Klipdrift Shelter, southern Cape, South Africa : preliminary report on the Howiesons Poort layers


previous item | next item

Citation
Klipdrift Shelter, southern Cape, South Africa : preliminary report on the Howiesons Poort layers

Material Information

Title:
Klipdrift Shelter, southern Cape, South Africa : preliminary report on the Howiesons Poort layers
Series Title:
Journal of archaeological science
Creator:
Henshilwood, Christopher S.
Niekerk, Karen L. van
Wurz, Sarah
Delagnes, Anne
Armitage, Simon J.
Rifkin, Riaan F.
Douze, Katja
Keene, Petro
Haaland, Magnus
Reynard, Jerome
Discamps, Emmanuel
Mienies, Samantha S.
Publisher:
Elsevier
Publication Date:
Physical Description:
1 online resource

Subjects

Subjects / Keywords:
Excavations (Archaeology) ( lcsh )
Caves ( lcsh )
Mesolithic period ( lcsh )
Genre:
serial ( sobekcm )
Location:
Africa -- South Africa -- Eastern Cape -- Howieson's Poort Shelter

Notes

Abstract:
Surveys for archaeological sites in the De Hoop Nature Reserve, southern Cape, South Africa resulted in the discovery of a cave complex comprising two locations, Klipdrift Cave and Klipdrift Shelter. Excavations commenced in 2010 with Later Stone Age deposits initially being recovered at the former site and Middle Stone Age deposits at the latter. The lithic component at Klipdrift Shelter is consistent with the Howiesons Poort, a technological complex recorded at a number of archaeological sites in southern Africa. The age for these deposits at Klipdrift Shelter, obtained by single grain optically stimulated luminescence, spans the period 65.5 ± 4.8 ka to 59.4 ± 4.6 ka. Controlled and accurate excavations of the discrete layers have resulted in the recovery of a hominin molar, marine shells, terrestrial fauna, floral remains, organic materials, hearths, lithics, ochre, and ostrich eggshell. More than 95 pieces of the latter, distributed across the layers, are engraved with diverse, abstract patterns. The preliminary results from Klipdrift Shelter presented in this report provide new insights into the Howiesons Poort in this sub-region and contribute further to ongoing knowledge about the complex behaviours of early Homo sapiens in southern Africa. Excavations at the Klipdrift Complex will continue in the future.
Original Version:
Volume 45
Original Version:
20 p.

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-05590 ( USFLDC DOI )
k26-5590 ( USFLDC 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

KlipdriftShelter,southernCape,SouthAfrica:preliminaryreporton theHowiesonsPoortlayers ChristopherS.Henshilwood a , b , * ,KarenL.vanNiekerk a ,SarahWurz a , b ,AnneDelagnes c , b , SimonJ.Armitage d ,RiaanF.Rifkin a , b ,KatjaDouze b ,PetroKeene b ,MagnusM.Haaland a , JeromeReynard b ,EmmanuelDiscamps a ,SamanthaS.Mienies b a InstituteforArchaeology,History,CultureandReligiousStudies,UniversityofBergen,Øysteinsgate3,N-5007Bergen,Norway b EvolutionaryStudiesInstitute,UniversityoftheWitwatersrand,1JanSmutsAvenue,Braamfontein2000,Johannesburg,SouthAfrica c UniversitéBordeaux1,CNRSUMR5199PACEA,EquipePréhistoire,Paléoenvironnement,Patrimoine,AvenuedesFacultés,F-33405Talence,France d DepartmentofGeography,RoyalHolloway,UniversityofLondon,Egham,SurreyTW200EX,UK articleinfo Articlehistory: Received23October2013 Receivedinrevisedform 29January2014 Accepted31January2014 Availableonline15February2014 Keywords: MiddleStoneAge HowiesonsPoort Homosapiens Modernhumanbehaviour Coastalsubsistence SouthernAfrica abstract SurveysforarchaeologicalsitesintheDeHoopNatureReserve,southernCape,SouthAfricaresultedin thediscoveryofacavecomplexcomprisingtwolocations,KlipdriftCaveandKlipdriftShelter.Excavationscommencedin2010withLaterStoneAgedepositsinitiallybeingrecoveredattheformersiteand MiddleStoneAgedepositsatthelatter.ThelithiccomponentatKlipdriftShelterisconsistentwiththe HowiesonsPoort,atechnologicalcomplexrecordedatanumberofarchaeologicalsitesinsouthern Africa.TheageforthesedepositsatKlipdriftShelter,obtainedbysinglegrainopticallystimulated luminescence,spanstheperiod65.5 4.8kato59.4 4.6ka.Controlledandaccurateexcavationsofthe discretelayershaveresultedintherecoveryofahomininmolar,marineshells,terrestrialfauna, oral remains,organicmaterials,hearths,lithics,ochre,andostricheggshell.Morethan95piecesofthelatter, distributedacrossthelayers,areengravedwithdiverse,abstractpatterns.Thepreliminaryresultsfrom KlipdriftShelterpresentedinthisreportprovidenewinsightsintotheHowiesonsPoortinthissubregionandcontributefurthertoongoingknowledgeaboutthecomplexbehavioursofearlyHomosapiensinsouthernAfrica.ExcavationsattheKlipdriftComplexwillcontinueinthefuture. 2014TheAuthors.PublishedbyElsevierLtd. 1.Introduction From1998to2009intermittentarchaeologicalsitesurveysby twooftheauthors(CSHandKvN)along60kmofcoastlinelocated intheDeHoopNatureReserve,southernCape,SouthAfrica( Fig.1 ) resultedinthedetailedmappingofmorethan160archaeological sites.In2010twoofthesitesthatcomprisetheKlipdriftComplex, KlipdriftShelter(KDS)andKlipdriftCave(KDC),wereselectedfor testexcavations( Figs.1 e 3 ).Theexcavationsformapartofthe Tracsymbols project,fundedbyaEuropeanResearchCouncilFP7 grant(2010 e 2015)( http://tracsymbols.eu/ ),withonekeyaimbeingtoinitiatenewexcavationsatLatePleistocenearchaeological sitesinsouthernAfrica.TheselectionoftheKlipdriftsiteswas basedontheirvisible, insitu LaterStoneAge(LSA)andMiddle StoneAge(MSA)deposits,thepreservedfaunaandtheirrelative accessibility.In2011testexcavationscommencedatKDS( Figs.2 and 3 )revealingc.1.6mdeep,wellpreserved,horizontalMSA depositsimmediatelybelowthesteeplysloping,erodedsurface ( Fig.4 c).Theclearseparationofstratigraphiclayersenabledthe accuraterecoveryofmaterialsfromdiscretedepositionallayers. Theanthropogenicassemblagecontainedmarineshells,terrestrial faunalremains,microfauna,ahumantooth,organicmaterials,ash lensesandhearths,lithicartefacts,ochreandostricheggshell.In 2012weinitiatedtestexcavationsatasecondMSAsitewithinthe complex,KlipdriftCaveLower(KDCL)( Figs.2 and 3 ). Herewereportonthepreliminaryanalysisofthematerials recoveredfromtheKDSlayersdatedat65.5 4.8kato 59.4 4.6kabysingle-grainopticallystimulatedluminescence (OSL)( Fig.4 ).Thelithicsaretypicalofthoseattributedtothe HowiesonsPoortIndustry(HP)insouthernAfrica.Theresearch emanatingfromthissitehasthepotentialofcontributingtocurrent debatesabouttheoriginsofmodernhumanbehaviourwitha * Correspondingauthor.InstituteforArchaeology,History,CultureandReligious Studies,UniversityofBergen,Øysteinsgate3,N-5007Bergen,Norway.Tel./fax: þ 27 214656067. E-mailaddress: christopher.henshilwood@ahkr.uib.no (C.S.Henshilwood). Contentslistsavailableat ScienceDirect JournalofArchaeologicalScience journalhomepage:http://www. elsevier.com/locate/jas http://dx.doi.org/10.1016/j.jas.2014.01.033 0 3 0 5 4 4 0 3 2014TheAuthors.PublishedbyElsevierLtd. JournalofArchaeologicalScience45(2014)284 e 303 Open access under CC BY-NC-ND license . Open access under CC BY-NC-ND license .

PAGE 2

speci cfocusonthe Homosapiens thatinhabitedthesouthernCape duringtheMSA.ExcavationsatKDSandatothersiteswithinthe complexwillcontinueinthefuture. 1.1.Sitebackground EvidenceforhumanoccupationoftheDeHoopareafromthe Acheuleaniscon rmedbyhandaxesfoundnearPotberg( Fig.1 ) andthenumerousLSAandfewerMSAsitesdistributedmainly alongthecoast.TheKlipdriftComplexisamajordepositoryforLate andTerminalPleistocenesedimentsandarchaeologicaldeposits thatarevisiblebothonthesurfaceandinerodedsections.The Complexisoneofseveralcavernsandoverhangsalongthe southernCapecoastformedwithinthe500 e 440Millionyear(Ma) TableMountainGroup(TMG)sandstones( DeaconandGeleijnse, 1988 ).MovementalongtheshearzoneswithintheTMGforms faultbrecciassusceptibletoerosionbyhighsealevelsleadingtothe formationofcaveswithinthenearcoastalcliffs( Pickeringetal., 2013 ).KDCandKDSareformedintheTMGsandstones,presumablyasaresultofthisprocess.IntheeasternsectionofDeHoop, 5MahardduneridgesofBredasdorpGrouplimestonein llthese TMGshearzones.Theseawardextensionofthelimestonehasbeen truncatedbymarineerosionandinthesecoastalcliffsanumberof vadosecaveshavedevelopedabovethecontactwiththeTMG ( MarkerandCraven,2002 ). TheKlipdriftComplex(3427.09630S,2043.45820E),islocated incoastalcliffs12 e 15mfromtheIndianOceanandc.19mabovesea level.Thelargerwesterncaveisc.21mdeepandcontainsatleast twosites,KDCandKDCL.KDSisac.7mdeepshelter,separatedfrom KDCandKDCLbyaquartzitepromontory( Figs.2 e 4 ).Thecomplexis Fig.1. LocationofKlipdriftComplexandsitesmentionedinthetext. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 285

PAGE 3

locatedwithintheeasternsectionoftheDeHoopNatureReserve ( Fig.1 b)onPortion20offarm516,Swellendamdistrictinthe OverbergregionofthesouthernCape.CapeTownisc.150kmtothe west;the Klipdriftfonteinspruit stream(namesakeofthecavecomplex)andNoetsiewaterfall( ScottandBurgers,1993 ),whichare perennialsourcesoffreshwater,lieabout200meastoftheKlipdrift Complex.TheextensiveBreedeRiverestuaryandBlombosCavelie respectively10kmand45kmeast/south-east( Fig.1 b). InKDCarchaeologicaldepositsareconcentratedbehindthe driplineandextendover280m2atac.25slope.Ac.15mtalus slopesseawardsat31.5.InKDSvisiblesurfacedepositsextendc. 7m2ataslopeofc.29behindthedripline.Thedepositsare severelytruncatedandthetalusliesat38.5.Itisprobablethatthe naturalandarchaeologicaldepositsinthecavecomplex,especially thoseinKDS,weretruncatedbymid-Holocene þ 2 e 3msealevels ( Batemanetal.,2004;Compton,2001 ).Aquartzitecobblebeach liesdirectlybelowthecomplexwithanextendedrockyshoreline andfewsandybeaches.InitialexcavationsinKDCin2010yielded TerminalPleistocenedeposits(AlbanyIndustry)radiocarbondated atc.14 e 10ka(reportinprep.).In2013severaltonsofrockfallwere removedintheareaofthedriplineinKlipdriftCave( Fig.3 ).A limitedtestexcavationintheKlipdriftCaveLower(KDCL)site revealedMSAdepositsunderlyingtheoverburden.Aprovisional minimumOSLageofc.70kawasobtainedforthebaseofthe overburden.FurtherexcavationsofKDCLareplanned. KDSwas rstexcavatedin2011withsubsequentseasonsin 2012and2013.Intotalavolumeof2.3m3overanareaof6.75m2hasbeenexcavatedatKDStodepthsfrom30cmto100cm(in individualquadrates)andmorethan20layersandlensesde ned ( Figs.3 and 4 ).Theuppermostdatedlayeryieldsanoptically stimulatedluminescence(OSL)ageof51.7 3.3ka,themiddle layerscontainingtheHPrangefrom65.5 4.8kato59.4 4.6ka andthelowermostexcavated,anthropogenicallysterilelayersgive anageof71.6 5.1ka( Fig.4 a). 1.2.Background:DeHoopNatureReserve DeHoopNatureReservecovers34000haandextendsfor60km alongtheIndianOceancoastline( Fig.1 b).ThePotbergrange,a 611mhighremnantofasynclineoftheCapeFoldedBeltcomposed ofhighlyresistantTMGquartzite,liestothenorth-westofKlipdrift. Amajorfaultatthebaseoftherangetruncatesittothesouth.The TMGquartzitesformseacliffswheretheyareexposedbeneaththe BredasdorpGrouplimestone.SedimentaryrocksoftheTMG (sandstones),BokkeveldGroup(shales)andUitenhageGroup (mainlyshaleconglomerates)formthebasementgeologyofthe area.Marinetransgressionshaveplanedthesoftershalesand conglomeratesintoagentlysouthwardslopingseriesofterraces. TheNeogenelimestonesoftheBredasdorpGroup,depositedas shallowmarineenvironments(thePlioceneDeHoopvleiFormation andthePleistoceneKleinBrakFormation,bothofwhichareshelly quartzosesandandconglomerate)andascoastaldunes(thePlioceneWankoeFormationandthePleistoceneWaenhuiskransFormation),underliethegreaterpartofthereserve( Markerand Craven,2002 )andcovermostoftheBokkeveldandUitenhage basementrocks.TheWankoeFormationformsthehigh-lying aeolianitesintowhichthecoastalplainwaserodedduringmarinetransgressions.Morerecentdunesystems(Waenhuiskrans formation)weresubsequentlyformedonthecoastalplain.The StrandveldFormation,depositedasastripofunconsolidateddunes duringtheHoloceneisthemostrecentmemberoftheBredasdorp Group( Batemanetal.,2004;Malan,1990;Robertsetal.,2006; Rogers,1988 ). ThereserveissituatedintheCapeFloristicRegion,oneofthesix oralKingdomsintheworld.Itfallswithinawinterrainfallarea thathasaMediterraneanclimate.Thecurrentmeanannualrainfall isapproximately380mmwiththemaximuminAugustandthe minimuminDecemberandJanuary.ThewarmAgulhascurrent resultsintemperatewintersandwarmsummerswithanaverageof 20.5Cduringthelatterandanaverageof13.2Cduringwinter. Thecontinentalshelf,knownastheAgulhasBank( Fig.1 b),begins asarelativelyshallowtopographicalfeaturesouthofPortElizabeth andextendstothesouthandwestbeyondCapeAgulhas,80km westoftheKlipdriftComplex.Atitswidestpoint,southofCape Infanta( Fig.1 b),theAgulhasBankextendsmorethan200km ( Batemanetal.,2004;Carretal.,2007;Compton,2011;VanAndel, 1989 ). Threemajorvegetationtypesoccurinthereserve,Limestone Fynbos,MountainFynbos,andDuneFynbos/Thicket( Lowand Rebelo,1996 ).Adiversityofplantsandanimals,bothterrestrial andmarineinacomplexmosaicofdifferenthabitattypes,isa resultofthesevariedgeologicalfeaturesandthecloselocationof thereservetoCapeAgulhas,themeetingpointofthewestcoast coldBenguelaandwarmeastcoastsubtropicalAgulhascurrents. Thisdiversityisillustratedbythe86terrestrialmammalspecies thatoccurhere,atleast250speciesof shinthemarineprotected areaandthemorethan260residentandmigratorybirdspecies. LimestoneFynbos,whichischaracterizedbylowshrubs,isthe predominantvegetationintheimmediatevicinityoftheKlipdrift Complex( Willisetal.,1996 ). 2.Excavationmethodology Twogridsystems,orientedonalocalnorth e southaxis,wereset upusingaTrimbleVXSpatialStation.The rstisathreeFig.2. KlipdriftCaveandKlipdriftSheltertowardsthenorth(upper)andwest(lower). C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 286

PAGE 4

dimensional,numericalcoordinatesystem,wheretheXand Y axes aregivenarbitrarynumericalvalues(50,100),andthe Z axisvalues refertoelevationabovesealevel.Thesecond,analpha-numerical system,consistsofacontinuoussquaremetregridstartingfrom A1,inwhicheachsquareisfurthersubdividedintofour50 50cm quadrates(nameda,b,candd)(see Fig.4 b). Eachquadratewasexcavatedindividuallybybrushandtrowel, followingstratigraphiclayers.Thelayerswithineachquadrate whichcontainsedimentsofseveraldepositionaleventswere principallyidenti edandde nedbytheirtexture,composition, colour,thicknessandcontent.Thespatialextentofindividuallayers variesthroughouttheexcavatedareaandlayerdepthsrangefromc. 2 e 30cm.Thelayersweregivenalphabeticallyorderednamecodes (PAL,PBA,PCAetc.)(see Fig.4 ).Namecodesthatsharethetwo rst letters(e.g.PAandsubdivisionsPAL,PAMetc.)wereinterpretedas havingclosecontextualrelationships.Amicromorphologicalstudy oftheselayersisinprogress.Spatialmeasurementstakenduring excavationrefertothenumericalcoordinateandweregivena three-dimensional(XYZ)spatialreference.Lithics > 20mm,identi ablebones,ostricheggshell,ochreandartefactsofspecialinterestwereindividuallyrecordedwithhighprecision(1/1000cm) andwithanaccuracyof 2mm.Recovered ndsorfeatureswere baggedinplastic,labelledwithprovenancedataandgivenaunique specimennumber.Allplotted ndswereclassi edonaprimary entryforminthe eldbyrawmaterial,species,tooltypeand specialcharacteristics.Non-plottedmaterial(deposit/sediments) wassievedthroughanested3.0 e 1.5mmsieveandretainedfor futureanalysis. Thetopographicsurfaceofastratigraphiclayerinaquadrate wasrecordedbyc.5003Dpoints(pointcloud)usingthe3D scanningfunctionontheTrimbleVXspatialstation.Thepoint cloudwaslaterconvertedintoa3Dmodeloftheentirelayersurfaceforremodellingoftheoriginalsurfacetopography.Thesurface ofeachquadratewasalsodigitallyphotographedwithasinglelens re excamera(NikonD4)withsurfacemarkers,permittingthe imagetobegeo-referencedandmodelledin3D.Similarphotos weretakenofsectionwalls,signi cantartefacts insitu andother relevantfeatures. Allsitemaps,crosssectionsandillustrationsoftheKDSstratigraphicsequencearegeo-referencedwithinthenumericalcoordinatesystemandmadebycombiningphotogrammetricmethods withtopographicdatarecordedbythetotalstation.TheKlipdrift Complexandsurroundsweremappedbyscanningthesitein3D. Thepointcloudthatwasgenerated(c.250000points)wasimportedintoTrimbleRealWorks6.5andconvertedintoa3Dmesh, fromwhichplanarmaps,cross-sectionofsurfacetopographyand elevationmodelswereproduced.Theseweresubsequently exportedasCAD lesandimportedintoESRIARCGIS10.1for furtherre nement,mapmakingandforcombiningwithgeoreferencedimages( Figs.3 and 4 ).Materialsrecoveredfromthe siteswereprimarilysortedandwashedatthebaselaboratorysituatedatPotbergintheDeHoopNatureReserve.Oncompletionof theexcavations,thematerialwasmovedtoourlaboratoryinCapeTownforcurationandfurtherinvestigation.Inthelongertermthe recoveredassemblageswillbecuratedattheIzikoSouthAfrican MuseuminCapeTown. 3.Opticallystimulatedluminescencedating TheMSAlayersatKDSweredatedusingsingle-grainOSL. Single-grainmeasurementswereperformedsincepreviousOSL datingstudiesconductedonsouthernAfricanMSAsitesdemonstratethatmulti-grainanalysesaresusceptibletoanumberof sourcesofinaccuracy(e.g. Jacobsetal.,2008 ).Theseinaccuracies maybeavoidedbymeasuringandanalysingtheOSLpropertiesofa sampleatthesingle-grainlevel( JacobsandRoberts,2007 ). 3.1.Samplecollection,preparationandmeasurement Sampleswerecollectedfromcleanedsectionsbyscrapingmaterialintoopaquebagswhileundertarpaulin.Samplelocationsare listedin Table1 .Usingtheprocedureoutlinedin Armitageetal. Fig.3. TopographicalfeaturesofKlipdriftComplexincludinglayoutofexcavatedarchaeologicalsites,KlipdriftCave,KlipdriftCaveLowerandKlipdri ftShelter. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 287

PAGE 5

Fig.4. a)StratigraphyofKlipdriftSheltershowinglayersandopticallystimulatedluminescenceages;b)LocationofexcavatedquadrateswithinKDS;c)e xcavatedlayersinsection showingtheslope. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 288

PAGE 6

(2011) 212 e 180mmdiameterquartzgrainswereextractedfrom bulksamples.Betaandgammadoserateswerecalculatedforeach sampleusingradioisotopeconcentrationsmeasuredbyICP-MS(U andTh)andICP-AES(K).Doserateswerecorrectedusingan assumedwatercontentof20 5%.Thisassumedvaluewas preferredtomeasuredvaluessincethelatterarestronglydependentuponthetimeelapsedsincethesectionwasexcavatedandthe antecedentweatherconditions.Theassumedvalueisclosetothe meanmeasuredwatercontent(19 6%)forasuiteof12samples fromKDS,whichwiththeinclusionofthe5%uncertaintyterm, givescon dencethatitapproximatesthetruemeanburialconditions.Gammadoserateswerecorrectedfora20%volumeoflowradioactivityclasts.Cosmicraydoserateswerecalculatedusing sitelocationandoverburdendensity,accountingforshieldingby thenearbyrockface( PrescottandHutton,1994;Smithetal.,1997 ). Aninternalalphadoserateof0.03 0.006Gy/kawasassumed. Equivalentdosesweremeasuredusingthesingle-aliquot regenerative-dosetechnique( MurrayandWintle,2000 )usinga RisøTL/OSL-DA-15instrument( Bøtter-Jensenetal.,2003 ) tted withasingle-grainOSLattachment( Dulleretal.,1999,2000 ). Single-aliquotdoserecoverytests( Robertsetal.,1999 )wereperformedoneverysample,andindicateinter-samplevariabilityin theoptimalpreheatingregime,aphenomenonalsoobservedat DiepkloofRockShelter( Triboloetal.,2013 ).Single-graindoserecoverytests,usingtheoptimalmeasurementconditionsidenti ed bythesingle-aliquotdata,wereperformedonfoursamplesand yieldeddoserecoveryratiosconsistentwithunity.Equivalentdose ( De)measurementswereperformedusingtheoptimalpreheating regimeidenti edforeachsample.Datawerescreenedusingthe grainrejectioncriteriaof Armitageetal.(2011) .Inaddition,grains wererejectedwherethesensitivity-correctednaturalluminescenceintensityexceededtwicethe D0valueofthesaturating exponential ttothegrowthcurve( WintleandMurray,2006; Chapotetal.,2012 ).Equivalentdoseswerecalculatedforgrains whichpassedtheserejectioncriteria. 3.2.Estimationofthesampleburialdose Allsamplesyieldedsuf cientdatatocalculateameaningful De. Wheretheoverdispersion(sd,therelativestandarddeviationof thetruepalaeodoses)ofsingle-grain Devaluesforasamplewas 20%orless,allgrainswereassumedtobelongtoasinglepopulation (following Olleyetal.,2004 ),andtheCentralAgeModel(CAM, Galbraithetal.,1999 )wasusedtocalculateanequivalentdosefor thatsample.Whereoverdispersionexceeded20%,itwasassumed thatmorethanonedosepopulationwaspresent,andthedataset wasanalysedusingtheFiniteMixtureModel(FMM, Robertsetal., 2000 ).AlldatasetstowhichtheFMMwasappliedwerebest tted withtwo Depopulations,andineachcaseasingledominant population( 87%ofacceptedgrains)wasidenti ed.The Decalculatedforthispopulationwasconsideredmostappropriatefor agedetermination.InsamplesKDS-DS7,10and11,theremaining grainsbelongtoasmall(2 e 8%)lowerdosepopulation,whichwas interpretedtoindicatetheintrusionoflowerdosegrainsfrom abovebybioturbation,thoughitisnoteworthythatsamplesoverlyingKDS-DS10(KDS-DS1,2and9)donotcontainsimilarpopulations.Thesmall(7 e 13%)higherdosepopulationpresentin samplesKDS-DS1,2and9wasinterpretedasindicatingthepresenceof “ partiallybleached ” grains. Although20%overdispersionhasbeenwidelyusedasa thresholdabovewhichtheFMMshouldbeused,ithasbeenargued thatthisthresholdisstrictlyonlyapplicabletothe Olleyetal. (2004) dataset.Inaddition,sampleswhichcannotcontainmore thanoneequivalentdosepopulationoccasionallyyieldoverdispersionvaluesabove20%(e.g. ArmitageandKing,2013 ).However,inspectionofradialplotsforsamplesKDS-DS9and10( Fig.5 a, b)indicatesthatboththeminorhighandlowDepopulations identi edbytheFMMareclearlydistinctfromthepopulation containingthemajorityofthegrains.Conversely,radialplotsfor samplesKDS-DS3and12( Fig.5 c,d),whichwereanalysedusing theCAM,appeartoshowasinglepopulationofgrains.Theseresults indicatethat,forourdataset,thecorrectstatisticalmodelmay accuratelybechosenusingtheoverdispersionparameter.Agesfor theKDSsamplesarepresentedin Table1 . 4.Culturalartefacts 4.1.Lithics Thispreliminarytechno-culturalinterpretationoftheKDS sequenceisbasedonthelithicsrecoveredin2010and2011.Layers PCAtoPAY,rangingfrom65.5 4.8kato59.4 4.6ka,provide highlysigni cantsamplesfora rsttechnologicalassessment,with 11,687lithics > 2cminthesevenlayersconsideredhere( Table2 ). Lithicrawmaterialsarecomposedof vemaingroups: quartzite,quartz,silcrete,cryptocrystallinesilicate(CCS)andcalcrete.Inalllayers,alargeportionofthestonefoundderivesfrom theshelter ’ swalls,andaremostlyquartziteandtoalesserextent quartz.Thesecoarseandpoorqualityrawmaterialswereoccasionallyexploitedbytheknappers.Quartzitealsoincludes negrainedtypesderivedfrompebbles,whilequartzispredominantlycomposedofgoodqualitytypes,withvery necrystalline structure.SilcretesusedbytheKDStool-makersarealmostexclusively ne-grainedtypes,frequentlywithinternalcracks.Colour variationsincludegrey,yellow-brown,brown,redandgreen.Primarysourcesofsilcreteandcalcretearepresentinabundance alongtheCapeFoldMountains(see Roberts,2003 )andnearKDS theyoccurasoutcropsinsmallrockyhillssome10kmnorthand Table1 SummaryequivalentdosedataandagesfortheKDSsamples.Samplesarelistedinstratigraphicorder:sddenotesoverdispersion,while n isthenumberofgrainswhichpass therejectioncriteria.TheagemodelsusedaretheCentralAgeModel(CAM)andtheFiniteMixtureModel(FMM).Uncertaintiesarebasedonthepropagat ion,inquadrature,of errorsassociatedwithindividualerrorsforallmeasuredquantities.Inadditiontouncertaintiescalculatedfromcountingstatistics,errorsdu eto1)betasourcecalibration(3%, ArmitageandBailey,2005 ),2)ICP-MS/AEScalibration(3%),3)doserateconversionfactors(3%),4)attenuationfactors(2%, MurrayandOlley,2002 )havebeenincluded. Sample(KDS. )SquareLevelsd(%) n AgemodelGrainsinmain component(%) Equivalentdose(Gy)Doserate(Gy/ka)Age(ka) DS11Q27BPAN/PAO25 3146FMM98 145.4 1.20.88 0.0451.7 3.3 DS12Q27BPAS18 3126CAM10052.1 1.40.86 0.0460.3 3.8 DS3R28CPAY19 381CAM10059.1 1.90.98 0.0560.0 4.0 DS2R28CPBA/PBB27 465FMM93 654.8 2.40.92 0.0559.4 4.6 DS1R28CPBC27 3113FMM87 945.2 1.90.69 0.0465.5 4.8 DS9R28CPBD21 3111FMM87 558.5 1.50.91 0.0564.6 4.2 DS10R28CPCA21 360FMM95 471.6 3.01.13 0.0663.5 4.7 DS7S30APE31 491FMM92 474.8 2.91.05 0.0571.6 5.1C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 289

PAGE 7

north-westofthesite.Someoftheknappedsilcretemayoriginate frompebblesourcesthathavenotyetbeenidenti ed. Signi cantchangesoccurintherelativeproportionsofthese rawmaterialgroupsovertime.Silcreteisdominantinthethree lowerlayers(PCA,PBE,PBD),whilequartzincreasessigni cantlyin thetwooverlyinglayers(PBC,PBA/PBB),andquartziteaswellas calcretebecomemoreabundantintheuppermostPAZandPAY layers( Table2 and Fig.6 ).Theseshiftsinthesequenceareeven morepronouncedwhenconsideringtherawmaterialdistribution ofthebladesandformaltools(backedtoolsandnotchedtoolsin particular)( Fig.6 ). Thelithic chaîneopératoire performedonquartz,silcreteand CCSisalmostentirelydevotedtotheproductionofblades,whichis con rmedbythestrongpredominanceofbladecoresinalllayers (PCA:16/21cores,PBE:9/12,PBD:35/47,PBC:21/29,PBA/PBB:26/ 43,PAZ:11/18,PAY:9/17).The akingmethodappliedtoblade productionisalmostexclusivelyunidirectionalandanumberof technicalattributes,e.g.platformedgeabrasion,weaklydeveloped bulbsandthinplatforms,indicatetheuseofdirectmarginalpercussionwithasofthammer,eithermineralorvegetal.Corevolume exploitationisvariedandincludesunifacialcoreswithprepared lateralconvexities,semi-rotatingcores, “ narrow-face ” coresand bipolarcores.Themeanwidthofbladesisquitehomogeneous acrossrawmaterialsandtendstobeslightlyhigherinthefour uppermostlayers(fromPBCtoPAY: Fig.6 ).Theelongationofblades ishighinalllayers,withnosigni cantpatternofchangeovertime (blades ’ length/widthinPCA:2.7,PBE:2.4,PBD:2.5,PBC:2.7,PBA/ PBB:2.3,PAZ:2.4,PAY:2.5).Blades( Fig.7 :1 e 12)rangefromvery small(lengthbetween10and20mm)tolarge(over60mmin length).Besidesbladeproduction,secondary akeproductionoccursonquartz,silcreteandcalcrete.Itconsistsmainlyofdiscoidal andLevalloisdébitage.Discoidalcoresoccurinsmallquantitiesin thewholesequence,unliketheLevalloiscoreswhicharelimitedto theupperpartofthesequence(layersPBC,PBA/PBB,PAZ,andPAY). TheexistenceofasecondaryLevalloisreductionsequenceis con rmedbythepresenceofLevallois akes.Theseareveryrareor absentfromlayersPCAtoPBA/PBBandamountto5Levallois akes inPAZ,and24inPAY.Thetoppartofthesequencethusprovides evidencefortheemergenceofanindependentandstructured ake reductionsequence.Incontrasttootherrawmaterials,quartzite waspredominantlyusedforproducing akes( Table2 )from informalandunidirectionalcores.Bladeproductiononquartziteis weaklydevelopedinalllayers,exceptinPAYwherequartziteblade productionisrelativelywellrepresented.Forboth akeandblade production,quartziteexploitationwasbasedonexpedientand shortreductionsequencesperformedwithdirecthardhammer percussion. Thetools( Fig.7 :13 e 27)aretypicaloftheHP;formaltoolsare composedofbackedtools,notchedtools,borers,retouchedblades, piècesesquillées andpoints.Retouchedtoolsaccountforlessthan5%oftheassemblages(PCA:3.5%,PBE:2.5%,PBD:2.8%,PBC:3.2%, PBA/PBB:3.2%,PAZ:5%,PAY:2.6%).Somemarkedshiftsoccurinthe Fig.5. Radialplotsofequivalentdosesfora)KDS-DS9,b)KDS-DS10,c)KDS-DS3andd)KDS-DS12ofremainingdeposits. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 290

PAGE 8

Table2 AssemblagecompositionatKDS(thechunkcategory,whichaccountsforc.40%ofthetotalassemblage,hasbeeneliminatedfromthequantitativeanaly sesasitincludesa numberofambiguousitems e naturalslabsorknappingdebris e especiallyforquartzite). PAYPAZPBA/PBBPBCPBDPBEPCA n % n % n % n % n % n % n % Flakes Quartz678.917833.963140.915223.61457.98110.210413.9 Silcrete&CCS9712.96011.4714.6446.825013.679109712.9 Calcrete344.5152.9593.810.20060.800 Quartzite31842.313625.931620.526541.156030.420425.822029.3 Blades Quartz222.9356.731220.26610.21508.111714.8668.8 Silcrete&CCS699.2326.1291.9446.855630.223830.118624.8 Calcrete15291.7171.10030.20010.1 Quartzite8811.7163291.9274.2804.3364.5364.8 Cores Quartz101.381.5412.7172.6140.860.891.2 Silcrete&CCS70.9101.940.391.4341.860.870.9 Calcrete50.70010.100000000 Quartzite40.510.20040.680.40000 Tools Quartz0061.1231.5101.670.40020.3 Silcrete&CCS91.2122.360.440.6321.7151.9202.7 Calcrete20.330.60000000000 Quartzite30.420.440.30040.210.100 Hammerstones 10.120.410.120.30030.420.3 Subtotal75110052510015441006451001843100792100750100 Chunks Quartz141 e 135 e 308 e 83 e 114 e 127 e 89 e Silcrete&CCS10 e 22 e 19 e 12 e 48 e 22 e 35 e Calcrete25 e 8 e 19 e 5 e 0 e 0 e 0 e Quartzite936 e 288 e 548 e 382 e 492 e 522 e 358 e Pebbles 14 e 12 e 11 e 11 e 8 e 17 e16 e Subtotal1126465905493662688498 Total187799024491138250514801248 Fig.6. TechnologicalchangesinlithicsatKDS,layersPCAtoPAY. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 291

PAGE 9

Fig.7. Bladesandformaltools:1:quartzblade,layerPBA/PBB;2,3,4,5,7,8:silcreteblades,layerPBD;6,9:silcreteblades,layerPBC;10,11,12:silcre teblades,layerPCA;13,14: quartzsegments,layerPBC;15:quartzsegment,layerPBA/PBB;16:quartzbackedtool,layerPBA/PBB;17:silcretesegment,layerPBD;18:silcreteb i-truncatedtool,layerPBD;19, 20:silcretetruncatedtools,layerPBE;21,22:silcretetruncatedtools,layerPCA;23,26:silcretestrangulatednotches,layerPBE;24:silcrete retouchedblade,layerPBD;25:silcrete strangulatedblade,layerPBD;27:silcretestrangulatedblade,layerPCA. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 292

PAGE 10

toolkitcompositionovertime,bothbetweenandwithintool groups( Table3 ).Backedtoolsincludedifferenttypes( Fig.7 :13 e 22),whoseproportionsvaryconsistentlyfromonelayertoanother. Segments( Fig.7 :13 e 15,17)arebestrepresentedinthemiddlepart ofthesequence(layersPBD,PBC,PBA/PBB),withapeakinPBC ( Fig.6 )wheretheycorrespondtoasmallsetofquartzsegments ( n ¼ 7)withstandardizedmorpho-dimensionalattributes.Truncatedblades( sensu IgrejaandPorraz,2013 )arepresentinalmostall layers(PCAtoPAZ).Withinthiscategory,afewhighlystandardized silcretetoolsarecharacterizedbyaproximalobliquetruncation oppositetoabrokentransversedistalpart( Fig.7 :19 e 22),which areonlypresentinthelowerlayers(PCA e n ¼ 3,PBE e n ¼ 2,PBD e n ¼ 1).Notchedtools( Fig.7 :23,25 e 27)arealsodiagnosticwith regardtopatternsofchangewithinthesequence( Fig.6 ).They representalargemajorityoftheretouchedtoolsinthelowerlayers (PCAandPBEwithrespectively16/22and12/16notchedtools/total oftools).Inthesetwolayers,notchedpiecesincludetypical strangulatedblades( Fig.7 :23,25 e 27)withmultipledeep retouchednotchesononeortwolateraledgesoflargesilcrete blades(PCA e n ¼ 7including1calcretetool,PBE e n ¼ 6).Theyalso occurinlesserproportionsinPBD( n ¼ 1)andPBC( n ¼ 2),butare totallyabsentintheuppermostlayers.Inalllayers,notchedtools arepredominantlymadeonsilcreteblanks. Theshiftfromanotchedtool-dominatedtoolkit(inPCA,PBE)to abackedtool-dominatedtoolkit(inPBC,PBA/PBB)isclosely correlatedwiththeinversionoftherelativeproportionsofsilcrete toquartzinthesamelayers( Fig.6 ).Fewothercategoriesofformal toolsarespeci ctocertainlayers.PBDinparticularcontainsborers insilcrete( n ¼ 2),quartz( n ¼ 2,including1crystalquartz)andCCS ( n ¼ 1).Silcretebladeswithmarginalcontinuousretouchonone lateraledge( Fig.7:24)arealmostexclusivelypresentinPBD ( n ¼ 8),andoccurrarelyinbothPCA( n ¼ 1)andPBC( n ¼ 1). UnifacialpointsonlyoccurinPAY( n ¼ 3)andaretypicalofthe “ post-HP ” periodinsouthernAfrica(seeforinstance Conardetal., 2012;Lombardetal.,2012;Sorianoetal.,2007;Villaetal.,2005 ). TechnologicalvariationsthroughtimefromPCAtoPAYrelateto threemainphasesthatcanbeincludedwithintheHPcomplex.The lowermostphase(PCA,PBE)ischaracterizedbythepredominant exploitationofsilcreteforbladeproduction,theprevalenceof notchedtools,thepresenceofstrangulatedbladesandofhighly standardizedtruncatedblades.Thefollowingphase(PBC,PBA/PBB) ismarkedbyanincreaseinquartzexploitationwhichbecomesthe mostcommonrawmaterial,whilebackedtools,includingtypical segments,constitutethemaintoolgroup.Thethirdanduppermost phase(PAY)isde nedbythepredominanceofquartzite,an increaseinthesizeofblades,theemergenceofanindependentand structured akeproductionbasedonaLevalloisconcept,adecrease intheproportionsofbackedtoolsandthepresenceofafewunifacialpoints.PAYcouldbeinterpretedasatransitionallayertowardsthepost-HP.Inbetweenthesephases,layersPBDandPAZ appearastransitionallayers,thuspointingtoaprocessofgradual changeovertime. 4.2.Ochre Mineralpigmentsrecoveredfromarchaeologicalcontextsare generallytermed ‘ ochre ’ andrefertorockswhichderivetheir colourfromhaematite(ae Fe)andgoethite(ae FeO(OH)) ( Eastaughetal.,2008 ).Thetermdescribesearthymaterialswhich consistofanhydrousiron(III e ferricorFe3 þ)oxidesuchasred ochre(unhydratedhaematiteorFe2O3),partlyhydratediron(III) oxide-hydroxidesuchasbrowngoethite(FeO(OH))orhydrated iron(III)oxide-hydroxidesuchasyellowlimonite(Fe2O3(OH) nH2O)( CornellandSchwertmann,2003 ). Anidenti edtotalof356piecesor1756gofochreousmaterial wasextractedduringthe2011 e 2013excavationseasonsatKDS. Ochreousdepositsdonotoccurwithintheshelterandnosources havebeenidenti edintheimmediatevicinityofthecomplex. Besidesaferricretesource400mtotheeastseveralochreous outcropsoccurwithin5 e 10kmofthesite.Ochresourcesaremore frequentwithina30kmradiusofKDS,themostconspicuousbeing theBokkeveldGroupdepositsoftheCapeSupergroup( Vorster, 2003 ).Thesecompriseredferruginousshales,siltstones,mudstonesandhaematisedshales.Theloweringofsealevels,for exampleduringMIS5e,wouldlikelyhaveexposedBokkeveld shaleswithin0.5 e 1kmfromthesite. Allidenti edspecimensheavierthan0.1gwereanalysedand aredescribedintermsofweightandsize,colour,geologyand processingtechniqueemployed.Theanalysedpiecescompriseboth complete(suchashardferruginous)andfragmentary(softershales andmudstones)specimens. 4.2.1.Stratigraphicfrequency ThebulkoftheassemblagederivesfromlayersPBA/PBBfollowedbyPCAandPBD( Table4 ).Bymass,layerPBEhasthehighest concentrationofredochre(847.6g)intheassemblage(48.3%).It shouldbenotedthatbyweightjustover90%oftheochreinlayer PBEconsistsofcoarseto nelyprocessedpiecesweighinglessthan 0.1geach.Intermsofaveragemassthehighestmeanweightsare recordedinPBC(4.3g)andPCA(3.2g).Thehighstandard Table3 RetouchedtoolcompositionatKDS(Q:quartz,S:silcrete,C:calcrete,Qi:quartzite)(backedtoolsmayincludelocalizedormarginalretouchandob liquetruncationsmayalsobe proximal). PAYPAZPBA/PBBPBCPBDPBEPCA QSCQiQSCQiQSCQiQSCQiQSCQiQSCQiQSCQi Segments1 eee 11 ee 41 ee 7 eee 13 e 1 e 1 eee 1 ee Backedtools e 1 ee 24 ee 171 ee 11 ee 34 eee 3 eeeeee Obliquetruncations eeee 1 eee 1 eee 21 eee 1 eee 2 eee 3 ee Singlenotches e 2 eee 11 e 32 ee 11 eee 1 eee 6 eee 8 ee Denticulates eee 1 eeee 3121 eeeee 2 e 2 eeeee 1 ee Strangulatedblades eeeeeeeeeeeee 2 eee 1 eee 6 eee 61 e Borers eeeeeeeeeeeeeeee 23 eeeeeeeeee Retouchedblades eeeeeeeeeeeee 1 eee 8 eeeeeee 1 ee Unifacialpointse 2 e 1 eeeeeeeeeeeeeeeeeeeeeeee Burins e 1 eee 2 eeeeee 1 eeeeeeeeeeeeeee Piècesesquillées eeeeee 1 ee 1 e 11 eee 15 eeeeeeeeee Scrapers eee 1 e 11122111 eeee 1 e 1 eeeee 1 ee Miscellaneous e 54145 e 113 e 21 eeee 12 eee 1 e 123 ee Total111448143231113515600741040190122410C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 293

PAGE 11

deviationsinweightforlayerPBCandalsoPCAindicatethat specimensrangesubstantiallyintermsofweightandtherefore size,andpossiblyalsointermsofintensityofprocessing.The lowestaverageweightsoccurinlayersPAY(0.78g)andPBE(1.6g). TheheaviestindividualpiecesderivefromlayerPCA(79.5g),followedbyPBE(38.5g),PBC(35.7g),PBA/PBB(29.2g)andPBD (17.3g).TheleastheavyexamplesoriginatefromlayersPBD,with 41piecesweighing < 0.5g,andPCAwith36pieces < 0.5g. Intermsofaveragesize,thelargestgroupingisthatfromlayers PBC(23.5mm)andPCA(19.1mm),followedbyPBD,PAZandPAYat 17.4mm,16.8mmand16.4mmrespectively( Table4 ).LayersPBA/ PBB(12.3mm)andPBE(12.9mm)containthesmallestmeansizes ofochrepieces.ThelargestpiecesarefromlayerPCA(74.6mm)and thesmallestfromPBA/PBB(1.0mm).NotethehighstandarddeviationsinsizeforlayersPAZ,PBCandPCA. 4.2.2.Geologicalpro lesandcolourcategories Sixrawmaterialcategoriesarediscerned,namely ssileshale, induratedshale,mudstone,ferricrete,haematiteandsandstone. Fine-grainedandsoft(2 e 3onMoh Â’ shardnessscale)sedimentary formsincluding ssileshale(53%),induratedshale(22.9%)and mudstone(14.5%)accountsfor90.3%oftherawmaterialassemblage( Fig.8 a).Harder( > 4onMoh Â’ sscale)andessentiallycoarsegrainedformssuchasferricrete(2.4%),haematite(2.4%)and sandstone(4.8%)constitutetheremainder(9.7%)oftheassemblage.LayersPAYtoPCAdisplaymarkedgeologicalvariability,with allsixgeologicalcategoriesoccurringinlayersPBC,PBDandPCA. LayerPBCexhibitsthehighestfrequenciesofferricrete(5.3%)and Table4 Ochrefrequencybyweight,sizeandstratigraphiclayer. LayerTotal(n)Total(g)Mean(g)Std.dev.Mean(mm)Std.dev. PAY1713.40.81.616.45.7 PAZ1221.21.83.216.810.4 PBA/PBB113248.72.25.415.48.3 PBC19126.74.311.523.514.4 PBD59182.21.75.117.49.1 PBE39847.61.67.512.98.8 PCA97316.33.28.919.110.5 3561756.12.2 e 17.3 e Fig.8. OchrerecoveredfromKDSindicatedstratigraphicallyandaccordingtoa)rawmaterialfrequenciesandb)colour.(Forinterpretationofthereferen cestocolourinthis gure legend,thereaderisreferredtothewebversionofthisarticle.) C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 294

PAGE 12

sandstone(31.6%).PBEdisplaysthegreatestproportion(91.3%)of redochrederivedfrom ssileandinduratedshales. Colourwascollapsedintotengroupsincludingred,maroon, purple,pink,white,orange,yellow,brown,greyandblack( Fig.8 b). Geologicalandcolorimetricrelationshipscouldnotbeobjectively ascertained,principallybecausedestructiveanalyticalmethodsare requiredtodeterminesuchvariables( Dayetetal.,2013 ).Basicvisualclassi cationandcomparisonwiththeNaturalColourSystem (NCS)DigitalAtlas( http://www.ncscolour.com ,2013)wasthereforeusedforcolourclassi cationinthisstudy.Althoughvisible spectroscopycanprovidetheabsorbancespectraandcolourparametersoftheochreassemblage,thismethodwillonlyprovide informationconcerningthecolorimetricpropertiesoftheexternal surfacesofthespecimens.Red(62%)isthepredominantcolour, followedbymaroon(15.3%),orange(4.5%)andpink(4.5%).The remainderoftheassemblage(13.7%)includeslighter(yellowand white)anddarker(brownandblack)categories.Themajority(77%) ofredpiecesarederivedfrom ssileshales. 4.2.3.Utilizationstrategies OchreatKDSoccursintheformofresidualpowder,nodules, andfragmentsorasinclusionsinlargerpiecesofrock( Fig.9 ).Some examplesshowsignsofgrindingonhardabrasivesurfacesor scrapingwithsharp-edgedimplements.Indicationsofochreprocessingbygrindingorscraping( n ¼ 20)orbydeliberateknapping ( n ¼ 31)havebeenidenti edatKDS. Theproportionofmodi edpieces(17.5%)iswellwithinthe rangeofotherMSAsites( w 14%)( Watts,2002,2009,2010; Hodgskiss,2010;Dayetetal.,2013 )( Table5 ).SimilartotheMSA atDiepkloof( Dayetetal.,2013 )Sibudu( Hodgskiss,2010 ),Blombos ( Watts,2009 )andPinnaclePoint( Watts,2010 ),grindingisthe primaryprocessingtechnique.Ofthegroundpiecesincluding crayons,67.7%comprise ssileshale,12.9%induratedshale,6.5% mudstoneandsandstonerespectivelyand3.2%haematiteand ferricreterespectively.Fissileandinduratedshalesappeartohave beenpreferentiallyprocessedbygrinding(80.6%).Inaddition, 81.8%ofochrecrayonscomprisesofttohardred ssileshales.At DiepkloofandSibuduscrapingisnotaprimaryprocessingtechniqueandthepresenceofonlyasinglescrapedpieceatKDS(layer PBD)isthereforenotunusual.Clearindicationsofknappingoccur on31piecesfromlayersPBA/PBB( n ¼ 20),PBD( n ¼ 5),PBE( n ¼ 1) andPCA( n ¼ 5),suggestingthatknappingmayhaveformedpartof thechaîneopératoireofochreprocessingintheselayers( Fig.9 e). 4.3.Ostricheggshell Wehaveidenti ed95fragmentsofclearlyanddeliberately engravedostricheggshell(EOES)recoveredfromlayersPAYtoPCA (3.8%ofthetotalnumberofOESfragments).Themajorityofthe EOESpiecesderivefromPBC(27%)andPBD(25%)( Table6 ).An additional6engravedpieceswererecoveredfromlayerPAX(not reportedhere),andnoEOESfragmentswererecoveredfromanyof thelayersabovePAX.TheEOESisspatiallydistributedacrossthe areawhereHPlayerswereexcavated(4.75m2)andupto50cm belowthesurface.TherearenoLSAdepositsinKDSandduring excavationtherewasnosignofdisturbancetothedepositsthat mighthaveresultedfromtheintrusiveburialofengravedeggsat thesitebyLSApeople.TheEOESfragmentsareunderstudybut preliminaryobservationscanbemade.Thedesignsentailvariationsofcross-hatchedorsub-parallellinethemes,andmostare similartothosereportedfromDiepkloofintheHPandpre-HP layers( Texieretal.,2010,2013 )andfromtheHPlayersatApollo 11( Vogelsangetal.,2010 ).Allthedesignsidenti edatDiepkloof ( Texieretal.,2013 ,Table4:3423)arepresentatKDS,exceptforthe “ sub-parallelintersectinglinesmotif ” .Onedesignpresentinthe upperlayersatKDS,notreportedfromDiepkloof,consistsofa nelycarveddiamondshapedcross-hatchedpattern( Fig.10 a,b), distinctlydifferenttothosefromlayersbelow,andfromthe “ crosshatchedgridmotif ” reportedfromDiepkloof( Texieretal., 2013 :3420).Thisdiamondshapedpatternispresentonlyin layersPAX,PAYandPAZ.InPAXandPAYthisistheonlyengraved Fig.9. ExamplesofprocessedochrepiecesfromKDS:a)coarse-grainedgroundpurpleshalecobble(PCA),b)groundandpolishedshale-derivedcrayon-likep iece(PBC),c)ground andscrapedsoftshale-derivedspecimen(PCA),d)groundhardshalechunk(PBE)e)knappedandgroundhaematite-richshalefragment(PBC).(Forinte rpretationofthereferences tocolourinthis gurelegend,thereaderisreferredtothewebversionofthisarticle.) Table5 Theprevalenceofprocessedochrepiecesperlayer. Layer n Ground%Crayons%Flakes% PAY1715.9 eeee PAZ1218.3 eeee PBA/PBB11363.933.92019.6 PBC19421.1210.5 ee PBD5913.423.458.5 PBE3912.612.612.6 PCA9756.232.155.2 356196.2113.0318.3C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 295

PAGE 13

motifpresent.The “ sub-parallelrectilinearorcurvedlines ” design atDiepkloof( Texieretal.,2013 :3423)isthemostcommonly occurringmotifinlayersPBCtoPAZatKDS.OurstudyoftheEOESis ongoingbutinitialobservationssuggestsimilaritieswithmanyof theEOESmotifsfoundatDiepkloof,withsomedifferences. 5.Fauna 5.1.Macrofauna ApreliminaryanalysisofthemacrofaunalremainsfromthePAY toPCAlayerswasconductedfollowing Driver(2005) and Kleinand Cruz-Uribe(1984) .ThecomparativefaunalcollectionsoftheDitsongMuseumofNaturalHistoryinPretoriawereusedtoidentify boneremains.Micromammals,de nedasspecieswhereadults weighlessthan750g,arenotincludedinthisanalysis.Becauseof thedif cultyindifferentiatingbovids,manyremainswereassigned onlytosizeclassesbasedon Brain(1974) .Sizeclass1includessmall bovidssuchasCapegrysbok( Raphicerusmelanotis ),size2includes southernreedbuck( Reduncaarundinum ),size3includesred hartebeest( Alcelaphusbuselaphus ),andsize4arelargebovidssuch aseland( Tragelaphusoryx )andAfricanbuffalo( Synceruscaffer ). Althoughelandissometimesidenti edas Taurotragusoryx ,we followtheclassi cationschemeof SkinnerandChimimba(2005) e basedongeneticstudies(e.g., Essopetal.,1997 ) e andclassify elandas Tragelaphusoryx .Wealsousethesize5classforverylarge bovids,suchastheextinctlong-hornedbuffalo( Syncerusantiquus ). Long-hornedorgiantbuffaloarealsoknownas Pelorovisantiquus butwefollowmorerecentstudiesthatassignthemtothegenus Syncerus ( Gentry,2010;RectorandReed,2010;Faith,2013 ).Dueto thefragmentarynatureoftheassemblage,manymammalremains suchasrib,cranialorvertebralfragmentscouldnotbeidenti ed beyondclass.Thesespecimensareclassi edas ‘ small ’ , ‘ medium ’ , ‘ large ’ or ‘ verylargemammal ’ basedonsize.Smallmammalsare de nedasindeterminatespecimensranginginsizefromtheCape dunemolerat( Bathyergussuillus )uptoandincludingsize1bovids,mediummammalsuptosize2bovids,andlargemammalsaresize 3bovidsandlarger( Brain,1974 ). ‘ Verylargemammal ’ includesa specimenthatcouldnotbecon dentlyidenti edtoorderandmay beblackrhinoceros( Dicerosbicornis )orlong-hornedbuffalo.Small carnivoresrangeinsizetothatoftheAfricanwildcat( Felissilvestris ),mediumcarnivorestothesizeoftheAfricancivet( Civettictiscivetta )andlargecarnivoresaslargerthan C.civetta .Afew shremainswererecoveredatKDS,mainlyvertebraeandjaw bones,butthesehavenotbeenstudied. 5.1.1.Assemblage Ofthe28,128fragmentsofbone,weighing11,758g,2129(7.6%) couldbeidenti edtoatleasttheclasslevel,while292(1.0%)could beidenti edtogenus/species.BonefromKDSisextensivelyfragmented:themajorityofidenti edfragments( n ¼ 1343;63.1%)are lessthan2cminlengthand19.7%ofidenti edbone( n ¼ 419)is lessthan1cmlong.Thisextensivefragmentationislikelythe reasonwhytheMinimumNumbersofIndividuals(MNI)inall layersislowerthanexpected( MarshallandPilgram,1993 ).Fragmentationisprobablyaresultofburningwithevidencepresenton 1761fragments(82.7%ofidenti edbone).Althoughburningwas notrecordedforunidenti edbone,theproportionofunidenti ed burntspecimensappearsmuchthesameasintheidenti ed sample.Theelevatedproportionofburntboneislikelyduetothe highnumbersofhearthsandhearth-likestructuresoccurringat KDS.Mostofthefaunalmaterialwasrecoveredfromwithin,or closeto,thesehearthssuggestingthatmostoftheburntbonecan beassociatedwithcookingevents.This,andtherelativelackof carnivores,isastrongindicatorthathumanswerethemainaccumulatorsofthefaunalassemblage. 5.1.2.Identi edfauna Tortoiseremainsarecommonandconstitute31%oftheidenti edfauna( Table7 ).Themajorityofidenti edtortoisebonesare carapaceorplastronbutduetotheirsmallsizeitwasnotpossible todifferentiatetortoisetaxabasedonshellfragments.Mostof thesearelikelyangulatetortoise( Chersinaangulata ),although somemaybetheCapetortoise( Homopus sp.).Rockhyrax( Procavia capensis )isthemostprevalentidenti edmacromammal.Layers Table6 Frequencyofengravedandunmodi edOESthroughoutthesequence. LayerEOES( n )OES( n )%EOES PAY51064.7 PAZ151878.0 PBA/PBB2212741.7 PBC233496.6 PBD2520212.4 PBE4904.4 PCA12820.4 Total9524903.8 Fig.10. EngravedOESpiecesfromPAZ(aandb),PBC(candd)andPBD(eandf).Notethatbothcanddconsistoftwore ttedparts. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 296

PAGE 14

PAYandPAZaredominatedbymicromammalremains,andsmall mammalssuchashyraxandCapedunemoleratwithafewidenti edbovidbones.LagomorphremainswererecoveredfromPAY, PBC,PBDandPCAwithonespecimenidenti edasscrubhare ( Lepussaxatilis ).Equids( Equus sp.)arecommoninlayersPBA/PBB andPBCandmanyofthe ‘ largemammal ’ ribandvertebralfragmentsintheselayersareprobablyequidremains.Basedonvariationinlongboneandmetapodiasizes,itislikelythatthequagga ( Equusquaggaquagga )orplainszebra( Equusquaggaburchellii )and mountainzebra( Equuszebra )maybepresentbutthefragmented natureofthebonespreventspositiveidenti cation.Itisunclear whethertheCapezebra( Equuscapensis )ispresent. Capegrysbokorsteenbok( Raphicerus sp.)occurinmostlayers andaremostcommoninPBC,PBDandPBE.Asingleoribi( Ourebia ourebia )phalangewasidenti edinPBA/PBBwithsuf cient morphologicaltraitstodistinguishthisspecimenfromgreyduiker ( Sylvicapragrimmia ),klipspringer( Oreotragusoreotragus )orthe morecommon Raphicerus .LargerbovidsarerelativelymorecommoninPBA/PBB,PBCandPCA.Blesbokorbontebok( Damaliscus pygargus )remainswererecoveredfromtheselayers.One Damaliscus toothfragmentwasnoticeablylargerthan D.pygargus butsmallerthantsessebe( Damaliscuslunatus )andmaybethe extinctblesbok( Damaliscusniro ).Reedbuck( Redunca sp.)occursin PBA/PBBandPBD.Thevertebralfragmentassignedto ‘ verylarge mammal ’ likelybelongstolong-hornedbuffalo.Regardingalcelaphines,hartebeestwasdistinguishablefromblackwildebeest ( Connochaetesgnou )bytoothmorphology.Forexample,enamel infolds,particularlyonthemesialregionofthebuccalsurfacesof molars,aremorepronouncedinhartebeestthanwildebeest.The relativelyhighnumberof ‘ largemammal ’ ribandvertebralspecimensinlayersPBA/PBBandPBCisprobablyrelatedtothealcelaphinesrecoveredfromthoselayers.Elandremainsarerelatively morecommoninPCAandsuggestthattheotherunidenti edsize4 bovidspecimensfromthatlayerarealsomostlikelyelandsinceno Africanbuffalohavebeenidenti ed. 5.1.3.Comparisonswithothersites AsisthecaseatKDS,smallmammals(particularlydunemolerat andhyrax),smallbovidsandtortoisearecommoninthepre-70ka MSAlayersatBlombos( Henshilwoodetal.,2001;Thompsonand Table7 TheNumberofIdenti edSpecimens(NISP)andtheMinimumNumberofIndividuals(MNI)formacromammalandtortoiseremains.Bovidsizeclassesexcludespecimensthat couldbeidenti edtogenus/species.Small,medium,largeandverylargemammalsincludespecimenssuchascranial,ribandvertebralfragmentsthatcouldnotbecon dently identi edbeyondclass.Linnaeanclassi cationbasedon SkinnerandChimimba(2005) exceptSyncerusantiquus( Gentry,2010 ). TaxaCommonnamePAYPAZPBA/PBBPBCPBDPBEPCA MNINISPMNINISPMNINISPMNINISPMNINISPMNINISPMNINISP TestudinidaeTortoise25912327125132512105124 Chersinaangulata Angulatetortoise21024282112212813 cf. PelomedusidaeTurtle eeeeeeee 11 eeee Lagomorpha Hare/Rabbit11 eeee 2413 ee 13 Lepussaxatilis Scrubhare eeeeeeeeeeee 11 Lepus sp.Hare eeeeee 1111 eeee Bathyergussuillus Capedunemolerat2811 eeeeeeeeee Procaviacapensis Rockhyrax322211113272301116 Smallmammal3412833337837828231 Herpestidaesp.Mongoose eeeeeeee 11 eeee Smallcarnivore eeee 1112 ee 11 ee Feliscf.caracal / serval Caracal/Serval ee 11 ee 11 eeeeee Arctocephaluscf.pusillus Capefurseal eeeeeeee 12 eeee Mediumcarnivore ee 12 ee 11 eeeeee Mediummammal216283134380264217330 Parahyaenabrunnea Brownhyena eeee 11 eeeeeeee Dicerosbicornis Blackrhinoceros eeeeee 12 eeeeee Equus sp.Zebra eeee 21032512 ee 12 Raphicerus sp.Steenbok/Grysbok111111261715 ee Sylvicapragrimmia Greyduikereeee 12 ee 12 eeee Oreotragusoreotragus Klipspringer eeeeeeee 14 eeee Ourebiaourebi Oribi eeee 11 eeeeeeee Peleacapreolus Grey(Vaal)rhebok ee 1113 eeeeee 12 Redunca sp.Reedbuck eeee 13 ee 11 eeee Reduncafulvorufula Mountainreedbuck eeee 11 eeeeeeee Reduncaarundinum Southernreedbuck eeeeee 11 eeeeee Alcelaphini sp.Hartebeest/Wilderbeest eeee 131211 ee 11 Connochaetesgnou Blackwildebeest eeeeee 111111 ee Alcelaphusbuselaphus Redhartebeest eeee 111112 ee 11 Damaliscus sp.Blesorbontebok/?D.niro eeee 13 eeeeeeee Damaliscuspygargus Bontebok/Blesbok ee 111112 eeee 11 Tragelaphusoryx Eland eeeeeeeeee 1114 cf.Syncerusantiquus Giantbuffalo eeee 11 eeeeeeee Bovid129112182153622714 Bovid1/2 eeeeeeeeeeee 11 Bovid231314342231330114240 Bovid2/3 eeee 121713 ee 14 Bovid31621032022924512140 Bovid3/4 eeee 27 ee 171116 Bovid4111213 ee 1413119 Largemammal25132253452232617 Verylargemammal eeee 11 eeeeeeee Total2519221814139740423386462118025230C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 297

PAGE 15

Henshilwood,2014 ),intheHPlayersatDiepkloof( SteeleandKlein, 2013 )andatDieKelders( KleinandCruz-Uribe,2000 ).Ofthelarge bovidsrecovered,elandisrelativelycommonatBlombos,Die KeldersandDiepkloofbutrareatKDS.Incontrast,equid,quite commonatDiepkloofandKDS,isonlypresentintheearlier ( w 100ka)M3phaseatBlombosandrareatDieKelders.The prevalenceofequidsatKDSandwithintheHPlayersatDiepkloof suggestsagrassierenvironmentduringthisperiod. Damaliscus doesnotoccurateitherDiepklooforBlombosbutispresentatDie KeldersandKDS.WhileAfricanbuffalooccuratBlombosandDie Kelders,remainshavenotbeenrecoveredfromKDS.Futurestudies ofKDSfaunawillincludeassessmentofskeletalpro lesandsurface modi cationpatterns. 5.2.Shell sh Theshell shdatapresentedareasampleofthematerial retainedinthe3mmsievefromanumberofquadrates(between fourandsixperlayer)spanningthesequencefromPCAtoPAY.This datarepresents32.4%ofthetotalvolumeexcavatedfromthese layers.Justover29kgofshell shhasbeenanalysedfromavolume of0.51m3and7layers.Shellswereweighedandquanti edby determiningtheminimumnumberofindividuals(MNI)perlayer, basedoncountingtheapicesofgastropods,theleftandright umbosofbivalveswiththemostcommonsidetakenastheMNI, andthehighestnumberofeitherfront,backormiddlevalves (middlevalvecountsweredividedby6)ofchitonswastakenasthe MNI.Bothapicesandoperculaofthegiantperiwinkle Turbosarmaticus werecountedandthehighestcounttakenastheMNI.The greatestdimensionsofintactlimpetsandoperculaweremeasured withdigitalcalliperstothenearestmillimetre.Shellsthatwere < 2cm(whole)werenotconsideredtobefooditemsandwere recordedasincidentalshellsorjuvenilelimpets. Intotal,excludingtheincidentalandjuvenileshells,14species ofshell shwereidenti ed( Table8 ).Notethatalthoughtwo periwinklespecies, Dilomasinensis and Dilomatigrina ,arepresent, theirdatahavebeencombinedasthecountableapicesarenot identi abletospecieslevelwhentheshellsarebroken.Overall,the mostcommonspecies,intermsofMNI(absolute,perm3andin termsofrelativefrequency),isthegiantchiton, Dinoplaxgigas , followedbythebrownmussel, Pernaperna ,and T.sarmaticus .By weight,themostcommonspeciesistheabalone, Haliotismidae , followedby D.gigas ,and T.sarmaticus . H.midae , D.gigas and T.sarmaticus areconsistentlythemost commonbyweightrelativetootherspecieswithinlayers,although frequenciesdifferbetweenlayers.Astheyareallrelativelylarge animalswithheavyshells,theirdominancebyweightisnotthat surprising,althoughtheytendtodominatetheassemblagein termsofMNIaswell.OnlyinPAYistheArgenville ’ slimpet, Scutellastraargenvillei ,whichisrareorabsentinotherlayers,the secondmostcommonintermsofweight(38%).Thiscouldbea functionofthesmallersamplesizeinPAY.IntermsofMNI,the rangeofthemostcommonspeciesismorevariedbetweenlayers, butinmostinstances D.gigas and T.sarmaticus aremostcommon, buttherelativepercentageof P.perna increasesinmostlayers.Four species,theblackmussel Choromytilusmeridionalis ,sandmussel Donaxserra ,kelplimpet Cymbulacompressa andbeardedlimpet Scutellastrabarbara arepresentinsuchsmallquantitiesthattheir contributiontothedietoftheKDSoccupantswouldhavebeen minimal. Thereisashiftintherelativepercentagebyweightofthethree mostcommonspecies e atthebaseofthesequence,PCA, T.sarmaticus isthemostcommon,inPBEandPBD H.midae isthe mostcommon,inPBC D.gigas isonlyslightlymorecommonthan H.midae ,andinallthelayersabove D.gigas isthemostfrequently occurringspeciesbyweight( Fig.11 ).Thesameshiftisevidentfrom thedensitydata(g/m3)ofthesespecies. 5.2.1.Densitiesbylayer Thereislittleornoshell shpresentinlayersbelowPCA.That whichhasbeenrecordedisthoughttoderivefromPCAabove, wherePCAandPCBcouldnotbeseparated.DensitiesareconsiderablyhigherinlayersPBCandPBD,with183kg/m3and181kg/m3respectively,thaninanyotherlayers( Table8 ),andgraduallydecreasestowardsthetopofthesequence,withlessthan3kg/m3in PAY.Shell shvolumesdecreasedrasticallyabovePAY( < 1kg/m3), andonlyincreaseagaininlayersPAQ(1.4kg/m3)andabove(upto 12.2kg/m3inPAL).Notwithstandingtheeffectsofvolumereductionovertimeandgeomorphologicalprocessesthatcannotbe accountedfor,PBCandPBDareverydenseshelllayers,bothrelativetootherlayerswithinthissite,aswellastootherMSAsites withshell shremainsandpublishedvolumes.AtKlasiesRiver,the highestrecordeddensityofshell shisintheMSAII,at162kg/m3( Thackeray,1988 ).TheHPlayersatKlasiesRivershowagradual declineinshell shvolumesthroughtime,startingat8.7kg/m3in thelowerlayers,andendingin0.8kg/m3intheuppermostHP layers.AtBlombos,shell shvolumesarehighestinlayerCIinthe M3phase,c.100ka,at163.8kg/m3( Henshilwoodetal.,2001 ),and lowerintheM2(c.80ka)andM1(c.75ka,StillBay)phases,with 31.8kg/m3and17.5kg/m3respectively.AtPinnaclePointCave PP13B,inlayersdatingbetween90and164ka,shelldensitiesare relativelylow,rangingfrom0.01kg/m3to8.7kg/m3( Jerardinoand Marean,2010 ).Shell shdatahavenotbeenprovidedfortheHP layersatPinnaclePointSitePP5 e 6,butdensitiesappeartobelow ( Brownetal.,2012 ). Althoughthedensityofshell shdeclineswithtime,thespecies compositiondoesnotindicateachangeinthedistancefromthe shoresigni cantenoughtoresultinchangesincollectionstrategies,forexampleanincreasein P.perna ,whichcanbetransported overgreaterdistances,oradecreaseinlargehighyieldspecies whendistancesexceed5km( Langejansetal.,2012 ). 5.2.2.Shell shsize Ithasbeenarguedthatreductionsinshell shsizecanbeused asaproxyforintensi cationofshell shgatheringandincreased groupsize(e.g. KleinandSteele,2013 ),althoughsomesuggestthat theroleofenvironmentalfactorsonshell shgrowthratesmightbe moresigni cantthanpreviouslyconsidered( SealyandGalimberti, 2011 ).Thenumberofmeasurableshellsfromthecurrentsampleis small,buthasbeenincludedhereforcompleteness.Veryfewofthe Cymbulagranatina shellswereintactenoughformeasurement,and allarefromPBDandPBC.Fromthesmallmeasurablesample ( n ¼ 10),themedianis67.5mm,mean67.4mm,minimum57mm andmaximum79mm.Thesesizesaresmallerthantheaverageof modern C.granatina fromunexploitedareasontheCapewestcoast ( Parkingtonetal.,2013 ),andsomewhatsmallerthanthosereportedfromMSAcontextsonthewestcoast,exceptforBoegoeberg 2,wheresizesaresimilar( SteeleandKlein,2008 ;exactmeasurementsarenotprovided). Thecurrentmeasurable Cymbulaoculus sampleisalsosmall (median72.5mm,mean71.2mm,minimum55mm,maximum 84mm, n ¼ 16).Thesesizesaresomewhatsmallerthanthatof C.oculus fromtheHPatKlasiesRiver,butbiggerthananypublished LSAdata( KleinandSteele,2013 ). Allmeasurements( n ¼ 63)of T.sarmaticus operculaarefrom layersPCAtoPBA/PBB.Themedianlengthis38mm,mean 36.9mm,minimum14mmandmaximum50mm.Thisissmaller thanthosefromtheHPatKlasiesRiver,largerthananypublished LSAoperculasizes,andmostsimilarinsizetothosefromtheMSAI andIIfromKlasiesRiver( KleinandSteele,2013 ).C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 298

PAGE 16

5.2.3.Collectionstrategies Theabundanceof D.gigas , H.midae and T.sarmaticus indicates thattheinhabitantsweretargetingspecieswithhighmeatyield rates( Langejansetal.,2012 ).Thesethreespeciescontributedthe highestaveragemeatweightperm3ineverylayer(cf. Avery,1976 ). Theyareusuallyonlycollectibleatlowtidesandintheinstanceof H.midae ,thekelplimpet C.compressa and S.barbara ,springlow tides.Thusitappearsthatthemajorityofshell shcollectionwas scheduledtocoincidewithlowtides. 5.3.Humanremains Anearlycompletecrownofanisolatedhumanleftmandibular deciduoussecondmolar(Ldm2)wasrecoveredfromquadrate S29b,layerPBEdatedatc.64ka( Fig.4 ,PBEliesbetweenlayers datedto64.6 4.2and63.5 4.7ka).Adescriptionofthemolaris inpreparationbyHavartietal. 6.Palaeoenvironment 6.1.Fauna Thehighdensitiesofshell shatKDSsuggestthatitwaslocated closetotheshoreduringmostoftheHPoccupation.Thecoldwater endemicshell shspecies, C.granatina ,orgranitelimpet,which doesnotoccuronthesouthcoasttoday,ispresentinrelatively smallquantitiesthroughoutthesequence,andmostcommonin termsofweight(8.4kg/m3)inPBC.Theirpresencesuggeststhatsea surfacetemperatures(SST)werecoolerthanpresent,althoughthe abundantpresenceofwarmerwaterspeciessuchas T.sarmaticus and D.gigas mitigatesagainstextremedifferencesintemperature. Thefewfragmentsof C.meridionalis ,aspeciesmostabundanton thecolderwestcoasttoday,couldalsosupportcoolerconditions, althoughitisprobablyonlyagoodindicatorofcoolerconditions whenitoutnumbersitswarmerwatercounterpart, P.perna ,which isnotthecasehere. Thespeciescompositionindicatesrockyshores,withthe exceptionofafewfragmentsof D.serra inPAY,whichisasandy beachinhabitant.Thesteadyincreasein D.gigas attheexpenseof H.midae and T.sarmaticus intheupperlayerscouldindicatean increaseinsandyconditions,as D.gigas ismoretolerantofsandy environmentsthantheothertwospecies( KilburnandRippey, 1982;Wood,1993;Yssel,1989 ). TheterrestrialfaunafromKDSconsistslargelyofspeciesthat occurredintheareahistorically( Skead,1980 ).Theabundanceof rockhyraxesindicatesrockyhillsidesassociatedwithshrubs, consistentwiththefynbosandrockycrevicessurroundingKDS today.Thepresenceofterrapinandreedbuckimpliesanearby fresh-watersourcesuchasawetlandorriverbed.Southernreedbuck( R.arundinum )prefertallgrassorreedbedsforcoverandare typicallyfoundingrasslandsadjacenttowetlandsor vleis . Table8 Minimumnumberofindividuals(MNI),weight(g)anddensity(MNI/m3andkg/m3)ofshell shfromlayersPCAtoPAY. SpeciesPAYPAZPBA/PBBPBCPBDPBEPCA MNIgMNIgMNIgMNIgMNIgMNIgMNIg Choromytilusmeridionalis ee 1 < 1 ee 1 < 1 eeeeee Pernaperna 15826144968123225432591823 Donaxserra 11 eeeeeeeeeeee Burnupenacincta 22210494563238464914 Haliotismidae 1214723341834452757391220521284 Cymbulacompressa ee 13121312025 ee Cymbulagranatina 1113126204178328163143 Cymbulaoculus 1111013119162202138010210559 Scutellastraargenvillei 148131 < 11115913142 Scutellastrabarbara eeeeeeeeee 1511 Patella spp. e 3112 e 29 e 36 e 65 e 49 e 89 Diloma sp. ee 1512417543363130 Turbosarmaticus 11331651757219707251265301263451290 Dinoplaxgigas 6542574211418772063612251193275466233 Shellfragments e 7 e 19 e 155 e 121 e 47 e 54 e 19 TotalMNIandg151354610461683281423906114191971264322802118 DensityMNI/m3andkg/m32943703161660328526183277118111594093125 Non-foodspecieseeeeeeeeeeeee Incidentalshells e 2 e 1 e 1 e 5 e 3 e 14 e 31 Juvenile Patella sp. eee 0 e 3 e 8 e 7 e 5 e 1 Fig.11. Relativefrequency(%)perlayerofthethreemostcommonshell shspecies basedonweight. C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 299

PAGE 17

Mountainreedbuckfavourdry,grass-coveredmountainslopesbut alsorequiretheavailabilityoffreshwater( SkinnerandChimimba, 2005 ).TheKlipdriftfonteinstreamliesclosetoKDSandcouldhave beenthesourceofthisfreshwater.Southernreedbuckisnot knowntooccurinfynboshistoricallyanditspresencecouldindicatemoisterthanpresentorhistoricalconditions( Skead,1980 ). Thepresenceofdunemole-ratsinPAZandPAYandtheabsence ofthisspeciesinallotherlayerscouldindicateachangeinthelocal environmenttomoresandyconditions.Dunemoleratsareassociatedwithmesiccoastal,sandy,environments( Bennettetal.,2009 ). Increaseinsizeisthoughttocorrelatewithincreasedmoisture ( Klein,1991 ).Unfortunatelymole-ratremainsfromKDSweretoo fragmentedformeasurement.Thelackofdunemole-ratremainsin otherlayersmayalsobeduetothesmallsamplesizeorchangesin taphonomicconditionsacrossthesequencethathaveyettobe identi ed.LikewisetheabsenceofCapefursealatKDScouldbe duetotaphonomicandpreservationalissues,ortherelativelysmall samplesize.Conversely,sealsmayhavebeenbutcheredclosetothe shorewithlittleosteologicalmaterialtransportedbacktothesite. Palaeoenvironmentalanalysesoflargeherbivorecommunities arestilltentativeforKDS,asidenti edremainsarerare.Thesample sizesforlayersPCA,PBE,PAZandPAYaretoosmall(totalNISP < 10) toallowforasecureinterpretation.Wethusfocushereonlayers PBD(totalNISP ¼ 20),PBC(40)andPBA/PBB(29).Althoughevidencefortemporalchangesshouldbetreatedwithcautiondueto thecurrentsmallsamplesize,someobservationscanbemade. Somepatternsareapparentwhenlargemammaldataisinterpretedintermsofgrazer/browserratiosandmainhabitatpreferences( Fig.12 ,interpretationsbasedonmoderndatacf. Rectorand Reed,2010;SkinnerandChimimba,2005;Sponheimeretal.,2003 ). While65%ofidenti edbonesinPBDcorrespondtoungulatesthat aremainlybrowsers(steenbok/grysbok,greyduikerandklipspringer),theupperHPlayersindicateaconsiderableincreasein ungulatesthataremainlygrazers(equids,redhartebeest,southern reedbuck,mountainreedbuck,blackwildebeest,bontebok/blesbok andoribi).InlayersPBCandPBA/PBBtheyrepresent80%and79%of identi edungulatesrespectively( Fig.12 ). Thissharpincreaseinpresenceofgrazersisparallelledbydata relatingtomainhabitatpreferences.Thereappearstobeachange fromslightlymorebushyterraininPBDtoanenvironmentdominatedbygrasslandsinPBC,potentiallyinterspersedwithwoodlandsandshrubs(asindicatedbythepresenceofblackrhinoceros). PBCdocumentsthedevelopmentofafullsuiteofungulatesthatare preferentiallyfoundinopengrassland/savannahecosystems(6out ofthe6taxaidenti edinthelayer),withequidsrepresenting63% ofidenti edungulates.TheenvironmentinPBA/PBBissomewhat intermediatebetweenPBDandPBC. Thedevelopmentofagrassland-dominatedecosysteminPBC around66kamaycorrespondtoanincreasedfrequencyofC4 plantsfollowinganincreaseinsummerrain.Isotopicstudiesofthe nearbyCreviceCavespeleothems( Bar-Matthewsetal.,2010 )supportthishypothesis:intheserecords,increasesind13Candd18O around68kahavebeeninterpretedasindicativeofcorrelatedincreasesinC4plantsandinsummerrainrespectively. 7.Discussion ThecomprehensivedatacollectionstrategyadoptedatKDS duringthe2011 e 2013excavationsoftheHPlayers,andthesubsequentandon-goinganalysisofthisassemblage,allowsforpreliminaryobservationstobemadeandprovideasoundbasisfor futureexcavationsatthesite.Thesmallassemblagerecoveredfrom theupperlayersofKDS(PAL e PAN/PAO)withanageof 51.7 3.3kaprovidesatentativeglimpseofthepost-HPlayersand willbeonefocusoffutureexcavations. TheHPlayersatKDShavebeendatedbyOSLto65.5 4.8kato 59.4 4.6ka.SimilarOSLdateshavebeenattainedforanumberof otherHPassemblagesinsouthernAfrica,suggestingthattheHPisa relativelyshort-livedindustry( Jacobsetal.,2008 ).However,the chronologyproducedby Jacobsetal.(2008) hasrecentlybeen questionedby Guérinetal.(2013) ,whoclaimthattheHPagesare erroneouslyprecise,andthatthe “ adjusteddoserate ” modelused by Jacobsetal.(2008) isincorrect.Itisbeyondthescopeofthis papertoattempttoadjudicateeithercriticism,butwenotethatthe Fig.12. Palaeoenvironmentalanalysesoflargeherbivorecommunities(grazer/browserratiosanddataonmainhabitatpreferencesexpressedasNISPpropor tionsandpresence/ absenceofeachtaxa). C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 300

PAGE 18

“ adjusteddoserate ” modelwasnotappliedtoKDSsamples,and individualagespresentedherehaverelativeuncertaintieswhich areconsistentwiththeexpectationsof Guérinetal.(2013) .New OSLandthermoluminescenceagesfromDiepkloof( Triboloetal., 2013 )alsocontradictthe ndingsof Jacobsetal.(2008) ,indicatingamuchlongerHPchronology,withanearlyHPatc.109ka anda nalHPatc.52ka.Thefullrangeoftheselatteragesforthe HPisnotevidentatKDS. ThelithicassemblagesfromPCAtoPAYcorrelatewiththeHP complex.Theyevidenceanumberofchanges,involvingrawmaterialcomposition,frequenciesandtypesofretouchedtools,which relatetothreemainphasesoccurringduringagradualprocessof change.Similarpatternsofchangethroughtimearedocumentedat KlasiesRiver( Villaetal.,2010;Wurz,2000 )andatDiepkloofinthe intermediateandlateHPlayers( Porrazetal.,2013a,2013b ).The lowerKDSlayers(PCA,PBE)shareanumberofsimilaritieswiththe lowerphaseatKlasiesRiverandtheIntermediateHPphaseat Diepkloof,whilethemiddleKDSlayers(PBC,PBA/PBB)correspond totheupperpartofthesequenceatKlasiesRiverandtothelateHP atDiepkloof.ThisdiagnosisforKDSisbasedonthelayerswhichare currentlyavailableforanalysis.Furtherresearchontheunderlying andoverlyinglayerswillundoubtedlycompleteandre nethis preliminaryassessment. Intermsofochreprocessingstrategiesandgeologicaldiversity, theKDSassemblageappearstoexhibitfourdistinctivephases. Whereasthelowestlayer(PCA)resemblesPBDandPBCintermsof displayingthestandardrangeofprocessingtechniquesand geologicalvarieties,layerPBEconsistsofadenseconcentrationof thoroughlyprocessedshale-derivedredochre.PBEcontainsthe highestconcentrationofredochrederivedfrom ssileshales,and therangeofgeologicaltypesarelimitedrelativetootherlayersin thesequence.LayerPBEthereforerepresentsabreakinstandard pigmentselectionandprocessingstrategiesdisplayedbythe samplesrecoveredfromlayersPAYtoPBD.Aspowderedochremay havebeenusedforvariouspurposes( Bonneauetal.,2012,d ’ Errico etal.,2012;Henshilwoodetal.,2009,2011;Rifkin,2011;Soriano etal.,2009;Wadleyetal.,2009 ),suchhighvolumesmaybe indicativeofthedeliberateprocessingoflargeamountsofochrefor veryspeci cpurposes.Followingthisemphasisonprocessing ochreinto nepowder,layerPBCexhibitsthelargestassemblyof ochrecrayonsandthewidestgeologicalvariability.InlayerPBA/ PBB,therawmaterialcompositionremainslargelyunchangedbut thereisincreasedevidencefor akingasaprimaryprocessing strategy.Theupperlayers(PAYandPAZ)displaytheleastvariability intermsofrawmaterialselectionandprocessingtechnique employed. TheprevalenceofsmallmammalsandtortoiseatKDSissimilar tothatfoundatmanyotherMSAsitesinthesouthernandwestern Cape.LargermammaldatafromKDS e particularlyalcelaphines andequids e suggestsanenvironmentwheregrassesfeaturemore prominentlythantheydidhistorically,ashasbeennotedfortheHP layersatDiepkloof( SteeleandKlein,2013 ).Inotherreportssigni cantfaunalchangesduringtheHPperiodarenotemphasized. AtKDS,whilemostlayers(PAZ,PBA/PBB,PBCandPCA)aredominatedbyremainsofmediumandlargemammals(mainlybovids andequids),othersaredominatedbytortoiseremains(PBDand PBE)andlayerPAYbysmallmammals(particularlyrockhyraxand Capedunemolerat).TheKDSsequencealsodocumentschangesin therelativeproportionofsmallbovids(e.g.Capegrysbok/steenbok, klipspringer,greyduiker;morecommoninlayerssuchasPBD)and ofequidsandlargerbovids(e.g.redhartebeest,blackwildebeest, bontebok/blesbok,eland e morecommoninlayerssuchasPBC).FurtherstudiesoftheKDSfaunawillincludetaphonomicalanalysestodecipherhowthesepatternscorrelatewithenvironmental and/orsubsistencechanges.Thesigni cantextentofthesefaunal changesmightimplythatHPhunter-gathererschangedtheir subsistencestrategiesandadaptedtovaryingenvironments,while notnecessarilymodifyingthemaincharacteristicsoftheirtechnicalandculturalbehaviours.KDScanplayaroleinfutureresearch focusedonunderstandingtheinterplaybetweenculturalchanges, especiallyinlithictechnology(see 4.1 above),andsubsistence strategiesduringtheHP. Theshell shdataherecomplementthatfromotherknownHP locationswithshell shsuchasKlasiesRiverandDiepkloof.Asat KlasiesRiver,thedensityofshell shdeclineswithtimethroughthe HP( Thackeray,1988 ),whereastheoppositeistrueatDiepkloof ( SteeleandKlein,2013 ).Thehighdensityofshell sh,particularlyin layersPBCandPBD,suggeststhatthecoastlinewasnearby,and lowerdensitiesintheyoungerlayerscouldre ectaretreatofthe coastduetoloweringsealevels.Conversely,thepresenceofdune mole-ratremainsonlyintheupperlayersPAYandPAZ,where shell shdensitiesarelowest,impliesthepresenceofdunesand andanearbycoastline.Moredataareneededtoaddressthese con ictingsignals.Thelowincidenceof shbonescouldbedueto taphonomicprocessesasboneisgenerallypoorlypreservedand fragmented,and shboneisevenmoresusceptibletodegradation thanmammalianbone( Szpak,2011 ). Ostricheggshellisabundantthroughoutthesite.Thepresence ofatleast95OESpiecesengravedwithabstractpatterns,similarto thatreportedfromonlytwootherHPcontexts,Diepkloofand Apollo11,extendsthegeographicextentofthisculturaltradition. 8.Conclusion KDSisanewlydiscoveredcoastalsiteinthesouthernCape containinglithicstypicaloftheHP.Itisthe rstknowntypicalHP site(see Henshilwood,2012 )locatedonthec.600kmofcoastline betweenNelsonBayCave,PlettenbergBayandPeersCave(Skildegat)ontheCapePeninsula( Fig.1 ).Noanthropogenicdepositswere recoveredatKDSthatpredate65.5 4.8suggestingthatthec.71ka earlyHP-liketechnologyreportedatPinnaclePoint( Brownetal., 2012 )andthe > c.72kaStillBayphasesfromnearbyBlombos Cave( Henshilwood,2012 )aretechnocomplexesthatpredatethe KDSHPdeposits.Nevertheless,theKDSassemblageprovidesa usefulcorollarywiththeearlierBlombosandPinnaclePointdataon coastalsubsistencepatternsduringtheMSAinthisregion.Future excavationsattheadjacentKDCLsitewithMSAdepositsthatpredatec.70ka,andofthepost e HPlayers(forwhichasingleageof 51.7 3.3kaiscurrentlyavailable)atKDS,willaddtothisknowledge.Theapparentabsenceofshell shatthissiteisworthnoting. ThecurrentfaunalsamplefromKDSistoosmallforde nitive statementsregardingenvironmentalconditionsduringtheHPin thisregion,althoughtentatively,themacromammalandshell shdatapointtosomechangesinrainfallregimesandlocalenvironmentswithinthesequence.Additionaldatafrommicrofauna,isotopicanalysisandlargermacrofaunalsampleswillcontributeto re ningtheseobservationsandtothegreaterpictureofenvironmentalconditionsduringthisperiod.Itisworthnotingthatthe environmentalchangethatisevidentinlayerPBCapparentlycorrespondstoachangeinlithicrawmaterials,fromapredominant exploitationofsilcretetoanincreasedimportanceofquartz exploitationandalsoamarkeddecreaseinpigmentexploitation. FutureresearchontheKDSHPwillfocusonunderstandingtherole playedbyenvironmentalchangesintheevolutionofrawmaterial andfoodprocurementstrategiesbyMSAhunter-gatherers.The developmentofanopenlandscapemighthavein uencedgeneral mobilitystrategies,affectingbothhuntedspeciesandaccesstoraw materials. WhethertheengravedOESfromKDSindicatescontinuityinthe practiceofmarkingordecorationofmaterialcultureintheC.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 301

PAGE 19

southernCape,asisevidencedatBlombos( Henshilwoodetal., 2009 )intheStillBayandpre-StillBaylayers,isnotclearatthis stage.Thisisespeciallysoasthereisananthropogenicallysterile sandlayerabovetheterminalMSAdepositsatBlombosandbelow the rstMSAdepositsatKDS.However,theplanneddetailed studiesoftheKDSengravedOESmayprovidefurtherevidenceon likelyculturallinkswithotherWesternCapesites. TherecentdiscoveryandexcavationofKDShelpsreinforcethe notionthat H.sapiens usingHPtechnologywerefairlywidely spreadinSouthAfricabetween66and59ka,wereabletoadaptto arangeofenvironmentalconditionsandyetproducedatechnology thatisfairlystandardized.Thelattersuggestsadeliberatecontinuityinmaterialculturestylesprobablyreinforcedbyfrequent contactamongandbetweenthegroupsthatrangedacrossthis region.Someoftheculturaltraditions,suchastheengravingof OES,appearinfrequently,buttheirpresenceinsitesonthewest coastandnowthesouthernCapereinforcesthisnotionofcontact withinafarreachingsocialnetwork. Acknowledgements FinancialsupportfortheKDSprojectwasprovidedtoCSHbya EuropeanResearchCouncilAdvancedGrant,TRACSYMBOLSNo. 249587,awardedundertheFP7programmeattheUniversityof Bergen,NorwayandbyaNationalResearchFoundation/DepartmentofScienceandTechnologyfundedChairattheUniversityof theWitwatersrand,SouthAfrica.AdditionalfundingfortheKDS excavationsin2013wasprovidedbytheNationalGeographicExpeditionsCouncil,grantnumberEC0592-12.Wewouldliketo extendthankstotheboardofCapeNature,andespeciallyTierck HoekstraandCallumBeattie,foraccesstotheKlipdriftComplex andthefacilitiesatPotberg.WethankKurtMuchleratNGS,and JohnComptonforusefulcomments.WethankGauthierDevilder forthelithicdrawings.ShawBadenhorstattheDitsongNational MuseumofNaturalHistoryprovidedhelpinthemammalianfaunal identi cationandallowedaccesstotheirfaunalreferencecollection.Wethanktheanonymousrefereesfortheirhelpfuland constructivesuggestions. ReferencesArmitage,S.J.,Bailey,R.M.,2005.Themeasureddependenceoflaboratorybetadose ratesonsamplegrainsize.Radiat.Meas.39,123 e 127 . Armitage,S.J.,King,G.E.,2013.Opticallystimulatedluminescencedatingofhearths fromtheFazzanBasin,Libya:atoolfordeterminingthetimingandpatternof HoloceneoccupationoftheSahara.Quat.Geochronol.15,88 e 97. http:// dx.doi.org/10.1016/j.quageo.2012.10.002 . Armitage,S.J.,Jasim,S.A.,Marks,A.E.,Parker,A.G.,Usik,V.I.,Uerpmann,H.-P.,2011. ThesouthernrouteoutofAfrica:evidenceforanearlyexpansionofmodern humansintoArabia.Science331,453 e 456. http://dx.doi.org/10.1126/ science.1199113 . Avery,G.,1976.ASystematicInvestigationofOpenStationShellMiddensalongthe SouthwesternCapeCoast.UniversityofCapeTown . Bar-Matthews,M.,Marean,C.W.,Jacobs,Z.,Karkanas,P.,Fisher,E.C.,Herries,A.I., Brown,K.,Williams,H.M.,Bernatchez,J.,Ayalon,A.,2010.Ahighresolutionand continuousisotopicspeleothemrecordofpaleoclimateandpaleoenvironment from90to53kafromPinnaclePointonthesouthcoastofSouthAfrica.Quat. Sci.Rev.29,2131 e 2145. http://dx.doi.org/10.1016/j.quascirev.2010.05.009 . Bateman,M.D.,Holmes,P.J.,Carr,A.S.,Horton,B.P.,Jaiswal,M.K.,2004.Aeolianite andbarrierduneconstructionspanningthelasttwoglacial-interglacialcycles fromthesouthernCapecoast,SouthAfrica.Quat.Sci.Rev.23,1681 e 1698. http://dx.doi.org/10.1016/S0277-3791(04)00043-5 . Bennett,N.C.,Faulkes,C.G.,Hart,L.,Jarvis,J.U.M.,2009. Bathyergussuillus (Rodentia: Bathyergidae).Mamm.Species828,1 e 7. http://dx.doi.org/10.1644/828.1 . Bonneau,A.,Pearce,D.,Pollard,A.,2012.Amulti-techniquecharacterizationand provenancestudyofthepigmentsusedinSanrockart,SouthAfrica.J.Archaeol. Sci.39,287 e 294. http://dx.doi.org/10.1016/j.jas.2011.09.011 . Bøtter-Jensen,L.,Andersen,C.E.,Duller,G.A.T.,Murray,A.S.,2003.Developmentsin radiation,stimulationandobservationfacilitiesinluminescencemeasurements.Radiat.Meas.37,535 e 541 . Brain,C.K.,1974.Somesuggestedproceduresintheanalysisofboneaccumulations fromsouthernAfricanQuaternarysites.Ann.TransvaalMus.29,1 e 8 . Brown,K.S.,Marean,C.W.,Jacobs,Z.,Schoville,B.J.,Oestmo,S.,Fisher,E.C., Bernatchez,J.,Karkanas,P.,Matthews,T.,2012.Anearlyandenduring advancedtechnologyoriginating71,000yearsagoinSouthAfrica.Nature491, 590 e 593 . Carr,A.,Bateman,M.,Holmes,P.,2007.Developinga150kaluminescencechronologyforthebarrierdunesofthesouthernCape,SouthAfrica.Quat.Geochronol.2,110 e 116 . Chapot,M.S.,Roberts,H.M.,Duller,G.A.T.,Lai,Z.P.,2012.Acomparisonofnaturalandlaboratory-generateddoseresponsecurvesforquartzopticallystimulatedluminescencesignalsfromChineseLoess.Radiat.Meas.47,1045 e 1052 . Compton,J.S.,2001.Holocenesea-level uctuationsinferredfromtheevolutionof depositionalenvironmentsofthesouthernLangebaanLagoonsaltmarsh,South Africa.Holocene11,395 e 405 . Compton,J.S.,2011.Pleistocenesea-level uctuationsandhumanevolutiononthe southerncoastalplainofSouthAfrica.Quat.Sci.Rev.30,506 e 527 . Conard,N.,Porraz,G.,Wadley,L.,2012.Whatisinaname?Characterizingthe “post-Howieson ’ sPoort ” atSibudu.S.Afr.Archaeol.Bull.67,180 e 199 . Cornell,R.M.,Schwertmann,U.,2003.TheIronOxides:Structure,Properties,Reactions,OccurrencesandUses.Wiley,Weinheim . Dayet,L.,Texier,P.-J.,Daniel,F.,Porraz,G.,2013.OchreresourcesfromtheMiddle StoneAgesequenceofDiepkloofRockShelter,WesternCape,SouthAfrica. J.Archaeol.Sci.40,3492 e 3505 . Deacon,H.J.,Geleijnse,V.B.,1988.Thestratigraphyandsedimentologyofthemain sitesequence,KlasiesRiver,SouthAfrica.S.Afr.Archaeol.Bull.43,5 e 14 . Driver,J.C.,2005.ManualforDescriptionofVertebraeRemains.CrowCanyon ArchaeologicalCenter,Cortez. http://dx.doi.org/10.6067/XCV8BP01K6 . Duller,G.A.T.,Bøtter-Jensen,L.,Murray,A.S.,Truscott,A.J.,1999.Singlegrainlaser luminescence(SGLL)measurementsusinganovelautomatedreader.Nucl. Instrum.MethodsPhys.Res.Sect.B:BeamInteract.Mater.Atoms155,506 e 514. http://dx.doi.org/10.1016/S0168-583X(99)00488-7 . Duller,G.A.T.,Bøtter-Jensen,L.,Murray,A.S.,2000.Opticaldatingofsinglesandsizedgrainsofquartz:sourcesofvariability.Radiat.Meas.32,453 e 457 . d ’ Errico,F.,GarcíaMoreno,R.,Rifkin,R.F.,2012.Technological,elementaland colorimetricanalysisofanengravedochrefragmentfromtheMiddleStoneAge levelsofKlasiesRiverCave1.S.Afr.J.Archaeol.Sci.39,942 e 952 . Eastaugh,N.,Walsh,V.,Chaplin,T.,Siddall,R.,2008.PigmentCompendium:a DictionaryandOpticalMicroscopyofHistoricPigments.ButterworthHeinemann,Oxford . Essop,M.F.,Harley,E.H.,Baumgarten,I.,1997.Amolecularphylogenyofsome bovidaebasedonrestrictive-sitemappingofmitochondrialDNA.J.Mammal. 78,377 e 386 . Faith,J.T.,2013.Taphonomicandpaleoecologicalchangeinthelargemammal sequencefromBoomplaasCave,westernCape,SouthAfrica.J.Hum.Evol.65 (6),715 e 730 . Galbraith,R.F.,Roberts,R.G.,Laslett,G.M.,Yoshida,H.,Olley,J.M.,1999.Optical datingofsingleandmultiplegrainsofquartzfromJinmiumrockshelter, NorthernAustralia:part1,experimentaldesignandstatisticalmodels. Archaeometry41,339 e 364 . Gentry,A.W.,2010.Bovidae.In:Werdelin,L.,Sanders,W.J.(Eds.),CenozoicMammalsofAfrica.UniversityofCaliforniaPress,Berkeley,pp.741 e 796 . Guérin,G.,Murray,A.S.,Jain,M.,Thomsen,K.J.,Mercier,N.,2013.Howcon dentare weinthechronologyofthetransitionbetweenHowieson ’ sPoortandStillBay. J.Hum.Evol.64(4),314 e 317 . Henshilwood,C.S.,2012.LatePleistoceneTechno-traditionsinSouthernAfrica:a reviewoftheStillBayandHowiesonsPoort,c.75 e 59ka.J.WorldPrehist.25, 205 e 237. http://dx.doi.org/10.1007/s10963-012-9060-3 . Henshilwood,C.S.,Sealy,J.C.,Yates,R.,Cruz-Uribe,K.,Goldberg,P.,Grine,F.E., Klein,R.G.,Poggenpoel,C.A.,vanNiekerk,K.,Watts,I.,2001.BlombosCave, SouthernCape,SouthAfrica:preliminaryreportonthe1992-1999excavations oftheMiddleStoneAgelevels.J.Archaeol.Sci.28,421 e 448 . Henshilwood,C.S.,d ’ Errico,F.,Watts,I.,2009.EngravedochresfromtheMiddleStoneAgelevelsatBlombosCave,SouthAfrica.J.Hum.Evol.57,27 e 47. http:// dx.doi.org/10.1016/j.jhevol.2009.01.005 . Henshilwood,C.S.,d ’ Errico,F.,vanNiekerk,K.L.,Coquinot,Y.,Jacobs,Z., Lauritzen,S.-E.,Menu,M.,García-Moreno,R.,2011.A100,000-year-oldochreprocessingworkshopatBlombosCave,SouthAfrica.Science334,219 e 222. http://dx.doi.org/10.1126/science.1211535 . Hodgskiss,T.,2010.Identifyinggrinding,scoringandrubbinguse-wearonexperimentalochrepieces.J.Archaeol.Sci.37,3344 e 3358 . http://tracsymbols.eu ,20/01/2014. http://www.ncscolour.com ,2013. Igreja,M.,Porraz,G.,2013.FunctionalinsightsintotheinnovativeEarlyHowiesons PoorttechnologyatDiepkloofRockShelter(WesternCape,SouthAfrica). J.Archaeol.Sci.40,3475 e 3491. http://dx.doi.org/10.1016/j.jas.2013.02.026 . Jacobs,Z.,Roberts,R.G.,2007.Advancesinopticallystimulatedluminescencedating ofindividualgrainsofquartzfromarcheologicaldeposits.Evol.Anthropol.IssuesNewsRev.16,210 e 223 . Jacobs,Z.,Roberts,R.G.,Galbraith,R.F.,Deacon,H.J.,Grun,R.,Mackay,A.,Mitchell,P., Vogelsang,R.,Wadley,L.,2008.AgesfortheMiddleStoneAgeofSouthern Africa:implicationsforhumanbehavioranddispersal.Science322,733 e 735 . Jerardino,A.,Marean,C.W.,2010.Shell shgathering,marinepaleoecologyand modernhumanbehavior:perspectivesfromcavePP13B,PinnaclePoint,South Africa.J.Hum.Evol.59,412 e 424 . Kilburn,R.,Rippey,E.,1982.SeaShellsofSouthernAfrica.MacMillanSouthAfrica, Johannesburg .C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 302

PAGE 20

Klein,R.G.,1991.SizevariationintheCapedunemolerat( Bathyergussuillus )and LateQuaternaryclimaticchangeinthesouthwesternCapeProvince,South Africa.Quat.Res.36,243 e 256 . Klein,R.G.,Cruz-Uribe,K.,1984.TheAnalysisofAnimalBonesfromArchaeological Sites.UniversityofChicagoPress,Chicago . Klein,R.G.,Cruz-Uribe,K.,2000.MiddleandLaterStoneAgelargemammaland tortoiseremainsfromDieKeldersCave1,WesternCapeProvince,SouthAfrica. J.Hum.Evol.38,169 e 195 . Klein,R.G.,Steele,T.E.,2013.Archaeologicalshell shsizeandlaterhumanevolution inAfrica.Proc.Natl.Acad.Sci.U.S.A.110,10910 e 10915. http://dx.doi.org/ 10.1073/pnas.1304750110 . Langejans,G.H.,vanNiekerk,K.L.,Dusseldorp,G.L.,Thackeray,J.F.,2012.Middle StoneAgeshell shexploitation:potentialindicationsformasscollectingand resourceintensi cationatBlombosCaveandKlasiesRiver,SouthAfrica.Quat. Int.270,80 e 94. http://dx.doi.org/10.1016/j.quaint.2011.09.003 . Lombard,M.,Wadley,L.,Deacon,J.,Wurz,S.,Parsons,I.,Mohapi,M.,Swart,J., Mitchell,P.,2012.SouthAfricanandLesothoStoneAgesequenceupdated(I). S.Afr.Archaeol.Bull.67,120 e 144 . Low,A.B.,Rebelo,A.G.,1996.VegetationofSouthAfrica,LesothoandSwaziland. DepartmentofEnvironmentalAffairs&Tourism,Pretoria . Malan,J.,1990.TheStratigraphyandSedimentologyoftheBredasdorpGroup, SouthernCapeProvince.UniversityofCapeTown . Marker,M.E.,Craven,S.A.,2002.ThekarstoftheDeHoopNatureReserve,Western Cape,SouthAfrica.CaveKarstSci.29,51 e 56 . Marshall,F.,Pilgram,T.,1993.NISPvs.MNIinquanti cationofbody-partrepresentation.Am.Antiq.58,261 e 269 . Murray,A.S.,Olley,J.M.,2002.Precisionandaccuracyintheopticallystimulated luminescencedatingofsedimentaryquartz:astatusreview.Geochronometrica 21,1 e 16 . Murray,A.S.,Wintle,A.G.,2000.Luminescencedatingofquartzusinganimproved single-aliquotregenerative-doseprotocol.Radiat.Meas.32,57 e 73 . Olley,J.M.,DeDeckker,P.,Roberts,R.G.,Fi eld,L.K.,Yoshida,H.,Hancock,G.,2004. Opticaldatingofdeep-seasedimentsusingsinglegrainsofquartz:acomparisonwithradiocarbon.Sediment.Geol.169,175 e 189 . Parkington,J.,Fisher,J.W.,Kyriacou,K.,2013.Limpetgatheringstrategiesinthe LaterStoneAgealongtheCapeWestCoast,SouthAfrica.J.Isl.Coast.Archaeol. 8,91 e 107. http://dx.doi.org/10.1080/15564894.2012.756084 . Pickering,R.,Jacobs,Z.,Herries,A.I.,Karkanas,P.,Bar-Matthews,M.,Woodhead,J.D., Kappen,P.,Fisher,E.,Marean,C.W.,2013.Paleoanthropologicallysigni cant SouthAfricanseacavesdatedto1.1 e 1.0millionyearsusingacombinationof U e Pb,TT-OSLandpalaeomagnetism.Quat.Sci.Rev.65,39 e 52 .Porraz,G.,Parkington,J.E.,Rigaud,J.-P.,Miller,C.E.,Poggenpoel,C.,Tribolo,C., Archer,W.,Cartwright,C.R.,Charrié-Duhaut,A.,Dayet,L.,Igreja,M.,Mercier,N., Schmidt,P.,Verna,C.,Texier,P.-J.,2013a.TheMSAsequenceofDiepkloofand thehistoryofsouthernAfricanLatePleistocenepopulations.J.Archaeol.Sci.40, 3542 e 3552. http://dx.doi.org/10.1016/j.jas.2013.02.024 . Porraz,G.,Texier,P.-J.,Archer,W.,Piboule,M.,Rigaud,J.-P.,Tribolo,C.,2013b. TechnologicalsuccessionsintheMiddleStoneAgesequenceofDiepkloofRock Shelter,WesternCape,SouthAfrica.J.Archaeol.Sci.40,3376 e 3400. http:// dx.doi.org/10.1016/j.jas.2013.02.012 . Prescott,J.R.,Hutton,J.T.,1994.CosmicraycontributionstodoseratesforluminescenceandESRdating:largedepthsandlong-termtimevariations.Radiat. Meas.23,497 e 500 . Rector,A.L.,Reed,K.E.,2010.MiddleandlatePleistocenefaunasofPinnaclePoint andtheirpaleoecologicalimplications.J.Hum.Evol.59,340 e 357 . Rifkin,R.F.,2011.Assessingtheef cacyofredochreasaprehistorichidetanning ingredient.J.Afr.Archaeol.9,131 e 158 . Roberts,D.L.,2003.Age,GenesisandSigni canceofSouthAfricanCoastalBelt Silcretes.CouncilforGeoscience,Pretoria . Roberts,R.G.,Galbraith,R.F.,Olley,J.M.,Yoshida,H.,Laslett,G.M.,1999.Optical datingofsingleandmultiplegrainsofquartzfromJinmiumRockShelter, northernAustralia:partII,resultsandimplications.Archaeometry41,365 e 395. http://dx.doi.org/10. 1111/j.1475-4754.1999.tb00988.x . Roberts,R.G.,Galbraith,R.F.,Yoshida,H.,Laslett,G.M.,Olley,J.M.,2000.Distinguishingdosepopulationsinsedimentmixtures:atestofsingle-grainopticaldatingproceduresusingmixturesoflaboratory-dosedquartz.Radiat. Meas.32,459 e 465 . Roberts,D.L.,Botha,G.A.,Maud,R.R.,Pether,J.,2006.Coastalcenozoicdeposits.In: Johnson,M.R.,Anhaeusser,C.R.,Thomas,R.J.(Eds.),TheGeologyofSouthAfrica. CouncilforGeoscience,Pretoria,GeologicalSocietyofSouthAfrica,Johannesburg,pp.605 e 628 . Rogers,J.,1988.Stratigraphyandgeomorphologyofthreegenerationsofregressive sequencesintheBredasdorpGroup,SouthernCapeProvince,SouthAfrica.In: Dardis,G.F.,Moon,B.P.(Eds.),GeomorphologicalStudiesinSouthernAfrica.A.A. Balkema,Rotterdam,pp.407 e 433 . Scott,H.A.,Burgers,C.J.,1993.APreliminaryInventoryofSensitiveHabitatsinthe DeHoopNatureReserve.CapeNatureConservation,CapeTown . Sealy,J.,Galimberti,M.,2011.Shell shingandtheinterpretationofshell shsizesin theMiddleandLaterStoneAgesofSouthAfrica.In:Bicho,N.F.,Haws,J.A., Davis,L.G.(Eds.),TrekkingtheShore.Springer,NewYork,pp.405 e 419 . Skead,C.J.,1980.HistoricalmammalincidenceintheCapeProvince.Departmentof NatureandEnvironmentalConservationoftheProvincialAdministrationofthe CapeofGoodHope,CapeTown . Skinner,J.D.,Chimimba,C.T.,2005.TheMammalsoftheSouthernAfricanSubregion,thirded.CambridgeUniversityPress,Cambridge . Smith,M.A.,Prescott,J.R.,Head,M.J.,1997.Comparisonof14Candluminescence chronologiesatPuritjarrarockshelter,centralAustralia.Quat.Sci.Rev.16,299 e 320 . Soriano,S.,Villa,P.,Wadley,L.,2007.BladetechnologyandtoolformsintheMiddle StoneAgeofSouthAfrica:theHowiesonsPoortandpost-HowiesonsPoortat roseCottageCave.J.Archaeol.Sci.34,681 e 703 . Soriano,S.,Villa,P.,Wadley,L.,2009.Ochreforthetoolmaker:shapingtheStillBay pointsatSibudu(KwaZulu-Natal,SouthAfrica).J.Afr.Archaeol.7,41e 54 . Sponheimer,M.,Lee-Thorp,J.A.,DeRuiter,D.J.,Smith,J.M.,vanderMerwe,N.J., Reed,K.,Grant,C.,Ayliffe,L.K.,Robinson,T.F.,Heidelberger,C.,2003.Diets ofsouthernAfricanBovidae:stableisotopeevidence.J.Mammal.84,471 e 479 . Steele,T.E.,Klein,R.G.,2008.Intertidalshell shuseduringtheMiddleandLater StoneAgeofSouthAfrica.Archaeofauna17,63 e 76 . Steele,T.E.,Klein,R.G.,2013.TheMiddleandLaterStoneAgefaunalremainsfrom DiepkloofRockShelter,WesternCape,SouthAfrica.J.Archaeol.Sci.40,3453 e 3462 . Szpak,P.,2011.Fishbonechemistryandultrastructure:implicationsfortaphonomy andstableisotopeanalysis.J.Archaeol.Sci.38,3358 e 3372 . Texier,J.-P.,Porraz,G.,Parkington,J.,Rigaud,J.-P.,Poggenpoel,C.,Miller,C., Tribolo,C.,Cartwright,C.,Coudenneau,A.,Klein,R.,Steele,T.,Vernai,C.,2010. AHowiesonsPoorttraditionofengravingostricheggshellcontainersdatedto 60,000yearsagoatDiepkloofRockShelter,SouthAfrica.Proc.Natl.Acad.Sci.U. S.A.107(14),6180 e 6185 . Texier,P.-J.,Porraz,G.,Parkington,J.,Rigaud,J.-P.,Poggenpoel,C.,Tribolo,C.,2013. Thecontext,formandsigni canceoftheMSAengravedostricheggshell collectionfromDiepkloofRockShelter,WesternCape,SouthAfrica.J.Archaeol. Sci.40,3412 e 3431. http://dx.doi.org/10.1016/j.jas.2013.02.021 . Thackeray,J.F.,1988.MolluscanfaunafromKlasiesRiver,SouthAfrica.S.Afr. Archaeol.Bull.43,27 e 32 . Thompson,J.C.,Henshilwood,C.S.,2014.Tortoisetaphonomyandtortoisebutchery patternsatBlombosCave,SouthAfrica.J.Archaeol.Sci.41,214 e 229 . Tribolo,C.,Mercier,N.,Douville,E.,Joron,J.L.,Reyss,J.L.,Rufer,D.,Cantin,N., Lefrais,Y.,Miller,C.E.,Porraz,G.,Parkington,J.,Rigaud,J.P.,Texier,P.J.,2013.OSL andTLdatingoftheMiddleStoneAgesequenceatDiepkloofRockShelter (SouthAfrica):aclari cation.J.Archaeol.Sci.40,3401 e 3411. http://dx.doi.org/ 10.1016/j.jas.2012.12.001 . VanAndel,T.H.,1989.LatePleistocenesealevelsandthehumanexploitationofthe shoreandshelfofsouthernSouthAfrica.J.FieldArchaeol.16,133 e 154 . Villa,P.,Delagnes,A.,Wadley,L.,2005.AlateMiddleStoneAgeartifactassemblage fromSibudu(KwaZulu-Natal):comparisonswiththeEuropeanMiddlePaleolithic.J.Archaeol.Sci.32,399 e 422 . Villa,P.,Soriano,S.,Teyssandier,N.,Wurz,S.,2010.TheHowiesonsPoortandMSAIII atKlasiesRivermainsite,Cave1A.J.Archaeol.Sci.37,630 e 655 . Vogelsang,R.,Richter,J.,Jacobs,Z.,Eichhorn,B.,Linseele,V.,Roberts,R.,2010.New excavationsofMiddleStoneAgedepositsatApollo11rockshelter,Namibia: stratigraphy,archaeology,chronologyandpastenvironments.J.Afr.Archaeol.8, 185 e 210 . Vorster,C.J.,2003.Simpli edGeology:SouthAfrica,LesothoandSwaziland.Council forGeoscience,Pretoria . Wadley,L.,Hodgskiss,T.,Grant,M.,2009.Implicationsforcomplexcognitionfrom thehaftingoftoolswithcompoundadhesivesintheMiddleStoneAge,South Africa.Proc.Natl.Acad.Sci.U.S.A.106,9590 . Watts,I.,2002.OchreintheMiddleStoneAgeofSouthernAfrica:ritualiseddisplay orhidepreservative?S.Afr.Archaeol.Bull.57,64 e 74. Watts,I.,2009.Redochre,bodypaintingandlanguage:interpretingtheBlombos ochre.In:Botha,R.,Knight,C.(Eds.),TheCradleofLanguage.OxfordUniversity Press,Oxford,pp.62 e 92 . Watts,I.,2010.ThepigmentsfromPinnaclePointCave13B,WesternCape,South Africa.J.Hum.Evol.59,392 e 411. http://dx.doi.org/10.1016/j.jhevol.2010.07.006 . Willis,C.K.,Cowling,R.M.,Lombard,A.T.,1996.Patternsofendemisminthe limestone oraofSouthAfricanlowlandfynbos.Biodivers.Conserv.5,55 e 73 . Wintle,A.G.,Murray,A.S.,2006.Areviewofquartzopticallystimulatedluminescencecharacteristicsandtheirrelevanceinsingle-aliquotregenerationdating protocols.Radiat.Meas.41,369 e 391 . Wood,A.D.,1993.AspectsoftheBiologyandEcologyoftheSouthAfricanabalone HaliotismidaeLinnaeus,1758(Mollusca:Gastropoda)alongtheEasternCape andCiskeiCoast.Rhodes,p.161 . Wurz,S.,2000.TheMiddleStoneAgeatKlasiesRiver.UniversityofStellenbosch, SouthAfrica . Yssel,S.G.,1989.Long-termTrendsinaPopulationofTurbosarmaticus,Linnaeus 1758,intheTsitsikammaCoastalNationalPark.UniversityofPortElizabeth, p.132 .C.S.Henshilwoodetal./JournalofArchaeologicalScience45(2014)284 e 303 303


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
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.