<<

Zafar, WatershedofLatePre-IslamicCulture

Introduction,StateoftheResearch Newresearchandthedisseminationofthisinformationtoascholarlyaswellasabroader publicarerevolutionisingourunderstandingofthenatureoflatepre-Islamic Arabiansocie- ty.IntheintroductiontoourconferenceinFebruary2006,andinseveralnewpublications, ChristianRobinhasarticulatedthatuntilrecentlytheHimyarite Agehasbeenlittleunder- stood;itslatepre-Islamicinhabitantsbeingpurportedlypoor,isolated,illiterate,lackinga stablepoliticalsystem,andlivingasnomadsinthedesert.Recentresearch,however,con- tradictsthisimage.Sincethe1970s,thetempoofresearchhasacceleratedand,withnew textualandinterpretationalstudies,onecanspeakofageneralreinterpretation,especially regardingHimyar.Isonetoexplainthedramaticandfar-reachingpost-Himyaritereligious, cultural,andeconomicdevelopmentsoflate6th early7 th century Arabiaasissuingfull-grown fromahistoricalvacuum?Dodecisiveculturalinfluencesin Arabiaatthistime,then,derive fromabroad,Himyar,andotherindigenousformativebodiesthatplayednorealroleorjust anegativeone? Today,noscholarwouldseriouslyconsiderIslamtohavegainedwidespread acceptancesolelywiththediplomaticandmilitaryeffortsoftheprophetMuhammedinhis generation.Otherfactorsplayedarole,butwhich?1 Followingcommentsonthestateofresearchoflatepre-Islamic ,thereaderwillfind abriefcharacterisationofhighlightsofarchaeologicalfieldworksince1998inZafar,capital ofancientHimyar. A shortoverviewoftheresearchonthislateOldSouth Arabian(OSA) centrecorrespondswiththegoalsofourconference.SincetheexcavationreportsonZafar arestillinpress(Yuleinpress),newexcavationresultsfromthesitearenotcommonly known. Themaintopicsofthepaperincludethenatureofthedefences,theappearanceof Zafar'scityscape,anditspost-Himyaritehistory. Thefinalsectiondiscussesbrieflyand fromanarchaeologicalpointofviewselectaspectsoflatepre-Islamicinthehigh- lands.

FewsitesyieldremainsoftheHimyariteperiod,andevenfeweroftheselieintheHimyar- iteheartland.Qaryatal-FawoftheKinda(al-Ansary1981)andNagran(Zarinsetal. 1983), keysitesthat haveyieldedsignificantrelevantcontextsandfinds,lietothenorthandout- sideoftheHimyaritehomelandperse .Ontheotherhand,ZafarandQanilieinsidethe generallimitsofthekingdom,thelatteraspoilthatfelltoHimyarafteritsconquestofthe kingdomofHadramawt. ThesettlementofthelowertownofQaniencompassesanareaof 100x250m.Inaddition,poisedonanextinctvolcano,the“fortofthecrow”measures some150x150m(allinall5ha).Inthe1st centuryCEQaniwasausefulprovisionssta-, tionbeforecontinuingtosailtotheeastorwest. Theexcavator, A.Sedov,describesthe presenceofalighthouse,storehouses,a“customshouse”aswellasasynagogueandatem- plededicatedtoalocaldeity(1998:276).Fromthe2to5nd th centuriesQani'stradeblos-, somed,judgingbyitsimportsfromNorth .Inthe6th and7 th centuriesonlythenorth-, westernpartofthetownwasinhabited(1998:278). Thefindingof80%ofamphorasherds, whichwerethebest-knownimportsdiscoveredintheexcavationsat Ayla/Aqaba(southern 1TheresearchbelowwasconductedmainlyduringaperiodofmunificentgrantsfromtheFritz ThyssenStif- tung(20042005)andtheDFG(2006). TheGeneralOrganisationfor AntiquitiesandMuseums(GOAM)gen- erouslysupportedourannualcampaignsinthe Yemen. Thepresentpaperisanabbreviatedversionoftheoral presentation.K.GalorcontributedtothesectionontheJewishpresenceinZafar. Wearesubmittingthispaper priortoreceivingtheresultsofseveralradiocarbonsamplesthatmayalterthechronologyinZafar. Weare gratefultoChristianRobinfortheinvitationtodeliverthepaperandfortheopportunitytodiscussseveralof thepointsraisedtherein.Papersoftheconference, Vieillesthèsesetnouveauxdocumentslejudaïsmeen Arabie,desoriginesàl'aubedel'Islam,inprintin Arabiarevuedesabéologie. Diacriticsnotpossible. 1 Jordan),informusthattradestillflourishedatthistime.Inthe7th century,Qanihadlostits importanceanddisappearedfromhistoricalview.

Atthesametime,Zafar,withitssubstantialruinswhichareattributablemostlytotheem- pire(c .270-523)andlate/postperiod(523-c .560/570),fillinthishistoricalpicture.Excav- ation,find-recordingandmappingatZafarsince1998haveshedconsiderablelightonthe natureofthecapitalandmaterialcultureoftheHimyar.Itisparadoxicalthatthecapitalof akingdom,whichdominateddirectlyandindirectlyanenormousareaencompassingmost ofpre-Islamic Arabiaforsome250years,untilrecentlyhasremainedamarginaltopicin Arabianstudies. Thisisallthemoresosinceearly Arabhistoriansandotherearlymedieval sourcesknewtheHimyarbetterthanthelong-vanishedkingdomsthatprecededthem.Most historianshaveshort-circuiteddirectlythisconnectionof Arabianhistory,preferringtohark backtotheageofthebiblicalandkoranicQueenofandcircumventthe“decadent” Himyar.

Excellentsyntheticarchaeologicalandhistoricalstudies(e.g. Vienna1998)haverevitalised theHimyarites.Forthoseofuswithoutextensivelibraryresources,C.Robin'sassembled relevantOSA textsinthepresentpublication(2006)arehelpful.Priortothis,numerous scatteredpublicationsdiscouragedallbutthemostdeterminedtowork,exceptsuperficially, inafewselectlibrariesonthe“lateperiod.” Asidefromtextualstudiesontheearlyhigh- lands,twosurveystookplace-onebroadlybasedinthearea(Wilkinsonetal. 1997)anda secondcentringontheagriculturalhistoryinandaroundZafar(Barcelóetal. 2000;2003). ThefirstcoversfromearliesttoHimyaritetimesandthesecondcoversessentiallythelatter periodtothepresentday.

Whatiswrongwiththeusualcharacterisationoflatepre-Islamic Yemen?Certainconserva- tivecolleaguesdenigratethisperiodatsomelength,contrasting,forexample,the“good,” linear,earlyartwiththe“bad,”organic,laterart.J.Schmidtreferstothe“sinnentleerte Bastelei”(1997:34-37)andS. Antoninitothe“late”or“decadent1st -4 th century”(herquota- tionmarks;2001:23). Atfirstglance,thishistoricalinterpretationisappealinginlightof thebiologicalorDarwinianmodelofthebud,flower,andwithering.Butnewfindscontra- dictthisconservative,ifnotold-fashioned,schematicmethod,whichinrealityfunctions predictablyinfewcategoriesofancientartandresultsinaself-fulfillingprophecy. There- by,artefactsofmediocrequalitycanbeassignedtothelateperiodignoringotherdatingat- tributes. WehaveherethepleasantopportunitytorehabilitatetheHimyarandtheirculture. ThedenigrationofHimyariteartresemblesoutofhandanalogouslythecastigationofLate Antiqueartasdecadent,discordantwithitsmodernperception.

AttheheartofmatterliesthesiteofZafar(14°13'N;44°24'E,125kmSSEofSan'aand 230kmNNW of ),wherecreativeculturalimpulsesaretobeexpectedowingtoitsas- sociationwiththeroyalhouse,itspatronageandasurplusofwealth.Zafarisnestledinthe chainofflattishcratersformedbyatleastfourirregularlyshapedextinctvolcanos. These riseabovethesurroundingrockylandscapetoanaltitudeofsome2800m. Thethreemain findareasoftheancientcityincludeZafarSouth(thepresent-dayvillage),theHusnRay- dan,andRaydanNorth(al-Gusrinthelocaldialect).Himyariteruinsareconcentratedmost heavilyonthesouthernandwesternslopesoftheHusnRaydan(Figs.1-3),particularly onabroadslopeknownasal-Jahw.

2 PrimaFacie AppearanceofHimyariteZafar Mappingoperationsilluminatethesize,shape,andlayofZafar. Themainruinsformaclus- terirregularinforminsidearectanglemeasuringsome800x1000m(E-W byN-S).One wouldexpectthemainpalacetostandatoptheHusnRaydan,themainlocalpromontory. In1998,afirsttesttrenchinsidea12x12mstructure(buildingz028)showedthatfurther workinthisprofoundlydisturbedareapresentedadubiousexcavationopportunityforthe investmentofprecious,limitedresourcesoftimeandmoney.Someyearslater, Abdullah Salihal-AnnabifromZafarinformedusthatthiswasthesiteofasub-recentsettlementthat wasallegedlylevelledbythe“imam”manyyears,evencenturies,agobecausethevillagers failedtopaytheirtaxes.Giventhevaguenessofthisstory,the“imam”easilycouldhave beenalocalgovernorofthefirst Turkishoccupation(945-1045/1538-1636),whowas knownforhisgreedandharshmethods(Smith2002:273b).Conditionsinthe Yemenand elsewheredeclinedgenerally,includingthepopulationdensity.Poorsitepreservationon theHusnRaydanprobablyresultfromtheaforementionedincidentinadditiontoextensive subsequentstonerobbing.

Fig.1MainruinareasofZafar.

3 Fig.2 Overviewofthelateantiqueperioddefencesandroads/waysatZafar.

Fig.3PlanofZafar/al-Jahwonthesouth-westernflankoftheHusnRaydan,condition:March2006.

4 BuildingruinsoftwokindsoccuratZafar:thoseobviouslyHimyarite,andrecent(<70 yearsold)stonehousesandtheirruins,thatarebuiltfromHimyaritespolia.Himyarite masonryoftheempireandlateperiodsischaracteristicallymadeofstone;itstimber-frame, wattle,anddaubconstructionhavenotsurvivedinZafar. Thismannerofbuildingisbest knownfromother,slightlyearliersites,particularly(cf.Breton1991:217fig.5). Accordingtothesizeandpurposeofthebuilding,Himyaritebuildingstonemayvaryin sizeandshape,butareregularlymasoned. Withinagivenbuildingstructuretheyarerela- tivelyregularinsizeandoverallappearance(Fig.6 ). Terracewallsbuiltinthistechnique belongtostructurescomparabletointegralpartsofthemainHimyariteperiodfortifica- tions.Clusteredonthesouth-westernedgeofZafar,afewfoundationruinsformoneofthe fewexceptions:Relativelyisolatedandirregularhouseplansshowlarger,somewhatcoar- serstonesthanthetypicalHimyaritehouses.

InZafarvillagethepopulationbegantoregeneratearound1950andhardlyabuildingthere predatesthis(Fig.4 ).Reportedlysome300yearsagowhateverinhabitantshadsurvived builtasmall(interiorsize3.7x3.4m)cubemosque(featureno.z070)ontopofwhatthe localpopulationconsideredtohavebeenanancientchurch(Fig.5 ,seebelow).Unfortun- ately,however,wehavelittlemeanstocorroboratethisdating. A phototakenofal-Jahwin 1970(Müller1979:cover)showsonlythecubemosque,z070,andasimplestoneshedover thelargeHimyaritetomb,z066.Inpastdecades,inhabitantsreturnedtoZafarfromtheout- lyingvillagestheyhadmovedtowhensubsistenceinZafarhadbecomeuntenable. Asking thelocalswhentheirhouseswerebuiltelicitsmostlytheanswerthatthishappened40to 50yearsago.However,theoldest“commonhouse”inthecentreofthevillageissaidtobe some300yearsold,perhapscontemporaneouswiththeoldmosque.Mostoftheinhabitants havesomeideaoftheageoftheirownhouse,butnotaclearoneaboutthoseoftheirneigh- bours. Thus,theareaaroundthecubemosque,knownasal-Jahw,was,infact,builtup mostlyinthelast30years,althoughyoungenthusiastsunwittinglyclaimafarhigherdate over100yearsold.Duringtheconstructiononthisspot,thevillagersexcavatedHimyarite ruins,whicharenowusedasstoragefacilities,particularlythetomb,z066(Yule2005:25 Fig.4).NicelytrimmedHimyaritebuildingstonesareamajorsourceofbuildingmaterial, notonlyforhousesinZafarbutalsofortheentireregionasfarawayas ,forexample, forthemainmosquethere. Transportbymeansofcamelanddonkeywaseasierthanactual stone-working.Inthisregard,itisdifficulttoenforcethepreservationlawsinZafar,asis alsothecaseinothercountries.SeveralHimyaritecontextsthatwerecordedarenearlyor completelydestroyed;manyofourphotosarealreadyhistoric.

Zafar/al-Jahw, Trenchz300 Intheyear2000(Yuleetal. 2007),wetestedtwoHimyaritebuildingsonthesouthern (buildingz178,Fig.5 )andeastern(z175)slopesoftheHusnRaydanbeforedecidingto partiallyinvestigatetheextensiveHimyaritecemeterysiteonthenorthernslopeofJabal al-Asabi(zc01). Theeasternextensionofbuildingz175haserodedaway(Fig.6 ).In2002, excavationcontinuedonthebuildingfoundationofz178,astructuremeasuring12x12m.

Subsequentlyin2003,work(trenchz300)commencedonthewesternslopeoftheHusn Raydan,partofwhichliesinal-Jahw. Thisandoursubsequentexcavationendeavourof 2004and2006,trenchz400,wereundertakeninthehopeoffindingremainsofamonumen- talbuildinginarchaeologicallymorepromisingareas,tojudgefromthesurfacefindsthere, mostlyrelieffragments.Some20meastofstructurez300(Fig.7 )thelargeanddated5th centurybuildinginscriptionofShurahbiil Yafur(siglum: zm1)cametolightaround1968 (UTMcoordinates:38P,e35541,n71533),inwhichherecountstherenovationoftheold4 15 5 Hargabpalace. Thisimportantfindgaveimpetustolookintheimmediatevicinity,evenif thefind-spotoftheinscriptionissecondary.Particularlyinvitingisthekingsdescriptionof decoratingthepalacewithnewfloralornamentinadditiontotheoldornament(Simain press). Theheavyoverburdenofstonedebrisandearth,whichprotectedthesetwocontig- uoussitessomewhatfromthedepredationsofquarrying,alertedustothepossibilityofbet- terpreservationhere. The20x10mtrenchz300containscubiclesbuiltofhewnstone blockswhichareassociatedstratigraphicallywithabroadstonepavement. Themasonryof thebuildingsisrustic. Whilethepavementslabsaresomewhatirregularinshape,theyfit sotightlytogetherthatonecannotputaknifebladebetweenthem. Theidentificationas magazinesisstrengthenedbythepresenceoflargestoragevesselsherds,howeverthese mayhavefallenintoz300fromthesettlementaboveintheHusnRaydan. A provisionaldat- ingforz300intheempireperiodisbasedonthequalityofthemasonryandthesurrounding contexts.

Fig.5Zafar/al-Jahwtowardthenorth,2005.

Zafar/al-Jahw, Trenchz400 Southofz300andadjacenttoit,atthetimeofwritingthis20x10mtrenchalsocutintothe western-facingslope.In2004and2005,excavationshererevealedabuildingbuiltoffinely cutstonewhichwaslaterintrudeduponbyacurvingfoundationwallfashionedoutof roughlybrokenstone(Fig.8-9 ). Thedateofthiswallispresumablylatepre-Islamic. The planinFig.8 showsthenorthernandwesternwallsofthisstonebuilding,bothofwhichap- peartohaveanentrance. Althoughtheeasternandsouthernsidesofthestonebuildingare yettobeunearthed,thestructureseemstobeacourtpavedwiththickstoneslabs. Atthe southernend,aheavyslagdeposit,whichyieldedradiocarbonsamples,coverstheStone Building,postdatingit.

6 Thefindcircumstancesofthispartlyexcavatedcourtseemedatfirsttoindicateafierydes- truction,butitismorelikelythatthelargeslagdepositderivesfromoneormorekilnsused aftertheStoneBuildinghadfallenoutofuse.Itcontainsaconsiderableamountofcharcoal, reminiscentofkilndebris.

Fig.5Zafar/al-Jahw,sketchplanofthebuildingfoundation,z178.

Fig.6Zafar/HusnRaydan,detailoffragmentarybuildingz175showingtypicalHimyaritedomesticmasonry.

7 RegardingthedateoftheStoneBuilding,first,aninscription(inRobin2006:1.A.b.1= GorgeduHaut-Bura'2)mentionstheHargabpalaceinZafararound (351CE).Giventhe proximityoftheabove-mentionedre-buildinginscriptionofShurahbiil Yafur(themonth dh-'I572Him/September457CE)describingtherenovationandredecorationofthepalacen (Simainpress),anidentificationwiththisbuildingseemedpossible,nowitisimpossible Moreover,carbon14datingprovidesa terminuspostquem fortheStoneBuildingintrench z400(ingreaterdetail, Yuleetal. inpress).Evenifsomeofthecarbonsampleslayonthe floor,theyneednotandseemnottobecontemporarywithit. Theavailabledeterminations clusterinthe4th century.Furtherassaysfromfavourablylocatedfind-spotsonandunder thefloormayshedmorepenetratinglightonthechronologyofthisattractivestructure.

Lab.Number14CAgeSTDCalAgep(68%)CalAgep(68%) [BP][calBC/AD][calBP(0=AD1950)] KIA268001736±23-290±401660±40 Hd-238781780±18-250±401700±40 Hd-238811680±28-340±401610±40 Hd-238831752±24-280±401670±40

Lab.number14CAgeSTDCalAgep(95%)CalAgep(95%) [BP][calBC/AD][calBP(0=AD1950)] KIA268001736±23210-370calAD1740-1580calBP Hd-238781780±18170-330calAD1780-1620calBP Hd-238811680±28260-420calAD1690-1530calBP Hd-238831752±24200-360calAD1750-1590calBP

Calibrated AMSandCdeterminationsfromz40014 .

A thirdmeansofdatingtheStoneBuildingisthemarginallydraftedpeckedmasonrywhich decorateit(Fig.10 ). ThisdecorationofG. W. VanBeeks Type6(1958:287295,esp.291b and295b)smoothedface;blocksandcoursessimulatedbyincisedlinesdefiningborders andpeckedareasoneachpanel;sparseandobliquepeckingisknown,accordingto Van Beek,fromaperiodbetweenthefirstcenturyCEandthe“lastPre-Islamicperiod,”presum- ably630CE.Fewcomparisons,datedorundated,areavailableforthisperiod,butsimilar- lydecoratedmasonryoccurredinthe1999-2000Italianexcavationsat Tamnaincontexts datingfromc. the3rd centuryBCEtothemid1st centuryCE(personalcommunication,S. Antonini,19May2006).Furtherexampleshavecometolightatthatdatefrom the4th to1 st centuriesBCE.SeveralothersuchdecoratedstonesexistatHuqqa,al-Sawda andal-Bayda,notallofwhicharepublished(Rathjensandv. Wissmann1932:49Fig.15-16, cf.alsoRadt1973: Taf.11,34a:“wohl ”).Parallelsforthiskindofstone-workingare foundinthelaterphasesoftheBarantempleinMarib(personalcommunicationB. Vogt). Thesefindssomewhatwiden VanBeek'sdatingforhis Type6,astheyderivefromcontexts whicharenotcloselydatable.Basedonthefinequalityandworkmanshipofthemasonry, theexcavatorsuggestsanoriginfortheStoneBuildingintrenchz400inthe1st or2cen- nd turyCE.Sincelittleiscurrentlyknownabouttheworkmanshipoflaterperiodsorofthe post-empire,thisconclusionmayneedtoberevisedinthefuture.

8 Fig.7Zafar/al-Jahw,isometricviewofbuildingintrenchz300.

Fig.8Zafar/al-Jahw,planoftrenchz400withstonebuildingand“steps”.

9 Fig.9Zafar/al-Jahw,stonebuildingduringexcavation,2006,towardthenorth-east..

Fig.10Marginallydraftedpeckedmasonryfromentrancez415ofstonebuilding;28x29x23.5cm.

ZafarsHimyariteDefences Althoughlessthan1%ofthecitywallsstillstand(ascertainedbymeansofsurveyandsate- lliteimagery),theoutermostcitydefencesareestimatedtohaveoriginallymeasuredsome 4.5kmincircumference. Thedefencesbynomeansconsistsolelyofmasonedwalls,but alsoofcliffsshearedofftogetherwiththem,trenches,andthenaturaltopography. Theinner defensiveringiseasiertoidentifythanthefragmentaryouterone(Fig.2 ). Whatremainsof thefortificationseludesprecisedating.Heavy,well-masonedandpatinatedcutstonewalls aregenerallyconsideredtobeHimyarite;thismeansofdatingstillhaslittleconfirmation

10 Fig.11Zafar/al-Shugagsadd z326towardthenorth-east. Anancientwellislocatedontheleft(north)sideof thedam.

Fig.12Zafar/Wadial-Haf,showingthefieldnames. bymeansofabsolutedates.Itisimpossibletodrawmorethanasketchyplanoftheentry zone,evenifitsroughoutlinecanbemadeout(Fig.2 ). Thuswedonotyetknowhowwide andhighthemainwallswere,andwemustestimatetheintervalsbetweentowers,fewof whichhavesurvived.Moreover,inthepost-Himyariteperiod,afterthefortificationshadfa- llenoutofuse,somewerefoundre-usedintheconstructionofthesadd (dam/retaining wall)inal-Uwar(Yuleetal. 2007).

Some30yearsago,abuildinginscriptioncametolightinZafar/al-Uwar(zm2263+2262+ 2264/2),presumablyoriginallysetintothecitywallorgateitdescribes. A.Simadeter- minedthatpartofthemmayhavebeenstrengthenedbymeansofadefensivetrench(2002). Intramuros,theHusnRaydanhaditsowndefensivewalls,themainonesofwhicharestill

11 visible. ThemainpreserveddefencesliesouthofZafarSouth,inal-JabubatLaqiyahandal- Hayfah.Certainanchoringpointsexistfortheirdating: Theaforementionedbuildingin- scriptionsfromal-Hayfahrecordarebuildingofthecitywallin347Himyar/232CE(a- ssumingabeginningfortheHimyaritecalendarin110BCE).Barcelóreferstothis“ep- igraphictablet”asdatingtothemidfourthcenturyCE(2003:141);thisdatingforthede- fencescanthusbecorrectedhere.

TheCityscape ThecoreofHimyariteZafarspreadsoverthesouthernandwesternslopesandthesummit oftheHusnRaydan,thefortifiedal-Gusr,andthepresent-dayvillage. Theancientcityex- tendseastwardandsouthwardoutsidethecitywalls. TotheeastoftheHusnRaydan,heavy accumulationsofstonedebrisarevisiblearoundthepresent-dayfieldswherehousesonce stood.ButtothesouthofQaryatZafar,aHimyaritedomesticquarterliesintactunderade- brislayeralongtheroadtoneighbouringBaital-Ashwal,especiallyinal-Hayfahandal-Jir- ajir. Withafewexceptions,littleevidenceexistsforsettlementtothewestandnorthofal- Gusr,andinalllikelihoodtherewaslittletherefromtheoutset. Thereisnolackofevidence forancientbuildingontopofal-Gusr,wherethefoundationwallsofanover100mlong structurecametolight.OntopoftheHusnRaydanbuildingremainsarescrappyanddiffi- culttodateowingtoalackofsurfacepottery.Numerousruinsinandaroundthepresent -dayvillageareofclearHimyaritedate,themostnotableconcentrationlyingjusttothe southintramuros onJabubatal-Laqiyah.InFebruary2004,whilewideningtheaccessto thevillage,Zafar'smainancientpavedentrance(z190)inthesouthwasextirpated. A little furtherintoal-Hayfah,bulldozingin2000and2005allbutobliteratedthemajorentrance tothecity.

Atpresent,fivemaincitygates(Fig.2 )canbelocalisedbymeansoftheirassociatedroads/ waysbutmoreprobablyexisted,judgingfromthesizeofthecityandal-Hamdani'sdescrip- tionofninegates.Noneofthemarewell-preserved.Fromtheheavilypreservedremains,it maybesurmisedthatthemaingateliesatthesouthernendofthepresent-dayvillage. At theoppositeend,tothenorth,liesasecondgate,attheeasternextremityofRaydanNorth. Atmidlatitudetotheeastandwest,thelayoftheruinsandpresent-dayroads/wayscon- formtotheancienttopographyandsuggesthereflankingentrances.Certainlytherewere gatesinthemiddlepartoftheirregularlyshaped1200m-longnorth-south-orientedsettle- ment. Thegateonthewesternsideliesadjacenttobabsuqal-layl, (gateofthemarketof thenight),whichhintsatthepresenceofagatewhereoneisexpectedforarchaeological reasons. A pathleadspastlargenumbersofrock-cuttombsdirectlytothemainpartofthe corearea,al-Jahw. A fifthgate,stillvisible,liesinthesaddlebetweentheHusnRaydanand al-Gusrandisvisiblyaccessibletothewest,butpresumablyalsototheeastjudgingfrom thelayoftheruinsandothercharacteristictraits.Usingtheareaofthesiteasabasisforcal- culation,J.Schiettecatte,inafinestudyonthedevelopmentofthe“ville ”(hisquotations marks),calculatesapopulationbetween50,600and75,900forZafar(2004:141, Table2, 10to15inhabitantsperstructure),makingitthemostpopulouscityin.But giventheunevensettlementdensityofthissite(themappedareais110ha),thenumberof dwellingsandpopulationwouldbelessthanthis.

ContinuityofPost-HimyariteSettlement Zafarisbotharewardingandfrustratingarchaeologicalsiteforthechronologist.Onthe onehand,stonerelieffragmentsarenumerousinourtrenchesatthesite,byvolumefive timesmorethanthepotterysherds.Particularlythereliefsraisehopesoffindingtheintact datablebuildingstructureswhichwereoncedecoratedbytheirwell-to-dopatrons. 12 Thechronologiesforthereliefsandpotteryoftheempireandlateperiodsarestillgenerally undeveloped(cf. Wilkinson,EdensandGibson1997:128-129,fig.19)andtheirpotteryse- quencerestsalmostentirelyonunstratifiedsurfacecontexts.Numeroushithertounknown andattractivesculpturesoftenderiveunfortunatelyfromdebris;furtherexcavationisneces- saryinordertolocatemoreusefulstructural/historicalcontexts. A few,however,doreveal theageorhistoricalcontextofthefindandhelptodatetheconsiderableholdingsofthesite museum. Whilenotemergingfromcontrolledexcavations,reliefsfromthesitemuseum, withbutfewexceptions,comefromtheimmediatearea,whichgivesanimpressionofthe artofthehighlands. Thequantityandartisticqualityofthenewartefactsinthesitemuseum placesheavyobligationsontheexpeditiontomakesomehistorical/chronologicalsenseof thisrichmaterial.

AsidefromtheStoneBuildingintrenchz400,whichisthemostintactsuchbuildingonsite (Fig.8-9),fewfindsandcontextsinZafardatetotheearlyHimyariteperiod.Inthecollection ofthesitemuseum,however,onemusnad inscriptionwrittenboustrophedonisclearlyearly (Simainpress).SeveralcolumncapitalsinHimyaritecontextsalsoappeartobe spolia lootedfromotherearlysites.Ontheotherhand,theempireandlateperiodsareknown fromthepotteryandfinds,andarestillunderstudy(seebelow). Thefirstradiocarbonassays alreadygivefixedpointsforimportantpiecesfromtheexcavationandfromthemuseumcol- lection.

Theevidenceforpost-HimyaritehabitationatZafar,atthelateendofthetimeline,iscontra- dictory.M.Barcelóindicatesacontinuityinthesettlementfromthe2centuryCEtothend presentdaybyvirtueoftheoccurrenceofarchaicplace-namesfordamsandtheirassoci- atedterracefieldsinal-Hamdani's(AH334/CE945)al-Iklil ,manyofwhichhereportedly hasidentified(Barcelóetal. 2000;2003:140-141).Hebaseshisproposedsettlementcon- tinuityonlong-livedplace-namesandcertainfinds.Paradoxicallyhowever,ourworkhas notyetyieldedmuchevidencefrompotteryandotherfindsevenpossiblydatabletotheIs- lamic Ageoranypartofit.

InscriptionsshedsomelightonthelatehistoryofZafar.In658Himyar/548CE,the Axum- iteruler wroteonhislarge 'Resgestae' steleinMaribthatayearearlierhesup- pressed,bymeansofthousandsoftroopswhatsoundslikeawidespreadinsurrectionjudg- ingfromthenumberofopposingtribeshenames. Whether Abrahachosetomovethecapi- talasaresultofthestampingoutoftherulerSumuyafa Ashwa'srevoltin535orbecause oftheinsurrectionof547isunclear.Noristheresubstantialevidenceforapost-Himyarite settlementatZafar.Butahintofthedevelopmentscomesfrom Abraha'sbuildingofthe cathedralinSanaac. 560/565CE,whichwasZafar'sswansongasacapital. Abraha'slong- knownintentiontomakeSana'aapilgrimagecentrewouldhavebeenarealstepforward forthecountry.Inanycase,withfewexceptions,theonlyfirmevidenceforthepost-560/ 565CEhabitationofZafarisfoundinthedescriptionsofal-Hamdani.

OnesiteatZafarwhichmayshedlightonthepost-Himyariteperiodistheaforementioned rollingsettlementruinmainlyinal-Jirajir(extramuros tothesouth),whichistransectedby theroadtoBaital-Ashwaltothesouth. Thissettlementseemslessdisturbedarchaeologi- callythanZafar'scentre. ThepotteryfindsthereappeartobeofHimyaritedatebutrequire furtherstudy.

13 Fig.13Sealringof YishaqbarHanina,from Zafar/al-Asabi/al-Salm;longestdimension Fig.14Zafar/al-Jahw,steppedentrancetotherock- c. 1.2cm. cutchamber(z071),viewtothenorth-east.

Fig.15 Above:plan;below:x-rayviewofthecubical Fig.16Zafar/al-Jahw,viewlookingdownintothe mosque(z070)androck-cutchamber(z071). cisternz096tothesouth-west

14 In2006,ourresearchgainedmomentuminanefforttoilluminatefossilplace-namesaround Zafarwhicharepossiblypreservedinthepresent-dayarchaiclocaldialect(Behnstedt 2002). Asyet,wehaveidentifiedbutafewsurvivingexamplesfromHimyaritetimestothe presentday(e.g. JabalHadaman). WedocumentedaHimyaritesadd/jirab combination (damandfield)inZafar/al-Shugag,whichismentionedinal-Iklil . Thisservedasatraining groundforourmulti-disciplinaryteamoftwosurveyengineers,anarchaeologist,anda semiticist. Theover1500yearsoldsadd correspondstowhatisconsideredtobetypical Himyaritemasonry(Fig.11 ):heavy,regularlycutandlaidstoneswhichhavebecomepat- inatedblackoverthecenturies.However,thechronologyfortheentireassociatedcomplex ofdamsandfieldsisproblematicasaresultofmultipleprovenre-buildingsoftheasdad or damsinthewaditothepresentday(Yuleetal. inpress1).Elderlyvillagerstellusthatcer- tainwalls,whichatfirstglancewetooktobeancient,infactarenewandarebuiltre-using ancientblocks.Surfacefinds,suchaspotsherdsinthedrainagebasin,areindeedrareandof nouseindating. Althoughseveralfieldnames(Fig.12 )arearchaic(e.g. dhial-Jadhub), suchformsmaycarryoverintothemainstreamofClassical ,whichistraditionally consideredtohavebegunc. 500CEandarenotlimitableindate.3

ThemainproblemforthedeterminationofhabitationsurvivalatZafaristhatintheabsence ofdefinablearchaeologicalcontextsand/orcarbon14determinationsfromtheearlyIslamic ageafterseveralseasonsofexcavationonlyasinglecontext(afortification/terracewallin al-Uwar)mightconceivablycloselyfollowthefallofHimyar(Yuleetal. inpress).4 Thus, ifthefieldsaroundZafarwereincontinuoususeoverthecenturies,whyistherenoevi- dencethereforsettlementinthefinaldeaththroesofthealreadydefunctHimyariteconfed- eracyinthemidlate6th century? The BookoftheHimyarites citesonepassageinwhichdepartsfromhis“capital”intoward.Otherevidencesuggeststhe moveoftheHimyaritecapitalfromZafartoSanaaaround560CE,whichjibeswith Abra- ha’sbuildingofthecathedralthere.CouldZafarsimplyhaveceasedtoexist?Ifso,onecan hardlyexpectroyalpatronageanddevelopmentofthesite(Lafitte2003:79).

Despiteal-Hamdanisearlymentionofseveralasdad inthevicinityofZafar,andthesurvi- valofplace-namestothepresentday,thissourceinitselfshedslittlelightonthecontinuity ofpost-Himyarsettlement. Although,thegeneralimpressiononegetsfromthedifferent sitesisadropinthepopulationdensityofthehighlandsafterthefalloftheHimyar,how drasticwasitreally? A smallpopulationwouldsufficetopreservetheplace-namestothe presentday.Surfacesurveyrevealslittleevidenceinthisregard.Futureworkcentresonlo- catingintactcontextsoftheempireandlatephases.

JewishPresenceinZafar: An ArchaeologicalPerspective C.Robinsrecentstudy(2003)onJewishpresenceinZafarenjoinsustosystematically evaluatetheexistingepigraphicandliteraryevidence. Thus,aroundtheyear380,mono- theismingeneralcomesstronglytothefore.Under Tharan Yuhanim(c.324-c.375),the Jewishcommunityinthe Yemenislargeandinfluential(2003:154),howevertheevents leadinguptothisblossomingremainobscure. WhiletheexcavationsandstrayfindsatZa- farshedsomelightonthismatter,overtlyJewishattributesarerareinthearchaeological

3 Certain Arabicinscriptionsarecenturiesolderthanthis,e.g. fromQaryatal-Faw,al-Ansary1981:146,Pl.1 =63(fromthetombof'AjlbinHaf'am). 4 A groupofheavyfoundationsofunusualformonthesouth-westernslopeoftheZafarvillageisnotdatable andisapotentialcandidateformedievalsettlement.

15 record.Moreover,severalpossibleindicatorsofJewishpresencecanbecited,although theirvalidityisoftenuncertainandquestionable. ThetombsandgravesexcavatedonJabal al-Asabi(Yuleetal. inpress1)belongpartlytothelateperiodbutoffernofirmevidence regardingthebeliefsoftheinterred.Norareburialandgravecustoms(e.g. grave/tomb forms)comparablewiththoseofwell-documentedoldJewishpopulations.Forexample, noexplicitJewishsymbolshaveasyetbeenidentified. Yet,inscriptionssuchasthatof Yehudah Yakkaf(siglum: Baital-Ashwal1)fromthefirsthalfofthe5th centurypossesssuf- ficientinformationforustolearnofhisconfessionevenwithouttheHebrewinscription writteninsidethemonogram.ButwhendidthefirstJewsarriveinthearea,andwhicharte- factswereproducedforJewsorreflectJewishcontent?

In2002, Ali Abdullahal-ZafarifromZafarshowedP. Yulearingthatwaspurportedly “found”afewyearsearlierinthesameplaceasourexcavationatal-Salm,onthenorthern faceofJabalal-Asabi. Theintaglioisanopaqueredcarnelian,flatinsectionandovoid en face,whichisdominatedbya TorahShrine. Thename YishaqbarHaninaappearsen- gravedinmirrorimage,andalthoughthepersonalname YishaqappearsinbiblicalHebrew, Haninadoesnot.Inhispublicationoftheringdevice,G. W.Nebenotesthatthenamewas knownbetween330BCEand200CE(Ilan2002:103-105)butalsolater,inthesynagogue inscriptionatel-Hammeh(Beyer1984:384(justbefore438CE).Inthecourseofour meetinginJerusalem, Yosef Tobipointedoutthatnumerousotherlateoccurrencesforthe nameareknown,thusloweringitslowerterminus.Different,butnotveryunusual,isthat theJewish-Aramaicowneriscalled“sonof,”notben butratherthe Aramaicbar . Tojudge fromthescript,names,andiconography,adatefromthe2centuryCEonwardseemsplaus-nd ible. Thekindofstone,itsreverseinscription,andthemotifallbespeakaseal.5 Thisringis undoubtedlyancientandprovidesevidencepotentiallyfortheearliestJewishsettlersinthe area.Eventhoughthesealringandits“Jewish”characterrepresentarelativelyunique findatthisjuncture,oneshouldkeepinmindtheoccurrenceofsimilarfindsinotherareas oftheMediterraneanlittoral,whereJewishpresenceisattestedintheRomanandByzan- tineperiods. Themajorityofinscriptionsuncoveredin“Jewishcontexts”indicateastrong assimilationtothelocalculturesbothintermsofthelanguagesusedandthenamesadopted. ThisisthecasenotonlyforinscriptionsfoundinregionswhereJewsconstitutedaminori- ty,suchasNorth AfricaorItaly,butalsoinregionssuchastheGalilee,wheretherewasa majorityJewishpopulation.6 Beyondtheuseofan Aramaicnameandthe Aramaicscript, the TorahShrinedepictedastheringscentralmotifgivesclearindicationoftheownersrel- igiousaffiliation.7 The TorahShrine,acommonlyusedsymbolinLateRomanandByzan- tinesynagogues,incarnatesthenewritualandspiritualvaluesthatbecamecentraltoJuda- ismafterthedestructionofthe TempleinJerusalemin70CE. Animalsacrificepreviously performedbyacoterieofpriestsintheJerusalem Templewasnowreplacedbythestudyof theholyscripturesandtheinstitutionofprayerthathithertocouldbeperformedinany-

5 EJ 1971:1072-1081;1074:onsealsfromthe2nd-5thcenturies.InBabylonia,from634to644,thecaliph forbadeChristiansandJewsfromcarryingseals,withexceptionofexilarchsand geonim. 6OnthemethodologicaldifficultiesassociatedwithdistinguishingJewishandChristianinscriptionsfrom paganones,seeKraemer1991:141-162.FortheJewishcommunityinRome,seeRutgers1995:176-184, whopointstotheoverwhelmingmajorityofGreekandLatininscriptionsandthealmostcompleteabsence ofSemiticinscriptions;forPalestine,seeLapin1999:239-268. 7 Incidentally,anotherLate AntiquecarneliansealringfromaclearlyJewishcontext,supposedlyownedby thelandlordofaRomanperiodmanorhouseatRamatHanadiv,featuresneitheran Aramaicinscriptionnor anexplicitJewishsymbol.ItbearstheimageofeitherPoseidonor Apollo(HirschfeldandFeinberg2005:29) and,alternatively,maybeanimportedRomansealstone.OnRomansealstonesinJerusalem,seePeleg2003: 52-67.

16 where,whereveracommunitywasformedandaplaceofworshipwaserected.8 Contem- porarydepictionsof TorahShrinesareknownfromnumeroussynagoguemosaicfloors suchasatHammat Tiberias(Dothan1983,Pls.26-27),Beth Alpha(Wilkinson1978:16, 20-21),Na'aran,HorvatSusiya,BethShean,andSepphoris(6th century, Weiss2005).Less commonisitsoccurrenceinreliefsculpturesuchasinthesynagogueatCapernaum(Loff- reda1985:36). A rarecase,giventhelowchancesofpreservationisitsappearanceona rd wallpaintingintheSyriansynagogueatDuraEuropos(3 century).9 MoststrikingatZafarisasteppedchamberlocatedbelowtheoldcubicalmosque(z070) onthesouth-westfootoftheHusnRaydan,knownasal-Jahw(Fig.14-15 ). Althoughprob- ablybuiltsome1000yearsaftertheundergroundcavity(z071),thesouth-westwallofthe mosquealignsitselfperfectlywiththenorthernfaceoftheinitialdescendingstaircase. This structuralcoordinationmaybeconnectedtothefactthatanearlierbuilding,possiblycon- temporaryandinfactrelatedtothesteppedpool,existedinthesamelocationasthecurrent houseofworship.Localsourcesassumeachurchtohavestoodinplaceoftheoldmosque, howevernoconcreteevidencecanconfirmthisclaim.Conversionofsynagoguestochurch- esandviceversa,asrecordedin“TheBookoftheHimyarites”(Moberg1924:cv,cix), may evensuggestthateitheroneorbothmayhaveexistedinplaceofthemosqueatanearlier time.Ifnumeroustombsuncoveredneartothemosquearecontemporaneouswiththeear- lierbuilding,itcouldnothavefunctionedasasynagogue.Incontrast,thejuxtapositionof tombswithapolytheisticorChristianhouseofworshipwouldhavebeenpossibleoreven likely.However,giventheuncertaintyofthesitechronology,thefunctionoftheearlier building/s,andtherelationshipbetweentheundergroundcavity,thebuilding/sontop,and thetombs,thediscussionremainsspeculative.

A moredetaileddescriptionoftheundergroundstructuremayhelptodeterminewhether additionalphysicaltracesofaJewishpresenceatZafarcanbedetected. Aftertheinitial16 steps,alandingredirectsthepositionofthestaircaseina90°angletotheleft(northward). Anadditionalthreestepsleadintothepoolproperwhichisofasymmetricformandcurves backtowardtheinitialstaircase,asifbeginningaspiral. Theinstallationisentirelycutfrom themaficbedrock. Theworkmanshipisofhighquality,typicaloftheHimyariteperiod,with smoothwallsandpreciselycutcorners(Yuleetal. 2007). Themaximumcapacityofthis undergroundstructure,whichwasevidentlydesignedtoholdwater,is1000litres.Upon discovery,rainwaterfilledapproximatelyhalfoftheinstallationandwasrecordedat1.6m depthabovethefloor. Atthisleveltheinstallationheldapproximately800litres.Ifthewat- erwereonly1mdeep,thevolumewouldhavebeenapproximately500litres.

Thereisclearlyneitherachronologicalnorfunctionalconnectionbetweentheinstallation andthemosque.GiventhepoolsclosestructuralsimilaritywithRoman-Byzantine miqvaot inPalestine,itcanbearguedthatitoriginallyfunctionedasaJewishritualpool. Anaston- ishingresemblancecanbedetectedifcompared,forexample,tothenumerousritualpools uncoveredatSepphorisinGalilee(HoglundandMeyers1996:39-43). Thoughoriginally believedtohavevanishedasaninstallationandpracticeafterthedestructionofthe Temple

8Levine(2003:91-97)pointsoutcorrectlythatthemeaningofacertainsymbolmayvarydependingonits context.ForacompleteandbrilliantstudyonsynagoguesinLate Antiquity,seealsoLevine2005. 9 Levine1999:335,fig.74(goldglass);216-217,fig.31and32(mosaicfloors);Goodenough1953-68,vol. 13:191-192.ButseealsoGöbl1973, Siegel-Motivklasse 98a,foranaediculasimilarinappearancetothe TorahShrine.

17 inJerusalem(R.Reich1988:102-107),miqvaot wererecentlyshowntohavemaintained theirfunctionwithinthecontextofatleastsomeJewishhouseholdsthroughouttheLate RomanandevenByzantineperiods(Galorinpress). Whilecautionisadvisedwhenusing talmudicsourcesandmishnaiclegalprescriptionsasviablehistoricaldocumentstorecon- structdailylifeandsocietyinRoman-ByzantinePalestineatlarge,theirvalueforunder- standingJewishethicsandrituals,atleastasperceivedbytherabbinicelite,isimportant. Fromahalakhic (Jewishrituallaw)pointofview,undergroundcavityz071couldcertainly havefunctionedasamiqveh .Itcanhold,evenifonlypartiallyfilledwithwater,themini- malrequiredquantityof40seahs .10 Thisquantityisnecessarytoallowfullbodyimmer- sion.Rainwatercaneasilyfillthecavity,thoughitisnotclearifitwouldhaveoriginally beenchannelledthroughapipeoraqueductor,alternatively,ifitwouldhavesimplyflown directlyoverthestairsandintothecavity.Intheflooroftheinstallationthereisnoprovis- ionforwaterdrainage.Multipleplasterlayerswereconsideredanimportantfactoriniden- tifyingmiqvaot inPalestine,howevertheirindispensablefunctioninallwaterinstallations inthatregionshouldbestressed.IntheregionalcontextofZafar,noplasterwasnecessary torenderwater-holdinginstallationsimpermeable.

. Fig.17Planofthecisternz096. Fig.18Crosssectionofcisternz096,viewtothesouth-west.

Anadditionalwaterinstallation,inthiscaseclearlyacistern(z096,Fig.16-18 ),islocated 10mwestofthesteppedpool(z071)andmeasures4.5minheightwithacapacityofsome 40,000litres. Althoughoncementionedasagranary(www.zafar-himyar.com),itsidentifi- cationasacisternismorelikelybecauserainwaterflowsintoitnaturally.Ifitwereagran- ary,theproblemwouldbetokeepthewaterout. Theproximityandsimilarcontextofthe twowater-holdinginstallationsinvitesacomparison.Inspiteofthedifferenceintool marks,thehigh-qualityworkmanshipcharacteristicoftheHimyarite Agesuggeststhatboth werebuiltatthesametime. Thedissimilarityinshaperesultsfromtheirdifferentfunctions.

InthediscussionofearlyJewishpresenceinHimyar,auniquedepictionofamaleface comesintoquestionowingtothedepictedside-lock(Fig.19 , Yule2005;Doe1971:107, pl.26). WithregardtoJewishside-locksinourhistoricalcontext,threepointsmustfirst beexplained. First,side-locks(peot inHebrew)aregenerallyconsideredasub-recenteasternEuropean

10 Estimatesforhowmanylitre sequal40seahs varyfrom250to1000litre sofwater.

18 Fig.19Himyaritemaskinalabaster(calcite), Fig.20Nichewithapomegranateandhatchedpattern, courtesyofMetropolitanMuseumof Art,1982.317.1; ZafarMuseum;zm0004;heightc.1.2m. height24cm. practicebasedontheteachingsofthebookofLeviticus(19:27).Butdelvingdeeperinto thematterbringsnewevidencetolight,whichisnotgenerallyknown.Firstofall,while Goiteinconcludedthatthewearingofpeot developedinthe Yemenbyadecreeinveighed againstJewishSabbatean-messianicactivitytherein1667(Tobi1995),Lecker(1997) foundanearlierreferencetoaJewwithtwoside-locksinPre-Islamic(Yathrib). Secondly,relevantinthiscontextisthedatingoftheHimyaritemaskthatshowscharacter- isticsdatablebymeansofRomancomparisonstothe2or3nd rd centuryCE(similarinstyle andtype: Antonini2001: Tav.49,C65). Thirdly,ourmaskhasonlyasingleside-lockand nottwo,whichindicatesthatthisisnotanearlyrepresentationofJewishside-locks.None- theless,thediscussionindicatesthattheyexistedatleastinthelatepre-Islamicperiod,even ifrepresentationshavenotsurvived.

A well-preservedHimyaritenichefashionedingreymarble,saidtohavebeenfoundin Zafar,maybementionedasapossiblecandidateforJewishiconography(Fig.20 ). Above thearchapomegranatemotifiscarved.Butwhilethepomegranateisasymbolsignificant toJewishart,italsoappearsincountlessotherhistoricNearEasterncontexts.Several niche fragmentsareincludedinthecollectionofthesitemuseum,butthisoneisintact. Othershavealsocometolightinrecentexcavations(Yuleetal. inpress1),forexample, fromtrenchz300.Motifsofsimilarformhavebeenidentifiedasdepictionsofa TorahShrine (Hachlili1998:364,Fig.vii-43,depictionsonclaylamps).Owingtothesmallsizeofthe relief,anattributionasa TorahShrineitselfseemsunlikely. Thus,othercontentualoptions existtointerprettheoriginandmeaningofthisHimyaritework. Tojudgefromthestylistic detailsofthestoneworking,thesemaydatetothe2to4nd th centuryCE. Theidentification asanIslamicmihrab isunlikelyowingtothestylisticdatingofthestoneintheHimyarite age.However,theSabaeanwordmhb inthissamehistoricalcontextistranslatedasachan- 19 cellery,ofakingorimportantpersonage,for example,butalsoasanarchitecturalfeature (texts:Robin2006:160).

Finally,yearsagoareliefcametolightwhich purportedlycamefromBaynuninthehigh- lands(Fig.21 =Radt1973: Taf.36.99). The subjectofawomanwithlonglocksbearinga birdonherleftforearmandabranchinher righthandisincisedinsimplelinearfashion intothecalcite(alabaster). Theextremelinear- ityofthisimageisunusualinOSA art(cf. however,anotherlinearreliefwiththesame Fig.21 Womanholdingabirdandabranch, motifalsofromBaynuninRadt1973: Taf. purportedlyfromBaynun;height14.5cm. 36.100).Pirennesattributionofthefigureas thebeneficentgoddess Atargatisanditsdatingto100CEareequallyunsupported(1977:I. 445-446). Thisdepictionmayhaveadornedatombstone,butaheightof14.5cmforthis artefactseemstoosmallforthispurpose. Theunusuallinearstyleandtechniqueseemsmore intunewiththe4th -6 th centuriesCEthanwithearlierperiods. Thebranchandbirdattributes inevitablybringtomindthedoveandolivebranch,astandardmotifinearlyChristianart whichreferstopeace,butalsototheholyghost,andhasnumeroussymbolicattributions. Particularlylater,thismotifisassociatedwiththefirstdoveNoahsentattheendofthebib- licalfloodthatreturnedwithafreshlypluckedoliveleaf(Gen.8:10-11). Whilethedove alsoplaysaroleinJewishiconographyasasymbolofpeaceandgiventhestrongassocia- tionswithChristianiconography,itseemslikelythatthisisoneofthefewsuchexamples fromthehighlands. Asunderstood,itmustbeoneoftheearliestusesofthissymbol.

Conclusion InspiteofthedifficultyinestablishingaprecisechronologyfortheJewishcommunityin ZafarandtheHimyariteregion,itspresencethereinthelatepre-Islamicperiodiswidely documentedintheliterarysources. Archaeologicallyspeaking,theidentificationofstruc- tures,materials,orsymbolicattributesthatcanbeassociatedwithspecificethnicandrelig- iousaffiliationsisrathercomplexasituationnotuniquetoSouth Arabia. Thisistruefor Late Antiquityasawhole,inparticularinregionstraversedbyinternationaltraderoutesto facilitatecommerce,contactwithforeigncultures,andoccasionallymigration.Exchange andexposurecanclearlyinspireandenrichalocalculturebutwill,attimes,completely disguiseitsoriginalfeaturesandidentity.Inadditiontothegeographicalconsiderationsof regionalmovementwithinandbeyondtheHimyaritekingdom,wearedealingwithother factorswhichobscureourunderstandingandrecognitionofreligiousidentity. Thelatepre- Islamicperiodisoneofthemostrevolutionaryphasesinthehistoryofreligions,during whichpolytheismwasgraduallyreplacedbymonotheism,which,inturn,maturedand branchedintomanydifferentstreams. Thoughitisthedifferencesbetweenthemthathave beenafocusofscholarship,itisthecommonalities,interconnections,andinterdependen- ciesthatarefarmoredominantandsignificant. Thetheologicalandideologicalrelation- shipsbetweenvariousreligiousgroupshaveclearrepercussionsforthedomainofthemat- erialremains.Evenwhereonewouldexpectmajordifferences,suchasfortheirhousesof worshipandburialcustoms,distinctionsareoftensubtleandmoresignificantlydetermined byregional-culturaldenominatorsratherthanbyreligiousbeliefs.Inspiteofthosedifficul- tiestoidentifythematerialremainsofaspecificreligiousgroup,somefindsatZafarcan almostbecertainlyclassifiedas“Jewish.” 20 Abbreviations EIEncyclopediaofIslam EJEncyclopaediaJudaica PSASProceedingsoftheSeminarfor ArabianStudies Vienna1998Jemen,Kunstund ArchäologieimLandderKöniginvonSaba,ExhibitionCatalogueVienna, Künstlerhaus,9November1998to21February1999 al-Ansary, A. 1981 Qaryatal-Fau, A PortraitofthePre-IslamicCivilizationinSaudi Arabia

Antonini,S. 2001 Lastatuariasudarabicainpietra, Repertorioiconograficosudarabico,t.1

Barceló,M. etal. 2000GoingaroundZafar(Yemen). TheBanuRuaynFieldSurvey:Hydraulic ArchaeologyandPeasant Work,PSAS 30:27-39

2003TheSearchfortheHararahasdadintheareaofZafár,GovernorateIbb, Yemen,PSAS 33:133-142

Behnstedt,P. 2002The ArabicDialectsofal-Yaman.In:EI 11:277280

Beyer,K. 1984 DiearamäischenTextevomTotenMeer

Breton,J.-F. 1991LechâteauroyaldeShabwa:notesd’histoire,Syria 68:209-227

Doe,D.B. 1971 Southern Arabia

Dothan,M. 1983 HammathTiberias:EarlySynagoguesandtheHellenisticandRomanRemains

EJ 1971Seal,Seals.In:EJ 14:1072-1081

Galor,K. 2003QumransPlasteredPools: A NewPerspective.In:J.-B.HumbertandJ.Gunneweg(eds.), Science and ArchaeologyatKhirbetQumranand'AinFeshka,Studiesin Archaeometryand Anthropology, vol.2:169-198 inpressThe WaterInstallationsoftheSepphoris Acropolis.In:D.EdwardsandC. T.McCollough(eds.), The ArchaeologyofDifference:Gender,Ethnicityandthe'Otherin Antiquity,StudiesinHonorofEricM. Meyers

Göbl,R. 1973 DersasanidischeSiegelkanon.HandbuchdermittelasiatischenNumismatik,4,Braunschweig

Goodenough,E.R. 1953-68JewishSymbolsintheGreco-RomanPeriod

Hachlili,R. 1998 AncientJewish Artand ArchaeologyintheDiaspora

Hirschfeld, Y.andFeinberg,M. 2005A CountryGentlemansEstate.UnearthingtheSplendorsofRamatHanadiv, Biblical Archaeology Review 31/2:18-31

21 Hoglund,K.andMeyers,E. 1996TheResidentialQuarteronthe WesternSummit.In:R.Nagy,C.Meyers,E.Meyers,andZ. Weiss (eds.), SepphorisinGalilee.CrosscurrentsofCulture:39-43

Ilan, T. 2002 LexiconofJewishNamesinLate Antiquity,pt.1:103-105

Kraemer,R. 1991Jewish TunaandChristianFish:IdentifyingReligious AffiliationinEpigraphicSources, Harvard TheologicalReview 84:141-162

Lafitte,R. 2003Surlezodiaquesudarabique, Arabiarevuedesabéologie 1:77-87

Lapin,H. 1999PalestinianInscriptionsandJewishEthnicityinLate Antiquity.In:E.M.Meyers(ed.), Galilee throughtheCenturies.ConfluenceofCultures:239-268

Lecker,M. 1997Zaydb. Thabit,“A Jewwith TwoSidelocks”: JudaismandLiteracyinPre-IslamicMedina(Yathrib), JournaloftheNearEasternSociety 56:259-273

Levine,I. 2003ContextualizingJewish Art. TheSynagoguesatHammat TiberiasandSepphoris.In:R.KalminandS. Schwartz(eds.), JewishCultureandSocietyundertheChristianRomanEmpire:91131

2005 The AncientSynagogue.TheFirstThousandYears.2ndedition.

Loffreda,S. 1985 RecoveringCapharnaum

Moberg, A. 1924 TheBookoftheHimyarites,FragmentsofahithertoUnknownSyriacWork (SkrifterutgivnaavKungl. Humanistiska VetenskapssamfundetiLund, VII)

Müller, W. 1979ZafarundHimjar, Yemen-Report 10:16-17

Peleg,O. 2003RomanIntaglioGemstonesfrom AeliaCapitolina, PalestineExplorationQuarterly 135/1:52-67.

Pirenne,J. etal. 1977CorpusdesinscriptionsdeantiquitésSud-Arabes

Radt, W. 1973Katalogderstaatlichen AntikensammlungvonSanaaundanderer AntikenimJemen

Rathjens,C./v.Wissmann,H. 1932 Vorislamische Altertümer,Rathjens/v.WissmanscheSüdarabien-Reise2, AbhandlungenausdemGe- bieteder Auslandskunde38

Reich,R. 1988TheHotBath-House(Balneum),theMiqwehandtheJewishCommunityintheSecond TemplePeri- od, JournalofJewishStudies 39:102-107

Robin,C. 2003LejudaïsmedeHimyar, Arabiarevuedesabéologie 1:97-172

2006 Himyarjuifetchrétien(c.380-560),Inventaireetanalysedesdescriptionslocaleslesplussignifica- tivespourlesévolutionspolitiquesetreligieuses,privatelycirculatedtext 22 inpress HimyaretIsraël

Rutgers,L. 1995 TheJewsinLate AncientRome:EvidenceofCulturalInteractionintheRomanDiaspora

Schiettecatte,J. 2004 Unedefinitiondela“ville”préislamiqueen Arabiedusud, Arabiarevuedesabéologie2:123-142

Schmidt,J. 1997TempleundHeiligtümerinSüdarabien.ZudenmateriallenundformalenStrukturenderSakralbau- kunst,NürnbergerBlätterzur Archäologie 14:10-40

Sedov, A. 1998 DerHafenvonQnidasTorzumJemeninfrühnachchristlicherZeit.In: Vienna 1998:274-278

Sima, A. 2002DieGeschichtederStadtZafar (lectureheldinBamberg29.06.2002) inpress DieInschriftendesZafarMuseums.in:N.Nebes(ed.),EpigraphischeForschungenaufder Arabi- schenHalbinsel,

Smith,G.R. 2002Al-Yaman,Historyfrompre-Islamic Timesto1962.In:EI 11:271-274

Tobi, Y. 1995MaamaroShelS.D.Goitein'alpe’otwe-'alsimonim, Têmâ:ketav'etleheqeryahudûtTêmânwe tarbûtah, AssociationforSocietyandCultureNetanya 32:97-109

VanBeek,G.W. 1958MarginallyDraftedPeckedMasonry.In:R.LeBaronBowenandF.P. Albright(eds.), Archaeological DiscoveriesinSouth Arabia,II:287-295

Weiss,Z. 2005 TheSepphorissynagogue:Decipheringan AncientMessagethroughIts ArchaeologicalandSocio- HistoricalContexts

Wilkinson 1978TheBeit AlphaSynagogueMosaic: TowardsanInterpretation.In: JournalofJewish Art 5:16-28

Wilkinson, T.J.,Edens,C.andGibson,M. 1997The Archaeologyofthe YemenHighPlains: A PreliminaryChronology, Arabian Archaeologyand Epigraphy 8:99-142

Yule,P. 2005Zafar TheCapitalofthe AncientHimyariteEmpireRediscovered, Jemen-Report 36:22-29

2006TowardaChronologyforHimyariteSculpture(lecture),Rencontressabéennes9,25-27.05.2005,Jena

2007 Himyar.SpätantikeimJemen/Late AntiqueYemen, Aichald

Yule,P. inpress1Zafar,CapitalofHimyar,FifthPreliminaryReport,February-March2005, Beiträgezur Archäolo- gieaußereuropäischenKulturen inpress2Zafar,CapitalofHimyar,SixthPreliminaryReport,February-March2006, ZeitschriftfürOrient-Ar- chäologie,inpress

23 Yule,P.P. Yule/K.Franke/C.Meyer/G.Nebe/C.Robin/C. Witzel, 2007Zafar,CapitalofHimyar,IbbProvince, YemenFirstPreliminaryReport:1998and2000,SecondPre- liminaryReport:2002, ThirdPreliminaryReport:2003,FourthPreliminaryReport:2004.In: ABADY 11,479-547,Pls.147+CD-ROM

P. Yule/K.Franke inpressZafar,CapitalofHimyar,SeventhPreliminaryReport,February-March2007andFebruary-March 2008, ZeitschriftfürOrient Archäologie, inpress

Zarins,J. etal. 1983PreliminaryReportontheNajranUkhdûdSurveyandExcavations1982/1402 AH,Atlal 7:22-44

Paul Yule SeminarfortheLanguagesandCulturesoftheNearEast SeminarforPrehistoryandNearEastern Archaeology D-69117Heidelberg Tel/Fax00499515191616 [email protected]

KatharinaGalor JoukowskyInstitutefor Archaeologyandthe Ancient World BrownUniversity Providence,RI02912USA [email protected]

Persistente URL: http://archiv.ub.uni-heidelberg.de/propylaeumdok/volltexte/2008/134/

24