/ÉS3 ш ELMALI PLAIN: A REVIEW OF ITS ENVIRONMENTAL SETTING AND ARCHAEOLOGICAL SETTLEMENTS

Volume I

The Institute of Economics and Social Sciences of Bilkent University

b y

y a s e m in ILSEVEN

In Partial Fulfillment of the Requirements for the Degree of MASTER OF ARTS

m

THE DEPARTMENT OF ARCHAEOLOGY AND HISTORY OF ART BILKENT UNIVERSITY ANKARA

June 1999 g s r . T

4 81 6 Í I certify that I have read this thesis and have found it is fully adequate, in scope and in quality, as a thesis for the degree of Master of Arts.

Doç. Dr. İlknur Özgen Supervisor

I certify that I have read this thesis and have found it is fully adequate, in scope and in quality, as a thesis for the degree of Master of Arts.

Dr. Ayşe Cfursan-Salzmat Examining Commitee Member

I certify that I have read this thesis and have found it is fully adequate, in scope and in quality, as a thesis for the degree of Master of Arts.

Dr. Jean Oztiirk Examining Committee Member

I certify that I have read this thesis and have found it is fully adequate, in scope and in aualitv?-as a thesis fq« the degree of Master of Arts.

Examining Commitee Member

Approved by the Institute of Economics and Social Sciences

Prof DrT^i Karaosmanoglu Director ABSTRACT

ELMALI PLAIN: A REVIEW OF ITS ENVIRONMENTAL SETTING AND ARCHAEOLOGICAL SETTLEMENTS tlseven, Yasemin Master, Department of Archaeology and History of Art Supervisor: Doç. Dr. İlknur Özgen June 1999

The Elmalı Plain is located in Northern , in the Teke Peninsula of

Southwest . The plain is geographically in a transitional zone between the forest rich Taurus Mountains and the Highlands of the Anatolian Plateau. Being a fertile upland basin, it is the closest large arable land mass to coastal Lycia. This thesis aims to place the Elmalı Plain in an environmental context and to compile the information regarding the history and archaeology of the region starting with the

Prehistoric periods up to the Islamic occupation of the plain. It evaluates the present evidence of literary sources, epigraphy, numismatic and archaeological remains. The thesis will also trace the history of traveUers accounts and scholarly research done in the region, and will try to understand the gaps that exist in our archaeological knowledge and identify issues that can be pursued further. The Elmalı Plain, with its natural resources, was able to hold substantial populations, which is evident from the

archaeological records for periods such as the Chalcolithic, Early Bronze Age, Iron

Ages and the Roman period. Existing evidence far from suggests a definite

continuity or discontinuity for all the historical periods, on which future

investigations, especially the ongoing project of Hacimusalar excavations and

survey, will throw more light. The present archaeological evidence for the Elmalı

Plain suggests that the region, rather than being a remote upland region, was a

transitional zone, both culturally and geographically between coastal Lycia and the

upland regions of , Kabalia and .

u ÖZET

ELMALI OVASI: ÇEVRESEL VERİLERİ VE ARKEOLOJİK YERLEŞİMLERİ ÇERÇEVESİNDE DEĞERLENDİRİLMESİ Hseven, Yasemin Master, Arkeoloji ve Sanat Tarihi Bölümü Tez Yöneticisi: Doç. Dr. İlknur Özgen Haziran 1999

Elmalı Ovası Kuzey Likya olarak bilinen Güneybatı Anadolu’nun Teke yarımadasında yer almaktadır. Ova coğrafik konumu ile bir geçiş havzası olarak nitelendirebileceğimiz, Toros Dağları ve Iç Anadolu Platosu arasında yer almaktadır. Bu çalışma. Elmalı Ovasını çevresel koşulları ve doğal kaynakları ile birlikte değerlendirmekle beraber, ovanın Prehistorik tarihinden başlayarak

Islami dönemine kadar uzanan zaman dilimine ı ş ı k tutacak tarihi ve arkeolojik verilerin derlenmesini amaçlamaktadır. Bu veriler bölgenin tarihi ile ilgili antik

kaynakları, epigrafik ve numismatik bilgileri, ve yapılan arkeolojik çalış maları

içermektedir. Bu çalışmada ayrıca bölgeyi geçen yüzyıllarda ziyaret eden

gezginlerin ve bölgeyle ilgili arkeolojik çalışm aların bir tarihçesi de verilmektedir.

Çalışma, ovanın sürekli bir yerleşim tarihinin belirlenmesi açısından

eksikliklerimizi ve yapılması gerekeni ortaya koymaktadır. Elmalı Ovası zengin

doğal kaynakları ile geniş nüfusları barındırabilecek alt yapıyı sahip

bulunmaktadır. Eldeki arkeolojik veriler ovanın Kalkolitik, Erken Tunç, Demir

Çağları ve Roma döneminde yoğun olarak yerleş ildiğini ortaya koymaktadır. Fakat

bu veriler ovada sürekli bir yerleşimin devam ettiğine dair kesin kanıt

sunmamaktadır. Ova ile ilgili arkeolojik veriler birçok boşluğa işaret etmektedir.

Gelecekteki çalışmalar, özellikle bölgede devam etmekte olan Hacımusalar K azısı

ve Araştırmaları ovanın arkeolojik tarihine ış ık tutacaktır.

m ACKNOWLEDGEMENTS

This work has been produced as a result of the works of various travellers and scholars who have visited and worked in the Elmalı Plain, which the prior acknowledgement should be given to. My greatest dept is to my advisor İlknur

Özgen, who gave me the chance to study in the Haclmusalar excavation under her guidence and shared her invaluable experience and ideas. Her farsightedness, intellectual stimulation and friendship has encouraged me at each stage of writing this thesis. My special thanks extends to Ayşe Gürsan Salzmann and Jean Öztürk for their generous comments and corrections on the draft of this work. I would also like to thank to Deniz Kaptan, Julian Beimett and Jacques Morin from the Department of

Archaeology and Art History. My thanks are extended to Gülcan Çelebioğlu for her help and morale support and to friends Aylin, Esen, Suna, Bülent, Tuğba and Elif who were always there to help and give morale support. Finally I am thankful to my family, and especially to my mother, without their encouragement and affection I would have never been able to succeed this work

IV TABLE OF CONTENTS

VOLUME I page

ABSTRACT u

ÖZ 111

ACKNOWLEDGMENTS IV

TABLE OF CONTENTS ...... v

LIST OF MAPS ...... X

LIST OF PLATES ...... x

LIST OF FIGURES ...... xvi

LIST OF CHARTS ...... xvi

LIST OF ABBREVIATIONS

INTRODUCTION 1

CHAPTER I: ENVIRONMENTAL SETTING 6

1.1 6

1.2 Climate 14

1.3 Vegetation 19

1.4 Modem and Ancient Lines of Communication 30

CHAPTER II. HISTORICAL REVIEW OF THE PLAIN IN THE LIGHT OF

EXISTING EVIDENCE 39

2.1 A Review of Travellers’ Accounts and Scholarly Research Carried

out in the Region 39

2.1.1 Accounts of Travellers’ 39

2.1.2 Excavations and Field Research in the 20th Century .. 44 2.2 Prehistoric Occupation in the Plain 48

2.2.1 Neolithic and Chalcolithic Occupation - Sites ofB ağbaşı and

others 50

(a) Late Neolithic and Early Chalcolithic 50

(b) Middle Chalcolithic 51

(c) Late Chalcolithic ...... 54

2.2.2 The Early Bronze Age Village of Karataş-Semayük ... 59

(a) The Early Bronze Age I (Karataş Village I and II).. 60

(b) The Early Bronze Age II (Karataş Village III-V) ... 63

(c) The Early Bronze Age III (Karataş Village V I)...... 71

2.2.3 Middle Bronze Age Evidence from Bagbaşı and

K arataş...... 72

2.3 Iron Age Evidence from the Plain ...... 74

2.3.1 B ay ın d ır Tumuli ...... 77

(a) Tumulus C ...... 77

(b) Tumulus D ...... 78

(c) Karaburun Tomb III ...... 79

(d) Karaburun Tomb IV ...... 79

2.3.2 The Painted Chamber Tomb of Kizilbel ...... 80

2.3.3 The Painted Chamber Tomb of Karaburun II ...... 83

Karaburun Tomb I ...... 86

Boztepe Tomb ...... 87

Müğren Tomb ...... 87

2.4. Classical Evidence from the Plain ...... 89

2.4.1 Sepulchral Evidence ...... 90

VI 2.4.2 Sculptural Evidence 91

2.4.3 Epigraphic Evidence 92

2.4.4 Numismatic Evidence ...... 93

2.5 The and The Early Principate (334bc - 43ad) 95

2.5.1 Epigraphic Evidence ...... 100

2.5.2 Numismatic Evidence ...... 101

2.5.3 Archaeological Evidence ...... 103

2.6 The Roman and Late Roman Periods ...... 107

2.6.1 Literary Evidence (with reference to Chôma, and Komba)

...... 116

2.6.2 Epigraphic Evidence ...... 116

(a) Outside the Plain with reference to Chôma and Podalia

The Opramoas Inscription of ...... 117.

Honourary Inscription of Jason of Cyanis ...... 118

(b) Archaeological Settlements with Epigraphic Evidence from

the Plain ...... 118

O vacık ...... 118

Bayındır (Çağıltemeller), Yarangediği ...... 120

Asar Tepe ...... 120

Semayük ...... 120

G ökpınar ...... 121

Küçük Söğle, Büyük Söğle ...... 121

B alıklar D a ğ ı (Ördek Beleni) ...... 122

Kapıkaya ...... 122

Güğü (Yapraklı) ...... 123

vu Hacımusalar, Sarılar ...... 124

Islamlar, Aytaş Mevkii ...... 125

Gömbe 126

Tekke ...... 126

2.6.3 Numismatic Evidence ...... 126

(a) Gordianus III coins of Chôma ...... 127

(b) Gordianus III coins of Podalia ...... 127

2.6.4 Local Cults attested from the Plain 128

(a) 12 Gods 128

(b) Kakasbos 130

(c) Dioskouroi and Helene 131

(d) Trinity Gods 132

(e) Lagbene 132

2.7 The Early Christian and Byzantine Periods 137

2.7.1 Ecclesiastic Evidence (With reference to Chôma, Podalia and

Komba) 150

(a) Hierocles Synecdemus 150

(b) Vita of Nicholas of Holy Sion 150

(c) Basil II Notitia 151

(d) Notitia Episcopatuum 151

(e) o f Chôma, Podalia and Komba ...... 152

(f) Constantinus Porphyrogenitus de Thematibus ...... 154

CONCLUSION ...... 156

VUl VOLUME II

BIBLIOGRAPHY 167

APPENDICES 191

A. Glossary ...... 191

B. Passages cited from the travellers’ accounts ...... 194

C. List of references to the ‘Milyans’ in literary sources...... 200

D. Literary sources with reference to Chôma, Podalia, and Komba

...... 204

E. Ecclesiastic Sources with reference to Chôma, Podalia, and Komba

...... 205

F. Epigraphic evidence from the Elm alı Plain ...... 209.

G. List of Onomasticon (as attested from the epigraphic evidence from

the plain) ...... 227

MAPS

PLATES

IX LIST OF MAPS

MAP 1 Elmalı Plain, Harita Genel Müdürlüğü, Ankara 1: 200 000

MAP 2 Elmalı Plain, Geological Map, M.T.A., Ankara 1; 500 00

LIST OF PLATES

PLATE 1 A view of the Kızlar Sivrisi (3086 m.) from the Beydağı Mountains,

overlooking the dense pine and cedar forests

A view of the Elmalı Plain and its surrounding mountains overlooking the

Avian Pass

PLATE 2 Lake Avian, prior to its drainage

Agricultural activity along the base of the plain

PLATE 3 An Elmalı House from the town

Mudbrick house from the villages of the Elmalı Plain

PLATE 4 Geobotanical Outline Map of the Near East

PLATE 5 Cedrus Libani trees reaching heights of 30 m. and age of ca. 600

PLATE 6 A Cedrus Libani from Çığlıkara, Elmalı

Typical cones from a Cedrus Libani

PLATE 7 Location of milestones and reconstructed Roman routes

PLATE 8 Map of Northern Lycia with ancient sites

PLATE 9 Remains of an ancient road from the skirts of Balıklar D ağı Remains of an ancient road (Roman ?) north of Elmalı

PLATE 10 Map of western and central Lycia with tentative reconstruction of Roman

road network

PLATE 11 Roman routes according to the Tabula Peutingeriana

PLATE 12 View towards the Avian Pass

View towards the Sinekçibel Pass

PLATE 13 Mounds from the Elmalı Plain

PLATE 14 Traditionally built beehive from Geçmen

PLATE 15 Map of Elmalı Plain showing areas surveyed by J. Mellaart

PLATE 16 Map of Elmalı Plain showing areas surveyed by C. Eslick

PLATE 17 Chronological Chart

PLATE 18 Elmalı Plain, Late Neolithic/Early ChalcoUthic sherds

PLATE 19 Elmalı Plain, Middle Chalcolithic sherds

PLATE 20 B ağbaşı, Late Chalcolithic remnants of wall built in mud slab technique

PLATE 21 Bağbaşı, Late Chalcolithic pottery

PLATE 22 Bağbaşı, Late Chalcolithic terracota objects

PLATE 23 Karataş Chronological Chart

PLATE 24 Village of Karataş, Periods H II

PLATE 25 Plan of the house and surroundings, level I

PLATE 26 Karataş, EB I Burial of AQ

PLATE 27 Karataş, EB I Pottery

PLATE 28 Karataş, EB I burial jar with incised decoration o f‘wooden hut’ sign,

goats, swásticas, and plant.

PLATE 29 Village of Karataş, Periods IV

PLATE 30 Village of Karataş, Periods V

XI PLATE 31 Comparative chart of Karataş Houses

PLATE 32 Yuvalılar, wattle and daub house construction

PLATE 33 Yuvalılar, wattle and daub house construction

PLATE 34 Karataş, EB II pithos and jar burials with circular grave markers

PLATE 35 Karataş, EB II Burials

PLATE 36 Karataş, EB II Pottery

PLATE 37 Karataş, EB II Pottery

PLATE 38 Karataş, krater-like vessel with tubular spouts

PLATE 39 Karataş, EB II/III Transitional Pottery

PLATE 40 Village of Karataş, Periods VI

PLATE 41 Karataş, EB III pottery

PLATE 42 Karataş, anthromorphic vessel

PLATE 43 Karataş and Bağbaşı, Middle Bronze Age Pottery

PLATE 44 Karataş, Early and Middle Bronze Age Burial jars

PLATE 45 Karaburun III Tumulus; dinos like urn which contained the cremation,

bichrome ware

Bağbaşı, Iron Age sherds Bağbaşı

Iron Age, bichrome painted sherds

PLATE 46 Bayındır Tumulus D, in-situ finds, silver finds

B ayındır Tumulus D, ivory and silver figurines

PLATE 47 Kizilbel Tomb, isometric view

K izilbel Tomb, fi’ont facade

PLATE 48 K izilbel Tomb, inside southeast comer, with the mythical scenes of

Gorgons, Medusa, Pegasus, Chrysaor and the ‘departure of the hero’

scene.

XU PLATE 49 Kxzilbel Tomb, inside, the winged female figure

Detail fi’om the ‘departure’ scene

PLATE 50 K izilbel Tomb, inside, boar hunt in the marshes, sea-going ship

PLATE 51 Karaburun II Tumulus, fi’ont facade, and front facade with retaining walls

Karaburun I, sarchophagus

PLATE 52 Karaburun II Tumulus, inside, main scene with the reclining dignitary

PLATE 53 Karaburun II Tumulus, inside south wall, the ‘ekphora’ scene

PLATE 54 Karaburun II Tumulus, inside north wall, the ‘battle scene’

Yalnxzdam Stele

PLATE 55 Müğren Tomb, completely robbed of its tumulus and reused as a church

PLATE 56 Islamlar, Lycian type of rock-cut tomb

Eskihisar, Lycian type rock-cut tomb

PLATE 57 Islamlar, rock-cut stele with banquet scene

Islamlar, sarchophagus with sculptural relief

PLATE 58 Kxzxlca rock-cut tomb with inscription

The Kxzxlca inscription

PLATE 59 Gilevgi, hillfort site with sketch plan

PLATE 60 Avian Gölü, hillfort site

Armutlu, hillfort site

PLATE 61 Hellenistic coins of Chôma, with the rider-god Kakasbos

Gordianos III coins of Chôma

Gordianos in coins of Podalia

PLATE 62 Kyaneai, Jason-inscription mentioning Chôma and Podalia

PLATE 63 Map of Elmalx Plain showing the main sites with epigraphic evidence

PLATE 64 Ovacxk, hilltop settlement, a view and sketch plan

xm PLATE 65 B alıklar D a ğ ı, gabled lid sarchophagus

Kapıtaş, rock-cut stele

PLATE 66 Güğü-Yapraklı (Buzağı Kayası)

Güğü-Yapraklı, ‘Trinity Gods’

PLATE 67 Islamlar, Dioskouroi and Helene

Islamlar, Aytaş rock relief

PLATE 68 12- Gods plaques from Gömbe, and Ahlatlıköy

PLATE 69 12- Gods plaques from Elmalı

PLATE 70 Map of Elmalı Plain showing main sites with evidence of local cult

worship

PLATE 71 Sarılar, sarchophagus lid with gabled roof

Girdev, lion sarchophagus.

HV LIST OF FIGURES

Page

FIGURE 1 Sketch map of the Elmalı Plain 7

FIGURE 2 Mean Annual Temperature (1953-1990) 15

FIGURES Daily difference between maximum and minimum temperature .. 15

FIGURE 4 Mean monthly precipitation ...... 16

FIGURE 5 Mean number of days with precipitation of 0.1 mm or more 16

FIGURE 6 Water balance ...... 17

LIST OF CHARTS

CHART 1 Tentative correlation of the vegetation history as taken from the pollen

diagrams for the region ...... 23

CHART 2 Crop yield for the Elmalı Plain as taken from the statistic accounts of

D.I.E., Ankara ...... 25

XV LIST OF ABBREVIATIONS

AA : Archäologischer Anzeiger

ActaArch : Acta Archaelogica (Kobenhavn)

AJA : American Journal of Archaeolgy

Anatolica : Anatolica. Annuaire international pour les civilisations de Г Asie

antérieure

Annuario : Annuario della Scula Archeologica di Atene delle Mission! Italiane in

Oriente

AntW : AntikeWelt. Zeitschrift fur Archäologie und Kulturgeschichte

AnzAW : Anzeiger für die Altertumswissenschaft

AnzWien : Österreichischer Akademie der Wissenschaften, Wien;

Philologischer-historische Klasse

AS T,: Araştırma Sonuçları Toplantısı

AthMitt : Athenische Mitteilungen

AS Atene ; Annuario della Scuola archeologica di Atene e delle Mission! italiane

in Oriente

BCH : Bulletin de Correspondance Hellénique

Belleten : Türk Tarih Kurumu, Belleten

BMC Lvcia : A Catalogue of the Greek Coins in the

BSA : Annual of the British School of Athens

BZ : Byzantinische Zeitschrift

САН : Cambridge Ancient History

XVI CIG : Corpus Inscriptionum Graecarum

CRAI : Comptes rendus des séances de 1’ Académie des inscriptions et belles-

lettres (Paris)

D.I.E. : Devlet istatistik Enstitüsü

D M I : Devlet Meteoloji istasyonu

DOP : Dumbarton Oaks Papers

D.S I. : Devlet Su işleri (State Water Power)

Elmalı - Karatas I : C. Estlich, Elmalı - Karatas I : The Neoloithic and

Chalcolothic Periods : B aabasi and Others Sites (BrynMawr

College 1992).

Elmalı - Karatas II ; J.L. Warner, Elmalı - Karatas II : The Early Bronze Age

Village of Karatas (Bryn Mawr College 1994).

EpigAnat ; Epigraphica Anatolica. Zeitschrift fur Epigraphic und Historische

Geograhpie Anatoliens

IGRR : Inscriptiones graecae ad res romanas pertinentes

IstForsch : Istanbuler Forschungen

IstMitt : Istanbuler Mitteilungen

IstMitt-BH ; Istanbuler Mitteilungen Beiheft

Jdl ; Jahrbuch des Deutschen Archäologischen Instituts

JFA ; Journal of Field Archaeology

JHS : Journal of Hellenic Studies

JMA: Journal of Mediterranean Archaeology

JNG: Journal of Numismatik und Geldgeschichte

JÖBG: Jahrbuch der Österreichischen Byzantinistischen Gesellschaft

JRS. Journal of Roman Studies

xvu K S T, Kazı Sonuçları Toplantısı

Lvcian Symposium I I ; J. Borchhardt and G. Dobesch (eds.) Akten des n.

Internationalen Lvkiens-Symposions Wien 6-12 Mai 1990

(Wien 1993)

M.T.A. : Maden Teknik Arama Enstitüsü

N Ç : Numismatic Chronicle

O D B : Oxford Dictionary of

Princeton The Princeton Encyclopedia of Classical Sites (ed. R. Stillwell et al.

Princeton 1976)

PPS: Proceedings of the Prehistoric Society

RA: Revue Archéologique

RE; Pauly-Wissowa, Real-Encyclopädie der klassischen

Altertumswissenschaft

R E G : Revue des études grecques

Reisen I : О. Bendorf und G Niemann Reisen in Südwestlichen Kleinasien I :

reisen in Lykien und Karien (Wien 1884)

Reisen II ; E. Petersen und F. von Luschan Reisen in Südwestlichen Kleinasien II

: Reisen in Lvkien. Milvas und Kibyratis (Wien 1889)

RRMAM : D. French Roman Roads and Milestones in Asia Minor BAR

International Series (1988)

SEG : Supplementum epigraphicum graecum

Sur La Lvcia Antique : H. Metzger (ed.) Actes du Colloque Sur la Lvcia Antique « Bibl. de L’Institut Français d’ etudes Anatoliennes d’.

TürkArkDerg : Türk Arkeoloji Dergisi

ТАМ : Tituli Asiae Minoris

XVUl T.T.K : Türk Tarih Kongresi Bildirileri

TTKY: Türk Tarih Kurumu Yayınları

ZfN: Zeitschrift fur Numismatik

ZPE: Zeitschrift für Papyrologie und Epigraphik

ABBREVIATIONS USED FOR LITERARY SOURCES

App. Bciv.: Appian Bella Civilia

Arist. Hist. An.: Aristotle Historia Animalum

Arr. Anab.: Arrian Anabasis

Cass. Dio.; Cassius Dio

Cic.Verr.: Cicero (Marcus Tullius) In Verrem

Diod. Sic.: Diodorus Siculus

Hdt.: Heredotus

Livy, Epit.: Livy, Epitomae

Pliny f f ii: Pliny (the Elder) Naturalis Historia

Pint.: Plutarch

Polyb.:

Ptol. Geog.: Ptolemaeus Geographia

Quint. Smym.: Quintus Smymaeus

Steph. Byz.: Stephanus Byzantinus

Strab.:

Theophr. Hist. PI.: Theophrastus Historia Plantarum

Vitr.; Vitrivius

XIX INTRODUCTION

The purpose of this thesis is to place the Elmalı Plain in its geographical, environmental, historical, and archaeological setting in the light of existing evidence, including a comprehensive survey of traveller’s accounts, historical and archaeological research and studies carried out in the region.

The Elmalı Plain is located in the southwestern part of Asia Minor, in the Teke peninsula, an area referred to as Northern Lycia. It is a typical southwestern Anatolian

alluvial plateau, with an elevation slightly above 1000 meters. It lies in the western

extension of the Taurus Mountains, in a transitional zone between coastal Lycia and

the southwestern ridge of the central Anatolian Plateau. The geological formation of

the plain has been mainly taken into consideration while determining the borders of

the area under study, whose natural boundaries are defined by the surrounding

mountains, which are the geological results of Mesozoic uplifts and foldings of the

Teke peninsula, and seem to have played an important role in the isolation of the

upland basins, forming various micro-environments whose economies possibly

depended on each other, as well as on coastal Lycia. The passages and river valleys,

on the contrary, form a number of natural links between these upland plains, which

play an important role in the establishment of communication and trade links, as weU

as cultural influences.

The Elmalı Plain has been selected as the focus of this study for several reasons:

first, the Hacimusalar Excavation and Survey Project led by the Department of

Archaeology and Art History of the Bilkent University under the direction of Dknur

Özgen, is located in the southern part of the ElmaliPlain. Second, although an

important series of excavations and surveys have been conducted by the Bryn Mawr

College with regard to the Prehistoric and Iron Ages of the plain, there is no

comprehensive research which has compiled all existing evidence at this level. A third reason is the vital importance of these upland transitional zones, especially in throwing light upon the history of Lycia as a larger region, as an integral part of its economy and later administrative system.

This thesis attempts to compile a wealth of information from different sources which includes both the enviromental and historical setting of the plain. It does not

aim to be a complete study of all its environmental, historical and archaeological

factors, which is the subject of the ongoing project of Hacxmusalar, and will be

pursued in depth by the specialists of their topics. The aim of this study has been,

rather, to pull together the accounts and research carried out in the region so far, to

draw conclusions based on the available evidence, and to understand the kind of

problems and gaps existing in the archaeological record that could be pursued further.

Organisation:

The thesis is in two parts: The first part discusses the environment and

geographical setting of the plain and its borders, followed by a discussion of the

modem geography, climate and vegetational composition of the plain, and a

discussion of its modem as well as ancient routes of communication.

The second part gives a historical and archaeological survey of the plain. It begins

with an overview of the history of modem travellers and scholarly research done in

the area. The accounts of the European travellers, beginning in the 19th century,

provide a very valuable source of descriptive narratives about the environment and

settlements which closely reflect the earlier times, and which are irreversibly

disappearing every day. After a discussion of the history of the scholarly research, a

chronological discussion of the plain based on the existing evidence is given. The

prehistoric sequence followed is the Neolithic, Chalcolithic, Bronze Ages; the

historical periods are the Classical, Hellenistic, Roman, Late Roman and Byzantine.

The Islamic Period will remain outside the scope of the thesis. Although the Islamic Period is an important period in understanding the general history of the plain, the data available for the period needs to be looked at from a different perspective, and necessitates an understanding of Ottoman sources.

Sources and research:

The evidence regarding the first part of the paper has been compiled from a number of sources including various State Institutions in Ankara. The geological data comprises the memoir and geological maps prepared by the M.T. A.(State Mineral

Research and Exploitation Institute), reports prepared by the D.S.I. (State Water

Institution) and the unpublished geological report of Sancar Ozaner. The climatic data has been gathered from Meteoroloji Genel Müdürlüğü, Ankara (General Directorate of Meteorology). For the chapter regarding vegetation and resources of the region, statistics prepared by D.I.E (State Statistical Institution) and reports prepared by the

Tarim Köy işleri Bakanlığı (The Ministry of Agricultural and Village Affairs), as well as publications of the State Forestry Department have been referred to. For modern routes the Antalya Map of Karayollari(State Highways Department) and the

1:200 000 series of the Harita Genel Komutanlığı Map (which was revised in 1943 on 19th century copies, and is out-of-print today) have been used.

Major sources consulted for the second part of the thesis include a number of 19th

century travellers’ accounts, and for the prehistoric occupation of the plain the

principle sources have been the articles and publications of the Karataş-Semayük and

B ağbaşı excavations have been referred to.Excavation reports, as well as discussions

in other works have been consulted for the Iron Age burial tumuli of Bayındır, and

the K izilbel and Kararaburun Tombs. The main types of evidence for the historic

periods are as follows;

• Literary evidence from ancient sources

• Epigraphic evidence

• Ecclesiastic sources (eg. Hierocles, Notitiae Episcopatum, Lists etc.) • Numismatic evidence

• Archaeological evidence

Three ancient cities which have been identified within the borders of the Elmalı

Plain have been discussed for their ancient and epigraphic sources. These cities are

Chôma (identified as the Hacimusalar Mound)’, Podalia (identified as a site west of

Avian Lake, as well as the mound site of Küçük Söğle)^ and Komba (identified as the modem town of Gömbe, located at the very southwest extremity of the plain).^

Limitations and problems encountered:

The environmental setting of the plain has been investigated in a very general sense, but there are a number of issues which need to be further pursued in order to throw more light upon our knowledge of how human groups have utilized and adapted to their immediate environment, for subsistence reasons.

One of the issues of primary importance is to define the ancient limits of the two lakes, namely Avian Gölü and Karagöl. These lakes, which occupied a large part of the plain, have fluctuated at regular time intervals over the past 100 years. The information would also be valuable for our understanding of the settlement patterns and the amount of arable land that could be utilized by the human communities.

Another important issue would be to do a pollen analysis, in order to reconstmct the

ancient vegetation of the plain,'* and would help us to understand the environmental variables available to its ancient inhabitants.

’ Bean and Harrison, 1967: 38-43. From this point on it will be referred as Hacimusalar-. ^ Reisen II: 161-162 identified it with a site near Avian Lake, known as Buralya; Bean, 1968 :157- 163, has prefered the the Söğle Mound. ^ Heberdey, 1898: 157-158. '* Such a research has been done for the Söğüt Plain (ca. 60 km. north of Elmalı Plain) which yielded a pollen diagram chart for the region, and an outline of the paleo-environment of the area has been reconstructed The research also included a number of surface samples fiom the Elm alı Plain Zeist et. al., 1979: 55-142. Two cote samples from Karagöl and Avian Gölü have been taken by Bottema and Woldring, 1986: 140-43, but these samples have not been dated based on C14, and remain at a limited time range compared to the Söğüt pollen diagram. The archaeological evidence from the Elmalı Plain is scanty, and unequally distributed through the various periods. Thanks to the Bryn Mawr College excavations under the direction of M. J. Mellink, we have a better understanding of the prehistory of the plain. Surveys have been conducted regarding the Neolithic,

Chalcolithic and Bronze Ages, which give us a general idea about settlement patterns during these periods, and shifts which occured in antiquity. The Late Bronze Age, in general, resembles a period of hiatus for the whole plain. The Iron Age on the contrary, displays a very rich and varied burial tradition, but without evidence for settlements. For the later periods of the Classical Age, Hellenistic, Roman and

Byzantine Periods no systematic research has taken place. The evidence regarding these periods includes the identification of certain classical cities, and the results of epigraphic surveys of the region since the 19th century. One of the reasons for the neglect of the region by scholars such as W.M.Ramsay and L. Robert might possibly be their interest in the rich sepulchral tradition of the rock cut monuments, which were abundantly represented in the coastal regions of Lycia and the upland regions of

Kibyra, Kabalia and Pisidia, but are visibly missing in the Elm alı Plain region. Also the absence of classical cities with monumental buildings may possibly have played an important role in the neglect of the region.

The ongoing project of the Hacimusalar excavations and survey promises to be an important project that will throw more light upon these periods. The preliminary results of the excavations have pointed to a wealth of pottery evidence for the

Classical, Hellenistic, Roman, Late Roman and later periods. Another problem encountered for the region is the lack of comparanda material, especially for the pottery finds, since no excavation regarding later periods has been conducted in the plain, and no pottery publications exist for the later periods. The Hacimusalar excavation will have a pioneering role for the establishment of a pottery sequence for the plain. CHAPTER I. GEOGRAPHICAL AND ENVIRONMENTAL SETTING OF

THE PLAIN

1.1 THE GEOGRAPHY OF THE PLAIN

The Elmalı Plain is a typical southwest Anatolian alluvial plain or polje depression, situated in the highlands of northern Lycia, dating to the Holocene period.^ The coordinates of the plain in a south-north direction are, 36° 32' - 36° 54 ', and in a west-east direction 29° 40' - 30° 0 5 ' The plain has a southwest-northeast orientation between the horse-shoe shaped mountain massifs of Akdağ and

Beydaglari. These moimtains separate the ecologically distinctive region of

Coastal Lycia from the northern highlands of central and northern Lycia. The provincial capital of the Elmalı Plain is Elmalı with a permanent population of 13,

202.’ Other large towns in the plain are Akçay and Gömbe.

The plain is separated into two main parts by two mountain ranges. Nohutlu

D ağı in the south, and Elmalı D ağı to the north, which converge on each other, leaving a narrow passage at the skirt of the town of Elmalı (see fig. 1).

The southern part of the plain: It lies at an altitude between 1030-1050 m.®, and is bordered by the following mountains; to the north is the Elmalı mountain, whose

^ According to the geophysical map of M.T. A of 1:500 000, Denizli sheet (refer to Map 2). ® The coordinates have been taken from the 1947 Harita Genel K om utanlığı Map of 1:200 000, Elmalı sheet (refer to Map 1). ’ 1997 census figure provided by D.I.E. Ankara, Türkiye. The statistic gives a total population of 35 662, for the province of Elm alı, with 22 460 people living in its villages. * Karagül Avlan I* *roiesi T aşkın Tesisleri Planlama Raporu, 1970:l.The Elmalı Plain is at an attitude of 1050 meters, with its surface slightly sloping towards east and south, the old surface levels of Karagül and Avian gülü, being at an elevation of 1030 meters. .'-UNTUV-.:,

G peak reaches 2505 m.; to the west are Elbisdag, Yumrudag and Akdağ with peaks reaching heights o f2604 m., 2741 m. and 3024 m., respectively. The southern edge of the plain is bordered by Susuz D ağı whose summit reaches 2465 m., and which meets the massif of the Beydaglari (or the ancient Massikytos®), also known as the

Eastern Akdağ. The Beydağları massif have several summits, the highest reaching

3070. To the east of the plain is Nohutlu D ağı, also adjoining the Bey D ağları, whose peak is at a height of 1707 m.(fig. 1)’°

The upland plains of the region have been rendered highly productive by drainage

and irrigation systems, and their climate is also suitable for the growth of a variety of

agricultural products. The plain of Elmalı especially proves to be one of the most

fertile with its products of vines, and fiuits, mainly apple, cereals and sugar-beet."

The town of Elmalı as the provincial capital, and with the largest regional market,

provides an outlet for agricultural products from the hinterland.

The northern part of the plain: It is in the form of a narrow valley to the north

of the town of Elmalı, and is bordered by the massif of Beydağları. It is also a polje

depression from the Holocene'known as the Golova or Gilevgi Polje, named after

its largest town Gilevgi, which is situated at the northeast edge of the plain. The

Gilevgi upland basin is 60-70 metres higher than the Elm alı Plain.'^ The north-

northeast orientation of the plain also provides one of the main lines of

communication, connecting the region of Elmalı to Korkuteli, and on to the southern

edge of the Anatolian Plateau.

® Pliny Ш V.28,10, Ptol. Geog. V.3.1 and Quint Smym. Ш. 234,8,107. The heights have been taken from the Harita Genel Kom utanlığı Map of 1:200 000, Elmalı sheet (refer to Map 1, also see sketch plan of the Elm alı Plain in fig. 1). " Dewdney, 1971: 177. See the geophysical m ^ 1. KaragQl Avian Projesi T a şk ın Tesisleri Planlama Raporu, 1970: 1. Recent improvements of the Elmalı-Korkuteli-Antalya route have provided another improved link to the Mediterranean coast. Other natural passes providing important lines of communication to the coastal regions of Lycia exist through the

Akdağ and Beydağ massifs, which border the plain along its western and southern parts.'''

Modern Geology:. 15

As mentioned, the plain of Elmalı is a polje depression, a basin of internal drainage that has formed within the karstic topography of the Western Taurus

Mountains. These comprehensive mountain ranges are of Mesozoic - Tertiary formation and they have been uplifted with recent tectonic movements.'® The oldest formation in the region is the autochthonous Mesozoic series (see Map 2). They form the Teke coastal ranges between the Elmalı - Akdağ mountain group and Fethiye, between the lower course of the Esen Çay and , and represent the extension of the autochthonous series of Beydağları and Korkuteli, which are situated further east

and shelter the upland plains of northern Lycia, including the Elmalı Plain. These

Mesozoic and Early Tertiary limestone formations, referred to as the ‘Comprehensive

series’ in the Denizli Sheet of the 1:500 000 geophysical map, precede the nappes

(overthrusted folds) which themselves precede the Eocene flysch.17

When viewed today, the Western Taurus Mountains form an impressive mountain

mass with several peaks above 3000 meters.'* This mountain range is characterized by

Discussed in detail on pg. 30-38. General information regarding the Elmalı Plain and the formation of surrounding mountain massifs of the Teke peninsula have been published in the 1 :S00 000 Denizli Sheet of the Türkiye Jeoloji Haritası and its accompanying memoir Pamir and Erentöz, 1974 ; 64-68. Additional information has been gathered from Sancar Qzaner’s unpublished report regarding the geomorphological setting of the Elmalı Polje and Hacimusalar Höyük, during his investigations with Bilkent University in the years of 1993 and 1994. Also Platt, 1994 :11-12, has been refered to. Geological terminology is explained in the glossary (.^^endix 1). '® МГА 1:500 000 geophysical map, Denizli sheet (refer to Map 2). Pamir and Erentöz, 1974:64-68. '* Zeist e t al., 1979: 56. The Western Taurus Mountain.«! referred to here are the mountains extending from Muğla in the west to Anamur in the east; see Brunn et. al., 1971: 225. 10

the fact that individual chains are not parallel to the main mountain ranges, but run approximately perpendicular to it. The strikes have a northeast-southwest orientation, as is the case with the Lycian Taurus.'® This phenomenon has been pointed out as the result of a complicated orogenic structure, referred to as the Courbur dTsparta' The presence of the nappe systems have been related to this orogenic structure.^®

Between İsparta and Fethiye, mountain massifs of Mesozoic limestone extend over

150 km., with a number of karstic polje depressions that are separated from each other by passages. The Elmalı Plain constitutes the southernmost part of these karstic depressions. The southwestern part of the plain has a northeast-southwest as well as a northwest-southeast orientation, in geological terms, with a distance reaching 30 km. in both directions. The actual width of the plain varies between 2-15 km. The extension of the two opposite geological orientations possibly points to the main lines of folds and fractures that existed with the earlier orogenic formation of the area. The present shape of the plain is a result of the karstic activities, involving the dissolution of the older carbonates by means of subsurface and underground drainage along the weaker fractures.^'

The surrounding bedrock of the Elmalı Plain consists largely of limestone, dolomite, shale and sandstone. Since the Miocene period these rocks have been above sea level and subject to erosion under warmer and wetter climates than at present.

Limestone, dolomite and shale tend to have low permeability. Most of the downward percolations of rain water seem to have occurred along fractures. The dissolution of the softer limestone and dolomite type of rocks has produced ‘dolines’ (sink holes),

caves and underground drainage instead o f surface drainage^^, which points to heavy

'® i.e. the Akdag and Beydaglari series of the Lycian Taurus Mountains, part of the Western Taurus Mountains. “ Zeist et al., 1979,57; Brunn et al., 1971; 225-55 gives a detail account of the geological structure of the Western Taurus mountains. Information firom S. Ozaner. “ Platt, 1994: 11. “ Ibid, 12. 11

karstic activity in the region.^'' When the hills were covered by plant foliage (most likely during drier times), alluvium, gravel, silt, and clay were deposited by streams on the irregular bedrock formed by the karstic activities, all of which resulted in the formation of the wide, flat, fertile mountain plains, typical of southwest Anatolia. ^

The southern and eastern slopes of the plain are flanked by steep limestone hills.

The western and northern slopes are flysch formations. The flysches are located on the higher slopes, above the alluvial fans, which make up the lowest part o f the hills that join the basin of the plain. The flysches comprise clay and sandstone layers which have a high permeability of water, and provide a good base for the village settlements, whereas the alluvial fans below the flysch settlements provide good agricultural land

for its inhabitants.^®

Major Water Sources:

The largest river of the plain is the Akçay River, which is 41 km. long. It originates

from the Gömbe and Beypinar springs in the Akdağ mountain massifs that border the

southwestern-western edge of the plain. ^ After leaving its source for ca. 10 km.,

Akçay enters the plain fi’om the southwest, and passes Hacimusalar mound at a

distance o f750 meters to its south. Running along the southern edge of the plain it

reaches the locality of the recently drained lake of Avian. The Akçay in the past was

one of the main sources of the Avian Lake.“

“ The dolines are mainly located along the eastern border of the plain, which comprises older limestone fractures of the Upper Cretaceous. The most well known is the ‘Bityiik Düden’, the Large Doline, with a del» of 10m3/sec, and a second is the YakaçifUiği Doline with a debi of Sm3/sec. Klmal 1 Akçav and Plains Hydrogeological Investigation Report, 1978: 22,28. The karstic activities of limestone still continue in the high attitudes of Susuz D ağı and Elmalı D ağı. Recent karstic activity has been observed on Susuz D ağı, where a doline o f400 meters in diameter and 500 meters in de(Xh has been located, the place known as Dokuzgöl Çukuru or the ‘Nine Lake Pit’ (information from S. Qzaner). “ Platt, 1994:12. According to the Elm alı. Akcav and Demre Plains Hydrogeological Investigation Report, 1978:7, the alluvium has been described as sancty and gravelly near the stream beds and cityey near the Eocene flysch. “ Information from S. Cteaner. ^ This river m ^ be the ancient Aedesa mentioned by Pliny V. 101, as identified by Bean and Harrison, 1%7: 34. “ Karagül Avlan Pmiesi Taşkın Tesisleri Planlama Raporu , 1970: 1. 12

The second important river of the plain is the Güğü or Eskihisar River, which enters the plain from the northwest and runs south into the Karagöl, providing its main water source. The river bed tends to dry out during the summer months, its main source of water being the snow from Akdağ massif.^

These rivers, having formed natural passages into the plain, provide access into the plain, especially through the mountains to the northwest, west, and southwest.

The two major lakes of the plain were Karagöl and Avian Gölü, which showed an ephemeral character and were partially marshland. They were dried by the D.S.I. in

1966 and 1974, respectively.^' The Avian Lake located along the southeast of the plain covered an area of 840 ha. before it was drained. Karagöl, located to the northwest of the plain, a lake with a larger swamp area around it, covered an area of

2440 ha.^^

Both of the lakes, as possibly was also the case in antiquity, played an important role in the subsistence of the inhabitants of the plain. The so-called ‘boar hunt’ scene

(Plate 50A), and the sea-going vessels (Plate 50B) from the K izilbel Tumulus

Burial, located to the north of the recently dried Karagöl lake have been suggested as

evidence for the subsistence of its inhabitants in antiquity.“ The draining of the lakes

Ib id , 1 30 See pg. 30-38 for a discussion of these routes. Karagöl Avlan Projesi T aşkın Tesisleri Planlama Raooru, 1970; i-v, 49-53. The decision to drain the lakes was taken in order to control the marshy land and the floods in the region, as well as to increase the amount of arable land in the plain. The lakes were drained by a series of drainage canals and subsurface tuimels ( eg. the S.S km. long Avian Tuimel), which cormected Karagöl to Avian Lake and later to the Büyük Düden and Beşgöz Dere valley. In order to prevent the water of the Akçay from filling the Avian Lake again, the river was also connected to the Avian Tutmel, and by means of the YakaçifUiği Düden was carried into the underground drainage system. “ Ibid, 2. Mellink, 1998; Plates XXTV, XTV and XV. For a more detailed discussion of the tomb, see herein page 82, footnote 329. 13

has altered the eco-system of the region considerably.^'* The recent reconsideration by the government authorities to fill the lakes to their former levels is a significant development for the region.

Other important water resources for the plain are its natural springs and low level wells, ca. 800 in number, which are concentrated to the west of the plain, in the area of the recently drained Karagöl lake and the Akçay valley. These wells are mainly used for the irrigation of the gardens and agricultural land.^^ One of the two important springs in the plain is located to the west of Balıklar D ağı and the other is the

Kazanpinari Spring located between the villages of Mursal and Kuzköy. The spring located west of Balıklar D ağı has a cumulative debi of 2.5 m3/ sec. and the

Kazanpinari Spring has a cumulative debi (the standard measurement for water

flow) of 0.1- 0.3 m3/sec. They appear to be between the older Mesozoic limestone formations and the Eocene flysch, from which they can be considered as karstic

springs. Both of the springs are mainly used for drinking as well as irrigation

purposes 36

Further discussion of the negative effects of their drainage on the climate as well as vegetation will be given under ‘Vegetation’, pages 26-27. Elm alı. Akcav and Demre Plains Hydrogeological Investigation Report, 1978: 26. ^ Ibid., 25, also see chart on pg. 12. A detailed research regarding the surface and subsurface water sources and their annual yields in the region has been conducted by the D.S.I. in the same publication. This publication can be referred to for further information regarding the nature and quality of the plains’ water resources. 14

1.2 CLIM ATE 37

The climate of the Elmalı Plain can be described as transitional between the coastal Mediterranean climate and the continental climate of the Central Anatolian

Plateau. The transitional character plays an important role in the vegetational make up of the region and the growth of certain agricultural products, especially cereals, which are more suitable to the continental climate of the interior plateau.

The coastal areas of Lycia, display a typical Mediterranean climate. The winters are mild with temperatures rarely dropping below freezing point. The summers are hot and almost without rain. Precipitation occurs in autumn, winter and spring from

October to April. Though there is no rain in the summer, humidity remains high due to proximity to the sea.^*

In the interior of Lycia the rate of precipitation decreases, with the annual rate for

Elmalı being around 500 mm. The northeast-southwest orientation of the mountain massifs and its polje depressions in the Lycian Taurus is important for the precipitation in the area, since the rains are usually carried by the south and

southwestern winds from the Mediterranean.^®

Climatic data regarding the region is available from four different meteorological stations, including the Elmalı and Akçay stations which function under D.M .I., and Gömbe and Güğü Meteorology stations under D.S.I. Annual statistics for the region exist from 1958-1990. The data for the graphs presented in this section has been taken from, Eievlet Meteoloji işleri Genel Müdürlüğü, Elmalı (Antalya) Ortalama ve Ekstrem Kıymetler Bülteni (Ankara, 1998). The article ‘Physical Geography and Settlements’ by Wagstaff, 1973; 197-215, has been used as a model for the information presented in this section. The annual precipitation rate for Fethiye, for example, is 1004 mm. ’®Zeistet. al., 1979: 57. 15

Fig. 2. Mean Annual Temperature, 1953-1990

The mean annual temperature (fig.2) for Elmalı is 12.1° C. July is the hottest month, with an average of 24° C. In June through August the temperature is over 20°

C, the months of April, March, September and October have an average over 10° C.

An average below 10° C. is seen with the months November till April. The month of

January constitutes the lowest rate with an average of 2.2° C. The months of

December and February show an average of 15 days with temperatures lower than 0°

C. January displays 20 days with temperatures below the freezing point. In the coldest month of January the soil temperature up to 5 cm., has an average temperature of 2.6 ° C, slightly above the freezing line. The average number of days of snow cover for the month of January is five, and the greatest thickness of snow cover at 50 cm. occurs in the same month.

Fig. 3 Daily differences between the extreme maximum and extreme minimum temperatures, 1953-1990. 16

Daily differences between maximum and minimum temperatures (fig. 3) seem to be highest in August with 27° C. The summer months of July, June and August have a daily temperature difference of 23° C, 24° C and 27° C respectively, which characterizes the continental character of these highland zones

Fig. 4 Mean monthly precipitation, 1953-1990

Fig. 5 Mean number of days with precipitation of 0.1 mm or more

JFMAMJJASOND

The mean annual precipitation (fig. 4) for the years between 1953 - 1990 has been

given as 504 mm. Starting with the month of December through March the monthly

precipitation remains over 45-50 mm. With November and December this amount

reaches 92 mm and January displays the wettest month with the highest precipitation 17

rate of 101.7 mm. A fall from 50 mm to 30 mm is observed between the months of

March and April. Precipitation rates for the summer remain under 10 mm. with

August presenting the lowest rate at 6.9 mm. From September to October the amount of precipitation steps up from 8.0 mm. to 32 mm. The wettest period of the year seems to begin in October, lasting till May, with a quite gradual increasing and decreasing cycle throughout the year. The mean number of days with a precipitation rate of 0.1 mm. or more does not exceed 11 per month, which represents the amounts for December and January (fig. 5).

The annual average of relative humidity is 56 %. Between the months of October to May it remains over 50 %, with the highest rate reaching to 72 % in December and

71 % in January. The high humidity rates result in mist formation of 3-4 days in these months. In the summer months of June, July and August humidity remains around 40-

45 %, contrasting with the high humidity rates of the coastal regions, which make the highlands a preferable location for the inhabitants of the coastline zone during the

summer months.

The mean annual rate of evaporation (fig. 6), which is 1401.3 mm, is in great

excess over the mean annual rate of precipitation (=504 mm.). The plain is water

deficient between the months of May and November (fig. 6), and has its peak of

deficiency with the months of June, July and August. Substantial irrigation of the plain

depending on its groundwater resources, becomes critically important during this 18

period. A surplus of water occurs between the months of November and March, which plays a dominant role in the plant growth during this period. The winter planting of cereals in Elmalı takes place at the beginning and in the middle of October till the end of November, and the harvesting starts at earliest with the beginning of June and continues till the end of the month. Summer planting of cereals, though practiced by few villages, starts in mid April, and continues till the end of May. Harvesting takes place in late August.'“

The strongest winds predominantly seem to be the North wind, especially in the months of March, April, May, September and December, and the north-northeast wind in January. The rain-bringing winds from the southwest and west appear to be strongest in February, October and November. The highest frequency of winds are the south wind, followed by the west-northwest wind and the west-southwest wind.

The continental character of the upland basin, with its cold winter months and high diurnal temperature differences, has neccessitated substantial dwellings for the inhabitants of the Elmalı Plain, especially if they aim to survive the winter months on the plain. The traditional Anatolian highland architecture, of stone foundation with mudbrick superstructure and timber framework, is also characteristic for the Elmalı

Plain. Although the modem brick-cement architecture has taken over in the main town of Elmalı (Plate 3a), the traditional timber framed mudbrick architecture (Plate

3b) still prevails in the older part of the Elmalı town and is still the dominating form

of architecture for the villages in the plain,^' its thick walls providing for summer

coolness and for warmth in the winter.

Information from K6v Envanter Etüdü Antalya 1981 (Village Inventory Summaries) : Tables 15.1 and 15.2. According to the Kdv Envanter Etudlerine GOre Antalva tF.lmali B u c a d i, 1964; Table 48; building material used in the villages of Elm alı constitutes 89% of mudbrick, 10% of stone,; 68% o f the houses have pitched roo& and 31% have flat roofr. For heating purposes mainly wood and animal dung is used. According to the 1981 statistics of the same publication; Table 3:3; presented in numbers, 3877 houses built of mudbrick, 933 of stone, and 388 of brick, and in percentages as 76% , 17% and 7% respectively, showing the recent increase in the usage of bride in the region of E lm alı. 19

Excavations in Karataş - Semayük have proved that mudbrick architecture has been the dominant form of architecture since prehistoric periods for the upland plains.

From the archaeological point of view this becomes an important factor, which enables the recovering of the remains of substantial buildings from various periods, their mudbrick architecture having resulted in the accumulation debris of the mounds of the upper plains. A large number o f mounds exists in these upland plains (Plate 13), but there are none in the coastal plains. The lack of mounded sites from the coastal regions has been attributed to the usage of more ephemeral building materials, such as wood, that would be more suitable for its mild climate. The existence of multi-period

mounded sites in the upland plains is very important for an understanding of the

earlier history of the region.

1.3 VEGETATION

Modern vegetation:

The contemporary vegetation of the region has been described by Zeist''^ as being

within the Oro-Mediterranean Vegetation Belt (Plate 4), and surrounded by Alpine

Vegetation, which exists over 2000 meters. Between 800 and 1200 meters, deciduous

forests with coniferous forests occur. Up to 1000 meters coniferous forests are made

up of Pinus brutia (Calabrian Pine - Kızılçam), and Firms nigra (Black Pine -

Karaçam). In warmer climatic zones Firms brutia may occur above 1000 meters and

1200 meters. The dominant tree in the deciduous forests is Quercus cerris (Turkish

Oak - M eşe). Quercus calliprinos is frequently found in the shrub layers of Pinus

brutia and also forms an undergrowth for Firms nigra forests.

Zeist et. al., 1979: 65-67. The quaternary vegetation of Southwestern Anatolia including the upper highlands of Northern Lycia has been stucUed by Zeist, Woldring and Stapert in 1973. The investigations have included both surface sampling and sediment coring from lake sites in order to reconstruct a pollen diagram of the late Quaternary including the last glacial and the postglacial periods. Together with a general description of the contemporary vegetation has been given. The following information has been cited from there. 20

Between the upper forest zones of 1200-2000 meters, mainly coniferous forests are found; the forest vegetation consists of Pirns nigra (upto 1800 meter), the Cedrus libcmi ( Lebanon cedar - Sedir), Abies cilicia (Cilician fir - K öknar) and Juniperus excelsa (Grecian Juniper - Boylu Ardıç). Extensive Cedrus and Abies cilicia forests, together with beautiful Juniperus excelsa and Cedrus libani forests, exist southeast of

Avian Gölü, to the south of Elmalı (Plates 5, 6). The shrub layers in these forests includes Quercus calliprinos (Oak - M eşe), Juniperus oxycedrus (Prickly juniperus -

Katran A rdıcı) , Lanicera (Hanımeli), Berberis (Karamuk), Ptacia ierebinthus/Palestina (Terebinth - Menengiç). As a degeneration stage of Cedrus -

Juniperus forests, a Juniperus excelsa - Quercus calliprinos shrub vegetation is found

in the same area. The Cedrus - Juniperus forests near Avian Gölü occurs at as low as

1050 meter. Dense forests of Cedrus Libani have been an important resource for the

subsistence of the inhabitants of the plain in the past, as today, and were used in

antiquity especially in ship construction, house architecture, furniture, and for heating

purposes. Reference to Cedrus libani occurs in ancient sources (eg.Pliny,

XII. 132, XIU 52, XTV 137, Theophr. Hist. PI. III. 12.3). Travellers’ accounts o f the

mid 19th century also mention the extensive use of cedarwood in Elmalı for heating

purposes, when a scent of cedar surrounded the whole town, especially in the

evenings.'*^ In Reisen I, Bendorf and Niemann give reference to a large caravan of 100

donkeys and camels loaded with cedar wood, east of the valley, heading

towards the Kasaba Plain (which is located just to the south of the Elmalı Plain).

The wood apparently would have been shipped to Alexandria in Egypt. The same

account reports the monopolising of wood-handling in the hands of the Lewissi of

Makri, the modem town of Fethiye.

Fellows, 1852: 382. Benndoif and Niemann, 1884: 127. 21

At present there is a large quarter of carpentry shops in the industrial center of

Elmalı, engaged in processing cedar wood which continues to be a source of important income for the inhabitants of the town, although its use is now limited to furniture making and luxury items. Its growth and cutting is strictly controlled by the

Forestry Department. Also, an Institution for the Research and Improvement of the

Growth of the Cedar in the region, under the Forestry Department, exists in Elmalı.

There are various publications on the Cedrus Li bani of the Taurus mountains, one of the most recent of which gives detailed accounts of the improvement programmes undertaken for Cedrus Libani in the region o f Avian Gölü.''^

Ancient Vegetation:

The Quaternary vegetation of the immediate region was studied by Zeist in 1973, based on a number of sediment corings of lake sites, including the lakes of Karamik,

Beyşehir, Hoyran, Köyceğiz and S ö ğ ü t.An additional study done by Bottema and

Woldring has included core saplings from the Karagöl and Avian Gölü, in 1977.^’ As

a result of the core sample from Karagöl/Elmalı three zones have been defined.“**

Zone one has been dated to a period between 4000-3000 BP and points to the

domination of cedar and pine forests in the region, with juniperus excelsa covering

slopes of the mountain at 1000 m. and above. Towards the end of the period (ca.

3000 BP) wallnut trees occure in small percentage. Both Karagöl and Avian Gölü

points to a hiatus of pollen between zone one and two. Farming is one of the most

critical types of human interference with natural plant cover. Evidence for man’s

interference with vegetation starts with zones two and three (no dates have been

given, but they are in the range of the second and first millenia b.c ), with forests being

Boydak, 1996. ^ Zeist, et. al., 1979: 55-142.The Söğüt Lake is located 60-80 km. as the crow flies to the north of the Elmalı Plain, and displays the nearest locality with a similar geographical and climatic set-up to that of the Elmalı Plain. Bottema and Woldring, 1986: 140-43. ““ Ibid, 141. The pollen samples could not been dated by C14, because of the non-existence of organic material. These zones have been dated by means of corrolations with pollen diagrams horn Söğüt, P ınam aşı and Beyşehir, and the sediment sample has been defined as circa 4000 years old 22

cleared over large areas. Palynological evidence for grain growing and cultivation of fruit trees with juglani (walnut), castanea, o/ea (olive tree), vitis (vine), and citrus exists in zones two and three (see Chart 1).'*’

Two zones have been defined for the Avian Lake, which range over a period of

8000 BP. During zone one dense cedar and pine forests seem to cover the slopes of the mountains to the southern of the plain, with pine overtaking towards the end of the period. Occurance of Cereaha-type of pollen (Plantago lanceolata, Rumex acetosa) indicated to agricultural activity in the region possibly along the northern part of the lake.*® With zone two, human involvement becomes clear. Pollen evidence of Plantago, Juglans, Vitis and Olea exists, but low in percentage. There seems to be a striking difference in the percentage of walnut trees between Elmalı and Avian

Lake, where in the the Karagöl/Ehnalı the percentage of walnut trees reach to 20%, pointing to extensive orchards in the area. Overgrazing of land and herding activities

around Avian Lake is appearant in zone two with the increasing Poterium and

Plantago lanceolata pollen (see Chart 1).**

The pollen diagrams from the Söğüt Lake (see Chart 1) suggest that during

20 000 - 10 000 BP, a drier climate with a fiall glacial and late glacial steppe

vegetation, consisting of Artemisia (yavşan, pelin otu, otçul bitkiler). With moister

phases, more or less open forests of pine-cedar and deciduous oak prevailed. Late in

8 500 BP, increasing moisture was the characteristic of Forest Vegetation, which

started to expand in the mountains of Southwestern . During 6 000 - 5 500

BP, forests completely took over the Oro-Mediterranean Vegetation Belt (at ca. 800-

2000 meters). Human interference with the environment is also evident around Söğüt

Lake, from as early as the second and first millenium b.c.“

'•’ Ibid, 141. “ Ibid, 142. Ibid, 142-43. 52 Zeistet. al., 1979:141. CHART 1 Tentative correlation of the vegetation history as taken from the pollen diagrams for the region

B.P. SÖĞÜT LAKE' ELM ALI^ AVIAN LAKE’

1000- pine forests forests (cedar and pine) forests forest regeneration 2000- forest regeneration forest clearing forest clearing NO POLLEN NO POLLEN 3000 predominantly pine forests forests predominantly pine forests 4000- (pine, cedar, juniper) forests (pine, cedar)

5000- forests (oak, juniper, pine) 6000- transition firom steppe and 7000- forest-steppe to forest 8000-

9000-

10 000- steppe with scattered tree stands

11 000 -

12 000 -

13 000- mosaic of steppe and forest stands (mainly oak and pine) 14 000-

15 000-

predominantly steppe with tree stands

20 000- ■

less forest than the earlier period

mosaic of steppe and oak-forest stands

25 000-·

steppe with forest stands (conifers and deciduous trees)

' Zeistet al., 1979:138. ^ Bottema and Woldring, 1986:145, fig. 16. ^Ibid- 24

Presently, Elmalı Plain proves to be one of the most fertile upland basins of the region. Its productivity has been increased by means of drainage and irrigation systems since the 1960s. Dry farming was practiced in the region before that, when its surface and subsurface waters were sufficient for this type of agriculture. Cereals and fhiits seem to be the dominating crops of the plain. Travellers’ accounts since the 19th century continuously point to the mild climate and the fertile lands of the plain, calling it the com capital of the whole of Lycia. These accounts give us a lot of information

on how the plain and its environs looked at in the turn of the century. These especially

the accounts of Fellows” , Spratt and Forbes*'*, Hoskyn” , Davis*® and Schönbom*^

prove to be valuable sources.

The distribution of the cultivated land in the present day Elmalı region is as

follows; forests 29%, grazing land 15%, dry farming fields 32%, irrigated fields 5%,

orchards and vineyard 5%.* ** Todays’s main crops and their aimual yields have been

grouped in see Chart 2. Cereals, vineyards, apple orchards and the industrial crops of

sugar beet and aniseed remain the main products of the plain. Another important

product of the region is honey, which was used to be collected from the behives on

built stone pillars, which the traditional beehives are still present today in the Geçmen-

Söğle region, southeast of Elmalı (Plate 14).

** Fellows, 1852; 381. The passages have been cited in Appendix B. 1. *'* Spratt and Forbes, 1847: 280, 286 ( Appendix B.2a-c). ** Hoskyn, 1842: 153-156 (Appendix B.3a-b). *® Davis, 1874: 237 (^ ^ n d ix B.4). *^ Ritter, 1859: 814 (A f^ndix B.5). *® K6v Envanter Etüdü Antalya 1981 59. CHART 2 Crop yield for the Elmalı nain as taken from the statistic accounts of D.LE., Ankara

Field Crops 1993 1994 1995 19% (gross yield in tons) Wheat 46 397 46 420 46 937 46 826 Barley 33 451 35 433 35 665 32 740 Oats 1013 1212 1227 869 Chickpea 4 988 4 783 7 000 6 300 Green pea 1752 1400 1 575 2 000 Onion 1 854 4 067 6 984 9 167 Patatoes 5 485 7 638 6 348 8 098 Tomatoes 29 500 31 500 33 7503 7 500 Water melon 14 152 15 623 20 000 18 000 Fodder (animal feeding) Luceme(yonca) 8 750 1700 1900 2 0 0 0 (Korunga) 2 250 450 600 0 (Fig) 5 895 1 115 1 400 0

Industrial crops 1992 1993 1994 1995 1996

Sugar beet - 63 381 77 345 65 001 74 225

Aniseed - 1479 2 237 1 600 1500 Orchards/fniit trees (num ber of trees) Apple 1400 000 1 137000 1 137 000 1 157 000 1205 000 Vine 2954 ton (315 ha.) 2945(325 ha.) 4 500(325 ha.) 5 425(325 ha.) 5425(325 ha.) Pears 198 000 198 000 1% 000 196 000 161000 Ouince 15 000 16 000 16 000 17 000 19 000 Apricot 5 750 6 500 6 600 6 600 6 9 0 0 Plum 8 140 9 000 9 300 9 300 11000 Oleaster 600 700 700 700 700 Peach 19 000 19 000 19 000 19 000 20 200 Morello cherry 2 0 0 0 2 100 2 100 2 100 2 200 Sweet cherry 3 0 0 0 3 250 3 250 3 250 3 250 Walnut 3 9 0 0 4 200 4 300 3 9 0 0 3 9 0 0 Almond 6 000 6 000 6 300 6 000 6 000 Mulberry 1300 1 000 1300 1 300 1300 Comellian cherry 1000 1 050 1050 1000 1000

Honey (in tons) 289 320 400 400 240 26

The drainage of Avian and Karagöl lakes in had a negative impact on the environment and economy of the people, prohibiting fishing, hunting and the usage of reeds, which had grown along the marshy portions of the lakes and provided the raw material for a variety of artifacts. These resources, which played an important role in the economy of the plain, can be also fairly safely attributed to the ancient inhabitants.

Information gathered from the Meteorology stations in the region points to the

regular flooding of the lakes every 10-20 years. Greater floods occurred between

1959-1969, with the greatest flood period recorded in 1969, showing that Karagöl,

from its regular size o f2440 ha., reached an area o f4500 ha., and the Avian Gölü,

from its regular size of 850 ha., reached 2000 ha.*® This factor has been pointed out as

a possible explanation for the scanty evidence of ancient settlements. The waters of

the lakes may have prevented the full exploitation in antiquity.** On the other hand, as

the lakes would retreat to their normal size, they would leave a naturally fertilized

land mass behind, which would play an important role in increasing the crop yield. A

variety of plants could be grown on the plain which are suitable both for wetter soil

and for drier lands. The project of the D.S.I. in the drying out of the two lakes,

proved to have a negative impact on the whole eco-system of the region, among

which climatic change was one of primary importance. A drier climate prevailed in the

region as a result of the drying out of the lakes. While the precipitation rate of the

1960s, was 540mm. it went down to 500 mm. in the 1990s. The subsurface water

level, which used to be at 5-18 meters, has gone down considerably, to 70-90

meters.*’ Cedar growth has stopped within the last five years, and the forest line of

cedar wood has started to retreat. A considerable increase in the number oïAcleris

Undulane (& kind of harmful insect known as the cedar flea among the locals) has

been recorded. This has been attributed to the extinction of the water birds in the

*® Karagöl Avlan Proiesi T aşkın Tesisleri Planlama Raporu 1970: 54. ** Fowden, 1990: 355. *’ Information from S. Ozaner. 27

region,“ that would hold the insect population in balance by feeding on them. The early accounts of Fellows“ refer to the existence of wild swans, large red ducks and waterfowl on the lakes, which do not exist today. Not only floral but also faunal changes in the region have occured, with the extinction of various wild species from the plain.

The climatic change has also affected the crop yield, especially of the apple orchards. It has been mentioned by the locals that over 220 000 apple trees had to be cut down due to drying out. This is also supported by the D.I.E. statistics for the years between 1993 and 1996. There have been recent attempts to fill up the lakes with water to their prior levels by D.S.I.

As has been stated by Garth Fowden“ , the area of Lycia today consists of various

micro-environments. The eco-system of the coastal regions differs from the upper

plains of Central and Northern Lycia, resulting in varied resources and agricultural

products. The semi-arid climate of the uplands resembles that of the Anatolian

Plateau, enabling extensive growth of wheat and cereals, which seem to have been the

main products of the region in the past as well. Dependence of the coastal regions on

the uplands plains for agricultural products must have been an important factor in the

past, as we know from accounts of the sixth century a.c. and the Ottoman period. In

this respect the life of Saint Nicholas, the Bishop of Sion,®’ is an important source, not

only for the students of hagiography but also for its value for studies of the economy,

popular religiosity and especially the topography of sixth century Lycia.®® Passages

51-53 from the Vita, give us the account of a conflict between the coastal and upland

parts of the region, denoting the importance of the upland products for the coastal

62 Ibid. 63 Fellows, 1852: 381. ®^ Fowden, 1990: 361-362. ®’ The Monastery of Holy Sion has been tentatively identified with the church of Karabel, located to the east of the Kasaba Plain and south of the Elmall Plain in central Lycia, by Harrison, 1963: 129- 35. ®® Sevcenko, 1984: 14. 28

region. As the bubonic plague of the Justinianic era hit Lycia, it started to spread around from its harbour of . The farmers, withdrawing themselves from the city, stopped bringing flour, grain, wine and wood from the uplands, and the rumor spread that the Abbot of Holy Sion had forbidden the farmers to go down into the city. The Philip, on hearing this, informed the chief magistrate of the

Province, and the governor ordered St. Nicholas to be chained and brought down to the city.®^

Ottoman period accounts inform us that the chronically deficient regions of the

Aegean and Ionian Islands and , from the 16th century, were allowed to

supply their need for grain from Southwestern Anatolia. The main centre of the Abdal

Musa Tekke, in the Elmalx Plain, was followed by the opening of a number of Tekkes

along the coastline, which operated in conjunction with their main center up in

Elmalı. These Tekke orders actually operated like large land estates during the 18th

and 19th centuries, which has been suggested by Faroqhi as the result of this

demand,®* with the grain being supplied particularly from the fertile upland basin of

Ehnali.

The early accounts of Beaufort in the 1810s tell us about the ventures of the Greek

vessels to load their boats with cereals along the coast of Lycia, to supply wheat for

the islands, the export of which was prohibited by the Ottoman authorities. The

demand had arisen due to the failure of wheat supplied by Sicily during the

Napoleonic wars. The following passage from Beaufort describes, in vivid detail, the

acquisition of wheat

„ ... The influence of commerce on the coast has been little felt till lately, but the immense demand

for wheat in the British garrison of the Mediterranean during the war, and the failure of supply from

67 Ibid., 83-85. ®* Faroqhi, 1981: 54. ^ Beaufort, 1818: 131. The passage has been quoted moreover to give an idea about the general situation of the region during the early 19th century. 29

that once plenteous granary, Sicily (now hardly adequate to its own consumption) had given such a spur to the enterprising islanders of Psara and Hydra, that in search of it they ransacked the whole surrounding coast of that sea. With $ in their hands every creek was explored, and a few quarters gleaned from each valley soon completed a cargo. The exportation of com is prohibited throughout the Turkish dominions, under penalty of confiscation and slavery; but this extreme severity only served to give fresh activity to the traffic: For the Aghas, being exorbitantly paid for their connivance, have a direct interest in prompting it and no Agha in the empire is proof against self - interest. In populous countries and in poor soils it may be a slow and difficult process to push the culture of com very much beyond its accustomed limits or to divert the necessary capital from other pursuits; but in the rich and thinly inhabited valleys of these coimtries, a single year is sufficient to produce exertions, which the stimulus of a free trade is alone wanting to perpetuate. The great plain of Adalia had begun to feel this effect of the impulse; and even from distant parts of the interior, camels, horses and asses were daily bringing their separate ventures, to load the Greek vessels which lay in the port.“

The Elmalı Plain, with its rich resources of woodland, and being the largest arable land of the region, has supported large populations and well established settlements with elite classes during the past. The main town of Elmalı, whose early Islamic

remains go back as early as the 14th and 15 th centuries, has been always referred to

as a well established market town with a population reaching 8000 during the mid

19th century.™ This is a high number when compared with the population of Antalya

in the 1810s, being also around 8000.^’ It is not a coincidence that one of the largest

and richest Tekke Orders of Anatolia, the Abdal Musa Tekke was located in the

plain.™ Excavations conducted in the region have revealed important prehistoric

settlements and a rich sepulchral tradition in the first millenium B.C., which again

points to the existence of a well established elite in these highlands.

Spratt and Forbes, 1847: 287; the number was given as 25 000 by Fellows, 1852: 380, when he visited the town few years before Spratt and Forbes. Beaufort, 1818: 131. ™ Faroqhi, 1981: 48 -77. 30

1.4 MODERN AND ANCIENT LINES OF COMMUNICATION

The upland plain of Elmalı, although well sheltered by a series of mountain ranges, has a number of natural passages that allow for outlets from the plain itself

The main modem lines of communication, which possibly existed in antiquity, have developed along these natural outlets (see maps on Plates 7, 10, also Map 1).

At present the best accessible passage into the plain is from its northeast extension.

The narrow polje depression of the Gilevgi Plain as a north-northeast extension of the

Elmalı Plain provides one of the easiest outlets, connecting it to the modem town of

Korkuteli and to the regions of Pisidia and the southern edge of the Anatolian Plateau.

The recently improved conditions of the Korkuteli - Antalya route, have provided a better communication line between Elmalı and Antalya.The motor traffic coming from Fethiye still makes a detour around the Elbis and Elmalı D ağı, and reaches the

Elmalı Plain over Korkuteli. This northeast passage into the plain possibly was the

main link in antiquity and for the population groups moving from the Anatolian

Plateau. No evidence regarding remains of ancient routes or milestones has been

recovered from the plain,^^ although the mins of a fort site at Gilevgi must be

mentioned; it stands right at the north eastern tip of the plain, pointing to its

strate^cally important location.

A number of passages connecting to the Söğüt Plain (or the ancient Kabalia)

existed to the north of Elmalı (Plate 8). One of these routes, as suggested by Naour,

may have left the plain from its northeast tip, over Mt. Tat, and reached the ancient

Anatolian archaeology is full of suprises. An inscription discovered from in 1993, also refered as Itinerarium Provinciae Lyciae gives the names of the Lycian cities, which are connected with roads, and their distances in stadioaThe inscription has been dated to AD 45/46, to the reign of emperor Claudius, and the Roman governor Quintus Veranius. Işık , 1994:254, fig. 2; 1995: 161-64. 31

site of Tyriaion (modem Kozağacı).^'' A second alternative may have left the plain from the skirts of Elmalı Mountain. The remains of an ancient Roman road (Plate

9a), to the north of the town, along the skirts of the Elmalı Mountain, exist today.

The short stretch of the road which is in good condition is one of the tourist attractions of Elmalı. Still another possible coimection between the Söğüt and

Elmalı Plains, coimecting Eskihisar to Kozağacı over Macun, has been suggested by

Naour.^^ The existence of a site near Macun^®, known as Macun Asarı, further supports this hypothesis.

The western part of the plain, together with the northern edge, is the most difficult

of access, but is not impossible. The Akdağ series bordering the plain along its

western and southern edge, does possess a number of natural passes which link the

Elmalı Plain with the neighbouring upland plain of Seki. Further routes would then

connect the Seki Plain to the coastal town of Fethiye, or upwards north to the Söğüt

Plain. Routes which have existed since antiquity have evolved along these natural

passages. These routes and passages have been moreover used by the pastoralist

groups who practise seasonal transhumance in the region. The routes are usually

closed during the winter months when the mountains are capped with snow. In the

winter months, the detour over Korkuteli becomes the only means of communication

for travellers approaching the plain from the western or northern regions. The various

passes over the Akdağ mountains, which exhibit a beautiful scenery, were much

explored by the mid 19th and early 20th century travellers. A description of the routes

has been presented by C. Ritter,^^ who gives a detailed summary of the accounts of

the earlier travellers. Fellows, Spratt and Forbes, Hoskyn and Schönbom.

74 Naour, 1980; 8. The identification of some ruins near the village of Ilyağ, (see especially footnote 26, in Naour, 8), has further supported the assumption. The passage, though difficult to pass, especially under snow cover, seems to be the shortest connection to the Söğüt Plain. Naour, 1980; 8,118-19; also see the article by the same author 1978; 184, for the discussion of milestones in the region. The milestones of Kozağacı have been referred to by French in RRMAM ; 109-20, under milestone nos. 288-291 (see map on Plate 7). Bean, 1978; 157. Ritter, 1854; 809 - 840. 32

The passes to the west of the plain will be describes starting from north to south respectively. The first route leaves the plain from the locality of Eskihisar, reaching the highest pass over the Akdağ massifs at the Güğü Pass (1810 meters), and descending to the Seki Plain, where the ancient city of , one of the Kibyratis tetrapolis, is located. The second route leaves the plain around the village of Yuva, winding straight up the Elbis D ağı, passing close to its summit, joining the route coming from Girdev Lake, and descending into the Seki Plain, as well. A southernmost alternative route leaves the plain between Gömbe and Akçay, reaching the Girdev Plain and its lake, and then either follows a southwestern route, reaching the river valley of the Esen Çay or Коса Çay, to coastland Lycian via the ancient city

of Xanthos, or leads north around Girdev Lake, joining the route coming from Yuva

and the Seki plain. The harbour town of Fethiye (Makri) is easily accesible through a

number of routes from the Seki plain. The wheat caravans which were used to carry

the agricultural product of the Elmalı Plain into the harbour of Fethiye, as mentioned

in the accounts of Hoskyn^®, used to take the northernmost route of Eskihisar, and

Güğü Pass into the Seki Ovası, referred to as the Şeydiler Yaylası in the early

accounts. Accounts of pastoralist groups and transhumance are given more around

the Yuva Yaylası and Girdev Plain, where various yaylas exist at the skirts of the

Elbis Mountain ( eg. Örtülü Yaylası, A ydınlı Yaylası, Kıldenli Yaylası ete.)

The recent stabilization of the Eskihisar - Güğü route by the State Highways, has

facilitated motor traffic between the Elmalı- Seki Plains and Fethiye. During the

winter months, however, the passage being covered with snow, a detour over

Korkuteli remains the main link into the plain.

In antiquity the region of Elmalı possibly had direct communication with the cities

of Kibyra and especially Oenoanda, which is located in the Seki Plain. Ancient sites as

78 1842: 24. 33

well as evidence for remains of roads do exist along the Eskihisar-Güğü route and in the vicinity of the Girdev Plain. As one leaves the plain from Eskihisar, ascending towards the Güğü Pass, a rock relief of an inscribed stele with a pediment, and figural and animal depictions, are to be seen. The site has been reported by travellers since the earliest journey of Fellows.^® Before one reaches the Güğü pass, at the village of

Yapraklı, the ruins of an ancient settlement have been identified as the ancient site of

Orpeeni.*®

Along the southernmost route which leads over the Girdev Polje, remains of a fortified settlement on the island of Girdev lake has been reported by travellers. A number of sarchophagi with lion lids exist in the vicinity of the lake. The site has been tentatively identified as the ancient Elbessus by Bean.®’ The name of Elbessus is mentioned in two inscriptions, one located in Oeneanda and the other in Güğü

(Orpeeni), shows affinities with the modem name of Elbis D ağı, along whose skirts the site is located.

Two milestones have been located to the southwest of Girdev Plain, possibly along the ancient route which led down to Xanthos. The names of Sevems and Caracalla are

preserved on both of them, with the additional mention of Geta and lulia Domna on

one.*^ Schönbom’s account in Ritter’s Erdkunde also refers to the traces of an antique

road in the vicinity.®®

FeUows ,1852: 384-85; Bean, 1971: 25. ®° Bean, 1971: 26 - 28. According to an Agonistic Inscription from Oenoanda, Огребгй was among the villages of Oenoanda, during the Hadrianic period ( see Wdrrle, 1988:47), pointing to an administrative link of the Kibyriatic city of Oenoanda with a village at the very northwest edge of the Elmalı Plaiii, during the Roman period. ®’ Bean, 1978; 175. ®® First published in TAM 2.3 ; 718-719:302 = milestone nos. 846,847 in RRMAM: 302 (Plate 7). ®® Ritter, 1854: 829. 34

Various routes leading north should at one point have met with the main Roman arterial route of - Laodicea - crossing the AnatoUan Highlands in a west - east direction.*'*

The other two important passages of the plain which provide its connection with the coastal regions are along its southern edge through the Akdağ and Beydağ massifs, known as the Sinekçibel Pass and the Avian Pass. The less accessible route

among these passages leaves the plain from the southwest edge via the towns of

Gömbe (ancient ) and Akçay. The river valley of Akçay provides a good

access to the Sinekçibel Pass which is located between Susuz D ağı and the Akdağ

massif The road leads into another highland plain of central Lycia, the Kasaba Plain.

The classical cities of Candyba (Gendere) and Ameai ( ) are located in this plain,

as well as the ruins of the 10th century a.c. church of Dereağzı. The route, passing

the sharp flanks of the Beydağları, is a less favorable passage to the route that passes

over Avian. After the plain o f Kasaba, the route diverges, one branch leading straight

down to Kaş (the ancient Antiphellos). A western route winds down to and a

further western extension reaches the valley of the Esen Çay and the Lycian

Metropolis of Xanthos. Recent improvements of the road have facilitated a regular

motor trafiQc and regular coach services between Elmalı and Kaş, especially during

the summer months. Transhumance and transportation of people and products on the

market days of the upland towns have kept the route busy all the year round.

According to the Village Inventory Summaries** the main exchange of market

produces of the villages of the Elmalı Plain is in the districts of Finike and K aş. **

*^ According to the Itineraiia Romana published by Miller, 1916; 706, which defined the main Roman roads based on the Tabula Peutingeriana, two routes were related with Lycia ( see Plate 11). One of these is the land route, which left Lycia out of its course, and connected ttie coastal cities of with Laodicea, including destinations of Cormassa () and Termissonio (Tefenni). The nearest access from the Elm alı Plain over Korkuteli to this route would have been from , which has been associated with the site of Örgütlü, ca. 25 km. north-northeast of Korioiteli. That would connea it to the Roman arterial route from , Ephesus, Laodicea, which was the nearest major cross - country road to Lycia, though it left the region out of its route. ** K6v Envanter Etüdü Antalya. 1968: Table 38, 88. ** Fethiye would be another important exchange town for Elmalı, though it is not in the same list because it does not belonging to the province of Antalya. 35

Remains of ancient roads (possibly Roman) have been encoxmtered during the aerial survey of the Kasaba Plain, in conjunction with the Dereagzi Project, and remains of a Roman road was encountered right below the Sinekçibel Pass by D. French.*^ It is not a coincidence that the second milestone located from the plain comes from the vicinity of Gömbe, which has been dated to a.d. 253.“

The other alternative route to the coastline leaves the plain from the southern extension of the valley of the Akçay. Today the main motor traffic leading to the

coastline travels along this route. As has been stated by Fowden“ , the ecological distinctions of Elmalı being a mountain plain and Finike a coastal plain have

complimented each other. Moreover the exchange of products, transhumance and

movement of seasonal migrant workers has made the route an important line of

communication today. Clearly this was also the case in antiquity. The location of the

two important classical sites of and along the route implies the

existence of an ancient road which would connect these cities to the main Roman

arterial road, over the plain of Elmalı. Evidence for remains of an ancient road does

exist along the route. A milestone was recovered to the north of Turunçova, dated to

the reign of Severus and Caracalla.®® The remains of a hillfort with a ring wall of

dressed masonry does exist along the shores of the Avian Lake,*’ pointing to the

strategical importance of the passage. One of the two milestones recovered from the

Elmalı Plain comes from the Tekke village, which is located circa five km. northwest ***

^ French, 1993: 87. The road has been pointed out as an evidence for the economical set-up of Roman, late Roman and Medieval Lycia, which included the exploitation of its hinterlands, especially Central and Northern Lycia, where the agricultural produces would have been brought down from the mountain plains to the coastal regions and coUected in centers near the main harbours, eg. and Patara. The economic system necessitated the formation of a wd)b of regional roads, which possibly also included the Elmall Plain, since the {dain constitutes the largest arable land for Lycia. “ Iplikcio^lu and Celgin, 1991: 184. ** Fowden, 1990: 356. ^ The milestone was first mentioned in R eisen ll: 74,75, and later added to the Kiepert Karte von Kleinasien Sheet D n and inIGRR 3.730. Also see RRAMS Vol.l: 75, milestone no. 196 (see map for the location of the milestone on Plates 7). ” Bean, 1978:154. 36

of the Avlan Passage. The milestone, first published by Harrison,” has been dated to the period of the emperors Constantius and Constans. Any threat to the highlands fi’om the coastal regions would possibly have arrived over this route.

According to the Tabula Peutingeriana, the second Roman road related with Lycia was the coastal road which connected Symma (modem Izmir) with Antalya, and followed the coast of Lycia, including the destinations of Thelmion (, modem Fethiye), Pataras, AntifiUon (Antiphellos, modem K aş), Apyrae (),

Myra, Limera (Limyra), Coridallos, , Oliva to Atalia.”

The eastern part of the plain, as mentioned before, is bordered by the Beydaglari

massifs, which prohibits any fixrther connections to neighbouring regions. Routes

along this part of the plain rather connect the plain settlements with higher seasonal

mountain settlements, which are used moreover as pastures for nomadic groups.

Within the Elmalı Plain itself routes tend to follow along the skirts of the

mountains that surround the plain. The route coming from Eskihisar, follows the

skirts of the Elmalı Mountain, reaching the main town of the plain. Another route

leaving Elmalı, again following the skirts of Noğutlu D ağı, reaches the site of

Avian Lake and continues to the Avian Pass. Another branch from the Avian lake

follows the southern edge of the plain along the skirts of Susuz D ağı, over Armutlu,

and reaches the Akçay river valley.

The early travellers have followed mainly these routes. They frequently mention

the difficulties in passing through the centre of the plain due to the marshes that

existed with the flooding of the lakes. Today a main route leads from Elmalı, through

the plain, to the Akçay district, passing by the Hacimusalar mound. The mound is

” Harrison, 1964:10. ” Miller. 1916: 706. Also see Hate 11. 37

actually situated at the junction of the roads that lead to the two important passages of the Avian and Sinekçibel passes that connect the plain to the coastline regions.

Remains of an ancient route have been reported from the surveys carried out by

Sancar Ozaner in 1993 near Balıklar D a ğ ı ( Plate 9b).The existence of a milestone in the garden of a school in Tekke village, along the southern edge of the plain, suggests the existence of an ancient route between Chôma and Arycanda..®^

Northern, central and coastal Lycia lie in a network of inter-regional, short distance roads. The upland basin of the Elmalı Plain was also a part of this network in the past, as it is today. It was connected by various roads which followed the natural landscape to its neighbouring upland plains as well as the coastal regions. As stated by Fowden®^ the intra-regional movement in Lycia usually followed short distance, narrow and peripheral roads, and occurred mainly in the form of ; l)Transhumance, 2)Exchange of produces.

The different micro-environments of coastal and upland Lycia have resulted in a variety of resources and products, resulting in turn in trade transactions where both regions relied on each other, whereas the coastal cities had also the advantage of enjoying a constant relationship with the wider world and especially with the Aegean

(notably Rhodes and Crete), the south coast of Asia Minor, Cyprus, Syria and Egypt, through their harbours and sea trade.^ This would have resulted in the infiltration of various products from this greater realm, into the ‘remote upland basins’. Accounts of goods sold on the Elmalı market, which arrived from the islands, particularly

Crete, have been given by the 19th century travellers.®^

94 Hairison, 1964: 10. Fowden, 1990: 364-65. Ibid, 362. 97 Davis, 1874: 251. 38

The only route that would connect Lycia to the cultural and economical heartland of Roman Asia Minor would be the Limyra-Laodicea route, which would pass over the Elmalı Plain, and where it would exercise a cultural influence.®® At this point we should emphasize the strategically important location of Hacimusalar Höyük in the

Elmalı Plain, at the crossing of the two major roads which reach the plain from the coastline.

As a conclusion, it may be stated that the Elmalı Plain, despite the fact that it has been referred to as a remote upland region, is a transitional region between coastline

Lycia, Western Anatolia, and the southern part of the Anatolian Plateau, more specifically the regions of Kibyra, Kabalia and Pisidia, with cultural influences arriving from both ends. A survey of the archaeological past of the region will possibly defend these points.

The next chapter will present a survey of the historical past of the Elmalı Plain, relying on the archaeological and historical evidence at hand. The chapter begins with a review of the travellers’ accounts since the mid 19th century, and continues by discussing the archaeological evidence existing for the Elmalı Plain, as early as the

Neolithic Period. The highlighted issues are;

• continuity and discontinuity of occupation on the plain,

• cultural contact zones for the Elmalı Plain (coastal and inland)

• the plain to be viewed as a transitional zone between coastal Lycia and upland

regions o f Kibyra, Kabalia and Psidia, rather than a remote upland plain.

®® Harrison, 1%3: 118. 39

CHAPTER H: HISTORICAL REVIEW OF THE PLAIN IN THE LIGHT OF

EXISTING EVIDENCE

2.1 A REVIEW OF THE TRAVELLERS’ ACCOUNTS AND SCHOLARLY

RESEARCH CARRIED OUT IN THE REGION

2.1.1 Accounts of travellers

The European travellers and scholars starting with the mid 19th century, prove to be a very valuable source for the history of the plain, as they describe an environment,

including settlements, traditions and events which may well reflect that of much earlier

times, and especially the period before the Effects of industrialization. These sources

portray a physical and cultural landscape which is irreversibly disappearing every day,

so a closer look at these accounts would provide a glimpse into the recent past.®®

The earliest accounts of a traveller to the Elmalı region is Evliya Çelebi,’®® the

well known Ottoman traveller of the 17th century. His visit to Elmalı and its vicinity

has been dated to 1671-72. He gives a detailed account of Elmalı and the largest

Bektaşi order of the region, the tekke of Abdal Musa. Just to mention briefly from

the accounts. Elmalı was a well established market town with its administration

system and guilds, as early as the 16th and 17th centuries, which is also suggested by

the earliest Islamic buildings that exist in the town.

The coastline of Lycia was investigated by European travellers as early as the 18th

century.'®’ For the first half of the 19th century, the interior part of Lycia remained

®® French, 1994: 54. ’®®Zillio9lu, 1985: 115-117. ’®’ Borchhardt et al., 1990: 228. Visits to Lycia began with Graf Choiseul Goufier the French consul in 1776, and later by Sir R. Ainslie in 1800 and Luigi Mayer, between 1800- 1810 by Joseph Freiherr von Hammer- Purgstall, the first president of the Aldmdemie der Wissenschaften in Wien. 40

unexplored due to lack of maps, information and risky traveling conditions. It was Sir

Charles Fellows’“ who made the pioneering journey into inland Lycia, during his second visit to the region in the year of 1840.’“

Fellows’ journey and his arrival in the Elmalı Plain displays the anxiety of a pioneer who arrives in unknown lands, but is surprised at finding well established settlements there. His accounts refer to Elmali as the largest town of Lycia, almost the largest in Asia Minor, with a population exceeding 25 000 (Appendix В la). His exaggerated accounts regarding the town of Elmalı and its population, is possibly the result of the unexpected encounters. Although Fellows’ visit to the plain lasted only two days, he gives detailed accounts of Elmalı, its houses, streets, bazaars, and monuments, as well as the economy and the products of the region. Information regarding ancient remains and sites from the plain remains very scanty in his accounts.

He m ainly notes some ruins near Eskihisar, which he identifies with ancient Podalia.

He also gives a description and sketch drawing of the rock relief at K apıtaş

(Appendix Bib).

At the end of 1840 and early 1841, Hoskyn, the master of the Admiralty ship of

Beacon also paid visits to Lycia and Karia, and later with Forbes.’“ Their visit of the

plain lasted for three days, in October. Passing through the plain, Hoskyn and Forbes

described it as barren, especially at that time of the year, when the rests of the com

We also have the accounts of Captain Beaufort, the Hydrographer of the British Admiralty, who had explored the coastline between 1811-1812 and published his accounts in his book Karamania, 1818. C.R Cockerell, the architect, started his journey alone but later accepted to be accompanied by Captain Beaufort, due to the insecurity of the region. The account of Charles Texier in 1836 examines mainly the coastline. ’“ Slatter, 1994; 3-8. Sir , a geographer, naturalist and trained in classics, was the son of a wealthy silk merchant His antiquarian interest played an inqx>rtant role in the persuasion of the trustees of die British Museum to transport the Tomb and the to London. Most of the British travellers, who had visited Lycia and could also make a joumty into the interior parts of the Lycian coast were engaged with the project ’“ FeUows, 1839; 1841. ’“ Hoskyn, 1842:153-54. Hoskyn first explored the valley of Xanthos, Cadyanda and Caunus. On his later visit in autumn 1841, accompanied by Forbes, he journeyed into the interior of Lycia. They arrived in the Elmalı Plain via Fethiye-Üzümlü-Seki Plain over the Eskihisar route. En route th^ encountered wheat caravans which serve between the town of Elm alı and Makri (modem Fethiye) Appendix B3a 41

fields had been burned down (this practice, despite the prohibitions set by the government, still exists today). Hoskyn’s visit included the site of Armutlu (Appendbc

B3c), the tombs and ruins of Kızılca. He also gives us a detailed account of the town of Elmalı (Appendix B3b). As they left the plain, ascending into the mountainous terrain of the ‘yayla’ country, they encountered a group of sarchophagi

of conglomerate rock and an abundant number of hewn stones and perhaps the ruins

of a fortress. Possibly they were describing the ruins of Gilevgi, the lake site in the

Elbis mountains. In the Yuva Yaylası Hoskyn foimd two inscriptions.’®^

In 1842, Lieutenant T.A.B. Spratt, naturalist Prof Forbes and Rev. Daniell paid a

visit into the interior of Lycia.’®® The accounts of their excursions in Lycia were

published in two volumes.’®^ During their four days of stay in the plain, they mainly

visited the sites around Eskihisar, where they encountered a hermit’s cave site

(Appendix B2e), the Kapıtaş and Güğü rock reliefs (Appendix B2d), and the ruins

of Armutlu. They identified the ruins of Armutlu as the ancient city of Podalia, and

suggested the identification of the ruins of Eskihisar, actually identified as Podalia by

Fellows, as .’®* Before reaching the village of Armutlu, they encountered a

mound site (Appendix B2f). Though no names from the immediate vicinity have

given, this is Hacimusalar Höyük, which stands out in the flat plain as a prominent

landmark, due to its size. The very same mound site is also referred to in the accounts

of later travellers.

’®* Ibid, 155. ’®® Spiatt was in charge of the geography and constmcting a detailed map. Forbes recorded the natu^ history, and Daniell, who unfortunately died during the course of the trip, was in charge of the antiquities. ’®’ Spratt and Forbes, 1847: 280-91. The first volume deals mainly with the route, modem and ancient sites, modem settlements and their inhabitants.The second volume, which was mainly prepared by Forbes, dealt with the geography of the region, giving accounts of the fauna and flora of both coastal and interior Lycia. ‘®* Ibid, 290; FeUows, 1852: 284. 42

In the same year of the visits of Spratt and Forbes, Julius Schönbom, a

German gymnasium teacher, and a student of the eminent Prof Carl Ritter made an extensive excursion into the mountains of Lycia as well as the plain of Elmalx.'“'

In May 1872 Rev. E.J. Davis, together with J. Seiff, visited the plain of

Elmalx."°They arrived in the Elmalı Plain from Antalya via Istanoz (modem town of

Korkuteli) and over the Gilevgi pass. Davis’ account of the town of Elmalı gives detailed information about the inhabitants, their costumes, culinary traditions, events in the town and neighbouring regions and the lively market day on which they happened to be there. Their decision to reach Makri via the Eskihisar and Güğü road was altered due to an event which occurred during their stay in Elmalı. The event

involved a group of mountain brigands who had attacked the town of Fethiye and held

it in their possession for a while (Appendbc B4b). Hostile relationships between the

mountains and coastline regions seem to have existed tUl recently. A case in point is

the mountain brigands of Kibyratis who were well known in terrorizing the coastline

in the Roman period.

The next series of accounts come from the expeditions of the Austrian Team, led

under the auspices of the Wissenschaftliche Akkademie Wien. The expeditions aimed

to make a detailed account of the sites, mins, monuments and inscriptions, later to be

formed into the corpus of the Tituli Asiae Nfinoris series with a separate volume

dedicated to monuments from Lycia. The first expedition o f 1881 was led under

Otto Benndorf, professor of Classical Archaeology in the University of Wien, and

His accounts have been reviewed by Ritter, 1841 ; 809-840. Ritter’s review also includes a detailed account of Fellows, Hoskyn and Spratt and Forbes. Another source which refers to the worics of Schönbom is by his nephew Charles Schönbom, 1868. Davis, 1874: 237-58. See discussion in footnote 413. Emestus Kalinka in the 1920s compiled the second volume as Tituli Lyciae, including the interpretations and inscriptions collected by the early travellers and scholars, especially from the woria of the Austrian expedition. Tituli Lyciae, though does not include northern Lycia, but mainly coastal and central Lycia. With the later works of Bean, 1971 and Neumann, 1979 under the additional issues to the Tituli Asiae Minoris, monuments and inscriptions from northern Lycia and the vicinity of Elmalı were included. 43

George Niemann, the architect. The expedition included the regions of coastal and central Lycia and Karia. During this trip Benndorf also paid a visit to the Elmalı Plain and its vicinity. The relief friezes of the Heroon of were transported to Vienna under the supervision of Benndorf at the end of these expeditions.

The expedition of 1882 was led by Eugen Petersen and Felix von Luschan, which mainly included the lands of the ancient and the Kibyratis."'’ The arrival of

Benndorf and Petersen into the E lm alı Plain was via Trysa (the modem town of

Gölbaş i), over the Akçay river valley and Gömbe. They identified the mins near the

modem town of Gömbe as the ancient city of Comba.”* Their account also refers to

the discovery of one of the relief plaques depicting the 12 Gods, a religious depiction

of the worship of local gods, dated to the Roman period."® Their expedition

continued into the mountains of Akdağ and the vicinity of Gilevgi Lake and Yuva

Yaylası, where they recorded a number of inscriptions, sarchophagi and architectural

mins. Returning back to the plain, they mention the mound site that had been noted by

Fellows, possibly the Hacimusalar Höyük (y^pendbt B6). Herewith we see the first

possible identification of the mound with the ancient city of Chôma being made by

Benndorf Further investigation of the plain included the hillforts of Armutlu and

Avian Lake, and they later preferred to identify the ancient Chôma with the mins of

Armutlu. They also identified the mins near Avian Lake, with the ancient Podalia, as

they had been told by a non-resident of Elmalı that the place was known as Podamia

or Buralya in earlier times.'" Their visits also included the Söğle region and Serkiz

Yaylası into the mountainous regions to the east of Avian Lake, where they

encountered the local tradition of built beehives on stone pillars (Plate 14 ), whose

architecture they linked with the pillar tomb tradition of Xanthos."® The northern

113 Benndorf and Niemann, 1884. 114 Petersen and von Luschan, 1889. Ibid, 156-57. "® Ibid, 157, including drawing of the relief plaque with its dedicatory inscription Ibid, 161. "* Ibid, 162. 44

most tip of the plain with its fort site of Gilevgi was also investigated, and a sketch plan of the fort site is included in their accounts."®

A further chapter in Reisen H, by Felix von Luschan, includes an anthropological study of the Yörük groups in the highlands of Lycia, including the Alevi groups known as the Tahtacılar in Lycia, the name usually given after their main occupation as woodcutters in the forest regions of Lycia. The local traditions and beliefs, as well as the architecture and costumes of these people, are reported in the chapter. Von

Luschan’s accounts are also interesting in exhibiting the 19th century traditions of

anthropological studies, where a table of skull dimensions of the locals are included as

evidence of ethnic differences encountered in the region.'^®

Charles Chipiez, in the early 1890s visited to the Elmalı Plain, arriving over the

Arikandus valley from the coastline. He mainly described the climate and vegetation

of the region, and he compared the wooded zone of Lycia to the Alpine scenery of

Switzerland and the Tyrol.

2.1.2 Excavations and field researches, in the 20th century

Ormerod and Woodward visited the Elmalı Plain in the summer of 1910, en route

from Antalya to Denizli. They followed the Korkuteli, Söğüt Gölü and Tefenni route,

making a detour into the plain from Korkuteli. Their visit included the survey of

mound sites, collection of pottery, and noting inscriptions and architectural remains.’“

Included in their record are two mound sites near the fort remains of Gilevgi, at

the very north-west tip of the plain.One of the mounds is situated right at the mouth

119 Ibid, 165. Ibid, 199-226. Perrot and Chipiez, 1892: 334-35. 122 Woodward and Ormerod, 1909-1910: 76-135. 45

of the Gilevgi pass, and the other one, forty minutes to the south of Gilevgj.’^^ The third mound located at the southern extremity of the plain, is the mound which was mentioned in the accounts of Spratt and Fellows.Woodward and Ormerod give a

more accurate location of the mound, as being at the foot of the low hill known as

Çataltepe, denoting the exact location of Hacimusalar Höyük (Appendix B7).

The next exploration of the region was by Sinclair Hood in 1949, which has been

included in the works of James Mellaart. Eluring the autumn o f 1951 and from April

to November in 1952 a survey of pre-classical remains of Southern Turkey, including

the Elmalı Plain was undertaken by Mellaart and the British Institute.'^“ From the

plain of Elmalı, a number of mound sites dated to the Chalcolithic, Early Bronze Age

and Iron Ages were recorded, with their pottery evidence.

The next survey in the region, in 1954, was led by M.J. Mellink, of the Department

of Classical and Near Eastern Archaeology of Bryn Mawr College. The Elmalı Plain

was surveyed in conjunction with the coastline and the main river valleys (mainly the

Xanthos river valley) of Lycia.’^* Nearly a decade later the first excavations on the

plain were undertaken by M.J. Mellink, between the years 1963 and 1974. The sites

excavated were the mounds of Karataş-Semayük and B ağbaşı, located at the

northern part of the Elmalı Plain. In conjunction with the excavations of Bağbaşı, a

survey was conducted by Christine Eslick in 1977, to recover the Chalcolithic and

Neolithic sites o f the plain.

Starting with 1969 a number of Iron Age and early Classical tumuli burials were

excavated as a rescue operation, together with the Antalya Museum. The excavated

tombs included the painted tombs chambers of Kizilbel and Karaburun n.

Ibid, 92. Mellaart, 1954; 174-239; 1955; 115-129. 125 Mellink, 1954; 326-27. 46

A further excavation of two tumuli burials from the B ayındır necropolis, located near Karaburun and Karataş-Semayük, and a small chapel was conducted by the

Antalya Museum in 1986 and 1987, the tombs yielding Phrygian and Lydian artifacts.

A survey to collect material for the chronological study of Early Christian architecture in Lycia had begun in 1959 by Martin Harrison*“ , supported by the

British Institute of Archaeology in Ankara. In 1963 a part of Harrison’s work included the epigraphical survey of the Elmalı Plain, accompanied by George Bean.

Resulting from their work, the mound of Hacimusalar was identified as the

ancient city of Choma.*^* Martin Harrison’s expeditions were interrupted in 1964 to be resumed in 1977 till 1980. Within these years work concentrated on three sites of

central Lycia and the Elmalı plain: O vacık, Arif and Alakilise. The 1978 season’s

work included another survey in the Elmalı Plain,and the identification of a hilltop

site at Ovacık, east of the Elmalı Plain, the results published in various articles.'“

Bean’s work concentrated on the epigraphical evidence and classical monuments

of the region. His findings were published in the series of Tituli Asiae Minoris.'“ Bean

identified the site of Söğle, located at the eastern skirts of the northern section of

the plain, as the ancient city of Podaha, rather than its otherwise suggested location

near the site of Avian Lake.He also identified the ruins of Güğü, on the Eskihisar

route, as the ancient city of Orpeeni, and the remains of Girdev Lake tentatively as the

ancient city Elbessus mentioned in inscriptions from Oenoanda and Yuva.

The research on Lycian tombs undertaken by J. Zahle and K. Kjeldsen, between

the years of 1970-76, and 1979-80 partly included the Lycian rock tombs located

'“ Harrison, 1%3: 117-63. Bean and Harrison, 1967: 38-43. '“ Harrison, 1979a: 205-206; 1979b: 235-237; 1979c: 525-531; 1980: 109-118; 1981a: 199-220; 1981b: 384-387; 1983:75-77. '“ Bean, 1971:22-33. Ibid, 28-33; 1%8:157-163. 47

along the southern edge of the Elmalı Plain, near the villages of Armutlu, Buralya,

Islamlar and K ızılca. Included was an investigation of the architecture of the local storehouses of wooden construction which still prevail in the upland region of

Elmalı.’^’ These were thought to represent the earliest domestic architecture of coastal Lycia, represented in the Lycian tomb architecture.

A fixrther epigraphical investigation of the region was conducted by a team from the University, Istanbul, under B. Iplikçioğlu, in the early 1990s. The results of this work have been published in two articles.’’^

These series of accounts, researches, surveys and excavations bring us to the recent survey and excavation project of the Hacimusalar Höyük which has been undertaken under the directorship of İlknur Özgen'” by the Department of

Archaeology and Art History of the Bilkent University, since 1993.

Kjeldsenand Zahle, 1975a: 312-350; 1975b: 312-50; Zahle, 1979: 343-45. I” İplikçioğu and Çelgm, 1991: 183-92; 1992. Özgen, 1997: 603-609. 48

2.2 PREHISTORIC OCCUPATIONS ON THE PLAIN

Documented knowledge of the prehistoric occupation of the Elmalı Plain mainly comes from the two excavated sites of Bağbaşı'^'' and Karataş /Semayük'^’, both of which have been explored in great detail, and have revealed a wealth of information.

There exists a continuous sequence of habitation from the Late Neolithic till the Late

Bronze Age on the Elmalı Plain. But our knowledge is neither consistent for the whole of the plain, nor for all the prehistoric periods. Despite the fact that the nature and development of the periods (eg. Late Chalcolithic, Early Bronze Age I, II, HI) have been well established archaeologically, some of the periods (eg. Late Neolithic,

Early Chalcolithic, Nfiddle Bronze Age) are less well known to us. Evidence for the

Late Bronze Age seems to be completely missing.

Interest in the prehistoric occupations of the plain started with the survey of

Sinclair Hood, in 1949.'^® Mellart’s extensive survey of Southern Turkey has also

-K a ra ta ş 1 : 1. The site o f B a ğ b a şı is located 7 km. northeast of E lm a lı. The mound is located on a natural rise at the skirts of the E lm al i Mountain, at an elevation of 1158 meters, overlooking the western part of the plain. A stream miming from the western hills provides a good water sui^ly, and today the region is covered with vineyards and orchards. B ağb aş i has revealed occupational evidence for Early Bronze Age, Middle Bronze Age, and Iron Age, but the main occupational level was of the Late Chalcolithic period. The settlement proved to be of short duration with an accumulation debris of 0.20-1.0 meters. Three superimposed levels were found, without a change in the pottery tradition, except the lowermost level of in , which did not reveal any architectural remains. An earlier Chalcolithic phase was determined on the basis of the pottery assemblage recovered from the excavations.. ’^^Elm al i -K a ra ta ş II: 5-6. The site of K a r a t a ş lies in Open fields between the modem village of Semayiik (Bozhöyük) and Gökpınar, 8 km to the east-noriheast of the town of E lm al i in the northern sector of the plaiir The mound is 100 meters in diameter and is a shallow rise of three to four meters from the surrounding fields. The soil cover over the bedrock is very thin (0.50-1.00 m.). Long periods of agricultural activity and plowing of the surface has resulted in the fragmentary nature of the archaeological data. The site was chosen to give evidence for the prehistorical occupations of northern Lycia, for a number of reasons. First of all due to its small size, no Roman or Classical occupation has appeared on the mound, other than few fragments of sherds. The site had been pointed out for its Early Bronze Age remains by Mellaart in 1954. The accidental recovery of the jar and pithos burials in the typical Early Bronze Age Western Anatolian burial tradition, which had been reported by the farmers, also added to the above reasons, for the choice of the site. The site has also revealed some Middle Bronze Age burials and Iron Age pottery extending well into the Roman era. MeUaart, 1954:176. 49

included the Elmalı Plain.'^^ As a result of his expedition a number of sites of the

ChalcoUthic, Early Bronze Age and Iron Ages were identified on the plain (Plate 15).

Two sites dated to the Chalcolithic period, Akçay I and Коса (Tekke) Höyük, were found to the south of the plain.'“ The ‘Chalcohthic Painted Ware’ found by

Mellaart, during his extensive survey of southern Turkey, proved to be widely distributed from Afyon down to the Elmalı Plain, as well as at Isakli Hisar (in

Bolvadin) and the Konya Plain (Seyeti Han). The fabric of the pottery is the same as the burnished ware. The surface color varies from a ,/ed, plum red to chocolate brown“. The matt paint has been applied on a buff, or on a cream or yellow slipped surface. Burnishing was done after the paint was applied. Inside, the pottery is usually red slipped or painted, and burnished. Decoration normally occurs on the exterior. 139

As part of the same survey, the Early Bronze Age sites Gilevgi, Mürsal, Beyler

(Hacımusalar), Akçay I and Коса (Tekke) Höyük were identified.’'”' The sherds collected from the sites fall into the category described by Mellaart as a red slipped and burnished hand made buff ware. Their surface color varies from a „salmon pink“

(which is typical for the Elmalı Plain) to a „bright red“.The surface usually has a crackly slip. The paste is grit and straw tempered, with a black core.’'"

The white painted pottery of the Early Bronze Age was also identified on a number of mound sites on the plain, including Gilevgi Höyük, Semayük, Karataş

Höyük, Söğrie and Beyler Höyük. The ware was briefly described as a buff ware with a fine pink burnished slip and matt paint.’'’^

137 Ibid, 174-239. ’“ Mellaart, 1954: 188. Ibid It is unfortunate that we have only the drawings of these sherds, where the description of their fabric remains unclear. ’'“ iWd, 192, map 3. ’'" Ibid, 190. Ibid, 220, ill. 187, 226, ill. 301, 238-239, ill. 434-458. 50

2.2.1 Neolithic and Chalcolithic OccupationsrBağbaşı and Other Mound Sites

from the Elmalı Plain

The pottery evidence remains to be the main class of archaeological data in reconstructing the Neolithic and Chalcolithic sequences of the Elmalı Plain. The pottery assemblage of Bağbaşı, other than being evaluated in its archaeological context, has been also evaluated against the pottery data which was collected from a number of sites during the 1977 survey season (Plate 16). Three main periods. Late

Neolithic/Early Chalcolithic, Middle Chalcolithic and Late Chalcolithic, have been

identified.'“'^ Comparanda material for dating these assemblages is taken from

excavations in Western and Southwestern Anatolia.'“'^ For the Middle Chalcolithic

period, which seems to be missing for Southwestern Anatolia, and therefore creating a

hiatus, comparanda material was taken from the Eastern Aegean sites (Plate

Evidence for architecture has been only encountered in the Late Chalcolithic period

in the Bağbaşı excavation.

(a) Late Neolithic/Eariy Chalcolithic

Pottery:

The earliest group of pottery from the plain has been identified as Late

Neolithic/Early Chalcolithic.’''’ One sherd from the Hacimusalar Mound was

identified as Neolithic.'““ The sites of Akçay and Tekke Höyük give evidence for Late

Rimal ı-Karataş 1 : 7 7 -7 9 . '“*“* Elmalı-Karataş 1 :67-80. Eslick, 1980:12-14. Including the sites of Baycesultan, Beybaba, Hacılar, Prehistoric , Kuruçay, and Karain Cave site for its Neolithic and Chalcolithic levels. Ibid., 67-80, Eslick, 1980: 10-12. Including the sites of En^rio/Chios, Tigani/Samos, Kalynmos, the cave site of Aspripetra/Kos, and Saliagos. 146 Elmalı-Karataş 1 : 76-77. Rimali-Karata$ I : 77. *■“ Ibid.,63, plate 80. Rim fiagment of a hole mouth jar was identified by S. Hood’s survey. 51

Neolithic and Early Chalcolithic, whereas at the site of Gökpınar there is no evidence for the Late Chalcolithic.'''®

The main forms are hole mouth jars, flaring bowls, inverted necks, jars and bowls with everted and verticle rims (Plate 18).'“ The pottery is handmade, using the coil technique. The handles are directly applied to the body. With the Late Chalcolithic assemblage this feature is replaced with lug handles which are plugged into the main body. The tubular lug handles of the Gökpınar examples show the characteristic feature of the Neolithic assemblage."" The color of the fabric varies from brown to orange, usually slipped with the color of the slip varying fi'om a brown to crimson, and usually a scarlet red.'“ The slipped surface is generally burnished. The fabric of the Late Neolithic/Early Chalcolithic pottery is grit tempered.*“ The black core of the sherds, some virtually as thick as the sherd itself, suggests little control of firing.'“

The Early Chalcolithic pottery fi'om the sites of Akçay and Tekke Höyük is considered to be a direct development and continuation of the Late Neolithic pottery.

The painted Chalcolithic wares of the Hacılar I and II pottery, although not adopted so much on the ElmalI Plain, are still visible.'“

(b) Middle Chalcolithic

Pottery;

*''® Ibid, 76-77. '“ Ibid, 77, for GOkpinar see plate 79, sherd nos. 1-8; for Akgay plate 76-77, sherd nos. 1-35. ib id , 67. '“ The streal^ look of the surface suggesting that the slip was applied with a cloth or brush rather than being dipped into the slip. '“ Ibid, 77,94. The size of the inclusions being between 0.15-0.05 cm. in diameter. Where the angular shape of the carbonate grains of the Ak^ay samples have suggested that they had been derived firom freshly crushed rocks and deliberately added into the clay. Chaff has been added but much less in comparrison with the Late Chalcolithic samples, which is suggested by the smaller amount of voids and the more compact nature of the &bric. *“ Ibid, 77. '“ Ibid, 77. 52

The second group of pottery has been identified as belonging to the Nfiddle

Chalcolithic phase of the Elmalı Plain. This group was mainly identified from the

Bağbaşı excavation.’“ The pottery appeared below the architectural levels of the

Late Chalcolithic village, in a burned debris with bones and pottery, without any architectural remains.'*^ Another group of pottery dated to the Middle Chalcohthic period, came from a fill of the Iron Age tumulus of Kizilbel, dated to the sixth century B.C.'“ The sites of Bayındırköy, Gökpınar and Arapkuyusu also give evidence for the Middle Chalcolithic period.’’®

The Middle Chalcolithic group of pottery has been described as showing the

characteristic traits of both the Late Neolithic/Early Chalcolithic and Late Chalcolithic pottery of the plain, with a closer resemblance to the Late Neolithic/Early Chalcolithic

assemblage. Its archaeological context, as being identified below the Late Chalcolithic

levels of Bağbaşı, has supported its dating.”®

Forms are limited to hole mouth jars, vessels with flat rims and flaring sides, and jars with short necks also appear. ’®’ Some stands and fragments resembling the

Late Chalcolithic assemblage also appear from Kizilbel (Plate 19C).’“ The

pottery is handmade, using coil technique, similar to the Late Neolithic/Early

Chalcolithic wares. The handles are still directly applied to the body, but a new feature

is deep scoring on both the handle and the body. This new form of strap handle, which

is to become the dominant element of the Late Chalcolithic ware, appears also in the

Middle Chalcolithic assemblage. It is usually placed either on the rim of a bowl or on

’ “ Ibid, 51,78. Eslick, 1980: 8. Especially from the lowest levels of the tenches 105 and 116. Elmalı -Karalas I: 1, 51. Ibid, 55. The site has been excavated for the Iron Age Tumulus, Chalcolithic sherds have been recovered over the tomb chamber, within its tumulus fill and along the ridge of the tumulus. No excavation for the Chalcolithic settlement has taken place here. ” ® Ibid, 78. Ibid, 51,78. Eshck, 1980: 8. ’®’ Elmalı -Karatas I: 51,52,56 and 78, for Lower Bağbaşı see plates 63-65, for K izilbel plates 74-75. Ibid, plate 75, sherd no. 14. 53

the body of a closed vessel. Some strap handles show a pinched profile along the upper end of the handle, these examples found in the Bağbaşı repertoire (Plate 19

B,D).’“ The surface color varies fi’om black to red and orange. While the

Bağbaşı"^ examples usually show a hght brown color, the Kizilbel*“ ones appear to be of a red/orange color. The surface is well smoothed giving a compact look, usually unslipped and hghtly burnished. Little decoration on the surface is observed.'*® Its fabric has been described as grit tempered, and coarse, though denser than the Late ChalcoHthic wares.'*^

The Middle ChalcoUthic period presents a chronological hiatus for Southwest

Anatolia.’*® The Middle Chalcolithic assemblage of the Elmalı Plain which has been identified as a transitional phase between the Early and Late Chalcolithic periods, has also been compared with the Middle Chalcolithic assemblages of the Aegean coast and islands. Main sites which were compared were Emporio of Chios (levels X-

Viil)i69, Tigani of Samos'™, Kalymnos'^', the cave site of Aspripetra on Kos'™ and

'*® Ibid., 78, for Lower Bağbaş i see plate 64, and for K izilb el see plate 74, sherd nos. 2,11. '*^ Ibid, 51. '** Ibid, 56. '** Ibid, 52,57 and 78. Single knobs on the body start to appear, one incised pot bears evidence of red paintod decoration, a bowl from K ız ı I b e l has diagonal impressions along the rim, and incised lines and dashes occur on some of the sherds, but do not show the white filling yet. Piercing below the rim has been also encountered with some of the vessels, for which parallels are to be found from the Eastern Aegean islands (eg. Aspipetra cave site/Kos, Kalymnos, Tigani/Samos). ’*^ Ibid, 51,56. Grit has been deliberately added to the clay, as by the Late Neolithic/Early Chalcolithic potters, but fiber inclusions seem to be less, possibly not deliberately added 1^ the potter. The pots are poorly fired with black cores still prevailing. '**Accordingto Mellink, 1992: 211-12; Esfick, 1980: 5-7. The latest levels of the Early Chalcolithic village of Hacılar I and n appear to come to an abrupt end which shows a hiatus with the earliest levels of XL-3ÖC, dated to the Late Chalcolithic period The sudden appearance of new cultural traits in Bqrcesultan’s Late Chalcolithic levels have been attributed to the arrival of new invaders into South-western Anatolia by their excavators Mellaart and Lloyd. But the appearance of a Middle Chalcolithic assemblage from the Blm al 1 Plain with affinities to the Aegean coast and islands may fill up this gap. '*® Elmali-Karatas : 69-70; Hood 1981: 240-300; 247, fig. 119 sherd no. 20,24 (for pots with internal lugs) and sherd no. 17-19,22 (for row of perforation below rim), 259 fig. 124 sherd no. 162 (rising handles), 268 fig. 127 sherd no. 179-181 (for hole mouth jars) and sherd no. 184 (for strap handle with a knob) 268 fig. 127 sherd no. 179-181 (for hole mouth jars), no. 184 (for strap handle with a knob), '™ Eslick, 1980: 10-11; Heidenreich, 1935-36: 133 F6-10,140-41F19-20,143 F31-33; Plates 34.6,43.6,48.3. Parallels especially with K izilbel repertoire could be observed with the incurved bowls, hole mouth jars, a bowl wiUi high str^ handle, piercing below the rim, and knobbed strap handles. 54

Saliagos'^^ Although no equivalents in fabric have been identified, technological and typological parallels could be observed between these assemblages.

On the basis of pottery evidence, Eslick points out a parallel development of certain production techniques between the E lm a lı Plain and the Aegean, in the

Chalcolithic period. Assuming that B ağbaşı pottery did not develop fi’om the latest phase of Hacılar Early Chalcolithic pottery, Eslick suggests contact between the

Aegean and the Elmalı Plain. She proposes two alternative routes; one being the overland route over Korkuteli, Burdur and following the Meander valley to the western coast of Anatolia, and the other a sea route over the southern coast of Lycia.

However both of these routes lack comparable evidence between the mentioned zones of contact.’^"

(c)Late Chalcolithic

Architecture:

Among the twenty trenches that were opened in B ağbaşı, substantial architectural remains were encountered in sbc of them.” ^ The remains of about five or sbc houses were revealed. Three main wall-building techniques were encountered.

These techniques have been referred to by the excavators as the mud slab technique

horizontal timber stacked or the Tog cabin’technique'and vertical timbers

Eslick, 1980:11; Fumey, 1956: 189 fig. 10, plate 18. High strap handles and a row of piercing below the rim are the main parallel features. Eslick, 1980: 11; Levi 1925-26: 294 fig.82,299 fig. 87.2-3. High strap handles and piercing below the rim are the main parallel elements. Eslick, 1980: 11; Evans and Renfiew, 1%8: fig. 45.6. Again high strap handles and piercing below the rim are the main parallel features. Eslick, 1980: 12-14. Eslick also mentions that along the Burdur and Meander valleys none of the sites have revealed Middle Chalcolithic pottery, but some sherds seem to occur from Aphrodisias, Dereköy L Hacılar n, and Kuruçay, ictentified as having the traits both of Late and ^ l y Chalcolithic pottery. Also the southwest coast and the off shore islands of Dodecanese have not revealed any evidence of Middle Chalcolithic pottery. F-lmali-Karatas 1:1-5, Eslick, 1988: 31. Namely the trench nuihbers being, 102,104,109,110,115,116, with trench nos. 105 and 106 evidence for superinqx)sed remains of houses were encountered. Mellink, l%9a: 328, plate 77, fig. 34 also see ill. 5 by D. Duryea; Elm alı-Karatas 1 : 15. Trench 105 excavated in 1%8 revealed the most complicated and advanced wall construction dated within 55

packed in mud plaster techmque.*^*Only the walls in trenches 110 and 116 (third level) had stone foundations.’^®

The comphcated technique of mudbrick and timber usage with mud plaster finish, shows us that the Elmalı people were able to make use of their environment efficiently, and had the technical skills as early as the Chalcolithic period.

No evidence for roofs has been identified, except some fi^agments of burned reed impressions in the fallen debris, which have been attributed to collapse fi’om the roof’*® Floors were made of compacted earth and were poorly preserved. A fixed hearth in a central position was encountered in nearly all of the houses. Cooking is thought to have taken place within the house.'*’ The flat bases of the cooking vessels suggest that they had been constructed to be suitably used on flat mud hearths. Most of the houses displayed large storage jars, suitable for storing grains and other food produces. Storage bins and pits had not been encountered in the houses.'*^

the Late Chalcolithic strata. The wall and its construction has been described in 1%9 excavation report as follows; „ ... Trench 105 had part of a burnt house consisting of a long wall fragment (pres.L nearly 4.00 meters) with a cross-wall (pres. L 2 meters) running in right angles and bonded into the main wall. The walls were averaged 40 cm. in thickness and were constructed, somewhat suprisingly at this stage of mudbrick recognizable shape and joins. The mudbrick was used along with an extensive amount of horizontal wooden ties, Idd both length wise and as cross pieces at various levels in the wall construction. In addition, the cross walls had layers of chaff tempered mud mortar between its courses of wood and/or mudbrick. ...“(Plate 20) ’®®Ibid, 329, plate??, fig. 38; Elmal i - K aratag 1 :15, plate ?b. The second technique known as the Tog-cabin’ was attested in the second level of trench 105 and in trench 115, in a more substantial form. Walls were made of timber laid horizontally and possibly held together with verticle props (which are not preserved even in the fallen remains of die wall). The timber wall was coated on its outside with mud plaster. The base of the wall in trench 115 was made of two layers of three logs which were embedded in mud packing. ’®* Ibid, 15. A third technique was encountered in trench 102, and the highest level of 105 and 116. This consisted of vertically and irregularly placed timber with chaff tendered mud pecked around them, the mud being smoothed on the outer face of the wall. With horizontal ties and wattling occuring occasionally. No post holes were found with this technique. ”®Ibid, 15-16, plate %b. '*° Ibid, 16, plate 9?b. ’*’ Ibid, 16. '*^ Eslick, 1988: 32. 56

Pottery:

The Late Chalcolithic pottery dominates the main assemblage of pottery finds, both fi’om the excavated site of Bağbaşı, and most of the surveyed sites on the plain. The sites which present Late Chalcolithic pottery can be listed as, Boztepe,

Karaburun, Gökpınar West, Maltepe, Semayük Bekleme, Dinsiz in the northern part of the plain; Tekke, Buralya and Yaka Çiftliği from the southern part of the plain.’”

The pottery forms include bowls, jars and jugs with a wide range of neck types with everted and flaring rims. (Plate 21)’” Baking pans with pierced impressions near the rim and stands are new in the repertoire. There is a change in the handle technology.’®’

Most of the forms find parallels in the Beycesultan and prehistoric Aphrodisias, although certain traits like the usage of white paint and deep necked tankards are missing features from the Elmalı Plain.’** The surface receives little treatment, only

10 % of the pots appear to be burnished. The paste appears to have had little preparation and lacks the fine grit temper of the earlier phases.’*’ The firing is poorly done, resulting in a black core.’**

Other finds from the Late Chalcolithic Bağbaşı can be grouped as the stone artifacts,’*® few metal objects, and clay artifacts.’®’ There are three artifacts which are not local to the region and may have been imported into the Elmalı Plain. They

Pimp.li - K a r a t a s I : 78,79. ’*'’ Ibid, 19-23, plates 16-56. ’*’ Ibid, 17,23-24. The handle is now attached to the bocfy by piercing a hole into the bodjy, and the handle with a plug at the end is inserted into the Ixxfy (ibid, plate 105, e,f). Strap handles together with loop handles start to be used Some of the strap handles are sharply terminated at the upper end Finger impressions and pierced holes sometimes appear on the lug and strap handles as a form of decoration (ibid, plate 49 sherd nos. 157,158, plate 50 sherd no. 169, plate 103c-f ’** Ibid, 78,79. ’*^ Ibid, 17,93-94. The grit particles seem to constitude the original clay, rather than being deliberately added by the potter. Fiber inclusions seem to appear in a higher percentage than in the earlier sanqples, leaving lots of voids in the clay. Fiber was deliberately added by the potter. ’** Ibid, 17. ’*® Ibid, 37-40, among which, we can count chipped stones, sickles, cutting tools, denticulated tools, scrapers, blades. ’®° Ibid, 40-41. Three copper objects, a composite object of an awl and chissel, another awl and a needle with a lo

are an obsidian tool, a green stone axe and a shell of the Mediterranean type, formed into a pendant.’®^

The existence of a clay stamp (Plate 22 with linear and geometric patterns is

peculiar and recalls the later Early Bronze Age stamp seals of Karat a?. The

existence of the stamp seal according to Eslick points to a complex social

organization, and that the inhabitants of the Bagba?i village were sufficiently

innovative to develop this method or adopt it.*®^

The location of Bağbaşı, being close to a water source on the fringe of the

hills with fertile soil, seems to constitute a suitable location for village settlements,

which is also evidenced in the continuous habitation of the area through the Bronze

and the Iron Ages. The Late Chalcolithic village, as suggested by the excavators, was

possibly a small village with its economy based on herding (cattle, sheep, goat) and

agriculture (though no evidence for grain in archaeological data has been recovered

due to the proximity of the levels to the surface). The thin debris accumulation

suggests a short lifespan for the village.'®* Eslick has suggested a different subsistence

pattern being adopted by the inhabitants of Elmalı, which is a form of swidden

agriculture that would necessitate moving often as land is left to fallow for long

periods of time before returning to it. 196

Archaeological evidence also indicates that the Bagba?i people were interacting

with neighbouring regions, as suggested by copper ftnds and the three imported

artifacts discussed above. An increased need for metal - copper mines lacking in

Southwest Anatolia - in manufacturing tools and objects, possibly, necessitated the

'®^ Eslick, 1988: 33; for the objects see Elmalı -Karataş I; 38, cat.no. 277 (obsidian blade), 40, cat. no 313 (green stone), 41, cat. no. 329 (shell). '®* Elmal 1 - K a r a t a ş T : 36, cat. no. 273, plate 106 e,f. '®^ Ibid, 88. '®* Ibid, 87-88. Especially when compared with the accumulation of the late Chalcolithic periods of Aphiodisias (six meters) and Beycesultan (eleven meters). '®^ Ibid, 88. 58

increasing interaction with other districts. This interdependence might have played a role, on the other hand, in reducing the self sufficiency of the villages.

Interestingly, most of the sites with pottery which resemble Late Chalcolithic wares seem to be concentrated in the northern sector of the plain, and usually along hilltops or, the skirts of the mountains that surround the plain. The Late Chalcolithic evidence from the lower plain areas seems to be missing, while for the Late Neolithic and Early Chalcolithic periods, we have evidence from the mounds located more or less on the plain’s basin. The shift in the settlement pattern can be attributed to climatic changes and the flooding of the two lakes located to the south of the plain and their marshy character enabling agricultural activity or settlements in the region.’®®This hypothesis needs to be supported by geomorphological analysis of the lake’s fluctuation limits in antiquity, by means of core drilling analysis. Another hypothesis for shifting settlement could be due to safety reasons, that the Late

Chalcolithic communities preferred to settle on higher ground, rather than the plain’s basin. Archaeological evidence from the Bagba?i site, however, points to a fairly peaceful village community where no circuit walls or weaponry has been attested; so the first explanation might be more plausible.

Bagba?i also presents a pre-stage for the development of the next phase of the

Early Bronze Age I period. The Early Bronze Age village of Karat a? is a direct

development from the Chalcolithic village of Bagba§i. The continuation of

architectural elements, pottery traditions, and the idea of using stamp seals and

decorated clay beads also supports cultural continuity on the plain.

Ibid, 88. Ibid, 88. 59

2.2.2 The Eariy Bronze Age Village of Karat aş - Semayuk'199

As a result of the Karata?-Semayuk excavation six major phases E B 1,11 and DI were identified. A summary of the Village chronology follows (Plate 23).^°° The earhest of the two phases that have been defined as the EB I period is represented by a single structure, the so-called central building. A small fire marked the end of the first period, after which the structure was rebuilt; the end of the second phase is marked by a greater conflagration than the first one. EB II period is represented by phases III to V.3. Karata? V.3 was defined as a transitional phase to the EB HI period. The EB II period shows an extensive remodelling of the central complex, with

a fortification system, after the levelling off of the central building. This period also

marks the existence of the earliest apsidal and megaron houses on the mound. The

main cemetery of pithos and jar burials belongs to this period, mainly to phase V. The

last occupation level of Karata?VI has been dated to the EB III period. The village of

Karata? does not appear to exist into the later phases of EB HI. Archaeological

evidence does not give any sign of violence or destruction or a hurried abandonment

for the latest phase. Nor is there evidence for a gradual decline.^'Within the western

Anatolian chronology, the first five levels of the K a ra ta ? village have been

associated with I period, and only the latest level of Karat a? VI with Troy

II. 202

The architecture of the K a r a ta ş Early Bronze Age village has been well discussed in E lm a lı - K a r a ta ş H, by J. L. Warner, which the information given here will mainly follow. The central mound, burials, pottery sequence and small finds will be published separately in a series of monographs, which are not available yet The discussion of these artifacts will mainly rely on the atmual reports published by M.J. Mellink in AJA. T.AD.. K.S.T. and on individual articles by MeUink, 1964c: 1-8; l%9c: 69-76; Mellink, l%9d: 290-99; Hesse and Perkins, 1974: 149-160; Wheeler, 1974: 413-425; Warner, 1979: 133-147; Esfick, 1988: 10-40. MeUink, 1966a: 247-255; 1984: 103-105: Elmalı-Karataş H : 7-10. Elmalı-Karataş II : 189. ^ Latest research and stratigraphical data at Troy may change some of the correlations, but basicaUy it wUl remain the same. 60

The next occupational evidence from the mound comprises the few burials of the

Middle Bronze Age and some pottery associated with it. The Late Bronze Age, as it appears throughout the plain, remains enigmatic for the mound of Karata?.

(a) The Early Bronze Age I (Karata? Village I and D)

Architecture:

Karata? I phase (see plan on Plate 24), has been suggested to follow after a short period of hiatus, at the end of the Chalcolithic period of Bagba?i.^°^

Karata? I level is represented by a single structure, a rectangular house surrounded by a courtyard and a buttressed wall, also referred to as the ‘central building’. The structure was supported with additional ramps and ramparts on its outer slopes, and a palisade which was made up of two rows of wattle and daub, with rooms existing on its inner (see plan on Plate 25).^

In the Karata? II phase (see plan on Plate 24) repairs were made and the burned walls were pulled down to level off“ the area. The central building was rebuilt, retaining its original floor level. A new ramp and retaining wall was built, and the

Mellink, 1969a: 329; E lma l i -K aratag 1 : 88-89: Elmali-Karata? I I : 9. ^ Mellink, 1984: 104; 1965a: 245-251; 1966a: 246-249; 1974a: 351-353; EsUck, 1988: 33-34. The rectangular building was constructed of carefully built mudbrick, which has been carefully plastered on its inner and outer fece. The rectangular building constitutes two floors, suggested by a staircase along the northeast comer of the house. The entrances for the building were along the short walls. The floors of the house and courtyard were finished with a buff plaster, similar to the walls. The lower floor level at least in its final stage was used as a storage area. Twelve large jars both in the pits and on the floors were found in situ, within the building, most of them being grouped in one comer. Four bins were located lined along the northern wall of the building. Post holes around the outside of the building suggest a support of the roof which partially covered the courtyard. A courtyard of three meters, which surrounded the rectangular building, is enclosed by an enclosure wall built in the same technique as the house itself. The enclosure wall was suj^rted with additional butresses in the interior. The butressed wall was additionally supported by a number of ramparts made of thick packing of rabble and pisé. The embankment around the oval court was built of herringbone technique of plano-convex bricks, embedded in cl^ . A rampart led around the south, east and west of the curvature walls to the central building. A palisade surrounded the oval courtyard and its retaining walls in a concentric &shion, following the same lines. A row of ‘fence houses’ forming a casemate wall existed to the interior of the palisade, possibly continuing all the way around the palisade.The walls of the houses were built of the wattle-and-daub technique on a row of posts, which were set into the cuts of the bedrock. The rooms of the houses showed evidence of domestic activity, with round hearths found in most of them. Pottery found on the floor was the domestic type of jars and cooking vessels. 61

northern entrance received certain alterations. The end of the second period came about with a larger conflagration, when the whole structure was destroyed.^®^

Although the specific function of the central building is unknown, perhaps a communal building, it has been suggested to be a residence of a person of high status. 206

Burials:

Some pithos and jar biuials have been dated to this period.^ The most distinctive burial of the Karata? II period is a built tomb (AQ 367)^°® which is definitely dated to this period(Plate 26). The unique nature of the burial has suggested the identification

of a chieftain or some one of higher rank, which may denote a stratified social

organization existing in the Karat a§ village of EB I.^ The special treatment of the

burial AQ suggests that at least some members of the society were differentiated from

the rest of the members.

An additional support for a differentiated society is marked by the existence of

stamp seals from the same level, which shows that a form of personal ownership

existed in the community. House 105 from the Late Chalcolithic level of Bagba?i,

its monumental complex and complex architecture as well as large number of vessels

and loom weights found in it, has led Eslick to suggest House 105 as possible stage of

development to the social stratification seen in EB I, Karata? village.^’®

Mellink, 1966a: 249. Eslick, 1988: 35. ^ Flmali-KflTata.q 11: 170. ^ Mellink, l%9a: 324-326, ill.3,4, plate 75; Eslick, 1988: 35. The rectangular built tomb had a surface maricer of a large stone circle with an outer diameter of six meters and the thickness of the enclosure wall of ca.one meter.The inner sur&ce of the enclosure wall was paved with smaller pebbles of white color. The five rows of large stones towards the east emphasized the orientation of the burial. The body of the deceased was laid on a thick gravel floor of 25 cm.. The position of the skeleton suggested a secondary arrangement, the burial had been transferred from a pretuninary tomb at some stage. The tomb pit was covered by earth in the form of a gable, its sloping edge packed with a mass of stones. Grave goods included a gold disc, some silver strips and casing, a stemmed bronze razor and jugs. Mellink, l%9a: 324-326; Esfick, 1988: 34-35. Ibid, 32. 62

Pottery:

The pottery assemblage from the EB I level follows closely the earlier Late

Chalcolithic tradition of Bagba?i the typical repertoire being curved bowls, ranging from shallow to deep, jugs and jars, two handled jars with vertical or flaring necks, vessels with elongated or bag shaped bodies, and pitchers with rising rims or beaked spouts beginning to appear (Plate 27).^" The wares are handmade, black or buff colored. Incised and relief decorations are commonly found among the pottery of the Karata? villages I and II. pithos jar (No.57), excavated in 1963, from the central part of the main cemetery, had the interesting combination of incised signs of swastikas, T-shaped standards, mountain goats, plant motif and a number of the wooden hut signs, which had been associated with the sign 24 of the Phaistos Disc, by

Mellink (Plate 28)^“*. Surviving forms of the portable wooden huts still prevail in

Northern Lycia. M.J. Mellink^*^ has convincingly identified it as a prototype of sign 24 on the Phaistos Disc, which has been dated into the Middle Minoan nib context of ca.

1600 B.C., much later than the Early Bronze Age context of Karata?, which the burial pithos appears from. Parrallels had been drawn by Sir A. Evans^'® between the

Lycian tomb facades and the very same sign, for which the Karat a? pithos seems to give further evidence. Ajar fragment in the burned debris of the central mound, from the Karat a? level II context, has proved to have similar incisions of goats, as it appears on the pithos jar no. 57, tentatively placing the pithos jar 57 in the Karat a?

II context. Even though for the time being it will be immature to make a statement for direct connections with the Aegean world, however, this detail needs to be

Mellink, 1965a: 249, plate 64, figs. 29-31, plate 66 figs. 41-44; 1966a: 249-50, plate 60, figs. 21,23. Elmali-Karatas I I : 8. Ibid, 8; Mellink, 1966a: plate 60, fig. 22, plate 62, fig. 37. Mellink, 1964a: 275. Mellink, 1964c: 1-7. Ibid, 6-7. Mellink, 1987: 22. 63

admitted for a possible connection, and fiiture studies may bring more evidence into this subject.

(b) The Early Bronze Age H (Karata? Village III-V)

Architecture:

The Karata? HI phase^’® (see plan on Plate 24), points to an extensive remodelling of the central mound with the levelling off of the central building.^'® The structures which may have been built over the leveled off area have been totally lost for the archaeological analysis, due to the heavy erosion of the central mound.

The earliest examples of apsidal houses and megaron type houses start to appear with Karata? III.^° The superimposed apsidal houses of level III show a rear antae in a curved tradition. Although whole structures have not been preserved, due to overlapping activity in the region, the curvature of the extending main side walls prove that they had been built in this fashion. In one of the apsidal houses a front

The Karataş Village in phase has been taken within the EBI period by Warner tElmali Karatas n : 169-71), however, it has been considered within the EB II period by MelUnk throughout the excavation reports, 1966a; 251; 1984: 104. Mellink, 1966: 250-51. After the big conflagration which sealed level II, the burnt remains of the walls were pulled down and leveled off" towards the inside of the oval coiut and partially over the surrounding ramps. The builders of level III have been held responsible for the destruction of the previous level. An interesting pit (pit no. 10) is located within the house, cutting through the destroyed levels and ruins of the earlier stmcture. Possibly before remodelling the house and its courtyards, the inhabitants of level III removed what was still usable firom the house. After the inspection pit had served its function, the pit was filled with the burned w all, as the rest of the area. The builders of level III had also smoothed and solidified the slopes of the southern edge of the central mound. Ash strata in form of deposits, layer after layer of clean wood, have suggested the careful accumulation of ash on the slopes of the central moundThe ash debris retains very little archaeological data, to suggest for any function. Few spindle whorls and beads with geometric designs, together with burnt bone occur in the ash debris. Suggestions regarding the function of the area have been made, as the possible existence of a pithos factory, which would have served the need for the abundant pithos and jar burials found at the site. The existence of the kilns would have also necessitated the regular cleaning and depositing of the ash. Neither the remains of a kiln, or misfired sherds, nor traces of any industrial activity has been recovered from the central mound The area has been also suggested to have functioned as a possible public place during the Karatas IV phase. Periodic gatherings may have taken place outside the entrance of the central building, and the open fire place may have served for the preparation of food during some kind of a festival (Elm alı- Karatas II: 171). Warner, 1979; 138-39. Three superimposed apsidal houses were recovered from trench area 35/37, northwest of the mound (refer to plate 31 for their plans). 64

porch with projecting antae exists. In only two of the houses entrance into the apsidal area has been attested through the main room.^^' The closest examples to the apsidal houses have been shown from Western Anatolia, Islands, Southern and Central

Greece, Thessaly, and Southeasten Europe.^^

In the Karata? IV phase (see plan on Plate 29), the central complex once again goes through a heavy remodelling with the occurrence of a fortification system. What the central area and the courtyard resembled in this level is totally unknown due to heavy erosion. Period IV comprised an enclosure much greater than the preceding three phases. ^

Three irregularly shaped megaron buildings appear from this level, located along the southern slopes of the central mound.The cemetery area of the Karata? IV village is located to the northwest of the mound, overlying the habitation levels of

Karata? Ill, and the three apsidal houses.“"

The Karata? V phase (see plan on Plate 30) shows the greatest expansion of

the village, with the most dense habitation of all the Early Bronze Age occupation of

the mound. A large area along the northwest side of the mound was inhabited.

Dwellings to the east and southeast of the mound were encountered for the first

F.lmali -Karatas II : 139-40. Warner, 1979:137,144-45. For Western Anatolia, Troy I (house no. 103): Blegen eLal, 1950: fig. 425. For Southern and Central fi’om Asine : Frodin and Person, 19: 92,93 fig. 68-69; from Lema : Caskey, I960: 293-94, which the apsidle house dated to EH n i appears slightly later than the Anatolian examples and fixjm : £>aux, 1%8: 859-61. For T h e ^ y fi^om Chasabaliand and Rachmani : Wace and Thompson, 1912: 25-53, fig. 17. From Southeastern Europe for Vucedol : Smith, 1945: 15-20 and for Sitagroi : Renfiew, 1970: 133, fig. I, plate XIX. Mellink, 1966a: 251. The enclosure was surrounded by a heavy wall or embankment of pisé, over the sloping ash strata of level Ш.ТЪе wall consisted of a pudding like substance, well mixed of buff mud and lime specks. The outer face was modelled in the form of a glacis, the wall strengthened with stones and thickly plastered with enormous amount of light colored speckled mud. The width of the wall averaged between three and four meters. The enclosure had a gateway on the south side. A road, with a gritty white floor level, in the form of funnel (five meters outside the gate and narrowing down to three meters in the interior of the gate), would pass through the gateway. Cuttings in the underlying ash levels have pointed to tower like projections on the east side of the wall. Warner, 1979: 138-39. 65

The estimated area of occupation was ca. 60 000 m2 (60 ha), with an estimated population o f640 based on the calculation of 128 households for level V.“ ®

The central complex of the mound is abandoned during this period without any sign of disaster or conflagration. Evidence of domestic activities is to be seen at this time, on the central mound.

During this period the consistent type of house architecture is the free-standing, rectangular house form of ‘megaron’ (Plate 3 1 ).^ Half of the total of 18 megaron houses on the site belong to the Karat a? level V. The location o f level V and level

VI, (which give the main bulk of evidence for the megaron houses) right below the plowing zone of the modem agricultural activities, has resulted in considerable destruction of these houses. In most of the cases only a line of foundation stones remained. Floor levels and any evidence for domestic activities, eg. platforms, benches, hearths and ovens were destroyed considerably. But sufficient traces of foundations remained in order to reconstruct plans of the houses.^

Rimaii-Karatas II : 172. Ibid, 177, also see chart 6. Mellink, 1966a: 251. ^ R im ai 1 -K a ra ta s II : 137. The well debated tern of ‘megaron’ for K a ra ta ç has been defined as following » ··· A building plan of a free standing rectangular structure. The basic characteristics are two structural long walls with cross walls inserted to form a main room and a front porch, which are entered axially on the short side. The long wall ends in antae at the front, the rear cross wall is usually set back from the ends of the long wall, which this projects as a rear antae....“ ^ A detailed description of the houses, their internal arrangement and constructions methods have been given in a separate chapter (chapter 4) in Elmal i-Karata$ n : 135-57, by Warner, which the information given here fallows. The main room constitutes the core of the building, where daily activities of sleeping, eating and domestic woric would take place. The size of the rooms vaiy between 3x4 meter to 5.3x9 meter. In few instances hearths which occupy a more or less central position have been located but in most cases the floors have been destrt^edThe porch or antae room is a roofed area, which would provide enough space for certain domestic activities or for storage in this part of the house. The depth of these rooms varies between 0.90 and 3.80 meters, with a common depth being between 1.20 and 2.40 meter. A hearth in this area has been located in some of the houses. The rear room, which has been sometimes additionally added or created by a division room within the main room of the megaron, would have been used as a storage area, as suggested from the large number of lined pithoi and storage jars found in situ. Though not in every case, but sometimes, a doorway is evident from the rear wall of the main room into the storeroom area. Annexed quarters in the form of X ’-shaped walls have been also located in some of the house plans, which suggest a later addition and some functional usage as storage areas. Screen walls which provide protection to the porch area have been also located. Their foundations occasionally being of stone, in most cases have left no traces. 66

Construction techniques:^°Stone foundations are used in EBII period, more often in the form of stone settings used in reinforcing the base of the walls. Regular stone foundations are not used till the EB IQ period, which occurs together with the use o f regular sized, mold bricks.“ ’ A variety of construction methods have been encountered for the superstructure, including traditional methods of wattle-and- daub“ ^, pise’^“, as well as a combination of both.

The evidence regarding the roofs of the houses remains fi’agmentary. But some surprising evidence of burned fragments of clay has suggested that a reconstruction of the superstructure had been undertaken. The hardened fiagments of a lump of clay, located in house 100a (Karat a$ level V.2/3) shows the traces o f a vertical wall beam, a horizontal cross beam (tie beam) and a slanted roof beam (rafter). The combination of the slanted roof beam and the horizontal cross beam gives the evidence that the so-called megaron houses of K a ra t a? may have displayed pitched roofs.“ '·

Parallels to the megaron houses mainly come from Troy, Southeastern Europe and

Southern Greece.“* The dominant form of megaron houses seen in the Early Bronze

^ Fim ali-Karatas n plates 198-206. Some contemporary forms of this technique, still being used by a group of modem dwellers south of the E lm al i Plain, near Yuvalılar (the ancient Limyra) have been discussed in the same chapter, the modem examples providing an ethno-archaeological data for the building tradition of the region (Plates 32, 33). Mellink, 1989: 269. Warner, 1979:139. The predominant method of wattle-and-daub remains the basic technique for the lniilding up of walls, sometimes used with the pisé technique. The wattle-and-daub technique is descritwl as following: „... A framewodc of densely packed solin g is covered on both sides with chaff-tempered mud walling material packed onto the sapling, without the aid of forms. This tf

“ * Ibid, 139. Pisé walls were constructed directly on the ground, with the mud walling giving the impression that it was packed and hardened between boards. Pisé stmcture on a stone foundation has been attested only with one house in K a r a t a ş level IV, house MS/a. Warner, 1979: 141, ill.4,5; Elmal i-Karataş n :149-154, fig.12,13. “ * Warner, 1979: 137,144-15. For Troy: Blegen (1950) 89-94 text, frgs. 418-419,426. From Southeastern Europe die main sites being Banjica, VugedofKaranova (see Warner footnotes.32 an 34 for further ret). From Thessally, for O lzaki: Milojcic, 1960: 12, 34 frgs. 2-4; for Dimini and Sesklo :Tsountas, 1908: 50 frg. 10,60 fig ll(for Dimini) and 90 fig. 18 (for Sesldo). With Dimini and Sesklo examples going as early as the Neolithic period and for Volo ‘PdScakia’ : Milojcic, 1974: 67

Age Karat a? village suggests a strong cultural contact with Western and

Southwestern Anatolia, Northern Lycia appearing to be an integral part of this cultural sphere.

Burials:

Pithos and jar burials are attested during the Karata? IV period; this tradition is frequently used in Karata? V. The EB inhabitants of Karata? used an extramural, inhumated pithos or jar type of burial. They were accompanied by tomb gifts, placed both inside and outside the jars. Habitation areas and cemeteries were strictly separated from each other in all the periods. A main cemetery area was located to the southwest of the mound, and different parts of the mound were used as a cemetery in different occupational periods, from time to time resulting in the overlapping of the habitation and cemetery zones. In few instances intramural types of burial occur, where isolated groups of children and adult burials were found outside and near some of the houses, within the Karat a? EB II and III context. Burials dug into abandoned houses have been also reported.^

Over 500 jar burials from the Karat a? site were located, with almost all of them coming from the main cemetery area, to the southwest of the mound (Plate 34). They have been dated to Karat a? V level, based on the tomb gifts and the typology of the burial jars which accompanied the body of the deceased.^^^ The burial jars have been grouped mainly into two; the pithoP® and the household jars o f the domestic type., 239

662-70, figs. 209,210. From Southern Greece for Lema, only one house (IV H) has been identified as megaron, see Warner, 1979:143 and footnote 13. MeUink, 1%7: 251-57. The chronology of the tombs and their relationship to the different habitation levels of the K a ra ta ? village will be published in the monograph series by Tamara Stech Wheeler. Mellink, 1964a: 273, plate 78, figs. 5,9.The pithoi are heavy, thick waUed vessels, with heights reaching over 1.50 and 1.80 meters and wall thicknesses of 6 cm. They have stump bases and thick flat bases, with lug handles on the shoulder. The fabric is a chaff tempered clay, coarse but fairly well fired. Some of the jars have decorations of thumb prints, coilings, relief medallions and incisions on the outer sur&ce, usually near the neck part and over the shoulders near the handles. Ibid., 273, plate 78, fig. 8. The second type of burial jars are the practical sized household jars with thinner trails. Sizes vary between 0.75-1.20 meters, with flat bases and loop handles of usually 68

The extensive, and well preserved Early Bronze Age cemetery of Karat a? has

provided specific details in reconstructing the burial tradition of the Early Bronze Age

people.^'“ The pithos or jar burials have been strictly placed in a fixed alignment,

where the opening of the jars were oriented towards the east. The mouths of the jars

are closed with large slabs, smaller stone piles, bowls or large fragments of broken

pithoi (Plate 35). The bodies are buried in a contracted position lying either on their

left or right sides. Most of the jars were reused on later occasions; the remains of the

earlier bodies would be pushed to the bottom of the jar and the new remains placed

above, again in the contracted position. The regular placement of the jars and their

reuse has suggested the existence of a form of tomb marker on the original surface,

which would have been destroyed with modem agricultural activities. A number of

large stone circles with diameters varying between one meter and 6.5 meters, and

encircling walls ranging from 40-50 cm., have been located right over a group of jar

burials (Plate 34 A,B). These stone circles gave evidence of the general nature of the

Early Bronze Age tombs, as they appeared in the past.^"' Tomb gifts also follow a

regular pattern, being placed both inside and outside the burial jars.^'*^

The Early Bronze Age burial tradition of Western Anatolia, which is more or less

homogeneous with its distinctive tradition of inhumated jar burials, has been attested

from various sites in Western and Southwestern Anatolia.

four, arranged symetrically around the shoulder. A smaller type of jar of the domestic size has been also attested for the usage of child burials (Plate 44 A). ^'^Mellink, 1969a: 319-24. Mellink, l%8a: 255-258, plate 81; 1969a: 321-324. Mellink, 1964a: 273-274. Tomb gifts generally include a variety of pottery (pitchers, jugs, small jars etc.) placed both inside and outside the pithos. Usually, in the inside a number of more precious objects such as metal objects, alabaster and marble idols, bronze-co|^er jewelry of bracelets, pins, toggle-pins, small rings, spindle whorls, beads and shells would be placed. A detailed discussion of the Early Bronze Age burial evidence from Anatolia, including the sites of Yortan, Babakôy, Ovabaymdir, , Kusura, Beycesultan, prehistoric /φhrodisias, lalysos. Outside of Anatolia, the nearest comparanda to K a ra ta ç burials appear to come fi'om the extensive cemetery of Byblos (over 1500 burials), dated to the Chalcolithic but continuing throughout the third millenium. A comparison with the neighbouring regions of and Aegean and Greece has been given in the article by Wheeler, 1979: 133-147. 69

Pottery:^""

The EB n pottery of Karata? is dominated by red polished ware, with white paint and relief decoration (Plate 37).^"^The white paint occurs in Karata? Level IV,

and becomes the dominant form of decoration on red polished ware in Karat a?

Level V. White paint on black burnished ware also occurs in level IV. The incised decoration, which has been a common type of decoration since levels I and n , is filled with white paint during Karata? Level III (Plate 36). Common forms for the EB II period are the beaked pitchers, bowls and four handled jars. Strainer spouts (Plate

37 B), though rare, appear in level IV .T w isted handles also show their first appearance in level IV (Plate 36 D).

An unusual vessel ( KA 954, see Plate 38)^'*®, which has been attributed to some ceremonial function, has been recovered fi’om the floor level of a rectangular structure, built of timber and pisé on stone socles, dated to the EB II period. Similar vessels have been pointed out from M ari^, Lema“ ’ and Cyclades^*^. This vessel brings to mind the ceremonial vessels attested from Mesopotamian cylinder seals.^253

Imitations of these seals appear also from Kültepe Kamm level II.

The Early Bronze Age typology of K a ra ta § pottery is being worked on by C. Eslick, which will be published in the monograph series. A short summary of the main features of the development of the EB K a ra ta § pottery has been given in Elmali-Karatag II8-9, by Eslick; the publication also includes a catalogue of finds, including pottery. Annual reports of the site have also discussed the pottery evidence of the various levels. Mellink, 1984: 104; see Mellink, 1964a: plate 78 fig. 8; 1966a: plate 58, figs. 11-14, plate 59, figs. 15-20; l%8a: plate 84, figs. 32, 34, 35, plate 85, figs. 39,41 for red polished ware with white painty and plate 83, figs. 24, 27 for relief ware with medallion. - K a r a t a s i l : 8, for plates refer to previous footnote.. Ibid, 8, plate 181c Ibid, 8. Mellink, 1969c: 69-70, fig. 1. The fully reconstructed pottery had a bowl shape with a flaring pedestal base. On the shoulder, four tubular spouts were attached symmetrically,with two grooved loop handles attached betwen each pair of the tubular spouts. Horizontal grooves and a continuous chevron band decorated the Ixxfy of the vessel. The &bric is the usual K a ra ta ? fabric, pointed out as a local production. The unusual size (40 cm.), heaviness and form of the vessel with its four spouts and capacious Ixxfy has led the excavators to interpret a ceremonial function for itThe vessel was not meant to be moved around firequently but had a fixed place, and a group of at most four persons could sip from the vessel through drinking tubes or straws. Mellink, l%9c: 74, fig. 4; Parrot, 1956: plate LXX, no. 674. Mellink, l%9c: 75, fig. 5; Caskey, 1956:162, plate 43. Mellink, l%9c: 75, fig. 6; Zervos, 1957: 58; fig. 13. Mellink, l%9c: 71-72, fig. 2,3; Frankfort, 1939:77. Mesopotamian Early Dynastic drinking scenes are well attested from cylinder seal designs that show such drinking ceremonies. These seals 70

K a ra ta ? level V.3 has been defined as a transitional phase into the EB in.

During this phase a number of new innovations in pottery manufacturing, typological change in household artifacts“ * become apparent, which is to continue into the next phase of EB HI. Wheelmade vessels start to be identified among some of the burial gifts. Among the pottery repertoire are the gray, grooved or red polished tankards with one handle“*, pitchers with exaggerated beak spouts teapots with basket handles^*®, and red lentoid flasks“ ®, which make their first appearance (Plate 39).“®

Among the metal finds, a distinctive metal artifact of a silver double-axe found at the central mound, has been dated to the EB H-III context.“ ’ The double-axe could be considered as the earliest ‘metal ’ found in Eastern Mediterranean. Its elongated slender form and the asymmetric treatment of the upper and lower edges, with slightly sloping curvature, has been shown to have parallels in the Anatolian examples from Kültepe IF®^ and ^®*, rather than the Minoan ones.^264

are attested in northwest and north Mesopotamia, up to Mari and Assur during the Early Dynastic II and m period. MelUnk, l%9c: 72; Ozgug, 1%5: 56, no: 39-41,46,49,70,74. The vessels have been shown in Hftfail on these cylinder seal impressions and their actual examples do appear in Kultepe Karum II period. “ * RlTTial 1 -K a ra ta s II plates 184,185. The spindle whorls which have been attested since K a ra ta § level I through level V, have the truncated biconical or globular profile (Ibid., plate 183). With the K a ra ta § level V.3, new forms of flatter spindle whorls appear in the awemhlage, which become dominant with the next phase. Typological change is also encountered in the shape of the grinding stones (Ibid, 9). ^ Ibid, 1%, see catalogue of finds for, Inv. no:KA 462, plate 175d Ibid, l% -98, see catalogue of finds for, Inv. no:KA 474, plate 174g, ,KA 567 plate 174h, KA 492, plate 175^ KA 675, plate 176h, KA 634, plate 176e, KA 519, plate 176a. “ ® Mellink, 1964a: plate 81, fig.20. “ ® Elmali-Karatas II196, see catalogue of finds for, Inv. no: KA 424, plate 172e. ^ Mellink, 1984:104. “ ’ Mellink, 1%7: 265, plate 84 fig.50a,b. ^®^ Ibid, 265; Ozgug, 1955: 70, fig. 41a and b, 43. The copper battle axe? has been revealed firom the Kultepe Colony Period level II. Mellink, 1%7: 265; Goldman, 1956: 289, fig. 425, 27. Mellink, 1967: 265. An interesting similar treatment of a double-axe appears from Sardinia, but from a rather later context, dated to 1200 B.C.. Pinza, 1902: 174-75, fig. 90, i^ate XVI, 30 and plate xvn.6. 71

(c)The Early Bronze Age E l (Karata? village VI)

Architecture:

Karata? VI phase (see plan on Plate 40): The latest level of the Karat a?

village has been considerably destroyed, due to its proximity to the plowing zone.

The best preserved megaron structures appear in this level, and this may be due to the

advanced technique of construction during the EB El period. These megarons are

more substantial buildings with stone foundations and regular sized mudbricks, and

with central hearths. The late appearance of mold made mudbricks in Karat a? has

been attributed to a number of factors by Mellink.^

Burials:

No burials were associated with Karat a$ level VI. Except for an inhumation

burial cut into a house, no burials could be safely attributed to this level.

Pottery:

The wheelmade pottery dominates the main assemblage of EB IE, in contrast to

the previous period’s assemblage, in which wheelmade wares made their appearence.

Two handled tankards^®^ depas cups, plates, platters and dishes, baggy shaped storage jars, red pitchers with exaggerated beak spouts are the common forms of this period

(Plate 41).^®An anthromorphic vessel has also been dated to this period (Plate 42).“ ®

One of the reasons being the strong reliance of the West Anatolian inhahitant-i: on the free molding of mud or clay substance on timber and wattle frameworks. A second reason has been shown as the difficult nature of the local brown and white clayey soil, especiaUy as a mixing agent Mellink, 1989: 378. Klmali-Karatasn: 174.. Ibid, Inv. no: KA 332, plate 165a. Ibid, 9; Mellink, 1984: 104-105; l%5a: plate 65, fig. 36; l%7a: plate 82, fig. 43 and plate 83, figs. 44,46. 1 -Karatas II: 197,see catalogue of finds for, Inv. no: KA 602, plate 166. 72

2.2.3. The Middle Bronze Age Evidence from the Bagba^i and Karatai? Sites

Architecture:

Although very scanty, the evidence for the Middle Bronze Age period comes

both from Karat a? and Bagba?i sites. Middle Bronze Age remnants of

architecture was evident from the surface levels o f Karata?.^™ From the site of

Bagba?i, remains of stone and timber structures were found.^^'

Pottery:

From the site of Bagba?i fragments of a quatrefoil unstemmed kantharos were

revealed (Plate 43 Parallels for the vessel are observed from Kultepe and

Yanarlar cemetery near Afyon^^^ and Beycesultan level IVa.^^'' Other types of pottery

forms include carinated bowls, basket handled jars, teapots on ring bases, cooking

pots with crescent shaped lug handles (Plate 43).^^^ No resemblance to the Troy VI

assemblage has been observed. The pottery data, however, does finds strong parallels

with the Tarsus Middle Bronze Age repertoire, and has been dated into the Old Hittite

period.^’® As pointed out by Mellink, as an indication of changing relationships of

Lycia, by about 1600 B.C.^”

Burials:

Burials were encountered both at Bagba?i and Karata?. The Karata?

inhumated jar burials o f‘D’ and ‘E’, were revealed along the slopes of the central

Mellink, 1984: 105 ; M ellink, 1985: 287. namely from trenches 119 and 120. Mellink, 1985: 290, fig.1, inv. no:KA 715. Ibid, 287; Emre, 1978: 98,116 ^ g, plate XXX. Ibid, 287; Lloyd and Mellaart, 1965: 136, plate 35. Ibid, 287, figs. 1-5. 276 Mellink, 1%9: 330; Goldman, 1956: fig. 302 (carinated bowls), fig. 297 (basket handles), fig. 299:926 (ring bases), fig. 310:1071 (cooking pot with crescent lug). Ibid, 330. 73

mound.^’® The continuation of the Early Bronze Age jar burials indicate the conservative tradition of burials in the region, with the only difference being that the mouths of the jars were oriented to the south at Karat a? and to the west in the

Bagba?i burials. The pithos jars are now ribbed and wheel-thrown (Plate 44 B).

The way the jars have been propped up with the aid of stones and the mouths have been covered with large slabs, is the same as the Early Bronze Age examples^^®. Tomb gifts include spindle whorls in the typical Hittite fashion,“ ® a bronze ring, a variety o f pottery including wheelmade bowls on ring bases“ ', a small pyxis and a buff jar or teapot.“ ^

The existence of the Middle Bronze Age occupations on higher hill sites has been suggested by Mellink to reflect a shift in settlement pattern. The Middle Bronze Age inhabitants of the plain, for some reason, might have preferred to settle on higher ground rather than the open, flat plain.“ ® The increasing influence of the Hittite traits in this region further points to an increased contact with Ifittite centers and the

Anatolian Plateau during this period.““

Mellink, 1966:257, plate 78, fig. 27 (for burial D); l%9a: 330, plate 71, fig 4 (for burial E). Mellink, 1967: 257. “ ® MeUink, 1985: 288. Inv. no. K A 122. The qnndle whorl with its convex top and deeply cut concave bottom reminds one of the Hittite examples with comparanda from Gordion (Mellink, 1950: 43) and Afyon Yanarlar ((Emre, 1978: 113, plate XLIV). “ ' Mellink, 1985: 291, fig.3: Inv. no.KA 368N. Ibid., 291 fig 4 (KA 707N), fig. 5 (KA 708N). “ ® Ibid, 289. “ “ Ibid, 289. 74

2.3 mON AGE EVIDENCE FROM THE PLAIN

Archaeological evidence from E lm a lı Plain presents a hiatus from the Middle

Bronze Ages till the eighth and seventh centuries b.c.. Archaeological investigations^^ of the B ayındır and Karaburun regions, and Kizilbel have documented tumulus burials exhibiting a variation of both inhumation and cremation in wooden chambers, sarchophagus burial and stone built chamber tombs. More well known burials, like the

Kizlibel and Karaburun II chamber tombs, display a rich wall painting repertoire.

The architectural style of the tombs, and their contents have suggested the infiltration of strong Phrygian and Lydian influence into the plain as early as the eighth and seventh centuries b.c., as it has been attested with the B a y ın d ır and Karaburun III and IV tombs. The sixth century tomb paintings of Kizilbel, as it has been observed from the subject matter, iconography and style, show strong East

GreekAVestem Anatolian/ Lydian influence with certain local traits prevailing. In the

Early Classical period of the first half of the fifth century bc., a mixture of Persian and

East Greek/Westem Anatolia influence becomes evident with the wall paintings of the

Karaburun II tomb. Other than the tumuli burial tradition attested from the northern section of the plain, the Lycian type of rock-cut tombs also appear, mainly along the southern edge of the plain.^ No substantial settlements related with the burials of the iron age have been found on the plain, except the very scanty remains from the

B ağ b aşı and Karataş sites.

A large number of tumuli burials were encountered, mainly concentrated to the north of the plain Most of these tombs were opened Roman grave robbers in antiquity, and are endangered by modem tomb robbers who seem to follow the activities of their Roman forerunners. A number of these burials were salvaged through excavations. In 1969 the tumulus and tomb chamber of K i z i l b e l started to be excavated, as a joint project of the Bryn Mawr College and the Antalya Museum. Its excavation and conservation programme continued into the 70s. In 1970, two more tumulus burials, Karaburun tumuli I and n, came to light The excavations of Karaburun HI and IV were accomplished in 1971, when in the same year, work at Karaburun II and K i z i l b e l continued. In 1972, an additional tumulus of Boztepe was investigated and partially excavated. See pg. 90 for a discussion of these tombs. 75

Architecture:

The site of Bağbaşı presents scanty evidence of architectural remains belonging

to the Early Iron Age. Burned fragments of walling indicate that architecture

continued to be of the traditional wattle-and-daub construction.

Burials

Tombs of common inhabitants were found towards the east edge of the

Karat aş-Semayük mound. These were the traditional types of inhumations found in

pithoi and jars. The pithoi and jars were the wheelmade banded type.“® The

continuation of both the traditional style of wattle-and-daub technique, and the pithos

burials point to the conservative nature of the Elmalı Plain inhabitants.

Pottery:

The evidence for Iron Age pottery from the Elmalı Plain has been mainly

attested through the expeditions of Mellaart’s survey of Southern Turkey. He has

mainly identified two types o f Iron Age Ware from the plain, what he refers to as

black-on-red II and bichrome ware, all of the sherds found at the site of Söğle.

The first group of black-on-red II ware has been dated by Mellaart with an upper

limit in the eighth century, but generally in the sbcth century b.c.. More specifically, in

Xanthos the black-on-red n ware seems to occur together with imported Greek

wares, suggesting a date for of seventh and sixth centuries b .c ..“® Mellaart describes

the black-on-red II ware as a less refined type of the black-on-red I. The usual

decoration of geometric patterns is applied as a grey black washy paint on a red or

brown surface, with the surface often unslipped. Forms usually include fiuit dishes.

Mellink, l%9a; 330-3İLİ970a: 250-51; 1984: 105 “ *Ib id Mellaart, 1955:115-36, for the bichrome wares from SO^le see pg. 133, plate 4, sherd nos. 64-66, pg. 135, plate 8, sherd no. 114 and for the black-on red n ware see pg. 135, {date 8,sherd nos. 112, 116. SOgle is located along the southwestern edge of the northern part of the plain, right across from the SemayOk-Karata§ site (Plate 15). Ibid., 122-123. 76

pedestal bowls, wide open bowls with ring or raised base, jars, juglets, small bowls,

small vessels with wide spouts.^'

The second group identified as the bichrome painted ware, has a decoration in matt black and red on a white or buflF slipped surface. Shapes mainly include bowls, jars and craters.^ It is identified as the typical Iron Age ware of the Southwestern

Anatolian Highlands.^ Other than the few specimens from Söğle, a complete urn

(Plate 45 A), with a dinos like shape was recovered from the Karaburun Tumulus ITT identified as belonging to the group of the bichrome painted ware of Mellaart’s, and dated within the sixth century B.C.^ Iron Age sherds of the bichrome ware have been also recovered from the site of Bağbaş i . They were described as having red zones between black bands combined with hatched triangles and concentric semicircles

(Plate 45 C).^^ Fragments of a large storage jar or amphora, of local fabric with reddish-brown paint, was found from the site of Bağbaş i . ^ Bichrome ware also appears in the stratified levels of Xanthos.^’

An East Greek import of a stemmed goblet; a buff ware with smoothed surface of blackish brown paint was identified from Söğle Höyük, pointing to East Greek contacts with the plain.^®* East Greek imported wares of the Wild Goat style, Fikelura,

Ionian cups and Rhodian plates are commonly seen in the seventh and sbcth centuries strata of Xanthos.^

nrid.,119. ^ Ibid, 120. ^ Ibid, 118,120. The Ware ^ has been defined as extending from the upper Meander valley to the south coast particularly being common to the south of Denizli-Burdur line, and especially located at Fuğla Höyük, north of Korkuteli and the Elmalı Plain ^ Mellink, 1972a: 261-62, plate 57, fig. 12. ^ Mellink, 1969a: 330, plate 76, fig.33. ^ Ibid, plate 76, fig.32; 1970a: plate 58, fig. 119, Inv. no:KA 752. The amphora after being reconstructed had a height of 32 cm., had designs of pendants, semicircles around the neck, and concentric circles on the shoulder zone. ^ Metzger, 1972:63-68, plates 23-24. ^ Mdlaart, 1955:135, plate 8, sherd no:l 17. Metzger, 1972:27-57, plates 2-19. 77

2.3.1 The Bayındır Tumulus Graves and Karaburun III and IV Tumuli

The tumulus tradition in Anatolia has been associated with the coming of the

Phyrigjans. The Phrygian elements of construction of a wooden chamber sunken into

a pit, sealed with layers of clay, pebble and earth is well attested from the excavations

at the Phrygian capital of Gordion.^“® The Iron Age cemetery of Gordion also gives

evidence for a variety of burial styles, varying from inhumation to cremation as well as

pithos and jar burials which seem to have their predecessors in the Hittite levels o f the

site 301

A whole range of tumulus burials, about a hundred in number, are concentrated in

the northern sector of the Elmalı Plain. They have been mainly located in the vicinity

of Bayındır, Karaburun, Söğle, Geçmen, and Karabayır.^'^They usually dominate

hiltops overlooking larger parts of the plain. The construction methods and contents of

the excavated tumulus burials of Bayındır and Karaburun very much resemble the

Phrygian examples.

(a) Bayındır Tumulus C ^ The tomb pit was dug into the virgin soil, and

covered with the tumulus mound. Four stratified layers were attested. Above the

virgin soil, into which the burial pit sunk, a layer o f clay would be placed, followed by

a course of fiat and larger stones, and a further course of pebbles of smaller size.^®^

Tomb C appeared to be a cremation burial. A burned layer was surrounding the

^ Young, 1981: 263-70. Mellink, 1976:22; Mellink, 1950: 1-15,45-51; Gunter, 1991: 4-7. ^®^ Geçmen is located 10-12 km. southeast of Elm alı, on the road to Söğle. Karabayir is lacated 30 km.north of Elmalı. MeUink, 1972a: 261. Özgen, 1988: 32-33. ^ Bayındır village is located along the eastern slopes of the Elmalı mountain in the northern half of the plain, few kilometers to the north of the Bağbaşı-Karataş sites (see Map 1).A group of built tumuli burials, numbering to 16, were located in the Bayındır village and the vicinity of Çağıltemeller in 1984. Some of these burials had been opened or tried on through illegal digs. In the two consequent seasons of 1986 and 87, tombs C and D were excavated respectively by the Antalya Museum. Dörtlük, 1988: 172. The dimensions of the tumulus were 36-38 meters and a height of 4.20 meters at its peak. The pit was 1.50x0.83 meter in width and 0.30 meter depth. 78

original pit. A group of tomb gifts was found in a damaged state due to the conflagration.^

(b)Tumulus D appeared to be an inhumation burial, consisting of a wooden chamber sunk into the virgin soil where the layers of the tumulus pile followed the same sequence of Tumulus C. Eight pit holes were discovered, that possibly would hold the beam supports for a wooden roof, which had collapsed under the weight of the stone pile. The skeletal remains belonged to a female. The mhumated body o f the deceased was laid on a wooden kline of iron construction. The tomb gifts consisted of personal valuables and ceremonial vessels.^ The tomb gifts of ceremonial character would well suit a banquet or symposium, very much reminiscent of the

Phrygian and Lydian examples of tumulus burials. Further evidence of an inscription in the Phrygian language on a silver cauldron and ladle has suggested that they have been imported or donated by Phrygian residents.^®® An important group of finds was the four figurines (Plate 46 C, D, and Ef'\ one silver priest figurine, and three of ivory. Holes above their heads has suggested their usage as perrirhanteria. The well- known ivory female figurine depicted with two children, a small girl on the right, and a small boy sitting on the left shoulder of his mother has been identified as wearing garments in the Western Anatolian/East Greek fashion, with a polos-like head gear and a long veU which covers the whole of the robe and is gathered at the front, tucked into the belt. This detail in garment has been also attested from Phrygian Kybele

^ özgen, 1988: 32-49,187-195 (especially with cat. no:29-31,40,43,44,47,60-62). The fragmentary grave goods included, electron pin heads in the shape of duck and lion, an iron tripod, 46 arrow h<>a<1s, two spear heads, an iron dagger with a golden handle, bronze protoms in the frm of griffon hpflds which would have been attached to cauldrons, ompholos bowl fragments, cups with swiveling handles and a ceremonial standard with phallus shaped knobs arranged in groups of three. 218 grams of melted gold, silver and electron appeared after the sieving of the dd)ris. Dörtlü, 1988:173-174. The dimensions of the tumulus was 40x45 meters with a peak height of 5.10 meters. A wooden chamber of 3.25x4.50 meters had been constructed in the pit. ^ Özgen, 1988: 33-49, 186-195 (especiaUy with cat. no. 32-39,41,42,45,46,48-59). SUver belts (one of them being worn by the deceased), a small silver cauldron with Phrygian graffito, omphalos bowls of silver and bronze, bowls with swiveling handles, ladles of bronze and silver were located from the tomb. Furniture inlays of ivory, a pyxis lid of ivory and horse gears of a horse bit and breastplates were also found in the tomb. ^ Ibid, 27. Ibid, 38-39, Cat Inv. no: 41,42. Ibid, 39, 190 no.42 . 79

figurines.^’^ The ivory figurines have been suggested to show closer afiBliation with

Western Anatolian works and especially Lydian craftsmanship. Parallels have been

shown as the priest figurines coming from the temple deposit of the Ephesus

Artemision.^’^

Both the architectural style and the contents of the tombs have dated them to the

eighth and seventh centuries but more likely to the seventh century b.c., due to

certain strong Lydian elements. The lavish usage of silver, gold and electron artifacts

also suggests a Lydian affiliation. Other than the figurines, jewelry including acorn

pendants, boat shaped earings, and animal figurines show parallels with the LFşak-

Güre material^'^

Similar burials have been excavated in the Karaburun vicinity.

(c) Karaburun The stone fill of the burial was partially lost with the

activities of the tomb robbers and erosion. The burial includes a two handled dinos

like vessel, placed in a pit lined with stones. The burial goods had been looted since

the Roman period. Two layers of artificial fill covered the boulders, which would

cover the tomb pit. The two handled dinos like urn (Plate 45 A), was complete with a

height reaching to 25 cm. and max. diameter of 36 cm. It represented a good example

of the bichrome ware of the South-western Anatolian Iron Age, as discussed above.^”

(d) Karaburun IV appears to have been totally dismantled and eroded of its

superstructure. The original boulders that covered the burial still exist near the

312 Ibid, 33. 313 Ibid, 33. 314 Ibid, 33. Özgen and öztürk, 19%: 27. Mellink, 1972: 261, plate 57, figs 1 1 - 12 . The site of Karaburun is located on a ridge between the village of Bayındır and Semayük-Karataş sites, to the west of the Elmal i-Korkuteli road (see map 1). Upto 30-40 tumuli have been located during the Bryn Mawr expeditions. See pg. 76. Mellink 1972a: 261-263, plate 57, fig. 12. The âbric is defined as an orange buff with a reddish slip. The upper part of the body was decorated in a matt dark color, which consisted of panels that were set between the handles on each side. Motiis consisted of wavy lines betwen horizontal straight lines, and a continuous band below the rim. 80

site. The estimated diameter of the tumulus is 20 meters. The traces of two pits, with a

mass of burned ash and cremated bones, still exist. Three fibulae of the Blinkenberg

XU type came fi'om one of the pits. An iron tweezer and whetstone with bronze

attachments were collected fi'om the pit to the north. The fibulae very much

resemble the B a y ın d ır D fibulae, though different dates have been given to the

tombs.^'® Both Karaburun III and IV have been dated to the sixth century b.c., by the

excavator.

2.3.2 The Painted Chamber Tomb of K iz ilb e l 321

The chamber tomb, which was covered with a tumulus, was described by the

excavator as follows (Plate 47): The K izilbel tumulus consists of a chamber tomb in

the top of a natural ridge. The rectangular chamber (2.45 X 2.00 meters with

maximum interior height of 2.30 meters) is of limestone construction, which is rough

on the outside, and shows a refined workmanship of „pseudo-polygonal“ technique

on the interior. The gabled roof is constructed of four long slabs laid parallel to the

ridge. The entrance was through a small doorway (1.00X0.80 meter) with a

„portcullis doorstop“. Retaining walls on both side of the doorway were built to hold

the fill of the tumulus (Plate 47 B). No dromos existed; after the inner chamber was

completed and the deceased placed in the chamber, the tumulus was covered and

filled on all sides.

The furmture of the chamber consisted of a stone kline which was smashed into

pieces. The tomb was possibly robbed in antiquity. It was completely emptied of its

contents. A few metal artifacts, some pottery, the rim of a limestone alabastron, a few

Ibid, 262-263, plate 56, fig. 8-10. Ibid 263. as stated by Melliok, this type of fibula was in popular use between the eighth and sixth centuries B.C.. Muscarella, 1967: 16-20, plate XII fig 7 14 Mellink, 1972a: 263. Kizilbel is located three km. to the southwest of Elmalı town, on a ridge overlooking the southern extention of the plain and the recently drained Karagöl (see Map 1 for location). Mellink, 1970a. 251, plate 59, fig. 25; 1969b: 141. Also see the chapter on the architecture of the tomb by R. A Bridges in the recently published book of Mellink, 1998: 7 -2 1 . 81

slivers of a Lydion, the neck fragment of a red-glazed Lydion, a bone fragment with

incision, and some skeletal remains of the deceased were recovered from the tomb.

The bones belonged to a male in his forties, medium tall and who had suffered from a

knee injury.

The inner walls of the chamber and ceiling slabs were covered with white lime

plaster, then painted.^^’ The ceiling was painted with a checkboard design filled with

floral patterns. The floors were partially decorated to mutate a rug.^^*The walls were

painted from the floor to the ceiling with superimposed registers of uneven height,

each displa)nng different subject matters.

The subject matter of the friezes shows a choice of Glreek subjects. They can be grouped as mythological (Plate ceremonial (Plate 48,49)^^® and every day life

scenes (Plate 50).^^® Subject matter of the paintings, their style and iconography has

Mellink, 1970: 251; Mellink, 1998: 4. Angel, 1973: 304; Mellink, 1972b: 128. Mellink, 1976: 27; 1971a: 248. The colors used for the paintings were different shades of red, light blue, black and white, with preliminary contours in r ^ and final contours in black. The skin of the males were colored in red, and white was used to color the skin of women. Mellink, 1976: 27. M ellink, 1970a: 252, plate 61, figs. 29-30; 1976: 28-29, fig.4; 1998: 57-58. The mythological scenes include the scenes of the Gorgons, Medusa, birth of Pegasus and Chrysaor and a possible Troilos scene, the choice of subjects being popular Anatolian themes, especially Pegasus, associated with the Bellerophon myth in Lycia, and Chiysaor with , and a Medusa-Gorgon scene with Perseus has been depict^ in the heroon of Pericle in Limyra. Mellink, 1998: 57. On the north wall of the tomb is a series of scenes arranged in boustrophedon; a processional scene of a series of attendants, horses, servants, Ixxfy guards with a central enthroned dignity has been suggested to depict events firom local or Lycian myth, unknown to us. Mellink, 1998: 59. Mellink, 1970a: 252, plate 60, figs. 27-28; Mellink, 1974b: 544, plate 165; 1976: 30 fig. 5; 1998: 59-61. The ceremonial scenes include the scene of ‘the departure of the hero’, which is displayed right over the kline.The scene continuing into a funeral banquet scene (a similar scene has been attested in Karaburun n, to be discussed on pg. 84, see footnote 345). As suggested by M ellink, the depiction of fimeral banquet scenes were more a ritual need than an artistic expression (1998: 60). A winged female figure holding a lotus-palmette, in the oriental style, has been depicted over the chariot scene (Plate 49 A). Mellink suggests a meaning of good omen; which winged sun discs occur over Neo-Hittite chariots, and an animated winged disc as a female figure holding lotus blossoms is evident on a Nimrud ivory (1998 : 59). The winged female figure has not been encountered either in Greek or East Greek, nor Southeastern Anatolian traditions. The scene also recalls the popular theme of the sirens, known from the funerary monuments of Xanthos, especially of the Harpies Tomb, the half bird half female figures (1996 : 30). Mellink, 1970a: 252-53, plate 59, fig. 26; 1971a: 248, plates 51-52. Still another category of paintings display various hunting scenes, which are more or less cormected to the mimdane world of the deceased, with details from the everyday life. A lion hunt has been displayed in the lintel of the southern wall. The archers in Scythian costumes have been suggested to follow East Greek 82

been suggested to show strong East Greek/ Western Anatolian influences. Anatolian

features are the unevenly superimposed registers^^, hunting scenes, and the lively

depiction of everyday life scenes that directly comes fi^om the mundane life of its

inhabitants. Certain orientalizing details are the lotus and palmette motifs, and the

usage of four color scheme, with the absence of the yellow and green, which

resembles Assyrian paintings. ™ Costumes of warriors and women, their facial

characteristics, and equipment point to an execution of the figures in the East Greek

tradition, to represent the works of a group of artists who must have worked in Lycia

and the neighbouring regions, experienced in drawing Greek conventions, and

following regional pictorial traditions.”^ The Kizilbel tomb has been dated on

styhstical grounds to the last quarter of the sixth century b.c. when Lycia had come

under Persian control, by general Harpagos the Mede in ca. 540 b c ..” “ But Persian

influence, as stated by Mellink” *, has not yet modified the Lycian and Phrygian

characteristics, which are to become apparent on the painted chamber tomb that will

be discussed next.

cxtnventions, whereas the depiction of the lion, and a group of running boys, acting as beaters, display local traits. A second hunting scene is a boar hunt in the marshes, where a group of boys appear on a boat, possibly describing the immediate environment of the deceased, as it would have appeared with the lakes and the marshland on the plain. The hunting scene in the tnar<;h

2.3.3 The Painted Chamber Tomb of Karaburun IP136

The tomb chamber (Plate 51 A)”\ covered under the tumulus mound, was made up of large quarried blocks of limestone, that were left unworked outside and worked on the interior, as in the Kizilbel example. The gabled roof was formed from two enormously large slabs, and was tightly fitted with reddish mortar. The roof was packed with a number of large stones which had been partially removed by tomb robbers in antiquity. The tumulus fill would have covered the whole structure. A peculiar architectural feature observed in the roof construction of Karaburun, suggested that Karaburun II was a stone imitation of the Phrygian wooden proto­ types.”® A pseudo-doorway existed along the eastern wall of the chamber, a limestone kline was set against the western wall, with a small table at the foot of the kline, which was destroyed. The kline had a raised edge and comer posts. The floor was paved with square blocks. The walls, kline, service table, and the floor blocks were decorated with paintings.”® Stone built chambers of tumulus burials are paralleled in the Bintepe necropolis of Lydian , and in the U?ak-Gtire region.^341

The organization and execution of the painted friezes differs from the Kızı Ibel superimposed fiiezes, in the manner that the Karaburun paintings have been executed in one continuous frieze, over a plain dado. The color scheme is more varied than

Karaburun n is located between the Karaburun m and I tombs. The rectangular tomb chamber had the dimentions of 3.00x2.61 meters, with a gabled height of 2.66 meter. Mellink, 1971a: 251, the interference of the Roman tomb robbers has been also indicated by the existence of a gnall Roman lamp fiagment in the chamber. Mellink, 1971a: 250-251. A central interior gable block was used in the roof construction, which differed from the Kızı Ibel tomb chamber. This was a triangular block of 33 cm. thickness, which fitted against the ceiling and rested in the sidewalls with horizontal tenons. This architectural feature is also o b serv ed in the wooden chamber of the Great Tumulus of Gordion. Young, 1958:148. Mellink, 1971:251. özgen and öztürk, 19%: 28-30. 84

with the K izilb el paintings.^^ Precise lines, especially in depicting the hands, feet, and drapery patterns shows a refine workmanship and conventions can be attributed to Greek influence,^“^ but the themes and subject matters show a strong Persian-

Anatolian influence.^

The tnain scene shows the tomb owner, who possibly was a local landlord, depicted with strong persianizing elements (Plate 52). The subject of the fiieze possibly indicates the preparations that would have taken place at a banquet or

symposium. The metal vessels depicted in the scene show strong Phrygian-

Achaemenid influence. Although the chamber has no remains of tomb gifts, due to being robbed in antiquity, the details of the tomb paintings give us an idea of the tomb

content. The Lydian tombs that have been illegaUy dug in the U?ak-Gure region

display a rich burial inventory, which has been recently returned to Turkish authorities by the Metropolitan Museum of Art. Some of the artifacts from the collection show

close parallels to the vessels depicted on the walls of the Karaburun II Tomb., 347

Two further scenes depicted on the walls of the Karaburun II Tomb are, one, a

battle scene between the victorious army of the deceased and the Greek soldiers.

Mellink, 1998:43,44. Solid colors are red, blue, white black, blends with the additional usage of green and purple color. Evidence for the real usage of fiesco technique has been also encountered. Mellink, 1974b: 545. Mellink, 1974b: 545. Mellink points out to the fact that some figural drawings have approached the Classical Greek features. ^ Ibid, 545. Mellink, 1971a: 252, plate 55, fig. 24-25; 1972a: 265, plate 58, fig. 16,18; 1974a: 545-546, plate 168,169. The reclining dignitary on a couch has been depicted wearing a robe with rosette boarders over which he wears a purple edged green himation with red and blue boarders.His jewelry consists of a blue and red ch ed ^ diadem, a lion head bracelet, earings, and a possible seal which is worn around his neck with only the strings showing underneath his garment. He is holding â phiale in his hand in a welcomming gesture, feeing a row of servants who are approaching him. The first servant, depicted in Persian costumes, with a long-sleeved tightly fitting belted robe, carries in his hands, a towel and a fen, followed by a second servant, in similar costumes, carrying a phiale, a two handled goblet with griffon finials and a highly decorated ladle hanging from his little finger. ^IW d., 254. özgen and öztûric, 19%: 47. Especially see cat nos. 24-31,107 for ladles, cat. nos. 35-42,122- 23 for phiaifts, cat nos. 75,77,86,228 for alabastra and 66 for a goblet 85

occupying the northern wall, (Plate 54) ^ the other is a funeral procession scene, on the opposite, south wall (Plate 53).^349

Stylistically the Karaburun II tomb has Greco-Persian, and Anatolian traits. As stated by Mellink^’®, the artists were possibly Anatolians who have learned from

Greeks; their hand-writing follows the Greek conventions, and in the more developed style of the Early Classical period. The subject matter is official and representative.

The owner of the tomb, possibly a local landlord, under the favourable patronage and guidance of the Persians, has strengthened his power and wealth in the region of the

Elmalx Plain.^*’ The landlord has chosen to be depicted as an Anatolian-

Achaemenid local ruler, affiliating himself with rites that would have taken place in the

Achaemenid court.^” His powerful position and wealth is conveyed in the scenes of

^ Mellink, 1972a: 267-68,plaie 59, figs. 21-22, plate 60, figs. 23-24. The battle scene, displayed on the northern wall, takes place between the foot soldiers of the dignitary and Greek hopUtes and archers The dignitary has been depicted in the centre of the frieze, on horseback, killing the leader of the enemy army. The costume and weaponry of the enemy suggest that they were Greeks, and the leader who is being killed by the spear of the dignitary, has been depicted wearing a short-sleeved blue tunic under a white corselet, and with a cuirassed helmet, which distinguishes him from the other enemy soldiers. The attire of the dignitary is quite un-Greek, with a purple long-sleeved tunic worn over purple trousers, which are tucked into his ankle-high blue shoes. The garment of the horseman has been to be strictly Anatolian, as is the attire of his soldiers. Their weapons are partly Greek (shields), partly Anatolian-Achaemenid (daggers), and the two-edged javilins are a possible local type. Mellink, 1972b: 127; 1973a: 298-299, plates 45-46. The oj^ site wall displays a contrastmg subject matpr to the batfie scene, where a (^ te and dignified f^eral procession is taking place. The dignitary, again depicted in a central position, is riding in a throne-chariot, accompanied by his aftAnHants Two servants follow his chariot, who are followed by a red domed chart. Two servants carrying furniture follow the domed cart The furniture, a piece of table, will be possibly placed in the tomb chamber, accompanying the deceased, together with other tomb gjfis. His black horse and a riderless horse lead the firont of the procession. In the throne chariot the dignitary is shown in ceremonial attire, wearing a purple robe, over which a white furlined mantle (kandys) with empty sleeves is draped He wears a cap with long lappets (başlık), that covers his head His right hand is raised in an artificial manner, and he is not holding the reins of the chariot The chariot is being directed by a servant behind it Mellink refers to the scene not as a symbolic depiction of the as he would have appeared during his life time, but as a realistic conveying of a local rite, where the deceased would be shown to his public for a last time, as he would have ai^)eared during his life time. Mellink, 1974b: 546-547; 1971a: 254. We should also not outlaw the possibility that he may be a Persian Ity race. (Keen, 1998:108) Trying to label an ethnic identity to the tomb owner would be dangerous at this point, unless we do not have the scientific evidence such as DNA analysis of the bones of the tomb owner. Ibid It has been discussed by Mellink that the the reclinig posture and the garments of the deceased are related to a Greco-Anatolian fashion, whereas the bearded features and the jewelry of the deceased show strong Persian elements, especially the diadem, which would be the type that would be worn by Persian nobles. The depicted rite, with the servants carrying attributes eg. &n. 86

the wall paintings, especially with the banquet and funeral scenes. The landlord has possibly fought for the Persian army, against the Greeks, as it is displayed in the battle field scene. This suggestion fits well into the historical context of Anatolia and

Southwest Anatolia, as the presence of Persians has started to be felt strongly in

Lycia, especially after their defeat of (494 b.c.), and the Greek expedition of

Xerxes (ca. 480 b.c.), for which Herodotus gives accounts of the Lycians and Milyans,

who had fought on the side of the Persian army.^^' The tomb has been dated on

stylistical grounds to 475 b.c.

Karabunin I Tomb The burial consisted of a simple, gabled roofed

sarchophagus built of limestone. (Plate 51 C) The sarchophagus had a northeast-

southwest orientation and was covered with pUes of rock. A few skeletal remains

of the deceased ware found, belonging to a female in her 30s or near 40s, and the

teeth suggested a good childhood diet, and health.^'® A group of pottery, smashed into

pieces by grave robbers, included an Attic black-glazed stemless kylix, a black figure

lekythos, with a frieze of dancing figures; a dark glazed trefoil oinochoe in East Greek

style with a reserved zone of ray motifs on its shoulder; and a fine ware in imitation

of the Achaemenian metal vessels, with red glaze, gray grooves, and horizontal flutes

on the body. A number of alabastron fragments were also located. The pottery finds

357 of the burial have dated it to ca.460 b.c..

towels and various vessels, recalls the Persepolis examples, and the type of rites that would have taken place in the Achaemenid court. Mellink, 1971a: 254. Herodotus (Book VII. 76,77), counts the Lycian and Milyan soldiers among the Xerxes army as he is getting prepared for his campaign against Greece. He also gives a description of the equipment of the Milyans a s; ‘The Milyans were carrying short spears, and were factpning their garments with brooches; and among them, some were armed with Lycian bows and wore helmets made up of copperAeather?’. Herodotus (Book IQ. 90) also gives the Lycians and Milyans among the tribute bearers to the Persian King Darius 1, among the peoples of Asia, with the lonians, who were subjected to pay 400 talents of silver. The built tumulus was located only 50 meters east to Karaburun Q. The proximity of the tombs of Karaburun 1 and Q has led the excavators to assume that they were related to each other. The tumulus was opened in the antiquity and also by recent grave robbers. Mellink, 1971a: 250, plate 53, fig. 17. The tumulus had a diameter of 35 meters and a preserved height of 2.50 meters. The dimensions of the sarcophagus were 2.18x1.03x0.94 meters (its box), and 2.50x1.40x0.48 meter (its lid). Angel, 1973: 304. Mellink, 1971a: 250. 87

Two further chamber tombs in the Elmalı Plain were investigated, and one of them partly excavated in conjunction with the excavations. These are the tomb chambers of Mügren (Plate 55)^“ and Boztepe.^359

To summarize, our information regarding the nature of Iron Age settlements is scanty, and depends mainly on the incidental recoveries at the Chalcolithic site of

Bağbaş i , and the Bronze Age site of Karat aş . The pottery evidence remains to be the Southwestern Anatolian Iron Age pottery of bichrome ware, and the so-called black-on-red ware of Mellaart, both dated within the eighth and sixth centuries b .c ...

The local usage o f inhumation in pithos burials is stiU prevailing, its predecessors going back well into the second and third millenium b .c .. They more or less represent the burial style of the common inhabitants. The tumulus burials of Bayındır,

Karaburun and K izilbel point to the cultural contact of the plain with its northern neighbours, the Phrygian highlands, as well as Western Anatolia. When we look at the tumulus burials of the Elmalı Plain, we can attest the general characteristics which reflect the similar origins of both Phrygian and Lydian tomb tradition.^®’ Besides, these tombs also point to a well established upper class in the first half of the first miUenium

Ibid. 249. Located to the north of Semaytik, the Müğren tomb is a gabled chamber, which has been completely exposed, and its tumulus completely gone. The interior walls are well smoothed, with traces of ancient paintings. The chamber was used as a Byzantine chapel, for which it was remodelled and painted and later it was used as a shelter for shepherds. With its architectural style it has been well dated into the 6th and early 5th century Mellink, 1973a; 2%-297; 1973b: 126-127, fig. 3. The Boztepe Tumulus is located eight km. to the northeast of the Elmalı Town. The tumulus and the stone chamber have been considerably disturbed, both by the ancient and modem activities, and has been also used for its building stones. The remains suggest a tumulus burial with ca. SO meters, and a preserved height of four meters. The rectangular built chamber, had the dimensions of 2.75x3.50 meters. The fragmentary remains of a wall block, which still preserved a cutting for a gable, suggested that the chamber had the usual gabled roof Different from the Kizilbel and Kar^xmm II tombs, the chamber of the Boztepe tomb was constructed of ashlar masonry, which was well smoothed in the interior, possibly to receive some kind of a decoration. Only one block showed traces of plaster and a patch of Ûue stucco, which was comparable more to the floor plaster of Karaburun n.The usage of ashlar masonry has dated the tomb to a later period than the Karaburun and Kizilbel tombs, but still in the first half of the fifth century B.C.. ^ Mellink, 1976: 24-25. The Bayındır Tumuli, with their wooden chambers sunken into a pit, and covered wifii tumulus fill of clay, pebbles and stones point to Phyrgian origins, as discussed. The Chamber Tumuli of K izilb el and Karaburun n are Lydian in the sense that their chambers are built of stone. However their gabled rooms, with no finishing on the exterior, and being covered with rode piles with no dromos provided, suggests Phrygian origins. 88

B.C., whose their wealth and existence should have depended on weU established

populations, and the utilization of the rich resources of the plain. As stated by M. J.

Mellink“ ’; „...Upland plains of Elmalı must have been a prosperous but somewhat

peripheral part of Lycia, culturally not as Lycian in classical times as the coastal zone,

but at least as closely connected with the coast as with the South -Pisidian

Plateau .The original residents of the Elmalı area may have been among those

called Milyans....“

The geographical limits of the Milyan territory have not yet been strictly identified.

The ethnic differentiation of these people has been attested from a number of classical

sources. Classical sources fi"om as early as Herodotus refer to the Milyans as a group

of people, ethnically distinguished fi-om the Lycians, Pisidians, Kibyrians and

Kabalians, which appear to be geographically in the neighbourhood of the Milyan

territory.^“

Ibid, 21. The idea has been also presented by Coulton, 1993: 81, where he compares the archaeological evidence of the first millenium, with coastal Lycia, Elmal i Plain and the Kibyratis, the upland regions to the west of the Elmal i Plain. No parallels to the rich burial tradition attested in the Elmalı Plain, in the first half of the first millenium appear fiom coastal Lycia (though coastal Lycia displsQ« a rich burial tradition, but different than the uplands), or the neighbouring highland regions of the Elmal i Plain, confimung the theory that, ethnically, a different group o f people inhabited the Elmalı Plain. ^ Hall 1986b: 137-157, in his article Hall discusses the literary evidence regarding the Milyans fiDm various periods (especially see pg. 142-151). In the light of a new discovery he extends the area of Milyans up north, till lake Askania (modem Burdur), also see fig. 1 . Three large rectangular blocks, with a dedicatory inscription which refers to the dedications of a group of Milyadeis and Roman merchants, dated to 5/4 B.C., to the reign of Augustus were located in the vicinity o f Burdur. One of the blocks has been located, used as spolia, in an Ottoman bridge, in the village of Boğaziçi, south of Burdur, fiom where the other two are also sui^iosed to come. Blocks A and B are in the Burdur Museum. A list of references to the Milyans and their territory fiom the classical sources is given in Appendix 3. These references, for the time being, aim to give an inright into the geographical location of the territory of these people, who have been ethnically distinguished fiom their neighbours as early as the fifth century B.C. 89

2.4 Classical evidence from the plain

The classical evidence from the fifth century into the third century b.c. on the

Elmalı Plain is scanty. When compared with the classical evidence from the coastal

Lycia, the Elmalı Plain appears to remain outside the immediate sphere of Lycian cultural influence,^^ although there is evidence for an increase in contact with the coastal regions. These ties, as suggested by Mellink, were possibly formed by the seasonal migrations that take place between the coastal zones and higher uplands, as also attested today.^ Growing contact with coastal Lycia between the fifth and third centuries can be observed with the occurence of Lycian type rock-cut tombs, and the recovery of the two coin hoards from the Elmalı Plain. The predominant coins are of the Lycian type; the so-called Elmalı hoard is dated to 460 b.c., and the Podalia hoard to 380-360 b.c.“ * The growing Persian influence of the coastal zones, which starts with the invasion of Harpagus, is also attested with the Karaburun II tomb, as discussed above. Two fiuther signs of Persian influence into the region are attested with the Yalnizdam Stela^“ and the so-called Pericles Inscription from K izilca^*\

Pottery evidence for the Classical period, except for an imported Attic type, is difficult to distinguish from the local wares of the Iron Age^®. Attic imports have been found in the Hacimusalar mound excavation. So far, excavations that have taken place in the neighbourhood have the tendency to publish the most significant finds, and have not concentrated on or published the local and more common ceramic groups according to their stratigraphical sequences. This has been a disadvantage for finding comparanda material for the ceramics from the Hacimusalar excavation.

^ Mellink, 1976: 21; Kjeldsen and Zahle 1975b: 315-318; Zahle, 1980: 37-47, fig. 1; in their articles they draw the borders of Classical Lycia, on evidence of the occurrence of Lycian type of burials and Lycian inscriptions; the Elm alı Plain remains outside this zone. ^ Mellink, 1976: 21. Coulton, 1993: 82. ^ Mellink, 1973a: 303; 1974a: 209. ^ Bean, 1971: 22; Neumann, 1979: 36, no. N 314. ^ Information conveyed by Mellink to Bean, stated in Bean, 1968: 11, see footnote 11. 90

whose future studies promise to have a pioneering role in understanding the pottery sequence of the region.

2.4.1 Sepulchral evidence

The Lycian type of rock cut tombs from the E lm a lı Plain have been mentioned in the various accounts of the 20th century European travellers. A more recent survey of the Lycian type of tombs, by K. Kjeldson and J. Zahle, has given mainly five areas, and ca. 20 examples of the rock-cut tombs in the region, located near the southern passages of the plain which leads into lower Lycia, and along the western edge of the plain. The localities of the attested rock-cut tombs can be listed as (from east to west respectively), the vicinity of Avian Lake (also refered as Buralya)“ ®, Armutlu

Islamlar (Plate 56 A)^^', K ızılca (Plate 58 A), and near the Eskihisar pass (Plate 57)

According to information given by S.Ozaner, the rock cut tombs seem to follow the lines of the limestone bases of the mountains which surround the plain. The rock

The rock-cut tombs at the Avian Lake site had been noted in the travellers' accounts and by Bean. Described as less than halfway up the north-east slope of the citadel are two small rock-cut tombs, but they are not of the Lycian ^pe. They are not elaborately decorated, and because of their unpretentious style can be attributed to common citizens. Bean, 1978:154. Hosityn, 1842: 153-54; Spratt and Forbes, 1847: 290, had identified the site with the ancient site of Podalia; Peterson and von Luschan, 1889:161, had favored the identification of the site with the ancient city of Chôma. The rock tombs which are located in the vicinity of Armutlu, have also been refered to by various travellers. Three rock-cut tombs of Lycian type have been noted on the calcerous slopes of the hill which lies to the south of Armutlu; however, there are more tombs in this region that need a thorough study. A group of rock-cut tombs, among which some represent the most elaborately decorated examples found on the Elmalı plain, appear near the village of Islamlar. The village of Islamlar is situated in the southwestern part of the Elmal i Plain, ca. 12 km. southwest of Hacimusalar Mound. The site of the rock-cut tombs and monuments are located one km. west of the village, on a hilltop, known as the Aytaş Mevkii. Some of the rock-cut fecades are accompanied by inscriptions, which them to the later periods (to be discussed under the Roman era). At the very top of the hillside a mfin^ftnpntal building, built of well masoned blocks exists. The building has been referred to as a heroon. The site has a dominating view over the southern part of the plain, the Hacimusalar Mound and the passage at whose entrance lies the village of Islamlar. The passage is one of the southern pa«ages of the plain, which leads over the towns of Akçay and (jömbe and the Sinekçibel and connects the plain fiuther south to central and coastal Lycia. The Lycian type of rock-cut tombs bad been first encountered by Mellink, and later by Zahle and the Hacimusalar team. A rock- cut tomb and a sarcophagus, with sculptural reliefs, have been published by J. Zahle, who dates them to the fourth century B.C. based on their reliefs. Zahle, 1979: 343-345. Kjeldson and Zahle, 1975b: 315,317. 91

cut chamber tombs of Lycia have been dated in general within the period of the fifth till third century b.c., and extending into the later periods.

2.4.2 Sculptural evidence

The rock-cut tomb of Islamlar consists of a two tiered rock facade, which depicts a banquet scene. The deceased is in a reclining position, and accompanied by a seated female figure, and two further figures, depicted in smaller size (Plate 56 The second evidence dated to the fourth century B.C., from Islamlar, is a sarcophagus socle that has reliefs along its three sides. The longer side exhibits a seated male figure with his mantle, and his right hand reaching out to a tropaion. To the left of the scene a rider in full gallop who is overriding a fallen enemy and his horse has been depicted.The two shorter sides exhibit combat scenes with wild animals. One of the scenes depicts the fight between a horseman and a bear. The second scene shows the piercing of a boar by a speared male figure. 375

Another stele was located in Yalnizdam, 8 km. To the northwest of Elmali.^^®

The limestone tomb stele (two meters high), which was found in two pieces, possibly functioned as a commemorative funeral marker near one of the tumuli or built tombs.

It had reliefs on both sides, dated into the early fourth centuiy b.c., within the Greco-

Persian style. The relief depicts a triumphant horseman over the fallen victim (Plate 54

349. Zahle, 1979, Kat. no.68, abb.68. Ibid, Kat. no.69, abb.69. Mellink, 1972a: 209, plate 60, figs. 25-26; 1973a: 303, the stele is in the Antalya Museum. 92

C, The other side of the relief shows a bearded man in a frontal position,

leaning on his staff and raising his right hand. A small boy stands by his side 378

2.4.3 Epigraphic evidence

The two rock cut tombs of Kızılca have been first encountered by Bean, who

mentions them briefly with the so-called Pericles inscription (Plate 58).^’® The higher

one to the left has no inscription. The lower one, which has the impressions of wooden beams depicted on its facade, has two groups of inscriptions, one on the rock below

and the other to the right These inscriptions appear to be the only examples of the

epichoric Lycian inscriptions found in northern Lycia."*° The epichoric inscriptions

have been translated by Neumann as; Part a; „This tomb, has been built by M., the son

of M., during the period of Pericles“. Part b: „ Whoever will damage the tomb, shall be hit by With the reference given to the Lycian Dynast Pericles, the tomb has been dated into his reign as ca. 360 b c..

"377 Mellink, 1972a: 209, pate 60, figs. 25-26. The relief has been described as follows; „... The Yalnizdam warrior wears a long-sleeved tunic, trousers, cuirass, and cloak knotted over his chest. The ribbons of his Persian cap and the cloak flutter in the breeze. The victim has lost his balance under the impact of the horse; he is falling on his shield and cloak, naked now except for his belt, ffis right bflnd is raised to push back the spear of his conqueror....“ Similar to the theme of the Karaburun 11 paintings, we have a victorious warrior (possibly Persian) in oriental attire, smashing his Greek enemy (Plate 54 A B). Mellink, 1973a: 303. Bean, 1971: 22-23. The rock-cut tombs are located 40 minutes walk above the village of Kızılca, to its west Other than the publication of the inscriptions by Bean, they also appear in Neumann, 1979: 36- 38 no N 314, where a revised publication was made upon squeezes taken by Borchhardt. 381 Neumann, 1979: 38. Bryce, 1986: 110-112; Child, 1986: 73-75; RE Suppl. ХШ ПА-215. At the beginning of the fourth century B.C. the Dynast Pericles, installed in Limyra, ruling in the eastern part of Lycia, is known to have expanded his territory towards the west, against the pro-Persian ruling of the country. His of the last Persian d|ynast of Xanthos, Artumpara, was noted in an inscription found in Limyra (TAM 1 ,104). This is followed by his capture of Telmessos (recorded in the fourth c. B.C. accoinits of Theopompos). Pericles is also accounted to have taken part in the revolt of Satraps, agüiıiirt Artaxerxes П, in 366 B.C.. With his defeat, the Lycian Dynasts come to an abrupt end, and Lycia is annexed to the rule of the Carian S a t^ Mausolos (Diodorus XV 90.3). As it has been noted by Mellink, 1976: 23, the modern K ız ılc a is the ‘yayla’ of the inhabitants of modem Finike, and the relationship of the area with Limyra can be possible confirmed witli the inscription 93

Both the Kizxlca inscription and the Yalnizdam stela have been taken as evidence by Borchhardt for the expansion of the Lycian Dynast Pericles into the

Elmalı Plain, in the early fourth century b.c. as a part of his expansionist policy in

Lycia.“^

2.4.4 Numismatic evidence

Two coin hoards, which had been recovered in illicit digs, have been identified to have come from the Elmalı Plain. Both of the coin hoards point to the possible interest and expansionist attempts of Lycian d}masts into the plain.

The ‘Decadrachm hoard’^®'* from Elm alı^*^ consists of 1700 silver coins, about

970 of which belong to Lycian Dynasts, and the rest from Athens, Northern Greece, the Aegean, and Asia Minor.^** It has been stated by Kagan, that the coins of the various groups had arrived from treasuries or a single treasury rather than from general circulation.“ ^ The hoard has been dated to a time between 480-460 b.c. The dominating group of coins belongs to Lycia, and to the Dynast Kuprlli of Xanthos, with a larger number of coins giving evidence for new dynasts’ names.“ ® The appearance of the coin hoard from the Elmalı Plain can be taken as evidence for increasing contact with southern Lycia, and the expansion of the Lycian influence into

389 these regions, as early as the beginning of the fifth century b.c..®

Borchhardt, 1976: 25 ; also see footnote 41. “ 4 carradice, 1987; especially the articles by Fried :l-20; Kagan: 21-28; Spier: 29-42. And especially see Table 1, pg. 9. The hoard which has been illicitly recovered and shi]q)ed abroad has been recently returned to TuAey. The find place of the hoard being unknown, is assumed to have arrived from somewhere northeast of Elm alı, in a pot. The hoard proves to be especially important for numismatics, since it has doubled the known number of Athenian decadrahms. Keen, 1998 107-108. Spier, 1987: 1, plates 1-5. Kagan, 1987:21. The coins seem to have no sign of wear, and are in near-mint condition. The hoard has been suggested by Keen to be an „emergency“ hoard, which had been collected and buried in a short time. Keen, 1998:109. “ * Spier, 1987: 29. Keen, 1998:111. It has been suggested that the large number of coins belonging to the Xanthian (fynast Kuprlli, may point out to the expansion of the Xanthian dtynas^ into the Elmalı Plain, especially after the battle of Eurymedon (470-465 B.C.), when Lycia was went over to Athens. The 94

The second hoard, which comes from the vicinity o f Avian Lake, was recovered in

1957. The hoard was reported to have been found in a pot, with about 1600 coins.^®®

All o f the coins were of silver mint, with 488 of them being Lycian and 242 from

Aspendus (among the Aspendus coins there were 40 Persian sigloi and 202 staters).

The Lycian coins were found to be silver staters, thirds and eights. They belonged to a number of Lycian Dynasts, dating the hoard between 390-360 b.c.^*’ One explanation suggested for the occurrence of the hoard near Avian Lake, where the ruins of a fortified stronghold exists at a strategically important pass"**^, is that it possibly had a military purpose - payment for a mercenary force?- and needed to be burned in a hurry.

The evidence from the Classical period for the Elmalı Plain remains scanty. The

Elmalı Plain does not appear to be a part o f‘Classical’ Lycia, as the evidence from the Lycian type of rock-cut tombs and inscriptions are very few in number when compared with the coastal regions of Lycia. But there is a growing infiltration of

Lycian influence into the plain, especially under the Persian dominance o f the region.

Elmalı Plain was still potentially in the Milyan territory, during the Classical periods, but also subjected to the expansionist interest of the Lycian dynasts, who had flourished under the patronage of the Persians. This is also evident in the accounts of

Arrian, at the eve of the Hellenistic era, where he gives the march of Alexander the

Great into the territory of Milyas, which actually belonged to Greater Phrygia, but was handed over to Lycia on the orders of the Persian kings( Appendix C2).^^

Elmalı Plain would have served as a buffer zone between the Achaemenid and Athenian forces, and its pro-Persian tendency is well attested from the paintings of Karaburun n tomb (pg. 84). CMc^ and Morkholm, 1971:1-29. The coins, recovered by an illicit dig, immidiately started to be distributed. Turkish authorities were able to recover 510 coins from the hoard, which are stored in the Istanbul Museum today. Ibid., 1; the names and the frequencies of the recovered coins of the (fynasts have been given in a list, which includes the Lycian Dynasts Vekhssere, Zagaba, Aruvatijesi, Mitbrapata, Trbbenimi,Vedevie,and Pericles. Bean, 1978:153-156; though the earliest pottery evidence noted from the citadel shows Hellenistic occupation, with Early Bronze Age pottery appearing on its slopes, and from the plain. Coulton, 1993: 82; Keen, 1998: 106-107. Arr. Anab. I. 24. 95

2.5 THE HELLENISTIC PERIOD AND THE EARLY PRINCIPATE

(334ac.-A-D.43)

Our knowledge for the Hellenistic period of the Elmalı Plain also remains very

scanty. Apart from the remains of some Hellenistic hillforts, from the surrounding

terrain of the plain, virtually no evidence of settlements exists, except for very scanty

remains of ruins, epigraphy, and coinage. With the Hellenistic period we learn more

about the history of Lycia through literary sources. Northern Lycia’s history may have

followed closely the histoiy of southern Lycia, in its general contours. In this section

archaeological evidence from the plain will be presented after a general history of

Lycia, which will aim to give an insight into the historical context of the region.

Alexander the Great arrived in Lycia foUovring his victory at the Battle of Granicus

in late 334 From Arrian’s accounts we learn that there was no battle. After

taking Telmessus, Alexander reached the Xanthos valley. The Lycian cities of

Pınara, Xanthos, Patara, and about thirty smaller places (unfortunately their names

are not mentioned in the sources) surrender to his army.

Alexander then advanced into the lands of Milyas, part of Greater Phrygia

(Appendix C.2). There he met a group of representatives of Phasehs, who offered him

the golden crown of their city as a sign of surrender and an offer of fiiendship,

followed by other representatives from lower Lycia making similar offers. Alexander’s

next move was to march to coastal Lycia, where he was the guest of the city of

Phaselis.^®®

Arr. Anab. 1.24. His general Parmenion having led part of the army straight into the Anatolian highlands from Sardis, to camp for the winter in the vicinity of Gordion, Alexander had arrived in Lycia with an army of 10-15 000 men. Arr. Anab. 1.25. In Phaselis he assisted the citizens in reducing a fortress built by the Pisidians used as a base for their raids from the mountainous terrain to the north of the city. 96

Strabo^*^ makes an additional comment as he is defining the mountain passages of

Solymia which connect the Milyan lands to the Pisidian city of , stating that

Alexander had to destroy the Milyan lands in order to control the passages to have a

better access to the city of Termessos (Appendbc C.5e).

Arrian’s account of Alexander’s campaign into the lands of Milyas, gives us no

information regarding the exact location of his route or the reasons for his campaign,

which was risky during mid winter, even for Alexander. His move into Milyan

territory was mainly to open a communication link with his general Parmenion, and

his army, now wintering in the Phrygian heartland of Anatolia, near Gordion, as

suggested by Freya Stark. An attempt to move north, trying to reach the Laodicea-

Isinda-Pamphylia route would, surely, serve this aim. Various routes have been

suggested; a most westerly route over the Xanthos valley^®®, and the shortest route

that would have left the coast from Kaş, over the Kasaba and Elmalı Plains to reach

Korkuteli (ancient ) seem to be the more plausible alternatives. Although it is

not known for sure which parts of Northern Lycia Alexander and his army passed

397 Strab. XIV.3.9. 398 Stark, 1958:102-120. The author has visited the places herself, and made detailed suggestions for possible routes which may have been taken by Alexander and his army into the Lycian uplands, during the mid winter, mainly taking the geographic and climatic factors into consideration. Staric, 1958:107-112, fig. 1 . The easiest route to lead such an army in mid winter into the northern highland«; of Lycia would be one of the most westerly routes from the Xanthos valley,, over Cadyanda or Aiaxa, to the region of Kibyratis (modem plain of Seki) and Karalitis (the region of Söğüt Gölü). This area however was not a part of the Milyas, and the route would not suit Arrian’s description. Starir ejqilained this by the fact that the term Milyan has been vaguely used, the exact territories of the Milyan land being as vaguely known to Alexander as today. '^Staric, 1958:113; HaU, 1986b: 144. An alternative route suggested by Stark, and supported by Alan Hall, is that with the help of local guides Alexander may have taken the shortest way up from the vicinity of Kaş, into the Kasaba plain. Alexander on his campaign frequently haH close relationships with certain locals of the lands which he had been conquering. This was certainly also the case for Lycia. We know from various accounts that his court seer, Aristander, on whom he relied a lot, was from Telmessus (modem Fethiye). His companion Nearchus the Cretan had some friends among whom was a local (fynast of Telmessus (Polyaenus, SttaL35). Passing through unknown territories, Alexander would be always provided by local guides, possibly through these connections. We do not know how far north he may have reached, before the delegates from Phaselis and lower Lycia reached him, and convinced him to come down. If he was able to pass over the Sineçibel Pass, he would have been in the Elmalı Plain (which would have brought him right into the Milyan territory). This would have enabled him easily to reach Isinda (the modem plain o f Korkuteli), and continue further into the Phrygian highlands. If he was not able to reach the Elmalı Plain an alternative would be to make a detour at the Kasaba Plain, under the safe gnirfanrp o f the Phaselian delegates, reaching the Finike Plain over the Arycandus valley. 97

through, it is certain that these uplands remained unconquered, an unsafe and hostile

territory for Alexander. This is implied from another event.^' After Alexander left the

region to continue with his campaign, he placed his fiiend Nearchus the Cretan in

charge of Lycia.'^

After Alexanders’ death in 323 b.c. in Babylon, however Lycia and Pamphylia

came under the rule of Antigonos, as lands were divided up by Perdiccas who had

assumed the supreme command of the Macedonian army.'“* With the re-distribution of

the satrapies in 321 b.c. by Antipater, Lycia and Greater Phrygia remained with

Antigonos, as before.'“^ As a result of the increasing interest of the in the region'“*, Lycia fell into their hands. In 309 b.c. I Soter took Xanthos and

Phaselis, starting the Ptolemaic rule in Lycia, which continued for another two centuries, till 197 b.c.'“® Under the Ptolemies a variety of taxes were installed on the

Lycians, which became a heavy burden and ruined the economy of the country.'“* An inscription from Telmessus gives us an insight into the Ptolemaic exploitation and remission. 408

401 Arr. Anah 1.26. When Alexander, during his stay in Phaselis, needed to send an important message to Parmenion, the messenger Amphoterus was accompanied by natives of Perge, and dressed in local garments, in order not to be recognized in these still hostile regions. He was only allowed to carry the message by word of mouth, without any written documents. '“* Arr Anah Ш.6. Arrian des^ibes it as ‘the lands of Lycia and the neighbouring territory as &r as Mount Taurus’. '“3 PF Ri^plementXin : 275; Diod. ХУШ.З.З. ^ Ibid, 275; Diod XVin.39.6; XDC.69.1. '“* Ibid, 275-76. Ptolemaic interest in the region was mainly due to the strategically important location of the Lycian peninsula, controlling the entrance into the Aegean, and providing sheltered outlet harbours for the Egyptian fleet. In addition it had valuable resources, which were thoroughly exploited by the Ptolemies, especially its timber, which was so important for the maintenance of the Egyptian fleet. '“4b id , 275; Diod. XX.27.1. '“* Jones, 1971: 99, and footnote 6. Among the taxes they payed was the purple lax and the ‘octroi’ which were collected at Alexandria. W6rrle, 1980: 63-72. The inscription from the fourth year of Ptolemy П, dated to 279 B.C. is a declaration of the assembly of Telmessos, who have hindered the attempts of Ptolemy П, and to give Telmessos and its territory as a ‘dorea’ -present - to Philokles, the King of SidoiL The inscription gives us an insight to the political situation, and the faith of the cities during this period Another inscription found a century ago in the Artemis temple (TAM П. 1) rrfers to Telmessos given as a ‘dorea’ to a member of the ^nastic family, Ptolemy, who was the son of , the daughter of Ptolemy I, and the Thracian King Lysimachos. ТЪе ‘dorea’ was granted by Ptolemy Ш, possibly his cousin, during the seventh year of his rule, which dated the inscription to 240 B.C.. 'The ‘dorea’ remained in the hands of the ИшАу for at least four generations. The inscription refers to tax ехепфйоп, which appears with the arrival of the new dorea-owner. 98

With the weakening control, and the ineffective rule of Ptolemy IV Philopator and

V Epiphanes, Lycia fell into the hands of the Seleucids. In 197 b.c. Antiochus III took

QYgi· iiig whole region of Lycia, Cilicia and Kana. The Seleucid rule of the region appears to have been short. In 189 b.c. the combined armies of Rome and Eumenes II

of , and the Rhodian fleet, ended the Anatolian conquests of Antiochus HI with the battle of Magnesia. The treaty of Apamea (188 b.c.) divided the lands of

Antiochus ffl between the allies of the winning side, under Roman supervision

(Appendix C3). According to the treaty, Lycia was given to Rhodes'"® and the Milyas together with Lydia, Laconia, and Telmessus''" to Eumenes n. The Pergamene dominance in the region lasted tiU the bequeath of the kingdom to Rome by its last king, Attalos m , in 133 B.C..

While the situation in Lycia during the turmoils o f the Hellenistic Period is quite clear, the situation in its upland regions presents a less vivid picture. Northern Lycia

According to the new legislations, tax from orchards, olive oil and grazing land were reduced to older rates ( possibly to the rates of the Persian period), announced to be 1/10 th of the aop and olive

^ RF <;iipplement XIII; 276; Livy, E ^ . XXXin. 19. Livy’s accounts tell us that Antiochus III first sent his two sons, Ardys and Mithridates, with an army overland, and he himself sailed along the coast. We do not know how eagerly the conquered lands accepted the change of power from the Ptolemies to the Seleucids, but from the fiagmentary accounts of the second century B.C. historian Agatharchides of , we learn that the Arycandians, because of their love of pleasure, were iinahlft to pay dept, and for this were willing to accept Mithridates as ruler, thinking that he would reward them by Öte cancelation of their ddjt. The event possibly refers to the Mithridates who is the son of Antiochus III, and points to the presence of the Seleucids in the region between 197- 190 B.C.. Jacoby, 196L 86.16(11), Athenaeus XII.35p.527F; also see Knoblauch and W itschel, 1993:233. The fote of Lycia imder the rule of the Rhodians, was also determined by the Romans. The Lycians heavily protested against the Rhodian dominance, and according to Polybius (XXX. 31.4) after three battles against the Rhodians, were set free by a decision of the Roman senate in 167 B.C. (Livy, Epit. XLIV.15.1, Polyb. XXX.5.12.). Lycian autonomy prevailed for another two centuries, till it 'was a under Claudius in AD.43. During this period the Lycian League flourished having its own federation, and starts to mint its own coins again, after a long period since the last Dynast under the Persian dominance. It is intp'^ n g that Telmessus is the only Lycian city given over to the Pergamene kings. As discussed by Wdrrle, 1970: 43-66, Telmessus seems to have held special status throughout its history. Rring hanHftH over to the Pergamenes, together with the Northern Lycian communities (eg. the Milyans), also brings into mind its strategically important location, as an ouqx>st into the sea for these countries.This was also the case during the Ottoman period, as ^tested from the various travellers’ accounts of the 19th century. Makri (Fethiye) and its port Levisi were the main ^^Acrinarinns for the wheat caravans from the Elm alı Plain and Seki Plaia 99

mainly constituted E buflfer zonc between the forces of the Seleucids snd the

Ptolemies, till they were given to the Pergamene King, Eumenes II, with the treaty of

Apameia. But, as during the period of Alexander, these mountainous uplands were quite autonomous and kept their ethnic identities throughout the Hellenistic period.

This is especially evident with the Kibyratic tyranny, which was established in the western part of Northern Lycia in the Hellenistic period."’" The last tyrant Moagetes was overthrown by the army of the Roman general Murena, who annexed the lands of

413 the tyranny to the Lycian federation in 82 b.c..

Further documentaiy evidence from the Republican period, prior to Lycia’s status as province of Rome, comes from Cicero"’", where he gives us the misdeeds of the

Roman quaestor Verres, who has looted the provinces including the 'Commune

Milyadum' (Appendk C.4) In the first century b.c.. MUyas was a land which was

412 Strabo (ХШ 3 17) gives us a detailed account of the Kibyratic Tetrapolis, established under Kibvra, which had the right of two votes and the cities of Oenoanda, Balboura, Boubon each of which had the right of one vote. According to Strabo, the tyranny expanded its territory to the borders of their neighbours the Milyans, Pisidians, and Lycians, and as far as the Rhodian Peraia. Strabo states further that the inhabitants of Kibyra were actuaUy of Lydian origin, and having conquered Kabalia and a part of Pisidia, had founded their city as a prosperous one through p«ftahiiching good laws. He also notes they used four different languages, Pisidian, Solymian, Lydian and Greek, which points to the ethnic diversity of its population. The city was able to field an army of 30 000 hoplites, and 2 000 horses, which is a great number during this period for a remote upland city. These c^es were always ruled by tyrants, but of genuine character. "’"Jones 1971:104; RF- Supplement ХШ : 277. The relationship between Lycia and the upland region of Kibyra to have been hostile throughout the Hellenistic period, which also extended into the Roman period with certain incidents, eg. the Oenoandans guiding the army of Brutus into the city of Xanthos during the siege of the city in 42 B.c. (Ад). Bella Civilia.4.10). The second century B.C. inscription discovered in (modem village of ören, north of the Xanthos valley), in 1946 refers to disputes between Araxa and the Kibyratic city of , which is accused for the raids by the tyrant Moagetes. The problem was taken to the court of the Lycian League, and solved through the succesfiil diplomacy of the ambassadors sent by the League. Another inscription from Araxa dated to 180 B.c. ( see B e ^ 1947-48: 46-58; Coulton, 1982:120) refers to fights between Araxa and the Termesians, in Oenoanda. During the second and first centuries B.C., Lycia suffered a lot from piracy, which possibly included also the brigands of the mountainious hinterlands, not to pif^rinn the well known activities of the Cilician pirate Zeniketes, whose piracy base on the mount of Olympos is noted by Strabo (Xrv.5.7) to dominate the whole of Lycia, Pamphylia, Pisidia and the Milyans (Appendix C.5f). His activities had terrorized Pamphylia and Eastern Lycia controlling his stronghold during the first century B.C., till he was stopped by the Roman general P. Servilius Isauricus in 78 B.C.. It is interesting to see that these uplands always were a potential problem, especially in the periods of lack of power, as is apparent even fix>m our near past fiT>m the observation of an event by the late 19th century traveller Davis (see pg 42). The decision of Licinius Murena (lieutenant of Sulla) to bring the Kibyratic lands under the control of the Lycian federation, was possibly a w ^ of rewarding the Ittyalty of the Lycians during the civil war between Sulla and Mithridates of , in 88 B.C.. "’" Cic. 1.95. 100

treated separately from Lycia, and it had its own tribal organization referred to as the

‘Commune Milyadum which could lay charge on the acts of a Roman quaestor.'"’

The Elmalx Plain possibly was a part of the territory of the ‘Commune Milyadum being one of the most fertüe upland plains especiaUy in producing grain, it would not

416 have escaped the exactions of Verres.

A review of the epigraphic, numismatic and archaeological evidence from the

Elmalı Plain, before it became a Roman province as a part of Lycia in a.d. 43 under

Claudius, will be given at this point

2.5.1 Epigraphic evidence

Among the inscriptions pubUshed from the Elmalı Plain, only a few have been

dated to the pre-Roman period. An interesting Hellenistic reference to one of the

ancient cities identified with the plain comes from the neighbouring region of

Kibyratis. This is the allotment list from Balboura'*” (near the modem village of Ibecik

at the spot known as Asar Tepe). One of the two new lot-holders, who have been

assigned an ethnic background, is a M aaaç K oppeuç (Masas from Komba'"*,

Hall, 1986b: 148. 416 Although we do not have any concrete evidence that the plain produced a surplus of grain as early as the Hellenistic period, we may attribute the establishment of an elite class in the first millenium B C (as attested by the rich sepulchral evidence from the plain), to the succesful utilization of its resources, o f which grain may have been one of the produces. Hall and Coulton, 1990:109-155.The inscription, originally the base of a stele, had been used as a spolia in one of the late fortification walls which ringed the upper city. The inscription had a long list of names (more than 280), that possibly referred to new lot holders (: 109-10). The names have been usually given by their patronymic or, sometimes, with name and ethnic, and even with name and occupation. The evidence gathered fi-om these names suggests that the new lot holders had strong Termessian or Pisidian connections (: 122-139). As mentioned above the two references given to ethnic origins come fix)m the city of (Domba and the Milyeis. The inscription has been dated from the middle of the second century B.c. to the middle of the fi^ century b.c., on the evidence of the lettering and the absence of Roman names (:121-122). It has been suggested by the authors of the article that the inscription points out to the western e^qiansion of Pisidians, some time after the Seleucids and the Ptolemies had lost their rulership in the region, and Rhodes (after 167 b.c. Lycian League) together with the Pergamene Kings were sharing the territory of Lycia and Northern Lycia. The Pisidian colonies could have certainly expanded during these unsettled conditions. It has been also suggested by the authors, that the western expansion of Pisidians m ^ have served in the communal organisation and urbanisation of their neighbours. The ancient city of Komba (Kop^a) has been mainly recorded as a name in the Byzantine Bishopric lists. The Geographer Ptolemy (Geog. V.3.5) has listed the city under the Lycian cities 101

identified with a mound near the modem town of Gömbe, along the southwestern tip

of the plain). The other one is a KöPıA,A,ıç A parnaç MıA,Xu8ç ( Kobilis Arapias from

Milyas).

Among the published inscriptions fi"om the Elmalı Plain, two have been dated to

the pre-Roman occupation of the region. The first inscription (Appendix FlOi) is a

dedicatory inscription found outside a house in Sarılar, a village close to the mound

of Hacimusalar, which has been identified with the ancient Chôma.A date before

AD 43 (before the Roman dependence) has been suggested. The dating of the

inscription was made according to the style of lettering. Further supporting evidence

given by the authors is that the martial exploits and war emblems do not point to any

mercenary service in the Roman army. The name of its mason has been also

mentioned in the inscription; he is known to us fi-om other inscriptions fi-om Side.''“

The second inscription (Appendbc F6c) is another dedicatory inscription fi-om Söğle.

The inscription upon lettering has been dated to the late Hellenistic period (first

century B.c.).The names of the dedicators have been given as Moles, Demetrias and a

certain Armasta, who honor the memory of their father. The name Armasta has been

suggested to be derived from the Luwian god name Arma.''421

2.5.2 Numismatic evidence

Coins with the mint of the two ancient cities Chôma and Podalia''“ , associated with

the Elmalı Plain, point to their Hellenistic existence. But the scanty evidence of the

near mount Kragos (associated with the modem Baba Dağ, to the east of Fethiye). R Heberdey was the earliest traveller who identified the location of the city. In his article, 1898; 157-158, he identified a site half an hour in distance from the modem town of Gömbe (as the nam^ ¿so suggests), as the ancient city of Komba. A limestone base with inscription from the site was published by Heberd^f, which bears the name of the city, as ‘KopPeoov’. Harrison and Bean, 1967; 34. Ibid, 43. tpiiifçingİH, 1992; 24, with Neumann’s suggestioa ''“ The identification of the location of the ancient site of Podalia has been more problematic, since no epigraphic evidence regarding the city has been located so fer. Two sites have been nominated for its possible location. Early travellers, especially starting with Petersen and von Luschan in their 102

Hellenistic coinage of Chôma and Podalia far from suggests that the cities were a part of the Lycian League during this period.

The Hellenistic bronze coin of Podalia is represented with an Artemis proton head on the obverse, and jumping ‘zebustieres’ stag? with the name of the city on its left as n o AAA, on the reverse.''^ The only bronze coin of Lycian League mint of Podalia'’^“ has an head and the letters AY on the obverse, and bow and quiver crossed in incuse square, with the letters HOA on the reverse The coin has been referred to as dubious, and not a League mint, by Troxell.

The first century b.c. bronze coinage of Chôma appears with the radiating head of an uncertain divinity, which also appears to be on the coins of Arycanda. The god has been identified as 08oa SoCov or Sabazios.'*“ The first century b.c. bronze coin of

Chôma, in the British Museum collection, has the Laureate head of facing left on the obverse, and a horseman wearing a crested helmet and clamys flying behind shoulders, riding, and raising a club in right hand with the letters XQ on the reverse

(Plate 61 The figure has been identified as the rider god Kakasbos, who has been frequently attested in the Lycian uplands of Kibyra and Kabalia, as well as

Pisidia The image of this god was reintroduced on Chôma coins, with Gordianus HI coins of the third century a.c..

Rftisen I I 161-162, and followed by later scholars, have identified the site at the Avian Lake with Podalia. Another alternative has been pointed out by Bean, who has identified the site of Söğle, at the north-eastern part of the Elmalı Plain, with the site of Podalia (Bean, 1968; 157-163). Aulock, 1974: 51, as it has been published in Л М 6 (1903) 192,122, plate XI.( non-vidi) Aulock, 1974: 34,54 as it has been published in Mionnet, 1806-37 :22,88. TroxeU, 1982: 240. BMC Lvcia Iviii. BMC Lvcia Plate XI. 1. 103

2.5.3 Archaeological evidence

Other than the ongoing excavation project of the Hacimusalar mound in the plain, no site has been excavated that has given evidence for the Hellenistic or later occupations/“

The only archaeological evidence that can be pointed as Hellenistic are the two hill forts at Gilevgi'*^ and Avlan^^°, which also shows later occupation (possibly well into the Byzantine period). The dates of these hill forts need to be taken cautiously, since no detailed research has been done. Pottery finds at these sites, and their building technique has suggested a date in the Helenistic period. These hill forts hold the strategically important passages into the plain. The Gilevgi (Plate 59) fort located

18 km. northeast of Elmalı, controls the important passage into the Korkuteli Plain and to Pisidia and the southwestern edge of the Anatolian highlands. The other hill fort is located south of Avian Lake (Plate 60 A), and controls the important passage of Avian that leads to lower Lycia, passing by the two important ancient cities of

Arycanda and Limyra, ending up in the plain and the coast of Finike. Another possible

Hellenistic hill fort is located at a site to the south of the village of Armutlu (Plate 60

B), which is located southwest of Elmalı, and four km. south of the Hacimusalar mound.

Hacimusalar-Choma, the ongoing excavations on the mound have yielded abundant Hellenistic and early Roman pottery, suggesting a Hellenistic occupation on the mound Podalia, according to two different views has been identified with the site of Avian Lake, and with the site of Söğle (see fii. no. 2). Both of the sites have yielded Hellenistic and later pottery, with the former site also exHbiting a Hellenistic fortification. No surv^ has been done at the mound site of Gömbe (Komba). '■^R e ise n fi: 164-65. ““ Bean, 1978: 153-54. 104

The construction of all of these hill forts exhibits polygonal masonry o f crude type, which has been also attested from Hellenistic fortification walls of the neighbouring region of Kibyra.''^'

The hill fort of Gilevgi was first visited by Beimdorf in 1882, and published by

Petersen and von Luschan together with a sketch plan/^^ The fort site is located 200 meters above the plain ground on a saddle known as the Çalca Tepe. The fort, as suggested from its drawing, has an irregular shape with an approximate diameter of

100-160 meters. Three of the towers were in a good state during the visit of the authors, and the remains of another one could be traced. Each tower had one entrance door, of 1.04 meter wide, opening to the inside o f the fort. The comers o f the towers were built of larger rectangular blocks of dressed stone. The thickness of the walls varied between 1.3-1.8 meters. The walls had polygonal and dressed blocks in the interior and exterior, filled with mbble in between.

The fort site of Buralya or Avian Gölü has been described by Bean"^^ as being located along the northern part of the saddle, the regular ashlar masonry of the ring walls standing up to 10 feet high, with a possible tower or citadel at its summit, built of smaller stones. He also mentions two plain rock-cut tombs, not resembling the

Coulton, 1993: 83-84. Several small fortified enclosures have been noted during the visits of Bean to Northern Lycia. Five of these enclosures are located near the villages of Karahasantaş, Çobanisa, Çaltılar, Doğanlar to the west of Balboura and Manay northeast of Söğüt Gölü in Kabalia. All of the forts exhibit a «milar construction of polygonal masonry with varying shapes but similar size. As siiggpgtpd by Coulton, they all are located near arable lands and possibly served to hold land, especially for the new immigrants attested in the Hellenistic allotment list of Balboura. In the main cities of Kibyra, Oenoanda, Balboura and Bubon, the earliest city fortification walls have been dated to the Hellenistic period. The Hellenistic fortification walls of Oenoanda, on the basis of their masonry have been described as; ‘quarry-to hammer-faced in the interior, and with well-nigh isodomic hammer-faced ashlar headers and strechers on the exterior’, and have been dated to the late third, early second century b.c. by McNicoU, 1997: 120-26, plates 51-53. The design and execution of the walls were similar to the Hellenistic walls of Pergamon, which has suggested a date after 188 B.C., with the treaty of Apamea and the increasing involvement of the Pergamene Kings in the region (Coulton, 1983: 5-6). But it also needs to be stressed that the polygonal Hellenistic tradition of masonry was attested in later periods, imder the Flavians. The earliest of the two bath buildings of Oenoanda, with the aqueduct system which was associated with the bath building, dated to the Imperial period, shows the polygonal building technique(6-9). Reisenll 164-165,fig.77, plate XXHI. "^Bean, 1978: 153-154. 105

Lycian type. Inside the ring wall he encountered considerable quantities of uncut building stones, but no squared blocks or marble. The pottery he suggests to be of

Hellenistic or a later date.

The hill forts of the Elmalı Plain are different in function from the hill forts of the

Kibyratic, as they mainly serve to protect the plain from intruders. The migrations which took place into the Kibyratic region during the Hellenistic period, have not been attested in the Elmalı Plain.

To summarize the Hellenistic period. Northern Lycia was more or less a buffer zone between the Seleucids and Ptolemies, and was later subjected to the Pergamene

Kings. Towards the end of the period Roman influence was felt more and more, as the

Roman generals and proconsuls had the last say in the most important decisions for the fate of the cities and regions. As the ongoing Hellenization of the coastal cities of

Lycia was taking place''^^ this was less evident for the sites of Northern Lycia. The upland regions were able to keep their ethnic identities, as is suggested by the ethnic diversities attested from literary and epigraphic sources from the region. The cities of

Kibyratis, under their tyrants or/as well as the new immigrant population from Pisidia, were more fortunate in developing into urban centers. This is less apparent with the ancient cities of the Elmalı Plain, which most likely was a part of the Milyan lands.

We have no literary evidence for any of the cities during the Hellenistic period identified with the Elmalı Plain. As suggested by Alan HalP“ , the Milyans were slow in urbanisation. Their cities were possibly small, located on mound settlements vrithout any fortification. Possibly, this was the reason for the location of the strong hill forts at strategically important locations, as had been attested from the two large fort sites

Coulton, 1993: 84. Jones, 1971: 99- 100, gives a discusión of the Hellenization of Lycia, which starts as early as the fourth century. The epichoric Lycian inscriptions and Lycian type of tombs come to an end with the third century B.C., and are not attested later. Also see Farrington, 1995:121-33, who gives a detailed discussion of the Hellenization and Romanization of Lycia. Hall, 1986b: 152. 106

of Gilevgj and Buralya that were protecting the main passages into the plain, during

the Hellenistic period. The Hacimusalar mound, identified with the ancient Chôma,

does give evidence for a Hellenistic settlement, as the prelimanary results of the

ongoing excavation show .The Hellenistic existence of the ancient cities identified

with the Elmalı Plain, is evident fi'om the scanty epigraphic, numismatic and

archaeological evidence. Literary sources do mention a commune Milyadum but

without a specific city. The first mentioning of the name of the cities of Chôma,

Podalia, and Komba in literary sources is no earlier than the first century by

which time now they are definitely a part o f the Lycian League and have become a

part of the Roman province together with Pamphylia, under the emperor Claudius.

The earliest literary reference to the cities identified with the Elm alı Plain comes fi-om the accounts of the first century historian in his book Natural History (Book V.28), who counts the cities of Chôma and Podalia under Lycia (see ^)pendix 4.1). Ptolemy in (Geography, BookV.3.7), while listing the cities of Lycia, makes a distinction for those cities of the Lycian Milyas, among which he counts Ntsa, Candyba, Chôma and Podalia (see .^jpendix 4.2). Though Chôma and Podalia are surely to be identified with the Elm alı Plain, the ancient cities of Can(fyba and Nysa have been identified with sites of the Kasaba Plain, the adjacent plain to the south. 107

2.6 THE ROMAN AND LATE ROMAN PERIOD

Following the integration of Lycia with Pamphylia as a double province, under

Claudius in a.d. 43, the Lycian League lost the independence granted to her in 167 b.c,

by Rome. From this date on, Lycia became an Imperial province, governed by a

praetorian leg a te.The first Roman governor Quintus Veranius as the legatus

Augustipro praetore was able to pacify the province and reform the administration

during his rule of five years.''^® Under Nero or Galba, Lycia was separated fi-om

Pamphylia for a short period, and once again was joined to Pamphylia, under

Vespasian, as late as a.d. 74.'"“ The Roman emperor visited Lycia personally in ca. a.d.

70 on his way fi’om Alexandria to Rome.'"" Other emperors who had paid a visit to

Lycia were Trajan in a.d. 113, on his way to the Parthian war, and Hadrian who visited

Lycia twice in a.d. 129 and 130, and Lucius Verus. Till A.D. 178 the province Lycia-

Pamphylia was governed by a legatus Augusti pro praetore, an imperial governor.

Sometime under Lucius Verus, the province was given to the Senate.“"*^ After A.D.

178 Lycia-Pamphylia was governed by Senatorial proconsuls, continuing till a.d. 313 when, with the reorganization of the provinces under Constantine, Lycia was separated from Pamphylia, and Myra became its capital and the seat of the metropolitan bishop."*^

438 Jones, 1971:105: RE Suupl. XIII : 278. The reason for the sudden annexation of Lycia by Claudius, was given as the breaking out of an internal riot in Lycia, which had ended with the murder of Roman citizens (Cass. Dio LX.7.3). '•^^Foss, 1996: 14. RE SuPDl. X in : 278-79. '"" Ibid., 279. '"'^Ibid. Ibid., 279. Possibly it was exchanged for the province Pontus-Bithynia, during the Parthian wars, because the later province lay on the main route to the eastern frontiers of the The exact date of Lycia-Pamphylia becoming a senatorial province is not known. According to Cassius Dio (LXIX. 14.4), this took place in A.D. 134, which falls into the reign of Hadrian, and has been considered to be an early date. An inscription from Arycanda published by şahin, 1984: 39-48, and dated to ca. AD. 160, gives the name of the proconsul who governed the province of Lycia- Pamphylia, in the name of the Senate. Ibid., 279. 108

The Lycian League continued to function under Roman rule, and it played an

intermediary role between the cities and the emperor. Under the Roman rule, it lost its

indépendance in international matters, and had to follow the foreign policies of the

Roman Empire. Its main responsibilities were to assure internal security, organize

religious festivals, reward public services for the member cities of the league, and

most importantly to organize the celebrations for the Imperial cult. Its courts

continued to function and collect the imperial tribute. The territory of the league

seems to have extended its area under the Roman rule, as Pliny counts 36 cities,

during the mid first century a.c., among the cities of the league, and Ptolemy gives 33

cities for Lycia.“^* In both accounts the ancient cities of Chôma and Podalia,

associated with the Elmalı Plain, were included among the Lycian cities. The first

century B.C. accounts of Strabo give us information regarding the administration and

functioning of the league, which basically seems to have continued under Roman rule,

with slight differences in its organization.'^

Lycia flourished during the period of the Pax Romana, with an accelerating

development encountered both in coastal and inland cities, which continued during the

Jones, 1971: 106. New cities eg. , Phaselis, in Pamphylia, the Cabalian cities of Oenoanda, Bubon, Balbuia, as weU as Chôma and Podalia in the north were added to the league. ^ Strab. XTV.3.3. TTie league consisted of 23 cities, and each city had the right of between one and three votes, the three vote cities being Xanthos, Patara, P ın a ra , Olympos, Myra and Ήos. The chosen delegates from these cities would form the confederacy which voted for a Lykiarch, and a number of other officials.The Lykiarch would be choosen for one year, and was assigned the highest civil and military position in the confederacy. The positions of the Strategos, Hipparch, Hypohipparch (which dealt with the millitary) and the Nauarch (who was reponsible for the navy) Hiiring the independent period of the league, were replaced with the Archiphylax and the Hypophlax, which functioned as the local police force of the league, being responsible for the intentai security of the member cities, under the Roman period. The Grammateus hmctioned as the secretary of the confederacy. Strabo mentions that a court could be found during the general meeting of the confederacy, named as the Dikastai, which would decide legal matters. The Dikastai seems to have cf^ntinnfttt under the same name in the Roman period as suggested by epigraphic evidence. During the ind«T^"riftnt period the tribute from the member cities would be collected the tamieusas, ‘ταμιευσοις του κοινού’, who was also responsible for the financial af&irs. Under the Roman rule it seems that the archiphylax was responsible for the collection of the imperial tax. The Lykiarch also fiiprtinnftH as the highest priest, whose duties were mainly to conduct the ceremonies of the Apollo cult which took place at the religious site of Letoon. With the Roman rule, the main role of the Lykiarch became to perform the ceremonies of the Imperial cult, holding the title of the ‘αρχιερευς των σεβαστών’. Dekretes would be dated to the official year of the Archiereus, and a list of Lykiarchs have been given in Magie II, 1609-1612. (RE Suppl. ΧΠΙ: 282-286). 109

late Roman period and well into the sixth century.^’ The change is especially attested in the coastal cities, Xanthos, Patara, Myra, which flourished due to their position along the main sea-lanes of the eastern Mediterranean, from Alexandria via Cyprus to

Rome/^ Lycia became a part of the grain trade that was so fundamental to the

Roman Empire. During the time of Hadrian, massive granaries were built at Patara and at Andriake, the port of Myra.'”® The grain which was stored in these granaries mainly came from the fertile valleys and delta plains o f the Xanthos river (Esençay) and the plain of Demre. With the increasing demand for grain in the Roman Empire, the upper plains of central and northern Lycia, which possibly also included the

Elmalı Plain, the largest arable area in the vicinity, were also involved in the trade.

Their produce would be brought down to Myra or Patara, and stored in these massive grananes.• 4S0

Timber was another important produce of the region. Especially with its rich cedar

(Pliny HN Xn. 132, XIII.52XVI. 137) and pine forests located in central and northern

Foss, 1994: 1-51. Clive Foss in liis article gives a detailed discussion of the flourishing of the coastal and inland cities of the late antique Lycia, whose economy was stabilized during the Roman period, and Lycia became an internal part of the eastern trade (;2-3). Archaeological evidence points to expansion of the main urban centers of the Roman period, as well as the establishment of new sites, both along the coast and inland, during late antiquity. The coastal cities, especially with their suitable harbours, were in close contact with the mountain villages of innCT Lycia, and their produces including wheat, wine, oil and especially timber would be traded through these harbours to other destinations. They were wd)bed in a network of roads, whose traces can be observed today. C. Foss gives a summary of the archaeological evidence of a number of regions and sites in discussing the evidence for this economical growth, including the districts of Telmessus with the harbour site at Ldiissos (Gemile adasi), Aparlea, Kyanaei with the harbour site of Tristomon, and Limyra, as well as the main urban centers of Xanthos, Patara, Myra and its hinterland. For seaways in general of the coast of Asia Minor see, Broughton, 1938; 857-60; specifically for Lycia see Zimmerman, 1992: 201-217 and the article by Keen, 1993; 71-77. Although the article disciissfx the periods earlier than the fourth century, it gives a good insight into the strategic importance of the Lycian coast Also see Parker, 1992; map 1, for the evidence of shipwrecks around the coastline of Lycia. Borchhardt ed., 1975:67-68.The conunentary of a buüding inscription has been given by Wörle which dated the ‘horrea’ of Andriake between A.D. 119-138, into the reign of Hadrian. For the inscriptions regarding the granaries also see, TAM П.397, for the granaries of Patara, TAM П.409- 412= IGRR Ш. 661-663, TAM n.419=IGRR Ш.660,665. Magie П, 1950; 1482, note 37. Wdrrle, in Borchhardt ed , 1975; 67, also states that the idea was first presented by Parabeni in nprimiis Princeots (1926) 175. Parabeni had stated that the possible reason for the massive horrea was the storage of grain from the fertile plains of Kabalia and Elm alı, as well as the Xanthos valley, to be exported further to Rome. An mscription from a sarcophagus found in Andriake points to an imperial hortearius, who possible was in charge of the storage of the products and their delivery. 110

Lycia, and sandal wood of Patara, which continued to be imported to Alexandria."451

Other products and natural resources mentioned by ancient writers from Lycia include"“ a special medicine called Lycium (Pliny, ffii XXIV. 124 and XXV. 67), wine of Telmessus (Pliny, XIV. 74), boars (Pliny, Kfrl XI.280), sponges which were especially well known (Arist. Hist. An. V. 16 and also Plinv.HN XXXI. 131), purple fish which is used in producing purple dye (Vitr. VII 13.2) and cimolian earth, a kind of white earth used for curing from Bubon (Pliny, №1 XXXV.201). Sculptors of Trysa have been mentioned by Broughton"“ , and a guild of tanners and leather workers from Kibyra as well as Termessus have been attested from epigraphic sources."“ Kibyrians have been also noted for their ability in working iron and the process of inlay (Strab. XIII.4.17).

Urban development seems to have accelerated under Roman rule, and the cities of coastal Lycia flourished."*^ This was also apparent with upland regions, especially in the Kibyratis, where larger cities such as Oenoanda, Balboura and Bubon were transformed into fairly large urban centers when compared to other north Lycian centers.

Oenoanda"“ (near the modem village of Incealiler, in the Seki Plain) displays a number of imperial buildings, dated to the reign of Hadrian, as well as into the Flavian

RF. Supol. xm 288 . "“ The information has been compiled by Broughton, 1938: 817-39, and in Magie II, 1950:1372- 1373; also see RE SupplXIII 288. "** Broughton, 1938: 833. lbid.,824 ( IGRR IV.907 mentions tanners from Kibyratis and IGRR. in.422, mentions leather workers from Telmessus). It is interesting to see that the taiming industry continued in the region into modem times. In the accounts of the early travellers the town of Elinali has been especially mentioned for its tanning industry. "** Foss, 1996:15-19 eg. Xanthos, Patara, Letoon, Pınara, , Myra, Limyra and others became urban centers adorned with civic buildings eg. tem^des, basilicas, theaters, baths, gymnasiums, odeons, council halls, , coUonaded streets, and monumental arches. Also see by Farrington, 1995: 121-36, for a discussion of the Romanization of Lycia. "“ The site of Oenoanda has been known since the mid 19th century by a number of travellers and scholars, including Spratt and Forbes, Heberdey, Kalinka, Petersen and von Luschan, Diehl, and French scholars such as HoUeaux and Cousin. The main point of interest for Oenoanda has been the fragments of many irtscriptions, of the Epicurean Diogenes, a native of the city, who became the main subject of an extensive survey of the ruins of the city that started under the direction of A Hall, J.J. Coulton and M.F. Smith, in 1974 (HaU, 1976:191-9^. The native philosopher of Diogenes was I l l

and Antenine periods. But most of the buildings appear to be of Severan date, of late second century The surviving city monuments suggest a sudden shrinking in the third century, when the late Roman city walls now enclose a much smaller area, in relation to the Hellenistic enclosure, leaving out certain parts of the earlier buildings, especiaUy the ‘esplanada’, the theater and the southern part of the city. An interesting enclosure, identified as a fort, within the late Roman fortification walls has been dated into the fourth century, together with an early basilican church. The plan of the fort very much resembles the hilltop site located over the village of Ovacık, in the Elmalı

Plain, as noted by M. Harrison.''^ The building history of Oenoanda has been summarized by J. J. Coulton, as a flourishing period under the early empire, with the

Lycia to its south being the source of influence, and later with the second and early third centuries, with influences coming fi-om Aphrodisias.''*® The site experienced a sudden contraction in the third century, followed by a partial recovery that continued till the early Byzantine period, apparently a pattern quite common for Asia Minor, and the region.'**® We need to keep in mind that this information relies on survey results and no excavation has taken place in Oenoanda. future excavations may change the information regarding the history of the site.

Oenoanda is also well-known for its agonistic festival, which was promoted under the emperor Hadrian, and founded by the local citizen C. lulius Demosthenes who privately sponsered the athletic games combined with a music festival.“*461 possibly a member of one of the wealthy and influential flimilies of Oenoanda who was bom in late second century and lived in the third century. His philosophy in the line of the Epicurean teachings advised against the pursuit of wealth and political power. His woiks and inscriptiotts have been extensively studied by M.F. Smith (1979: 69-88; 1980:73-87; 1993 also see the note by Hall, 1979:160-163). The irtscriptions of Diogenes, which mainly display the two treatises of his Physics and Ethics, that extend at least over 80 meters, is thought to have been inscribed on the walls of the south stoa which flanked the area known as the Esplanada, a flat polygonal open space, with an unknown function, flanked by the two stoas to its south and north, which is located on the higher flanks of the hill, north of the main ruins (Coulton, 1982:58-59). “**^ Coulton, 1983: 4-14. “*** lbid.,17, fig.7. For a discussion of the Ovacık hill top site, see pg. 118-19 and Plate 64. '**^ The architectural order and decoration of the porticoes are paralleled by features from Aphrodisias. For a more detailed discussion see Coulton, 1983: 10-12. “*^Ibid,17. “**' Important epigraphic evidence dated to the reign of Hadrian (ca. A.D. 120), gives the detailed account of the celd)ration of the festivals and the duties of the various feast officers and the 112

The ancient city of Balboura'“^(on a mountain top known as Asar Tepe, near

Dirmil) also possesed a number of imperial buildings. The ruins of the ancient city spread over an extensive area, including an with parts of the Hellenistic fortification walls of polygonal masonry remaining, dated tentatively to the second century B.C..Aqueducts carry water into the city from a nearby spring located outside the city, possibly to the bath building'“^ dated to Vespasian by inscriptions. There is also a paved with triple arched gateway dated to the Severan period, the late

Roman fortification walls, which refortified the acropolis hill in a smaller circuit, and three churches with basilican plan, mainly dated to the fifth century. Pottery evidence from the site points to habitation till the seventh century, with an expansion into the lower plains during the Roman period.'^ The upper theater of Balboura has been tentatively dated to the early Roman period, and the unfinished lower theater has been dated into late second and early third centuries.'“^ Buildings of special interest for the social structure and history of the city include three small buildings located along the south-western comer of the agora; an exedra of Onesimos, a city slave, and his temple of ;'^ the exedra of Antoninia Meleager son of Castor, a member of the leading families of Balboura, and an associated inscription of Meleager, which mentions the games he organized."“^ The building and the inscriptions, as suggested by institutions of the Boule of Oenoanda, by the native C. Julius Demosthenes. The inscription has been translated and published by Wdrrle, 1988, giving a detailed account of the social and administrative system of Oenoanda and its territory. Also see the article by M. S. Smith, 1994: 59-64. The site has been also known and visited by travellers and scholars since the mid 19th century, together with the other Kibyratic sities of Bubón, Oenoanda and Kibyra. The site similar to Oenoanda, was the focal point of researches due to its inscriptions, which gave reference to the athletic games and celdjrations that were organized by the wealthy donor Meleager. To have a better ^i^rfffrctanding of the inscriptions and the buildings associated with the inscriptions, as well as the general building programe of the site, an extensive survey of Balboura was started in 1984. (Coulton et. al.,1988a:121-145; 1989: 41-62 also see by Coulton, 1986:171-178; 1987a: 205-211; 1988 : 255- 231). Farrington and Coulton. 1990: 55-61. ^ Coulton, 1986: 172-175; 1987a: 205-211; 1988b: 226-227. '«'5 Bier, 1994: 23-46; 1990: 69-79. Onesimos was a city slave (démoslos), who was wealthy enough to dedicate two buildings in the city agora, right next to the building of the aristocrat Meleanger ( Coulton, 1987a: 208). All of the buildings have been dated into the mid-second half of the second century, to the reign of . Though none of the statues have been located, it has been suggested from the bases of both of the exedrae, that they were of bronze, similar to the Bubón Sd>asteion examples (Coulton et.al.,1989:61-62).The Meleager inscriptions have been completely published by 113

the authors of the articles, point to a considerable elite class in the remote upland region of Lycia.This eUte class sponsored the celebration of the athletic games and competitions, which seems to have constituted an important part of the social life of

468 the city, and also served for the celebration of the imperial cult/

An important piece of evidence for the celebration of the imperial cult in these upland regions comes from the Kibyran city of Bubon“'^®, which lies only 18-20 km. west of Balboura (known as the Dikmen Tepesi, near the modem village of Ibecik). A rectangular building excavated by Jale Inan has revealed statue bases with inscriptions, which were dedications to the various emperors and empresses starting with Nero, Poppaea Sabina, Nerva, possibly Commodus, Septimus Severus, Julia

Domna, Caracalla (twice), Gordian III, Gallienus and Salonina.The building was identified as the Sebasteion of Bubon, housing at least 8-10 bronze statues of the emperors and their empresses, which had been looted during this century.'*^’ Bubon is also known from literary sources to have acquired the right of three votes in the

Lycian League, through the emperor Commodus, who praised and rewarded the city for having succesfully defeated the local brigands.“', 472

The »lain Kibyran cities of Oenoanda, Balboura and Bubon were well integrated into the Roman culture. There was a considerable, wealthy elite class Uving in these

Milner, 1991: 23-62. The inscriptions include 12 bases and a number of other inscriptions, all of which have been found in secondary context. The bases, which have been located upturned, used as internal piers of the basilican church nearby, have been pointed out as being peculiar. The inscriptions gives us the details of the athletic games and music festival, referred to as the Meleagreia, named after its founder, and receiving the ai^roval of Antoninus Pius, in A.D. 158, which dates the first victors of the festivals, foUowed by the 4th, 6th, 7th, 9th and 11th celebrations. The pmfts were designed on the Demostheneia which their neighbour, tire city of Oenoanda, was celd>rating.The main content of the inscriptions gives us the victors of these athletic games, and especially their association with members of the leading fam ilies of Balboura. ‘'*®^ulton etal., 1989:62. The site which has never been excavated or surveyed, has became the target of illegal digs. A rescue excavation of a building took place in the year of 1% 71^ Jale Inan, after a bronze torso revealed fi^om the illegal digs was brought to the Burdur Museum during Âe same year. Inan, 1977/78: 267-73. Ibid., 274-87. A number of bronze statues in the museums of the United States, have been pointed out as possible candidates for having arrived from the Sdrasteion of Bubon, on stylistical grounds and similarities with the Burdur bronze torso, by Jale İnan Bean, 1978: 164. 114

cities, which possibly had their roots well in the Hellenistic period'*^\ and with the

peaceful period of the Pax Romana, became rich donors of the buildings that

decorated their cities. They became the leading organizers of celebrations of agonistic

festivals, which had become an important aspect of the social life and were related

with the imperial cults, and this was a way to show their loyalty to their new rulers.

The effect of the crisis of the mid-third century also had an impact on the cities, as we

see the cities contracting behind late Roman enclosures. Whether the upland

populations did significantly shrink, as suggested by C. Foss, is still a debated issue,

and mainly lacks evidence on both sides. But at least the pottery evidence, as

suggested by J. J Coulton'’^“ seems to continue well into the seventh century in the

region.

Unfortunately, fi’om the classical sites of Elmalı we do not posses any

archaeological evidence for massive civic buildings. Epigraphic evidence, however,

gives reference to the donation of certain civic buildings for the ancient city of

Chôma, and archaeological evidence of surface finds such as pottery and inscriptions

suggests that the plain was abundantly populated during the Roman period, and well

into the medieval periods. Lack of evidence for larger civic buildings might be due to

the nature of the sites (mostly mound settlements) or, as has been suggested by M.

Harrison,''’^ due to the heavy Islamic occupation of the plain starting as early as the

11th and 12th centuries. The Elmalı Plain appears to have received the earliest

Islamic settlers in the region.

During the crisis years o f the third century, with the decreasing stability in the

Roman Empire, Lycia seems to be once again affected with bandits, fi-om Isauria,

Especially see the Licinii of Oenoanda, a leading family of Oenoanda which Haimc descent from the Hellenistic period. The long genealogical inscription of Licinnia Flavilla, rlaimg descent from the Spartan Cleander, who is said to be the founder of Kibyra ( Jameson, 1966:125-137). Also the article by Kearsley, 1988:43-51. Coulton, 1988b: 226. Coulton also states that entire reliance on pottery evidence for this critical period would be over-optimistic. Harrison, 1981b: 387. 115

especially during the reign of Probus in a.d. 276-282/'’® An inscription from Ovacık,

a mountain site southeast of the Elmalı Plain, points to a settlement which was

founded during these troubled periods, which was guarded by a duke of the Guard-

posts. The inscription praises a certain person who has slain the brigands, but the

problems with them seem to continue for the very same site, which proves to be short

477 lived. The inscription was dated to before a.d. 361.''

For late antiquity and the early medieval period we have two opposing views given by Harrison and Foss. According to Clive Foss, after the crisis years at the end of the third century, and into the sbdh centuries and the later dark ages, the remote uplands of northern Lycia, which includes the plain of Elmalı and the Kibyratis, severely suffered and were depopulated to a great extent.''’* Martin Harrison on the contrary, debates for a shifting of populations into the uplands, especially to central Lycia from the coast, with the early medieval period. His evidence to support the theory mainly refers to the early Christian evidence from central Lycia, which will be the subject of the next section.

At this point I would like to give the literary, epigraphic and archaeological, and numismatic evidence from the Elmalı Plain, related to the Roman and Late Roman periods. No excavation or detailed study of the period has taken place in the region.

The ony archaeological work conducted for the region are the epigraphic surveys and identification of some of the mound sites as the classical cities of Chôma, Komba and

Podalia, which are mentioned in literary sources.

“'’® Foss, 19%: 15. Lydius, the Isaurian bandit leader, had led a major insurrection in Pisidia, threatening the provinces of Lycia and Pamphylia, during the reign of Probus. His m ain base in was defeated by a major Roman campaign. See Mitchell, 1995:177-219. Harrison, 1979c: 530-531. ''’*Foss, 19%: 20. 116

2.6.1 Literary Sources (with reference to Chôma, Podalia and Komba)

The only reference that mentions the cities of Chôma, Podalia and Komba comes from Pliny The Elder, who mentions Chôma and Podalia as Lycian cities. Pliny makes an additional remark concerning the geographical location of Chôma, indicating that it was located by the river Aedesa (Akçay river), which has supported the identification of the Hacimusalar Mound with ancient Chôma (Appendbc D.l).

The second century a.d. account of Ptolemy (Appendbc D.2) counts Podalia and

Chôma, together with and Nysa among the cities of Lycian Milyas. Kandyba and Nysa identified with sites in the Kasaba Plain. The classical city of Komba, which has been identified with Gömbe, has been listed among the Lycian cities near Mt.

Kragos (Babadağ).'·^®

2.6.2 Epigraphic evidence

The Roman period has produced abundant epigraphical evidence from the Elmalı

Plain. Since the earliest travellers to Lycia, inscriptions have been recorded and published from the northern regions of Lycia and the ancient regions of the Kibyratis,

Kabalia and Milyas. The inscriptions from the Elmalı Plain have been mainly published by G.E. Bean'^, M. Harrison and G.E. Bean*“’ and C. Iplikçioğlu"“^. Most of these inscriptions are dated to the second, third and fourth centuries, with few inscriptions dated to the Early Imperial period. As has been attested in general, no inscription is evident after the fifth century.““^ Most of the inscriptions are funerary in nature and honour family members or are dedications to certain deities. The frequently mentioned deities give us an insight to the cult practices of the inhabitants of the

See footnote 1 and 418 for the discussion of the identification of Komba with Gömbe. 'Bean, 1971:22-28. Bean and Harrison, 1%7: 34-44. ^ Iplikçioğlu and Çelgin, 1992. 483 Foss, 1996: 21. 117

Elmalı Plain. A list of the persons’ names attested from the inscriptions has been also included in Appendix G, to give an insight into the ethnic makeup of the population. The frequently mentioned names seem to be common names for the uplands of Kibyra, Kabalia and Pisidia.

(a)Epigraphic evidence outside the plain, which refers to Chôma and Podalia The Opramoas inscription from Rhodiapolis'*^''

Both Chôma and Podalia received funds from Opramoas of Rhodiopolis, during the mid second century a.c.. The Rhodiopolis Heroon inscription refers to a payment of 7 000 deneraii, for Chôma, for the construction of a stoa and an Augusteum. The donation to Podalia, on the other hand, does not give any details. The mentioning of an Augusteum and Stoa for Chôma suggests that we should expect to find certain civic buildings, possibly from the vicinity of the Hacimusalar Mound. It is also evident that celebrations of the Imperial cult did take place in the plain.

IGRR Tomus III.739, cap 63 = TAM 11.704 32 ‘δενάρια, ς και τη Χωματέων εις στοάν [και Σεβα]στειον | δηνάρια, ζ και τη Π[ο]δαλιωτσν δηνάρια....’

The Opramoas Inscription from the Heroon of Rhodiapolis, was first published by Hdjerdey, 1897, with following publications in TAM 11.905 and IGRR Ш. 739. The grave monument belonged to a local citizen of Rhodiapolis, Opramoas, who was a local millionaire and philanthropist who had dedicated around 1000 000 denari to the various Lycian cities for the rebuilding of their buildings, which were especially hit by the earthquake in a d . 141.HİS charities other than the building programs of the various cities included free distribution of grain, education for children, dowries for the girls of poor families and fimerals for the citizens. The cities which were endowed by his gifts included Patara,Xanthus, Pınara, Tlos, Telmessus, , Cadyanda, Oenoanda, , Bubon, Choma,Podalia, Arneae, Arycanda, Corydalla, Phellus, Antiphellus, ^»rlae, , Myra, Limyra,, Olympus, Phaselis and Acalissus. Opramoas had also served as a Lykiarch and had close family relations to Senate members in Rome. A detailed discussion of the deeds of Opramoas is given by Balland, 1979: 173-224. Also see the article by J. Coulton, 1987: 171-178. A list of Opramoas’ donations has been given by Broughton, 1938: 780. 118

Honorary inscription of Jason (son of Nikostratos) of Kyaneai'*;485

The inscription located near Kyaneai (Plate 62), written on a rock crop, gives the letters and decrees of fourteen Lycian cities which honor Jason, including Chôma and

Podalia. The inscription refers to a dispute between Jason and a Moles, which ends with the declaration of the emperor Antoninus Pius, who approves Jason and accuses

Moles. The letters suggest a date ranging from A.D 137 to 146.

(b)Archaeological settlements with epigraphical evidence from the Elmalı

Plain (Plate 63) Ovacık (plate 64)''**

The site has been identified as Late Roman/Early Medieval in date, from the surface pottery and a number of inscriptions found there. The badly robbed site has an approximate measurement of 300 meters in diameter, without any perimeter walls, displaying the remains of a number of buildings with finely dressed ashlar masonry, including two churches and a cemetery area with the remains of sarcophagi, among which at least one had an inscription. A fragmentary Dioskouroi relief was also spotted. The villagers also showed bronze coins which ranged from Early Empire to

Arkadius. Another site known at the Ovacık Asarı, some distance from the earlier one, has been identified as a hilltop settlement, which displays a walled enclosure of approximately 300 meters in diameter. The rough triangular shape of the enclosure

The inscription was first published by Heberdey and Kalinka in 1896, and is included in IGRR Tomus in.704,cap.l; also see Kolb, 1993: Tafel 4.3-4. Jason son of Nikostratos, similarly to Optamoas, had donated money for a number of buildings, which includes a portico, theater and the beautification of an Apollo Temple. Compared with Opramoas his gifts remain limited, but it is «fignifirant that both of the wealthy donors come from lesser known, smaller towns of Lycia and possibly have accumulated their wealth from the countryside, rather than the main metropoleis of the For information regarding Jason see the appenrUx in Zimmermarm, 1992: 252-270, and the article by Coulton, 1987b: 171-178, where he makes a comparison between Opramoas and Jason, and gives a general review of other donors of the time. The site known as the Ovacık öreni is located approx 40 km. north-east of the town of Elmalı, 4 km. to the south of the modem village of Ovacık, on the northern slopes of the Beyda^lari at an elevation of 1500 meters. The site was first visited by Schönbom and F.Von Luschan in 1882 and refened to in Reisen II: 164 A more detailed survey of the site was done by Harrison, 1979b: 235- 237; 1979c: 529-531; 1980: 112-114; 1981a: 199-200; 1981b: 384; 1983: 75-77. 119

exhibits a tower-like projection to its south and narrow oblong rooms on three sides

of the enclosure, looking into the central court. The function of the building has not

been identified, but suggestions have been made for its use as a military installation, a

refuge place or a monastery.

From the site of Ovacık comes the lower part of a stele, inscribed on all four

sides.The block had some 6 lines of inscription, which starts with a letter fi"om the

Duke of the Guard-posts to a certain Hermaios, concerning a selected group of young

men. The inscription on its less well preserved sides gives reference twice to

Termessos and once to Cremna.The best preserved face appears to be a part of a

municipal resolution, which decides between the ones who shall live in the , and the ones who shall act as the guard, among whom one is stated to be a well-bom brigand chaser. Among the ones who were allowed to stay in the polis was a man who had sent grain supplies to the guard post. The inscription was dated before a d . 367, due to the title used to define the duke, as the ‘SiotarinoTaToi’.''®®

A second inscription (Appendix FI) is a long oblong base with two columns of inscriptions. One column of the base refers to a letter sent by the praefectus Valerius

Euthios to the village-head man Kliortes, and the second column is a Decree of the

Council and People who honour the same person, Kliortes.The inscriptions have been dated between the years of a d . 324-337, based on the Praefectus Valerius Euethios, known to have served in a certain post in Egypt between a d . 302-304 during the reign of Constantinus I. ^ The inscription has been pointed out as significant by B.

Iplikçioğlu, for giving evidence for the continuation of the Imperial Cult, well into the fourth century in the region. This is suggested by the person who is honoured by the

^ Harrison, 1976: 530-531. Ibid., 531, see footnote 17. The title ‘Diasdemotatoi’ for a Duke, attested in the inscription was in usage till 361 A D ., with the earliest usage of the title ‘Dux’ in 364/7 A D ‘^'Iplikcioglu, 1992: 12-13. 120

inscription as having served as the high-priest of the Imperial Cult, under

Constantinus I.

Bayındır-Çağı Itemeller, Y arangedig r;490

A statue base and a column have been located from a church, which were used as spolia. The inscriptions have been dated to the early and mid third century ad.. Both of the inscriptions give evidence for personalities with an ethmcon of Terponelli

‘TepTioveXlEis' (Appendix F2a,b). A further evidence for a settlement named as

Terponelli is a border mark, located on a rocky outcrop at Yarangedigi, 5 km. north­ west of Bayındır. The border mark refers to the common border of Philips and

Terponelli (Appendix F2c).

Asar Tepe''®'

Traces of an ancient settlement have been reported to the north of the location of the inscriptions, but no detailed study has been carried out. Two inscriptions on limestone columns published by B. Iplik9ioglu (Appendix F3) are of fimerary character and come from an area that has been pointed out as a possible necropolis.

They have been located within the collapsed remains of round burial structures.

Semayiik

The Early Bronze Age site of Semayuk has yielded a dedicatory inscription, fragmentary in nature (Appendix F4), pointing to the habitation of the site during the

Roman period.

Çağıltemeller is located 750 meters south-west of the village of Bayındır, known for its tumuli burials. A chapel which was excavated by the Antalya Museum in the vicinity of Tumuli C and D, has given evitknce of two inscriptions. ''®' The site of Asartepe is located 13 km. north-east of Elm alı, and 2 km. to the east of Yarangedigi, where the border stone of the Terponelli had been located. 121

G o k p in ar A fragmentary inscription has been located from Gokpinar, which has shown habitation during the Chalcolithic, Bronze Ages and Iron Age (Appendix F5).

Kûçük-Büyük Sogle''^^ The mound has been identified as the classical city of Podalia, instead of the site by the Avian Lake, by Bean/” The mound settlement of Sogle points to a continuous settlement from the third millenium B.C. onwards, with a concentration of Roman and

Medieval pottery. A large number of inscribed and sculptured stones have been located from the mound, and from the villages, used as spoha (Appendbc F6).

Most of the inscriptions are dedicatory in character. Noteworthy are a limestone block with the Dioskouroi figures and the veiled goddess, located in the school of the village, which is said to have come from the mound (Appendix F6e), and an inscription consisting of 15 lines, which gives the list of the names of local men who have served as soldiers of Legion II. Bean has suggested the Legio II Parthica which was created by Septimus Severus, and on several occasions had accompanied emperors to the East. After it had been separated from the garrison of Rome in the fourth century, it was stationed in Mesopotamia. The names recorded in this inscription are purely Greek or Anatolian, and do not point to any Roman citizens.

Only some of the names have patrynomics. Bean suggests a date of fourth or fifth century, on the evidence of the nature of the names and character of the script

(Appendix F6g). It is significant to see this kind of an inscription from the Elmalı

Plain, since mountainous regions were considered as a potential resource for raising armies, during the Roman period.“”

The villages of Küçük Söğle and Büyük Söğle are located ca. 13 km. to the southwest of the town of Elmalı. There is a mound site which is located right in the cemetery grounds and entrance of the village of Küçük Söğle. The sites of Söğle were visited by F. von Luschan and E. Petersen Reisen n 163-164, andby Schönbom in 1841. Bean, 1%8: 158-163. Whifly, 1995:104. Also see Milner, 1993: 3-4, regarding the preference for country men as soldiers in the Roman army. 122

Balıklar Dağı - Ördek Beleni''®^

A necropolis site consisting of a number of smashed sarchopagi has been identified along the north foot of the passage (Plate 65 A). The existence of a Roman road has been also noted in the reports of S. Ozaner as part the Hacimusalar Project. Two of the inscriptions on sarchophagi have been included in Appendix F7.

Kapikaya"®^ The inscription (Appendix F8) praises the deeds of a person named ‘OiX-oX-oye’, who is assumed to be a freed slave and the chief huntsman on a private estate. The name Philologus is a common name given to freed slaves, and the missing patronymic also suggests a freedmen.'*^ The title ‘apxiKuvriyov’ has been suggested by the author to refer to a chief huntsman, and the decoration of the relief with various animals of pray would confirm this fact (Plate 65 B). The relief has been described by

Bean as follows; the main relief, accompanied by inscriptions, shows a stele with pediment. It is much weathered, the only identifiable figures being an eagle in the pediment, in the middle panel a boar (though also identified as an ox). On the rock- face to the right are two reliefs one above the other; the upper seems to depict a rider, the lower (very indistinct) apparently a man in pursuit of an animal. To the left of the stele are four small panels, quite plain, and between the upper panel and the stele there is what appears to be a man in energetic motion.

The rite is located 9 km. west-southwest of Elm alı, on a hill which rises from the ground of the plain. The hill, consisting of two peaks, gives way to a passage which connects the peaks known as ‘ördek Beleni’ or the Duck Pass. The Kapikaya inscription with its relief is located approximately 18 km. to the west of the town of Elm alı, on the road betwen Eskihisar and GüğüAf'apraklı. The rock relief had been noted since the earliest travellers to the region, see Fellows, 1852; 233; Spratt and Forbes, 1847: 284; Reisen n 225. More recently the inscription was studied and publish^ by Bean, 1971; 25, no.42. Bean, 1971: 25 123

GûğûA^apraklı''*’

The village had been identified as the ancient site of Orpeeni, due to the reference given in the inscription of a sarcophagus. The village displays a number of rock-reliefs in the cliffs above the village and pieces of architectural blocks and sculpture in the village itself .The site has been visited since the earliest travellers. Some rock reliefs accompanied by inscriptions have been published by B.Pace*“* and G.F.Hill who has published the inscriptions from the copies of Daniell and Fellows“ ‘,and republished by

G.E.Bean.*®"

Three of the inscriptions (Appendbc F9 a,b,c) carved on tall rock-cut steles, dated to the Early Imperial period are located next to each other, at a dominating view overlooking the village, the passage and a part of the Elmalı Plain. The central relief displays a crouching bull, which the local name of the cliff follows as the ‘B uzağı

K ayası’ (Plate 66A). The line of the inscription gives us the name o f the person,

Trokondas, son Mideis, grandson of Oxadeis. Next to it is another high stele, surmounted with a vase, where also the name of the wife of Trokondas is given as

Armastan. A third stele, which is located slightly recessed to the other side of the stele with the crouching bull, is totally smashed, and only few letters have survived.

Two other adjacent rock-cut reliefs with aediculae and inscriptions have been located in the cliffs above the village. One of the inscriptions (Appendix F9d) in an aedicula refers to the goddess Artemis Lagbene. The relief accompanying the inscription does not present the Goddess, but a trophy of shield, breastplate, helmet and spear, with a boar below. The boar seems to be a common attribute of the

Goddess. The second rock-cut relief (Appendix F9e) mentions the ‘ Geois aypiois’

All of these sites are located at the very northwest tip of the plain, which gives the passage over the Gtigtibeli, into the next upland plain of Seki, the ancient region of Kibyratis. The sites have been briefly discussed on pg. 31-32. Pace Annuario 3,1916-1920: 64, no.62-66; 2 680-682,684-685. Hill, 1915: 121, P.302, no.7,8a,8b. ^Bean, 1971: 26-28, no.43-48. 124

referred to also as the ‘trinity of deities’in an aedicula.The relief represents three male deities with short tunics, each carrying a double axe over their shoulder, with three snakes at their sides (Plate 66B).

The last inscription (Appendix F9f) gives evidence for the ancient name of the site. The representation of three figures in a banquet scene on a sarcophagus with a crested lid, gives the name of a person called Armonanis, formerly a native of

Elbessus, who has later settled in Orpeëni, which is taken to be the ancient name for the site.

Both Orpeëni and Elbessus have been recently identified within the city boundaries of Oenoanda, rather than Choma“\ at least during the Hadrianic period. The evidence comes from the Demosthenes inscription of Oenoanda, which Worrle**^ has used in discussing the eastern borders of Oenoanda. He suggests that the recently drained

Karagöl constituted a natural border to the city of Chôma, which occupied the southern part of the plain. The triad figures which are so abundantly represented in the

Kibyratic region are found to occur only at Güğü in the Elmalı Plain, pointing to a possible cultural connection.

Hacimusalar, Sarılar villages^^

The largest mound on the plain, which gives evidence as early as the Neolithic and

Chalcolithic periods. Early Bronze Age and Iron Ages, and with substantial remains in

^ Naour, 1980: 7. ^ WOrrle, 1988: 47. The Demosthenes inscription gives the name of a number of villages of Oenoanda, which participate in the festivals. Lines 73 and 75 give the names of OpTiEvva and EX-psaaoCT, respectively. As suggested by Wbrrle, during the period of Hadrian the whole massif of Elbis Dağ (which borders the west and north-west part of Elmalı Plain) was within the province border of Oenoanda. This has also been encountered from other sarcophagus inscriptions located in the area of Girdev Lake, which has been tentatively suggested to be Elbessus, by Bean, 1978: 175. The lion lid sarcophagi, which are abundantly found in Kibyra, and especially in Balboura, have been also found at the site, suggesting further cultural influence into the Kil^ratic region. Another inscription located in Yuva (located to the northwestern ridge of the plain), denotes the demos of Oenoanda.(CIG HI, 4380.o) Both of the villages are located to the south and close to the mound of Hacimusalar, which is located 18 km. south-west Elmalı. 125

the Roman and later periods, is located at the skirts of the villages. Large numbers of

inscribed and scuptured blocks have been located in the nearby villages of the mound

(Hacimusalar, Sarılar and Beyler). Sbc of the inscriptions published in M. Harrison

and G.E. Bean in JRS 57 (1967) 34-44, mention the name of the ancient city of

Chôma, identified with the mound (Appendix F 10). Most of the inscriptions are

dedicatory and fiineral in character.

Islamlar, Aytaş mevkii^

The cliffs behind the village, also known as the Aytaş Mevki were used as a

necropolis area, from the classical period. The Lycian type of rock-cut tombs had

been encountered by M. Mellink, and also referred to by Zahle during his extensive

study of the Lycian tombs, where he studied two of the tombs as evidence for fourth

century B.C. sculpture.*®’ The remains of a temple-like structure with an ante-room,

constructed of large limestone blocks, exists at one of the highest points of the cliffs, overlooking the plain.

A number of the rock reliefs have inscriptions. At Aytaş Mevkii a shield cut into the rocks is accompanied by a dedicatory inscription (Appendix FI lb, Plate 67 B).^®*

Also at Aytaş Mevki, 100 meters south of the temple-like building, the shape of a prism was carved on a double facade. The accompanying inscription of the ‘prism relief is a dedication fi'om a certain Molesis to the Gods who were worshipped in the same temple and altar (Appendix FI Ic). This part of the Elmalı Plain is also known for the worship of the 12-Gods, which are also referred to as the 12-Gods of Gömbe.

Can this be a possible reference to the same gods?^°® A third relief has been carved into the rocky outcrop in the form of a high stele, at the Aytaş Mevki, facing the

^ The village of Islamlar is 18 km. southwest of Elmalı, on the way to Akçay and Gömbe, leading to the Sinekçibel Pass. See discussion pg. 90-91. ^ The nature of the shield, which resembles a fiill-moon has led the locals to define the vicinity as ‘Aytaş’ or the moon stone. ^ Gömbe is 10 km. south of Islamlar. A discussion of the cult is given on pg. 128-29. 126

village of Islamlar (Appendix FI Id). A limestone block with the depiction of the

Dioskouroi flanking the veiled goddess, has also been located in the village of Islamlar

(Appendix File, Plate 67 A).

Gömbe''“

3.5 km.south-east of Gömbe at Çayboğazı mevkii, near the village of Kapaklı a milestone dated to the year of a.d. 253 has been found (Appendix F 12). The milestone located near one of the important passages leading to the Lycian coast, can be taken as evidence that the Elmalı Plain was part of the network of Roman roads in Lycia.

Gömbe is also known for the cult of the 12-Gods.

Tekke'"

The Constantinian milestone (Append« F13), dated to the first half of the fourth century was found near the most frequently used passage from Elmalı Plain to the coast of Lycia. The milestone’s proximity both to the passage and the Hacimusalar mound points to the existence of Roman roads connecting Chôma with other centers, especially on the Lycian coast.

2.6.3 Numismatic evidence

During the Roman period, in the double province of Lycia-Pamphylia, whereas

Pamphylian cities continued to possess the right of minting their own coinage, the right of minting their own coinage was taken away fi’om the Lycian cities as soon as it became an Imperial province under Claudius. The right of minting their own coinage was granted only during the reign of the emperor Gordianos EQ, during

A mound settlement in the vicinity has been identified with the ancient city of Komba. See herein footnote 418 for fiuther discussion. The village of Tekke is five km. southeast of the Hacimusalar mound, on the way to the Avian Pass, which is one of the important passages which connected the plain to coastal Lycia, via the ancient cities of Arycanda and Limyra. Aulock, 1974: 14-15. 127

whose reign (a.d. 238-244), 20 Lycian cities, including Chôma and Podalia, were given the right to mint their own coinage. The privilage was taken away by the succeeding emperor Philip the Arab.*’^

(a) Gordianus Ш coins of Chôma (Plate 61 B)

Gordianus Ш is shown wearing a laurel wreath, with the inscription of AYT K A IM

ANT rOPAIANOC CE (obverse); the rider god Kakasbos, wearing a cap, a short tunic and a flying chlamys over his shoulder is depicted with a club in his raised right hand, with the inscription o f ΧΩΜΑΤΕΙΤΩΝ (reverse).’’^' Another coin depicts the Emperor with a crown o f projecting sun rays, with the inscription AY К M ANT roPAIANOC CEB

(obverse); the standing Goddess Tyche with her polos head-dress, comocopia and a steering rod, with the inscription ΧΩΜΑΤΕΙΤΩΝ (reverse).*’*

(b) Gordianus Ш coins of Podalia (Plate 61 C)

Gordianus Ш is depicted with a laurel crest, with the inscription AY KAI M ANT roPAlANOC CEB (obverse); the God Apollo is standing to the right, with a cloth wrapped around his left arm and lower body, holding a plektron in his right, and a lyra in his left hand, resting on a pedestal, with the inscription ΠOΔAΛЮTΩN (reverse).*’*

Another coin depicts standing, wearing a helmet and breastplate, with his right hand resting on a spear, and holding a shield in his left hand, with the inscription

ΠΟΔΑΛΙΩΤΩΝ.*’^

*’* Ibid.,22 *’^ Aulock, 1974:60, no.54, Tafel 4. *’* Ibid, 60, no.55, Tafel 4. *’* Ibid, 79, no.282,283, Tafel 15. *’■' Ibid, 79, no.284, Tafel 15. 128

2.6.4 Local Cults attested from the Elmalı Plain

Certain cult practices attested from the Elmalı Plain indicate that the inhabitants of the plain were a part of the cultural sphere of both the upland regions of Kibyra,

Kabalia and Pisidia, and coastal Lycia. The Elmalı Plain in this respect can be viewed as a transitional region. The discussion of the M owing cult practices attested from the region (though the evidence for some of them is very fragmentary) aims to give an insight to the cultural sphere of the regions (Plate 70).

(a) The 12 gods518

A total of eight votive fragments of the ‘12 gods’, ‘the Dodeka Theoi’, have been encountered from the Elmalı Plain, five of which have been reported to have come from Gömbe (Plate 68 A), two from the town of Elmalı itself (Plate 69) and one from the village of Ahlatlikoy, near Akçay (Plate 68 B).*'® The relief plaques have been found in a range of other places, including the island of Lesbos in the Aegean, but mainly they have been encountered in coastal and central Lycia, eg. Kaş, Üçağız,

Finike, the island Megiste, Fethiye, Demre, Salardis (west of ), Sarılar (west of

Kyaneai) and as far as Burdur and Korkuteli in the interior plateau.

The votive reliefs are usually carved on thin local limestone plaques of 3-9 cm thickness, which have been roughly worked and the reverse usually left unworked.

They all depict a similar scene, on two rows, the upper row depicting 12 identical gods in frontal position, usually equiped with spears and sometimes also with

The first discovery of the votive fragments of the 12 gods, was by Lushan and Petersen Reisen II: 157. The first extensive stutfy of the votive reliefs was done by Weinreich, 1913, and with the most recent work, by Freyer-Schauenburg, 1994: 29-44, a conqjlete list of the votive reliefs has been given, with a detailed study of the style and iconography of the reliefs, which have been mainly mouped into five stylistical groups. Fieyer-Schauenburg, 1994, see the list at the very end The Elmalı and Ahlatlıköy reliefs are today in Antalya, but the location of the Gömbe reliefs is unknown. 129

shields/’^® A single figure, more prominent than the others, is always depicted in the

center of the other gods. The central figure is usually a male, but on some occasions a

female“ ’, usually standing under a cenotaph like roofed structure, which is sometimes

crudely decorated with geometrical figures like triangles and dot in circle motifs.“^ In

the lower row are 12 dogs, again with a central figure, of smaller size than the figures

above. A Greek inscription, which appears between the two rows and beneath the

second row gives the name ‘AoSeKa Oeoi?’ , ‘to the 12 gods’, and the name of the

dedicant of the inscription.

The original cult center of the 12 gods has not been located, nor the individual

names of the gods.“^ The gods have been stated to be of native Anatolian origin,

which possibly has its roots in the Bronze Ages“ ^ Their names have been

encountered ¡from the Xanthos Inscribed Stele, which indicates that the stele was set

up in honour of the 12 gods.^“ As suggested by Bryce, the fact that a very important

stele had been dedicated to their name suggests that they had a prominent place in

Lycia, as early as the fifth century b.c.“ ® The relief plaques have been generally dated

to the third century

Ibid.,52-60, for a detailed discussion of their garments and war equipment. The name of the central figure has been given on four occasions. From a relief plaque from Kaş today in Antalya Museum (Bean, 1958. 77-78, no.96).The relief plaque depicts a female, instead of the usual male, in the centeral position. The accompanying inscription shows the dedication to be to Artemis, the 12 gods and their fether. With this reference the central male figure has been also identified as the fether of the 12 gods. Three more reliefs have given the nam(> ‘Artemis Kynegetis’ these located in Finike, Museum and one near Kyaneai. See Zimmermann’s article in Kolbe, 1993; 146, Taf. 28,2. Freyer-Schauenburg, 1994; 63-68, for the architectural details of the plaques. The gods are often referred to as „The 12 Gods of Komba“ as the result of the very first fragments reported by Petersen and von Luschan fi-om Gömbe. An original cult place in Phellos, Central Lvcia has been suggested by Frei ANRW 11, 1990; 1830,1833 (Kolbe, 1993; 146). Though in an inscription located by Bean, 1978; 81, used as spoUa in the mosque of the village of Dodurga, in the vicinity of the ancient site of , the names of 12 gods have been given, among which were the more well known ones eg. Zeus,.^x>llo,Artemis, , .The inscription has been taV<»n as a possible evidence for the names of the gods. Gates, 1974; 165-166. They have been affiliated with the 12 underworld deities of the Hittites, ducted in the open air sanctuary of Yazilikaya, Boğazköy. ^^TAM 1.44; Demargne, 1958; 82; and as the reference has been taken from Bryce, 1986; 97. TL 44, lines 2 1-22, of the Greek version of the stele. Bryce, 1986; 180. B^n, 1978; 159; Fr^er-Schauenburg, 1994;69-88, especially see pg. 88. 130

(b) Kakasbos,528

No Kakasbos relief has been encountered in the Elmalı Plain. This is unusual,

because the coins of Chôma bear the image of the god as early as the Hellenistic

period, and then in the third century ac (Plate 61 A, B). The rider god Kakasbos has

been well attested through dedicatory inscriptions and relief depictions, in Lycia and

Southwestern Pisidia.""“ It has been considered as a cult widely distributed in

Southwestern Anatolia, especially around Telmessus, and the Kibyratic uplands north

of Telmessus, with an example also from Rhodes. The god, who is always depicted

with a club in his right hand, has been identified with the god Herakles, and some of

the dedicatory inscriptions have referred to Herakles instead of the rider god

Kakasbos. Though the god has also been affiliated with the Greek god Sozon and the

Thracian ridergod, L. Robert in his article denies these anologies and debates on the side of Bronze Age origins for the god, and the god being a native Anatolian deity.

From the linguistic point of view, three further forms of the name have been attested

KaKa0ıP(öi,KaKao|3aı and KaKaaPei as well as another god depicted in the same style, located in Telmessus, mentioned under another name TpiKooPoç, which has also been identified with the same god. The Lykian word esbe- could be equal to asba- (meaning horse). And kakasbos- to the fast?or strong? horse or rider.^. 531

Robert, 1946: 38-73 Though one limestone stele from Elmalı (Sekiköy) has been mentioned in the Doctoral Dissertation on the Kakasbos Figures by Delemen, 1993: 172, no. 111. Sekiköy is located in the Seki Plain, west of Elmalı Plain. Ibid Most of the examples are located in the Antalya, Fethiye, Burdur and Izmir Museums. They have a wide distribution area as far as Eskişehir and Konya in Central Anatolia, Sardis and Izmir in Western Anatolia. ^^’Neumann, 1979:265. 131

(c) Dioskouroi and Helene,532

Dedications to the Dioskouroi have been attested from three sites in the Ehnali

Plain these being Ovacık, Sögle and Islamlar (Plate 67 A). The depicted figures are usually two riders, which flank a draped, veiled, frontal, statuesque female figure. Her name has never been mentioned, with the only exception of a coin from Akalissos which has the same depiction of the riders flankmg a veiled goddess, who has been attributed to Helena. But some of her depictions have been accompanied with the attributions of the moon goddess, which has caused some scholars to believe that she is the Anatolian Artemis. The cult of the Dioskouroi has been abundantly represented in the regions of Lycia, Phrygia, Pisidia, Kabalia, Kibyratis, Milyas as well as the

Elmalı Plain. The Dioskouroi have been located at a number of sites including

Telmessus (Fethiye), Korydalla () in Lycia; Oenoanda, Y azir from the vicinity of Balboura in Kibyratis; Dengereköy, Çaltılar, Alifahrettin, K izilbel,

Tyriaion (Kozagaci), Kayabaş, Çıvgılar in Kabalia; Macun Asarı and Isinda

(Korkuteli) north-east of Elmalı Plain; Termessos, and near Lake Kestel in Pisidia.

Their depictions besides on the stone relief have been attested on Lycian coinage from

Akalissos as stated above. The Dioskouroi reliefs are geographically well distributed in the eastern Mediterranean, and have been found in Sparta, and as far as Gaul.

L. Robert argues for the roots of the cult to be in Anatolia and Near East.’^^ The rider god also appears in the region, depicted alone, or in pairs, triads and dozens without the Goddess in the center. These rider gods sometimes receive the names of Greek deities, eg. , but more often are referred to as the rider god Kakasbos. As suggested,*^ they may also represent a body of young men with military affiliations who are grouped around the Anatolian Goddess, since a similar idea also appears with the 12 gods, which in some cases flank a female figure, who has been in few instances associated with Artemis.

Robert, 1987: 553-583; Milner and Smith, 1994: 65-67. Ibid, 576-79. Milner and Smith, 1994:66-67. 132

(d) Trinity Godsi535

Only one depiction of the gods has been located from the plain, this being the

Güğü example (Appendix F 9e, Plate 66 B). The accompanying inscription of the

Gügü example has identified these gods as the θεοί άγριοι ‘wild gods’, which has

been pointed out by Alexander Polyhistor to be associated with a cult place on Mt.

Kragos, the modem Baba d ağ ı, in western Lycia (though no such relief has been

identified there). The so-called triad figures have been mainly identified in the

Kibyratis and Pisidia.’^

(e) Artemis Lagbene'537

Only one example of the depiction of the Goddess has been identified in Güğü

(Appendix F9d). They have been mainly attested from Korkuteli, Söğüt and the Seki

Plain. The most recent example from Korkuteli, depicts the goddess mounted on a boar and holding a crown and a torch. The cult centre for the goddess has been identified with the ancient city of , as the title of the goddess suggests. The site at Mevkii near K irkpinar has been identified with the ancient city of Lagbe 538

To summarize, the Roman and late Roman period for Northern Lycia and the

Elmalı Plain, with the peace and stability brought to the country under Roman mle, there is a general prosperity encountered both with the coastal regions and inland

Robert, 1946: 75-76; 1955: 5-12. Dedications to the gods have been encountered from other places, including Dont, Oenoanda, Balboura (Asar Tepe), Tyriaion, Kibyra (Horzum-Gölhisar), Y azir near Korkuteli, Osmankalfelar, and from the southern skirts of Bey Dağları, the site of Idebessus. Robert, 1987: 594-597; French, 1994: 88. The Goddess was first encountered on a dedicatory inscription found in Manay (Söğüt), during the mid 19th century. The inscription (CIG 3 .4318b)referrs to the dedication of Demos of Lagbe, to the goddess Artemis Lagijene Other dedications to the goddess have been encountered in the vicinity of (Denoanda, and in Korkuteli, by Robert. For Lagbe see the article by French, 1994: 87-92. 133

Lycia. The northern uplands have now become administratively connected to the

double province of Lycia-Pamphylia. The cities of Chôma, Podalia, and Komba, all

associated with the Elmalı Plain, are among the cities of Lycia, they have definitely

become a part of the Lycian league, which, though being pacified under Roman rule is

still functioning as an intermediary between the Empire and the member cities o f the

league. From the economic point of view, Lycia was possibly enjoying its most

prosperous and wealthy age, partly due to the strategically important location of its

harbours (Patara, Andriake etc.), and partly due to the rich hinterland, whose

resources eg. timber, olive oil, grain etc are now fully exploited. Central and northern

Lycia have become an integral part of the larger Lycian economy, their resources

brought down to certain collection centers, and shipped to more distant destinations.

The cities and villages of the Elmalı Plain with a high possibility have become a part

of the system as well, since the plain exhibits the largest arable land for coastal Lycia,

apart from the smaller valley plains of central Lycia. As the archaeological evidence

proves, the whole of Lycia was webbed in a network of roads during the Roman

period, which were possibly built over earlier Lycian roads. The evidence for the

Elmalı Plain being a part of the network, apart fi-om the traces of ancient roads, is

evident from the two milestones located along the two important passages of the

plain, which links it to the coast. The first milestone is from Gömbe, which is located

at the very southwestern extremity of the plain, and leads into the Sinekçibel passage

and further down to the Kasaba Plain, then reaches via a network o f roads either

Phellos (modem Kaş), and Patara, or over the river valley of Myra, down to

Andriake, one of the two important harbours of Lycia during the Roman period,

which evidently was a collection center for products of the uplands, as the large

Hadrianic horrea suggest.”® The second milestone is located in the village of Teke,

right before the Avian pass, which links the plain via the cities of Arycanda and

539 The lines of communications have been discussed by French, 1993: 87-90, especially from the Kasaba plain to coast, as well as its connection to the Elmalı Plain. Especially see pg. 89 for the function of the road connecting Elmalı Plain to the harbours of lower Lycia, and Andriake, and pg. 87 for traces of the road in the Sinekçibel passage. 134

Limyra to the coast, and is the best known passage to connect the uplands to the

coast. At this point I would like again to stress the important location of ancient

Chôma, identified as the Hacimusalar mound, which lies in the southern part of the

plain and right at the conjunction of both of the passages that connect the plain to the

coast.

No archaeological investigation of the Roman or later periods of the plain has been

carried out, apart fi’om the recent Hacimusalar mound excavations and survey of the

surrounding plain. The main studies done in the region have been limited to the

epigraphic surveys and the identification of the larger classical sites, which have been

encountered from literary sources. Unfortunately we don’t have evidence of typical classical sites, with civic buildings and centers, as has been encountered in other upland regions of Lycia, eg. the Kibyrian cities of Oenoanda, Balboura and Bubon

This may possibly be due to the nature of the evidence, all of the sites being mound settlements, and the intensive settlement of the region after the medieval period, with the infiltration of the Islamic groups into the Elmalı Plain. Epigraphic evidence, as well as surface evidence, indicated that the plain was abundantly populated during the

Roman period. How well the area was integrated into the Roman culture, and especially involved in the celebrations o f the Imperial cult, is again obscure, due to lack of evidence. With the evidence of the Opramoas inscription, we have the reference to an Augusteum in Chôma, built together with a stoa, from the funds received from the wealthy Lycian donor of Rhodiapolis. Epigraphic evidence from

Ovacık also points to the celebration of the imperial cult well into the fourth century.^'“ In contrast to the neighbouring upland region of Kibyra, we do not have the rich evidence for agonistic festivals with athletic or musical competitions, which also served for the celebration of the Imperial cult, as well as evidence for a local wealthy elite, which took the leading role in the organization o f these festivals. As it has been recently suggested from the agonistic inscription of Demosthenes, the city

540 See .^jpendix F 1. 135

boundaries of Oenoanda reached well into the Elmalı Plain, and constituted the area northwest of the Lake Karagöl, which has recently been dried out.

Sepulchral evidence and the evidence of votive reliefs indicates that the Elmalı

Plain is still a transitional zone between Lycia and the upland regions. The Lycian type of sarcophagi, with the characteristic convex ridged lid have been encountered mainly to the south of the plain, and survey of the Hacimusalar mound has revealed a number of these sarcophagi, scattered around the mound site (Plate 71 A). The lion lid sarchophagi^’, which are typical for the Kibyratic uplands,especially encountered in Balboura, and Oenoanda, are virtually unknown from the Elmali Plain.^''^ They have been mainly encountered around lake Girdev (Plate 71 B), in the mountains southwest, and west of the Elmalı Plain, at Macun A sarı, and the site in the mountains north of Orpeeni (modem village of Güğü), both of which sites probably belonged to the Oenoandian territory . The votive reliefs of the triad figures,

Dioskouroi and Artemis of Lagbene, which are typical again for the upland regions of

Kibyra, Kabalia and Pisidia have been attested from the Elmalı Plain, but are again virtually few and mainly coming from the northern part o f the plain. The Kakasbos figure has been attested only from the numismatic evidence. The 12 god votive dedications have been located in plenty from the southern part of the plain, and associated with the ancient Komba. The dedicatory reliefs of these gods show a concentrated distribution in central Lycia, and around Antiphellos (modem town of

Kaş) and to the east of coastal Lycia. The cultural influences attested in the plain seem to be closely related with the geographical setting of the Elmalı Plain and with its lines of communications with the neighbouring territories, eg. central and coastal

Lycia, Kibyra, Kabalia and Pisidia.

Money, 1990:29-53. Ibid, 38-39; Coulton, 1980: 84. Ibid,38-39. 136

The effect of the crisis years and the later transformations brought in with the rise

of in late antiquity, again lack evidence for any historical reconstructions

for the plain. The opposing views of Harrison and Foss, regarding a shift o f

populations or a complete depopulation of the upland Lycia, each lacks concrete

evidence for the argument. Although archaeological and epigraphic evidence of the

Ovacık hilltop site has pointed to a re-establishment of a mountain site, possibly due

to increasing brigand problems, it would be dangerous to conclude that there was a

shift of population to mountain sites in general. The evidence from the plain, though

mainly dependent on pottery finds, points to an abundant amount of late Roman

wares, which is also attested from the preliminary results of the Hacimusalar

excavation. As it has been attested from coastal Lycia, where both urban and rural life

continued to prosper in late antiquity, the Kibyratic cities of Oenoanda and Balboura

point to shrinking of the sites behind late Roman fortification walls, though life seems

to have continued without a break well into the medieval period, in Northern Lycia.

As suggested by S. Mtchell*^, with the resihence of the countryside, rural life

probably continued without serious interruptions through the crisis years. It probably

made any difference to the rural communities in the countyside if they were exploited

by local aristocrats, civic councillors, ofHcials or soldiers, since they had less to lose.

The abundant evidence of inscriptions generally attested in Asia Minor, dated between

the mid second century and the fourth century, suggests a continuity in the late

Roman period, and a similar picture is presented for the Elmalı Plain, at least as

attested from the epigraphic evidence, which seems to be quite abundant between the

suggested periods. Further systematic investigation through the regional surveys and

the Hacimusalar mound excavation will possibly throw more light upon the;issue.

^ MitcheU 1, 1993: 239. 137

2.7 THE EARLY CHRISTIAN AND THE BYZANTINE PERIODS 545

Lycia had become a separate province under Constantine, in A.D. 311-312, governed by Zipraeses, with Myra becoming the civil and ecclesiastical metropolis.

Under Theodosius H (408-450), the praeses was promoted to a comularis.^

The early Christian history, and the spreading of the new religion, is little known in

Lycia ’“'Trom the fragmentary evidence it has been understood that the turmoils of the early Christian period, which started in the third century and accelerated in the first half of the fourth century were also encountered in Lycia. These turmoils seem to settle down towards the end of the century, when churches and monastic complexes start to spread both in coastal and upland regions of Lycia.

During the religious controversies of the early Christian Church, Lycian bishops seem to have supported Arius, in the Council of Nicea in A.D. 325, which ended with the condemnation of Arianism.^"’That Lycian bishops took the side of the Arians against Orthodoxy, at least during the early periods, is apparent from two further accounts.In the Seleucian Council of 359, Lycian bishops once again supported the

545 Qjjg of the difficulties in stnictering the thesis was setting the chronological and historical time limits of the Late Roman and Byzantine eras. What may to the reader appear as chronological overlapping of events under different periods, is a decision based on presenting the archaeological, epigraphical or literary evidence in an unbroken and complete sequence. Foss, 1994: 2. A stucfy of the early Christian history of Lycia, based on literary sources was included in the work of Schultze, 1926: 188-209. A summary is included in the article by Harrison, 1963: 118-24, under ‘Historical Summary’ PF Snppl. XIII294. The earliest known bishop from Lycia is Methodius of Olympus, who was actually assassinated during the Diocletianic persecutions. During the persecutions, Nikandros, apperantly the first bishop of Nfyra, together with a prebyster, was also martyred. Other martyrs from the period were a certain Dioskurides and Crecens, a missionary from Myra. The turbulences of the early fourth century are also apparent from an inscription found at Arycanda. The context of the inscription (TAM П.785), dated to A.D. 311-312, gives the request of the Arycandians for imperial support from the emperors Galerius Valerius Flavius (Donstantinus and Valerius Licitmianus, against the confusions created by the Christiarts.The destruction of temples and statues must have taken рЬгр during this period. The famous bishop Nicholas of Myra, also lived during this period, and is associated with the destruction of the temple of Artemis Eleuthera in Myra, which he also used as a foundation for his church. Schulze, 1926: 190; Harrison, 1963: 119. 138

Arians. Another account comes from a letter sent in 375, by Basil of Caesareia, who is known for his strong view on the side of Orthodoxy. But during the Council of

Constantinople in A.D. 381, nine Lycian bishops give their votes on the side of

Orthodoxy, including the bishops of Chôma and Podalia, who are the earliest bishops attested from the bishopric centers of the Elmalı Plain. With the fifth century Lycia seems to have been brought into line with Orthodoxy.^^’ In the Councils of Ephesus

(A D 431) and (A.D. 451), and the related letter Epistola ad Leonem

(A D 457), Lycian bishops take the side of Orthodoxy, now against the Monophysite belief, which found its roots in Syria and Palestine.^“ Among the bishops’ names, bishops both from Chôma and Podalia appear. The siding of Lycian bishops Avith

Orthodoxy has been suggested to be controversial when viewed through the archaeological evidence. The early churches and monastic complexes in Lycia show an increasing influence from Palestine and Egypt, probably because o f the strong connections of Lycia to these regions, principally through sea trade.

The ecclesiastical sources of Ifierocles Synecdemus, and the Notitiae number 30-

35 bishop seats in Lycia, starting with the sbcth century and continuing well into the twelth century. The lists of cities given in the Notitiae remain quite constant throughout this period. Podalia, Chôma and Komba are included as bishop seats of

Lycia, under the metropolis of Myra, in most of the lists.

Archaeological evidence suggests that the classical cities of Lycia, which prospered during the Roman period in the first, second and partially in the third ****

Ibid. The letter, sent to the bishop of Iconium asks for an investigation into the bishops of Lycia, for the names of the bishops who could be safely approached for accepting Orthodoxy. The answer to the letter counts only the bishops of Corydalla, Phellus, Patara and Telmessus, with four presbyters from Limyra and Myra. *** Harrison, 1%3: 119. Ibid. Ibid,120. Harrison further suggests that this controversy m ^ have been based on the friction between monasticism (which was rising in the countryside and in general sided with the monophysite views ), and the ChurcL 139

centuries, continued to prosper throughout late antiquity, well into the sixth century.

The transformations brought in with late antiquity, with social and administrative

changes appearing under the new order of Christianity, became especially evident in

the cities. The public buildings of the Roman period started to deteroriate, because of

falling into disuse, new churches and monastic complexes were built over the earlier

public spaces and buildings, and the use of mortar and rubble, and the dependence on

spolias of the classical period are typical for this period. The change in the general

appearance of the cities has been taken as a sign of a declining urban setting, and a

shift of wealth and population from the city to the countryside and from the coast to

the uplands, sometime in the late antique/early Byzantine period by M. Harrison, on

evidence from his extensive surveys conducted in the hinterlands of central and

partially in northern Lycia.^^

In the Early Byzantine period Lycia continued to hold its important position in the

sea-trade, with its well sheltered harbours, and being on the main sea route between

the eastern Mediterranean, Alexandria and Constantinople, which the imperial grain

fleet was now serving. The sixth century has been pointed out as a flourishing period

for especially the countryside of Lycia, by literary and archaeological evidence.

The sixth century hagiography of Saint Nicholas of Sion gives us valuable

information regarding the topography of Lycia during the period. As stated by

Foss’*\ in the Vita of St. Nicholas one gets the impression that Lycian society and

'^Harrison, 1977; 11-15; 1979c: 525-531; 1980: 109-118; 1981b; 383-387. His idea of a decline in the uiban centers of Lycia, and a shift of population to the uplands has been disputed by later scholars, eg. Foss, 1994: 46, especially see footnote no. 185, and Hellenkemper 1993: 106.Though his valuable studies conducted in the uplands of central Lycia, has thrown li^ t upon the display of the growing wealth which took place in the countryside, especially accelerating into the sixth century, and has provided ample evidence for economical relationships between coastal and hinterland Lycia, both parts of which were well-off during the period. Harrison, 1%3: 117-151; Foss, 19%: 19-25, 334-337. Anrich, 1913-1917; Sevcenko, 1984; commands of Robert, 1955; 188-222· and Foss 19%· 303- 337. Foss, 1996: 334-337. 140

economy flourished, especiaUy in the form of vUlage life. Around 20 villages^“ were mentioned in the Vita, most of which have not been identified, and are unknown fi-om the Notitiae. They are located in central Lycia, between Myra and Ameai in the

Kasaba plain, but there is also reference to sites in the distant region of the Kibyratis

(eg. Oenoanda is mentioned on one occasion), and there is a possible reference to

Komba, in the Elmalı Plain. All of the mentioned viUages seem to possess their own churches, and be in close contact with their neighbouring vUlage societies, monasteries, as well as Myra, the metropoUtan seat. The income of these upland villages closely reUed on their revenues fi-om agricultural produce (eg. grain, vine, olive etc.) which they brought down to Myra, to be consumed by the metropolis itself or exported to Constantinople through its harbour in Andriake. It is also clear that, during the bubonic plague that hit the coast, Myra very much relied on the produce brought down from the hinterland.’’^ From the Vita we also learn that St. Nicholas had visited Palestine three times and Egypt once. Interesting hints regarding the sea- lanes and ports of Lycia are also given in the Vita.

The archaeological evidence seems to be closely related to the social and economic picture given to us through the Vita. The extensive survey of M. Harrison, starting in

1959 and continuing till the beginning of the 80s with interruptions, has shown that there is an abundant evidence of churches, monasteries, and village settlements with traces of agricultural activities especially in central Lycia, the Kasaba plain and the

Arycandus valley, and mainly in the mountain plains north of Myra.'560

Sevcenko, 1984: 18. Three territorial terms were used in the accounts, being ‘ χώρα, κώμη, and Ktopvov’, which have been translated as district, village and hamlet respectively. **^or the details of the story, see pg. 27-28. ^Harrison, 1960a: 26-27; 1960b: 25; 1%3: 117-151; 1972: 118-197. In his 1%3 article, Harrison presents with a compact list of churches, chapels and monastic buildings in the form of a catalogue. These include the domed shrine of Asardnii, a complex of ecclesiastic buildings at Alakilise (a basilica with an adjoining stucture of cross-in-square plan, two chapels and a rock-cut chapel with a triconch plan), the Dikmen churches (with the earlier being a triconch church), a chapel and church at Muskar, the Monastery church of Karabel (which has been identified with the church of Holy Sion, associated with St. Nicholas of Sion, on epigraphic evidence; see pg 130-136 of the article), the large triconch church at Alacahisar, the triconch church at Devekujoisu, the church and small basilica at Qi^nan, a small church at Karabel, the two churches at Ameai, the remains of a church 141

The tradional type o f church architecture in Lycia has been summarized by

Harrison as being the three-aisled basilican form, with an atrium, narthex and a single projecting apse, and covered with a wooden roof The timber-roofed basilican type of church is usually divided by piers in the counttyside, and by reused columns in the cities. Another canonical feature becomes the annexation of a small chapel, on the east end of one of the aisles The ashlar masonry of the early churches was succeeded by mortared nibble constniction, with a feeing of smaU stones and stucco. The intrusion of two new types of church in the sixth century has been suggested as pointing to the strong cultural and economical connections o f Lycia with Palestine and Egypt. One o f the new types of church, is the cross-in-square church, as at Alakilise, and the second is the triconch church,""’ The important location of Lycia between the Aegean and the

Levant, and its economy, which was dependent on the commerce of the eastern

Mediterranean, had placed it in close connection with and Alexandria, as

by Harrison, being places torn which the early Church architecture of

562 Lycia had derived its form."

With the sixth century an abundance of architectural sculpture is also to be

bserved especially with the interior decoration of the churches. M. Harrison has suggested the flourishing of a local sculpture workshop in the vicinity of Alakilise-

Muskar""" Three examples from the same workshop have been identified in the •

• r.ömbe basilica in the Demre gorge, a basilica and church in Kökbumu, a small chapel in *h Demre lagoon, a basihea at Gürses, five churches at Andriake and the churches of Sura (the ® ches of Andriake and Sura have been later restudied and published ly Feld and Peschlow in ^rchhardt ed., 1975; 398-425), and the churches of island and Ü çağız. m triconch churches being from Alakilise, Dikmen, Alacahisar and Karabel. Harrison, 1963; 148-149- The triconch church is constructed of an apse, with three walls in which each wall is broken by an apse. The architecture is familiar from Roman tomb and palace architecture, and was exploited in the fifth and early sixth century in Egypt and Palestine. Ibid., 151. Harrison, 1972; 187-198. A group of capitals have been attributed to the same atelier, coming from Alakilise, Muskar, Demre, Antalya and three examples fi’om the Elmalı Plain. The homogeneity of the group has been identified by Harrison, in the usage of local limestone, the same technical use of chisel and drill, and the reoccurrence of certain moti6 eg. the windblown acanthus. 142

Elmalı Plain. One of these was found in the school building of Tekke village.^ The other two are located in the village of Müğren^^ and the site of O vacık.^ The significance of the workshop, which has been identified as serving the ecclesiastic buildings in the countryside, has been stated by M. Harrison to point „to the shift of well-to-do patronage from the classical cities to the monastic countryside“.^®^

Archaeological evidence also points to the abundant usage of the countryside.

Although a date in the sixth century has been suggested, one needs to keep in mind that to obtain a definite date for ancient agricultural activities will be too optimistic.

Village compounds with traces of agricultural terracing and ancient field systems have been located around Karabel, Dikmen, Muskar Alakilise®®, Arifr®* and Limyra®™.

^ Ibid. 197, no.5, fig.20. The capital is decorated with two tiers of windblown acanthus, and a third row of ¿antous irâves, set oblique between volutes. Deeply drilled and chiseUed. Harrison, 1980:1 lO.The limestone capital is decorated with windblown acanthus, and an abacus with two opposed Maltese crosses, wreathed. Ibid, a small capital with windblown acanthus, carved in one piece with the column shaft. Tran^jorted to the local museum in Elmalı. 5®' Harrison, 1972: 198. Harrison, 1978: 9-10; 1979a: 205; 1979c: 526-529; 198b: 386-387; 1983: 75.The sixth century basilica, the CToss-in square church and the monastic buildings of Alakilise dedicated to the Archangel Gabriel, as given in an inscription, were first published by the Austrian expedition of R eisenll: 38-40, and studied by Rott, 1908: 316-324, and restudied by Harrison, 1963: 126-13. to the monastery complex, three more chapels, remains of ancient roads, and a whole village complex with ca. 30 houses, freestanding, of two storeys, some terraced into the hillside, were lo c a te d in the valley floor and on its slopes, as well as traces of ancient field ^ tem s. The houses had living quarters in the second storey, and most of them possesed a rock-cut winepress and some had cistern in their courtyards The site seems to be a new site, with no earlier occupation, and shows a homogeneous construction, identical to the first phase of the church, dated to ¿he sixth century. ®®® Harrison, 1979a: 205-206; 1981a: 198; 1981b: 384-386; 1979d: 13-17; 1980: 118. The medieval site of Arif was investigated between 1978-1980. The Early Byzantine ruins have been named after the closest village to the site, to distinguish it from the classic^ site of Arycanda, which lies only two km. to its south. The site has been enclosed (an area of ca. 2.5 hectares) with a fortification system, with towers flanking its northern side, which seem to be more strongly built.The inside of the enclosure has been clearly laid out, with a main street connecting the east and west gates, and subsidiary streets running in a north-south orientation. Public buildings including a bath, palace complex, a market place in the form of a small piazza, and three churches were identified. Most of the buildings appear to be private houses. A great amount of spolia has been used from the classical site of Arycanda.The uniformity of the site has suggested the plaimed construction of a single period site, which pottery evidence has dated to some time in the sixth century. The site was built to replace the classical site of Arycanda, which according to its excavator ProfiDr. C. Bayburtluoğlu, was abandoned in the sixth century. The shift in the settlement pattern has been taken as further evidence for the general shift in settlements and populations seen in the late antique/early medieval period in Lycia. The medieval site of Arif is estimated to have held a population of approximately 1000 people. The number has been suggested by C. Foss to be less than that of the classical ci^, and evidence for the depopulation of the uplands, with its inhabitants possibly attracted by the urban centers of the 143

The rich ecclesiastical hoard found in a buried context in Kumluca, is also dated to the sixth century/’* The hoard consists of over fifty objects, including liturgical vessels of patens, chalices, censers, amphorae, ewers, book covers, an asterisk and

ecclesiastical furnishings of lamps and polycandelas/’^ The inscriptions and stamps on

some of the objects have given the names of their clerical donors, and mention the

Church of Holy Sion, which has caused scholars tentatively to identify the intended destination of the hoard as the church of Karabel, as this had been identified by

Harrison with the Holy Church of Sion, built by St. Nicholas.*’^ The burial context of the hoard in which the objects were found carefully smashed and flattened, perhaps for the purpose of easy transportion, has suggested that they were destined for the melting pot. This could have been done by the hands of the clergy of the Church, for the purpose of raising cash in a crisis situation, or by a group of looters. The Arab raids of the mid-seventh century have been suggested as a possible reason for the event But other alternatives, eg. an earthquake or a possible period of drought, have

574 been also suggested.

With the seventh century a complete reversal of the continuous flourishing of the urban and rural life in Lycia is attested, and it marks the start of the so-called Dark

coastline. Foss, 1996: 24. His view, as stated earlier, contrasts with the view of Harrison, also in regard to the direction of demographic change. 57o®^cobsen, 1992:287-292. A survey was lead by the Austrian team working in Limyra, under R. Jacobsen, a**"*ng to throw light on the late antique/early medieval settlements around Limyra, an area of 16 km. along the flanks of the Gülmez dağı being investigated in summer 1990. Similar results have been achieved with the location of three sites at Keştik Plateau, vUlage of Asarönû and Gûlmezasarı. The sites possesed village settlements, with remains of fortification ^rstem, and a church or basilica. Field terracing, cisterns, oil and winepresses, were also spotted with the domestic architecture. Sculptural pieces found at the sites dated the site to the late fifth or sixth centuries. 5·'! Boyd and Mango eds., 1992: 5-18. The treasure was located in the village of Kumluca, near a hill known as Büyük Asar, identified with ancient Korydalla. The treasure is housed in the Antalya Museum, and a part of it for stutfy purposes in Dumbarton Oaks Washington. 5^2 Ibid, 9. Ibid, 8-9; Harrison, 1%3: 132-35,150. Ibid, 7. 144

Ages, a period defined as the seventh to mid-ninth century. Neither coastal nor inland

Lycia shows any signs of urban flourishing, except for short periods of recovery under

the Macedonian Dynasty and the Comnenids, after this period. *^^The Dark Ages

become evident in the archaeological data, where in all the larger cities of coastal

Lycia there appears to be a considerable shrinkage, together with a dimimshing

population, and settlements shifting to hilltop fortresses.* The crisis of the seventh

century has been generally attributed to the increasing instability brought into the

empire through the continuous attacks of the Persians and later the Arabs , who

terrorized the country and main sea routes of the eastern Mediterranean well into the

tenth and eleventh centuries. It has also been pointed out, however, that other factors

may have played a role in bringing in the Dark Ages, and earlier clues to a degrading

economy and increasing problems in the sixth century do exist. 578

The Persian attacks of the sixth century have not been literally or archaeologically

accounted for in Lycia, but possibly the Lycian shore was threatened, especially

Foss, 1994; 48-51; 19%: 27-30. Foss' 1994: 2-4, where he also discusses separately the evidence from Xanthos, Telmessus, Patara, Myra and Limyra for the depopulation and contraction of the sites in the seventh century. See also Foss, 1996: 25-29. Foss, 1996; 26, in general ddiates for the impact of Arabs as a cause of the Dark Ages. Hellenkemper, 1993: 99-100. Hellenkemper argues for a number of causes which may have laid the ground and played a combined role in the crisis of the seventh century. He points to certain clues from the sixth century accounts of the Vita of Nicholas, which refers to a famine two years after the well known bubonic plague which had especially hit coastal Lycia, and probably was responsible for a considerable loss of population.The miracles of St. Nicholas include his regular visits to remote villages in the countryside, which he provides with grain, oil and wine, during the famine years, as well as his miracle, which took place in Myra, where after the distrubution of the grain from the imperial fleet that had arrived from Egypt, there remains no grain for Constantinople, and miraculously he fills the grain depots of the ship with grain. Another reason for the degradation of the Lycian economy, as suggested by Hellenkemper, is the economical impoverishment and destruction in Syria and Palestine brought with file Persian attacks starting in the mid sixth century, that were responsible in the breaking of the trade chain in the eastern Mediterranean, on which the economy of Lycia also relied. A third reason has been suggested to be a series of environmental factors, such as increasing silting of the harbours of Patara and Andriake, and the plain of Myra. The accelerating silting of the coast, besides the tectonic movements (eg. sinking of the land masses), has also been suggested as a result of the increasing erosion which may have been caused by the extensive exploitation of the mountains through terracing and the deforestation because of human activity, especially starting in the Roman period. The impoverishment of the soil quality in the plains may have played a role in impoverishing production and economy. Rather than attributing the cause of the Dark Ages to the impact of the Arabs, he favors a coihbination of reasons, which gives a more plausible explanation for the crisis. 145

during their sen campaigns against Samos and Rhodes. As discussed above, the

Lycian economy may have indirectly been affected by the disturbances caused by the

Persians in Syria and Palestine, which later spread through the eastern Mediterranean.

Lycia was part of the Cibyrraeotic Theme in 732, which included both Pamphylia and Lycia under one strategus.^^° During the reign of Herachus, or Constans II, Asia

Minor was gradually reorganized into military small-holdings known as Themata, which were ruled under one commander, the strategus, who was both the civil and military governor of the particular Theme.

The first Arab impact on the Byzantine Empire arrived via land routes over Cilicia, and hit central Anatolia, especially , in 644 .The first wave of Arab sea expeditions under Mu’aviya took place in 649 and 653 , and mainly hit Cyprus.’«'

The second expedition o f653 had the main aim of looting the coasts of Cos, Crete and Rhodes, and also possibly included the coast of Lycia. This would have been the period where the inhabitants of coastal Lycia migrated inland in order to avoid the

Arab raids, as suggested by H. Hellenkempler.’«« Following the attacks of the Arab fleet, the sea battle of Phoinix (modem town of Finike) occured in 655, between the

Byzantine fleet and the forces of Mu’aviaya. This famous sea battle, which is also known as the ‘battle of masts’, ended with the defeat of the Byzantine fleet, leaving

Foss, 1994: 49; 19%: 26. During 602-628. Harrison, 1%3: 121;Morgenstem, 1983; 21. Oxford Dictionary of Byzantium: 2034-203. The stategoi would function also as civil administrators and judges. The Themes, as they had been originally formed with the purpose of strengthening the defense of the empire against the Arabs and Bulgars, later became centers of revolts, with strategoi pretending for the throne, and weakening the power of the central government. As stated by Treadgold, 1995; 23, the Themes were simply mobile armies of the previous periods, which were now stationed in specific regions. Their foundation has been attributed to the reign of Constans П. The system of themes existed till the end of the empire. Hellenkemper, 1993:100. Hellenkemper also discusses the literary evidence which gives accounts of the extent of the raids which had taken place in early 649, in Cyprus, ending with the death or slavery of ca. 120 000 people, the news of which had possibly terrified the inhabitants of the coastal regions. Ibid., 100. 146

the coastline totally unprotected against the exploitations and raids of the Arabs, and opened the sea routes into the Aegean and up to Constantinople.^*^

The second wave of Arab expeditions came a decade later, and resulted in the siege of Constantinople (663-677). During this period Lycian harbours probably continued to be a stopping off point for the Arabs on their way to Constantinople. An

Arab fleet in search of timber was reported in 717, in the eastern Mediterranean.

Theophanes gives accounts of the Arab fleet, which had sailed fi-om Alexandria to

Phoinix (in Phoenicem), with the aim of finding cedar wood. However, this reference to Phoinix (other than the modem Finike) has also been suggested to be the Phonician coast which is also well known for its cedar forests, throughout early history.**’ If we accept the identification of with Finike, we would have evidence for the Arab exploitation of inland Lycia. In 677 and 717 the Arab fleet was defeated by the

Byzantine fleet, and in 790 the Cibyrraeotic strategus Theophilus, sailing fi-om Myra, defeated the Arab fleet in the gulf of Attalia. In 808, one of the admirals of Hamn ar-

Rasid on his way fiom Rhodes stopped at Myra/Andriake, where the church of St.

Nicholas was looted by the Arabs. Also in 842, a whole Arab fleet was shipwrecked at Cape Gelidonya, due to bad weather.The decrease of the power of the Abbasid

Caliphate in 840 brought a relief to the Byzantine Empire, though the threat of Arab fleets and piracy continued to terrorize the coast well into the 11* century.*S86 ***

584 ibid.,101-102. Hellenkempler gives a detailed discussion of the sea battle and its consequences, especially for the exploitation of the hinterland Lycia for its timber, which was so valuable for the maintenance of the Arabic fleet. Foss, 1994: 49, gives the evidence for the total disappearing of the maritime culture in Lycia, where small ports eg. Lebissos, the island of Gemile, ^)erlae etc., were totally deserted in the seventh centuries, with some of them eg. L^issos no habitation occured till the twelth century. The interesting account of an Arabic graffito on a wall Of a basilica on the Gemile Island, dated to late seventh or early eighth century also gives evidence for the Arabic existence on the Lycian coast (:9). *** Ibid, 102; Theophanes, Chronographia 1 384. *** Ibid, 103. In 860 Antalya was taken by an Arab fleet, and in 1034 the church of St. Nicholas was once again looted by an Arab flotilla. 147

The recovery under the Macedonian dynasty^ (867-1056) lasted for a short

period, and right after the battle of Manzikert, in A.D. 1071, won by the Turks,

Lycia was overrun by Turkish troops and the infiltration of nomads, especially fi-om

the northeast, who occupied northern Lycia, mainly the plains of Elmalı, Korkuteli

and Söğüt. With the movement of the nomads down to the coast, at the latest in

1191 the detailed accounts of the Third Crusaders refer to the lands east of Mt.

Cragus (modem Babadağ) and Xanthos, as caput Turchiae, the lands of the Turks,

which had actually marked the frontier between the empire and the sultans. ^**The

Byzantine control over Phaselis survived untill 1158, and over Antalya untill 1207,

and one by one the main coastal strongholds were also overtaken by the Selçuk

sultans.

Eiuring the period of Turkish infiltration, only under the Comnenus dynasts (1081-

1185), did coastal Lycia live a short period of prosperity, which is evident from a number of building activities at sites, including Telmessos, Xanthos, Patara, Kyaneae,

Limyra, Phoenix and Myra. 589

Two central Lycian sites, in contrast to the coastline settlements, gives evidence for continuity through the crisis years and recovery in the ninth century. Both

Alakilise, and a new settlement at Dereagzi with a grand scale church enclosed within a fortification system built in the ninth century, continue well into the eleventh and twelth centuries. The sixth century church and village complex of Alakilise, which

Foss, 1994: 50. Under the Macedonian dynasty Crete, which had constituted an important base for Arab activities, was taken from the Arabs, in %1. During this period Byzantium defeated most of its enemies bringing stability to the country. The recovery under the Macedonian dynasts has been considered to be of small scale, and evident at a number of sites, with the rdxiilding of churches, usually within the larger church or basilica structure, in a smaller scale. Buildings dated to this period are the small church at Xanthos, which replaced its large basilica, a church in Cydna, and with repairs to its fortification wall, and in Myra, the repair of the church of St. Nicholas and a new church built below the acropolis.The church of DereaQzi, in the Kasaba Plain of central Lycia, points to a major scale of recovery, together with the rd>uilding of Alakilise during the same period. Ibid,3. ^»^Ibid,51. 148

has been surveyed by Harrison^®®, gives evidence for the rebuilding of the cross-in­

square church of the sbcth century in a grander scale in 812, as it has been attested

from an inscription. The village complex also gives evidence for continuity without

any break during the crisis years of the seventh and eighth centuries. With no sign of

increasing population, it would be dangerous to conclude that the coastland

population moved inland. But we may state that the inland settlements were less

vulnarable to outside threats, and had a better chance to survive the crisis years. Three

phases of renovation in the church also suggest its continuing function well in the

eleventh and twelth centuries.

The settlement complex of the church and fort of Dereagzi and its vicinity has

been recently studied by J. Morganstem.^®^ The strategically important location of the

settlement, at the head of the Demre gorge in the Kasaba plain, has been suggested to

be the reason for its survival in the ninth century. Its location guards access into the

interior, especially to the Kasaba and Elmalı plains, with their rich resources of grain

as well as timber, which had played an important role in the Arab looting activities. It

has been suggested that the reflourishing of the settlement was a counter reaction to

the Arab threat with the aim of providing a substantial base for the fleet of the

Cibyrraeotic Theme, which had played an important role in the defeat of the Arab

forces.^”

The archaeological evidence has suggested a more or less continuous occupation,

with periods of interruption, from the Iron Ages throughout the Ottoman period, in

Harrison, 1963; 127-131; 1978; 9-10; 1979a; 205-206; 1981b; 386. Ibid The site was first reported by a number of 19th and early 20th century travellers. It was first studied by H. Rott and RM. Michel in 1906, and an extensive re-studying of the monument and its surrounding was begun by J. Morganstem in I%7 supported by the American Research Institute and continued in the I970’s.The preliminary publication of this work was given in a number of articles presented in the 1968.217-225, 1969-70; 383-393, 1967; 161-169, and see 1983. A recent publication is Morganstem ed, 1993. In Morganstem ed, 1993; 165. 149

the region of the Dereagzi settlement . ^®''The church of Dereagzi, which was a part

of the settlement and the fort, is a monumental building of a cross-domed church,

which is flanked to the north and south with additional buildings of octagonal shape.

The dating of the church has been problematic, but a date within the second half of

the ninth century has been suggested.

For the northern part of Lycia, and the cities of the KibjTatis, virtually nothing is

known for Late Antiquity or the Byzantine period, apart from the mentioned basilican

churches of Oenoanda and Balboura, which have been tentatively dated to the fourth

or fifth centuries. These sites have not been excavated, but only surveyed, and the

given dates are subject to change as a result of excavations in the future. From the

plain of Elmalı we only have the excavation of an unpublished church from

Cağıltemeller in the village of Ba yindir, which was excavated by the Antalya

museum in 1989. Harrison refers to two churches from Ovacık, without giving any

detail or dating.^^ Also mention must be made of the Müğren chamber tomb of the

sixth century b.c., subsequently converted into a church.

594' Ibid, 1 162-163. Mainly relying on the evidence of pottery, and coins, as well as the Lycian type of rock cut tombs it has been suggested that there was occupation in the region as early as the Middle and Late Iron Ages, as well as in the late sixth or early fifth century, and from the fourth to the first century B.C.. Coins dated to the first half of the fourth century B.C. belonging to the Lycian (fynast Pericles were encountered Also coins belonging to Ptolemy I and II have been encountered pointing to Hellenistic existence at the site. Sherds and coins suggest continuation of occupation fi-om the Hellenistic period well into the seventh century AC. Late Roman coins dated to third till mid-filth century AC. were abundant at the site. After a period of a gap coins appear fixrm the Justinianic period of the mid sixth century. Coins from the first half of the seventh century appear to belong to and Constans n. Coins of the eighth century, rare from any other site, have not been encountered at DereaQzi either. The first coins encountered after the gap of the eighth century belong to Leo V, with increasing activity in the ninA century throughout the tenth, and into the eleventh centuries. Ottoman coins from the twelth till twentieth century have been suggested as pninring to a continuation of occupation in the form of a modest settlement at the site. ^ Morganstem, 1969/70: 390-393. Preliminary investigations of the church have given a date in the second half of the ninth century, fiom the evidence of the mosaic of the diaconicon, the representations on the mosaic can be attributed only to a time before the beginning of Iconoclasm, in 726 or some time after the Feast of Orthodoxy in 843. C 14 dating of beams from the church have given an earliest date of 794 ± 40, which rules out a date of earlier than 726. The fort, which is thought to be contemporary with the church complex, has been dated with C 14 daring of its mortar to the last two-thirds of the nineth century, and qrigraphic evidence from the fort has ciggpstfd a date of the end of the ninth or early tenth century ( Morganstem, 1983:163) Harrison, 1979c: 530. 150

Interestingly, despite the turmoils of the Byzantine and Medieval periods, the number of bishopric centers conveyed to us through the Notitiae remains fairly constant throughout the sixth century, into the twelth century. We need to bear in mind, however, that the mentioning of the bishop centers did not always mean that there existed a whole town or community, nor did it mean that the bishop necessarily resided there.

At this point I would like to give the literary evidence that mentions the main cities or bishopric centers of the Elmalı Plain.

2.7.1 Ecclesiastical Sources

(a)Hieroclis Synecdemus (ed.A. Burchardt, Leipzig 1893)

Hierocles’ list, which has been dated to a period before A.D. 535, presents a geographical list of cities of the Eastern Empire, which have been arranged by province and in a rough geographical order. The list is assumed to have been based on a secular administrative document from the mid fifth century, with later additions under Justinian I, under whom the author compiled the list (Appendix El). Only

Chôma and Podalia have been mentioned, right after Limyra and Arycanda, which fits into the geographical order.

(b) Vita of Nicholas, the late Archimandrite of Holy Sion (ed. N.P.and I.

Sevcenko, The Life of Saint Nicholas of Sion, Hellenic Coll. Press 1984)

The source possibly gives reference to the ancient city of Komba, in the sbcth century (;79); „... And there were grief and wailing and lamentation and great

597 Oxford Dictionary of Byzantium:930. 151

mourning. And Paul (son) of Hermaios, the most devout deacon from the hamlet of

Oumbe (χωρίου Ουμβή)^®*, arrived himself, for that honorable wake. Having consoled all the holy fathers and brethen, he went down to Myra to meet Philip,

God’s most beloved bishop of the Phellites, that Philip might bury Nicholas. . . .“

(c)Basil n Notitia (ed. H.Gelzer, Descriptio orbis Romani Leipzig 1890)

This source has been compiled by a seventh century geographer, George of

Cyprus. His list, similar to that of Hierocles, has been compiled from secular administrative documents and re-compiled by the ninth century Armenian Basil of lalimbana (Appendix E2).^^ The three ancient cities associated with the Elmalı Plain have been mentioned under the Eparchia Lycia, and the Metropolis of Myra. Podalia has been mentioned right after Appillon, and before Arycanda, Chôma after Oenoanda and before Kandanon, and Komba in conjunction with Patara and Nysa.The sequence implies that the lists were compiled in a geographical order.

(d)Notitiae Episcopatum (ed. J.A.A. Darrouzes Notitiee Episcopatuum Ecclesiae Constan tanae Paris 1981).

The Notitiae are lists of ecclesiastical , which are arranged in a hierarchical order; as metropolitan sees, then autonomous archbishoprics, and the bishoprics clustered under a metropolis. For Lycian bishoprics this has always been the main

Though the name has not been written in the form we have been encountering in the ancient sources, as ‘Κόμβων’ , the name closely resembles the modem name of the town. Gömbe. Oxford Dictionary of Byzantium: 837-38. ^ If Podalia is to be identified with the site of Buralia, near the Avian Lake, Arycanda (modem Arif) would be the neighbour to the south, once passing the Avian Pass. Chôma and Oenoanda probably had a common border, which has been alreatfy discussed (see pg. 124). Komba,identified with modem Gömbe, and Nysa (identified with a site near Sütleğen, right at the foot of the Sinekçibel pass) would also fit into the picture. 152

metropolitan seat of Myra (modem town of Demre). The earliest of these lists were compiled in the reign of Heraclius and continued into the Turkish period (Appendix

E3).®°’ All three cities of Chôma, Podalia and Komba have been mentioned in the lists of the seventh through twelth centuries. The order in which they have been mentioned does follow the order in the lists of Hierocles and the Basil II Notitia.

(e) Bishop names for Chôma, Podalia and Komba

We do have the record of bishops’ names as early as the fourth century, both for

Chôma and Podalia, which can be pointed out as significant for a region commonly considered as remote. Bishops from Podalia and Chôma took part in the Council of

Constantinople, in A.D. 381, and have sided with the Orthodox view, as mentioned earlier (pg. 138).The same Bishop (II. Eudoxius) of Chôma was present at the

Councils of both Ephesus and Chalcedon, and signed the Epistola ad Leonem, siding with Orthodoxy, against the Monophysites. We have the name of the Bishop from

Podalia, who signed the Epistola ad Leonem, but did not take part in the Councils.

We do have the names of Bishops from Chôma as late as the ninth century and from

Podalia as late as the sbcth century. From Komba we have only the date of a Bishop in the ninth century. The following quotations have been cited from Le Quin, to give the exact names, titles, and dates of the Bishops from Chôma, Podalia and Komba.

Oxford Dictionaiy of Byzantium : 14%. Notitia 21, compiled by Darrouzes, has been dated to the Tinkish period, with an antequem of 1453. In the list under the metropolis we encounter the city of Kabalia (η Χριστον>πολισ ήτοι Καβαλα). Though traditionally the region north of the Elmalı Plain, around Söğüt Gölü, has been identified with the ancient region of Kabalis, there have been also attempts to identify the modem town of Elm alı with Kabalia. We know that with the Islamic period, the town of Elm alı and its vicinity started to be come important centers. Attempts by scholars (eg. Ostrogorsky) have been made to interpret these lists with the economical and political history of certain regions, assuming that the lists display a close situation of the periods. But this approach has been criticised by others. 153

Bishops of the Ecclesia Chomatis in Provincia Lvciae I. PIONIUS (Le Quien I: 983; Mansi ΙΠ:570)«“

(A.D. 381) „ ...Inter episcopos Constantinopohtanae synodi prioris deprovincia

Lycia, legjtur P/0/I/M5 Comatensis, seu potius Chomatensis.

II. EUDOXIUS (Le Quien I;983; Mansi IV:1226,Mansi νΠ;437,580)

(A.D. 431,451, 458) „ ... Ephesinae synodo primum interfiiit & subscripsit Eudoxius episcopus Chomatis provinciae Lyciae, Εύδοξιος Επίσκοπος Χώματος έπαρχίας

Λ υκίας, qui prosit annos viginti şedit iterum in concilio Chalcedonensi, quin & concibo Myrensi anno 458. cujus synodiae ad Leonem Imp. perinde subscripsit,

Eudoxius episcopus Chomatenus.... “

III. NICOLAUS (Le Quien 1:983)

(879/880 AD)“ ... Photianae synodo sub Loanne Papa Vin.Mco/ûfM5

Chomatis,Νικολάου Χώματος.... “

Bishops of the Ecclesia Podaliae in Provincia Lvciae

I.CALLINICUS (Le Quien 1:978; Mansi 0:570)

(A.D. 381) „... Inter episcopos provinciae Pisidiae & Lyciae, qui

Constantinopolitanae synodo primae intersuerunt Callinicus Podalienus vel

Podaliensis.“

B AQUILINUS (Le Quien I: 978; Mansi VH:580)

(A.D. 458)“ ... Epistolae synodi Myrensis ad Leonem Imp. de nece S. Proterii

Alexandrini subscripsit, Aquilirms episcopus Podalienus vel Podaliensis...."

602 Le Quien, 1740; Mansi et. al., 1759/1927. 154

III. JOANNES I. (Le Quien I: 978; Mansi ΥΠΙ; 1147)

(A.D. 536) „ ...Synodo Constantinopolitanae sub Menna aderat Joannes episcopus

Podaliae, Ιωαύυης Επίσκοπος Ποδαλίίχς.

IV.JOANNES Π. (Le Quien I: 978)

(A.D. 879/880) „ ... Synodo, Photianae, Joannes Podaliae,Ιωαύνου Ποδαλείας,

Ιωάννης Ποδαλείας...."

Bishops of the Ecclesia Comborum in provincia Lykiae

I. JOANNES (Le Quien I; 991)

Date ? „ ...Quinisextae synodi canonibus cum Lyciae episcopis subscrittus legitur

Joannes episcopus Combi, Ιωαύνης Επίσκοπος Σόμβον, legendum Κόμβον..... “

II. CONSTANTINUS I. (Le Quien I: 991)

Date ? „ ... In Séptima synodo generali şedit & subscripsit cum aliis de Lycia episcopis, Κωνσαντίνος vel Κώνςας Επίσκοπος Κόμβων. Constantinus episcus

Comborum.“

III. CONSTANTINUS II. (Le Quien I: 991)

(A.D. 879/880) „ ... In Photii synodo sub Joanne Papa VIL Constantinus Combae,

Κωνσταντίνον Κόμβας, Κωνσαντίνος Κόμβας. ..."

(f) Constantinus Porphyrogenitus de Thematibus

Under the military system of the Themes, Northern Lycia came under the

Kibyraeotic Thema. Among the list of the cities given to us by Constantine

Porphyrogenitus n, Podalia was mentioned among the cities of the Kibyraeotic

Theme. 155

„Thema XTV, dictum Cibyrraeotarum

... transit deinde Stratoniciam, et quae vocantur Mogola, et urbem pisyam; (percurrit

etiam earn quae Hagia dicitur, et Tauropolim: pertingit vero etiam Tlon, et

Oeniandem usque. Transit deinde Phileto, et imsam Podaliam.)“

The history of Lycia, and especially Northern Lycia during the Byzantine period,

remains very sketchy. Lycia was separated from Pamphylia under Constantine in the

early third century, and Myra was appointed as its ecclesiastical metropolis. In 732 it

was included in the Kibyraeotic Theme, together with Pampylia. The Arab attacts hit

the Lycian coast in 655, with the sea battle of Phoinix, which opened the whole

coastline to the exploitations of the Arab fleets. The prosperity of the coastal and

inland Lycia, reaching its peak during the fifth and sixth centuries, encountered both

from archaeological and literary evidence, came to an abrupt end during the seventh

century. The signs of a degrading economy and rising problems can be vaguely

observed in the Vita of Bishop Nicholas, during the sixth century. A combination of

reasons, economic, social and environmental, besides the impact of the Persian and

Arab attacks can be counted among the start of the crisis years of the seventh till

early ninth century. The shrinking of urban settlements on the coast and inland has

been taken as evidence for a mass decline in population . The question regarding

what happened to the inhabitants remains unanswered. A shift of population to the

inland as suggested by Harrison lacks concrete evidence. The scanty evidence from

Alakilise and Dereagzi which point to continuity during the crisis years, remains far

from suggesting a mass movement into the hinterland. Lack of archaeological

evidence or any sign of building activity should not be taken as absolute evidence for

depopulation of the region. Short periods of recovery have been observed in

archaeological evidence during the ninth and twelth centuries, under the Macedonian

Dynasty and Comnenids. After the twelth century Lycia completely falls into the hands of the Turkish troops, which most probably were infiltrating the upland basins 156

of Elmalı, Söğüt and Korkuteli, and the plains displayed a suitable environment for the arriving pastoralist groups.

Our knowledge for the Byzantine period of the Elmalı Plain remains sketchy, both due to lack of evidence, and the lack of investigations on the plain. We have the

ecclesiastical lists which give the names of the cities of Chôma, Podalia and Komba from the sixth century onwards, till the twelth century. We also have the names of

Bishops, as early as the fourth century till the tenth. Mounds identified with Chôma,

Podalia and Komba have been suggested to give pottery evidence into the medieval periods. Sculptured capitals dated to the sixth century have been found in the villages of Ovacık, Müğren and Tekke. The ongoing excavations at Hacimusalar have pointed to substantial building activities after the late Roman period. With the very recent excavation season of 1998 a church with a mosaic floor has been recovered, which is still in the process of investigation. The Çağıltemeller church, the Müğren tomb converted into a church, and the Ovacık churches are the only archaeological evidence for ecclesiastic buildings from the plain. The ongoing excavations and survey project of Hacimusalar will throw ample light upon the history of the plain and, moreover, on the region of Lycia in general. 157

CONCLUSION

The present work has aimed to give a review of Elmalı Plain’s archaeological and historical evidence in its environmental setting. The conclusion section evaluates the archaeological evidence at hand and makes some suggestions for issues that can be derived from the archaeological data. Suggestions have done for issues that can be pursued further with archaeological investigations, for a better understanding of the history and nature of human occupations on the plain.

The archaeological evidence is silent for the periods before the Neolithic. Starting with the Late Neolithic for the prehistoric occupations on the plain we have a general idea for the existence of settlements.^“^ The evidence for archaeological reconstruction of the prehistoric settlements comes from the two excavated mounds of Bağbaşı and Karataş, both extensively studied, which, although adding a wealth of information to our understanding of the Late Chalcolithic and Early Bronze Age periods of the plain, these two examples are barely suflBcient to reconstruct the general pattern of settlements for the whole plain. Our knowledge regarding the pattern and nature of the prehistoric settlements during the Chalcolithic and Bronze

Ages remains full o f gaps.

Especially from the survey results of Hood, Mellaart and Eslick, see pg. 48-51 for their discussion. 158

After a discontinuity of archaeological evidence for the Late Bronze Age, we have

a rich burial tradition that becomes evident through a number of incidental recoveries and salvage excavations, dated to as early as the eighth and seventh centuries b.c . and continuing well into the fifth century b.c.. The excavation of the tumuli of Bayındır,

K izilbel and Karaburun has proved the existence of rich burial customs where both

Phrygian and Lydian burial traditions can be observed, with local traits prevailing. The long-standing Early Bronze Age tradition of inhumation in pithos and jar burials also seem to continue, as we understand from the scanty evidence from the Karataş mound. If we consider burial traditions as an important indication in reconstructing the cultural, political and economical conditions of the associated communities, we may state that the first half of the first mülenium b.c. displays an increasing contact with the „Southern Phrygian“ and Anatolian Highlands. This may indicate the arrival of new groups or population with mixed ethnic backgrounds in the first half of the first millennium b.c.. The Elmalı Plain was possibly inhabited by the Milyans which are a mentioned in the accounts of Herodotus, as early as the fifth century b.c. .^'* The contents of the B ayındır Tumuli and the richly decorated chamber tombs of

K izilbel and Karaburun, possibly belonging to the members of rich land owners, point to an established elite class living in the Elmalı Plain. They were possibly engaged in military activities and adapted themselves well to the changing political conditions of Southwestern Anatolia. We have the possible depiction of the K izilbel tomb owner in Greek armor (Plates 48, 49) and fifty years later, we see the owner of

Karaburun who has chosen to be depicted as an Achaemenid aristocrat (Plate 52), and

^ It is significant that for the rich tumuli tradition showing Pyhrigian and Lydian influences, our only evidence from Lycia comes from the Elm alı Plain. Although tumulus tombs dated to the first half of the first millennium do exist in Central Lycia, around Phellos (Çukurbağ), they are in the so- called Lelegian/ West Carian tradition. Their discussion has remained b^ond the scope of the present work, but they also indicate the arrival of different ethnic groups, sometime in the first millennium B.C. in Lycia, probably via the coastline. Borchhardt and Wurster, 1974: 514-538; Zahle, 1975: 77-94. 159

fighting against Greek soldiers (Plate 54 A, B). Unfortunately, we do not have the

evidence of any settlements fi-om this period that can be related to the owners of these

tombs. This is an issue that will be clarified with future investigations and excavations

in the region. So far, only Söğle and Hacimusalar mounds have yielded Iron Age

pottery. Both of these sites appear to have substantial occupation levels of later periods, extending well into the medieval periods. Understanding the later periods and the Iron Age levels, besides being a critical addition to the archaelogical record, is a long term investment and a commitment. Excavations at Hacimusalar, the largest mound on the plain, possibly will reveal the Iron Age phases, the nature of their settlements, their interactions and relations with neighboring cultures and their continuity fi’om the Bronze Age onwards.

No archaeological work has taken place on the plain regarding the classical periods. The few studies have concentrated only on the location of classical cities mentioned in the literary sources, which have been identified with sites in the plain, and the documentation of some of the inscriptions. With the scanty evidence at hand, any reconstruction of these periods will remain immature at this stage. We may only hypothesize an increasing contact of the Elmalı Plain with the coastal regions, and

Lycia in general, with the evidence of few Lycian types of tomb, the only epichoric inscription located at K ızılca, and the two coin hoards which include coins of Lycian dynasts. If these could be attributed to the expansion of the political powers of coastal

Lycia into the uplands, again, it remains premature to make a definete statement at this stage.*^*

Other than political contacts, we also need to keep in mind the factor of transhumance, which Northern Lyda, as it ai^iears today, may have served as a summer pasture for the people of coastland. 160

For the Hellenistic, Roman, Late Roman and later periods, virtually other than the

epigraphic evidence, virtually no archaeological research has been conducted in the

region. From literary and ecclesiastic lists we see that the classical cities identified

with the Elmalı Plain, Chôma, Podalia and Komba, have been connected

administratively to Lycia and possibly also became an integral part of the economy of

Coastal Lycia, which has flourished during the Roman periods, with its well situated

harbours and rich hinterlands. However, once again it would be premature to try to

reconstruct these periods for the Elmalı Plain. At this point, the ongoing excavation

and survey project of the Hacimusalar Höyük promises to be a crucial project in throwing more light upon our archaeological knowledge of all these periods.

However, one of the basic tool of archaeology, which is the study of ceramics, lacks comparanda material in general from the excavated sites in Lycia.

Herewith two generalizations will be pursued further. First the plain with its rich resources, was able to hold substantial populations, with elite classes, which possibly comprised rich land owners, whose wealth depended on their successful utilization of the resources of the plain. Second, the Elmalı Plain, despite being referred to as a remote upland plain, was a transitional zone where cultural influences arrived fi’om both the coastal zone and the Anatolian Plateau.

The Elmalı Plain and its surrounding mountains offer a variety of resources, such as timber, available throughout the dense cedar and pine forests that surround the plain. Rich pastures, referred to ‘Yayla’ exist at different elevations in the surrounding mountains. We can also consider game fi'om the surrounding mountains 161

and forests as a possible additional source for subsistence, as it is today.^^ The fertile

alluvial base, with well supplied water sources of two lakes and a number of rivers,

offers a good set up for agriculture in the plain. The lakes would also have provided a

variety of resources for the inhabitants of the plain. The Elmalı Plain, in general,

would provide a good setting for human groups. The Late Chalcolithic and Early

Bronze Age periods point to substantial habitations on the plain. For the Iron Ages,

and the first part of the first millenium we do not have the evidence of settlements but

the rich sepulchral tradition which points to the existence of an elite class which could not have survived without the existence of a substantial population. These populations also might have been a resource for raising armies.“ ^ The stability brought with the

Roman period possibly resulted again with a substantial population in the plain. The archaeological data at hand at this point is far from suggesting a continuity or discontinuity through the Roman, Late Roman and Byzantine periods. No intense archaeological investigations have taken place on the plain regarding these periods, but the abundant evidence of sites and epigraphy indicates that the plain was well populated. However the epigraphic data does not give any evidence for a wealthy elite class, such as is apparent from the neighbouring region of Kibyra in the Roman period. From the published inscriptions we have the scanty evidence of these upland populations being used as a source of mercenary.^*

^ The preliminaiy result of the zooarchaeological investigation of the Hacimusalar Mound points to an of abundance of wild animal remains, which suggests that the inhabitants of the mound were hunting in the past. In both of the tumuli the owners have taken part in battles, as they have been depicted marching off to a war, and as soldiers. We have the literary account of Herodotus (VII.76,77), that the Milyans accompanied the Persian King Xerxes, in the war against the Greeks. The suggested expansionist move of the Lycian (fynast Pericles in the first half of the fourth century B.C., into the Elmalı Plain, was possibly an attempt to bring the upland population imder his control, and be able to utilize their resources (both natural and human). The P o ^ ia Hoard of the late fourth century B.C., which constitutes Persian sigloi and staters, has been suggested to be a possible payment for a mercenary force. Epigraphic evidence from the mound Söğle (Appendix F6g), which lists the names of soldiers for a mercenary troop, and the mention of persons who have served as soldiers (Appendix F2b, FlOa), 162

The Elmalı Plain being the southernmost upland basin of northern Lycia, can be

viewed as a transitional region, both geographically and culturally, between coastal

and central Lycia, and the southwest edge of the Anatolian Plateau. Despite the fact

that it has been referred as an isolated remote upland basin, archaeological evidence

indicates that it was well within the Southwestern Anatolian cultural sphere. The

Akdağ and Beydaglari massifs make the plain a difficult place for access; however,

passages do exist which have become the main communication links to the coast and

Anatolian highlands. The Elmalı Plain as early as the Chalcolithic period had

connections to the outside world, as suggested from the scanty archaeological

evidence fi"om the B ağbaşı site. Pottery evidence for the Middle Chalcolithic period

points to possible connections to the Aegean Islands. The Karataş Village which

proves to be a direct development out of the Chalcolithic period, is a part of the

Southwestern Anatolian cultural sphere. The burial tradition attested from Karataş

shows strong cultural connections with Western Anatolia, as well as pointing to

parallels as far as Byblos in the Eastern Mediterranean. The fi’eestanding houses of the

megaron type have a wide distribution in Eastern Europe, Thessaly, Greece, the

Aegean and Western Anatolia. As the Early Bronze Age evidence points to strong

cultural ties with Western Anatolia, with the Middle Bronze Age evidence, though

scanty, we see growing cultural ties with Central Anatolia and Hittite centers. No

change in the burial tradition discards the possibility of new arrivals into the plain. The

Middle and Late Bronze Ages need to be better examined for the region. The rich

burial traditions attested fi'om the Bayındır, K izilbel and Karaburun tombs point to possible new comers to the region, especially after the long period of the so-called

can be taken as evidence that Elm alı Plain, as it was usual for highland regions, was a source of raising armies in the Roman period. 163

‘Dark Ages’. The strong Phrygian and Lydian influences are significant for the Elmalı

Plain, and do not appear in the neighboring regions. Starting with the Classical period,

an increasing influence of the coastal regions becomes apparent. The extensions of the

strong Persian influence can be observed in the plain during the period of Persian

existence in Lycia. The typical Lycian tomb architecture is observed especially along

the southern part of the plain, with sites which are proximate to the main passages that connect the plain to the coast. With the Roman period the Elmalı Plain becomes economically and administratively connected to the province of Lycia. Cult practices attested from the plain (such as the worship of the twelve gods, the Dioskouroi and

Helene, Artemis Lagbene, the Trinity Gods and Kakasbos) proves that the Elmalı

Plain was both a part of the cultural sphere of the upland regions of Kibyra, Kabalia and Pisidia, as well as Coastal and Central Lycia The geographical location of the plain appears to be one of the dominating factors for the cultural influences attested from the plain.

Another period that deserves scrutiny at, but remains beyond the scope of the research and is not included in this thesis, is the Islamic occupation of the Elmalı

Plain. During the Islamic period the Elmalı Plain was well populated, first by the arrival of pastoralist groups, and later by a substantial urban population in the main town of Elmalı, with the Ottoman period. As early as the twelfth century, first

Northern then Coastal Lycia was lost to the Turkish troops. The Lycian peninsula was populated by pastoralist Türkmen groups which arrived from two important centers,

Ladiq and Honaz.®®^ The Elmalı, Söğüt and Korkuteli plains were the first places to

Flemming, 1964:1-2. Ibn Sa’id’s accounts (1274-1286) tell us that there were 200 000 black tents of the Tüikmen tribes had settled around Ladiq (Denizli). If this number is not exaggerated, it points to a huge population. These bases constituted the border between the Rum-land (Byzantine) and the Türkmen groups, which the Comnenid emperors had to fight against, but were hopeless under the increasing preasure of new arrivals. 164

receive Turkish nomadic groups.^‘® One of the most well-known Bektaşi centers, the

Tekke (Derviş centers) of Abdal Musa, is located in the Elmalı Plain. Abdal Musa is

assumed to be one of the most important followers of Hacı Bektaş Veli.^“ Although

legends state that he lived and died in Elmalı, where apparently his tomb is to be

found, we do not know how reliable this information is. Our earliest account of the

Tekke is in the 16th century, where Evliya Çelebi gives a detailed description of the

foundation.^*^ Other Tekke orders located along the coastal Lycia, and an even earlier

one in Turunçova (Limyra), have existed in conjunction with the m ain center in

Elmalı. The main center in Elm alı, till it was closed in 1826, functioned as a large

land estate, and was considered one of the richest Derviş orders in Anatolia.

The town of Elmalı quickly developed into an administrative center. Earliest in H.

822/1419 century we have the mention of the name of the town in a ‘v ak ıf plots

account, known as the Toroman Vakfı, which refers to a donation of 32 pieces of

ibid.,3-4. Ibn Sa’id gives us the accounts of the Uğ-Türionens around Makri and Antalya, also in the region of Karia, in the thirteenth century. Köprülü, 1973: 198-207; Akçay, 1970: 360-73; Birge, 1937: 55, 211-12. His life and legends are known to us from the 15th century Ottoman historian Aşıkpaşazade.He is assumed to have arrived from Hoy (Xoy) in Horasan together with the 40 Abdals, and lived some time in the late thirteenth and first half of the foiuteenth century, and took part in the wellknown siege of Bursa, in 1326, under Orhan Gazi, founder of the Ottoman Empire Although his belief later differed from that of the Haci Bektaş ’, his ideas were followed by the Alevis, and especially spread among the nomadic Türkmen tribes, more than among the educated settled elite conununities of Bektaş i . It is significant that his main center was found in Elm alı, in a region where Türkmen nomads had settled as early as the twelth and thirteenth centuries, and remained there. Also see the stucfy by F. von Luschmi in Reisen I I 199-226, where he gives a detailed account of the Alevi groups also known as the Tahtacılar, who retained their strong Türkmen identity as late as the 19th century. ®’^Zülioğlu, 1985: 114-15. The Kafi Baba Tekke in Turunçova is believed to have existed since the fifteenth century, and is the earliest known Tekke in the region. The Tekke has been excavated by the Limyra team, see Bauer, 1988: 343-62. Faroqhi, 1981:48-77. Especially the 18th and 19th century Ottoman ‘tahrir’ tax registers point out that the Abdal Musa Tekke functioned like a land estate, with tenures, and share croppers, and leasing out money against interest rates. When the Tekke was closed in 1826, the inventory list consisted of 890 ha. land with additional 5 ha. of orchards and vineyards, 40.5 tons of wheat, 16 tons of com (which was circa 3% of the total revenue of the provice of Antalya). Also see the chapter by Faroqhi, 1986: 69-%, for general information regarding tire system of the agricultural activities in a Bektaşi center. 165

land around Elmalı.^** The traditional Elmalı wrestling contests, which are held

annually, have celebrated their 646th anniversary this year, which suggests that this

tradition went back as early as the fourteenth century. The 16th century accounts of

EvUya Çelebi, though sometimes exaggerated, refer to a well established

administration system and organization of guilds. He counts 32 districts with 32

mosques (assigning one mosque to each district), 8888 houses (?), 300 shops, 70

mills, seven schools, four Medreses, five bath houses, and three inns.*^^® From the 19th

century travellers’ accounts we leam that the town had a population around 10 000,

including an ethnically diverse population of Greeks, Jewish and Armenians.

According to the first census figures of the Turkish Republic 4777 Muslems, 239

Greeks and 430 Armenians lived in Elmalı, which points to a considerable shrinkage of the population, especially with the Non-muslim groups, that can be viewed as normal, especially after the 1922 exchange of populations with mainland Greece.

Finally I would like to stress the importance of certain issues that need to be

answered for a better understanding of the history of the plain. The environmental

data presented in the first chapter does point to the potentials of the Elmalı Plain.

The present day model should be tested against the environmental conditions of the

ancient times. One of the primary issues would be to determine the lake limits of

Karagöl and Avian in the past, which proves to be very important to our

understanding of the settlement patterns in the plain. These lakes which played an

important role in the subsistance of the human groups, possibly also had negative

The information has been kindly given by Abdullah Ekiz, whose investigations regarding the history of Elmalı Town has been gathered in his unpublished book Dünden Bueûne E lm alı. The next known account which mentions Elmalı is from the ‘Tekke Livası Defleri’ a regional administrative account dated into the first half of the 16th century, which gives 21 districts for the town of Elmalı. ‘"'^Zillioglu, 1985; 115-17. 166

effects, especially during periods of flood, which would enable the usage of

agricultural land and foundation of settlements. Other environmental issues would

be the preparation of a pollen diagram and reconstructing a paleovegetational map

of the plain, and the preperation of soil map of the plain. A climatic study of the

ancient periods also would prove to be very important.

The ongoing excavations of Hacimusalar will help to determine the pottery

sequence for the different periods and establish local fabrics. The ongoing survey

project of the Elmalı Plain, in conjunction with the Hacimusalar excavation, aims for the location of sites from all periods, which will add to our understanding of the settlement patterns vrithin the historical sequence of the plain. It would be important to see if human groups settled prior to the Chalcolithic and Neolithic periods in the plain. For the Early Bronze Age period, which points to a flourishing period of village life, it would be helpful to know the exact distrubution o f settlements in the plain.

Locating Iron Age settlements on the plain is another important issue, as well as understanding the settlements distribution of Classical, Hellenistic and Roman periods.

It is important to determine the extent of relations of the Elmalı Plain with the coast during the Roman period, which points to the increasing involvement of northern and central Lycia with the coast. Investigation of remains of roads along the main passages to the coast and neighbouring regions, which is one of the objectives of the ongoing survey is another important issue. The excavations at Hacimusalar are revealing large amount of amphoras and storage jars; the nature of these vessels may tell us about the traded items and trade connections. Another idea could be the investigation of the sherd yards around the Horrae of Andriake and Patara for comparanda material. Determining the Late Roman occupations on the plain proves to 167 be another crucial issue for the history of Lycia in general. The preliminary results of the excavations point to substantial building activities during late antiquity. These results would be very important in checking the opposing theories of population movement into the interior or considerable shrinkage starting with the Late Roman period, for upland Lycia. Understanding the extent of inhabitation of the plain during the Byzantine and Medieval periods, and to see the extent of Islamic intrution and nomadic involvement into the plain, are other issues that need to be examined. With the Islamic periods it becomes necessary to look at Ottoman documents related with the town of Elm alı. At this point an ethnographic study of the remaining pastoralist groups, which are rapidly disappearing from the scene, becomes very important. We can also try to trace their existence from archaeological records. It would be also useful to understand if periods of discontinuity and lack of archaeological evidence is a result of shifts from a sedentary way of life to a more pastorolist subsistence on the plain.

We hope that the systematic survey and excavations on the plain will at least answer some of these questions and emphasize the role of Lycia in the archaeological context of Anatolia.