Subsistence economy and land use strategies in the Burdur province (SW Anatolia)

14 Pages·2017·1.61 MB·English

by  B. De Cupere

QuaternaryInternational436(2017)4e17
ContentslistsavailableatScienceDirect
Quaternary International
journal homepage: www.elsevier.com/locate/quaint
Subsistence economy and land use strategies in the Burdur province
(SW Anatolia) from prehistory to the Byzantine period
B. De Cupere a,*, D. Fre(cid:1)mondeau b, E. Kaptijn a, E. Marinova a,b, J. Poblome c, R. Vandam d,
W. Van Neera,e
aRoyalBelgianInstituteofNaturalSciences,Vautierstraat29,1000,Brussels,Belgium
bCentreforArchaeologicalScience,UniversityofLeuven,Celestijnenlaan200e,3000,Leuven,Belgium
cSagalassosArchaeologicalResearchProject,UniversityofLeuven,BlijdeInkomststraat21/3314,3000,Leuven,Belgium
dDepartmentofAnthropology,StateUniversityofNewYorkatBuffalo,380MFACEllicottComplex,Buffalo,NY,14261,USA
eDepartmentofBiology,UniversityofLeuven,C.Deberiotstraat32,3000,Leuven,Belgium
a r t i c l e i n f o a b s t r a c t
Articlehistory: Areconstructionoftheancientsubsistenceeconomyandlandusestrategiesispresentedhereforthe
Availableonline28December2015 province of Burdur, more specifically the area south and southeast of Lake Burdur, in southwestern
Turkey.Thisreviewisbasedontheresultsfromarchaeozoological(includingbonecollagencarbonand
Keywords: nitrogen stable isotope analyses) and archaeobotanical analyses available from seven sites, dated to
Bioarchaeology distincttimeperiodsrangingfromtheNeolithictotheMiddleByzantineperiod.Thedata(bothpub-
Subsistence lishedandunpublished)arecompiledwithinformationavailableonsettlementdevelopmentanddy-
Settlementdevelopment
namics,inthestudyarea.Resultsshowchangingagrarianandanimalhusbandrypracticesforaperiodof
Diachronicperspective
morethan 8000years,rangingbetweenself-sustainingeconomies tospecialisedhusbandrypractices
Burdur
Turkey andanintensiveagriculturalexploitationofthelandscape,inrelationtohumansettlementactivities.
©2015ElsevierLtdandINQUA.Allrightsreserved.
1. Introduction raisingofanimalsethroughtimeanditslinktolanduseintensity,
technologicaldevelopmentsandsocialcomplexity.Providingdirect
With the onset of the Neolithic, human societies emerged in evidenceonthescaleandnatureofearlyplantandanimalculti-
whichtheexploitationofcerealsandanimalhusbandryprovided vationisnotonlyrelevantforarchaeologicalresearch,butalsofor
the staple subsistence resource. This inevitably resulted into a other disciplines like paleoecology, geomorphology and land use
modification of the natural environment (Goudie, 2013: 11e20). reconstructionsandmodelling.Also,itcanbeofgreatimportance
Theseearlysubsistenceeconomiesledtotheriseofsedentarylife, forsimulationsofpre-industriallanduse(Kaplanetal.,2011).
theabilityofproducingfoodsurplusesandanincreaseofpopula- Humanimpactontheenvironment,relatedwiththebeginnings
tiondensities(Colledgeetal.,2005).Inthesubsequentperiods,the anddevelopmentofagrariansubsistence,tookplaceinAnatoliaas
furtherdevelopmentofeconomicsubsistencestrategies,thatwere earlyasthe9thmillenniumBC(AsoutiandFairbairn,2010;Düring,
depending on cultural, biotic and abiotic backgrounds, caused 2011:16e17).Humancommunitiesunderwentthoroughchanges
considerableimpactonthelandscape(Dearingetal.,2006;Goudie, during prehistory, starting from the Neolithic (8500e6100 B C),
2013; Walsh, 2014). Studies of ancient subsistence strategies are continuingthroughtheChalcolithic(6100e3000BC)andfollowed
thereforeofgreatimportanceforunderstandingtheorganisationof bythedevelopmentofurbancommunitiesandmorecomplexso-
agriculturalproductioneincludingthecultivationofcropsandthe cieties in the Early Bronze Age (3000e2000 B C) (Sagona and
Zimansky,2009;Düring,2011).TheLakeDistrict(SWAnatolia)e
includingtheBurdurregioneplaysamajorroleinthediscussions
on diffusion and/or migration processes from the Near East into
* Correspondingauthor. Europe during Neolithic times (Thissen, 2010; Brami and Heyd,
E-mail addresses: [email protected] (B. De Cupere), Delphine.
[email protected] (D. Fre(cid:1)mondeau), [email protected] 2011; Düring, 2011). Also in later periods, the Burdur region
(E.Kaptijn),[email protected](E.Marinova),jeroen.poblome@arts. continuedtoplayapivotalrole,becauseofitslocationalongoneof
kuleuven.be (J. Poblome), [email protected] (R. Vandam), wvanneer@ the major north-south routes through this mountainous area as
naturalsciences.be(W.VanNeer).
http://dx.doi.org/10.1016/j.quaint.2015.11.097
1040-6182/©2015ElsevierLtdandINQUA.Allrightsreserved.
B.DeCupereetal./QuaternaryInternational436(2017)4e17 5
well as being connected to routes leading west, in combination TotheeastandsoutheastofLakeBurdur,alargezoneofmarlde-
with the presence of a vast area of fertile land (Mitchell, 1993; positsoccurs,theso-called‘BadlandsofBurdur’.
Talloen,inpress).DuringtheRomanImperialperiodthearchaeo- The climate in this region is considered to have an Oro-
logicalsiteofSagalassosbecamealargeurbancenterthatfulfilleda Mediterranean character, with typically dry hot summers and
centralroleinthewiderregion(Waelkens,2011;Waelkensetal., longwetwinters(Paulissenetal.,1993).FromJulyuntilSeptember
2011). theareaissubmittedtoaridconditions,whereastheothermonths
Archaeological research in the Burdur province (SWAnatolia) are quite humid. Temperatures frequently drop below freezing
documentstheriseandcontinuouschangeofsettlementsfromthe pointduringthewintermonthsandmostoftheprecipitationfalls
Neolithic(7thmillenniumBConwards)uptotheByzantinetimes inJanuaryandDecember,primarilyintheformofsnow.
(11th/13th century AD) (e.g. Vanhaverbeke and Waelkens, 2003; The natural vegetation in the study area belongs to the Oro-
Kaptijn et al., 2012, 2013; Vandam et al., 2013; Vandam, 2014, Mediterraneanzone,mainlycharacterisedbyforestsofPinusbru-
2015;Vandametal.,inprep.).Furthermore,thereisanincreasing tia, P. nigra and deciduous trees such as Quercus cerris (Zohary,
numberof environmentaland bioarchaeological studies available 1973).However,thevegetationconsistsnowadaysofamixtureof
forthis region. Archaeobotanicalreportsfromthisarea (Helbaek, Mediterranean forests andwoodlands, mainlycomposed of Pinus
1970; Nesbitt, 1996; Martinoli and Nesbitt, 2003) provide infor- nigra, Juniperus excelsa and J. oxycedrus, and to a lesser degree
mationonthecropsgrownintheLateNeolithicandChalcolithic Quercuscoccifera(Bakker,2012,p.9).Todaythemostfertileareas
periods. Studies on faunal assemblages have been published by arevirtuallyallundercultivation,whileintensivegrazingpressure
Westley (1970), Deniz and S¸entuna (1988), De Cupere (2001), De hastakenplaceonthelowerslopes,aroundtheagriculturalfields,
Cupere and Duru (2003) and De Cupere et al. (2008). Several andresultedinmaquisorquiteopenanthropogenicsteppewith,
palynologicalandpaleoecologicalstudies(BottemaandWoldring, amongotherspecies,Q.coccifera,J.oxycedrus,Artemisiaherba/alba
1984; Eastwood et al., 1998; Vermoere et al., 2002; Vermoere, andAstragalus(Kaniewskietal.,2007).
2004; Kaniewski et al., 2007; for a more recent review see;
Bakkeretal.,2012)focussedonthedevelopmentoflanduseand
the human impact on vegetation, as well as on past climatic 3. Materialandmethods
changes. Complementary information on the exploitation of ani-
mals within this region was obtained through geochemical evi- Asalreadymentionedintheintroduction,thedataofprincipally
dence (Vanhaverbeke et al., 2011a). Human diet and animal twodisciplineswillbeusedtomakeinferencesonthesubsistence
husbandry practices in the Classical/Hellenistic (400e200 B C), andlandusestrategieswithinthestudyareathroughtime.These
Roman Imperial (25BCe450 AD) and Byzantine (450e1200 AD) are archaeozoology and archaeobotany. The chronological frame
periods were reconstructed through bulk collagen stable isotope addressedinthispaperrangesfromtheNeolithictotheByzantine
analyses(Fulleretal.,2012). periods.Itwastriedtogatherasmuchinformationaspossiblefrom
The environmental archaeological studies, available for the thesetwodisciplinesforthedifferentperiods.Dataofownresearch
Burdur region, can, if combined, be highly informative on past wasused,aswellaspublisheddatabyotherresearchers.Materials
subsistence practices in the region, but are until now not suffi- fromthefollowingsiteshavebeenconsidered:DüzenTepe,Hacılar,
cientlyexplored.Therefore,theobjectiveofthiscontributionisto Hacılar Büyük Ho€yük (here after called Hacılar BH), Ho€yücek,
compileand,asfaraspossible,tointegrateallavailableinformation Kuruçay, Sagalassos and Bademag(cid:3)acı. The site of Bademag(cid:3)acı is
ebothpublishedandunpublishedefromarchaeozoologicaland located at about 50 km south of our study area; nevertheless its
archaeobotanical research, together with relevant archaeological data were also included, as it is situated within an environment
evidence, in order to provide a diachronic (and where possible whichis verysimilar tothe one discussed here.By including the
spatial)reconstructionofsubsistenceandlandusestrategiesinthe dataofthissite,itispossibletomakeinferencesabouttheNeolithic
studyareafromprehistoryintotheByzantineperiod.Thisarticle inthewiderregionofBurdur.
seekstoelucidatetheinteractionbetweenhumansociety,subsis- Faunalandmacrobotanicalassemblagesarewidelyrecognised
tenceeconomiesandtheenvironmentalsetting. tobeavaluablesourceofinformationonpasthumanactivitiesand
theyareoftenusedtoreconstructthehumandietor,moregener-
ally, subsistence in the past (Jacomet and Kreuz,1999; Reitz and
Wing,1999).Plantremainscanindeeddocumentthepresenceof
2. Physicalsettinganddescriptionofthestudyarea locallycultivatedcrops,weeds,importedplantsorplantscollected
in the wild, and provide information on the agricultural de-
The areafor whichthe datawill be presented in this paper is velopments, vegetation and land use (Jacomet and Kreuz,1999).
locatedwithintheprovinceofBurdur(SWTurkey),~100kmnorth Due to the soil conditions in the study area, only charred plant
of the Turkish Mediterranean coast, and covers a surface of remainsarepreserved(withcharringbeingmostoftentheresultof
approximately1200km2(Fig.1).ThestudyareaispartoftheLake dailysettlementactivities).
District and characterised by the presence of mountains, inter- Faunal remains document the importance of hunting versus
mountainbasinsandthelargesalineLakeBurdurwithintheBur- herding,thecompositionoflivestockandtheexploitationofani-
dur basin. The highest peaks in the study area range between mals (Davis, 1987; Reitz and Wing, 1999). The contribution or
2000 m (Bes¸parmak) and 2275 m (Akdag(cid:3)). The intermountain importance of each animal species or group is considered to be
valleys (like those of Ag(cid:3)lasun, Hisarko€y, Çanaklı, and Çeltikçi- reflected by the number of their respective remains (Reitz and
Bag(cid:3)saray) are generally filled with clay sediments (Dusar et al., Wing, 1999). In this paper the quantification of the faunal re-
2012). The valley of Ag(cid:3)lasun is located at a height of approxi- mains is based on the Number of Identified Specimens (NISP),
mately1100m.Moretothesouth,theplainnearBucakissituated whichrepresentsthetotalnumberofidentifiedelementsofeach
at an altitude of about 750e800 m a.s.l. To the southwest of the speciesortaxonomicgroup.Workedbonesorantlerfragmentsare
Burdur lake (about 850 m a.s.l.), the Burdur plain can be found, notincluded,sinceartisanalactivityimpliesselectionandpossible
between the modern villages of Düg(cid:3)er and Kuruçay; its altitude concentration(e.g.DeCupere,2001)andcould,therefore,leadto
risesinaccordancetothedistancefromthelakeupto950ma.s.l. falseconclusionsabouttheimportanceofacertainspeciesinthe
6 B.DeCupereetal./QuaternaryInternational436(2017)4e17
Fig.1. Mapofthestudyarea,withindicationofthepresent-daylocalities(dots)andarchaeologicalsites(stars;1:Sagalassos,2:DüzenTepe;3:Ho€yücek;4:KuruçayHo€yük;5:
Hacılar;6:HacılarBH)mentionedinthetext.
Fig.2. ArchaeobotanicalevidenceontheNeolithictoEarlyBronzeAgeplantsubsistenceinthestudyarea.
subsistence. Similarly, remains of intrusiveanimals and carcasses arenotasrepresentativeforthewholerangeofplantsusedatasite
havenotbeenconsidered. as the flotation samples are, but they can be used to prove the
An overview of the available archaeobotanical and archae- cultivationoruseofcertainplantsatthesiteswheretheyarefound
ozoologicaldataisgiveninTable1andsummarisesthemajortraits (Jacomet and Kreuz,1999). As most of the archaeobotanical data
oftheconsideredassemblages,inordertoevaluatethecompara- from the prehistoric sites are semi-quantitative, they are repre-
bilityofthedifferentdatasets.Regardingthearchaeobotanicaldata, sented on an abundance scale (Fig. 2); the fully-quantified data
we have considered in this paper samples that were obtained from the other sites (Düzen Tepe and Sagalassos) are given as
through flotation (i.e. at the sites of Hacılar BH, DüzenTepe and percentagesofthemainecologicalandeconomicgroups(Fig.3).
Sagalassos), aswellas plantremains thatwere hand-collected in Inthecaseofthearchaeozoologicalassemblages,allsiteshave
the excavation trench from visible concentrations, usually from yieldeddataonhand-collectedfaunalremains;atDüzenTepeand
storagefacilities(i.e.atthesiteofHo€yücek,HacılarandKuruçay). Sagalassos material has in addition also been retrieved through
The archaeobotanical assemblages from concentrations evidently sieving.Inordertoguaranteethecomparabilitybetweendatafrom
B.DeCupereetal./QuaternaryInternational436(2017)4e17 7
Table1
Overviewofthesitesfromwhichdatahavebeenusedinthispaper(AZ:archaeozoology;AB:archaeobotany).
Site Sitedescription ReferencetoAZ/ABdata Collection Samplesize:numberof
method bones/numberofABsamples
Bademag(cid:3)acıMound;Neolithiclevelswhicharecoveredbyquite AZ DeCupereetal.2008; Hand Morethan12,000bones
largeEarlyBronzeAgesettlement DeCupereetal.,inpressa
AB
Ho€yücek Smallmound;siteconsistingof2rectangularrooms, AZ DeCupereandDuru,2003 Hand 626bones
separatedbysomesmallerstorageareas; ABMartinoliandNesbitt,2003 Hand 11samples
datedtotheNeolithic
Kuruçay Impressivefortificationwallwithapairof AZ DenizandS¸entuna,1988 Hand Approx.1500bones;dataon
semi-circulartowers,surroundingsmallvillage;dated Neolithicmaterialmissing;
fromtheNeolithictotheEarlyBronzeAge identificationnotalwaysreliable
ABNesbitt,1996 Hand 25samples
Hacılar Mound;uninterruptedsettlementwith9phases, AZ Westley,1970 Hand Numbersnotalwaysgiven
datedfromtheNeolithictotheendoftheChalcolithic ABHelbaek,1970 Hand 9samples
HacılarBH Mound;large-sizedsettlement,surroundedby AZ DeCupere,unpublisheddata Hand Approx.3000bones
fortificationwall;mainlydatedtoEBAIandEBAII ABMarinova,unpublisheddata Hand/Flotation 12/6samples
DüzenTepe Large,fortifiedvillagesite,onhill;datedto AZ DeCupereetal.,inpressb; Hand/Sieved About36,000bones
Classical/Hellenisticperiod(500e200BC) unpublisheddata
ABFulleretal.,2012;Marinova, Flotation 36samples
unpublisheddata
Sagalassos Urbansite,onhillslope,withlargeterritory; AZ DeCupere,2001;unpublishedHand/Sieved Verylargecollectionof
mainlyinhabitedduringRoman-EarlyByzantineperiod dataFulleretal.,2012 animalbones
ABFulleretal.,2012;Marinova, Flotation 96samples
unpublisheddata
thedifferentsites,onlyhand-collectedremainshavebeenconsid- Inadditiontothearchaeobotanicalandarchaeozoologicaldata,
ered and will be discussed in this paper. The publication on the resultsofpreviouslypublishedpalaeoecologicalstudieswithinthe
faunalremainsfromHacılar(Westley,1970)isbasedonaverysmall regionareconsideredinthispaper.Areviewonthepalynological
amountofanimalremainsandexactnumbersarenotalwaysgiven; andotherpalaeoecologicalstudiesintheregionhasalreadybeen
theyshouldbeinterpretedwithcare.ThefaunalreportonKuruçay publishedbyBakkeretal.(2012)andwillbeusedhere.
(DenizandS¸entuna,1988)mentionstheidentifiedfaunalremains Thesubsistencedatawillbediscussedinviewofthesettlement
accordingtophase,butthelistoftheNeolithicmaterialislacking. patternintheregionthroughtime;asummaryofthisispresented
Inaddition,mostoftheidentificationsdonotgobeyondthe‘small/ in Fig. 6. The knowledge on the settlement patterns in the study
largeruminant’-level.Therefore,thesedataareoflimiteduse.The area is based on two types of survey, i.e. intensive and non-
archaeozoologicaldataofthesitesarepresentedinFigs.4and5. intensiveorextensivesurvey.Duringtheearlyyearsof theSaga-
Bonecollagenstableisotopeanalysis(d13C,d15N)hasbeencarried lassos Project, the entire territory was investigated by non-
out on animal remains from Düzen Tepe and Sagalassos assem- intensive survey (Waelkens et al.,1997, 2000; Vanhaverbeke and
blagesanddocumentschangesinlivestockfeedingstrategiesfrom Waelkens,2003),followedbyaseriesofintensivesurveys,firstin
the Classical-Hellenistic to the Middle Byzantine periods (Fuller the direct vicinity of the city (1999e2006) (Vanhaverbeke et al.,
etal.,2012);resultsofthisstudywillalsobeincludedinthispa- 2007), later in more remote valleys of the territory, i.e. around
per. For the moment there are no archaeobotanical or archae- Bereket (2008) and Bag(cid:3)saray(2009) (Vanhaverbekeet al., 2011b;
ozoological data on the study area for the Middle and Late Kaptijnetal.,2013)andintheBurdurPlain(2010e2012)(Kaptijn
Chalcolithic,LateBronzeAgeandIronAge. etal.,2012;Vandametal.,2013;Vandam,2014).
Fig.3. ArchaeobotanicalevidenceontheplantsubsistenceatDüzenTepeandSagalassos(Classical/HellenistictoMiddleByzantine).
8 B.DeCupereetal./QuaternaryInternational436(2017)4e17
Fig.4. Compositionofthelivestockintheprehistoricfaunalassemblages(NeolithictoEarlyBronzeAge).Numbers(n)indicateNISP's.
Fig.5. CompositionofthelivestockinthefaunalassemblagesofDüzenTepeandSagalassos(Classical/HellenistictoMiddleByzantine).Numbers(n)indicateNISP's.
4. Diachronicsynthesisofdata observed,whilepigandcattlearemuchlessimportant.Giventhe
slaughterdata,ithasbeenarguedthatthesmalllivestockewith
4.1. NeolithicandEarlyChalcolithicperiod(7000e5500BC) sheep outnumbering goats e at this site were (at least partially)
exploitedfortheirmilk;thepresenceofdairyfatwasalsoattested
Indicationsofhumanactivitypreceding6500BCintheBurdur throughresidueanalysiscarriedoutonceramicsherds(DeCupere
province are scant. The possibly aceramic phase excavated at et al., 2008; De Cupere et al., in press a). The evidence for milk
Hacılar (Mellaart, 1970, pp. 3e7) remains controversial (Duru, production, together with their predominance in faunal assem-
1989). The earliest levels of Bademag(cid:3)acı (EN I levels) could indi- blages,underlinestheimportanceofsheep/goats,andespeciallyof
cate that the Neolithic way of life had probably already been sheep(DeCupereetal.,inpressa).Dataonslaughteragessuggest
establishedbefore6500BC,asseenmostclearlyatUlucakHo€yük thatherdingstrategiesforcattleweredirectedtowardsmeatpro-
(Level VI; Çilingiroglu et al., 2012, p. 141). However, the limited duction (De Cupere et al., 2008). During the Late Neolithic-Early
amountoffindsandabsolutedatesoftheselevelsmakeitdifficult Chalcolithic, both at Bademag(cid:3)acı and Ho€yücek there is a strong
toassesstheirpossiblepre-6500BCdate.Thefirstconclusiveev- increasedimportanceofcattle,whichcanbelinkedtoanintensi-
idence of human occupation derives from the Late Neolithic ficationofbeefproductionor,asindicatedbythekill-offpattern,a
(6500e6100BC).TheknownLateNeolithicsettlementsappearto preference towards milk exploitation (De Cupere et al., 2008). In
besimilarincharacter,consistingofsmall/medium-sizedvillages addition to the domesticates, remains of wild animals were also
(1e2ha), with no evidence for smaller, contemporarily, dispersed found in the faunal assemblages. Their share represents the
farmsorhamletsyetidentified(Vandam,2014;Vandam,accepted). importanceofhuntingversusherdingandbreedingthroughtime.
These Neolithic villages were spread throughout different inter- Among the hunted species are mainly deer, including all three
mountain plain areas of the Burdur region (French, 2012, maps species native to this region, namely red deer (Cervus elaphus),
11.16e11.17) and located in favorable places in close proximity to fallow deer (Dama dama) and roe deer (Capreolus capreolus). Oc-
water sources and fertile land (Vandam, accepted). A similar set- casionally, other large mammals such as wild boar (Sus scrofa),
tlement pattern can be reconstructed for the Early Chalcolithic bezoargoat(Capraaegagrus),mouflon(Ovisorientalis)andbrown
(6100e5500 B C) but now, in addition, smaller isolated nuclei of bear(Ursusarctos)werecounted.Remainsofsmallwildmammals
habitation, probably farmsteads or hamlets, seem to have been that contributed to the subsistence, including hare (Lepus euro-
presentinthelandscape(Kaptijnetal.,2012;Vandam,accepted). paeus), and birds, are scant in the hand-collected faunal assem-
TheemergenceofNeolithicculturesinvolvedtheintroduction blages.Altogether,theNeolithicassemblagesofBademag(cid:3)acıshow
and establishment of domestic flocks (comprised of sheep/goat, quitesmallnumbersofwildremains(upto5%).
cattleandpigs)withthefirstsettlers(DeCupereetal.,2008).For No stable isotope data are available for this period in the
theearlylevelsatBademag(cid:3)acı,agreatrelianceonsheepandgoatis geographicalareaconsidered.However,thefaunalassemblagesof
B.DeCupereetal./QuaternaryInternational436(2017)4e17 9
ARCHAEOLOGY ARCHAEOZOOLOGY PALYNOLOGY ARCHAEOBOTANY
GEOMORPHOLOGY
Genetic data show a great reduction in population
early 13th C.
End of Sagalassos (and most sites in countryside). mid 12th C. Abrupt end of anthropogenic ac(cid:415)vity in palynological data.
Construction of fort on hilltop in front of Sagalassos: 12th C.
violently destroyed (mid 12th/early 13th).
AD 1000 Small scale resurgence of agriculture visible in pollen cores.
(re-)appearance of fortified hilltop settlements: insecurity?
Sagalassos looses urban character. Great decrease in c. AD 900
activity in countryside. Human indicators in pollen cores reduced. End of BO-phase at Gravgaz.
AD 650
Sagalassos: reduction in habitation + monumental character. Decrease in ca(cid:425)le importance in favour of sheep/goat.
Functional change of buildings (incl. agricultural activities). Pig, sheep and goat similar C3 plant-based diet, ca(cid:425)le Cereals remain staple crop, but barley more important.
Introduc(cid:415)on of millet/rye.
seems to be managed differently (grazing in marshy areas?).
AD 450
BO-phase ends at Bereket (paludifica(cid:415)on?), reduc(cid:415)on in
Large scale building activities at Sagalassos. AD 300 agriculture at Gravgaz: areas now primarily used for grazing.
Sagalassos becomes 'first city of Pisidia'.
Large scale activity in the countryside. Intensive agriculture, hor(cid:415)culture and arboriculture.
Greater diversity taxa and hence cul(cid:415)vated areas.
25 BC Staple crops: free threshing wheat and barley, pulses less
Settlements move from hilltops to valley edges. important, appearance of several fruit and nut species.
Increase in number of settlements.
Sagalassos becomes central settlement in region. 200 BC
300 BC BO-phase also visible in valleys of Bereket and Gravgaz.
Enhanced erosion (deforestation) and sedimentation.
By 300 BC valleys filled with sediments: suitable for extensive crop
cultivation, but limiting the possibilities of livestock grazing on hillslopes.
Emergence of large well-defended hilltop settlements.
Burdur plain densely occupied by small villages. High proportion sheep/goat (milk, wool, meat), less cattle
Free threshing wheat, pulses and some barley, viniculture.
and pig. Stable isotopes: all animals grazed in same area
or fed more or less the same. Winter cereal cul(cid:415)va(cid:415)on in rota(cid:415)on with several pulses.
700-800 BC
Palynology: start of Beyşehir Occupa(cid:415)on-phase in Ağlasun valley.
800/1000 BC
Deforesta(cid:415)on in valleys of Bereket and Gravgaz.
1000/1200 BC
Hardly any information available
2000 BC
Rise settlement density, both small-scale (and short-lived?) Cereals and pulses important. Both drought
hamlets as well as larger villages and first large Predominance of sheep/goat (more sheep than goats), resistent + drought intolerant crops: risk reduc(cid:415)on
settlement (3.5-4ha): emergence of settlement hierarchy? but increase in reliance on ca(cid:425)le (milk and/or labour). through diversifica(cid:415)on or evidence for irriga(cid:415)on?
4000 BC
Hardly any information available
5500 BC
Similar settlement pattern but now with small farms Increase of ca(cid:425)le (meat).
in between the villages.
6000 BC
Husbandry of sheep/goat (milk), more sheep than goats; Cul(cid:415)va(cid:415)on and storage of cereals
First farmers: network of villages (1-2ha). to a lesser extent ca(cid:425)le and pig (meat). (esp. free threshing wheat) and pulses.
6500 BC
Fig.6. Synthesisofthebioarchaeologicalandarchaeologicaldata,presentedthroughtime(Timescaleisnotlinear).
several sites from southcentral (Çatalho€yük: 7500e6000 cal BC) resources and/or to a changing environment. Similar processes
and south-eastern (Çayo€nü Tepesi, 9the7th mil. cal BC) Anatolia mighthavetakenplaceinourstudyarea.
andnorthernSyria(TellSabiAbyad,6800e5800calBC)havebeen Theprehistoricplanteconomywasbasedonavarietyofcereal
subjected to bone collagen carbon and nitrogen stable isotope crops (barley, free threshing and hulled wheat) and pulses (pea,
analysis (Pearson et al., 2007, 2013, 2015; van Der Plicht et al., lentil,bittervetch,grasspea,chickpea).Theimportanceofpulses
2012).Mostdataconcernsheep/goat,whilecattleandpigareless correspondswiththeobservationsmadeforcropassemblagesfrom
documented. Depending on the site and period considered, the theAegeanandCentralAnatolia,datedfromtheNeolithicuptothe
proportionofC4plantsvariesgreatlyinthedietofdomesticani- Early Bronze Age (Asouti and Fairbairn, 2002; Riehl, 2014). How-
mals:fromabsentoranecdotic(asisgenerallythecaseforpigs)toa ever,theassemblagespresentedherewere,withtheexceptionof
likelysteadyamount(e.g.forsheep/goatsfromÇatalho€yükandthe some samples from Early Bronze Age Hacılar BH, collected from
earliest level of Tell Sabi Abyad). Changes in the management of accidentallyburntstorages.Therefore,theyarenotrepresentative
sheep/goats through time are observed, including a shift in the forthewholerangeofplantsusedandtheabundanceofeachplant
environments exploited. A specific management strategy is sug- group(i.e.cereals,pulses,oilcrops,etc.)doesnotnecessarilyreflect
gestedforcattleatÇayo€nüTepesi(wherecattlestandoutwithvery its relative importance. Nevertheless, the presence of storage fa-
highd15Nvaluescomparedtopigs,sheep/goatsandhumans),while cilitieswiththeremainsofpulsecropsshowthattheywerealready
a diversity in cattle management is proposed for Çatalho€yük. A available(Fig.2).Freethreshingwheatalsoplayedasignificantrole
diachronicshiftintheenvironmentsexploitedasgrazingareasis in the subsistence, again similar to tendencies that have been
alsonoticedatthelattersite.Pigvaluestestifytoaratherherbiv- observed at Neolithic sites in centraland eastern Anatolia (Riehl,
orousdiet, indicatingextensiveherding rather than a permanent 2014). Free threshing wheat is less drought stress tolerant than
presencewithinthevillages.Inshort,itappearsthatalreadyduring barley and hulled wheat, and requires more care (Nesbitt and
theNeolithic,differentherdmanagementstrategieshaddeveloped Samuel, 1996). It is usually grown separate from hulled wheat,
according to the site and/or the period considered. This may implying different cultivationplots and reliance on a diversity of
reflect different processes: specialisation, adaptation to available crops.
10 B.DeCupereetal./QuaternaryInternational436(2017)4e17
4.2. MiddleandLateChalcolithicandEarlyBronzeAge The diversityofcereal cropsknownfromthe Neolithicperiod
(5500e2000BC) continues during the Chalcolithic and Early Bronze Age period;
pulsesalsocontinuetobeanimportantelementofthediet.But,in
For the Middle Chalcolithic and the beginning of the Late contrast to the Neolithic period, they now mostly represent
Chalcolithic(i.e.theperiodbetween5500and3500BC)noremains drought resistant species, such as bitter vetch (Vicia ervilia) and
ofhumanactivityhaveyetbeenfoundinthestudyregion(Kaptijn grass pea (Lathyrus sativus/cicera). Further, oil/fibre crops (Linum
etal.,2012;Vandametal.,2013;Vandam,2015).Palynologicaland usitatissimum) are found for the first time within the region, in
sedimentological data also point towards a low anthropogenic storage facilities at Kuruçay (Late Chalcolithic) and Hacılar BH
impactandpronounceddryconditions(Bakkeretal.,2012).From (Early Bronze Age). The presence of this plant species is also an
3500BConwardsariseinsitenumbersintheBurdurregionand indicationforfieldswithdroughtintolerantcrops.Thesetwoele-
SW Turkey in general can be noticed (French, 2012, maps mentspointtowardstheexistenceofanagriculturalsystemwith
11.16e11.17).TheBurdurplainischaracterizedbyahighnumberof disparate cultivated areas. This may be considered an attempt to
sites,whichweremainlysmall-scale,flatandprobablyshort-lived, spread the risk of crop failure. Another possibility would be that
andco-existedwithlargervillages,likeKuruçayHo€yük(Vandam, certainplotswereirrigated.Suchpracticeshavebeensuggestedfor
2014; Vandam et al., in prep). The appearance of Hacılar BH the Late Chalcolithic and Early Bronze Age (3000e2000 BC) pas-
(Umurtak and Duru, 2012, 2013) in the Early Bronze Age I toralcommunityvillageofArslantepe(south-easternTurkey)(Masi
(3000e2600BC)isthefirstindicationfortheemergenceofamore etal.,2014).Thegrowthconditionsofbarleyandemmerhavebeen
complex society. This settlement, which is currently being exca- investigatedatthissitethroughisotopeanalysisandtheirD13C(the
vatedandseemstoberadialinshapelikeBademag(cid:3)acı,isparticu- carbonisotopediscrimination;seee.g.Arausetal.,1999)variation
larly large and may even, for the first time in the settlement patternsindicatedthatthebarleyfieldsweremostlikelynotirri-
history of the Burdur Plain, have surpassed the size of a village. gated(Riehl,2008;Masietal.,2014).Emmer,onthecontrary,might
Towhatextentitspresenceinthelandscapeimpliedsitehierarchy have benefited from anthropogenic water input or its cultivation
is currently unknown. During the Early Bronze Age II wasrelocatedtowater-richareas(fromwhichbarleywouldhave
(2600e2400BC),afurtherriseinnumberofsitescanbenoticed, beenexcluded)duringtheperiod3000e2750BC(Masietal.,2014).
alongsidetheappearanceofcemeteriesandformalburialsinthe Wild plant resources also continuously played a role in the
landscape(Vandametal.,2013).Incontrastwithotherregionsin subsistencewithin the studyarea. AtHacılar BH, the presence of
Western Anatolia, there is currently no evidence to assume that concentrations of Brassicaceae (cf. Descurainia), almonds (Amyg-
large complex sites with lower and upper settlements such as dalus/Prunus), terebinths (Pistacia sp.) and other wild food plants
LimanTepeorTroyoccurredinthestudyregion(Vandam,2014). showsacontinuityofNeolithictraditions(asforexampleobserved
Only very few sites, dating to the following Early Bronze Age III at Çatalho€yuk, see Filipovich, 2014; Fairbairn et al., 2006). The
(2400e2000 BC), havebeen identified and this might illustratea numerousgrapefinds,foundinallflotationsamplesofHacılarBH,
shifttowardssettlementnucleation. may reflect the increasing importance of this fruit in the plant
Sheep/goatsarethemostabundantatthesiteofHacılarBH,but economy of the region. Parallels are known from nearlycontem-
fromEBAItoEBAIIanincreasingrelianceoncattleisnoted;cattleis porarysitesinEastAnatoliaandWestAsia(Miller,2008).
predominantintheEBAII-levelsofthesiteofBademag(cid:3)acı(Fig.4).
Bothsheep/goatsandcattlewerekeptwellintomaturityduringthe 4.3. Archaic,ClassicalandHellenistictimes(750e25BC)
EarlyBronzeAgeatbothsites;thischangeinslaughterage,especially
ofcattle,suggestsaherdmanagementadjustedtoadifferentecon- Inboththeresearchareaandthewiderregiononlyfewsettle-
omycomparedtotheLateNeolithic/EarlyChalcolithic,mostprobably mentremainshavebeenidentifiedfromtheMiddleandLateBronze
amulti-purposeuse(principallyincludingmilk,fleece,labourand Age (2000e1650/1650e1200 BC) and the Early Iron Age
meat)ofthesedomesticates(DeCupere,unpublisheddata).Ingen- (1200e750 BC). This scarcity of remains makes it impossible to
eral,thecompositionoflivestockisdependsonenvironmentaland developareliablereconstructionofhumansubsistenceduringthese
political/socialconditions.Cattlewereimportantineconomic,sym- periods. In the Archaic period (750e500 BC) sites appear that are
bolicandritualpracticesduringtheBronzeAgeandageneralcorre- locatedontopofmountainsanddisplayelaboratefortificationwalls.
lationbetweencattleandsitestatushasbeenobserved:economically Additionally,several,apparentlyagricultural,villageswererecently
theywerehighlyvaluedfortheirsecondaryproducts,includingpo- discoveredintheBurdurplain,locatedatcloseintervals(i.e.~1km)
wer.Theywereusedasbeastsasburden,asshownindepictionsand on river banks (Kaptijn et al., 2012). Contemporaneous with the
LateBronzeAgetexts(Arbuckle,2014).Thehighproportionofcattle reappearance of visible human presence around c.1000e800 BC,
inthesubsistenceatHacılarBHcanmostprobablybelinkedtothe pollen cores inthe Ag(cid:3)lasunvalleyshowanincrease in human in-
sizeofthesite.Althoughthepresenceofmainlyold/senileindividuals dicators.Thisisconsideredtobethestartoftheso-calledBeys¸ehir
points towards the possible use of cattle as draught animals, no OccupationPhaseinthisarea(Vermoere,2004).Inanothervalley,at
pathological deformationsthatcanberelated tohard labourhave Gravgaz,adistinctperiodofhumaninduceddeforestationoccurred
beenobservedbyusinthefaunalassemblagesofHacılarBH,norfrom startingatmoreorlessthesametime,i.e.c.800BC(Vermoereetal.,
theEBA-levelatBademag(cid:3)acı.Theintensityoftheiruseasbeastsof 2002). Geomorphological research has identified a peak in slope
burdenwaseithertoolowtoinducepathologiesontheirbonesor erosionandcorrespondingvalleysedimentationwithastartingdate
cattlewere mainly kept for theirothersecondaryproductsinthis around 800/700 B C (Dusar et al., 2012). Furthermore, geo-
region.Noisotopedatapertainingtoanimalhusbandryareavailable morphologic modelling at Gravgaz has shownthat climate played
for the Chalcolithic/Early Bronze Age in the geographical area only a negligible role in this erosion, and that human impact, i.e.
considered,andinthewidergeographicalframeofAnatolia,faunal deforestation, was the dominant factor (Dusar, 2011). The reap-
stableisotopeanalysesaretooscarcetobeusable.Comparedtothe pearanceofsignificantevisibleinthearchaeologicalrecordehu-
Neolithic period, larger numbers (between 5 and 10%) of wild manpresenceintheregionduringtheArchaicperiodhadclearlya
mammalremainsareobservedfortheEarlyBronzeAge.Considering largeimpactonthenaturalenvironment.
the possible meat yield of these wild mammals (which exceed in IntheClassicalperiodthesettlementpatternremainedmoreor
almost all cases these of sheep and goat), they represent a non- lessthesamewithmostsitescontinuingornewonesbeingfoun-
negligiblecomponentofthesubsistenceeconomy. dedonsimilarlocations.OneofthenewlyfoundedsitesisDüzen
B.DeCupereetal./QuaternaryInternational436(2017)4e17 11
Tepe.ThissitefitsallthecharacteristicstypicalforArchaic/Classical periods(AD300e450).Bytheendofthe2ndcenturyADSagalassos
sites;itislocatedonasteepmountain,hasasubstantialfortifica- couldboastofatheatre,ahugebathcomplex,severalmonumental
tionwallandisofconsiderablesize.Excavationshaveshownthatit fountainsanditwasallowedtocallitself‘thefirstcityofPisidia’
wasalargedenselysettledvillage,occupiedbetweenthe5thand (Waelkensetal.,2011).TheprosperityofSagalassosismirroredin
the2ndcenturyBC,thatincorporatedcraftproduction(Braekmans, thecountrysidewherethenumberofvillages,hamletsandvillas
2010; Vanhaverbeke et al., 2010; Vyncke, 2013). At the site of hasgrownexponentially.
Sagalassos,locatedonaneighbouringslope, activityalsoappears AtSagalassos,theRomanImperialperiodischaracterisedbya
duringthisperiod.AlthoughthecharacterofClassicalSagalassosis dominanceofsheep/goat(Fig.5)egoatsarenowmoreimportant
notentirelyclearyet,itwasatthistimemostprobablysmallerthan thansheepewhichwereslaughteredatanoldage,showingagain
DüzenTepe(Waelkensetal.,2011;Poblomeetal.,2014). thattheywereprimarilyraisedfortheirmilkandwool(DeCupere,
TheHellenisticperiod(333e25BC)marksachangeinthesettle- 2001,p.87,p.140:Fig.100).However,theproportionofcattlein-
mentpattern.Insteadoflargefortifiedhilltopsettlements,newsites creasesduringtheEarlyandMiddleRomanImperialandLateRo-
are now located in or on the edge of valleys. This location can be manperiods(Fig.5).Almostexclusivelyoldcattlewereslaughtered
(partially)explainedbytheslopeerosionthathadbeenongoingsince at this site (De Cupere, 2001, pp. 93e94) and pathological de-
c.800/700BC.Itisestimatedthatbyc.300BClimestoneslopeshad formations that could be related to the use of these animals for
lostmostoftheirsedimentandwerealmostcompletelybare(Dusar, heavy work, such as ploughing arable land and transport, were
2011). This not only made crop cultivation impossible, but also observed(DeCupereetal.,2000; DeCupere,2001,pp.105e116).
severely limited grazing in the higher areas. However, the valleys, Theincreaseofcattleremains,togetherwiththeiroldslaughterage
wheretheerodedsedimenthadbeendeposited,nowcontainedlarge andpathologicaldeformations,canalsobeinterpretedasapossible
areasoffertilesoilhighlysuitedforcropcultivation.TheHellenistic reflectionoftheintensificationofagricultureintheterritoryofthe
villagersinthevalleywereintensivelyfarmingthevalleyplains.Attwo siteeasmorecattlewereneededtoworkontheland(seealsoDe
otherpollencorelocationstheBeys¸ehirOccupationPhaseisnowalso Cupere,2001).Possiblythesestronganimalswerealsobroughtinto
visible,i.e.Gravgazfrom400to260BConwardsandBereketfrom actionforotherpurposes,e.g.transportofgoods(Corremansetal.,
280BConwards(Bakkeretal.,2012).DuringtheHellenisticperiod 2012).So,cattlewerenotonlyexploitedfortheirmeat,butalsofor
SagalassoshadoutgrownDüzenTepethatceasedtoexisttowardsthe their power. This stands largely in contrast with the prehistoric
endofthisperiod.Instead,Sagalassosbecamethepoliticalcenterofthe sites,andevenwithDüzenTepe(DeCupereetal.,inpressb),where
regionandpartofthestudyareaturnedintotheterritoryofthistown. noorveryfewlabour-relatedpathologieswerefound.Further,from
ThefaunalassemblageofDüzenTepe(500e200BC)showsalow the Early and Middle Roman Imperial periods onwards, stable
proportionof cattleandpigs,andahighproportionofsheep/goat isotopeanalysisalsoclearlyindicatesthatspeciesdifferfromeach
(Fig.5),goatsbeingasnumerousassheep(DeCupereetal.,inpressb). otherbothbytheird13Candd15Nvalueseinsharpcontrasttothe
Accordingtotheirslaughteringpattern,sheepandgoatsweremainly Classical/Hellenisticperiod.Cattleandsheepshowaslightincrease
raisedfortheirmilkandwool,andultimatelymeat(DeCupereetal., in their d13C mean values (though not statistically significant),
inpressb).Stableisotopeanalysisofcattle,sheep,goatandpigbone whichsuggestsanincreasedcontributionofC4plantstotheirdiet
collagenprovided mean d13C and d15N values close to each other, andchangeinthefeedingstrategyofthesespecies,potentiallykept
whichindicatesthatallthesespecieswereallowedtograzeinthe inthesameareasorfedasimilardiet(Fulleretal.,2012).Pigshada
same area or were fed a nearly identical diet (Fuller et al., 2012). mainlyC3terrestrialsignatureanda15N-enricheddietthatlikely
Altogetheritissuggestedthat,regardinganimalmanagement,the includedhumanfoodrefuse.Goatsstandoutbytheirlowd13Cand
inhabitants of DüzenTepe had developed a small-scale rural self- d15N meanvalues,which indicatea ratherdifferentmanagement
sustainingeconomy,withaherdmanagementbasedonsheepand comparedtotheotherspecies,likelyincludingkeeping,herdingor
goats.Inaddition,theshareofwildmammalsatDüzenTepeisvery grazing in other areas (possibly along the mountainsides; Fuller
lowanddoesnotevenexceed1%(DeCupereetal.inpressb),indi- etal.,2012).Huntingcontinuestoplayaminorroleinthesubsis-
catingthathuntingplayedaminorroleinthefoodeconomy. tence: as a whole, the share of wild mammals in the faunal as-
ThearchaeobotanicalevidencefromDüzenTepeindicatesthat semblagesofSagalassosisverylowandhardlyeverexceeds1%(De
the main cereal crops are free threshing wheat and to a lesser Cupere,2001;DeCupere,unpublisheddata).
extent barley, accompanied by a quite high variety of pulses, The archaeobotanical assemblages from the cultural layers of
includinglentil,bittervetchandchickpea(Fig.3).Pulsesarenearly Sagalasssosshowthatthestaplecropsarefreethreshingwheatand
asimportantastheywereintheprehistoricperiodandcomprise, barley(Fig.3).Pulsesbecomelessimportantbutavarietyoffruit/
basedonthenumberofidentifiedremainsintheflotationsamples, nutspecieshavenowappearedonthespecieslist.Indeed,pollen
approximately 20e25% of the crop plants. Considering the cores and counter-weights from olive or wine presses (found in
numerous finds of grape (including pips, skins and stalks), vini- severalvalleys throughout the countryside) also suggest that the
culturewaspartoftheagriculturaleconomyintheregion.Almost inhabitantswerepracticingintensiveagricultureincludingarbor-
nofurtherevidenceforthecultivationoffruits/nutswasfoundat iculture(Vermoereetal.,2003;Vermoere,2004;Kaniewskietal.,
thesite.Hulledwheat(einkornandemmer),typicalfortheAegean 2007).Basedonthewildweedsthatarefoundtogetherwiththe
(RiehlandNesbitt,2003),iscompletelyabsent.Thecompositionof cereals e (such as Agrostemma githago, Anthemis cotula, A. pseu-
themaincropscorrespondswelltowhatisgenerallyobservedin docotula,Bupleurumcf.rotundifolium,Bromuscf.arvensis,Galiumcf.
the eastern Mediterranean. The identified weeds point towards spurium, Lolium cf. temulentum, Ranunculus arvensis, Sherardia
wintercerealcultivation.Consideringthehighproportionanddi- arvensis, Vaccaria cf. pyramidata, Vicia hirsuta/tetrasperma) e it is
versityofthepulsecrops,thesewereprobablygrowninrotation assumed that the cereals were intensively cultivated as winter
withthecerealsinordertoincreasesoilfertility. crops. Most of these weeds point towards good soil quality and
moistureavailability.Amongthefruitsandnuts,it appearedthat
4.4. EarlyeMiddleRomanImperialandLateRomanperiod(25BC mostof theseweregrownlocally(i.e.Amygdalussp.,Ficuscarica,
eAD450) Juglans regia, Malus/Pyrus, Punica granatum, Prunus sp., Vitis
vinifera);somespecies(includingCorylussp.,Pinuspinea),however,
The growth of Sagalassos continued in the Early and Middle couldeconsideringtheirecologicalrequirementsenotbeculti-
Roman Imperial (25 BCeAD 100/AD 100e300) and Late Roman vatedlocallybutwereimported.Thereisalsodiverseevidenceof
12 B.DeCupereetal./QuaternaryInternational436(2017)4e17
vegetables, aromatic herbs and spices (like Anethum graveolens, Anatolia(Ertug(cid:3)-Yaras¸,1997).RemainsofC4plantsthataretypicalof
Apium graveolens, Coriandrum sativum, Cucumis melo, Cuminum wet habitats (such as Carex sp., Eleocharis sp. or Salsola sp.) have
cyminum,Saturejahortensis,Trigonellaphoenum-graecum),mostof been identified in the botanical assemblages (Marinova, unpub-
which, although already present in earlier periods in the Eastern lished data) and palynological data show increased moisture
Mediterranean,becameatypicalelementofthedietandagricul- availability during the Early Byzantine period with marshes and
tureduringtheRomanperiod.Theintroductionandestablishment wetlands potentially expanding in certain valleys (Bakker et al.,
ofextensivearboricultureandhorticulture,eachwiththeirspecies 2013). Both alternatives could have resulted in higher d15N
andspecificrequirements,indicatethattherealsomusthavebeen collagenvalues;inthefirstcaseasaresultofadjunctionofcattle
anincreasingdiversityofcultivatedareas. manuretotheprairieeincreasingthed15Noftheplants(e.g.Bol
et al., 2005); in the second case linked to the fact that moisture
4.5. EarlyByzantinetimes(450e700AD),ByzantineDarkAges canlocallyincreasesd15Nplantvalues(Handleyetal.,1999).
(700e850AD)andtheMiddleByzantineperiod(850e1200AD) Inthefirsthalfofthe7thcentury,anearthquakedestroyedlarge
partsofthecity.Habitationinthecitycontinued,butonareduced
Habitationintensitywasstillhighduringthesecondhalfofthe scale(Poblome,2014).Fromthemiddleofthe7thcenturyonwards,
Late Roman (AD 300e450) and the Early Byzantine periods (AD areductionintheintensityofhabitationisalsovisibleintherestof
450e700). In certain valleys intensive crop cultivation and arbori- theterritory.Onlyasmallportionofthesettlementscontinuedinto
culture disappeared or greatly decreased and grazing grew in the Byzantine Dark Ages (AD 650e850/900), mostly the larger
importance during this period (Bakker et al., 2013). However, the Romanvillages,suggestingacertainlevelofcontinuity.Theexact
settlementpatterncontinuedseeminglyunchanged(Kaptijnetal., reasonforthischangeinsocialorganizationisunknown,butfol-
2013), indicating that this change of landscape use is likely the lows wider contemporary trends and patterns in Asia Minor.
reflectionofneweconomicchoicesratherthanadecreaseinhuman Palynologically,allintensivecropcultivationseemstohaveended
control/pressure. At Sagalassos, a reduction in the monumental and an even increased importance of animal husbandry is sug-
character of habitation becomes visible in the late 6th century gested (Bakker et al., 2013). The end of the so-called Beys¸ehir
(Jacobs,2015).Thearchaeobotanicalassemblagesshowthatcereals OccupationPhase,signalledbyasharpdecreaseinanthropogenic
remainthestaplecropintheEarlyByzantineperiod,althoughbarley indicator species, is recorded at several places in the territory of
is now more important than wheat (Fig. 3). Rye and millet are Sagalassos, though the timing differs from location to location
introduced into the area and, considering the ecological re- (Gravgaz: between AD 600 and 700; Ag(cid:3)lasun: ca. AD 1000 AD;
quirements of certain weeds and wild growing plants and their Bereket:aroundca.AD300AD;Bakkeretal.,2012).
presenceinthearchaeobotanicalrecord,thereisincreasingevidence MostoftheByzantineDarkAgesitescontinuedintotheMiddle
oftheuseofnitrogenrich,disturbedhabitats(Fulleretal.,2012).In Byzantineperiod(ca.AD850/900e1200),butnowseveralfortified
addition,remains ofcrop processing by-productsare found at the sites, located on steep mountain tops have also been identified.
site,whichwerenearlyabsentduringRomanImperial/LateRoman Most, but not all, are reoccupied sites from the Archaic/Classical
times.Thiscouldbeconsideredasanindicatorfortheincorporation period. These locations suggest a concern for security, i.e. these
ofagriculturalpracticesintheurbancenterofSagalassosintheEarly locations are often poorly accessible and far removed from both
Byzantinetimes.ItisalsoduringthecourseoftheEarlyByzantine potential agricultural fields and perennial water sources. During
periodthatthepubliclatrineoftheImperialBathswasrepurposed this period, palynology shows a limited resurgence of crop culti-
asaproductionsiteofruminantdung-basedmanure(Baetenetal., vation(Bakkeretal.,2013)andatSagalassosevidenceofoccupa-
2012). Flocks of sheep and goats became more important again tion has been attested at several locations in the ancient city
during the Early Byzantine period (Fig. 5) and continued to be (Waelkens, 2005, 2006, 2009). Only small archaeozoological as-
slaughteredatanoldage.JustasinRomanImperialandLateRoman semblagesareavailablefortheMiddleByzantineperiodandthese
times, goats were more common than sheep (De Cupere, 2001, p. datashouldthereforebeinterpretedwithcare.Nevertheless,from
140:Fig.100).Asforthecollagenstableisotopeanalysis,pigs,sheep thebonecollagenstableisotopeanalysis(Fulleretal.,2012),only
and goats display similar lower d13C values in accordance with a slightdifferencesindomesticanimals'dietcomparedtotheearlier
mainlyC3terrestrialsignature,whilecattlestandoutwithaslightly, phasecanbeinferred.Themostnoticeableoneisthatsheepseem
butstatisticallysignificant,higherd13Cmeanvalue,indicatingthatat characterizedbyisotopevaluesthatareclosertocattlethantogoat
leastsomeofthemhadaccesstoC4plants.Bothpigsandcattleare values (more alike Roman Imperial and Late Roman than Early
characterizedbyd15Nmeanvaluessignificantlyhigherthanthoseof Byzantine periods), suggesting they were kept with cattle rather
sheepandgoats.The15N-enrichedcollagenofpigscanbeexplained thanwithgoat,althoughthisobservationisbasedonafewremains
by their omnivorous diet. Cattle most likely benefited from a ofspecimensonly.InthevalleybelowSagalassos,alargevillageis
differentmanagement,implyingagreatercontributionofC4plants nowpresent,locatedclosetotheagriculturalfields(Vionis,2009).
totheirdietandtheirfeedingon15N-enrichedpasturesorfoodre- Duringthe12thcenturyAD,afortwasconstructedatAlexander's
sources(Fulleretal.,2012). Hill, a steep hill in front of Sagalassos, but this was violently
AlikelypossibilityforthegreatercontributionofC4plantstothe destroyed somewhere between the middle of the 12th and early
dietofcattlecouldbemixedgrasslands,particularlyastrampling 13th century AD (Vionis et al., 2010). All habitation at Sagalassos
and increased grazing pressure further the development of C4 endedatthistimeorshortlyafter.Thevillageinthevalleybelow
plants(C(cid:4)arniandMucina,1998;Wang,2002).Plantremainschar- continued, however, and Selçuk influence has been found in the
acteristicofsuchdisturbedvegetation(e.g.Setariaspp.andPortu- nearby village of Ag(cid:3)lasun (Vanhaverbeke et al., 2005). Like other
lacaoleraceaforC4plants,Polygonumaviculare,Loliumperenneand sites in the region, Sagalassos and its territory was part of the
(cid:4)
Plantago spp. for C3 plants; Carni and Mucina, 1998) have been Selçukempireintheearly13thcenturyAD.
identifiedwithintheEarlyByzantinebotanicalassemblages(Mar-
inova, unpublished data). This could perhaps be related to the 5. Discussion
changefromintensivecropcultivationtograzing,observedinsome
valleys.Alternatively,themarshyareasoftheterritorycouldhave Wildfaunaincludedmainlytheremainsofcervids.Whengoing
been exploited, as already proposed by Fuller et al. (2012). The moreintodetailandanalysingtheproportionsofthevariousdeer
gathering of fodder in wetlands has been attested in Central species at the prehistoric sites, it appears that roe deer has only
B.DeCupereetal./QuaternaryInternational436(2017)4e17 13
been identified at the sites of Bademag(cid:3)acı and Ho€yücek. Pre- the valleys and people were more inclined to keep goats on the
liminaryresearchshowsthatitisabsentatthesiteofHacılarBH.In mountain slopes: even though the upper slopes were severely
addition,atthelattersitereddeerismuchmoreprominentthan eroded,theyremainedsuitableforgoatherding.
fallowdeer,thisincontrastwiththesituationatBademag(cid:3)acıand Archaeozoological data are used to indicate the exploitation of
Ho€yücek, wherefallowdeer is more common than reddeer. Roe domesticanimals.Nexttomeat,milk,woolandpower,anotherby-
deer is mentioned onlyonce at Hacılar (involving an antler frag- product of these animals to be considered is manure. Although
mentwithpedicle)withinlevelIX(LateNeolithic)(Westley,1970). archaeozoologicaldatadonotallowtoestablishtheuseofmanure,the
AtKuruçay,allantlerfragmentshavebeenlabelledasroedeerand collectionofhumanandanimalmanurehasbeenattestedatEarly
thisidentificationmightbequestionedsincenootherdeerspecies ByzantineSagalassosthroughfaecalbiomarkeranalysis(Baetenetal.,
have been listed, not among the antler remains, nor among the 2012).Inarchaeologicalsurveymanuringoffieldswithanimaland
post-cranial elements (Deniz and S¸entuna,1988). Remains of roe habitationrefusehasoftenbeenattestedintheeasternMediterranean
deer are also completely missing and fallow deer is underrepre- andNearEastthroughthepresenceoflargehaloesor‘carpets’ofworn
sentedatClassical/HellenisticDüzenTepe(DeCupereetal.,inpress potterysherds(Wilkinson,1982;Bintliffetal.,2007;Kaptijn,2009).In
b). At Roman e Early Byzantine Sagalassos, remains of roe and the intensively surveyed Burdur Plain such ceramic evidence was
fallow deer are occasionally found, while red deer occurs a little absent.Onthecontrary,thesites,fromallperiods,stoodoutbecauseof
morefrequentlyinthefaunalassemblages. theirsharpboundariesbetweensiteandoff-site(alsocorroboratedby
Thetotalabsenceofroedeerandthedominanceofreddeer,in geophysicalevidence).Awayfromsites,theoff-sitescatterwasalmost
theBurdurplain(atHacılarBH)duringtheEarlyBronzeAge(and non-existent(Vandam,2014,p.82).However,itisveryunlikelythat
mostprobablyalsoduringtheNeolithic)isverystriking,especially theinhabitantsofSagalassosdidnotusethecollectedmanure.Instead,
whencomparedtothesituationinthesouthernpartofourstudy manure of Sagalassos was most probablyapplied as a fertiliser on
areawhere roe deer is well represented and fallow deer is more fieldsmoreintheimmediatevicinityofthetown.
abundantthanreddeer.Ecologicalrequirementsofthethreecervid Geochemicalanalysesoftraceelements(PbandCu)inarchae-
speciesareoverlapping,sinceallthreepreferablyliveinlandscapes ologicalbonefromgoat,cattleandpigofSagalassoshavebeenused
whereforestsalternatewithopenareas(Broekhuizenetal.,1992). toreconstructlandusechangesinthevicinityofthetown,aheavily
Reddeerandroedeeroccurbothinlow-andhighlands(VanDen pollutedzone(Vanhaverbekeetal.,2011a).Itwasshownthatthe
Brink,1978).BeforeitsintroductionanddistributionintoEurope, distancefromwhereanimalswereobtained,standsinrelationto
themainpopulationoffallowdeerlivedinwesternandsouthern theurban-ruralintegrationofthetownanditsterritory.Whenthe
Anatolia, where they must have been common on the coastal relationbetweenSagalassosandthecountrysidewasatitsclosest,
wooded plains and in the lower elevations of the mountains during the Late Roman period, the animals consumed in town
(Kumerloeve,1975; Uerpmann,1987, p. 57; Masseti and Vernesi, originated from beyond the immediate vicinity. During the Ear-
2015, Fig. 2.1). At Çatalho€yük, a predominance of red deer over lyeMiddle Roman Imperial and Early Byzantine periods the city
fallow deer is reported and also roe deer is only occasionally wasprimarilyexploitingtheareaclose-by,asshownbythehigher
recorded(RussellandMartin,2005,p.43:Table2.5),whichisvery levelsoftraceelementsintheanimalbone.
similartothesituationatHacılarBH,DüzenTepeandSagalassos. Increasing social complexity through time is also reflected by
One can assume that people targeted the available wild fauna. the archaeobotanical remains. Although no quantative data from
Therefore,itisverylikelythatthesitesinthenorthernpartofour theNeolithic/EarlyChalcolithicperiodcouldbeused,thenumerous
study region fall outside the natural distribution range of fallow grapefindsatHacılarBHcanbeanindicationforgrapecultivation
deer (and also roe deer) and that people had access to different duringtheEarlyBronzeAge,acropwhichrequireslongtermin-
gamedependingonthezonewithinthestudyarea. vestment of labour and certainly more than annual corps. In its
Sheepandgoatsareamongthemajorcomponentsofthefaunal turn,thiscanbeinterpretedasasignforspecialisationinsubsis-
assemblagesthroughoutallperiodsconsideredinthiscontribution. tence strategies. At Classical/Hellenistic Düzen Tepe, viniculture
It is interesting to note that the ratio sheep:goat varies through certainlytookplace,whilepulsesandcerealcropswereprobably
time: from the Neolithic to the Early Bronze period sheep grown in rotation. Clear archaeobotanical indications for a more
outnumbergoats;bothspeciesarepresentinaboutequalnumbers complexsocietycomefromthecityofSagalassoswheretherecords
during the Classical/Hellenistic period, and from Roman Imperial pointtoaplantsubsistencesysteminvolvingnotonlyspecialisa-
times onwards goat becomes the most abundant. This might be tion(includingannualcrops,arboriculture,vegetablesandspices),
related to the different areas preferentially inhabited and the but also consumption of imported food (such as date palm and
exploitationofthelandduringthevariousperiods,combinedwith stonepine),whichthereforecouldhavehadaluxurycharacter.
the feeding habits of sheep and goats. Sheep are preferentially BotanicalrecordsshowthatduringtheNeolithicuptotheEarly
grazers while goats are more likely to be herded and browse on BronzeAgethemainstrategytoreduceriskincludedthediversifi-
mountainslopes.Theprehistoricsiteswerelocatedinthevalleys; cation of crops grown: the population relied on three to four
accordingly,peopleweremorelikelytokeepsheepconsideringthe differentcerealcropsandonuptofourpulsecrops.Someofthese
vicinity of the pastures. At the hilltop settlement of Düzen Tepe crops,suchasemmer,weremoreproductive,whileothers,suchas
(Classical/Hellenistic), stable isotope analysis show that all do- barley, einkorn, bitter vetch, were more resistant to drought or
mesticates(sheep,goat,cattleandpig)hadaccesstosimilarfodder unfavourableconditions.DuringtheHellenisticandRomanperiods,
(Fulleretal.,2012).Undersuchconditionsbothsheepandgoatscan the staple cereal crops included only two species: free threshing
bekept,whichmightexplaintheincreaseofgoatcomparedtothe wheatandbarley.But,consideringatthesametimethepresenceof
Early Bronze Age. Following the evolution of soil erosion, which planttaxawhichwereinterpretedasweeds(cf.supra),thismightbe
startedintheArchaicperiodandculminatedinHellenistictimes, anindicationfortheintensificationofgrowingthesestaplecrops.
villages were built on the edge of the valleys (Dusar, 2011). The TheoverallcropdiversityincreasessignificantlyduringtheRoman
intensification of land exploitation e from the Roman Imperial andByzantineperiodsandisrelatedtoamoresophisticatedagri-
period onwards e put as much as possible of the landscape into culturalsystemincludingnotonlyannualstaplecrops,butalsofruit
cultivation, including the lower slopes (Vanhaverbeke and treesandvegetables.Millet,asummercropthatwasintroducedin
Waelkens, 2003, 306). Due to the expansion of arable land and theareaduringtheByzantineperiod(Fig.3),providedanotherback-
theincreaseofcattleherds,therewaslessplaceforsheepherdsin upincaseoffailureoftheothercerealsgrown.