.Theseinterpretationsaresupportedby Fig18 whichillustratesdorsalscarpatternandshapeoftheblankforhornfelsanddolerite.Ascanbeseen,doleriteflakes withexpandingandconvergingedges(whichareconsistentwithflakeproduction)havea higherpercentageofcentripetal/subcentripetalandunidirectionalconvergentscarpatterns. Thesamelogicappliestohornfelswhereconvergingandexpandingflakeshapeshaveanotable representationofunidirectionalconvergentandcentripetal/subcentripetalscarpatterns. Furthermore,theChi^2analysisofthescarpatternandshapeoftheblankinthethree mainrocktypesdemonstratesthatthedifferentcategorieshaveremarkablydifferentgroupings.Thismightimplydifferentmanufacturingstrategiesforhornfelsanddolerite(sandstone frequenciesweretoosmalltoperformtheChi-squaretest)[ScarPattern:Dolerite:Deg.freedom:4,Chi^2:188.42,p(same):1.1594E-39.Hornfels:Deg.freedom:4,Chi^2:60.173,p Fig16.Flakescarpatterningfordolerite,hornfelsandsandstoneinSibudu,GRlayer. doi:10.1371/journal.pone.0143451.g016 Fig17.Lateraledgefordolerite,hornfelsandsandstoneinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g017 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 27/52

PAGE 28

Fig18.DorsalscarpatternandshapeoftheblankforhornfelsanddoleriteinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g018 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 28/52

PAGE 29

Fig19.TypeofplatformanddorsalscarpatternforhornfelsanddoleriteinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g019 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 29/52

PAGE 30

Fig20.TypometricaldistributionofplatformsflakesbyrocktypeinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g020 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 30/52

PAGE 31

(same):2.6681E-12.Shapeoftheblank:Dolerite:Deg.freedom:4,Chi^2:501.3,p(same): 3.5144E-107.Hornfels:Deg.freedom:4;Chi^2:91.323,p(same):6.8922E-19]. Thefrequenciesofplatformtypesgivecluestothetrendsofknappingforthedifferentrock types.Forexample,hornfelshas18.35%offacetedplatforms,probablyrelatedtotheblade/ Fig21.CortexpercentageandscarpatternforhornfelsanddoleriteinSibudu,layerGR. doi:10.1371/journal.pone.0143451.g021 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 31/52

PAGE 32

bladeletproduction.Doleritehasahigherpercentageofdihedralproduction,whichIemphasizedearlier.Finally,sandstonehasthehighestpercentageofplainplatforms,butthismightbe acharacteristicofthisrocktypewhichisverydifficulttoshape.Theseobservationsareconfirmedin Fig19 whichshowdorsalscarpatternandplatformtypeforhornfelsanddolerite.As canbeseen,flakeswithfacetedplatformsinhornfelshavepredominantlyunidirectionalor bidirectionalscarpatterns,whichisconsistentwithblade/bladeletknappingmethods.Onthe otherhand,doleriteflakeswithdihedralplatformsshowpredominantlyunidirectionalconvergentandcentripetal/subcentripetalscarpatterns,whichareconsistentwith Levallois /discoidal reduction. AgaintheChi^2analysisoftheplatformtypesinthethreemainrocktypesimpliesthatthe differentcategoriesappearingroupsthatareremarkablydifferent,whichmightpointtodifferentstrategiesforeachoneofthesethreerocktypes[Dolerite:Deg.freedom:5,Chi^2:1409.9,p (same):9.7987E-303.Hornfels:Deg.freedom:5,Chi^2:251.67,p(same):2.4077E-52]. Theknappingtechniquesappliedtothebladeproductionaredifficulttodetermine.These aretechnologicaltraitsthathavebeenmentionedinpreviousHPtechnologicalstudies[ 6 , 25 , 33 , 35 ].Itmustbetakenintoaccountthathornfelsanddoleritehaveveryparticularmechanicalproperties[ 56 ];andthatthemacroscopiccharacteristicsrequiredtodistinguishknapping techniquesusuallycomefromflintexperimentalprogrammesand,therefore,theyshouldnot beuncriticallyappliedtohornfelsanddolerite.Indeed,recently,eventhequalitativeapproach torecognizingknappingtechniquesforflinthasbeenquestionedfromastatisticalanalysis pointofview[ 88 ].Idecidednottodrawfurtherconclusionsaboutdoleriteandhornfelsknappingtechniquesuntilknappingexperimentsaredone.Thedifficultiesassociatedwithdeterminingknappingtechniquesofthesetworocktypesarealsoevidentwhenwelookatthe typometricaldistributionoftheplatformflakesbyrocktype.Ascanbeseen( Fig20 ),dolerite hassystematicallybiggerandbroaderplatforms,whichisprobablyduetoitsmechanical properties. Itisalsousefultoexaminetherelationshipbetweencortexpercentageandscarpatternfor hornfelsanddolerite.Ifwefocusonthelaststageofreduction,whentheflakeshave1to10% orevennocortex,adifferentpatternforhornfelsanddoleritecanbeobserved( Fig21 ).Doleritehasamoreabundantpercentageofcentripetalandsubcentripetalscars( > 50%),whereas hornfelsunidirectionalscarsaregreaterthan50%.Inotherwords,itseemsthatforhornfels, eveniftherewasflakeproduction,thepursuitofblade(orelongatedflakes)wasvaluedmore. Meanwhilefordoleritetheoppositetrendwasobserved(flakesweremoreimportantthan bladeproduction). DuringtheanalysisofcompleteblanksImadethedistinctionbetweenflakesandblades usingtechnologicalcriteria( vid . supra ).Ascanbeseenin Table9 thepercentageofbladesin doleriteisgreaterthanthatofhornfelsandsandstone,whichgoesagainstmypreviousstatementsaboutthefrequenciesofscarpattern,shapeofblanks,platformtypesandcortexpercentages(withdoleritehavingagreatertendencytowardsflakeproductionthanblade production,andtheoppositetrendforhornfels).However,thehigherproportionofdolerite Table9.Sibudu,layerGR.Technologicalrecognitionofflakesandblades/bladeletsforhornfels,doleriteandsandstone. Technologicalrecognitionof akesandblades/bladeletsDoleriteHornfelsSandstone N%N%N% Blade26524.5823422.6321321.7 Flake81375.4280077.3776878.3 TOTAL10781001034100981100 doi:10.1371/journal.pone.0143451.t009 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 32/52

PAGE 33

(technological)bladesmightberelatedtothefactthatdoleritetendstofractureaccidentally muchlessthanhornfels[ 89 ]andsandstone;andthatinthisregardthefrequenciesofscarpatternandplatformtypesarereliable. Whatisthetypometryofthecompleteblanksshowingus?. Inregardtothetypometry ofalltheblanks(withoutdistinguishingbytechnologicalcategory, Table10 )( Fig22 )theShapiroWilknormalitytestshowshowthethreemainmeasurements(length,breadthandthickness)forflakesdonotshowanormaldistribution( Table11 ).Thisisnotsurprisingbecause,as Ihaveexplainedinthemethodologysection,inafirstanalysisoftheblankswithoutretouchI didnotmakethedistinctionbetweenflakesandbladesfromatechnologicalpointofview (evenifthetechnologicaldistinctionwasrecordedinmydatabase)andIconsideredeverything asflakes.Aswashighlightedintheprevioussections,itisclear,fromthequalitativecharacteristicsofthecores,thecorerelatedby-productsandtheblankswithoutretouchthatdifferent knappingmethodswereinvolvedinsideeachoneofthemanagementstrategiesofthedifferent rocktypes.Thereforetheseoverarchingsetsofdata( Table10 )resultfromdifferentknapping methodswithinthesamerocktype.Forexample,fordoleritewehaveseensofarthatthere probablywasbigbladeproductionfromprismaticcores,blade/bladeletproductionfromHP cores,bladeletproductionfromcoresonflakesandflakeproduction.Consequently,itisnot surprisingthatthelengthorthebreadthofthedoleriteflakesdoesnothaveanormaldistribution,becausedifferentpopulationsofblankscomefromdifferenttypesofknappingmethods. Itisinterestingtoobservethathornfelsanddoleriteblankshavelengthsandbreadthswith averysimilardistribution( Fig22 ,upperpart).Doleriteis,however,alwaysthickerincomparisontohornfelsandsandstone;butthiscanbeanintrinsiccharacteristicofthedolerite.Ascan beseenin( Fig22 -upperpart),doleritedisplaysaclearlyunimodaldistributionwhilehornfels andsandstoneproduceabimodalcurve(whichisclearerforsandstone). Itisalsonoteworthythatifweonlytakeintoaccounttheblade/bladeletblanksindolerite, hornfelsandsandstone(usingatechnologicaldistinction),theShapiroWilktestforlength, breadthandthicknessdonotgiveanormaldistribution(exceptforthethicknessofhornfels blade/bladelets)( Fig22 -lowerpartand Table12 ).Thisisprobablybecausetherearedifferent groupsofblade/bladeletsthatcomefromdifferentbladeknappingmethods(coreonflakes, bladeletcoresfromsmallnodules,bigprismaticcoresandHPcores).Indeed,thehistogramof theblade/bladeletby-productsisnotgivingunimodaldistributionsforeitheroftherocktypes Table10.Sibudu,layerGR.Univariatestatisticsoflength,breadthandthicknessofallcompleteblanksinthethreemainrocktypes. Length Dolerite Length Hornfels Length Sandstone Breadth Dolerite Breadth Hornfels Breadth Sandstone Thickness Dolerite Thickness Hornfels Thickness Sandstone N 710.00150.00119.00710.00150.00119.00710.00150.00119.00 Min 9.8011.6312.437.038.2314.091.841.543.24 Max 99.6963.0193.0395.7148.87127.2323.4913.3032.28 Mean 28.7525.5533.5426.8521.5432.586.794.859.29 Std.error 0.460.801.550.360.591.770.110.170.45 Variance 148.0696.08285.0994.2151.64374.198.174.2624.03 Stand.dev 12.179.8016.889.717.1919.342.862.064.90 Median 25.7323.6827.4824.7121.5526.786.274.677.97 Skewness 1.711.351.411.500.763.061.141.542.12 Kurtosis 4.301.971.674.551.2811.222.123.715.99 Geom. mean 26.6823.9630.1625.3420.3729.256.254.488.36 Coeff.var 42.3238.3750.3436.1533.3759.3842.1042.5452.79 doi:10.1371/journal.pone.0143451.t010 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 33/52

PAGE 34

Fig22.Sibudu,layerGR. Above:Histogramsofallcompletedblankswithoutdistinguishingbetweenflakesandblades.Below:Histogramofonlythepieces recognizedasblade/bladeletstechnologically. doi:10.1371/journal.pone.0143451.g022 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 34/52

PAGE 35

(seeforexamplelengthandthicknessfordoleriteorbreadthforhornfels).Amixtureofanalyseswasappliedforthelengthofdolerite,hornfelsandsandstoneblade-blanksinorderto explorethepossibilitythatdifferentsetsofnormaldistributionsweremixedwithineachoneof thesegroups.Ascanbeseenin Table13 forthethreerocktypesthreegroupsofnormaldistributionweremorecoherentthanonegroup,whichsupportsthequalitativeideathatmorethan onemethodwasusedfortheproductionofblade/bladeletsineachoftherocktypes.TheRetouchedBlanksWhatisthegeneralrepresentationofretouchblanksbyrocktype?. Whentheretouched piecesareexaminedbyrocktype,thepercentagessuggestadifferentdistributionfromwhatwe haveseensofarinothertechnologicalgroups(cores,corerelatedby-products,etc.)Inthiscase doleriteandquartzarethemostretouchedrocktypes,followedbyhornfels.Inaddition,there aresomeretouchedpiecesinquartziteandCCS( Table14 ). Itisclearthatsandstone(thethirdmostcommonrocktyperepresentedinthislayer Fig3 ) isnotretouchedatall.Thisfactprobablymeansthattherewasahugeamountofrockknapped (andpotentiallyused)butveryseldomretouched.Inordertogiveanideaofthegeneraltrends forthemanagementofretouchedpiecesinthethreemainrocktypesIhavegroupedthemorphotypesintothreebigcategories:domestictools(withthiscategoryIrefertotoolsshapedby simpleretouchsuchasend-scrapers,borer,strangulatedpiece,notchandburin)( Fig23 ), backedtools( Fig24 ),andbifacialpieces( Fig25 ).Iacknowledgethatthesecategoriesare completelyartificialgroups.Nonetheless,thisdivisionofmorphotypes( Fig26 )simplifies Table14 considerablyand,thus,showseloquentlythatinfacthornfelsanddoleritehaveavery similarpatternofmorphotypepercentages,withbackedpiecesaround60 – 70%,followedby domestictools(3 – 5%)andaveryminorpercentageofbifacialpieces.Incontrast,quartzis mainlydirectedtowardtheproductionofbifacialpieces,asstatedinapreviousspecialized publicationaroundthisHPmorphotype[ 19 ]. Whattypesofblankswerepreferablyretouched?. Itisworthnotingthatthepreferred blanksforconversiontoretouchedpiecesarealwaysbladesforbackedpiecesanddomestic tools,andflakesforbifacialpieces.InallcaseswhereIcouldidentify(technologically)thetype ofblank,itwasalwaysabladeforbackedpieces.Amongstthedomestictoolstherearesome flakesusedasblanks,butremarkablyfew(n=8,6indoleriteand2inhornfels).However,for Table11.Sibudu,layerGR.Shapiro-Wilknormalitytestsoflength,breadthandthicknessbyrocktypeofalltheblanks(withoutdistinguishingby technologicalcategories). Length Dolerite Length Hornfels Length Sandstone Breadth Dolerite Breadth Hornfels Breadth Sandstone Thickness Dolerite Thickness Hornfels Thickness Sandstone W 0.87030.89270.86040.90570.96150.66980.93080.89040.8077 p (normal) 8.25E-245.19E-093.27E-091.57E-200.00033965.40E-151.38E-173.94E-093.53E-11 doi:10.1371/journal.pone.0143451.t011 Table12.Shapiro-Wilknormalitytestsoflength,breadthandthicknessbyrocktypeofallthebladeandbladeletsblanks(distinguishedbytechnologicalcriteria). Length Dolerite Length Hornfels Length Sandstone Breadth Dolerite Breadth Hornfels Breadth Sandstone Thickness Dolerite Thickness Hornfels Thickness Sandstone W 0.89440.90490.91770.96970.87260.92180.93110.97270.8523 p (normal) 2.712E-080.00090570.039640.0043429.163E-050.049533.813E-060.3208 * 0.001577 Thecaseswhichshownormalityarehighlightedwithanasterisk. doi:10.1371/journal.pone.0143451.t012 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 35/52

PAGE 36

thebifacialpiecesonquartzmostoftheblanks,whentheycanberecognizedare,onthecontrary,onflakes[ 37 ].Therefore,thisreinforcestheideathatmostoftheflakeproductionfor hornfels,doleriteandsandstonewasperformedwithoutthenecessityforshapingorresharpening(retouch). Whattypeofmorphotyperepresentationdowehavewithinthecategorybacked pieces?. Forthethreemainrocktypesitisclearthatsegmentsarethemostfrequentbacked pieces( Table14 , Fig24 ),followedbytruncations(see Fig5 tounderstandthesubtypesdistinguishedinthisstudy).Itwouldbeinterestinginthefuturetoinvestigatewhetherthereisa functionaldistinctionbetweenthesetwosubtypesofmorphotypes. Whatdoesthetypometrytellusaboutthebackedpieces?. Itisnoteworthythatthe lengthandbreadthofbackedpiecesshownormaldistribution( Table15 ).Thisisremarkable becausepreviouslyithasbeenshownthattheblade/bladelet(withoutretouch)blanksdonot showanormaldistributionandprobablytherewereatleastthreegroupsofbladeproduction (afteramixtureofanalyses, vid . supra ).Consequently,therewasaselectionofblanksforbacking.Indeed,oneofthemeansproposedbythemixtureanalysisofdoleritelength(mean32.66) matchesquitewellwiththemeanofthelengthofdoleritebackedpieces(mean:34.68)(cf. Tables 13 and 15 ). Furthermore,thestandarddeviationislessthanhalfthevalueofthemeanforlengthand breadthindoleriteandhornfels( Table16 ),whichmeansthatthevariabilityforthesetwo parametersislow.Forthicknessthisrulealsoapplies,butitislesspronounced. Inaddition,aFtestperformedforbreadthandthicknessshowedthattherearenotsignificantdifferencesinbreadthparameterfordoleriteandhornfels(see Table17 and Fig27 ).Iused thebreadthtocomparethemasmostofthepiecescangivemaximalbreadthbut,onthecontrary,manyofthesepiecesarebrokenintheirdistalorproximalpartand,therefore,lengthitis notsosuitabletocomparethem. IhavenotcomparedhornfelsanddoleritebackedpiecestoquartzonesbecauseinGRthere areveryfewbackedpiecesinthisrocktypeandnoneofthemiscompleted.However,the backedpiecesonquartzarenotablysmaller,aswasalreadynotedinGSlayer(wherethereisa biggersample,see[ 37 ]and Fig15 withinit)andalsoinPGS[ 42 ]. Howisthebifacialproductioncarriedout?. TheGRquartzbifacialpointsaremadeon flakesandaregenerallyunstandardizedinshapeandsize[ 19 ].Detailedtechnologicalanalysis Table13.Mixtureanalysisforthelengthofdolerite,hornfelsandsandstoneblade/bladeletsblanks. Dolerite:threegroups.Log:-412.4.Akaike:837.5 ProbMeanStdev 0.6011132.6587.9396 0.2739256.74217.059 0.1249621.7081.2309 Hornfels:threegroups.Log:-127.AkaikeIC:268 ProbMeanStdev 0.31845.7867.8336 0.2754521.8471.8279 0.4065529.0143.531 Sandstone:threegroups.Log:-74.9.AkaikeIC:166.2 ProbMeanStdev 0.1613866.64510.514 0.5781541.6665.5817 0.2604625.9671.7993 doi:10.1371/journal.pone.0143451.t013 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 36/52

PAGE 37

ofGRpointsrevealedashortreductionsequencewithoutsuccessivethinningstages(thisis mainlyduetothesizeofthequartzpebbles,~5-10cm).Thebrevityandsimplicityofthepoint reductionsequencemaypartlyreflectthehighriskofknappingaccidentswithquartz.Inapreviousworkfourphasesforthemanufacturingofthesepointswerehypothesized: 1.Selectionofoptimally-sizedflakeblanksmadefromdiscoidalorunpreparedcores;2. Minimalthinningwherebyunfinishedblanksdisplayattributesofeitherunifacialreduction (knappingoneflakefaceinitially,thentheother)orabifacialstrategywithalternatingblowsto thetwofaces,astepthatconflatesCallahan ’ s[ 90 ]thinningstages3and4;3.Shapingofthe blanktoarefinedpreform,probablymadewithacombinationofhardmineralhammerand softorganichammer,resultinginregularscars;4.Finalshapingofthepreformandfinished product,possiblywithasoftorganichammerandpressureflaking. InthisnewdetailedstudyofthetechnologyofGRIfoundsevenmorebifacialpiecesthat mustbeaddedtotheinventorypublishedindelaPeñaetal.[ 41 ].Fiveofthesepiecesare quartzandtwoarehornfels.Furthermore,oneofthenewpieces(aquartzbase)refitswithatip fromoneofthepointspublishedin[ 19 ]( Fig25 #2). Table14.FormaltoolsandretouchedpiecesinlayerGRaccordingtorocktype. DoleriteHornfelsQuartzQuartziteCCS N%N%N%NN Denticulate11.20 End-Scraper22.41 Borer11.45 Notch22.90 Micronotch 89.09 Burin22.90 Strangulatedpiece11.2022.90 Simpleretouch ake33.6122.9011.14 Simpleretouchblade67.2357.251 Marginalretouch ake11.2034.35 Ind.Retouch(simpleretouch)56.0257.2522.27 Ind.Segment33.611014.4933.411 S11315.66811.5922.27 S289.6422.9011.141 ST211.20 ST311.20 ST411.45 Tr111.2011.45 Tr211.2011.45 DT111.2011.45 T211.45 IndeterminateTruncation1315.6611.45 Rectilineartruncation11.20 Backedbladelet < 12mm11.201826.09 Ind.Backed1720.4866.821 Bifacialfragmentorcompletepieces33.6134.356573.861 TOTAL 83100.0069100.0088100.00 41 %byrocktype 33.8828.1635.921.630.41 doi:10.1371/journal.pone.0143451.t014 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 37/52

PAGE 38

Thequartzretouchproductionforthenonbifacialpieces. Quartzbladelettechnologyis presentinGRfirstfromfreehandprismaticbladeletcoresandthenfrombipolarcores.The productionwasfocusedonobtainingtinybladeletstobeusedwithoutretouch,ortobeconvertedintosmallbackedpiecesorothermorphotypes(suchassmallsingleordoublenotches).DiscussionTheGRtechnologicalstudyhasshownagreatvarietyofknappingmethods.Thehallmarksof theGRlithictechnologyare: Differentvarietiesofbladeknappingmethods(fromHPcores,bigbladeprismaticcores, coreonflakes). Astrongcomponentofflakeproductionthatforhornfels,doleriteandsandstonewasnot retouched. Theproductionofbackedtoolsofdifferentsizerangesandshapesforhornfels,doleriteand quartz.Segmentsarethemostfrequentmorphotype. Quartznodulesaremaximizedthroughreductionoffreehandbladeletprismaticcores,then bipolarcores. Fig23.DifferentmorphotypesfromlayerGR,generallyreferredtoas ‘ domestictools ’ inthispaper. 1,3and6Retouchedbladeonhornfelsand dolerite.2and4Strangulatedbladeonhornfelsanddolerite.5.Pieceswithmacrotracesondolerite.7and8Burinsonhornfels.Piecenumber8isabur in madeonfalsesemicrestfromabigprismaticbladedoleritecore. doi:10.1371/journal.pone.0143451.g023 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 38/52

PAGE 39

Theproductionofbifacialpiecesonquartz. Theuseofcoarsegrainedrocktypessuchassandstoneordolerite. Aftertheindepthstudyofthemanagementofthedifferentrocktypesitisclearthatdifferentobjectivesandstrategiesweredevelopeddependingontherocktypeathand,aspostulated inthemethodologicalanalysis(nullhypothesis,see Methodology section).Moreover,itseems thatthelownumbersofdomestictoolsisprobablyowedtothefactthatalotofactivitieswere performedwithnon-retouchedblanks,whereasshapedblanks(formaltools)suchasbacked implementsorbifacialpieceswereprobablydevotedtohuntingstrategies,andthissuggestion issupportedbypreviousresearchonSibudu ’ sHPbackedtools[ 41 – 44 ]. ThistechnologicalanalysiscomplementstherecentstudyofSorianoetal.[ 35 ]becausedifferentbladeknappingmethodshavebeenhighlighted.Moreover,inthepresentanalysis,flake productionhasbeendemonstratedashighlysignificant.Finally,therepresentationofquartz, evenifitisnotthemostfrequentlyusedrocktype,hasbeendemonstratedassubstantial becauseofthehighrepresentationofretouchedpiecesinthismineral(bifacialpiecesand backedpieces)andbecauseofthestrategyofmaximizingthequartzpiecesthroughreduction bybipolarknapping. Oncethedescriptionofthemaintechnologicaltrendsofthisso-calledHPassemblagehas beenperformed,thequestionslackingananswerare:howdoesthistechnologicaldescription Fig24.Backedtoolsinhornfels(left)anddolerite(right). Forhornelsontheleft:1.Trapeze.2,4,3,6and7.Segments.5.Singletruncation.Fordolerite (ontheright):1,5,7and8.Segments.2,3,4and6.Truncations. doi:10.1371/journal.pone.0143451.g024 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 39/52

PAGE 40

Fig25.BifacialpiecesinGRlayer. 1,2,3,4and7onquartz.Pieces#2and8arerefittings.5and6on hornfels. doi:10.1371/journal.pone.0143451.g025 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 40/52

PAGE 41

relatetootherassemblagesalsodefinedasHP?Moreimportantly,whatisthe technological idiosyncrasyoftheHP? InordertodiscussthevariabilitywithinHPassemblagesitisappropriatetopointoutthat therearedifferencesandsimilaritiesbetweenandwithinHPassemblagesalreadypublished fromothersites. FiveofthemaintechnologicalknappingmethodshighlightedforGRhavebeenalso detectedinotherHPsequences;letusfirstreviewtheevidence: ThesocalledHPcoresdescribedinWurz[ 25 ]andinVillaetal.[ 34 ]arepresentinhornfels anddoleriteinGR,buttheyarenottheonlybladeknappingmethodsforbladeblanksinHP assemblages.Aspointedoutpreviouslyprismaticbladeproductionandcoresonflakesplayan importantroleinthebladeproductionintheGS[ 36 ]andGRlayersofSibudu. Theflakeproductiondetectedinthisstudycoincideswiththediscoidaland Levallois flaking methodsdescribedforthemostrecentlayersofotherHPcontextssuchasKlasiesRiver[ 25 ], RoseCottage(layersETHandSUZ)[ 33 ],Diepkloof(layersascribedtothe ‘ LateHP ’ )[ 6 ]and Klipdrift(layersPBC,PBA/PBB,PAZ,andPAY,eventhoughthesediscoidalcoresoccur throughouttheHPsequence)[ 39 ].Theincreaseofflakeproductioncouldbeatemporal markertowardstheendoftheHPtechnocomplex. ThebipolarknappingproductionforquartzinGRseemstobearecurrentstrategyforthe maximumreductionofthisrockanditalsosupposescontinuitywithGSandPGSwithinthe Fig26.Representationofdomestictools,backedpiecesandbifacialpiecesforthethreemainretouchedrocktypesinGRlayer. doi:10.1371/journal.pone.0143451.g026 Table15.Shapiro-Wilknormalitytestsoflength,breadthandthicknessfordoleriteandhornfelsbackedpiecesinGR. Quartzwasnotincludedasit hasareallysmallnumberofcases. LengthDoleriteLengthHornfelsBreadthDoleriteBreadthHornfelsThicknessDoleriteThicknessHornfels W 0.94580.85840.92390.94120.73710.816 p(normal) 0.5762 * 0.09216 * 0.005128 * 0.1076 * 2.635E-083.439E-05 Caseswhichshownormalityhavebeenhighlightedwithanasterisk. doi:10.1371/journal.pone.0143451.t015 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 41/52

PAGE 42

HPSibudusequence[ 37 , 84 ].Theuseofbipolarknappinghasbeeninterpretedhereasthe mostefficientmethodofreducingquartzsmallcoresand/ornodules.Quartzwasprobably highlyvaluedasarockbecauseofitsmechanicalpropertiesanditsrelativescarcityintheVerulamarea.Somereasonsforitschoicearedifficulttoelucidate.Althoughabundantinthis area,thephysicaloccurrenceofquartzisonlyinsmallpebbles(5-10cm).Thus,thisknapping strategywasasolutiontomaximizetheproductionofbladeletsorsmallflakes,ashasbeen emphasizedinothergeographicareasandperiods[ 91 , 92 ].Bipolarknappinghasbeenalso observedinotherHPcontextsofsouthernAfrica.Wurz[ 25 ]mentioned outilsécaillés inthe HPsequenceofKlasiesRiverfromtheSingerandWymer1967/8excavations.Sheinterpreted themasproductsofextendedcorereductionandnotastools.Mackay[ 32 ]pointedoutthatin Diepkloofbipolarcoresarecommoninthe > 74,70 – 65,65 – 62and62 – 60kalayers.Meanwhile atKleinKliphuisbipolarcoresarecommonbetween62and60ka. AsstatedintheintroductionbackedpiecesaretheprincipaltypologicalmarkerintheHP [ 26 ].InthisstudyIhavegivenaspecificclassificationforthesetypesofpieces( Fig5 ).They maypointtodifferentfunctionalsolutions[ 16 ]andtheirdescriptionandrefineddefinition canhelpusinthefuturetounderstandbetterfunctionalvariationintheHP[ 42 ].Moreover, theymightbepointingouttheimpositionofattributesofstyle[ 26 ]andmayberegionalvariationswithinHP.Forexample,theobliquelybackedpointsconstitutethemainbackedmorphotypeintheHPlayersofRoseCottage[ 17 , 38 ],whereasinDiepkloofdifferentbacked morphotypesarepredominant,suchasthegeometricbackedtoolsintheearlyHPorthelarge backedpiecesintheMSAtype ‘ Jack ’ [ 6 ].InPinnaclePointtheHPisidentifiedinlayer ‘ Dark BrownCompactSand ’ (DBCS)(withthreeOSLdatesranging58±4to65±4ka)bybacked toolsandnotchedtoolforms.However,thissitehasanolderlayer, ‘ ShellyAshyDarkBrown Sand ’ (SABDS)withaweightedmeanOSLdateof71.1±2.3ka,alsowithbackedpieces;which Table16.UnivariatedatafordoleriteandhornfelsbackedpiecesinGR. LengthdoleriteLengthhornfelsBreadthdoleriteBreadthhornfelsThicknessdoleriteThicknesshornfels N 12.009.0046.0029.0052.0036.00 Min 24.5315.778.908.202.502.05 Max 44.4853.3630.6622.9716.5011.68 Mean 34.6830.1116.6013.855.474.41 Std.error 1.963.390.570.680.310.33 Variance 46.28103.6214.9513.435.083.81 Stand.dev 6.8010.183.873.672.251.95 Median 34.3328.6715.9113.765.133.89 Skewness 0.091.421.240.672.882.00 Kurtosis -1.313.843.210.6611.495.03 Geom.mean 34.0628.7416.2013.395.154.10 Coeff.var 19.6233.8023.2926.4741.2144.22 doi:10.1371/journal.pone.0143451.t016 Table17.FtestandTtestforbreadthofbackedpiecesinhornfelsanddoleritefromlayerGR. TESTS F: 1.1132p(same):0.77579 t: 3.0628p(same):0.00307 Uneq.vart 3.1007p(same):0.002902 Permutationttest(N=9999): p(same):0.0027 doi:10.1371/journal.pone.0143451.t017 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 42/52

PAGE 43

ispresentedasan ‘ unrecognizedadvancedstonetooltechnology ’ ( sic )withintheMSA[ 93 ]. ThemaindifferencestatedbetweenthesetwoPinnaclePointlayers(DBCSandSABDS)is, first,thetypometry/length-widthratioofthebackedpiecesand,secondly,SABDSbacked piecesarestatisticallydifferentfromtheKlasiesRiverHPbackedpieces.However,itmustbe stressedthatthereisagreatdealofmorphologicalandtypometricalvariationinHPbacked pieces[ 42 ].ThiswashighlightedinSibudu ’ sGSlayerwherequartzanddoleriteandhornfels backedpiecesdemonstratemorphologicalandtypometricalvariation,see[ 37 ]and Table12 and Fig15 withinit;andthisvariationalsooccursinGR(althoughisnotprovidedbecause mostofthequartzbackedpiecesareincompleteandhavemultiplefractures).InSibudu ’ sGR layerweseeapredominanceofsegments,butothermorphotypesarealsorepresented(see Fig 5 and Table14 ). BifacialpiecesarealsoacommoncharacteristicinboththeSibuduandDiepkloofHP sequences[ 6 ].However,theGRandGSbifacialpiecesaremainlymadeonquartz[ 41 ]whereas Fig27.Box-plotoflength,breadthandthicknessofdoleriteandhornfelsbackedtools.D=Dolerite, H=Hornfels. doi:10.1371/journal.pone.0143451.g027 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 43/52

PAGE 44

Fig28.ComparisonofpercentageofGRandGSbackedpieces. doi:10.1371/journal.pone.0143451.g028 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 44/52

PAGE 45

Fig29.ComparisonofpercentageofGRandGSdomestictools. doi:10.1371/journal.pone.0143451.g029 RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 45/52

PAGE 46

thepiecesfromDiepkloofaremadeonsilcrete.Itisdifficulttocomparethesetwoassemblages ofbifacialpointsasmostofthepiecesfromSibuduseemtobefailuresorfragmentsproduced duringtheknappingreductionsequenceandthereareveryfewexampleswhichappearfinished[ 41 ].Inotherwords,itisdifficulttovisualizeastandardizedpatternfromtheirshape. Apartofthesecommonandimportantfeaturesitmustbehighlightedthatnoneofthe techno-typologicalmodelsmentionedintheintroductionentirelyworksforSibudu,evenif thereareremarkabletechnologicaltrendsincommon.Oneofthemostevidentreasonsissimplytherepresentationofformaltools.Forexample,thebifacialproductionatSibuduappears mainlyinGR(themostrecentHPlayerinthissequence)andinGS[ 36 ].IftheHPfrom Sibuducorrespondstotheso-called ‘ classicHP ’ sensu Porrazetal.[ 6 ],itisquitecontradictory thatthebifacialproductionappearsinitsmostrecentlayers,whereasatDiepkloofthisispresentedasoneofthehallmarksoftheearlyphaseofHP( vid . supra tocomparethechronology fortheselayers),togetherwith piècesesquillées (forthissiteinterpretedastools,see[ 46 ]butsee [ 32 ])andthetruncatedpieces.Itcouldbearguedthatthebifacialpiecesanddifferentpoint productionatSibuduandDiepkloofconstitutevariationswithintheHPtechnocomplex which,indeed,seemsmorethanlikely. Anotherreasonforthediscrepancycouldbethatthe piècesesquillées inSibuduhavebeen interpretedasbipolarcores[ 36 , 37 , 72 ].IntheSibudusequencetheyarerepresentedinsimilar percentagesinallthreelayers[ 72 ],whereasinDiepklooftheyseemespeciallyabundantinthe earlyHPphase.InRoseCottagethebipolarknappingisprominentonlyinthepost-HPlayers [ 33 ]. AthirdtechnologicalcharacteristicwhichdoesnotcoincidewithDiepklooforwithRose Cottageisthedeteriorationofthebladeproductionpointedoutinthesetwosequences[ 6 , 33 ]. InSibuduIcouldnotdetectanysuchtraitinGR,Indeed,inGRthebladeknappingmethods seemverysimilartotheonesdocumentedinGS[ 36 ].Thevariationsbetweenthesetwolayers seemmainlyfunctional,expressedinGRbytheabundanceofflakeblanksorthepercentageof formaltools(Figs 28 and 29 ).Moreover,toevaluateknappingperformanceintheseterms seemtomehighlysubjective. AfourthSibuducharacteristicwhichdoesnotcoincidewiththeotherHPsequencesisthe mainknappingmethod.Forexample,atSibuduthebladeproductiondoesnotonlycomefrom HPcores(asinKlasiesRiver[ 34 ]),eventhoughthistypeisrepresented.Moreover,prismatic bladeproductionseemsmoreprominent.Wehavealsoseen,fromthestudyofRoseCottage, thatthischaracteristiccanbeadaptedtotherawmaterialavailablelocally[ 33 ]. Aspointedoutintheintroductionthesediscrepanciesbetweensitesmightcomefromdifferentreasons:differentdiachronicvariations,functionalvariationsormaybedifferentregional variations(differencesinstyle)withinwhathasbeenrecognizedtraditionallyandtypologically asHP[ 5 , 94 ].Furthermore,itshouldberememberedthatsomeofthetechnologicalcharacteristicspointedoutherecoincidewithothertechnologicaldevelopmentswithintheMSA[ 95 ]; forexampletheflakeproduction( Levallois anddiscoidal),thebifacialpieces,thebigprismatic bladeproduction,theimportanceoflocalrocktypesandthebipolarknapping. Afterallthesecomparisonsthequestionis,onceagain,whatdefinesHP?Moreover,whatis HP?Ashighlightedintheintroduction,HPhasbeendefinedasanindustry,MSAentity,horizonmarker,technocomplexoratechno-tradition.Moreover,evenifanimportanteffortto definethelithictechnologyassociatedwithithasbeenmadelately,thetruthofthematteris thatitisstillmainlyrecognizedbythetypology.IdonotrecallanySouthAfricanMSA sequencenamedHPonlybyitstechnologywhenthereisalackofbackedpieces.However, withsomeofthetechnologicaltrendshighlightedhereandinpreviousworks(suchasHPcore reductionsequence,bigprismaticbladeproduction,coreonflakes,flakingmethods,bipolar knapping,theimportanceofbackedmorphotypesandthepresenceofabifacialreduction RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 46/52

PAGE 47

sequence)maybeinthefuturewecangofurtherinordertorecognizethistechnocomplex uniquelybyitstechnology.Inotherwords,thetechnologicaltrendshighlightedatthebeginningofthisdiscussion,andthisspecifictechnologicalanalysis,basedoninter-sitecomparisons,supportsthecontemplationofHPasatechnocomplex. RecentlyMackayetal.[ 96 ]:36highlightedvariousbroadinter-regionalconsistenciesinHP assemblagesthatgivethemaunity:thebasicformsofimplements,theproductionofsmall bladesusingsimilarflakingsystems,assemblagesizesandtheon-sitereductionofcoresand productionofimplements. ThenewtechnologicalstudiesarerefiningourunderstandingofthisHPtechnocomplex(as thispaperhastriedtocontribute).ThefactthatGRsharestechnologicalcharacteristics,but alsoshowssomeremarkablediscrepancies,mightbepointingoutregionalvariationofthis technocomplex.Moreover,whenwewanttocomparedifferentassemblages,thedifficulty comeswiththecurrentchronologicaldebatewhichmakesthistaskhighlyproblematic.Aswe arediscussingvariabilitybetweendifferentsitesitmustbepointedoutthatitisextremelycontentioustocorrelateHPassemblageschronologicallybecause,asexplainedearlier,withthe currentchronologicalmodels[ 3 ]vs[ 47 ]thereisabigdifferencebetweenatechnocomplex consideredasa ‘ shortlived ’ culturalphenomenon(5ka)(in ‘ prehistoric ’ terms)asopposedtoa technocomplexlasting45-50ka.FollowingtheDiepkloofmodel[ 6 , 47 ]GRlithictechnology shouldbecomparedwiththeIntermediateorLatephaseofHP.Moreover,withthecurrent discussionaroundthechronologyofHPandSB,itisdifficulttoknowwhichclimaticfactors influenced(andifeffectivelytheyhadaninfluence)theappearance,developmentanddisappearanceoftheHP.Thisisquiteabigissuebecause,aspointedoutbefore,themajorityof interpretationsoftheHPhavebeenmadewithinanecologicalparadigm. Inanycase,GRtechnologydemonstratesaspecificmomentintimewithintheHPtechnocomplexand,asspecifiedbefore,probablyaregionalexpression.Inthissenseitisrepresentativeofthisareaatthisspecificperiodwhich,followingJacobsandRobert[ 49 ],is61.7±2ka, obtainedfromsinglegrainopticallystimulatedluminescenceonsedimentfromGR2.The SibuduGRlayeraddsinterestinginputtotheinterpretationofthevariabilityofHPinthiseasternpartofsouthernAfrica.AftertheanalysisoftheGSlayer[ 36 , 37 ]andthisnewGRstudy,it isevidentthatmostoftheknappingmethodsrepresentedinGSappearedalsointhismore recentlayerGR.Thus,thereisanevidentcontinuityinthisregard.Moreover,thesamegroups offormaltoolsarerepresentedwithslightvariations.Therefore,thereisanevidentcontinuity intermsoftechnologyandthevariationsofformaltoolpercentagescouldbeinterpretedin termsoffunctionality(Figs 28 and 29 ). ThedetailedtechnologicalanalysisoflayerGRhasshowntheimportanceofbladeproductionbutalsoofflakingmethodsincoarsegrainedrocktypes.Moreover,newstrategiesofbifacialproductionandmicrolithismwerereallyimportant(asthemanagementofquartzhas shown).Thus,GRlithictechnologyshowsareallyversatileexampleofreductionstrategiesthat werehighlyinfluencedbythecharacteristicsoftherocktypes.Afterthisdetailedanalysisthe challengeistoevaluatehowtheHPchangesthroughtimeandwhy.Thisisalsoimportantfor understandingtheputativechangefromHPtotheso-calledpost-HP[ 97 , 98 ]. TheGRlithicindustryrepresentsadiversemanagementofrocktypes;severalstrategies highlightedinthispapercoincidewiththoseinothersequencesinsouthernAfrica.Incontrast, othercharacteristics,suchasthemanagementofquartz,seemsitespecificwithinthesame technologicaltradition.However,thegeneralcharacteristicspointedoutbyMackay[ 96 ]for HP(thebasicformsofimplements,theproductionofsmallbladesusingsimilarflakingsystems,assemblagesizesandtheon-sitereductionofcoresandproductionofimplements)can befoundinallSibuduHPlayers.Ontheonehand,thetechnologicalstudiesfromsiteslikeKlasiesRiver,Diepkloof,KleinKliphuis,RoseCottageandSibudu,demonstratethatthereare RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 47/52

PAGE 48

commontechnologicaltrendsduringthelatePleistocene,buttheystillneedtobeproperlycircumscribedchronologically.Ontheotherhand,HPcharacteristicssuchasthebifacialproductioninquartzarereminiscentofproductioninsomeSBorpre-SBindustries(e.g.theso-called Pietersburg)andtheflakeproductionortheprismaticbladeproductiondescribedherecould beapointincommonwithpre-SBandpost-HPindustries[ 99 ]. InthispaperIhaveinquiredintothetechnologicalcharacteristicsofaso-calledHPassemblage,andcompareditinordertoseewhetherthe ‘ technocomplex ’ definitionresiststhiscomparison.IbelievethatadetailedtechnologicalandcomparativestudyofHPandotherMSA industriescanattenuatethecharacterofextremeoriginalitypresupposedforHPwithout removingitstechnicalpersonality.Moreover,thefactthattypologicallythisindustryhas attractedsuchattentionmighthavecreatedatendencytohighlightitsuniquestatuswithinthe MSA[ 100 ].Inthesamesensethatstudiesfromotherarchaeologicalbranches,suchasarchaeobotanicalremains,arepointingoutadvancedtechnologieswithinotherphasesoftheMSA [ 101 , 102 ],technologicalstudiescandemonstratethatitistheentireMSA,ratherthansimply theHP,thatisexceptional.SupportingInformationS1File.Theblankswithoutretouchdatabase.TableA. (XLSX)AcknowledgmentsLynWadley,WilliamE.Banks,TammyHodgskissandGötzOssendorfkindlyhelpedmewith thediscussionandtheEnglisheditingofthismanuscript.KelseyGlennonkindlyhelpedme withthestatistics.Fiveanonymousreviewershelpedmetoimprovenotablythispaper.Any errorsoromissionsaremyresponsibility.AuthorContributionsConceivedanddesignedtheexperiments:PdlP.Performedtheexperiments:PdlP.Analyzed thedata:PdlP.Contributedreagents/materials/analysistools:PdlP.Wrotethepaper:PdlP.References1. StapletonP,HewittJ.Stoneimplementsfromarock-shelteratHowieson'sPoortnearGrahamstown. SouthAfricanJournalofScience.1927;24:574 – 587. 2. StapletonP,HewittJ.StoneimplementsfromHowieson'sPoort,nearGrahamstown.SouthAfrican JournalofScience.1928;25:399 – 409. 3. JacobsZ,RobertsRG,GalbraithRF,DeaconHJ,GranR,MackayA,etal.AgesfortheMiddleStone AgeofSouthernAfrica:implicationsforhumanbehavioranddispersal.Science.2008;322(5902): 733 – 735.doi: 10.1126/science.1162219 PMID: 18974351 4. DeaconHJ.LatePleistocenepalaeoecologyandarchaeologyinthesouthernCape,SouthAfrica.In: MellarsP,StringerCeditors.TheHumanRevolution:BehaviouralandBiologicalPerspectivesonthe OriginsofModernHumans.Edinburgh:EdinburghUniversityPress;1989.pp.547 – 564. 5. WurzS.TechnologicalTrendsintheMiddleStoneAgeofSouthAfricabetweenMIS7andMIS3.CurrentAnthropology.2013;54(S8):S305 – S319. 6. PorrazG,TexierPJ,ArcherW,PibouleM,RigaudJP,TriboloC.Technologicalsuccessionsinthe MiddleStoneAgesequenceofDiepkloofRockShelter,WesternCape,SouthAfrica.Journalof ArchaeologicalScience.2013;40(9):3376 – 3400. 7. HenshilwoodCS.LatePleistocenetechno-traditionsinsouthernAfrica:areviewofthestillBayand HowiesonsPoort,c.75e59ka.JournalofWorldPrehistory.2012;25,205 – 237. 8. Clarke,D.L.AnalyticalArchaeology.London:Methuen;1968. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 48/52

PAGE 49

9. delaPe ñ aP.ApropósitodelGravetiense ... Elpasodeculturaatecnocomplejo:uncasoejemplarde pervivenciaparticularista.Complutum.2012;23(1):41 – 62. 10. d ’ ErricoF,HenshilwoodCS.AdditionalevidenceforbonetechnologyinthesouthernAfricanMiddle StoneAge.JournalofHumanEvolution.2007;52:142 – 163.PMID: 16996574 11. d ’ ErricoF,BackwellL,WadleyL.IdentifyingregionalvariabilityinMiddleStoneAgebonetechnology: thecaseofSibuduCave.JournalofArchaeologicalScience.2012;39:2479 – 2495. 12. MackayA,WelzA.EngravedochrefromaMiddleStoneAgecontextatKleinKliphuisintheWestern CapeofSouthAfrica.JournalofArchaeologicalScience.2008;35:1521 – 1532. 13. TexierP.-J,PorrazG,ParkingtonJ,RigaudJ-P,PoggenpoelC,MillerC,etal.HowiesonsPoorttraditionofengravingostricheggshellcontainersdatedto60,000yearsagoatDiepkloofRockShelter, SouthAfrica.ProceedingsoftheNationalAcademyofSciences,USA.2010;107:6180 – 6185. 14. CochraneGW.ThetransitionfromHowieson ’ sPoorttoPost-HowiesonsPoortindustriesinSouthern Africa.SouthAfricanArchaeologicalSociety.Goodwinseries.2008;10:157:167. 15. ClarkJD.AreviewofprehistoricresearchinNorthernRhodesiaandNyasaland.In:ClarkJD,ColeS editors.Proceedingsofthe3rdPan-AfricanCongressonPrehistory,Livingstone,1955.London: ChattoandWindus.pp.412 – 32. 16. SingerR,WymerJ.TheMiddleStoneAgeatKlasiesRiverMouthinSouthAfrica.Chicago:UniversityofChicago;1982. 17. WadleyL,HarperP.RoseCottageCaverevisited:Malan'sMiddleStoneAgecollection.TheSouth AfricanArchaeologicalBulletin.1989;44:23 – 32. 18. KaplanJ.TheUmhlatuzanaRockSheltersequence:100000yearsofStoneAgehistory.Natal MuseumJournalofHumanities.1990;2:1 94. 19. BeaumontPB.BorderCave.MAthesis,UniversityofCapeTown,CapeTown.1978. 20. WendtWE. ‘ Artmobilier ’ fromtheApollo11Cave,SouthWestAfrica:Africa ’ soldestdatedworksof art.SouthAfricanArchaeologicalBulletin.1976;31:5 – 11. 21. DeaconHJ.TwoLatePleistocene-HoloceneArchaeologicalDepositoriesfromtheSouthernCape, SouthAfrica.SouthAfricanArchaeologicalBulletin.1995;50(162):121 – 131. 22. WiessnerP.StyleandSocialInformationinKalahariSanProjectilePoints.AmericanAntiquity,1983: 48(2):253 – 276. 23. WadleyL.HuntersandGatherersoftheLaterStoneAge,SouthernTransvaal.(CambridgeMonographsinAfricanArchaeology25.)Oxford:BritishArchaeologicalreports;1987. 24. McBreartyS,BrooksA.Therevolutionthatwasn't:anewinterpretationoftheoriginofmodernhuman behaviour,JournalofHumanEvolution,2000;39(5):453 – 563.PMID: 11102266 25. WurzS.TheMiddleStoneAgeSequenceatKlasiesRiver,SouthAfrica.Ph.D.thesis.Universityof Stellenbosch.2000. 26. WurzS.TheHowiesonsPoortBackedArtefactsfromKlasiesRiver:AnArgumentforSymbolic Behaviour.TheSouthAfricanArchaeologicalBulletin.1999;54(169):38 – 50. 27. VillaP,SorianoS,TsanovaT,DeganoI,HighamT,d ’ErricoF,etal.BorderCaveandthebeginning oftheLaterStoneAgeinSouthAfrica.ProceedingsoftheNationalAcademyofSciences.2012;109 (33):13208 – 13213. 28. AmbroseSH,LorenzKG.SocialandecologicalmodelsfortheMiddleStoneAgeinsouthernAfrica. In:MellarsPA.(Editor),TheEmergenceofModernHumans.Edinburgh:UniversityPress,pp.3 – 33. 29. MinichilloTJ.MiddleStoneAgeLithicStudy,SouthAfrica:anExaminationofModernHumanOrigins. Ph.D.thesis.UniversityofWashington.2005. 30. BousmanCB.Hunter-gathereradaptations,economicriskandtooldesign.LithicTechnology.1993; 18:59 – 86. 31. ThackerayAI.TheMiddleStoneAgesouthoftheLimpopoRiver.JournalofWorldPrehistory.1992; 6(4):385 – 431. 32. MackayA.HistoryandSelectionintheLatePleistoceneArchaeologyoftheWesternCape,South Africa.PhDthesis.AustralianNationalUniversity.2009. 33. SorianoS,VillaP,WadleyL.BladetechnologyandtoolformsintheMiddleStoneAgeofSouthAfrica: theHowiesonsPoortandpost-HowiesonsPoortatRoseCottageCave.JournalofArchaeological Science.2007;34(5):681 – 703. 34. VillaP,SorianoS,TeyssandierN,WurzS.TheHowiesonsPoortandMSAIIIatKlasiesRivermain site,Cave1A.JournalofArchaeologicalScience.2010;37:630 – 655. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 49/52

PAGE 50

35. SorianoS,VillaP,DeganoI,PollaroloL,LucejkoJJ,HenshilwoodC,etal.TheStillBayandHowiesonsPoortatSibuduandBlombos:UnderstandingMiddleStoneAgeTechnologies.PLoSONE10 (7):e0131127.doi: 10.1371/journal.pone.0131127 PMID: 26161665 36. delaPe ñ aP,WadleyL.NewknappingmethodsintheHowiesonsPoortatSibudu(KwaZulu-Natal, SouthAfrica).QuaternaryInternational.2014;350:26 – 42. 37. delaPe ñ aP,WadleyL.QuartzKnappingStrategiesintheHowiesonsPoortatSibudu(KwaZuluNatal,SouthAfrica).PLoSONE.2014;9(7):01534.doi: 10.1371/journal.pone.0101534 38. HarperPT.TheMiddleStoneAgesequencesatRoseCottageCave:asearchforcontinuityanddiscontinuity.MAdissertation,UniversityoftheWitwatersrand.1994. 39. HenshilwoodCS,vanNiekerkKL,WurzS,DelagnesA,ArmitageSJ,RifkinRF,etal.Klipdriftshelter, southernCape,SouthAfrica:preliminaryreportontheHowiesonsPoortlayers.JournalofArchaeologicalScience.2014;45:284 – 303. 40. JacobsZ,RobertsRG.Testingtimes:oldandnewchronologiesfortheHowieson ’ sPoortandStill Bayindustriesinenvironmentalcontext.SouthAfricanArchaeologicalSocietyGoodwinSeries.2008; 10:9 – 34. 41. delaPe ñ aP,WadleyL,LombardM.QuartzbifacialpointsintheHowiesonspoortofSibudu.South AfricanArchaeologicalBulletin.2013;68(198):136 – 199. 42. WadleyL,MohapiM.Asegmentisnotamonolith:evidencefromtheHowiesonsPoortatSibudu, SouthAfrica.JournalofArchaeologicalScience.2008;35:2594 – 2605. 43. Lombard,M.Archaeologicaluse-traceanalysesofstonetoolsfromSouthAfrica.Phdissertation.UniversityoftheWitswatersrand.2007. 44. PargeterJ.HowiesonsPoortsegmentsashuntingweapons:experimentswithreplicatedprojectiles. SouthAfricanArchaeologicalBulletin.2007;62:147 53. 45. WadleyL,HodgskissT,GrantM.ImplicationsforcomplexcognitionfromthehaftingoftoolswithcompoundadhesivesintheMiddleStoneAge,SouthAfrica.ProceedingsofNaturalAcademyofScience. 2009;106:9590 9594. 46. IgrejaM,PorrazG.FunctionalinsightsintotheinnovativeEarlyHowiesonsPoorttechnologyatDiepkloofRockShelter(WesternCape,SouthAfrica).JournalofArchaeologicalScience.2013;40(9): 3475 – 3491. 47. TriboloC,MercierN,DouvilleE,JoronJL,ReyssJL,RuferD,etal.OSLandTLdatingoftheMiddle StoneAgesequenceatDiepkloofRockShelter(SouthAfrica):aclarification.JournalofArchaeologicalScience.2013;40(9):3401 – 3411. 48. FeathersJ.LuminescenceDatingatDiepkloofRockShelter – NewDatesfromSingle-grain.Quartz. JournalofArchaeologicalScience.2015:inpress. 49. JacobsZ,RobertsRG.AnimprovedsinglegrainOSLchronologyforthesedimentarydepositsfrom DiepkloofRockshelter,WesternCape,SouthAfrica.JournalofArchaeologicalScience.2015:in press. 50. WadleyL.Examplesoftheuseofspace77,000to62,000yearsagoatSibudu,SouthAfrica.In:ConardNJ,DelagnesAeditors.SettlementDynamicsoftheMiddlePalaeolithicandMiddleStoneAge. Tuebingen:KernsVerlagVolIV;2015:pp.11 – 26. 51. WadleyL.TheHowieson ’ sPoortindustryofSibuduCave.SouthAfricanArchaeologicalBulletin. GoodwinSeries.2008;10:122 – 132. 52. WadleyL,JacobsJ.SibuduCave:backgroundtotheexcavations,stratigraphyanddating.Southern AfricanHumanities.2006;18(1):1 – 26. 53.WadleyL.SomecombustionfeaturesatSibudu,SouthAfrica,between65,000and58,000yearsago. QuaternaryInternational.2012;247:341 – 349. 54. GoldbergP,MillerCE,SchieglS,LigouisB,BernaF,etal.Bedding,hearths,andsitemaintenancein theMiddleStoneAgeofSibuduCave,KwaZulu-Natal,SouthAfrica.ArchaeologicalAnthropological Science.2009;1:95 – 122. 55. ClarkeBM,SchoemanP,KotaMW.2007.ThegeologyoftheVerulamarea.CouncilforGeoscience: D&VPremierPrint. 56. WadleyL,KempsonH.Areviewofrockstudiesforarchaeologists,andananalysisofdoleriteand hornfelsfromtheSibuduarea,KwaZulu-Natal.SouthernAfricanHumanities.2011;23:87 – 107. 57. MourreV.LesindustriesenquartzauPaléolithique:terminologie,méthodologieettechnologie. Paléo.1996;8:205 – 223. 58. deLombera-HermidaA.Quartzmorphostructuralgroupsandtheirmechanicalimplications.Annali dell ’ Universit à degliStudidiFerrara.2008:101 – 104. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 50/52

PAGE 51

59. MartínezCortizasA,LlanaRodríguezC.Morphostructuralvariablesandtheanalysisoftheireffecton quartzblankcharecteristics.In:MoloneyN,RaposoL,SantonjaM,editors.Non-FlintStoneToolsand thePalaeolithicOccupationoftheIberianPeninsula.Oxford:TempusReparatum,649;1996.pp. 49 – 53. 60. DelagnesA,WadleyL,VillaP,LombardM.CrystalquartzbackedtoolsfromtheHowiesonsPoortat SibuduCave.SouthernAfricanHumanities.2006;18(1):43 – 56. 61. LemonnierP.Ladescriptiondeschaînesopératoires:contributional ’ analysedessystemestechniques.TechniquesetCulture.1976;1:101 – 150. 62. KarlinC,BoduP,PelegrinJ.Processustecniquesetchaînesopératoires.Commentlespréhistoriens s ’ appropientunconceptélaboréparlesethnologues.In:BalfetH,editor.Observerl ’ actiontechnique, leschaînesoperatoires,pourquoifaire?Paris:CNRS;1991.pp.101 – 117. 63. PelegrinJ.Technologielithique.LeCh à telperroniendeRoc-De-Combe(Lot)etdelaCôte(Dordogne).Paris:CahiersduQuaternaire;1995. 64. DibbleHL.BiacheSaint-Vaast,LevelIIA:Acomparisonofanalyticalapproaches.In:DibbleH,Bar YosefO,editors.TheInterpretationandDefinitionofLevalloisVariability.Madison:PrehistoryPress; 1995.pp:93 – 116. 65. ShottMJ.C haîneopératoire andreductionsequence.Lithictechnology.2003;28(2):95 – 105. 66. TostevinGB.SeeingLithics:AMiddle-RangeTheoryforTestingforCulturalTransmissioninthe Pleistocene.AmericanSchoolofPrehistoricResearchMonograph;2013. 67. ShottMJ.Sizeandformintheanalysisofflakedebris:reviewandrecentapproaches.Journalof ArchaeologicalMethodandTheory.1994;1(1):69 – 110. 68. KlaricL.L ’ unitetechniquedesindustriesaburindurayssedansleurcontextediachronique.ReflexionssurladiversiteculturelleaugravettienapartirdesdonnesdelaPicardie,d ’ Arcy-Sur-Cure,de BrassempouyetducirquedelaPatrie.Ph.DissertationUniversitéParisI.2003. 69. LeBrun-RicalensF,BordesJG,BonF(eds).Productionslamellairesattribuées à l ’ Aurignacien:chaînesopératoiresetperspectivestechno-culturelles.Li è ge:Symposium6 – 7,XVeCongr è sdel ’ UISPP, 2 – 8Septembre2001.UniversitédeLi è ge.Archeologiques1.MuseeNationald ’ Histoireetd ’ Art.Luxembourg.2006. 70. CallahanE.,1985.AnEvaluationoftheLithicTechnologyinMiddleSwedenduringtheMesolithicand Neolithic.Uppsala:SocietasArchaeologicaUpsaliensis;1985. 71. Díez-MartínF,SánchezP,Domínguez-RodrigoM,BarbaR.WereOlduvaiHomininsmakingbutcheringtoolsorbatteringtools?AnalysisofarecentlyexcavatedlithicassemblagefromBK(BedII,OlduvaiGorge,Tanzania).JournalofAnthropologicalArchaeology.2009;28,274 – 289. 72. delaPe ñ aP.Aqualitativeguidetorecognizebipolarknappingforflintandquartz.Lithictechnology. 2015:Inpress.73. ShimelmitzR,KuhnS.EarlyMousterianLevalloisTechnologyinUnitIXofTabunCave.PaleoAnthropology.2013:1 27. 74. SotoSebastiánA.ProducciónygestióndelaindustrialíticadeAtxostxe(Alava):unaaproximacióna lassociedadesepipaleolíticasymesolíticasdelAltoEbro.UniversidaddelPaísVasco,Vitoria-Gasteiz.2014. 75. RiosJ,delaPe ñ aP,SanEmeterioA.EstudiodelasindustriaslíticasyóseasdelaCuevadeAitzbitarteIII(zonadelaentrada).In:AltunaJ,MariezkurrenaK,RiosJ.editorsOcupacioneshumanasen AitzbitarteIII(PaísVasco)33600-18400BP.2012,EKOB,5:79 – 345. 76. SoleckiL,SoleckiRS.AnewsecondaryflakingtechniqueattheNahrIbrahimCavesite,Lebanon. BulletinduMuseedeBeyruth.1970;XXIII:138 – 140. 77. DibbleH.L.ThemousterianindustryfromBisituncave(Iran).Paléorient.1984;10(2):23 – 34. 78. NishiakiY.Truncated-facetedFlakesfromLevantineAssemblages.UniversityofTokyoBulletin. 1985;4:215 – 226. 79. Goren-InbarN.Toosmalltobetrue?ReevaluationofcoresonflakesinLevantineMousterianassemblages.LithicTechnology.1988;17(1):37 – 44. 80. Hivernel-GuerreF,NewcomerMH.Nucléussuréclat:technologieetutilisationpardifférentescultures préhistoriques.BulletindelaSociétépréhistoriquedeFrance.1974;71:119 – 128. 81. BoëdaE,GenesteJ-M,MeignemL.IdentificationdesChaînesopératoireslithiquesduPaléolithique AncienetMoyen.Paléo.1990;2:43 – 80. 82. BoëdaE.Ledébitagedisco ï deetledébitagelevalloisrécurrentcentrip è te.BulletindelaSociétépréhistoriquefrançaise.1993;90(6):392 – 404. RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 51/52

PAGE 52

83. BoëdaE.Levallois:avolumetricconstruction,methods,atechnique.In:DibbleHL,Bar-YosefO,editors.TheDefinitionandInterpretationofLevalloisTechnology.Madison:MonographsinWorld Archaeology;1995.pp.41 – 68. 84. delaPe ñ aP.TheinterpretationofbipolarknappinginAfricanStoneAgestudies.CurrentAnthropology.2015:Inpress. 85. BordesF.(1961).OutilsLevallois.TypologieduPaléolithiqueancienetmoyen.ImprimériesDelmas, Bordeaux.1961. 86. VanPeerP.InterassemblageVariabilityandLevalloisStyles:TheCaseoftheNorthernAfricanMiddlePalaeolithic.JournalofAnthropologicalArchaeology.1991;10:107 – 151. 87. PeresaniM.DiscoidLithicTechnology.Advancesandimplications.BARInternationalSeries1120. Oxford:Archaeopress.2003. 88. DriscollK,García-RojasM.Theirlipsaresealed:identifyinghardstone,softstone,andantlerhammer directpercussioninPalaeolithicprismaticbladeproduction.JournalofArchaeologicalScience.2014; 47:134 – 141. 89. CochraneGWG.Ananalysisoflithicartefactsfromthe~60kalayersofSibuduCave.SouthernAfricanHumanities.2006;18(1):69 – 88. 90. CallahanE.AcommentonNami'scomment.PlainsAnthropologist,1991;36:367 368. 91. HiscockP.MobilityandtechnologyintheKakaducoastalwetlands.IndoPacificPrehistoryAssociationBulletin.1996;15:151 – 157. 92. HiscockP.MakingitsmallinthePalaeolithic:bipolarstone-working,miniatureartefactsandmodels ofcorerecycling,WorldArchaeology,doi: 10.1080/00438243.2014.991808 93. BrownKS,MareanCW,JacobsZ,SchovilleBJ,OestmoS,FisherEC,etal.Anearlyandenduring advancedtechnologyoriginating71000yearsagoinSouthAfrica.Nature.2012;491:590 – 593.doi: 10.1038/nature11660 PMID: 23135405 94. ClarksonC.RegionaldiversitywithinthecoretechnologyoftheHowiesonsPoortTechno-Complex. In:LycetSJ,ChauhanPR,editors.NewPerspectivesonOldStones:AnalyticalApproachestoPaleolithicTechnologies.NewYork:Springer;2010.pp.43 – 59. 95. HarperP.T.N.TheMiddleStoneAgesequencesatRoseCottageCave:asearchforcontinuityand discontinuity.SouthAfricanJournalofScience.1997;93:470 – 475. 96. MackayA,StewartBA,ChaseBM.CoalescenceandfragmentationinthelatePleistocenearchaeologyofsouthernmostAfrica.JournalofHumanEvolution.2014;72:26 – 51.doi: 10.1016/j.jhevol. 2014.03.003 PMID: 24746546 97. WillM,BaderGD,ConardNJ.CharacterizingtheLatePleistoceneMSALithicTechnologyofSibudu, KwaZulu-Natal,SouthAfrica.PlosOne9(5):e98359.doi: 10.1371/journal.pone.0098359 98. ClarkJL,PlugI.AnimalexploitationstrategiesduringtheSouthAfricanMiddleStoneAge:Howiesons Poortandpost-HowiesonsPoortfaunafromSibuduCave.JournalofHumanEvolution.2008;54(6): 886 – 898.doi: 10.1016/j.jhevol.2007.12.004 PMID: 18234284 99. WadleyL.Thosemarvellousmillennia:theMiddleStoneAgeofSouthernAfrica.Azania,2015;50 (2):155 – 226. 100. LombardM,ParsonsI.WhathappenedtothehumanmindaftertheHowiesonsPoort?Antiquity. 2011;85:1433 – 1443. 101. WadleyL,SieversC,BamfordM,GoldbergP,BernaF,MillerC.MiddleStoneAgebeddingconstructionandsettlementpatternsatSibudu,SouthAfrica.Science.2011;334:1388 1391.doi:10.1126/ science.1213317 PMID: 22158814 102. LennoxSJ,BamfordM.Useofwoodanatomytoidentifypoisonousplants:CharcoalofSpirostachys Africana.SouthAfricanJournalofScience.2015;111(3/4)Available: http://dx.doi.org/10.17159/sajs. 2015/20140143 . RefiningOurUnderstandingofHowiesonsPoortLithicTechnology PLOSONE|DOI:10.1371/journal.pone.0143451December3,2015 52/52


printinsert_linkshareget_appmore_horiz

Download Options

close


  • info Info

    There are only PDFs associated with this resource.

  • link PDF(s)



Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.