7/29/2019 20004olteanphd

1/274

Later prehistoric and Roman rural settlement and land-use inwestern Transylvania

loanaAdina Oltean

Submitted or the degreeof Doctor in Philosophy

University of GlasgowDepartment of Archaeology

May 2004

r-; W-f.

@ oanaA. Oltean,2004

7/29/2019 20004olteanphd

2/274

THESIS CONTAINS

VIDEO CD DVD TAPE CASSETTE

7/29/2019 20004olteanphd

3/274

Abstract.-

ThepresentstudyanalysesRoman-native nteraction rom a landscapeperspectiven a core erritoryof both Iron Age andRomanDacia. The study area ncludes heroyal Dacianheartland the OrastieMountains)and ts surroundingowlands,and also hehinterlandsof Ulpia TraianaSarmizegetusaandApulum,thetwo most mportant Roman ownsin theprovince.The research onsidershenatureanddistributionof lowcr-ordersettlementsn thepre-RomanandRomanperiods,human mpacton the local landscapeandthechangeswhich occurredas a result oftheRomanoccupation.Also, it addressesreviousbiasesof interpretation hroughre-evaluationofearlierdataandconsideration f newdatasetsprovidedby the interpretationand mappingof recentobliqueaerialphotographs.New detailedplansof the sitesdiscoveredhroughaerialphotographyhavebeen ntegratedwithin a significantamountof scattered ublisheddata(excavationandfieldwalking reports; gazetteers) ndrelevant nformation from historicalmaps. All thematerialhasbeenanalysedutilising a relationaldatabaseinked to a GIS. In this way the spatialdistributionofsettlements f various ypesandfunction,or therelationshipbetweensettlements ndnaturalresources r topography,has evealedpatterns ndicativeof the factorswhich mayhave nfluencedtheir evolution.

Theresultsprovidea complex econsiderationon a more realisticand up-to-datebasisof previoustheories egarding henative settlementpatternand he impactof Romancolonisation n thechronologicaland geographical ontextspecified.Also, throughtheresultingdatabaseandGIS, itprovidesa methodological rameworkand a custornisedool for finther analysisof the landscape ndof theevolutionof thesettlement atternwhich canbeextendedhroughout heprovinceof Daciaand nto neighbouringareas.Finally, it createsa usefulsourceOfanalogyor contrast or Empire-widestudiesof romanisationandRoman-nativenteraction.

7/29/2019 20004olteanphd

4/274

Chapter breakdown

Chapter 1: Background, limitations, aims, approachesand methodsChapter 2: The study area: natural environmentChapter 3: The historical backgroundChapter 4: Late pre-Roman Iron Age landscape;Iron agerural settlementpattern and societyChapter 5: The Roman landscapeChapter 6: Romanisation of the rural landscape

7/29/2019 20004olteanphd

5/274

Acknowledgments:

This study would not havebeenpossible n Romaniaonly a few yearsago,before he first aerialreconnaissance rogrammeexclusively or archaeological urposeswasstartedby Prof W. S.Hansonandmyself in WesternTransylvania.Apart from providing thevital photographsand mapsfor this study, heproject,financedby theLeverhulmeTrust andtheBritish Academy,hasintroduceda new perspectiveo archaeologyn theregion. This has beenpromoted hroughoutmythesis.Many thanksmustgo to Prot W. S.Hanson or setting up this project,andto all thecollaboratorsand ocalauthorities hatensuredts running.This new approacho thearchaeology f Daciawasapparenthroughout he sessionon RomanDaciaorganisedby Prot HansonandDr. L Haynes London)at theRomanArchaeologyConferencenGlasgow n 2001.Thepaperspresented ndthediscussionsn that sessionhave beenextremelybeneficialto this study-,Prof.HansonandDr. Hayneshavebeenextremelykind in making availablethematerialssubmitted or publication n the SupplementarySeriesof theJournalof RomanArchaeologyprior to theirpublication(HansonandHaynes2004).Moreover,Dr. L Haynes,andDr.K. Lockyearfrom theApulurn HinterlandProject,kindly allowedmeto include n this study heirgeophysicalsurveying esultsat SeusaandOarda.I would like to thankto Prof Hansonalsofor thehigh standardof academicexcellencen thesupervisionof my Ph.D. researchwhich hasgreatly nfluenced he waymy own thoughtsand deashavebeenexpressedhroughout his study-, lso,for patientlyproviding his full supportand ensuringeverything neededor its completion.Many thanks o Dr. K- Brophy, Dr. P. vanDommelen,Dr. R.Housley,Prof A. B. Knappand all theacademicstaff and o my fellow postgraduate tudentsn theDepartmentof Archaeologywho providedanextremelychallengingandstimulatingenvironmentthroughoutmy time there.Also thanks o Prot L PisoandDr. A. DiaconescuBabes-BolyaiUniversity of Cluj Napoca) or providing me with a solid basisof knowledgeof Romanhistoryandarchaeology, ndfor guiding my first attemptsat academicesearch.Many thanks, oo, to R. Palmer Cambridge) or additional nstruction in air-photo nterpretationandmapping,usefulcommentsand advice,printing of photographic llustrationsfor thethesisandhissenseof humour,to him, and also o Prof 1.Scollar(Bonn),Dr. M. DoneusandDr. W. Neubauer(Vienna),H. Corley (Glasgow) hanks or help in persuadinghe computers o help me in sitemappingandGIS designand analysis.Also, thanks o Dr. D. Ruscu(Cluj Napoca) or usefulcommentson partsof this thesis; o him andto Dr. L. Ruscu,Dr. C. Ciongradi,V. Radeanu,Dr. E.Bota,my gratitude or help in gathering herelevantbibliography for this studyandfor constantmoral support.Dr. C. CiongradiandDr. G. Gheorghiu Cluj Napoca)facilitatedmy accessorelevant nformationcontainedn their PhDthesesat theCologneUniversity andtheBabesBolyaiUniversity of Cluj Napoca espectively.Many thanks o LorraineMcEwan andthesupportstaff oftheDepartmentof Archaeologyn Glasgow-,ast but not least, o my family for all their moralsupportand patience.The financialsupportwhich made his study possiblewas providedby theUniversity of Glasgow,through ts 2001 Postgraduate esearchScholarship,andby the OverseasResearchStudentAwardsscheme.Various fiinds from theGraduateSchoolof Arts andHumanitiesandfrom theSchoolofHistory andArchaeologyat GlasgowUniversityhavealso supportedvisits to siteson thegroundandto relevantmuseumsand ibrariesin Cluj-Napoca,DevaandAlba lulia, andproductionofphotographicllustrationsnecessaryor my thesis.The Derrick Riley Bursary for aerialarchaeologysupportedheacquisitionof Coronasatellite mageryused n this study.

IoanaAdina OltcanMay 2004,Glasgow

7/29/2019 20004olteanphd

6/274

Dedicatedto those ongsummernightsat Sar7nizegetusa

7/29/2019 20004olteanphd

7/274

List of illustrations

1.1 Romanremains on the 1870-1875Austriancadastralmapof Transylvania: he auxiliary fort atMicia; thecolonia Sartnizagetusand surroundingssections183and23 1; Copyright theAustrianStateArchives, Vienna).1.2 Remainsof theRoman oad n Tara HateguluibetweenUnciuc andSubcetatesection232)and oftheRomanroadalong heMuresfrom theriver crossingnearGelmar, owardsSibot (CopyrighttheAustrian StateArchives,Vienna).1.3 Wpia TraianaSarmizegetusa-generalview of themodemvillage from the eastwith theIronGatesof Transylvania n thebackground1.4 Wpia TraianaSartnizegetusa-the forum.1.5Apulum-general view anddetailof theeasternenclosurewall andrampart) visible as cropmark1.6 MappedRomanbuildings from cropmarkevidencen thecolonia and hemunicipium atApulum1.7 Low-altitude aerialphotographof a Romanbuilding (asa negativecropmark) somewhere orth ofAlba Iulia by V. Barbuta1.8Flight track logs in thestudyareasince19981.9Map of Daciaand ts neighbours1.10Generalview of theGradisteaMuncelului hill, thesite of theDaciancapital,Sannizegetusa egia2.1Map of themain topographicalunitsand he locationof modemsettlementsmentionedn text2.2River meandersonMuresandStreivalleysat theMures-StreiconfluencenearSimeriaand nearCalan2.3Aerial photographof areascoveredby industrial refuseat HunedoaraandnearDeva2.4Waterreservoirs n Tara Hategului2.5Extensivemodemdrainagesystemn TaraHateguluieastof Sarmizegetusa4.1 Dacianhouseplans n uplandsettlement4.2 Man-made erracesn theGradisteaMuncelului area,probably ateIron Age in date4.3 Aerial photographof thevillage of sunkenhousesand pits at Vintu deJos4.4 Siteplan: Multi-prehistoric village atVintu de Jos4.5 Planof theDacianhillfort at Cigmauoverlainon a DTM of thearea4.6 Aerial photographof thehillfort at Costesti-Cetatuie4.7 Aerial photographof thehillfort at Costesti-Blidaru4.8 Aerial photographof thearea sacraat Sarmizegetusa egia4.9 Aerial photographof thehillfort and umuli at Cugir4.10 Aerial photographshowing heremainsof the fortified siteatBreteaMuresanadestroyed ymodemquarrying4.11Aerial photographof themulti-periodsettlementat Simeria4.12Aerial photographof thehillfort on theSampetruhill, Hunedoara;he largeenclosuredatesn theHallstattperiod,but a smallercircularenclosures visible towards ts edge,perhaps elated o theDacianoccupationof thesite attestedby artefactdiscovery4.13Aerial photographof the ate Dacian nfant cemeteryexcavationat Hunedoaranearby hemedievalcastle4.14 Aerial photographof thecitadelhill at Deva4.15Quarryon Bejanhill nearbyDeva4.16Magura Calanuluiquarry4.17Distribution of Daciansettlement4.18Distribution of pre-Romansettlementn relationto topographyand current anduse;settlementbuffersat2 kilometres4.19Distribution of Dacianhoards n relationto settlementocations;settlementbuffersat2 kilometres

7/29/2019 20004olteanphd

8/274

4.20 Prc-Romanndustrialactivity4.21Distributionof pre-Romanspiritualactivity5.1Planof theAiudul de Susvilla houseshowingdifferences n walls width and a possibleinterpretationof internalspacedivision5.2Non-villa building atAlba lulia5.3 Multiple buildingsof theRomanvilla atOarda evealedas negativecropmarksn cereals5.4 Planof theRomanvilla atOarda5.5 Geophysicalsurveying results rom Seusaand rom a secondvilla siteat Oarda (courtesyof Dr. IanHaynes)5.6 Generalview of the site of Romanvilla adjacento thepit andsunkenhousevillage eastof VintudeJos5.7Roman eguldeand prehistoric pottery from thesite5.8 Siteplan:Romanvilla nearby heprehistoricvillage eastof Vintu de Jos5.9 Multiple buildingsvisible as cropmarksundermodemcultivationat Vintu deJos ndicatingapossiblevilla or village5.10probableRomanvilla or village atVintu deJoson thesouthernbankof the river Mures (map)5.11Romanandmedieval materialcollected rom the site of possibleRomanvilla or village atVintude Jos5.12Streivilla andearly medievalchurch5.13Villa siteat Sibotpartially revealedasnegativecropmarksbecause f cultivation in long narrowstripsof differentcrops, someof themnot influenced n their growthby the archaeologicaleaturesunderneath5.14Siteplanof the Romanvilla at Siboton thenorthernbankof theriver Mures5.15Probablebuilding of Romandatenear heRomanroadline at Sebes5.16Siteplan:PossibleRomanbuilding alongside oadline5.17 Building of a probableRomanvilla atHobitavisible in pasture5.18 Siteplan:ProbableRomanvilla at Hobita5.19 A secondpossibleexampleof villa at Hobita5.20 Planof possiblevilla eastof Hobitaon Sucionihill5.21Generalview overTaraHateguluifrom Hobitavilla (fig.5.17)andmillstonescomingfrom thesecondsite(fig. 5.19)5.22Possiblevilla building visible undergrasscoverageat Salasude Sus5.23Siteplanof thepossiblevilla at SalasudeSus5.24Plansof villa complexesandbuildings in Dacia5.25Aiud-Cetatuiemound n theMuresvalley settledat multiple moments hroughout ime; in theRomanperioda homestead r villa was ocatedat the easternend5.26Distribution of Romanvillas and ndividualhomesteads5.27Tracesof quarryingand associated ettlementat Coziaof probableRomandate; hecircularstructurecouldbe latePrehistoric n date5.28Extensive emainsof gold quarryingactivity sinceRoman(possiblyDacian) ime atPianude Sus5.29Limestonequarryof probableRomandate o thewestof Sannizegetusa 7pia5.30 Romanandesitequarry atUroi (possiblePetris) in theMuresvalley5.31Excavatedareaof themilitary vicusatMicia -thermae andamphitheatre5.32Partsof thenorth-eastemareaof thevicusatMicia revealedascropmarks n successive tripfields5.33Micia: domesticbuildingsto thewest of thefort5.34Micia: buriedbuildingsandroads n thesouthernsectorof thevicus5.35 Siteplan:Vetel-Micia Romanmilitary vicus5.36Cigmau-general view of theauxiliary fort andof thecentralpart of theWcusshowingbuildingsand nternalroads; n the foreground centre) hetriple ditch andrampartenclosuresystemof thepre-Romanhillfort arestill visible partially extant5.37Cigrnau-view of the road eading rom themain coreof thevicuseastwards,o GeoagiuandGelmar,faint tracesof buildings insidethevicus alongtheroadandelsewhereare alsovisible5.38Siteplan:Cigmau-Germisara,omanauxiliary fort andmilitary vicus

2

7/29/2019 20004olteanphd

9/274

5.39Layoutand errainmodelof thefort and military vicusat Cigrnau-Germisarain ArcView GIS)andexaggeratedeight factor3-dimensionalerrainmodelof the locationof the fort and military vicusat Razboieni in AirPhoto)5.40Razboieni-vicus buildingsandroads o thenorth of the fort5.41Razboieni-vicus buildingsand roads o thenorth-eastof the fort5.42Razboieni-vicus buildingsand roads o thesouth of the fbrt owards heriver5.43 Siteplan:RazboieniRoman ort andmilitary vicus5.44 Distribution of aggregated ettlementn thecontext of Romansites5.45 Possible dIns n thevicusatMcia5.46 Romanspolia in thewall of theearly medievalchurch atDensusand a funeraryaltar re-usedas apilasternsidehechurch5.47Romanndustry5.48Romanerrestrial nd iverine ransportetwork5.49Distribution f Roman onstruction aterialsn relationo roads5.50RomanarchaeologicalandscapeatApulum5.51Roadssystemas possible ndicationof centuriationsouthofApulum, eastof Vintu.de Jos5.52Calan-Bai Aquae)Romanstonepool nearbymodem nstallations5.53GennisaraRomanspaandritual complexnear modempoolsat Geoagiu-Bai5.54Romansub-urbanvilla at Sarmizegetusa5.55Distributionof Romanspiritualactivity6.1Comparativediagramshowing hesettlementype andhierarchy n the ate Iron Age andtheRomanperiods6.2Distributionof Daco-Romansettlement6.3Comparativedistributionof DacianandRomansettlements6.4 Distribution of military sitesandmaterials

7/29/2019 20004olteanphd

10/274

Chapter 1: Background, limitations, aims, approachesand methods

1. The current state of research on rural life in Roman Dacia:Theparticularityof RomanDacia s that its rathershort ife within theempire covers he2d and mostof theP centuryAD. Therefore, t covers heperiod whentheRomanPrincipateexperienced oththepeakof its power,as well as he onsetof decay.Createdby Trajan, in theage of thegreatestterritorial extentandpower of theempire, t endedup beingabandonedater in theP centuryADbecause f political infighting within the empireandthe failure to control its boundaries.Thereasonsfor theconquestwereboth strategicandpolitical. By theendof the V centuryAD theempirehadrefinedits approacho conquestandfurtherorganisationof newprovinces.Theperiodcovering he2nd ndearlyP centuriesAD is generallyacceptedasthemostcharacteristicof Roman mperialruleand, herefore,servesas amodel in variouscomparisons.Consequently, tudying he implementationand mpactof Romanadministrationwithin newterritories,as exemplifiedby Dacia,givesa uniqueopportunity o reveal hepatternof conquestat theheightof Rome'spower.Oneof thegreatexpressionsOfRomanpower in theconquerederritories s the impacton thenaturaland cultural landscape. hecurrentorthodoxyconcerning he mpactof RomanoccupationonDacia,the implementationof Roman ule anddevelopmentof theprovincial landscape, ouldbesurnmarisedn four main pointsthat seem o havebeengenerallyaccepted.Firstly, theRomanconquests seenas adramaticevent, nvolving massivecolonisation.Several iterary sourcesdescribe heprocess,amongwhich CassiusDio andEutropiusseem o give mostdetails.CassiusDio(L XVHI 14,4) refers o Trajan'spolicy of colonisation: " Kai o'brcoC ,dwda 'PC0,Ua!0V WrIK00Ceyevero, ai ; 6Aczc vvrq- Tpaiavbqmr0moev " C'In this way Dacia becamesubject o theRomans,andTrajanfoundedcities there").Later Eutropius (VIII 6,2) wrote,concerningHadrian'spolicy onTrajan'sconquests t thebeginningof his reign: "Idem deDaciafacere conatumamicideterrueruntne multi civesRomanibarbaris traderenturproptereaquia Traianusvicta Dacia extotoorbeRomanonfinitas eocopiashominum ranstuleratad agroset urbe-s olendas,Dacia enimdiuturno belloDecibali virisfuerat exhausta (hewanted o do the same n Dacia but his friendspersuaded im not to deliver largenumberof Romancitizens o thebarbariansbecause fterhisvictory againstDacia, Trajantransfereda considerable umberof peoplefrom all aroundRomanworld to occupy ts landsandcities; Daciawas ndeeddepopulated y the long war againstDecebalus).Both examplessuggesthat thephenomenonwas mportantenough o come o theattentionof ancienthistorians for themostrecentdiscussionseeRuscu2004)and,at leastatasuperficial evel, thearchaeologicalevidenceappearso support hese ecords.Secondly, t is statedthat thenativeDacianpopulationwasmovedfrom their hill forts andsettled n the lower areas,

4

7/29/2019 20004olteanphd

11/274

which wereeasier or theRomanarmyto keepunder control. The archaeological tudy of themajorDacianhill forts seemso indicate hattheir violent destructionoccurredwithin thechronologicalcontext of theconquest Glodariu 1993,15). Moreover, it seemshat there s no evidence hatany ofthesesitescontinued o beoccupiedduringtheRomanperiod.Thirdly, it is generallyassertedhattheRomanpolicy of colonisation.wasratherbrutal,with theauthoritiesdeprivingthenativesof theirlands or town foundations,or colonistsandveteran's and holdingsthatestablished illa estates,orarmyneedsandfor imperial estatesProtase1968,510).Finally, however, t is believed hat thenatives ived in harmonywith the colonistsand romanised hemselves uring the two centuriesofRomanoccupation.Together heyconstitutedhepopulation hatcontinued o setfle hese andsaftertheRomanadministrativeand military retreat rom Daciaat theendof P centuryAD (seeProtase1980,228-252).The inconsistencyand nternalcontradictionswithin theseassumptionswill behighlightedbelow.I Kolendo(1994)stateshat theRomanprovinces alongthe line of Danubewere all characterisedya rural typeof civilisation, becausen theseareashe modelof theancientcity couldbe appliedonlyin part.His conclusions basedon a comparisonwith theurbandevelopment atternseen n theprovincesof Northern Africa or Gaul.In orderto makean analysisof theevolution of thearea,heusesmainlyepigraphicsourceshatprovidehim with informationon theurbanstatusof settlements.Also, through he interpretationof theuseof thenatio (natione)for designatingheorigo of thepersonwithin the text of severalnscriptions rom theDanubearea,hetries to explore hementalitybehindthesepeoples'attitudes.Unfortunately,this results n his supportingargumentbeingbasedentirely on the study of theurbancontextandnot on a study of rural sitesand he andscape.He,aswith many others, akes nto accountas alling into an urban categoryonly thosesiteswith municipalstatus,whethermunicipiaor coloniaeprovenby the epigraphicsources,without any considerationofthereal functionof thesettlementshemselves.Also, even houghhe s awareof someof thefactorsthat wouldhave nfluenced henumberof charteredsettlementswithin a province,suchas hepresence f imperial estates, r ciWtates ereg?nae, or thenatureof thepre-Romanculturalbackground,he fails to addresshe issueswithin abroader andscape ontext and produceaconclusionbasedon ananalysisof thesettlementpattern.As a result,his quantitativeudgementscannothavea realisticbasisand someof his interpretationsseem o push he evidenceo extremes.I havechoseno beginwith this particularstudybecauset is, in my opinion, symptomaticof thekind of archaeological tudy hathasbeen n vogue n Romania or some ime now. In order ojustify theapproachaken n my own research, will try to analyse heway rural life in RomanDaciain particular, as well as provincial archaeologyn general,has beenapproached o far. Despite heeffortsmadeduringthe20thcentury,archaeologicalesearchon rural settlementn RomanDaciasofar hasbeenapproached nly sporadicallyandwith a low priority. Research as ended oconcentrate n military andurbansites, he latterrestrictedpredominantly o siteswith provenmunicipalor colonial status seeabove).Thus,a significant numberof sites hat,asfar asweareaware,did not achievemunicipalstatushavebeensomehowneglected.This categorywould coversites hatare acceptedelsewhereashavingat leasta semi-urban unction (e.g.BurnhamandWacher1990).Thefirst decades f the2CPcenturysaw hebeginningof considerations f rural settlements

7/29/2019 20004olteanphd

12/274

with excavations t severalvilla sites,many of themstill extantat that time, suchasManerau n1912,Apahidaor Garbou n 1913 seeMitrofan 1973,127-150,with full bibliographyfor thefirstpublication of theseearlyexcavations).More attentionwas givento therural sitesafter theWorldWar II though or different, this time political, reasonsinked to the agendaof thecommunist egime.The Romansweredepictedas he imperialistrobbersof Dacian landwhosepreviousoccupantsbecomeservants n theestates f theRomancolonists.Also, attention ocusedonthenativeccworking masses"ather hanon thesurvival of thenative elite (detailed ater in this chapter).Excavationswere conductedon sitesof vici or cemeteries elongingto them(eg.Obreja,SoporudeCampie,Bratei, Cristesti,Micasasa.-seeProtase1998;Protase1980,38-85with bibliography-,HusarandMan 1998,Mitrofan 1999).Similarly, severalvilla sites werealsoresearched,uch asHobita,Deva,SantamariaOrlea,Aiud, Cincis, Chinteni (Mitrofan 1973,1974and 1976;Floca 1953;Popa1972;Winkler et aL 1968;FlocaandValea,1965;Alicu 1994and 1998).Efforts havebeenmade o recordall archaeological ites within theterritory, improvingtheolderdatawith newinformationobtained hrough ield walking. Unfortunately,an ambitiousprojecttoproducea generalarchaeological azetteer f Romaniawhich wasstarteddecades gowasnevercompleted,hough n somecasest resulted n thepublicationof regionalgazetteers f severalcountiesor geographical reas.For mostof thediscoveries,however, he information s scatterednstudiesat various evelsof comprehensivenessublished n variousRomanianarchaeologicalperiodicals.At the time of its appearance,udor's book 'Orase, targuil si sate n Dacia Romana'(1968)was he argestcollectionof published nformation on Romansettlementsn Daciathatspecialistscould rely upon,despite ts badreviews(Daicoviciu, 1969).But now aftermorethan30years he informationneedso beupdated,and hesameapplies o the Tabula mperii Romani(L 34-BudapestandL 35-Bucharest)whose nformationcontinues o beusedby theeditorsof historicalatlasessuchasTalbertet al. 2000). n 2000, heInstitut deMemorie Culturala,Bucharest clMeC),startedanotherambitiousprojectto createa largedatabase f thearchaeological itesof Romaniaaccessible n the nternet o scholarsnternationally, hrougha EuropeanUnion fundedproject(http: /archweb. imec.o/ -lastvisited08.05.2004).The antiquitiesof RomanDaciahaveattractednterestsince heRenaissance,nd t is the reportsofantiquariesand scholars,along with earlymappingprojects(seeFigures 1.1-1.2), hatconstitute hemain testimony o thedegreeof preservationof archaeologicalsites. ndeed,up to the20' centurythesewere still visible on theground,despitesporadicor sometimesextensivedamagesuchasstonerobbingor marbleburning. But only a hundredyears ater,at thebeginningof the2 1" century, helandscapes very muchchanged.The two World Warsof thepreviouscenturyresultedn greatdamageo archaeological ites n Romaniaas n therestof Europe.Moreover,successive oliticalleadershipmovementshereaftergeneratedhangeswithin thepropertysystemand, herefore,nallotment systems ndploughingpatterns.Modemdevelopment,andespecially hedeliberatepolicyof heavy ndustrialisationof thecountry mainly after the SecondWorld War, also producedgreatdisruption n landscapeOltean2002,224).The transformationof the agriculturalprocessnto acommunistsystem,which startedn 1949and ended n 1962,wasbasedon thecreationof newagriculturalunits ownedandexploitedcollectivelyby groupsof individuals (CAP)or the state IAS)

7/29/2019 20004olteanphd

13/274

(Constantinescu.tal. 1970,386). In fact, this involved on the onehand herecognised xpropriationof largerprivateestates y law (The Nationalisationof the Main ProductionResources ct of IIJune 1948),and on theotherthe persuasion'of thepeasantso gather heir propertiesnto CAPs.Asa resultnew field boundarieswereestablished.Within the 'plannedeconomy'system,ntensiveexploitationreplacedhetraditionalextensiveapproacho agriculturalproduction,which meant hatarablecultivationcovered63%of thesurfaceof Romania Constantinescuet aL 1970,410)and hatdeepploughing echniques,de-forestation,e-forestationsandthe conversionof marshesnto arablelandswereall takingplace.On theotherhand, hemain focusof the 'plannedeconomy'wastheacceleratedndustrialisationof thecountryside,with a specialpreferenceor heavyor chemicalindustrieswhich registereda dramatic ncrease n therateof development rom only 4% in 1938 o13.4%since1960 Constantinescuet aL 1970,410). Massive ndustrial installationssuch asHunedoara,Calan,Mintia, OrastieandOcnaMuresemergedn the landscapeseechapter2 andfigure 2.3). Artificial lakescoveredmodemsettlements s well asarchaeological ites.The biggestexample s the Iron Gates on theDanube,but within my researcharea hereare similar examplesata smallerscaleat Cincis (seeFlocaandValea 1965)OstrovandHateg(Popa1989;see igure 2.4).The road,gasand electricitynetworksalsodevelopedon a largescale.All thesecombined actorshave mpactedon thevisibility of siteson thegroundor evendestroyedhemcompletely.NowRomania s goingthrougha newphaseof massivedevelopmentespeciallyn transport acilities,withthe constructionof motorwaysandthedevelopmentof the current roadnetwork,and ndustrialexploitation.Themostfamousexampleof the atter is thegold exploitationconcessionsn theApuseniMountains hatwill completely emoveseveralarchaeologicalsitesof variousdates,including theRomansmall town andgold minesatAlburnus Maior (modemRosiaMontana).Therefore,a study of thesettlementpatternsof RomanDacia from a landscape erspectives nowrequiredbeforemodemdevelopmenthasanirreversibleaffect on boththenatural andhistoricallandscape f thecountry.

2. Biasesand limitations of current research:Becauseresearch asbeen ocusedmainly on military and urban sites,all the resthavebeenseriouslyneglected,even allowing for thebrief attention given to rural life particularly in the 1960's.The limited numberof specialistsand hefundsavailablehave beenconcentrated n themorepromising and spectacular ites,which havegenerallynot includedrural sites(Alicu 1998,127-128).As a resultthereare numerousbiaseshatapplynot only to the quantity,but mainly to thequalityofcurrent nformation.Therefore,sinceonebiascanbea direct consequencef another,heendresultis that thetheoriesgeneratedrom thedatacannotbeotherthan ill-founded.

2.1. The nfluence of histoty andpolifics on archaeologicalresearch.First of all, I shall consider hemannern which archaeologicalevidences currently perceivedandacceptedn Romania.Archaeological esearch asbeensubserviento established istoricaltheoriesgenerated y literary sources.This attitude s deeplyrooted n conceptsof thepastheldby modem

7/29/2019 20004olteanphd

14/274

individuals,andespeciallyprofessionalscholars.Theultimate goalof archaeologicalesearchs, ofcourse,directed owardsexplaininghistoricalevolutionsandphenomena.Unfortunately, hevisibletendencyoverthepastcenturyhasbeen o rely primarily on theexisting literary sourceswithouttaking into considerationheir inherentbiaseswhoseeffectcould potentiallybe reducedby referenceto the totality of theevidence.The fact thatarchaeological videncehasbeenconsideredonly whenit supportedhehistorians'arguments,ather hanattempting o build up a theorybasedalso on thearchaeologicalevidence,s a generalproblem n the study of Roman Empire.A greatdanger n sucha situation s that the iterary sourcesalone end o generatehe conclusions.Consequently,hemostrecent comprehensive tudieshave ried to addresshis issue.In Dacia this situationmight be explainedsometimes y thepaucity of archaeologicalnformation,but unfortunately,evenwhere hatevidences available,otherfactorsdistort its considerationandresultingtheconclusions.There s nothingnew n therecognitionof the importance hathistoricalmodelshave or political discoursen general,but for a long time in Romaniahistory itself wasentirely subordinatedo politics. As a result various subjectsof archaeologicalesearchwereapproached ndfundedonly whentheywereseen o serve hepolitical fashionof themoment. Thisattitudewasvisible in Romaniaespeciallyafterthe SecondWorld War, when politics becamemoreintrusively and, ndeed,aggressivelynvolved in differentaspectsof researchas with life in general.Above all, theway of thinking andwriting historyhadto beMarxist (basedon the theoryofhistoricalmaterialism),andno otherapproachwasperceivedas suitable'. Eversince1947,whenthecommunist epublic wasformed,andespeciallyn the 1950sand 1960s-during the eadershipofGh.Gheorgiu-Dej-, hekey role of historicalresearchwasto feedthediscourseof communisttheories egarding hesocial classes'antagonisms ndtherejection of westernmperialism.In thiscontext researchat thema or Romansites, ncluding Wpia TraidnaSannizegetusa,heprovincialcapital, ceasedo be financed,andboth fundsandpersonnelwerere-directed o undertake esearchon nativesettlements f pre-Roman,Romanor post-Romandate.On the onehandsucha decisionhadapositivevalue,giventhe fact that it determinedthere-direction of researchowardssites hathadnot beenapproached efore,or not adequately, ndaddressed biasof previousresearch.But onthe othertheconclusionsof theresearch adto fit thegeneral heoryof theoppressionexperiencedby the native masses nderthe imperialistoccupation.This idea,bom in themiddle of the industrialera,hadto besustained y history througharchaeologicalestimoniesandprovenot ust that itexistedfor a long thetime, but also hat socialantagonisms xisted oo. This is themomentwhentheRomanconquestandoccupationof Daciawasdescribed,at least n part, asa negativeeventboth inrelation to the treatmentapplied o thenatives,but alsobecauseof the introduction nto theconquerederritory of a socio-economic ystembasedon slavesandthe exploitationof the lowerclassesConstantinescuel al 1970,53).Accordingto thesescholars,who were projecting deasusually connectedwith theconceptof modemnations nto the contextof anancientsociety, henatives ost their liberty/sovereigntyas a peopleand heir properties n favourof their Romanconquerors.Theopinionexpressedy M. Macrea 1969,457) is that "the Romanconquestbroughtfor Dacianotjust economic,socialandculturalprogress, ut alsoanentirecortegeof sufferance,pillage and unmercifulexploitationfor theslavesand oweststrata,all inherent o an antagonisticsocietybasedon slaveexploitation"(my translation).Similarly, the authorsof 'Inscriptiile Daciei S

7/29/2019 20004olteanphd

15/274

Romane'(Russuet aL 1975,22)wrotein their historical introduction: "Ibe Romanoccupationbrought[ ... ] realprogressn modesof organisation,materialand spiritual culture,technologyandproduction, and n thewayof life; but it was at thesame ime, a regimeof pillage androughexploitationof the ower,working classes f thesociety andespeciallyof themasses f nativesubjugatednhabitantsandworkers,expropriatedn their own country" (my translation).The othermain political andhistoricalthemewas heconcern o producehistoricalargumentsosupport hepolitical actthroughwhichTransylvaniabecamea part of Romania n December1918.This theme s recurrentbeforeandafter 1918,and s thesubjectof a long-termdebate n RomanianandHungarianhistoriographywhich remainsunresolved the Romanianpoint of view is mainlysummarisedn Daicoviciu el al., 1963,while theHungarian hesis s presentedmostrecentlybyVikony 2000).During thecommunistyears hesubjectwasparticularly in fashion n the 1970s-1980s,duringtherule of N. Ceausescu, hencommunistpropagandan Romaniagaineda strongnationalistmessage.Within this context one of themainconcernsof Romanianhistorianswas obring forwardargumentsor therapidanddurable omanisationof thenatives n Daciaasmajorelement n the contextof the 'theory of Dacian-Roman ontinuity'. This thesiswasdevelopedduringthepast3 centuriesmainly asa responseo thecorrespondingHungarianpropagandawhich attemptsto sustain ts political position towardsTransylvaniaandtries to proveon a scientificbasis hat thisterritory wasdevoidOf anypopulationon thearrival of theHungarian ribes. t is argued hat theRomaniansemergedn theregion througha massivemovementof population rom southof theDanube ateron,becauseheDacianshad disappeareds a resultof thewarsof AD 101-102andAD105-106andtheRomanshadwithdrawnall thepopulation n the YdcenturyAD. Therefore,Romanianscholarsocusedon disprovingthis thesis,bringing fowardargumentso support heromanisationof thenativeDacians;underRomanrule and he continuity of life in Transylvania romprehistoryto theMiddle Age.The currentlyacceptedheorieson theRomanconquestand rule of Daciaandmostof all, on theromanisationprocess,hadto be fitted in accordingly, n spite of several lagrantcontradictionshatIwish to address riefly here.The archaeological videnceso far attestshepresenceof colonisingelementsn both urban and rural contexts.But if in theurbanand military sitesfunerary,epigraphicand othertypesof evidenceseemso indicatequitea largepopulation with anorigin otherthanindigenous seealsoCiongradi2004a andb), thecurrent evel of archaeologicaldata or the ruralareasof RomanDaciamakes t impossibleat thispoint to asserta similar percentage f colonisedelementshere.Thedestructionof theDacianhill forts andthemovementof thepopulationhasbeenacceptedwith little if anysurprise,giventhehugeeffort of the Romans o conquer hesesites.But ona closerexamination,current heorieson thenatureof theconquestandthe treatmentapplied o thenativesare n blatantcontradiction.Oneof themsuggestshat thenatives ived on thetopsof themountainsand wereforced o settle he fertile lowlands,while anotherstateshat theRomanscameandtook the fertile lands or their own propertiesandforcedthenatives o move awayor work ontheir propertiesas cheapabour.Unfortunately, t is often forgottenthat thesole reason or theexistenceof thehill forts is aselite sitesand hepurposeof their location is strategicandstatus-related.Accordingly,after themilitary defeatand he introductionof Romanrule, suchsitesno

9

7/29/2019 20004olteanphd

16/274

longerhada reasono exist.The extrapolationof this modelof hill fort destruction o theentiretyofthe Daciansettlementpatternnot only pusheshis interpretation o anunsubstantiated xtreme,butalso would havebeena questionable olicy decisionon thepart of theRomans.The idea hat theRomancolonistsdeprivedall thenativesof their lands doesnot sit well with thebenefitsof a rapidanddurableromanisationhroughcloseandpeaceful elationshipsbetweenall the inhabitantsof thenew province,whethernativesor colonists,civilians or soldiers. If true,theresultingattitude of theDacians owards heir conquerorss likely to havebeenresistanceo acculturation, atherthanreceptiveness.t certainlydoesnot take nto considerationotherfactors- culturalor economic, orexample which musthave nfluencedattitudesand he whole processneedsamorerefinedanddetailed nterpretation.Therefore, t is no surprise hat the theoryhasbeenchallengedn the 1990snot ust by theHungariancolleagues,but alsoby someRomanianscholars,especiallyarchaeologists.t is not my intentionhereto analysehispolemic in moredetail, but to show hat it has nfluenced heperceptionevenof theexistingevidenceand n theend t has distorted hegeneralview of RomanDacia n termsofconquest,colonisation,administrationand, ndeed,romanisation.The eventual eactionof someRomanianarchaeologistso this theory s an attempt o evaluatehe existingarchaeologicalevidenceat its truevalue, without dismissing heelementshatmight offer adifferentpicturethan theonedesiredby political discourse.t would bewrongto acceptany form of political interferencen eitherarchaeological r historicalpracticeanylonger.Thepolitical attempts o influence nterpretationsofthepast usuallyreflect a failure to find solutions o presentssues a distracting actor from realpolitical issues).Moreover,addinga supplementary iasto those hat currentlyapply toarchaeological esearchwould distortour perceptionof thepastevenmore.

2.1 Reflability of and accesso infomation.Reliableevidences surprisingly imited. This statementmight soundoddgiventhe amountof datarecordedsofar for RomanDacia, but unfortunately or themostpart the informationprovides onlythe generalocationof archaeological iscoveries.Many effortshavebeenmade o identify andsystematically esearchhearchaeological itesof Transylvanian 20'hcenturyand n somecasesevenearlier,especially hroughfield walking andexcavationsat various scales.Both TIR L-34(Budapest)andTIR L-3 5 (Bucharest),along with D. Tudor's book (1968)andtheseveralcountyarchaeological azetteers,efer to a significant numberof sites.Unfortunately,thevast majority ofthesesiteshavesomewhatstereotypicaldescriptions,most of the time mentioningceramicfragments,sometimesassociatedwith constructionmaterials. n fact, thesestereotypes elate o themethodsof collection,interpretationand, ndeed,evaluationof the data or thewhole territory. Thereis still a lackof systematiccoverageof theterritoryandno unitary methodof recording.Only some10%of reportedsiteshavebeen he subjectof more extensiveexcavationprojects.The remaining90% areust indicatedby finds (artefactsor buildingmaterialsscatteredon the groundsurface).Thereare no site plansavailable or themand, herefore, heir sizeandsignificancehasnot been ullyappreciated ecause f a failure to applymodem echniquesof siteprospection.This issuewill bediscussedn moredetail lateron in thischapter,but it is responsiblealsofor the fact that seveml 10

7/29/2019 20004olteanphd

17/274

categoriesof sitescharacteristicof a landscapearestill overlooked.Potentiallythis couldbetheexplanation,or instance,or the fact thatno Iron Age, or indeedRoman, and-usesystemshavebeenfound in Dacia.But even he informationthat is available s extremelydifficult to considerandevaluateas a whole,in theabsence f anaccessible ationalsystemof record of archaeologicalsites.Dealingwith a typeof informationthat is, by its nature,accumulated ver a longperiod of time, naturallybringsproblemsof storageand accessibilityof data. n Romania, he systemof publicationof theresults snot veryhelpfid in termsof access.TIR L-34 andL-35, andTudor (1968)arewell out of date.Morerecentdiscoveriesare accessed rimarily throughsporadic, andom publication n variousRomanianperiodicals.Thematerialmight have beenaddressedn a few casesn more generalstudies,whichaim to collect thedatadiscoveredover a wider area,usuallyin termsof historicalgeography,chronologyor specific categoriesof sites(e.g. Wollmann 1996 or mining and quarrying;Popa 1989for TaraHategului).Therearealsofew casesof modem regionalarchaeologicalgazetteerse.g. AlbaCounty-MogaandCiugudean1995).Thereare no regularlyupdatedarchaeological atabases.he recentefforts to completea generalarchaeological atabaseby cIMeC -seeabove)are extremely mportantandthevalueof suchactionhasbeenprovedalreadyby theavailability of informationon-line under he formatof a searchabledatabaserom excavationsrom the years1983-1997,1999and2000.Theuseof theInternetenvironment or information,aswell as heavailabilityof digital data o supportcomputerisedquantitativeand settlementpatternanalysis,will providea valuablesupport or futureresearch.Unfortunately,even his latestattempt o preservearchaeologicalnformationdoesnot includetheprecisegeographicalocationsof the sites,continuing o use ocationdescriptionsby place-nameswhich havealreadyproved o be inadequateseebelow).

2.3. Site location.The experience f the ast 100yearsshows he importanceof theaccurateransmissionofinformation in thecontextof successive hangesn archaeologicalmethodological equirementsoreven n the territorial administrationsystemandplace-names. he failure to locatearchaeologicalsitesby their geographical o-ordinatesandrelianceon place-names roducessignificantdifficultiesin attempting o locatesomesites hathavebeenpreviouslyreported.This occursespeciallywithplacenamesof very local significancewithin theareaof aparticularvillage, for example,which arenot in useanymore,nor traceablehrougharchivemapsor documents.As a directconsequencefthis failure to locatearchaeological itesprecisely,someof themarevery imprecisely ocatedwhenreferred o by variousauthors.This confusionpersistseven n the mostrecentpublications.For example,N. Gudea n his studyontheRomanmilitary campsn Dacia (1997,101-2) ocates heRoman ort of Cigmau(Germisara)and ts civilian settlementapproximatelyonekilometreto theN of its trueposition(HansonandOltean2002,114).Theprocessof alterationof informationis visible in thecaseof D. Benea'sarticleon military vici from Dacia(2000). Acknowledging hedifficulty of accesso information(see

II

7/29/2019 20004olteanphd

18/274

above),she ried to assemble ll the dataavailable or civilian settlementsoutsideRomanauxiliaryforts in Dacia,and husproducea useful tool for bothRomanianand nternationalarchaeologists.Shemakesuse,amongstothers,of thepublishednformation from therescueexcavationsatVetel(Micia) (Marghitan1970).Unfortunately, he ocationof thebuilding complexesn questionsmisplacedby some250metres o theEastand at least50 metres o thenorth.Sucha mistakehighlights thedifficulty of assemblingdata,especially or thesites whereanoverall site planhasyetnot beenproduced.As a resultaccuratearchaeologicalmapsof largerterritoriesnecessaryo supportlandscapestudiesand settlementpatternanalysisare argelymissing.

2.4. The quality ofexcavation methodology.The excavationmethodsapplied o theRomansitesof Daciaoverthe lastcenturyhavevaried.Asshownby CondurachiandDaicoviciu "after thesecondworld war thedistinctivefeaturesofRomanianarchaeologicalmethodweretheabsolutepriority it gave o stratigraphyand ts preferencefor the exhaustiveexcavationof largesites o themaximum extent hat thecircumstancesermitted"(1971,20). Unfortunately heseprincipleswerenot applied o Romansitesuntil much later.This factis well illustratedby the list thatthey give of siteswhere modemmethodshad beenappliedby thedateof theirpublication,which containsnot a singleRomanexample,andcanbe furtherconfirmedby consultingotherexcavation eports.Thus, for someof theexcavations,especiallyof villa sites, hepublishedresultsgive little indicationof constructionphases.This is mostunusual,especiallysince repairsOrchanges f plan withinbuildings,or evenchanges f useof buildings,arefrequentlyrecorded n civilian archaeologicalcontextselsewhere.At Hobita,despite hefact that is sofar theonly exampleof a villa siteto havethewhole enclosedareadelimitedand ts internalarrangementsefined, here s no indicationof anyphasesof constructionandrepair (Floca 1953). Publishedexcavation eportsrarely expressanyconcernabout heir failure to identify earlier phases f archaeological omplexes.n the caseof atleast 2 civilian buildings (no. I andno. 3) in thevicus excavated y MarghitanatMicia some30years ago(1970),earliertimberphaseswererevealedby excavation.Despite hefact that therewereother indications,suchastheexistenceof artefacts,ncludingconstructionmaterials tiles),withinthe filling layersunder he floors, thesediscoveriesoccurredonly incidentallywhile theexcavatorswere trying to reach hedeepestevel of thestonewall foundationswhenthe earlierconstructiontrenchesor the timberwalls intersectedheir trenches.However, he excavatorsmadeno attempt oestablish heextent and ayoutof these eatures,o considervariationsof planor internalspacedivision from one phase o another,or to make a studyof thequantitativeand qualitativeor stylisticevolutionof thedifferent categoriesof finds.Excavationssuchasthesehave, herefore,producedincompletesiteplans wherechronologicaldevelopmentsarenow impossible o pursue.Similarly, incomplexeswhere he stonephasewent throughsuccessiveransformationsor repairs, heseareapparentneither n thereports,nor siteplans f available.This issuewill beconsideredn moredepthlater,whendealingwith specificclassesof sites,but it is worth statingfor themoment hat in suchcaseshe interpretationof the nternalarrangementss almost mpossible o establishaccurately.

12

7/29/2019 20004olteanphd

19/274

2.5. The imited applicationof archaeologicalprospectionmethods.Accordingto Bewley andRaczkowski, Excavation s still synonymouswith archaeologyn manycountries"(2002b, 3) and hathasbeenvery muchthecase n Romaniauntil very recently. t is truethatexcavationcangive themostpreciseanddetailed nformation aboutarchaeological ites,but notevery site canbeexcavated.This is not ust dueto thehigh costs nvolved or theamountsof timenecessaryo completeanexcavation.Excavations,by its nature,an invasivemethodof researchwhich damageshe site tself in theprocess.Theareasaffectedby excavationwill no longerrepresentan ntacttestimonyof thepast.Excavationstandards volve over time, and evenwhat isperceivedat this moment o be technicallyfirst rate, couldbeconsiderednsufficientor inappropriatein future.Therefore, hemodemapproacho archaeological eritageconservations focusedon non-invasivemethodsof research, nd mostarchaeological ites end to beexcavatedonly whentheyareendangeredy development.Within this context,modem methodsof archaeological rospectionhave developedas a necessity, oth for understanding nd monitoring archaeological ites,butalsofor prior evaluationof thesiteand ts potential as accuratelyas possible.But perhapshemajorcontributionof modem methodsof prospection,and especiallyof aerialphotography,s thepossibility of providingbetterunderstanding f sites rom a landscape erspective,and ndicatingtheway thathumansettlementsnteractedwith thenatural andscapeandwith eachother.Traditional field walking is theonly methodof archaeological rospectionwhich hasbeenappliedhistorically in thestudyof RomanDacia(seeabove).Unfortunately, in its applicationno attentionhasbeengivento theunitaryplanningof the field coverage,or to systematicdatacollectionandanalysis.Only very recentlyhas heApulurn HinterlandProject nternational eamstartedsystematicfield walking in thevicinity of Alba lulia in collaborationwith theUniversity of Alba Iulia(informationDr. Ian Haynes).But evenat its best, ield walking alonecangive only limited cluesasto thenatureof thesite.Whatcanmake hedifference s understandinghe site's full extentandmorphology,allowing thedifferentiationbetweenanindividual homesteadfarm) or a nucleatedsettlement,or example. n thecaseof individual homesteads,he layout of the internalbuildings,their individual plan, or thebuilding materialsusedcandistinguishbetweena villa siteanda nativefarm. Field walking alone can ocatea site, provide somendicationof its extentand, rominterpretationof finds,evaluatets chronology or offer suppositionsabout ts nature. t doesnotprovidea clear deaof themorphologyand ayoutof thesite, which canbeknown only from theinterpretationof aerial photographs r geophysicaldata.Geophysicalsurveyhasonly recentlystarted o beapplied o Dacia.The usual ackof funds, rainedspecialistsor surveyingequipmentare probablythemain reasonsor this, asfor aerial archaeology(seebelow).As in thecaseof excavations,hefirst attemptsat geophysicalsurveyhave ocusedonurbanandmilitary sites.Some esults rom a geophysicalsurveyfinancedby TheCountyMuseumofHistory andArt of Zalau n 1996,which applied magnetometryandresistivity to a 2.2hectarearea othe south of theRoman ort atPorolissum,wererecentlypublished(Scurtu 1997).The samespecialisthasconductedmore imited surveysat Cigmau ort (information from Dr. A. andMr. E.Pescaru).At Apulum, geophysicalsurvey n thecoloniaandthemunicipiumby Lockyear(information from Dr. 1.HaynesandDr. A. Diaconescu,heApulurn Project)hasrecentlystartedo

13

7/29/2019 20004olteanphd

20/274

extend nto theterritory around heRoman own with surveysat villa sitesat SeusaandOardawithinthenew frameof theApulurnHinterland Project. see igure 5.5). But geophysicalsurvey,whetherutilising magnetometry,esistivityor groundpenetrating adar, s by its nature argelyconfined o thelimits of individual sites.Despite ts recognised alue n recordingdetailsof site layout, t offersonlylimited opportunitieso evaluatehesitefrom a landscapeperspectiveandconsiderotherpossibleadjacent eatureshatmight berelated.The aerialview giveshumanperceptionabroaderperspective.Archaeologicalsitescanberecognisedevenwhentheir degreeof preservations very poor, whetherstill visible to someextenton theground surfaceor even otally buried.Overthe60 yearsor more it hasbeenproved onnumerousoccasionshat,givensuitablesoil conditions,buriedarchaeologicaleaturescanberecognisedrom theair ascropmarks(Wilson 2000,16-23, Bewley 2002).This hasmadeaerialreconnaissancextremelyvaluable,especially or the identificationof previouslyunknownarchaeologicaleatures.Furthermore,a trained nterpretercanacquireconsiderablenformationabouta site,bothin termsof its morphologyand ts probabledate, hroughanalogieswith similarsiteswhosechronologyhasbeenestablished y other methods Wilson, 2000,65-67,84-87).Also,the speedof coverageandconsequently f analysisof even arge erritoriesis significantlyhigherthanthroughfield walking, or indeedgeophysicalmethods.Both thesecharacteristicsmakeaerialphotographyhepreferredmethodof archaeological rospectionn Europe,especially or landscaperesearchandmanagement. heseadvantagesave determined he initiation of programmesnvolvingaerialreconnaissanceo acquirenew imageryandevaluationof available mages rom aerialphotographicarchivesn severalcountriesof Europeat national scaleandon permanentbasis seeBewley andRaczkowski,2002b,Fig. 1).Most recently,satellitephotographic,multi- andhyper-spectral magery,or airbornescanningechniques avebeenaddressedoo. Therefore,modemsurvey of a givenarea or its hiddenarchaeology hould ncludeat least hestudy of verticaloroblique aerialphotographs, eophysicalsurveyandfield walking.Probably he first aerialsurveyandphotographyof an archaeological ite in Romania ook place asearlyasMarch 1918,whenCarl Schuchhardtook aerialphotographsof theRoman rontier wall inDobrudja. Despite he fact that thesephotographsemainedunpublisheduntil 1954 Crawford,1954,208and plateVI), Schuchhardt sed hem to correcthis own publishedmap of thewall.Unfortunately,this remainedvery muchtheonly exampleof its kind for a long time. Theonly aerialphotographicsurveyprogrammes verthe following decadesweremade or differentreasons,whethermilitary duringtheworld wars and after,or civilian for mappingpurpose.Access o theseaerialphotographicarchives or archaeologistsemained imited for a long time, thougha startwasmade n utilising themfor archaeological urposesmainly on Romanmilitary sites(e.g. Bogdan-Cataniciu.1981), but concentratingn mostcases n extantor known archaeologicaleatures. nRomania hewholeprocesshasbeenconstrained y legaldifficulties, lack of financial resources nda shortageof specialistsexperiencedn the interpretationof aerialphotographsseeBraasch2002;Oltean2002;HansonandOltean 2002; HansonandOltean2003).Until the last decade,CentralandEasternEuropewas houghtnot to besufficiently responsiveo cropmark formation because f theheavily alluvial topsoil thatcovers he arable egionscombinedwith a lesscontrastingdeeper

14

7/29/2019 20004olteanphd

21/274

geologicalbackgroundBradford 1957,15and23).Moreover, therecognitionof crop marks wouldhave beenmoredifficult from archivevertical imagery.Apart from thevariablequality of thephotographsand heir scale,suchdata s likely to be of more limited archaeological aluebecausethe imageswere obtainedoriginally for purposes therthanarchaeologyandthevery specificconditionsnecessaryo recordmanyarchaeologicaleaturesmay not havebeen n force.Therehavebeena few attemptso makeuseof aerialphotographic nformation in archaeologicalstudiesand n theoccasional ublicationof sites,but without a clearly structuredmethod ofapproach.Unfortunately, n all of thesecaseshe nformation availablewas nsufficiently exploitedbecauseof the limited trainingandexpertiseof thesepioneers.They were either archaeologistswithvery little or no experiencen mappingor photo-interpretation,or topographicand geographicphoto-interpreterswithout archaeologicalexpertise.A group of specialists rom theRomaniancartographicinstituteof Bucharestbelong o the lattercategory.Theypublishedshortarticlesutilising principlesof photo-interpretationo identify possiblearchaeological itesat Sanislau SatuMare) andDersida(Salaj) in north-westernRomania,or at SarighioldeDeal, SatuNou andIsaccea.longtheDanube nthe south-eastern omania,where heyprovide nterpretationandmappingof thearchaeologicalfeatures Rada,CochinaandManea1989,RadaandCochina*, Rada,CochinaandCorcodel**).Unfortunately, heir archaeologicalandevenphotographicnterpretationof thevertical imagess notalwayscorrect.Also, in someof thecasesheyseem o focuson the identificationof thearchaeologicaleaturesobservedwhich in most of the casess hypothetical),rather hanon thedescriptionof more methodologicalssuesof archaeological ite recognitionand generalprinciplesof photographicnterpretation.n thecaseof thesmallBronzeAge siteenclosed y a systemofmultiple ditchesand rampartsat Sanislau,or example, he interpretersailed to recogniseheexistenceof ditchesn front of theramparts,aswell as a fourth externalditch with its rampart Rada,CochinaandManea1989,203andphoto 1). However, heir effort is still of valuebecauseof theattempt o identify previously unknownarchaeological ites and, n somecases, roduceinterpretationsof features n theform of maps. n herstudy of the Romandefences f Dacia, .Bogdan-Cataniciu ublishedsomeexamplesof extant military enclosures isible on existing archivevertical photographs1981).But, as wasthecasewith thematerialpublishedby Rada,CochinaManeaor Corcodel, hequality of thereproductionsor in somecaseseven of theoriginalphotographss so poor that thereadermust rely for themostpart on the interpretationsprovidedbythe author.Archaeologistshaveusedaerialphotographs poradically o illustrate ectures e.g. C.Craciunmentionedn Ardevan1998,76),publicationsof sites(e.g. Alicu 1998,plate3; Tamba1997,plate8) or as exhibition illustrations.Othershaveundertakenoccasional imited flightsphotographingknownsites rom theair for illustrationpurpose,suchasG. Stefan n the SouthEastof Romaniaor E. Pescaru,n Hunedoara ounty, he latter usinga helicopteras an aerialplatform.Lessfortunateenthusiasts,uchasV. Barbutawereconstrainedby lack of funds o makephotographsusingkitesasaerialplatforms(figure 1-7),with somesuccess specially or theidentificationof the templeof Liber Pater at,4pulum information Dr. A. Diaconescu).Becauseof restrictionson civilian air traffic in forceduring the Communistyears, he inability to fundaerial reconnaissancerogrammes, nd he ackof relevantequipment,experienceandexpertiseo

15

7/29/2019 20004olteanphd

22/274

undertakesuchwork, theseeffortswere no morethansporadic.However, in 1998 heUniversityofGlasgowstartedaprogramof systematicaerialreconnaissancen the mid-MuresRiver valley andTaraHateguluiundertaken y Prof W. S. Hanson, undedby the LeverhulmeTrust andsubsequently y theBritish Academy.Thework wasundertakenn co-operationwith theNationalMuseumof Transylvaniain Cluj, theMuseumof DacianandRomanCivilizations in Deva andtheUniversity of Alba Iulia. Thepurposeof theprojectwas o establishheparametersor the applicationof aerialreconnaissancen thedifferent environmental, oil andagriculturalconditionspertaining n Romaniaandalso o increaseunderstanding f thehistoryanddevelopment f the andscapeof westernTransylvania,particularlyfrom laterprehistory o the immediatepost-Roman eriod.The geographicalocusof theproject wasSouth-WesternTransylvania,particularlythemiddle andupperMuresvalley andtheplain of Hateg othe south,which liesat theheartof both theIron Age kingdomandthesubsequentRomanprovinceofDacia. This projecthasestablishedhefirst relationaldatabase f aerialarchaeological itesof differentdateandhasprovidedvaluable nformation for this thesis(seebelow). However, t is generallyrecognisedhat thebestresultsare obtained hrougha combinationof thesevariousmethodsof survey,andhopefully future financingwill finally allow such a programme.

3. Aims and approaches: andscapeand romanisation.The relevanceof ancient andscapestudiesn finding theanswer o general, heoreticalor evenmorefocused ssuesof theancientworld hasbeenbroadlyrecognised.Datainterpretationhasalwayspre-supposed n evaluationof thedistributionof similar sites.But understanding f thegeneralpatternsor, indeed,unique eaturesrevealedby thesitesmustbeconsideredrom abroaderperspective."Archaeologicalmapsof settlementdistributioncan ell us much more thanwhere hesitesare.Theycanbereadassummaries f a whole complex of ecological consequences,ynamicand argelypredictable.Takingeachpoint asrepresenting vortex of socialand economicactivities, t is readilyapparenthow thecumulativeeffectsof thatactivity over time canleadto more or lesspermanentenvironmentalchanges"Delano-Smith1996,174).The landscape uffers a continuousevolutionwhich, from an environmentalist iew, wouldberegardedas a processof degradation.Someof it iscausedby naturalerosion,but themostof it is, in fact, dueto humanor animalactivity. Amongstallspecies,humanshavebeen hemostsuccessfuln usingthenaturalhabitat.Moreover,humanshaveadaptedhathabitat o create he landscape, sing ts resources ndtransforming t according o theirnecessities. rom thispoint of view, the landscape ears hemarks of thehumans hathave nhabitedit throughtime, reflectingtheir needsand way of life, or their level of technologicalknowledge.Apart from active ecologicaldestructioneading o moreor lesspermanentenvironmentalchangescausedby sitesoccupiedby man(Delano-Smith 1996,167-8),the developmentof landscapess ableto tell a differentkind of story,thatof thecultural evolutionof humancommunities hrough ime.The natural environmenthasbrought nto theequation actorssuchastopographyandclimate, heavailability of resourcesor life (either beneathheearthor on its surface,as ocal flora and auna)or, indeed, hedefensivepotentialof particular ocations.To this, humanshaveaddedheir ownleadingmentality,needsand echnological esources,nowledge,socialregulations,politics,which

16

7/29/2019 20004olteanphd

23/274

has mpactedon theuse,and herefore, e-shapingof the andscape.The approaches resented boveareacceptinghe andscapetself as anobjective reality transformedby the subjectivityof man.But, whatever hestudyapproach,t is acceptedhat the interactionbetweenhumansand heenvironmenteventually esultedn a transformationof bothparts, aprocessthatwould have eft significanttraces,dentifiable by modemarchaeologicalmethods.Whenreadingthese races,however, t is possible o go Rirther han ust an ecological mpactevaluationof pasthumanactivity. Thenumberof humans hatwereaccommodatedwithin a given landscapeateachmoment n time is another actorthatwould have nfluenced he amount of physicalspaceaffected.Going beyond hephysicalboundariesof eachsite, hestudy of settlementpatternss far from beingsolely a mapping or locationalprocess.t would normally involve analysing hespatial,along withthe functional,relationshipsof contemporary iteswithin particular cultures(Knapp 1997,5). Bothof thesedimensions,spatialandfunctional,are not to beunderstood,however, n a strict economicsense,but also n political, social,religious,or culturalterms(KnappandAshmore1999,vanDommelen,1999).The evolutionof landscapesoffersa differentunderstanding f thoseancientMediterraneancivilisations thatwere argelytown-based, uchas hoseof GreeceandRome,both in theirMediterraneanheartland Shipley,1996,8) andbeyond e.g. Dark andDark, 1997).Apart from theobviouseffect of revealingwhatprovincial settlementswould have ookedlike, thestudyof thesettlementpatternwithin Romanprovincescanaddressmore general ssues.Thedecision o settleand useaparticular spacewasbeingtakenby people, n groupsor as ndividuals,in direct relation otheir interests.Consequently, y studyingtheresulting mpactof their actionon the landscape,onecanpresumably ell whether heoriginal effort involvedwas madeby several ndividualsor by anorganisedgroupfollowing a certainpolicy. Therefore,henatureof thecolonisationprocesscanbeanalysed rom theway thenewRoman-typesettlements mergedwithin theprovincial territoryandtheir effect on thepreviousnativepattern.Accordingto thecurrentorthodoxy,aftertheRomanconquestDaciaexperiencedhe first large nflux of populations rom outside ts culturalboundaries,a phenomenon escribedby ancienthistoriansandre-enforcedby the epigraphic evidence.Thesenewcomers,mostly from other partsof theRomanworld rather hanRome tself, whethergrantedRomancitizenshipor not,hadto beaccommodated ithin Dacianterritory, asdid themanifestationsof the new legaland administrativesystemand hemilitary. The native settlershadto comply withthesituation.Subjectof debateor decades,heapproacho romanisationhasbeenmarkedby severalsuccessivetheoretical rends materialism,colonialism, post-colonialism),all trying to find a satisfactoryexplanation or an extremelycomplexsocio-culturalphenomenon.Traditionally, thetwo partiesinvolved, thenativesandtheir conquerors, avebeenpresentedasfacing eachotherfrom different,sometimesevenconflicting or antagonisticpositions, eflecting of modempolitical (national) houghton ancientsocieties.But romanisationstill givesunexplainablydifferent, even contradictorypicturesnot ust when subjectedo differenttheoreticalapproaches,ut alsowhen seen rom different comersof theRomanEmpire.Indeed,one of the mainquestionsn defining romanisationas a processwhethertheRomans omanisedheprovinces e.g. GamscyandWhittaker 1978)or thenatives

17

7/29/2019 20004olteanphd

24/274

romanised hemselvese.g. Millett 1990)-if answered, as ound different responses.On theonehand, his makesa globalunderstanding f theprocessvery difficult and,therefore, esearch nromanisationhas ended o remainat a levelof local studies.On the other, it hasresulted n extremeattitudesand sometimests natureor very existence,n particularregions(e.g. Africa -see

B6nabou1976; Dacia-e.g. V6kony2000)or as a concept e.g. Hingley 1996;Barrett 1997),hasbeenchallenged.However, Woolf (1995)rejects deas ike conflict, competitionor interactionandadvancesa newinterpretative i=ework, with thecreation of a new imperial cultureasa structuredsystemof differencesreplacingbothpreviouscultures Romanand native), its spreadcomparablewith the growth of anorganism hatmetabolisesmatter.A similar approach o romanisation,as"alargely consciousprocessby which sectionsof the indigenouspopulation sought o emulateRomanculture,at least n the form in which theyexperiencedt, motivatedby theneed o establishheir ownsocial statusanddirectly assisted y theRomanauthorities" wasearlier employedby Hanson 1994)in his own interpretationof thephenomenonn British context.This understanding f romanisationcombinesbest he theoreticalandempirical approacheso thesubjectand s theone hathasbeenemployed hroughout his study.Romanisationas consideredhere s theway thatRoman ule affectednot ust thenativepopulations,but alsothewhole landscapen theconquerederritories. Ibe emergence f Roman-typeowns, hebroaddiversificationof therangeandfunctionof settlements, ndtheparticularwayof organisingspaceprobablyhadamoresignificant impacton thepre-existing system hanany otherpreviouschangesduringprehistoryandcanreveal, n comparisonwith other provincesof theempire,particularaspectsof theromanisationprocess,aswell as giving thereal scaleof thewholeprocesswithin the territory. By combiningarchaeologicalnformation,historicaldataand nformationregarding henatural andscape,t is possible o understand etterthegeneralevolutionof thelandscapeandthehuman mpactupon t, both in thepre-RomanandRomanperiods. In such acontext, t shouldbepossible o distinguish rom theamountof dataavailable exactly whatconstituteshegeneralpatternand what canbeconsidered nusual.Moreover,it is possible oidentify evolutionarypatterns,as well as consider heoccurrenceof specialcases,whetherdictatedby naturalor man-made auses,rom a more realisticstandpoint.The natureof currentresearch n rural RomanDacia describedabovesignificantly biases heevidence or anysuch analysisof theeconomicand social ife of theprovince.Since hisbias hasnotbeenrecognisedbefore, t raisesseriousdoubtsabout hevalidity of currently acceptedheoriesaboutthe developmentof this landscape. hepotentialdensityof humansettlementn theperiodhasnotbeen ully appreciated nd hetypologyof rural sitesmight not becomplete.Accordingly, thenativepre-Romancomponentn the life of theprovincehasprobablybeenmisinterpreted.All thesessuesaffect theevaluationof theRoman mpacton theconquerederritory andthe natureof theromanisationprocessn Dacia.Thepurposeof my researchs to redress omeof thesebiases. intendto focusmy studyon theeffectsof theRomanoccupationon the indigenoussettlementpatternandland-use.Fromaninterpretativepoint of view, thereare severalquestionso beaddressedLn what waydid theRomanconquestaffect thenative andscape?Whatwerethemechanisms ehindthechoiceof settlement

18

7/29/2019 20004olteanphd

25/274

locationandwhich of the factors hat influence hedecisionarepredominant n thecaseof differenttypesof settlementsn RomanDacia?Canwe detect heevidenceo support he deaof a state-directedpolicy of settlementemergence ndpattern n thecaseof RomanDacia,ashasrecentlybeensuggested, r is the impactof theRomancolonists heproductof multiple small-scalendividualstrategies?Did theconquestesult n any perceptible esistance henomenaamongsthenatives?Finally, how did theprocessof romanisationdevelop n Dacia?Tbroughthesequestions,t will bepossible o addressa numberof currentdebatesand assumptions:whetherarchaeologicalevidencebearsout the literary referenceso depopulation,whethermuch of thehinterlandof Sarmizegetusawasunoccupiedn thepre-Romanperiod,or whether andwasparcelled out andgivento thecolonists.Theunderstanding f therealRoman mpacts whethermilitary or civilian) and of the truenatureof thesocialrelationshipestablished etweenheconquerorsand nativeswill ultimately leadto abetterunderstand f romanisationn Dacia.

4. Methodology:My study encapsulates an area situated within the territory surrounded by the Carpathian Mountains,which was in fact the geographical core of both pre-Roman and Roman Dacia. It covers thehinterland of Sarmizegetusa (Colonia Upia Traiana Samizegetusa), the provincial capital andpossible former legionary base in the area of Hateg (figures 1.3-1.4), extending further along theupper Mures Valley, beyond the colony and legionary base at Alba Iulia (Apulum) (figures 1.5-1.6).Although the region hasbeen chosen to include in particular the lowlands of the Tara Hategului, andmid-Mures valley, it also covers the surrounding uplands that are structurally related (figure 2.1).More extensive upland areas ncluded in this study are the Orastie Mountains asthe main core of IronAge Dacia (figure 1.10), and the Roman iron-mining district from the Poiana Rusca mountains (inorder to balance the Dacian focus of iron extraction in the Orastie Mountains). However, the mostextensive mining area of the Roman Dacia, the gold-mining district located in the ApuseniMountains just to the north and west of the mid-Mures valley, hasbeen deliberately excluded. Thereason for this is that its extent and exclusive focus on mining makes the area a specialised landscapein its own right without much comparison with the either the lowlands or the uplands included here.Given the current bias of research, I have included a larger range of sites than might normally beaccepted as rural, extending my study to sites with an urban function (vid) but lacking the explicitproof of a municipal status. Such sites have traditionally been perceived by Romanian archaeologistsasnon-urban. Therefore, even though I will operate within the terminological framework currently inuse for Roman Dacia, my main concern will be to addressthe ftinction of each site wheneverpossible. This should provide a more realistic basis of study and redress the previous biases createdby an approach focused on status, without totally dismissing it as an issue.From what I have shown so far, it is clear that the traditional approach is very much out of date,especially when it comes to rural issues.The perspective for analysing the data must now focus onthe evolution of the landscape. This is the primary aim of my study on the evolution of the settlementpattern. In order to achieve it, I have augmented the current data-set with new information obtained

19

7/29/2019 20004olteanphd

26/274

throughmodemmethodsof archaeological rospection,particularly aerialreconnaissance.partfrom the fact that it haspreviouslybeen ittle applied n Romania, hismethodhasbeenshown oprovide someof thebestresults or landscapestudies e.g. Palmer1984;Stoertz1997).My researchmakesprimary useof thephotographicnformationfrom theaerial reconnaissancerogrammeconductedby Prof. Hanson seeabove)since1998which covers he samestudyarea figure 1.8),andwith which I havebeendirectly involvedas a part-timeresearchassistant.I haveshownabove hat thearchaeologicaldatacurrentlyknown is hardly satisfactory.The quantityand quality of informationon particularsites and heproblems n accessingt hasmademy taskextremelydifficult. In quantitative erms,a total of 627sites database ntries)havebeenconsideredin this study,but 60%of themare merelyaccidentaldiscoveries,mainly in the form of artefacts orthe way of their discovery/researchasnotbeenmentionedn publication); a ftu-ther10%come romantiquarian eports.Some 17% have beensubject o excavationat varyingscalesanda finther 8%have been eported hroughfieldwalking (althoughmany cannotbeaccuratelyocated).Only 3%representnew or augmented iscoveries rom theair. However, hequalitativebalanceof thedataprovidedin thesedifferentcategoriess effectivelyreversedseebelow).Nonetheless,all theinformationhasbeenbroughttogether nto a coherentsystem o permit its evaluationasa whole, norderto facilitate bothoverall anddetailedanalysis,and producegeneralconclusions.n parallel,theevidencehat formedthefragile basisof theprevious heorieshasbeenreviewed n orderto seewhatis reliable and what is not from a21' century archaeological erspective.This would haveallowedme to seehow muchof what is currentlyasserteds, indeed,basedon factsandhow muchhasstartedashypothesisbut endedup asacceptedact.The sourcesor this typeof dataare hevariousreportsof occasional ield walking, the existentgazetteers, xcavation eportsandotherpublications,mainlyfrom the20thcentury,but alsoof earlierdate.Of course,since he alreadypublisheddata-sethasbeen nadequatelyadministered,anynewadditionshighlight the need or exhaustive,lexible andinteractivemanagement f data,and ts analysiswithin a computerisedenvironment.Nowadays hisis possible hrough the ncreasinguse of computer acilities, in the form of databasesandGIS(Geographicnformation Systems).

4.1. The conversionof information into a relational database:The useof archaeological atabasess widely recognised sthebestway of handling argeamountsof informationfrom varioussources,ocations,anddates.Sinceno digital archaeologicaldataexistedwhen this studybegan, hadto collectand processahugequantityof published nformationscatteredn numeroussources.Accordingly, a relationaldatabaseof theknown sites rom variouspublicationshasbeendesigned,createdand subsequently ugmentedwith all thenewsitesof provenandpotential ateIron Age andRomandaterevealedby aerialarchaeologyn theareasince1998. Acopyof this databases providedon CD with this thesis.Theprinciple of thedatabase esignwasthe fragmentationof information into its parts inkedtogether hrougha complexsystemof relationships.Apart from thepracticality of ensuringbetterstorageof thesetsof information included, hemainbenefitof this system s that it imposesa clear

20

7/29/2019 20004olteanphd

27/274

standardisedseof termsanddefinitionsto describehesites,which allows advancedquantitativeanalysis.Thedatabaseontainsa main tableof siteswith detailsof their location,alongwithspecificationon thecharacterof previousresearch, n thesizeof the settlementor on the typeofhousesdocumented. or eachsite, related ablesprovide nformationon their chronology siteoccupation),sitetypeandfinds.The informationon site chronologycontains hespecifieddaterelated o that site entry,alongwith a schemeof theoccupationon that sitethroughout hemainprehistoricandhistoricalperiods.Becausewhenevert was possible,a singleentrywasused orcomplexsites, he tablecontaining hesitetypesprovidesa list of typeoptions;settlement(unknown);small town,hillfort, tower-house, illage, individual homestead,cemetery,ndividualgrave/tumulus, oad,religioussite/temple,aqueduct,quarry/mine,or artefactdiscovery.Oneorseveralof theseoptions can, herefore,beviable for one siteentry.A final entry contains heinterpretationof thestatustfunctionof thesite.Particularcategories f finds have beenconsidered ssignificant for thepurposeof this study:stonewalls,murusDacicus,timber,bricks,mortar,daub,wall plaster,paintedwall plaster, oof tiles,pavement,ceramicpipes,hypocaustpilae,architecturalpieces/statues,nscriptions,military diplomas, unerarymonuments, toragepits, hearths,agriculturaltools/millstones, umaces/kilns,hoards,coins, ewellery, styli, pins.A general ield was alsoprovidedto includeanyotherdetailsor commentson finds. Potteryfinds were ncluded n a separatetable,to include(wheneverpossible)detailson their manufacturingechnique,colour, clay texture,as well asbroad ndicationof date(Dacian,Celtic, Roman).Terminologicalstandardisationasnot beenseenasparticularly important n thecontextofRomanianarchaeology, ohavingto apply it now for all of thereportedsites n theareaprovedextremelydifficult. On theonehand herewas an unevenquantity andquality of informationfor thebetter-knownsites, eflectingthe interestsof their researchers.While abundantanddetailedinformationwasprovidedfor someaspects uch as ortificationsor artistic decorationson artefacts,considerablyess,or evenno informationwas providedon chronology, unction,or sitestructureandlayout. Moreover, hevariousdefinitionswhich havebeenapplied o siteshave beenparticularlyfocusedon reflecting their status, ather hanfunction(seee.g. chapter5). But thearchaeologicalevidences in mostcasesextremelyscarceandtheresearchmethodology raditionally employedhasbeen ar from satisfactoryn defining thenatureof thesettlementn the largemajority of thereportedsites n both the late Iron Age andRomanperiods.Many of thesites are represented nly by artefactdiscoveriesand n numerous ases,with theexceptionof theirbroad date, heir naturehasnot evenbeenspecified.Accordingly, theyhave beenconsideredwithin thepresentanalysisas settlements funknowntype'. Generalchronologicaldefinitionsused n thepublisheddatasuchas Iron Age, TaTene', Tacian' or 'Roman' is thereasonor havingto operaten themain with extremely oosechronologicalboundarieswithin thedatabase; ence,hepossibility of defining contemporaneoussites wasvery limited. Furtherdifficulty wasencounteredn including artefactdiscoveries.Especially n thecaseof those elated o the lateIron Age, thereweremany caseswhere henatureoftheseartefactsdid not explicitly indicatea settlement e.g. hoardsor isolatedartefactdiscoveries).However, hepresence f potterywas considered ereasanindicatorof potentialsettlement.Alsofunerarysites recordedoutsidea known settlementcontexthavebeenconsideredas ndicatorsof

21

7/29/2019 20004olteanphd

28/274

possiblesettlementn thevicinity.

4.2. Building an archaeologicalGIS:But simply analysing abular nformationon archaeological ites s insufficient when t comesounderstandinghe andscape.nformationrelated o the geographicandtopographicsettingof sites,thoserelated o theroadsystemor theproperconsiderationof site plans cannotbe handledby thetabularformatof a database. herefore, hedatabaseonly provideda frameworkfor handlingsiteattributeswithin the argerframeworkof aGIS.This is amoreadvancedool capableof complexmanagement f information in relationto its precise ocationwithin thephysicalworld, facilitatingtheplacing of sites nto their spatial elationshipwith eachotherandultimately offering a landscapeperspectiveof theevolution of settlementpatterns rom the latepre-Roman o Romanperiod.Thesystem createdusingArcView GIS 3.2) wasdesignedo includethepublished nformationonarchaeological ites n their relationshipwith thenatural andscape, ut especiallywith otherdata-sets.Oneof theseadditionaldata-setswasprovidedby historical geographicalandarchaeologicalinformation.The TabulaPeutingetianaoffers a uniqueaccountof settlementevidencerom Daciathathaspreviouslybeenconsidered, ut thecorrelationwith archaeologicalevidenceandespeciallywith the locationof thesesiteshasprovidedtheoverall analysiswith additionalgrounds orinterpretationof the functionsof thesesites and heir place within the landscape.Archive mapsavailable o meconsistof sheetsof the 1:28800Austriancadastral urvey ofTransylvania (1870-1875),which containa significantamount archaeologicalnformationthatneedsto beconsideredandevaluatedn connectionwith otherdata (figures 1.1-1.2).But apart rom directinformationon archaeological ites,archive mappingand early aerialviews provide informationonthe evolutionof the landscapeprior to much modemdevelopment.For this purpose, havecomparedvariouseditionsof modemmapswith the latesteditions available.Archive aerial photographs aveunfortunatelynot beenavailable;however,for anareaalongtheMuresvalley theequivalenthasbeensuppliedby first-generationsatellite magery(CORONA) declassifiedby theUnited Statesn1995.CoronaKH4A satellite sequenceDS 022-21104DF025s adeclassifiedntelligence mageacquired n 2e of July 1965. t covers heMuresriver valley betweenVintu de JosandZamwith awidth of 10.6mileson theground,and provides abestground resolutionof approximately3 metres.But despite ts poor resolutioncomparedo conventionalaerialphotographs,hevalueof theJuly1965recording s high becauset givestheopportunity o evaluatehe landscape efore t wasaffectedby the ater developmentof the area.This informationwasof particularhelpalso n theprocessof interpretationof archaeologicaleaturesvisible from theair and n theestimationof theirpotentialdate.GIS is capableof establishinghemorphologicalbasis or therecognitionof differentsitetypes,orfacilitating analysisof settlementdistributionpatterns,both in relationto eachother andagainstothernaturalfactorssuchas opography,soil or vegetationcoverageand and use. Ideally suchinformation is providedby mappingagenciesn digital formatat thedesiredscale.But the

22

7/29/2019 20004olteanphd

29/274

developmentof digital mapdata nformation in Romania s no less n its infancy than t is forRomanianarchaeology.n theseconditions,everythinghadto beproduced rom scratch.Theappropriatebackgroundor the locationof sites n thenatural andscapeand n relationto modemlandusewas providedby LANDSAT-5 satellitedata reely availablefrom theNASA internetwebsite(tti2s:/zulu.ssc. asa. ov/mrsid/-lastvisited 10.05.2004).With a pixel resolutionof 28.5metres considerablyess han heearlierCORONA images), his multi-spectral mage aken n theearly 1990s s far too poorto beused or detailedmapping, n relation to the interpretationof aerialphotographs.However, t providedsufficientdetailed nformation to support hecreationof a mapbaseof thearea, ncludinganindicationof modem and-use,againstwhich couldbeset all thesitesincluded n thedatabase,hemainRomanurban centresandthesitesdiscoveredhroughaerialphotography all aspoint themes),and hehydrographicalnetworkoverlain(asa line theme).Butdespiteall efforts to locate hesitespreviouslyidentified and publishedby theRomanianarchaeologists,ossiblemistakesmade n theoriginal publicationsare ikely to havebeenpreservedin thepresentstudy, unlesssubsequentataacquired hroughaerial reconnaissance ndtheassociatedmappingmethodologyemployedhave mprovedour knowledgeof the location (e.g.Cigmau,seebelow chapter5 andOlteanandHanson,2001). But improvementn site ocationwasonly a minor benefitof theuseof aerialphotographsn this study, n comparisonwith thehugecontributionto theclarificationof thecharacterof theoccupationand even ndicationsof sitefunctionprovidedby their facility to generate ite plans.The GIScreatedor thepurposeof this studyhas been hereal support hathas acilitatedthere-interpretationof archaeological vidence.Simply to say hat its mainpurposewastheproductionoftheamountof distributionmapsprinted and ncludedherewould mean o minimise ts contribution,aswell aspotential.Throughadding hespatialdimension o the attributesof eachsitecontainednthe databaset enablesamultitudeof analyses.Only a numberof themformedthesubjectof mystudy andtheset of printedmaps ncludedhasbeen imited. However,a set of thedigital data hatwasused hroughout his studyhasbeen ncludedwith thetext in orderto facilitate furthercustornised ueriesby thereadcr/user.Finally, occasionaluseof theGIS capacity or handlingattributes n relationto spatialrepresentations asemployed o facilitatereinterpretationof individual publishedsitesof both lateIron Age andRomandate. This facilitatedtheproduction of improvedplanswith a consistent caleandorientation(e.g. for Dacianhousesn theuplands see figure 4.1) or with a differentiateddisplay/legendaccording o different featuresndicatingpossibleconstructionphases r useofinternalspace e.g. for excavated illa sites see figure 5.1). This exercise esulted n providingsomenewunderstanding f theuseof internalspacewithin lateIron Age andRomanvilla housesn Dacia.

4.3 Interpretation and mapping of oblique aerial photographs:GIS is an appropriateool for interpretationand mappingof the aerialphotographsand, urther,forthe creationof site plansand regional archaeologicalmaps.Settlementsitesand other relevantlandscapeeaturespotentiallyof Iron Age or Romandatediscovered rom the air havebeen

23

7/29/2019 20004olteanphd

30/274

7/29/2019 20004olteanphd

31/274

boundary.An additionalproblemof thecadastralmapswhich hadto beused n a numberof cases(e.g. Sebes,Vintu de Jos,Sibot) is thattheydo not record mportant topographicalnformation,suchasthecontours,sothatdigital terrainmodelling couldnot beproduced.At timesthecultivation systemhasproved o bean mpediment o sitevisibility from theair,particularly in areasof strip-fields,since he creationof cropmarks s heavily dependenton the typeof vegetationcoverageandnot all croptypesare equally responsive o archaeologicaleaturesbeneathheground.For example,atVintu deJos(site409) reconnaissancen the summerof 2000identified theplan of a settlementwith sunkenhousesand storagepits immediatelyadjacento aRomanvilla both visible ascropmarksn a field whenthis was under wheatcultivation.Thecomsown in thefield in 2002made he site nvisible from theair, thoughit facilitatedgroundreconnaissancey ProfessorHansonandmyselfwhich led to the discoveryof archaeologicalmaterial, ncluding Roman egulde,supporting hepositive identification of thesiteas a Romanvilla(see HansonandOltean,2003;Oltean2004). At Oarda, ike at Vintu de Jos,someof thefields aremoreextensivesothata reasonable roportion of the remainsof thebuilding complexwas morereadily visible. However, n 2003only some50%theareaof Oardawasvisible becauset wascoveredby two differentcropswhich had different responseso cropmarkformation. Themosteloquentexampleof thestrongbias n site recognition and mapping nducedby cultivationofdifferent crops n adjacentong narrowstrips s visible at Micia (figure 5.33),where heextensiveplan of thesite included n this study(Figure5.35) s theresultof sustained urvey overseveralyearsduringwhich differentpartsof thesitebecamevisible ascropswererotated.Therefore,some100photographswere examinedandseveraldozenwererectified andtheir archaeologicaleaturesmappedn 3 different stages in 2000,2002and2003). Similar conditionshavebeenencountered tCigmau;while theextensivevicusatRazboieni,despiteregular surveillance,only startedo reveal tsfeatures n 2002andmoreextensively n 2003.Additional problems or site identificationfrom theair werecreatedby thespecific climateand soilconditions.Thepartial floodsalongtheMuresriver in the surnmerof 1998andthegenerallywetweatherconditions he following yearbadlyaffected hecreationof cropmarks n thesurveyarea.By contrast,extendeddrought n 2000also affected hevegetationand crops:on theonehand,parchmarksof buried buildingswere visible in earlyJune,while on the other,on extensivecultivatedareascropsdid not grow at all. At both CigmauandMicia different buildings havebeenvisible notonly in different summers, ut at differenttimesduringeachsummerbecauseof variations n cropandsoil conditionsacrosseachsite(OlteanandHanson2001).Becauseof thenatureof the ocal soils(alluvial clays, retainingmoisturebetterthansandysoils),mainly negativecropmarksare visible indicatingthepresence f stonebuildings and husfavouringthe discoveryof Romansites(HansonandOltean 2003). Positivecropmarks epresenting itches,pits, drainsor sunken-housesre rarelyrecorded.Theyaremorecommonto agricultural villagesandpre-Romansettlements,which meanshattherecoveryof native-typeof settlements y aerialphotographyhasbeensignificantly reduced,creatingapotentialbiasthatneeds o betaken ntoconsideration.Also, asdemonstrated tMicia (Olteanel aL forthcoming) this bias in thedataprovidedby aerialreconnaissance eanshat thecropmarkevidence eflectsthe stonephases f

25

7/29/2019 20004olteanphd

32/274

constructionat thesettlements,or only in theverybestconditionsare he slight tracesof theconstruction renchesof timberbuildingsvisible as positive cropmarks.Also, only rarelycancropmarks ndicatedifferentphasesn stoneconstructionson basisof differences n alignmentandfeaturesoverlapping e.g. Cigmau,Oarda,probablyRazboieni-see chapter5).It is inevitable,therefore, hattherecoveryof siteplansbasedon aerialphotographss partial.Giventhechangingclimatic conditionsandvegetationcoveragen each season, erial surveyshouldbecomea regularannualactivity, as t is in Britain, andthroughoutdifferent seasonsn order o beableto detectarchaeologicalemainsalso hroughshadow,soil andfrost marks; also,futureresearchcould augment hesesiteplanswith additional nformation, ncluding precisedating.The detailedmappingbasedon the interpretationof aerialphotographs onstitutesa separateayerofinformationwithin theGIS. Other ayerscover general nformationonpublishedarchaeological itesandthe satellitedata.Themappingwasgenerated singtwo different grid systems.The satellitedatais calibratedaccording o theUTM-WGS84,which seemso bemoreand moretheuniversallypreferredmappingsystem.Accordingly, it was usedalsofor thegeneralviews/mapsof theareaandthe representation f thesitesaspoints(in relationto databasenformation).But for individual siteplansof aerial photographs,heRomanianNational Grid systemwasused ProiectieStereografica1970 Dealul Piscului).This wasdonepartly becausehemapsavailable(1:5000)were usingthissystemand partly becauseof therecognition hat this of data-setwould beused n subsequentieldresearch,ncludingexcavation,by Romanianarchaeologists. or thesesites,a certain amountofinformation (e.g. topography,major rivers and streams,modemroadsandrailways)wasdigitisedfrom therelevant1:5000papermaps n order o providebackgroundnformation for thesitemaps.However,all thesitesdiscoveredhroughaerial photographyare present n the generalandscapeviews as a dot themeand, herefore,have been ncluded n thegeneral andscape nalysis.Theresultantbasemapsof theareaprovidean overview of thedevelopmentof the landscape eforeandafter the Romanconqueston which someof thesomeof theanswerso themorecomplex ssuesofcolonisationand,ultimately, romanisationn Daciahave beenbased.It would beunfair to say which of themethodsused n thepresentanalysishasproduced hegreatestoutcome.Theconversionof published nformation into database r GIS format haseachbroughtsignificantbencfitstowardsgeneralanalysisby increased ossibility to handlehugeamountsofinformation. Nevertheless, tarting rom a set of dataalready lawed by imprecision(seeabove), tcould only perpetuatehem.Furthermore,erminologicalstandardisationmposedby theconversionof data n thedigital databaseormatcanby itself introducebias,simply by losing a certainamountof informationwhich at somepoint couldbecomerelevant n someaspects.n orderto minimisesuchdanger, he designof thedatabase as ried to find an adequatebalancebetween ields ofattributesallowing limited options(yestnoor pre-defined ist) andotherswhereno such estrictionshave been mposed,but onemustbeawareof suchpossibility. The aerialstudyhas mprovedourknowledgeof sites ocationand ayoutwhich determinedobjectiveevaluationson thecharacterofthe occupationandsitefunction,but because f its nature(seeabove) t provided anyresultfor alimited numberof cases ndfor particular ypesof settlementsnot ableto identify settlementswhichby their natureandmorphologicalcharacteristics ada reducedmpacton thenatural andscape-e.g.

26

7/29/2019 20004olteanphd

33/274

possibleunenclosedDacianor non-villa Daco-Romanndividual homesteads seebelow chaptersand5).

27

7/29/2019 20004olteanphd

34/274

jk-

000

:32M xjik-- - tA;

Figure 1.1:Romanremainson the 1870-1875Austrian cadastral mapofTransylvania:the auxiliary fort atMicia (above)andof thecolonia Sarmizegetusa,with the Romanroadeastwardsstill in use at the time (in red) (sections183and231;

28

C the Austrian StateArchives, Vienna)

7/29/2019 20004olteanphd

35