LANDSCAPE ARCHAEOLOGY IN SOUTHERN EPIRUS, GREECE I

HESPERIA

SUPPLEMENTS

1* S. Dow, Prytaneis: Studyof the Inscriptions A HonoringtheAthenian Councillors (1937) Well Graves a Seventh-Century in theAgora and 2* R. S. Young,Late Geometric (1939) 3* G. P. Stevens, TheSettingof thePericlean Parthenon (1940) 4* H. A. Thompson, TheTholos AthensandIts Predecessors (1940) of 5* W. B. Dinsmoor, Observations theHephaisteion on (1941) 6* J. H. Oliver, TheSacred Gerusia (1941) 7* G. R. Davidson and D. B. Thompson, SmallObjectsfrom Pnyx:I (1943) the LeslieShear(1949) Studiesin Honorof Theodore 8* Commemorative Real Security, Land Tenure Ancient and in 9* J. V. A. Fine, Horoi:Studiesin Mortgage, Athens(1951) 10* L. Talcott, B. Philippaki,G. R. Edwards,and V. R. Grace, SmallObjectsfrom the Pnyx:II (1956) 11* J. R. McCredie, FortifedMilitary CampsinAttica (1966) 12* D. J. Geagan, TheAthenianConstitution afterSulla (1967) 13 J. H. Oliver,MarcusAurelius:AspectsCivicand Cultural of Policyin theEast (1970) 14 J. S. Traill, ThePoliticalOrganization Attica (1975) of 15 S. V. Tracy,TheLetteringof an AthenianMason(1975) 16 M. K. Langdon,A Sanctuary Zeuson MountHymettos (1976) of 17 T. L. ShearJr.,Kalliasof Sphettos theRevoltofAthensin 268 B.C.(1978) and 18* L. V. Watrous,Lasithi:AHistoryof Settlement a HighlandPlain in Crete(1982) on 19 Studiesin Attic Epigraphy, Presented EugeneVanderpool to History,and Topography (1982) 20 Studiesin AthenianArchitecture, and Topography Presentedto Homer Sculpture, A. Thompson (1982) 21 J. E. Coleman, Excavations Pylosin Elis (1986) at 22 E. J. Walters,Attic GraveReliefsThatRepresent in Women theDressoflsis (1988) 23 C. Grandjouan, Hellenistic ReliefMoldsfromtheAthenianAgora(1989) 24 J. S. Soles, ThePrepalatialCemeteries Mochlos Gournia theHouseTombs at and and of BronzeAgeCrete(1992) 25 S. I. Rotroff and J. H. Oakley,Debris from a PublicDining Placein theAthenian Agora(1992) 26 I. S. Mark, TheSanctuary AthenaNike in Athens:Architectural of Stagesand Chronology(1993) 27 N. A. Winter, ed., Proceedings theInternational on of Conference GreekArchitectural Terracottas the Classical Hellenistic and December Periods, 12-15, 1991 (1994) of 28 D. A. Amyx and P. Lawrence,Studiesin ArchaicCorinthianVase Painting (1996) 29 R. S. Stroud, TheAthenianGrain-TaxLaw of374/3 B.C. (1998) 30 J. W. Shaw,A. Van de Moortel, P. M. Day, and V. Kilikoglou,A LMIA Ceramic Kiln in South-Central Functionand Pottery Crete. Production (2001) 31 J. Papadopoulos,Ceramicus Redivivus:TheEarlyIronAge Potters' Field in theArea * Out of the ClassicalAthenian Agora(2003) ofprint

32 HesperiaSupplement

LANDSCAPE ARCHAEOLOGY IN SOUTHERN EPIRUS, GREECE I

EDITED JAMES

BY WISEMAN AND KONSTANTINOS ZACHOS

Studiesat Athens TheAmericanSchool Classical of

2003

Copyright ? 2003 The American School of Classical Studies at Athens All rights reserved.

Out-of-print Hesperiasupplements may be purchased from Swets & Zeitlinger Backsets Department P.O. Box 810 2160 SZ Lisse The Netherlands E-mail: backsets@swets.nl

Cover illustration: The eroded landscapeof Kokkinopilos above the Louros River gorge

of Data Library CongressCataloging-in-Publication in Landscape archaeology southern Epirus,GreeceI / editedbyJames Wisemanand Konstantinos Zachos. p. cm.-(Hesperia Supplement;32) Includesbibliographical references (p.). ISBN 0-87661-532-9 (alk.paper) 1. Preveza (Greece)-Antiquities. 2. Excavations (Archaeology)-GreecePreveza. Landscape 3. archaeology-Greece-Preveza.4. Arta(Greece:Nome)Antiquities.5. Excavations (Archaeology)-Greece-Arta (Nome) 6. Landscape James. II.Zachos,Konstantinos archaeology-Greece-Arta(Nome) I.Wiseman, L. III. Hesperia(Princeton, NJ.). Supplement;32. DF9oI.P72L36 2003
938'.2-dc2I 2002044060

CONTENTS

List of Illustrations List of Tables PrefaceandAcknowledgments 1 Chapter

THE NIKOPOLIS PROJECT: CONCEPT, AIMS, AND ORGANIZATION

vii xii xv

Zachos byJamesWisemanand Konstantinos Chapter2

THE ARCHAEOLOGICAL
STRATEGIES AND FIELD

SURVEY: SAMPLING
METHODS

byThomasF.Tartaron 3 Chapter
THE EARLY STONE LANDSCAPE AGE OF THE AND NOMOS OF PREVEZA: SETTLEMENT

23

by CurtisN. RunnelsandTjeerdH. vanAndel 4 Chapter

EARLY UPPER PALAEOLITHIC SPILAION: ARTIFACT-RICH SURFACE SITE AN

47

and Karimali, BrendaCullen 135 by CurtisN. Runnels,Evangelia 5 Chapter

THE COASTAL EVOLUTION AND ITS SETTINGS OF THE AMBRACIAN TO EMBAYMENT RELATIONSHIP

ARCHAEOLOGICAL

by ZhichunJing and George(Rip) Rapp

157

VI

CONTENTS

6 Chapter
THE LOWER ACHERON RIVER VALLEY: ANCIENT
ACCOUNTS AND THE CHANGING LANDSCAPE

by Mark R. Besonen, George (Rip) Rapp, and Zhichun Jing

199

Chapter7
SUMMARY OBSERVATIONS

by James Wiseman and Konstantinos Zachos References Index

265 269 283

ILLUSTRATIONS
Illustrationsareby membersof the projectexceptwhere noted.

1.1. 1.2. 1.3. 1.4. 1.5. 1.6. 1.7. 2.1. 2.2. 2.3. 2.4. 3.1. 3.2. 3.3. 3.4. 3.5. 3.6. 3.7. 3.8.

regions Map of Epirusand adjacent Map of surveyzone with selectedtoponyms Multispectral image(SPOT) of the northern part of the surveyzone Multispectral image(SPOT) of the southernpartof the surveyzone The erodedlandscape Kokkinopilos of Aerialview of the fortifiedtown site at KastroRogon Aerialview of the waterchanneland aqueduct bridgesacrossthe LourosRiver Map of southwestern Epirus Archaeological surveytractform of betweentractsand Examples spatialrelationships site/scatters Generalview of the site at Grammeno(SS92-6) areas Map of Epirusand surrounding Tectonicsof northwestern Greeceandthe IonianSea tectonicfeatures Possiblyactive(LateQuaternary) of westernEpirus Presenttectonicactivityin westernEpirusas indicatedby freshstriaeon faultplanes bedrockmapof westernEpirus Simplified Formation a doline (sinkhole) of Diagramof the genesisof loutsesand poljeson a karstic peneplain Poljesandloutsesin westernEpirus

2 3 14 14 16 18 19 29 36 41 44 49 55 56 56 57 58 59 60

VIII

ILLUSTRATIONS

3.9. View ofValtos Kalodiki 3.10. The eponymous loutsa on the raised peneplain south of the lower Acheron valley 3.11. The polje of Cheimadio 3.12. Red sediments and paleosols

63 63 63 64-65 66 67 71 72 73 74 74 75 76 77 79 79 81 81 82 84 87

3.13. Terra rossa redeposited in fan complex 3.14. Typical grain-size frequency diagrams of terra rossa redeposited in poljes and loutses 3.15. The raised polje of Kokkinopilos 3.16. 3.17. Badland erosion at Kokkinopilos Cross section through the incised polje deposits of Kokkinopilos

3.18. Morphi polje outcrop with paleosols forming hard, protruding benches 3.19. 3.20. Composite profile of Ayia loutsa Stratified lower section of the Ayia loutsa looking west; detail of Mousterian artifacts in situ

3.21. The Adriatic Sea during the last glacial maximum 3.22. Global sea-level variations for the past 140,000 years

3.23. The emerged coastal plain off Epirus at six key moments 3.24. Two sea-level rise curves for the deglaciation interval of late OIS 2 3.25. 3.26. Locations of raised paleoshore deposits of the last interglacial in coastal Epirus Cumulative grain-size distributions of coastal sediments of the last interglacial and early Holocene

3.27. The raised Tyrrhenian beach at Tsarlambas 3.28. Climate and vegetation changes during the last two glacialinterglacial cycles

3.29. Maturity stages and approximateages of the Mediterranean paleosol chronosequence 3.30. Relationship between paleosol maturity,terra rossa deposition rate, and Palaeolithic stone tool age in poljes and loutses Palaeolithic site/scatters in the Thesprotiko valley

94 99 100 101

3.31.

3.32. Palaeolithic and Mesolithic site/scatters in the Acheron valley 3.33. View of a stone cluster at Alonaki

ILLUSTRATIONS

IX

3.34. 3.35. 3.36. 3.37. 3.38.

Early Palaeolithic artifacts from Alonaki Early Palaeolithic choppers from Alonaki Early Palaeolithic core-choppers from Alonaki Early Palaeolithic core from Alonaki Early Palaeolithic biface (handaxe) from Ormos Odysseos

102 102 103 103 104 104 104

3.39. Interglacial sand dune (SS92-25) at Ormos Odysseos 3.40. 3.41. Ormos Odysseos, biface findspot (W94-20)

Early Palaeolithic biface or bifacial core from Ayios Thomas 105 109 110 110 111 111 112 112 115 116

3.42. The Palaeolithic site of Ayia and its setting 3.43. Middle Palaeolithic (Mousterian) artifacts from Ayia 3.44. Middle Palaeolithic (Mousterian) artifacts from Ayia 3.45. Palaeolithic findspots in the vicinity of Kastrosykia

3.46. Anavatis site/scatter 94-13, looking northeast 3.47. View of Rodaki (SS92-15)

3.48. Middle Palaeolithic artifacts from Rodaki 3.49. 3.50. 3.51. Early Upper Palaeolithic end scrapersfrom Spilaion Late Upper Palaeolithic backed blades

Palaeolithic and Mesolithic site/scatters in the Preveza area 118 120 120 121 122 123 123 124 136 139 139 140 141

3.52. Mesolithic artifacts from Tsouknida and Ammoudia 3.53. Mesolithic trapeze from Ammoudia 3.54. View of Ammoudia, looking northwest, with stone feature visible at left 3.55. Mesolithic artifacts from Loutsa 3.56. Typical Preveza Mesolithic findspot (SS94-23), looking southwest 3.57. Typical Mesolithic artifact scatter near Preveza (SS94-22) 3.58. Mesolithic artifacts from the Preveza area 4.1. Map showing the location of Spilaion at the mouth of the Acheron River 4.2. Map of Spilaion showing topographic contours 4.3. View of Spilaion, looking southwest 4.4. View of the rugged karst surface on the southeast slope of Spilaion at the time of collection 4.5. Sample grid on the southeast slope of Spilaion during collection

ILLUSTRATIONS

4.6. Lithic artifacts from Spilaion 4.7. Lithic artifacts from Spilaion 4.8. Lithic artifacts from Spilaion 4.9. Lithic artifacts from Spilaion 4.10. 4.11. 4.12. 4.13. Lithic artifacts from Spilaion End scrapersfrom Spilaion Spatial distribution of lithic debitage and retouched tools at Spilaion Spatial distribution of individual categories of retouched tools at Spilaion

145 145 146 146 146 146 151 152

5.1. Geology and geomorphology of the Ambracian embayment 158 and its vicinity 5.2. Locations of geologic cores and cross sections 5.3. Map of the Nikopolis isthmus showing the location of geologic cores and cross sections 5.4. Map of Ormos Vathy showing the location of geologic cores and cross section 5.5. 5.6. 5.7. 5.8. Stratigraphiccross section D-D', parallel to the axis of the Nikopolis isthmus Stratigraphiccross section E-E', parallel to the axis of the Nikopolis isthmus Stratigraphiccross section A-A', perpendicularto the axis of the Nikopolis isthmus Stratigraphiccross section B-B', perpendicularto the axis of the Nikopolis isthmus 159 163 163 165 166 170 171 172

5.9. Stratigraphiccross section C-C', perpendicularto the axis of the Nikopolis isthmus 5.10. Paleogeographic reconstruction of the eastern side of the Nikopolis isthmus showing the shorelines at different periods Stratigraphiccross section along the west arm of Ormos Vathy Paleogeographic reconstructions of Ormos Vathy indicating shoreline changes from the Neolithic through modern periods Stratigraphiccross section near the Grammeno plain

173 175

5.11. 5.12.

176 178 180 181

5.13.

5.14. Map of Kastro Rogon and vicinity showing the location of geologic cores and cross sections 5.15. Stratigraphiccross section C-C' at Kastro Rogon

ILLUSTRATIONS

XI

crosssectionB-B' nearKastroRogon 5.16. Stratigraphic crosssectionA-A' nearKastroRogon 5.17. Stratigraphic

183 185

Gulf crosssectionnorthof the Ambracian 5.18. Stratigraphic across and showingsedimentary sequences environments 187 the entirecoastalplain-lagoon-barrier system of reconstructions KastroRogonand 5.19. Paleogeographic the changingcoastlinesand environvicinityshowing 190-191 mentsfrom7000/6500 B.P. through1000/500 B.P. 5.20. Changesin relativesea level as indicatedby the radiocarbon-dated samplesfromswampdeposits peat northof the Ambracian Gulf reconstructions the Ambracian of 5.21. Paleogeographic embayment showingthe shorelinechangesfrom 7000/6500 B.P. through1000/500 B.P. 6.1. Areamapof Epirus

193

196-197

200

201 6.2. Areamapof the lowerAcheronvalley beachridgessurrounding 6.3. View of concentric accretionary 202 PhanariBay lake 6.4. Suggestedlocationsof the Acherousian in the lower Acheronvalley 6.5. Satelliteimageof Epirus 6.6. Simplified geologyof the lowerAcheronvalley 6.7. Corelocationsin the lowerAcheronvalley 6.8. Topographic mapof the lowerAcheronvalleybottom 6.9. North-southcrosssectionthroughthe Mesopotamon/ Tsouknida valleyconstriction 6.10. East-westcrosssectionthroughthe Mesopotamon/ Tsouknida valleyconstriction 6.11. Northeast-southwest crosssectionthroughthe valley bottom(areaof formermarineembayment) reconstructions the lowerAcheron of 6.12. Paleogeographic for 2100 B.C.andthe 8th centuryB.C. valley of 6.13. Paleogeographic reconstructions the lowerAcheron for 433 B.C.and 1 B.C. valley 6.14. Paleogeographic reconstructions the lowerAcheron of
valley forA.D. 1100 andA.D. 1500 223

203 206 207 210 211 218 219 220 221 222

6.15. Paleogeographic of reconstruction the lowerAcheron for A.D. 1809 and a mapof the modernlandscape valley

224

TABLES

1.1. Project Staff and the Yearsof Their Participation 1.2. Field School Students and Their Home Institutions 2.1. Stratified Sample and Systematic Survey Coverage, Lower Acheron Valley,1992-1994

10-11 12 31 33 61 67 68-69 70 71
78

2.2. Typical Daily Work Assignment, June 28,1994 3.1. Dimensions and Elevations of Poljes and Loutses in Western Epirus 3.2. Composition of the Fraction >0.064 mm in Redeposited Terra Rossa 3.3. Grain-Size Distribution of Redeposited Terra Rossa 3.4. Mineral Composition of Redeposited Terra Rossa at Kokkinopilos 3.5. Mineral Composition of Redeposited Terra Rossa from Poljes and Loutses in Western Epirus 3.6. Approximate Paleoshoreline Depths and Coastal Plain Widths, 140 kyr B.P. to Present 3.7. Mineral Composition of Modern and Last Interglacial Coastal Sands in Western Epirus 3.8. Maturity Indicators of the B Horizon of Greek Quaternary Paleosols 3.9. Short Descriptions and Maturity Stages of Paleosol Bt Horizons at Key Sites in Coastal Epirus

83 87 88 91 92 98

3.10. Thermoluminescence and Infrared Stimulated Luminescence Sediment Dates for Western Epirus 3.11. 3.12. ChronostratigraphicDiagram for Archaeological Sites, Sediments, and Paleosols in the Preveza Region Early Stone Age Chronology

TABLES

XIII

4.1. 4.2. 4.3. 5.1. 6.1

Categories of Flintknapping Debitage Types of Retouched Tools Degree of Association between Pairs of Classes of Flintknapping Debitage Radiocarbon Dates from the Ambracian Embayment Radiocarbon Dates from the Acheron River Valley

143 144 150 168 210

PREFACE

AND

ACKNOWLEDGMENTS

As editors of this volume we wish to thank the Hellenic Ministry of Culture for the approval of the permit to conduct archaeological surface investigations in southern Epirus, and to thank as well the directors of the 12th Ephoreiaof Prehistoricand ClassicalAntiquities and the 8th Ephoreia of Byzantine Antiquities, Angelika Douzougli and Frankiska Kephallonitou, for their positive recommendation to the Central Archaeological Council and their cooperation for the entire duration of the project. We also want to thank Evangelos Chrysos, then Professor of Byzantine History of the University of Ioannina (now at the University of Athens), for his many different contributions to the success of the project, and Nikolaos Yiannoulis, Mayor of Preveza during our investigations, who helped us in the resolution of a variety of problems that arose in the course of the project. We acknowledge the significant help in geological matters of Panayiotis Paschos, geologist of the Institute of Geology and Mineralogy Exploration (Preveza branch) and an expert in the geomorphological investigations of Epirus. During the fieldwork and the subsequentresearch in the facilities of the Archaeological Museum and the Byzantine Museum of Ioannina, to which the ancient artifacts collected in the surface survey had been brought, the project enjoyed substantial help from the scientific, technical, and security personnel of both ephoreias, to whom we express our warm thanks. The American School of Classical Studies at Athens approvedthe proposal for American participation in this cooperative project, and staff members of the project annually benefited from the superb library and other facilities of the School in Athens. We are grateful to the School, its staff, and its director during those years, the late W. D. E. Coulson. The former comptroller of the School, Joanna Driva, and the School's Administrator,Maria Pilali, were particularlyhelpful on numerous occasions, and it is a pleasure to acknowledge their congenial advice and cooperation. The project was sponsored in the United States by Boston University through its Department of Archaeology, the Center for Archaeological Studies, and the Center for Remote Sensing, all of which provided equipment and facilities to the project, and whose faculty, staff, and students have been supportive in many ways. Boston University also provided fi-

XVI

PREFACE

AND

ACKNOWLEDGMENTS

its nancial logistical and through Officeof International Programs, support field school as part of the projectin which sponsoredan archaeological of of 1992-1994.The Americancodirector the project(JW) was director and the fieldschool,andThomasF.Tartaron CarolA. Steinwereteaching instruction guidand also all assistants; seniorstaffof the project provided studieswere fully inteance to the students,whose field and laboratory All activities. staff and field school studentsare gratedinto the project's listed in Tables 1.1 and 1.2. Thomas L. Sever,now of NASAs Global Hydrologyand ClimateCenterin Huntsville,Alabama,and FaroukElwereboth of Centerfor RemoteSensing, Baz,director BostonUniversity's on and with advice remote-sensing of supportive helpful aspects theproject. Fundingfor the NikopolisProjectwas providedby grantsfrom the EarthObservingSystem,NASA in 1991;the NationalGeographic Soci1993-1995; and contribuety, 1992;the Institutefor AegeanPrehistory, tions throughout yearsof the project a number private the of individuby als, the Friendsof the Nikopolis Project,who are listed below. Special thanksaredueto fourof the Friends, MarthaSharpe and Joukowsky ArtemisA. W.Joukowsky, and JamesH. OttawayJr., MalcolmHewittWiener, for theirsupportandencouragement fromthe inceptionof the projectto its conclusion. for and and Equipment geophysical topographic survey for aerial was photography provided throughgrantsby the W. M. KeckFoundationto the CenterforRemoteSensing. Autodesk gavethe Nikopolis Inc. AutoCAD, Version12, for Projectcopiesof its superbdrawing program, eachof the threecomputer used Macintosh, DOS, platforms bythe project: andUNIX. Trimble lent the project GlobalPotwo Navigation Company CorsitioningSystemsforthe 1994 season.In 1993,the AppleComputer four to two porationcontributed computers the project, Quadra950s and two PowerBook 160s,which servedmanyof the computingneedsof the both in Greeceand in Boston.The Archaeometry of project, Laboratory the University Minnesota,Duluth,provided of substantial in personaid nel timeandsupport analyses. for we Finally, thankCarolA. Stein,a memberof the NikopolisProjectstaffandManuscript Editorat the American Schoolof ClassicalStudiesat Athens, for her congenial,thoughtful, and perceptive help in editingthis volumeandguidingit throughthe publication process.On behalfof the entirestaffof the project, acknowledge we with deepgratitude help andcontributions all. the of JamesWiseman Konstantinos Zachos

PREFACE

AND

ACKNOWLEDGMENTS

XVII

FRIENDS OF THE NIKOPOLIS PROJECT

BENEFACTORS Lloyd Cotsen and the NeutrogenaCorporation Dr. Martha SharpeJoukowskyand Dr. Artemis A. W. Joukowsky James H. OttawayJr. Malcolm Hewitt Wiener
PATRONS

Ms. Betty Banks Elizabeth Buntrock Leon Levy Dr. Anna MargueriteMcCann and Mr. RobertTaggart Professor P. Sullivant andJ. L. Godfrey J.
SPONSORS

Anonymous Mr.James R. JamesJr. Philip J. King Dr. William Ruf and Mrs. Elizabeth Ruf J. Robert Sewell
SUSTAINING MEMBERS

Anonymous Dr. Barbara Bell Doreen C. Spitzer Susan and Stephen Wiseman MEMBERS CONTRIBUTING Dr. PatriciaAnawalt ProfessorApostolos Athanassakis Robert S. Carter ProfessorMarian B. Davist Ernestine S. Elster,Ph.D. Dr. Howard Gotlieb In memory of StuartHaupt ProfessorG. L. Huxley Mr. Robert F.Johnston Michaelt and Susan Katzev Norma Kershaw Tom Lucia W. V. MacDonald

KatherineNordsieck LeonardV. Quigley,Esq. Eleanor Robbins Susan PetschaftRothstein Jane Ayer Scott Jane Dunn Sibley Judith P. Sullivan Professorand Mrs. Homer A. Thompsontt Dr. George Udvarhelyi Elizabeth LydingWill Donald and Rae Wiseman

CHAPTER

THE

NIKOPOLIS AIMS,

PROJECT: AND

CONCEPT,

ORGANIZATION
byJames Wiseman and Konstantinos Zachos

Human societies at all times and in all parts of the world interact with the landscape they inhabit: it could not be otherwise, even if the interaction were somehow limited to the selective exploitation of natural resources. Human activities alter the landscape and the natural environment, often in dramaticways; the alterations may occur as the result of human design, as in clearing a forest to plant crops, or may be incidental, as in the destruction (or reshaping) of a mountainside by Roman miners of precious metals. Conversely, humans at various times in the past have physically adapted to changes in their environment (especially in the distant past), or responded to environmental change in a variety of other ways. Some of these responses, such as migration or technological innovation, have been drastic and revolutionaryin their effect and are often recognizable in the archaeologicalrecord,while other responseswere more gradual,even subtle, and are more difficult to detect. To acknowledge the importance of the natural setting, of the environment at large, in studying change in human society is not to deny the importance of interculturalrelationships, or the role of the individual intellect or collective social conscience in the evolution of ethical, spiritual, or other sociocultural phenomena in human affairs.The point is that to understand and explain changes in human society over time, it is critically important to study society in relationship to the changing environment in which it existed. Through this approach to the past archaeologists are able to provide insights into the factors that underliechanges in human-land relationships,sometimes over a short timespan or even regarding specific events, but especially over the long term. And they can explore those intercultural relationships and sociocultural phenomena cited above, which themselves evolve within specific environmental settings and change. We have sought to apply these concepts in the formulation and conduct of the Nikopolis Project, an undertaking in landscape archaeology focused on the human societies that inhabited southern Epirus in northwestern Greece from earliest times to the medieval period. More specifically, the project has employed intensive archaeological survey and geological investigations to determine patterns of human-activity areas, and

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

Figure1.1. Map of Epirusand adjacent regions.The surveyzone is indicatedby crosses. what the landscape and other features of the natural setting were like in which those activities took place, in an effort to understand and explain observed changes in human-land relationships through time.1

THE CHOICE STUDY

OF SOUTHERN

EPIRUS FOR THE

Southern Epirus was selected for this broad diachronic study in part because, at the time, it was only in Epirus and in Thessaly that there was material evidence for something approaching the full range of prehistoric periods. Palaeolithic stone tools, for example, were first attested in Greece in the Louros River valley of Epirus.2 The area is also topographically diverse, including coastal regions, marshy lagoons, inland valleys, high upland plains in rugged mountain terrain, and mountain passes,3thereby providing a variety of environmental settings for different types of human activities that might be investigated by the project.What is more, prior to the Nikopolis Project there had been no large-scale, systematic, modern survey of the region, and most of the previous archaeological excavations were limited in a variety of ways.4The Nikopolis Project thus could be expected to enlarge our knowledge of a region that was not well known archaeologically. Another important considerationwas the existence in the surveyzone of Nikopolis, the "city of victory" founded by Augustus to celebrate his

1.This introductory section an is version the statement of of expanded

aims set out in Wiseman 1995a, p. 1, and uses some of the phrasingof that earlierformulation. 2. Dakaris,Higgs, and Hey 1964; Higgs and Vita Finzi 1966; Higgs et al. 1967. 3. Etudegdologique. 4. See below,"Previous Archaeological Work in the SurveyZone."

THE

NIKOPOLIS

PROJECT

3
I

*..11,-

Louros River

Acheron River

,~~\ ~
, Parga ,Kiperi Phnr^ . (Ammoudia Ephyra Vouv taos

f Voulista

~ ~Panayia

Yeoryios Ayios Thesprotiko Kastri kinopilos 'N2manteion KastroRizovouni , *Spilaion \ *Loutsa Aloaki. V Ch mLadiouros .:- dio Rogo, i;':: Palaiorophoros? Louros-Kastro ' / -' - :, Arachthos Cassope* ) '"X * Strongyli \ River K'
Kastkrosykm Grammeno Archan los

IC mian Sea

*1o
Nikopolis

Chlts .Tmas

SaIaor

OrmosVathy

y7."\ Prey .a ;:, Actm

Ambracian Gulf

Figure1.2. Map of surveyzone with selectedtoponyms

10

15

20

25 KM
..:.

victory in 31 B.C. over Antony and Cleopatra in the Battle of Actium. The creation of the urban population by the officially encouraged migration or forced removal to Nikopolis of populations from other cities of Epirus, Acarnania,Leucas, Amphilochia, and Aetolia,5 and the long life of Nikopolis as the metropolis of Epirus, raised a number of challenging problems regarding the relationship between the city and its territory to which the project'sresearchconcepts were directly applicable.The project thus takes its name from Nikopolis, the best-known toponym in southern Epirus. Finally, there was an urgent need for interdisciplinary survey before certain types of evidence, including some of the culturalremains, vanished as a result of various activities:land reclamation near the coast, the growth of the modern town of Preveza and several other smaller communities, industrial and agriculturaldevelopment, limestone quarrying,and other development activities related to tourism. These activities had wrought major changes on the regional landscape since 1950, and the pace of change in recent years had accelerated.

THE SURVEY ZONE

5. Kirsten(1987), Murrayand Petsas (1989, pp. 4-5), and Purcell (1987) all discussthe founding of Nikopolis and cite the most important sources.

The survey zone (Figs. 1.1, 1.2), about 1,200 km2, includes the entire nomos (administrativedistrict)of Preveza,a modern town on the Nikopolis peninsula, extending from the straits of Actium almost to the walls of the ancient city. On the east the survey zone extends into the nomos of Arta,

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

so that the entire deltaic, lagoonal area of the Louros River after its exit from its gorge at the modern town of Philippias was included; not included was the course of the Arachthos, a larger river east of the Louros which flows through the city of Arta (the ancient Ambracia) before emptying into the Ambracian Gulf, also known today as the Gulf of Arta. It is the western part of the north coast of the gulf, therefore, that lies within the surveyzone, from Salaoraon the east to the southerntip of the Nikopolis peninsula. The other boundaries follow those of the nomos of Preveza. That is, the western boundary of the survey zone is the shoreline of the Ionian Sea, from the straits of Actium on the south, where the Ambracian Gulf is linked to the sea, extending north beyond Ammoudia Bay (= Phanari Bay), at the mouth of the Acheron River, to Parga.The northern boundary of the survey zone runs east from Parga, along the middle Acheron River, and across the mountains to the narrowsof the Louros River gorge near the modern town of Kleisoura, below the ancient acropolis known locally as Voulista Panayia. The geology and geomorphology of southern Epirus are discussed in detail in Chapters 3, 5, and 6, so comments here are limited to observations of an introductory nature, primarily focusing on features providing general constraints on communication and exploitation of resources. A series of north-south Mesozoic limestone ridges, 600-1,000 m high, extends across the region from the Louros gorge to the Ionian coast, alternating with Tertiary flysch basins at elevations of 150-600 m, so that the basins provide now, as they did in the past, corridors of varying convenience for traveling north-south; fortified town sites of Archaic, Classical, and Hellenistic times are situated along the routes. Access to these natural corridors on the south is via passes through or between a series of mountains along the Ambracian embayment: from west to east, Mts. Zalongo, Stavros, and Rokia (see Fig. 5.1). The Louros River valley was an important communication route from early prehistoric times to the present; the principal road from Arta to Ioannina, present-day capital of Epirus, still passes through the gorge. The next basin on the west is most easily entered from the south between Mts. Rokia and Stavros, and a travelerwould pass near a fortified Classical and Hellenistic town site (Kastro Rizovouni) en route to the north and the passes that lead eventually into the valley of Dodona. The next basin to the west includes access to the upper Acheron River, and can be entered over a low ridge between Mts. Stavros and Zalongo. A bit furtherwest, the naturalroute is over a ridge of Mt. Zalongo, by the Classical and Hellenistic town of Cassope, and from there through a winding pass to the modern town of Kanallakion in the eastern part of the plain of the lower Acheron River. Agriculture is now practiced throughout the region, wherever it is possible to do so, in the upland valleys, along the courses of rivers and streams, and in the coastal areas. In the latter regions, especially around Ammoudia Bay and along the north coast of the Ambracian Gulf, swamps and marshyareashave been drainedduring the past half-century and flooding has been further controlled by the construction of canals, which also serve as conduits for irrigationof fields. Dams were built on both the Louros and Arachthos Rivers.There has been extensive work also in some of the

THE

NIKOPOLIS

PROJECT

upland basins; for example, a small lake (Lake Mavri) was drained in the basin east of Kastro Rizovouni to provide more arableland, and the deep waters of Lake Ziros in the same area are now being tapped for irrigation. The whole lower Acheron and the valley of its chief tributary,the Vouvos (ancient Kokytos) River,as far as the modern town of Paramythia(outside the survey zone) are now lush with vegetation, including a variety of cash crops and orchards.

PREVIOUS ARCHAEOLOGICAL SURVEY ZONE

WORK IN THE

account previous of 6. A detailed

investigationsin southernEpirus is

beingprepared K.Zachos. by
7. Dakaris 1971, 1975b, 1977,1978, 1979, 1980, 1981, 1982, 1983. 8. Dakaris 1958, 1960, 1961, 1962, 1963, 1964, 1975a, 1975b, 1977, 1993; Wiseman 1998. 9. Dakaris,Higgs, and Hey 1964; Higgs and Vita-Finzi 1966; Higgs et al. 1967. 10. Bailey et al. 1983a, 1983b; and Bailey,Papaconstantinou, Sturdy 1992. The investigationsin Epirusby

The most significant archaeologicalactivities in the largerregion in earlier years6were excavationsby Greek and German scholars at the ancient town of Cassope;7 Greek excavations at a site near the mouth of the Acheron identified by the excavatoras the Nekyomanteion, the Oracle of the Dead;8 and investigations by British scholars of Palaeolithic sites in the Louros River gorge to the northeast of Nikopolis.9 Recently the British renewed their interest in some of Eric Higgs's early work at Kokkinopilos and its environs (e.g., Asprochaliko),and carriedout limited surveyfor Palaeolithic remains along the coast.10Little was known of Neolithic, Bronze Age, and early Iron Age developments in the region, but the historical period was somewhat better represented in the scholarly literature.Important, useful studies of the region in antiquity were published by N. G. L. Hammond11 and by Sotirios Dakaris.12Both authors included copious topographical observations in their books and their researchinvolved some survey,which was, however, neither systematic nor intensive. Other archaeological investigations in the area have been limited to small-scale operations, usually involving salvageor preservationby the ephoreias,and have been briefly Deltion of the Greek reported over the years in the annualArchaiologikon Archaeological Service.

BACKGROUND PROJECT

AND ORGANIZATION

OF THE

as G. Bailey his colleagues, wellas and otherrecent worksomewhat further afield(e.g.,by K.Petruso Albania), in arediscussed, additional and publicationscited,by Runnels vanAndel and
in Chapter3. 11. Hammond 1967. 12. Dakaris 1971, 1972. 13. Paperspresentedat the symposiumwere publishedin Chrysos 1987. 14. Wiseman 1987, p. 413.

The Nikopolis Project had its origins in the First International Symposium on Nicopolis in 1984.13A paper presented by one of us (JW) focused on the need for the study of Nikopolis in its topographic setting, and suggested approachesto such a study.One specific recommendation, particularly relevant to the eventual development of the Nikopolis Project, was phrased as follows. A survey both of the naturalresources and the cultural remains of the region will be required if Nikopolis is to be studied in its regional context. What is more, the ancient topographic profile, including the changing coastlines, must be determined, along with climatic changes and the palaeoecology generally.14

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

and Remotesensing,includinggeophysical prospection, computer-aided as were discussedin the samepresentation usefultools to aid in analysis as such an undertaking, well as in the investigationof Nikopolis itself. methodwas in prospection, particular, citedas an important Geophysical ology by which at least somepartsof the city planof Nikopolismight be was beforeany excavation initiated.Symposiumparticipants established and interestedin the investigation preservawere deeply and organizers tion of the great city itself, and a coordinated, multifaceted, long-term effortwas formallydeclaredby the symposiumboardto be a desirable outcomeof the symposium.15 led Continuedconcernfor Nikopoliseventually to the appointment of MelinaMercouri, a special in 1986 by the GreekMinisterof Culture, Committee for the Preservationof Nikopolis, which was headed by of Evangelos Chrysos(nowbasedat the University Athens),who wasthen of Professor ByzantineHistoryat the Universityof Ioanninaand one of The committeemembersrepresented of the organizers the symposium. in Greecewith concernsor responsibilities the groupsand organizations for Nikopolis,includingthe GreekArchaeological Service,the Archaeoof the of Athens,the city of Preveza, University Ioannina, logicalSociety hiredby the committeeweregiven an officein the andothers.Architects and Town Hall of Preveza, they beganthe important jobs of mappingall and visible remainsin Nikopolis and its periphery, of documentingthe zone of Nikopolis. within the archaeological ownershipof all properties in The committeewas reconstituted occasionally the 1990s to reflectpolitical(bothlocalandnational)andinstitutional changes,but Chrysosreof the permutations the committee tainedthe chairmanship throughout until the completionof the NikopolisProject. in of Wisemanbegandiscussions With the encouragement Chrysos, 1988 with AngelikaDouzougli,the newly appointed (direcproistameni and of and tor)of the 12th Ephoreia Prehistoric Classical Antiquities, her in Konstantinos Zachos,seniorarchaeologist the sameephoreia, husband, in which collaboration a project the Nikopolisregion, on possible regarding lies within the purviewof that ephoreia. The 8th Ephoreiaof Byzantine also directed Frankiska by Antiquities, Kephallonitou, becameinvolvedin the earlyplanning,becauseLate Antique and Byzantineremainsin the of The decisameregionwere amongthe responsibilities that ephoreia. in both basedin loannina, sion was reached 1990 that the two ephoreias, of wouldjointlyshare responsibilities the project, the andBostonUniversity for when finalized, for a joint underwas so that the proposal the project, in The of taking,synergasia Greekterminology. directors the two ephoreias of with Wiseman,the Ameriand K. Zachoswerecodirectors the project of were canPrincipal and Investigator, otherrepresentatives the ephoreias first wasthen submitted to alsomembers the staff. of The project proposal the AmericanSchoolof ClassicalStudies,as then required Greeklaw by or for a projectinvolving Americansponsorship cosponsorship. basedon Therewas for a time consideration a collaborative of project that would carry with the group Nikopolisitself,workingin cooperation The out the regionalstudy,as envisionedat the Nikopolissymposium.16 at aimsof workat Nikopoliswouldhavebeen to determine least principal

15. Chrysos 1987, pp. 417-418. 16. Wiseman 1987.

THE NIKOPOLIS

PROJECT

17. van Andel and Runnels 1987; Jameson,Runnels,and van Andel 1994.

and the generaloutlineof the city plan throughgeophysical prospection froma tethered otherformsof remotesensing; photography blimpboth to of help in detectingthe townplanandto aidin the documentation aboveand intendedto providea stratigraphic groundremains; test excavations an controlfor regional ceramics, urgentneedbecausetherewerethen few These planswereabandoned in published groupsof well-datedceramics. as it becameclearthatthereweretoo manyconflicting competand 1991, rightsat Nikopolisitself for any one group, ing claimsto archaeological a Council especially new one, to obtainthe supportof the Archaeological for archaeological for approving in Athens, the responsible body permits of investigations any kind in Greece.The proposalas finallysubmitted wasfor a combinedarchaeological geologicalsurvey the region,but and of For not includingNikopolis,conductedin synergasia. 1991, the project of would involve mainlyground-truthing satelliteimageryand gaining with by greater familiarity the landscape the American staff,andfinalizing The the aims and methodologyof the regionalinvestigation. subsequent permitwas for threeyears,1992-1994, duringwhich the archaeological and geologicalinvestigations were carried There were studyseasons out. in the summers 1995 and 1996,when seniorstaff,basedin Ioanninato of were able to materialscollectedduringthe survey, study archaeological revisitthe surveyzone with staff reportsin hand and to discussproject and resultsandinterpretations. analyses studyboth of the artiLaboratory factsandthe archives havecontinuedsincethat time. A numberof scholars Greece,the United States,the United Kingin research contributed the eventual to dom, andothercountries design,inboth specificresearchaims and methodologiesadoptedby the cluding Nikopolis Project,especiallythose who have devoted so much of their time andeffortas members the staff.George(Rip) Rapp,a geoarchaeof at the University Minnesota,Duluth,with extensive of field expeologist riencein Greeceandotherpartsof the eastern was Mediterranean, one of the firstscholars and invitedto join the staff;he organized directed much of the project's and shorelinestudies. coringprogram, geologicalsurvey, at CurtisRunnels,an archaeologist BostonUniversity, broughthis expertisein the early of to prehistory Greeceandin survey the NikopolisProject. He wouldleadthe Palaeolithic of with the aid andcooperation his survey, Priscilla Fellowin Archaeology BostonUniverat Research wife, Murray, of Unisity,andTjeerdvanAndel, a geoarchaeologist formerly Stanford of then (andnow) of the University Cambridge. Runnelsandvan versity, Andel would now applysurveytechniques they hadjointlydevelopedon in Greeceto the investigation earlyhumansandhomiof projects southern nids in Epirus.17 Their survey, which supplemented, was conducted but carried by otherstaff,inout from,the intensivesurface separately survey volvedintensivegeomorphologic studiesin the detectionof Pleistocene whichtheythensearched. Bothwouldalsojoin in otherproject landscapes, of for in stone responsibilities-Runnels, example, the analysis prehistoric andvanAndel in geomorphology all periods,as well as providfor tools, concerns. ing counselandinsightfor allgeoarchaeological LucyWiseman of BostonUniversity's CenterforArchaeological Studieswas alsoa member of the stafffromthe beginning,servingboth as projectadministrator

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

and registrarof artifacts.Three advancedgraduate students in archaeology at Boston University were also part of the senior staff. Thomas Tartaron and Carol Stein were the primary team leaders in archaeological survey, and provided both supervision and guidance for others who subsequently became survey team leaders. Tartaron also developed a specific sampling strategy for the Acheron River valley and Ayios Thomas peninsula, reflecting the overall stratified sampling strategy of the project, and carried out a special study of the Bronze Age sites and materials, part of which was included in his doctoral dissertation.18Melissa Moore oversaw the study and registration of ceramics, and part of her research has been included in her Ph.D. dissertation.19Other staff and consultants included geologists, computer scientists, archaeologists, and specialists in various other fields; all staff and their affiliations during the Nikopolis Project are provided in Table 1.1. Students enrolled in a Boston University Archaeological Field School were invaluablemembersboth of the field surveyteams and the geological coring and survey units in 1992, 1993, and 1994. As a part of their archaeological training, they participated in all activities of the project in Greece, including the processing of artifacts,data processing on computer, digitizing of maps, ground-truthing of satellite imagery,topographical survey,geophysical prospection, aerial photography by tetheredblimp, and other investigations.Their names and the institutionswhere they were studying at the time are listed in Table 1.2.

SPECIFIC RESEARCH AIMS

Research aims, nested within the larger conceptual framework described above, relate mainly to specific time periods and include the following topics, phrased as questions, which much of the project's fieldwork was intended to answer. 1. What forms do the cultural remains of the earliest inhabitants of southern Epirus take, and how may we explain their distribution in the different periods of the Palaeolithic?What resources were exploited by the early humans and hominids, and what was the environmental setting? 2. What is the evidence for the shift from hunting/gathering groups to agriculturalsocieties? Can that shift be related to changes in the landscape? 3. What was the nature of the contacts between peoples of this region in later prehistoric times, especially in the Late Bronze Age, and groups on the shores of the Ionian Sea, in other parts of Greece, and more generally in the eastern Mediterranean? Do these contacts differ in quality during fully historical times? 4. How are colonial activities of southern Greeks manifested in this region? 5. What were the effects of the development of political leagues and interregional alliances on settlement patterns, sizes of sites, religious centers, and resource exploitation in Classical and Hellenistic times?

18. Tartaron1996. 19. Moore 2000.

THE

NIKOPOLIS

PROJECT

6. What were the effects of the historically documented Roman intrusion into Epirus (which was also the earliest intervention by Romans in Greek affairs) in the 3rd and 2nd centuries B.C., and how may they be identified in the landscape?How intrusive into local society were the Romans, and what activities (military,industrial, commercial, social, etc.) are indicated by the cultural remains? involved in the 7. What was the regional effect of the synoecism founding of Nikopolis by Octavian, later Augustus, first emperor of Rome? How are the new patterns of settlement and communication related to changes in the landscape itself? 8. What was the nature of the exploitation of the countryside in the Late Antique period (4th-6th centuries A.c.) and how was it related to the socioeconomic transformation into medieval times? More specifically,what was the economic basis of southern Epirus in late antiquity and in medieval times? When did the extensive exploitation of wetlands along the Ambracian Gulf begin, and when the deliberate reclamation of land from coastal lagoons?

METHODOLOGIES
The research design called for the archaeological sampling by intensive surface survey of all environmental zones: coastal plains, inland valleys, mountainous terrain,and upland valleys.The large size of the surveyzone precluded archaeological survey over the entire region. The selection of the areasto be surveyedwithin each environmental zone would be guided primarilyby acquiredknowledge of the region. Geological surveyand other geomorphologic investigationsprovidedimportantinformation,both negative and positive, influencing the selection of fields and transectsto survey; fieldwalking teams, for example, could avoid areas of recent alluviation where remains (if any) of prehistoric-medieval times would have been covered over and not detectable.The location of early historical or even Pleistocene landscapes exposed by erosion, on the other hand, offered opportunities for survey with greater expectation of detecting archaeological remains. Even so, occasional surveyswere conducted to test negative indications from geomorphology or satellite imagery,20 when fieldwalking as teams spent a day walking transects across the presumed relict coastlines of Ammoudia Bay that were formed by long-shore deposition in recent historical times. The negative results of the intensive survey confirmed the geomorphologic conclusions and the interpretations of imagery.The degree of visibility was recorded for all areas surveyed. Fields where vegetation was too dense for archaeological remains to be seen during preliminary reconnaissance were not selected for survey. This practice is an important consideration in evaluating the results of the survey,because in some other year, or some other time of year,those fields might be clear of vegetation, and might, of course, yield archaeological materials. On the other hand, in some instances fieldwalking teams were able to return to a region to survey fields that had been too densely covered for survey in a

20. A practicerecommendedin Sever andWiseman 1985, pp. 70-71.

IO

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

TABLE 1.1. PROJECT

Name
CODIRECTORS

STAFF AND THE YEARS OF THEIR

1991 1992

PARTICIPATION
1994 1995 1996

1993

Angelika Douzougli/KonstantinosZachos, 12th Ephoreiaof Prehistoricand ClassicalAntiquities Frankiska Kephallonitou,

8th Ephoreia of Byzantine Antiquities James Wiseman
ADMINISTRATION AND INVENTORY

0*
0 0 0

Lucy Wiseman (registrar of artifacts, administration) Melissa Moore (registrar of ceramics,archaeology) Lia Karimali (lithics, survey) Dimitra Papagianni, University of Cambridge (lithics, survey)

* 0

* *

* * * * * * e * * 0 *

KaterinaDakari,8th Ephoreiaof ByzantineAntiquities

(survey, Late Antique ceramics) Ricardo Elia (associatedirector,archaeology) Asymina Kardasi, Athens (Byzantine ceramics) * *

Stavroula Vrachionidou,12th Ephoreiaof Prehistoricand

Classical Antiquities (administration, survey) ARCHAEOLOGY, SENIOR STAFF * * *
*

Timothy Baugh (remotesensing, ground-truthing) Brenda Cullen (survey, remotesensing) Priscilla Murray (survey, drafting) Curtis Runnels (field director,Palaeolithic survey; lithics) Carol Stein (survey, remotesensing) Thomas Tartaron (survey, ground-truthing)

* * * *

* *
0

* *

* *
S

StavrosZabetas,Greek ArchaeologicalService (survey)

GEOLOGY AND GEOPHYSICS

Mark Besonen, Universityof Minnesota, Duluth

(geologicalsurvey, coring) Richard Dunn, University of Delaware (geologicalsurvey, coring)

ZhichunJing, Universityof Minnesota, Duluth

(geologicalsurvey, coring) Jon Jolly, Seattle, Washington (oceanography,instrumentation) * *

George (Rip) Rapp,Universityof Minnesota, Duluth

(geology,geoarchaeology) * *
0

Apostolos Sarris,Athens, Greece (geophysics)

Marie Schneider (geology,survey)

Tjeerdvan Andel, Universityof Cambridge

(Pleistocenegeology, geomorphology,geoarchaeology) * * *

Sytze van Heteren (geology) John Weymouth,Universityof Nebraska(geophysics) Li-Ping Zhou, Universityof Cambridge
dating) (geology, thermoluminescence

THE NIKOPOLIS

PROJECT

II

TABLE 1.1-Continued
Name
COMPUTER SCIENCE

1991

1992

1993

1994

1995

1996

Robert DeRoy (computerscience,remotesensing) Daniel Juliano (computerscience,remotesensing)

Rudi Perkins,Bangor,Maine (computer science)

PHOTOGRAPHY Michael Hamilton (aerial photography,generalphotography) Eleanor Emlen Myers' (aerialphotography) *

J. Wilson Myers (aerialphotography)

TOPOGRAPHICAL SURVEY AND DRAFTING

Theodoros Chazitheodoros,Greek ArchaeologicalService,

Athens (topographicalsurvey, drafting) David Clayton (topographicalsurvey, drafting) *

Athina Kotsani,Preveza(drafting)
Kostas Papavasileiou, Preveza (architecture,drafting)

a a

Anne Van Dyne, Seattle,Washington

(topographicalsurvey, drafting) GENERAL STAFF
0 a

Stephen Agnew (ground-truthing)

KaelAlford (survey)
Alesia Alphin (survey, inventory)

Betty Banks, Spokane,Washington (survey,inventory,

data entry) Mark Greco (survey) Cinder Griffin, Bryn Mawr (survey, inventory) *
a a

Nikola Hampe, Universityof Miinster (survey) Alan Kaiser(survey) PetraMatern, Universityof Miinster (survey)
Michele Miller (ground-truthing, survey) Lee Riccardi (survey, inventory) * S

0
*

KatrinVanderhuyde, Universityof loannina (survey) ElizabethWiseman, Littleton, Colorado

(photography,ground-truthing) CON SULTANTS

Wilson College (ceramics) VirginiaAnderson-Stojanovic,

Evangelos Chrysos, University of loannina (Byzantine history) * * *

*
* * *

HarrisonEiteljorgII, Bryn Mawr (databases, AutoCAD)

Panayiotis Paschos, IGME, Preveza (geology)

Staff memberslisted without an institutionalaffiliationor city were from Boston University.

I2

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

TABLE 1.2. FIELD SCHOOL

I992

STUDENTS

AND THEIR

HOME INSTITUTIONS

KaelAlford, Boston University AlexandraBienkowska,Boston University Anne Cockburn,Williams College Todd Gukelberger, SUNY, Albany Deborah King, RensselaerUniversity Dawna Marden,Universityof SouthernMaine Thomas Matthews, Utica College of SyracuseUniversity RichardRotman,Boston University Bayleh Shapiro,Boston University Jane Sontheimer,Boston University Anita Vyas,Boston University ErikaWashburn,Boston University 1993 AlessandroAbdo, Boston University Evie Ahtaridis,Universityof Pennsylvania TracyBarnes,Texas ChristianUniversity Arlyn Bruccoli,Bard College ChristinaCalvin, George Mason University Scott deBrestian,Boston University Antonina Delu, Universityof California,Riverside KatherineDemopoulos, Universityof California,Los Angeles Cheryl Eckhardt,Boston University JenniferFisher,Boston University Lorena Freeman,Universityof the South Stephani Kleiman,Loyola MarymountUniversity Noah Koff, Boston University

Natalie Loomis, TulaneUniversity Michael Marton, Franklinand MarshallCollege Martin McBrearty,FurmanUniversity Scott McCrimmon, Boston University Sean Mulligan, Boston University Wendy O'Brien,Boston University Dena Pappathanasi, Universityof New Hampshire Rudolph Perkins,Boston University Duke University Jamie Ravenscraft, JonathanWood, PrincetonUniversity KellyYounger,Loyola MarymountUniversity 1994 Lisa Davis, HarvardUniversity Mely Do, Universityof Pittsburgh Aviva Figler,Boston University Mike Gaddis, PrincetonUniversity Amy Graves,Miami University Leslie Harlacker, Boston University KarlaManternach,Loras College Joe Nigro, Boston University Anne Maxson, Duke University KathyMontgomery,Boston University JenniferMurray,SUNY, Buffalo VersalliusCollege, Brussels StephanPapageorgiou, T. J. Reed, Cornell University YasuhisaShimizu, Boston University Alison Spear,Mount Holyoke College

previous year.The methodology of the surface survey is discussed in detail by Tartaron in Chapter 2, but it is important to note here that surface surveys included both transects within large regions and intensive sampling, or complete coverage, of human-activity areas ranging from small single-activity sites to extensive settlements. In addition, one fortified town site (Kastri, in the lower Acheron valley) was selected for intensive urban survey. Geomorphologic studies formed part of the central core of the project, as required by the research concept. If we were to study the interaction between humans and their environment,we reasoned,one of the first steps must be to determine what that natural setting was-that is, what the landscape and other aspects of the environment were like over time. A number of investigations, therefore, were planned to provide the needed evidence. An extensive coring program was initiated in 1992 and continued through 1994 that was aimed at determining changes in shorelines over time both in the Ambracian Gulf and along the Ionian coast. The analyses of the cores, most of which were carriedout in the Archaeometry Laboratoryof the University of Minnesota, Duluth, also made it possible to establish a sequence of local change and, through radiocarbondating, to determine the chronology of change. Cores also provided microfauna,

THE

NIKOPOLIS

PROJECT

I3

21. See the discussionsin Wiseman 1992b, pp. 3-5; 1993a, pp. 12-13. 22. The following brief accountis intended mainly to explainwhat kinds of remote-sensingimagerywere acquiredand used by the project,and why they were used. 23. Wiseman 1996a, 1996b. 24. Stein and Cullen 1994; Wiseman 1996a, 1996b.

macrofauna,and pollen for paleoenvironmental reconstruction. Geomorphologic investigations involved geological survey in all parts of the survey zone, and intensive work, including coring and mapping, at selected sites or regions. Geological survey and coring were coordinated as closely as possible with the archaeological survey, so that field teams often comprised both geologists and archaeologistsworking together. We had planned offshore investigations to supplement the study of shoreline change, and there was a promising beginning to that research. The Hellenic Navy dispatched a research ship, the Pytheas,to work with project staff for two weeks in 1992. A Klein side-scan sonar and a Klein subbottom profilerwere towed behind the ship both in the Ionian Sea and in the Ambracian Gulf, the former recording features on the surfaceof the sea bottom, the latter detailing the depth and nature of sediments below the sea floor.The survey,in perpendiculartransects forming a grid pattern, produced data covering some 300 linear kilometers, which to this date have received only preliminary analysis21 because they were subsequently another bureau of the Greek government. sequestered by Remote sensing from space was determined to be a potentially useful tool for our surveywell before the initiation of the project,as noted above.22 We did not, however, expect remote-sensing imagery to play a significant role in the detection of archaeological sites because at that time most remote sensors were known to be unsuccessful in penetrating dense vegetation, which covered much of our survey zone.23What is more, although the resolution of satellite imagery had been improved, the smallest picture element (= pixel) of available multispectral imagery was 20 meters to a side, too large to be helpful in detecting the small features and artifacts of most archaeologicallandscapes. It is an interesting sidelight on the development of archaeological methodologies that remote sensing in the end proved to be quite useful in detecting Pleistocene landscapes,which could then be located and searchedby ground-truthing survey teams, and which resulted in the discovery of five prehistoric sites.24Its greatest value, we thought at the time, would probablylie in its ability to provide imagery of the entire region that would permit the classification and identification of present-day land cover. It could, therefore, help in defining the environmental zones of the survey area;show currentconditions that might affect the conduct of surfacesurvey;and perhapsprovide some insight into routes of communication among known (or subsequently discovered) ancient settlements. The imagery would also serve as a layer in the computeraided GIS (geographic information system) maps to be generated by the project, and we hoped to develop spectral signatures-that is, a characteristic spectralresponse identifiable in the imagery-for features of archaeological interest. Both multispectral (MSS) and panchromatic imagery of the entire surveyzone was acquiredfrom the French satellite company SPOT before the beginning of fieldwork in 1991. SPOT imagery was selected primarily because its spatial resolution was the finest availablefor general researchat that time: MSS at 20 meters, panchromatic at an even finer 10 meters. The United States'Thematic Mapper (TM) satellite imagery,in contrast, has a resolution of 30 meters. Since spatial resolution on the ground is a

I4

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

Figure 1.3. Multispectral image (SPOT) of the northern part of the survey zone

Figure 1.4. Multispectral image (SPOT) of the southern part of the survey zone. Leucas (lower left) and other regions south of the Ambracian Gulf lie outside the survey area.

THE

NIKOPOLIS

PROJECT

I5

function of altitude as well as the type of sensor, we could have achieved finer resolution from sensors mounted on aircraft, instead of spacecraft. The only airborne platform available to the project, however, was a tethered blimp, which, although excellent for individual sites and smaller areas, was not appropriatefor such a large regional survey as ours because of the time and other logistical difficulties such coverage would require.Full coverage of the survey zone required two images, both in MSS and panchromatic. The northern image (Fig. 1.3) included almost the entire survey zone, and the second (Fig. 1.4) added the southernpart of the Nikopolis peninsula, along with areas outside the survey zone: Actium, Leucas, and other areas south of the Ambracian Gulf. Multispectral imagery is particularlyuseful in showing different types of landcover because landcover types have a different reflectance value in each band of the electromagnetic spectrum. The combination of these numeric values in the bands used by the sensor (SPOT uses green, red, and near infrared) constitutes a spectral signature, which may be represented by a (false) color assigned in a multispectralimage generated by the computer.This assigning of colors, or classification of images, is a process whereby each land area having the same kind of cover receives the same (false) color in the image. The researcher,then, after identifying on the ground at least once the class representedby a particularcolor as a particular landcover (e.g., class 12 = red = limestone outcropping), may reasonably expect other patches of red in that image to represent the same kind of landcover;in the example just cited, more limestone outcrops. In practice, however,the classification of an image may result in the combining of several signatures into a single class, or the subdivision of a signature into more than one class, depending on the number of classes the researcher chooses for the image and on other physical aspects of the landcover.Making use of the facilities of the Center for Remote Sensing at Boston University, Carol Stein classified the MSS imagery of the Nikopolis Project into fifty classes,with all unclassifiedlandcoverassignedclass 0. The number of classes was considerably larger than proved useful in the field because the fine distinctions the classification made possible resulted in the identification of many kinds of landcover that were irrelevantfor our research. For example, there was no reason for us to be able to distinguish kiwi plants from maize, which our classification enabled us to do. In retrospect, we now see that fewer landcover classes (say,fifteen to twenty) would have been preferable,because such a classificationwould have resulted in a beneficial lumping together of rock outcroppings, and would have created other continuous zones-as in fact they were-of barrenland, instead of a number of separate units in the classified imagery.The finer distinctions involved in developing a spectral signature of an archaeologicalfeature, or archaeological feature combined with a particularvegetation, would still have been theoretically possible. The relevant portions of the MSS images were then subdivided by Stein into twenty scenes, each representing about 100 km2on the ground, and printed for field use. Transparentoverlays at the same size were also printed, five for each scene, each displaying ten of the fifty false colors of classes of landcover,so that field teams were able to use them conveniently

I6

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

Z ~ ~. W

of Figure1.5.The erodedlandscape ~ .......Kokkinopilos abovethe Louros . _ ~~~~~ ~ __ ............ .River gorge

to determine what on the ground was actually being represented by each false color; this kind of fieldwork is called "ground-truthing."The hard copy of the scenes and transparencieswere at a scale of 1:50,000, so they could be used in conjunction with our topographical maps of the same scale; the transparenciescould be used as overlays of the maps, just as they were on the printed scenes. Ground-truthing, a focus of our fieldwork in 1991, required precise location of the observed landscape, so the field teams were also provided with copies of the panchromatic scenes, and even more detailed subscenes. Locations were marked on 1:5,000 topographical maps, and aerialphotographs (scale: 1:20,000) also were used to help locate specific features in the landscape; both maps and photographs were obtained from the Geographic Service of the Hellenic Army. Additional locational information was obtained by 1) global positioning systems (GPS), which provideUTM as well as longitude/latitude readings through communication with the navigational satellites (21 in number in 1991) that constantly orbit earth; 2) altimeter readings (more accurate at that time than GPS in determining altitude), when benchmarks are not readily available;and 3) readings by electronic laser theodolite, for still more precise location in three dimensions, as appropriate.These ground-truthing expeditions, which were led by Timothy G. Baugh during the first, preparatoryfield season, resulted in the identification of 27 of the 50 classes of landcover.An additional 12 classes were created for areaswith distinctive features related to human activity whose spectral signatures might serve as guides to the location of other similar areas:e.g., quarries or ancient sites. One of those new classifications was the eroded Pleistocene landscape of Kokkinopilos (Fig. 1.5), which eventually led to the discovery of five other similar landscapes, and prehistoric sites, as mentioned above. The experience gained in using GPS, satellite imagery,and topographic maps in 1991 was invaluable in developing standardproceduresfor the surveyteams of 1992-1994.

THE

NIKOPOLIS

PROJECT

I7

25. Hemans, Myers, and Wiseman 1987. 26. Hemans, Myers, and Wiseman 1987.

of What is more,the ground-truthing several expeditions 1991 provided with the Epirotelandmembers the staffwith a fundamental of familiarity scape. Anotherkind of remotesensing,aerialphotography from a tethered wasemployed the project documentsomeof the larger to known by blimp, ancientsites.Foursiteswerephotographed radio-controlled with cameras in 1992 by field teamsled byJ.Wilson MyersandEleanorEmlenMyers: the fortified townof Kastro Rogonsouthof the LourosRiver gorge;Kastro a fortifiedtown in an enclosedplain northof KastroRogon; Rizovouni, the Romanaqueduct nearAyios Georgiosin the LourosRivergorge;and VoulistaPanayia, Hellenisticsite overlooking narrows the same a the of further northat Kleisoura. MichaelHamilton,who wasthe project's gorge staffphotographer, the blimp-photography led teamin 1993 that photothe and graphed largefortifiedClassical Hellenisticsite at the abandoned modernvillageof Palaiorophoros, northof the townof Louros. The use of this techniquewas limitedby a numberof factors. The necessityfor permits frommultiplecivilianand military authorities resulted numerous, in of anddisruption schedules in costlydelays (e.g.,blimpphotography 1991 had to be cancelledand the 1993 seasonwas severely The excurtailed). and increased 1993whenwe decided, in pensewas significant, wasgreatly for safetyreasons, use heliumin the blimpinsteadof less expensive, to but flammable In addition, therewerethe normaldelaysand highly hydrogen. logisticalproblemsimposedby the techniqueitself, such as the need to awaitfavorable winds (thatis, none or verylight) andotherclimaticconditions.The photographic resultsof this technique,however,are highly are useful,especially when, as on the NikopolisProject, multiplecameras usedto provide both in blackandwhite andin color.A particular coverage of froma tethered advantage photography blimpis thatthe viewsarevertical and so can be used in mapping,unlikethe obliqueviews frequently It gatheredby camerason aircraft. is also possiblein a single flight to obtainphotosat a seriesof elevations to a maximum 800 m, thereby of up both close-upsand extensivecoverage Fig. 1.6).The aerial (see providing also can be scannedand then combinedwith the multispecphotograph tralimage of that area,a techniquewe used in the studyof the fortified town site of Palaiorophoros. The BostonUniversity blimp-photography systemwas designedbyJ. Wilson Myers,who modeledit on the systemhe had developedearlier, and is described detailelsewhere.25 multispectral in A video camera, sucusedon a tethered cessfully blimpin the Corinthia a BostonUniversity by teamin 1986,26 not usedby the NikopolisProject, couldusefully was but be deployed the future, in sinceit canprovide in high spatialresolution six bandsof the electromagnetic spectrum. of kindswas carried at a number out Geophysical prospection various of sites,primarily providedataon possiblesubsurface to in features areas where surfacesurveysuggestedsignificanthumanactivity. Only limited waspossiblein 1992 becauseof staffingandequipment prospection probwere conductedin 1993 lems, but successfulprogramsof investigation underthe direction JohnWeymouthof the University Nebraska of of and

I8

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

Figure1.6. Aerialview of the fortifiedtown site at KastroRogon

from an elevation of 400 m. Photoby
J. Wilson and Eleanor Emlen Myers

in 1994, when Weymouth was succeeded by his protege, Apostolos Sarris. Instrumentation included a proton magnetometer, electrical resistivity meter, and electromagnetic conductivity meter, of which the first was most frequently used. Weymouth and Sarris are preparing a report on their investigations for volume 2 of this series, and the results are also being incorporated into reports on the town sites where geophysical prospection detected significant subsurfacefeatures such as probable kilns and buildings. The permit of the Nikopolis Project was for survey, not excavation; indeed, under Greek law a single permit might cover only one or the other. As a result, the project had an arrangementwhereby one of the cooperating Greek ephoreias would perform excavation if a site was discovered by the project to be in need of emergency attention. The discovery at the Roman villa site of Strongyli, for example, that burialshad been plundered by clandestine diggers and parts of floor mosaics had been exposed A prompted excavations by the Greek ephoreia to ensure conservation.27 similar situation arose at Frangoklisia, probably another Roman villa, on the Ionian coast near Loutsa.28The project did carry out limited excavation in 1991 at the request of the Prehistoric and Classical Ephoreia in the Roman aqueduct below the village of Ayios Georgios, so that details of the water channels and the chronological sequence of aqueduct bridges across the Louros River might be studied and drawn (Fig. 1.7). Our work here resultedin, among other conclusions, the confirmation that the northern bridge was built and utilized for the aqueduct after the earlier,Augustan bridge had been damaged and abandoned.

27. Douzougli 1998a, 1998b. 28. Zachos 1998.

THE

NIKOPOLIS

PROJECT

I9

Figure1.7. Aerialview of the water channel(right)andaqueduct bridges acrossthe LourosRiverfroman elevationof 320 m. Photo by
J. Wilson and Eleanor Emlen Myers

i .
'
s

t
a f

DOCUMENTATION
All team leadersand individual investigators kept a dailyrecordof their activities observations bound,hardback and in which alsoconnotebooks, tainedphotographic printsanddrawings, wereindexeduponcompleand tion.The notebooks werenumbered This historisequentially. permanent cal record,partiallyin narrative form,was supplemented an arrayof by printedformsthatwerefilledout in the fieldor laboratory, appropriate, as providingdetailedinformation all aspectsof the investigations, on from surface surveyto artifact These two kindsof writtendocumeninventory. tationwerecross-referenced a dailybasis,but it wasprimarily series on the of printedformsthat providedthe bulk of the information that was enteredinto the computer I databases. summarize belowthe principal databasesof the NikopolisProject. formswere numbered yearand seAll by with an quentialaccessionwithin the year,e.g., 92-1. Databasesmarked asterisk dealtwith in greater are detailin Chapter2. 1. Ground-Truthing Form(GTF). A GTF was filledout at every locationwhereground-truthing conductedto identifythe was landcover classesin the satelliteimagery. of *2.Tract(T). The tract,an areaof arbitrary is the project's size, or primary surveyunit whetherin the countryside within a large site.The database includeslocation,size, description, conditions of the survey, total artifact results. counts,and summary

20

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

*3. Site/Scatter(SS). An SS is anylocationwheretherewas a or of concentration artifacts that is marked visible,in situ by This categoryincludesanylocationfroma small remains. includes scatterof lithicsto a fortifiedtown.The database and location,size, description, chronology, surveydata. *4.Walkover (W). A W indicatesa nonintensive surveyor a visit or eitherfor reconnaissance reexamination. includesthe description, The sampledatabase 5. Sample. counts, collectedduring material dates,and otherdetailsof all cultural of are survey. Samplenumbers identicalto the numbers the surveyunitswheretheywerecollected. and werecatalogued Artifactsselectedfor inventory 6. Inventory. into Selectioncriteria storedaccording material/function. for cluded,amongothers,significance datingor functional as or analysis, the likelihoodof publication a type artifact. a recordof the context This database 7. SpecialAnalyses. provides fromclay and natureof samplestakenfor laboratory analyses, samplesto geologicalcores. A and 8. Photo Inventory. recordof all black-and-white color
photographs taken by the Nikopolis Project, in the field, photo studio, or laboratory. 9. Drawing Inventory.A record of all drawings made by and for the
project.

Relational databases 1-6 were all created in FoxBase+ for Mac, which seemed to the staff, including the computer scientists and engineers, the most suitable at the time. Unfortunately,when the softwarewas redesigned as FoxPro in 1993, databases in earlierversions of the software could not be upgraded;all windows for data entry would have had to be redesigned and the data reenteredto use FoxPro,a duplication of effort we declined to do. The program, therefore, lacks some of the flexibility and ease of some of the more recent databases, but still has served the project well. The design of the relational databases reflects the archaeological concerns and experience of the senior staff, and there was much (both fruitful and lively) discussion between the archaeologists and the computer experts who put it all together. The various forms and notebooks were supplemented by copies of maps, primarily the 1:5,000 topographical maps, on which field teams marked surveylocations and other observations. Each member of the staff also prepareda staff report at the end of each season, which summarized the activities each person performed, the forms and notebooks in which the records were kept, and whatever other comments the staff desired to make.There were numerous other logistical records,including logs to keep track of the forms assigned for field use, and extensive cross-referencing. We hold redundancy in archaeological records to be a virtue because it makes it possible to discover the inevitable recording errorsthat occasionally creep into databases, however carefully they are kept. All databases and other archives of the Nikopolis Project are stored in the Center for Archaeological Studies at Boston University.

THE

NIKOPOLIS

PROJECT

2I

POST-FIELDWORK ANALYSES
During study seasonsin 1995 and 1996, materialscollectedduringthe and werereexamined studied Ioannina. Byzantine in The surveys Ephoreia former madeavailable studyspace secularized for the mosque, FetiyeDzami, of locatedon the highestpartof the fortress Ali Pashaandadjacent the to new Museumof Byzantineand Post-Byzantine The Archaeology. glorious view fromone side of the mosqueincludedthe lake of Ioanninaand the PindosMountains,andthereweretreesnearby that offeredshadefor staff memberswho might be workingoutside.The staff is particularly for sucha splendidplaceto gratefulto the ByzantineEphoreia providing and ClassicalEphoreiafor permittingthe study,and to the Prehistoric acrosstown from the Archaeological surveymaterialto be transported Museumto the Kastroduringtwo summers. During each of the two studyseasons,the seniorstaff also had the to areas teams precious opportunity revisit survey unaccompanied survey by to direct,and not burdened with surveysto conductor detailedformsto fill out. The staff,then, were ableto contemplate the spot the obseron vations of previousyears, and had the leisure to discuss observations andinterpretations eachotherin the midstof the landscape were with we studying.

PRESENTATION OF RESULTS
Preliminary reports of the Nikopolis Project appeared regularlyin Greek in the Archaiologikon Deltion29and in English in Contextand the Nikopolis Newsletter,publications of Boston University's Center for Archaeological Studies.30Papers by several members of the staff have appearedin full or in abstract form in the published transactions of the several conferences and symposia at which they were presented,31and a few special reports have been published in journals and edited volumes of essays.32 addiIn tion to the doctoral dissertationsof Moore and Tartaron,which were based mainly on project results and have been cited above, a dissertation by Dimitra Papagianni also includes researchon material from the Nikopolis Project.33Chapter 5 in this volume, written by Mark Besonen, George (Rip) Rapp, and ZhichunJing, is based in part on Besonen's M.S. thesis.34 The present book is the first of two volumes of final reports. Chapter 1, by Wiseman and Zachos, provides a history of the Nikopolis Project, and discussions of the research aims, the interdisciplinary methodologies employed, the databases,and the organization of staff and responsibilities. In the second chapterTartaronpresents in detail the methodology of the diachronic surface survey and places both the methodology and the aims within the historical and theoretical context of survey archaeology,especially as that field has evolved in the archaeology of Europe. These two chapters,which constitute an introduction to the work of the project, provide a historical, theoretical, and methodological frameworkwithin which the results of the overall interdisciplinaryproject may be understood and evaluated. They are not intended to be summaries of the results them-

29. Wiseman, and Zachos,

Kephallonitou1996, 1997, 1998. 30. Wiseman 1991,1992a, 1992b, 1993a, 1993b, 1994, 1995a, 1995b, 1997a. 31. Rapp andJing 1994; Runnels 1994; Stein and Cullen 1994;Tartaron 1994;Tartaronand Zachos 1999; Wiseman 1997a, 1997b;Wiseman and Douzougli-Zachos 1994;Wiseman, Robinson, and Stein 1999; reportsby severalstaff membersrecentlyappeared in Isager2001. Articles and abstractsin presshave been omitted here. 32. Runnels and van Andel 1993b; Tartaronand Runnels 1992;Tartaron, Runnels, and Karimali1999. 33. Papagianni2000, which is based on her (1999) dissertationat the Universityof Cambridge. 34. Besonen 1997.

22

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

selves, which arepresented in the reports that follow in this volume and its forthcoming companion volume. In Chapter 3 Runnels and van Andel present the results of the Palaeolithic survey,which they conducted as a supplement to the diachronic survey. Their methodology, developed over some fifteen years of survey in southern and central Greece, was based first on the investigation of the paleoenvironment, especially the geological history of Pleistocene sediments and other landforms.Their report thus deals comprehensivelywith the geomorphology and changes in the environment of southern Epirus in early prehistoric times, as well as the cultural evolution of its human inhabitants, from the Lower Palaeolithic to the Mesolithic. One of the of most remarkable the open-air Palaeolithicsites investigatedby the project is Spilaion, an Early Upper Palaeolithic site near the currentmouth of the Acheron River, where the ground surface was littered with an estimated 150,000 lithic artifacts. Runnels, Evangelia Karimali, and Brenda Cullen report in Chapter 4 on their study of the Spilaion assemblage, including the results of a spatial analysis of the distribution of the artifacts. Chapters 5 and 6 carrythe discussion of the geomorphology of southern Epirus and its relationships to archaeologicalsites from the end of the Pleistocene to the present. Both reports are based on extensive geologic coring programs and intensive laboratory analyses of the cores, as well as other geomorphologic investigations in the field. ZhichunJing and George (Rip) Rapp document the changes over the past 10,000 years in the coastal landscape of the Nikopolis peninsula and the area to its east, which comprises most of the north coast of the Ambracian Gulf. The locations of the important Classical, Roman, and medieval town sites in this region, and of human habitation generally, are related to the dramatic changes in the landscape,which are themselves shown to result from a variety of environmental, geomorphologic, and cultural factors. Besonen, Rapp, and Jing report in detail on the post-Pleistocene geologic history of the lower Acheron valley,tracing the changing course of the Acheron River,the creation and demise of the Acherousian lake, and the gradual change over time of the deep embayment known to Strabo as the Glykys Limen, where large fleets of ships found anchorage both in Greek and Roman times, to the small bay of the present day at the mouth of the Acheron River.The historical implications of the coastal changes are also discussed. In a final chapter the editors comment briefly on the results reportedin this volume. Volume 2 of Landscape in will inArchaeology SouthernEpirus, Greece clude a catalogue of sites/scatters and all tracts surveyed; reports on the pottery,lithics, and other artifacts;and a chronological presentation of the cultural remains in their environmental contexts.

CHAPTER

THE

ARCHAEOLOGICAL
STRATEGIES

SURVEY:
AND

SAMPLING

FIELD

METHODS

F by Thomas Tartaron

1. Keller Rupp1983;Barker and

1991; Cherry 1983, 1994. 2. Alcock 1993; Cherry 1994;

and Alcock, Cherry, Davis1994;

Kardulias1994a; Bintliff 1997. 3. Cherry 1994, pp. 92-95. 4. Binford 1964. 5. Fish and Kowalewski1990; Trigger1989, p. 311. 6. Fish and Kowalewski1990; but see Alcock, Cherry,and Davis 1994, pp.137-138. 7. Kintigh 1990; Plog 1990. 8. Parsons1990; Sumner 1990. 9. Fish and Kowalewski1990.

Systematic surface survey has been practiced and refined in the Mediterranean region for more than a quarter century,1and there is no longer serious controversy about the legitimacy of survey as a robust methodological tool for regionally focused research, or about the contribution it has made to the study of all periods of the Mediterranean past.2 Although the many achievementsof surveyprojectsareself-evident and arouse much optimism,3 few would suggest that a state of disciplinary maturity has been attained. The developmental years have witnessed continuous and serious challenges to many of the theoretical and methodological foundations upon which surface survey rests, as archaeologists have increasingly recognized the complexity of the surface archaeological record, and the inadequacy of many of our methods and conceptual frameworks for analysis and interpretation. Vigorous debate continues on a range of theoretical and practical matters. Recently,the validity of probabilisticsampling schemes and quantitative methods, once regardedas powerful means of characterizingentire regions from carefullychosen samples,4has been called into question. Exoften fail to capture the true perimental data suggest that such "samples" variabilitypresent in the archaeologicalrecord,making suspect the notion that patterns discerned for a portion of a region are necessarily valid for the whole.5 Fish and Kowalewski are particularlyvocal in advocating "total"regional coverage to offset the problem of sampling,6but this approach fails to solve-and in some cases to address-a range of problems, which are well documented by Kintigh and Plog.7 Among these is the fact that surveys ignore the powerful effect of survey many of these "full-coverage" intensity; thus, one project that employed a 30-m spacing interval between walkers,and another in which intensive and systematic coveragearetreated as secondary concerns, hardly point the way forward to revealing the fullness of human activity upon a landscape.8In more practical terms, while the general principle of covering a region, however narrowly or broadly defined, in its entirety would seem unimpeachable, the immense increase in costs entailed in such coverage must be justified by suitably enhanced results. In view of the cases presented by Fish and Kowalewski,9we must at present conclude that sometimes they are, and sometimes they are not.

24

THOMAS

F. TARTARON

At minimum, the critical parametersof intensity and systematic data collection must be integral, not independent,'? variablesin full-coverage survey design. Most Mediterranean surveys, while acknowledging potential problems with sampling, have relied on some type of stratification of the survey universe, typically incorporating samples of a full range of environmental zones with survey locations derived from known distributions of archaeological remains.T" Perhaps more disturbing is the fact that whereas archaeologists ask ever more expansive and complex questions of the archaeological record, the development of increasinglyrefined methods capable of providing the answers has failed to keep pace. Archaeologists have not been able to resolve a range of difficulties that stem, on the one hand, from the inherent complexity of the surface record and, on the other, from an inability of existing methods to record the scatters in a way that faithfully represents their distribution, density, and degree of clustering. The issues are both observational and analytical in scope. The intrinsic complexity of the surfacearchaeologicalrecordhas been measured in a number of recent studies. It is well understood that surface scatters of artifacts at a given location are constantly modified by diverse naturaland culturalagents over time, as replicationand experimentalstudies Ammerman's work in particularreminds us have clearly demonstrated.12 that the local circumstancesand timing of an inspection strongly influence the results, and that repeatedvisits over a period of years may be necessary to capture the fullness of the archaeological record. (This became abundantly apparent to us at locations such as Grammeno and Ormos Vathy; see below.) Yet the precise effects of erosion and deposition, water action, plowing, and other processes on forming and transforming the surface record are not alwayswell understood. Simulation studies have provided a number of promising approaches,'3but they seem not to have been widely applied, in part because it is difficult to control a broad range of variables in nonexperimental situations, and because for each survey,unique conditions pertain. Some relationship between surface scatters and subsurfaceremains is usually assumed, but rarelydemonstrated.14 In recent years, however, survey archaeologists have developed a battery of techniques designed to measure the relationship between surface scatters and the subsurface remains with which they are presumed to be associated. These techniques not only evaluate our measurements of this relationship but improve upon them. Thus, the application of long-term replication studies,'5 geophysical remote sensing,16 phosphate studies,'7 and controlled collections followed by limited, targeted excavation'8all contribute positively to the measurementof subsurface phenomena from surfaceor plowzone scatters,either by identifying subsurfaceremains directly or by isolating the variables affecting the surface/subsurfacerelationship.i9The most promising results have emerged when these techniques are practiced in combination. At one Fort Ancient site in southwestern Ohio, the patterning of surfacematerial was found to supply information that was lacking or ambiguous from excavation, with the result that an anomalously early village of circularplan The Laconia Survey applied controlled collection, phoswas recognized.20

10. Fish and Kowalewski1990, p. 2. 11. Alcock, Cherry,and Davis 1994, p. 138. 12. E.g., Ammerman 1981,1985, 1993; Shott 1995. 13. Odell and Cowan 1987; Shott 1995; Dunnell and Simek 1995. 14. Dunnell and Simek 1995, pp. 306-307; Downum and Brown 1998, p. 111. 15. E.g., Ammerman 1981, 1985. 16. E.g., Weymouth and Huggins 1985; SarrisandJones 2000. 17. E.g., Cavanagh, Jones, and Sarris1996. 18. E.g., Shott 1995. 19. See Dunnell and Simek 1995; Odell and Cowan 1987; Shott 1995. 20. Hawkins 1998.

THE

ARCHAEOLOGICAL

SURVEY

25

21. Cavanagh, Jones,andSarris

1996. 22. Downum and Brown 1998. 23. Downum and Brown 1998, pp.119-120. 24. Wandsniderand Camilli 1992. 25. Bintliff and Snodgrass1988a, p. 506. 26. Alcock, Cherry,and Davis 1994, p. 141. 27. E.g., Odell and Cowan 1987; Stoddartand Whitehead 1991. 28. See especiallythe contributions to Sullivan1998.

phate analysis, and geophysical methods to a number of small ruralsites in southern Greece.21A notable finding of this work was that the extent of habitation sites tends to be largerthan the scatterof surfaceartifactswould suggest. In a large culturalresourcemanagement (CRM) project in southern Arizona, certain artifact types were found to be more reliable predictors of subsurface remains than others.22This study also found that in cases where post-depositional disturbances are great, subsurface remains may have largely or completely vanished, making the surface assemblage the sole remaining source of information.23 Wandsnider and Camilli concentrated on the interface of survey design, survey performance,and the physical properties of the archaeological record, asking in effect what we are measuring with our survey methods, and how this impacts the investigator's aim to faithfully document the Specifically, they sought to measure the disparity archaeological record.24 between the archaeological record (the total population of artifacts that is availableto be found), and the document (the actual population of artifacts discovered). In a series of controlled collections, they measured the effects that intensity and interval of transects,ground visibility,and the size, color, and shape of artifactshave on the document that is produced.Their results indicate that discovery is biased toward obtrusive and highly clustered artifacts;low-density scatters are acutely underrepresentedbecause the typical CRM surveyin the United States is not designed to detect them. Thus, the apparent clustering of surface material may be more an artifact of the measurement technique than an inherent property of the record itself. In Greece, however, where there exists a long tradition of intensive, nonsite surveys, the data reveal striking regional variability in artifact density and clustering. While the Boeotia survey reports an "almostunbroken carpet" other intensive surveyshave recordeda lowerof off-site pottery scatters,25 density, more discontinuous pattern in which artifacts tend to occur in discrete clusters with little intervening scatter.26 The success of the discipline in finding viable solutions to these challenges holds obvious implications for the validity of inferences from the surface record.The present state of progress toward that goal depends in part on one's perspective (is our cup half empty or half full?), but it is also important to recognize that all field situations are not equally amenable to the kinds of innovative approaches that appear with ever-greater frequency in the literature.Thus, in the event of poor preservation of surface materials, less-than-ideal conditions of site visibility, or restrictive permit regulations that preclude excavation and other complementary operations, rather pessimistic assessments of the utility of surface data can be expected.27 a guarded but growing optimism is apparentin reYet cent years-based on an increasing archive of successful applications in a wide range of settings-that carefulresearchdesign and field methods can unlock the intrinsic interpretive potential of the surface archaeological record.28 In conclusion, rather than constituting a cause for alarm,the discomfiture over the limitations and uncertainties surrounding surface survey reflects a phase of critical self-awareness in survey archaeology,and a willingness to tackle the problems head-on rather than simply bemoaning

26

THOMAS

F. TARTARON

them.29The Nikopolis Project, mindful of a host of problems and potential solutions, sought to introduce certain refinements, in part responding to some of the issues raised above, and in part designed specifically for the unique conditions of the Epirote landscape. Though it is not the aim of this chapter to examine in detail the contingencies of archaeological survey, an attempt has been made to come to grips with many of themsometimes successfully and sometimes not. Instead, its purpose is to explain in specific terms the principles on which the survey was designed, and the means by which data were collected.

THE NIKOPOLIS PROJECT AND REGIONAL STUDIES IN GREECE

There are compelling reasons that the sampling strategy and methods of data collection be describedin explicit detail for each surveyproject.While it is certainly true that the range of research strategies and field methods must be flexible enough to respond to widely varying conditions of local topography,vegetation, and access, as well as past archaeological investigation and currentresearchgoals, it is nonetheless imperativethat a framework be provided by which survey results can be evaluated on their own merits, and compared to those of other surveys. The emergence and proliferation of systematic, intensive survey techniques in Greece provide the potential for such a frameworkby introducing methods, using well-defined, quantifiable parameters,which form a basis for comparison of data among projects.30 While acknowledging the complexities of establishing objective criteriaby which data can be evaluated and compared (and achieving that objectivity in one's fieldwork),31 it cannot be doubted that comparability of information, more than a commendable ideal, is in fact a matter of great urgency. Surface survey comprises an ever-increasing proportion of archaeological researchin Greece for severalreasons, among which are the moderate cost and logistical complexity of surveys relative to researchexcavations;the perception that surand the growing interest in landscape vey is a less destructive technique;32 archaeology and regional dynamics, which are best investigated using survey methods. As the pace of survey research quickens and wide tracts of the Greek countryside are explored, some thought should be directed to the legacy of information that is to be left to future generations as the combination of surface survey activity and modern development diminishes the country'sarchaeologicalresources.33 The most ominous prospect of ending up with a patchwork of projects whose data are not comparable is that we shall never learn much about interregional,diachronic trendsprecisely the sorts of issues about which regional archaeological survey ought to be informative. The detailed publication of the major theoretical and methodological components of a systematic survey design-research goals, sampling schemes, and data collection methods-plays a key role in constructing a basis for evaluation and comparison. The considerable attention devoted

29. Cherry 1994,p. 105.

30. E.g., Cherry 1982; Bintliff and Snodgrass1985; Wright et al. 1990; Cherry,Davis, and Mantzourani1991; Jameson,Runnels,and van Andel 1994; Wells and Runnels 1996; Davis et al. 1997; Mee and Forbes1997. 31. Kellerand Rupp 1983, pp. 4344; Bradley,Durden, and Spencer 1994. 32. Surfacesurveycannot be considereda nondestructive technique, however.Under certaincircumstances, as in the case of a field that is plowed artifactson the surfacemay frequently, be replenished,redistributed, or fragmentedacrossthe surface.But in many cases,the tracesof human activity discoveredon the surfacehave no subsurfacesources,or corresponding the mechanismsfor bringing additional materialto the surfacein the short term arelacking.In these instances,artifact collection may have the effect of permanentlyremovingevidence. Another concernis the confounding effect of the piling up or scatteringof artifactsleft behind by archaeologists and others making surfacecollections. At the very least, this action adds a post-depositionalstratumfor which allowancewill have to be made in all futureresearch.On these issues, see Lloyd and Barker1981, p. 390; Ammerman 1981, 1985, 1993; Cherry 1983, pp. 397-400 (discussion). 33. Runnels 1981.

THE

ARCHAEOLOGICAL

SURVEY

27

34. E.g., Bintliff and Snodgrass 1985, 1988a;Wright et al. 1990; Cherryet al. 1991; Wells, Runnels, and Zangger 1990; Wells and Runnels 1996, pp. 15-22; Davis et al. 1997. 35. Cherry,Davis, and Mantzourani 1991, p. 53. 36. Jameson,Runnels,and van Andel 1994. 37. Cherryet al. 1988;Wright et al. 1990. 38. Bintliff 1985; Bintliff and Snodgrass1985. 39. Wells and Runnels 1996. 40. See Chapter 1;Wiseman, Zachos, and Kephallonitou1996, 1997, 1998. 41. E.g., Gaffney andTingle 1985; Bintliff 1985. 42. Wandsniderand Camilli 1992, p. 183.

to laying out field methods (and their essential linkages to research aims, analysis, and interpretation), particularlynotable in the reports of recent cannot help but encourage a replicative or selfintensive surveyprojects,34 effect. Methods that work well in the field and are of sound perpetuating theoretical basis will be recognized, imitated, and refined, with the result that an evolution toward methods yielding statistically valid data amenable to comparison with other regions is set into motion. Recent experience has shown that it is both desirable and possible to devise such field methods, even though the exact replication of methods from one survey to another is rarelypractical and often undesirable.35 The Nikopolis Project surface survey,accomplished in three field seasons from 1992 to 1994, was, like any other, a particular response to a unique set of research interests, environmental conditions, and logistical limitations. Methods and innovations that were developed in earlier surveys in Greece and elsewhere were nonetheless incorporated, and adapted for use in the context of southern Epirus. Particularly influential were those employed in systematic, intensive surveys of recent the years:the Argolid Exploration Project,36 Nemea ValleyArchaeological and the Project,37 Cambridge/BradfordBoeotian Expedition,38 the BerbatiLimnes Archaeological Survey.39 positioning our survey methodology By squarely in this (now well-established) tradition, we acknowledged the validity of the tradition, and sought to produce data that will be, as much as possible, directly comparableto those recordedin other regions of Greece. Yet in putting together the methodological package described below, we made a conscious effort to addresscritically some of the shortcomings we perceivedin previoussurveypractice,and to develop methods that would work well on the Epirote landscape, though perhaps not elsewhere. A first area of concern was to create a program of geomorphological investigation that was more closely integrated with the intensive survey than was typical at the time.40Because southern Epirus contains a high percentage of erosional landscapes, it was essential to establish control on the movement of soils and sediments so that we did not misunderstand the depositional contexts of cultural material we encountered. Coarse-scale geomorphological mapping of the survey areawas supplemented by fine-scale analysis of the contexts of many sites and other locations of interest. In cases where built features were known or suspected, geophysical survey often augmented the results of surface collections. Coastal geomorphology was studied in the lower Acheron valley and on the northern shore of the Ambracian Gulf to measure the change in coastlines over time. We were also convinced that not enough emphasis had been placed on the resolution of quantitative density data from off-site locations, although the benefits of high-resolution data collection were not unknown.41 We agreed with Wandsnider and Camilli's recommendation to decrease survey transect intervals and overall survey pace as a way to ensure that both the low- and high-density surfacerecords are documented.42 purTo sue this objective,we designed a method of close-interval surveywith high-

28

THOMAS

F. TARTARON

resolution data recording that could be used, with simple modifications, for discovery-phase reconnaissance,investigation of rural sites, and urban survey.

THE PURPOSE AND PLACE OF INTENSIVE SURVEY IN THE NIKOPOLIS PROJECT

In many ways, the intensive survey of the Nikopolis Project was different in scale and purpose from the earliersurveys in which it found inspiration. First, the culturallandscape of southern Epirus was investigated by a number of means, of which intensive survey was but one. Other methods by which activity areas were discovered and investigated included 1) extensive survey, comprising systematic but nonintensive "walkovers" (see below), scouting, and the independent Palaeolithic survey;2) a wide-ranging season of ground-truthingof satelliteimageryin 1991, duringwhich known sites were visited and unknown sites were noted (but not investigated); 3) geomorphological studies, in which naturalprocesses affecting sites and landscapes were examined, and unknown sites sometimes found; 4) aerial balloon photography;5) geophysical survey;and 6) documentaryresearch.43 A consequence of this full, multidisciplinaryprogramwas that crew members were shared among the teams listed above, and additionally assigned to laboratoryand data input tasks. Furthermore,the Nikopolis Projectwas conceived and operated as a field school intended primarily for undergraduate students, requiring the senior staff and graduate assistants to engage in many hours of instruction, and requiring students to spend substantial time learning other components of the overall project. Whereas no apology is offered for our researchdesign or educational model, in practice these circumstancesplaced limitations on availableperson-hours, and most notably on the total territory covered by the intensive survey,which was also constrained by the deliberately intensive methods we employed (see below). The principal purposes of the intensive surveywere to test its feasibility in the Epirote landscape; to reveal the overall characteristics of the region's archaeological resources;and to examine rigorously certain locations of particular prehistoric or historical interest, such as the lower Acheron valley and the Ayios Thomas peninsula (Fig. 2.1). It was not known whether the southern Epirote countryside would be as well suited to intensive surface survey as had been the southern Greek mainland and islands, where most such surveys have taken place. In fact, the climate and topography did present unusual challenges.The southern Epirote climate, transitional between Mediterranean and temperate, is characterizedby a much higher annualrainfallthan that of southernGreece and, consequently, the vegetation is more lush. The terrainis, on the whole, more rugged and mountainous, although such topography is by no means lacking in the south. Furthermore, the land is less developed agriculturally,so there is less open terrain.A significant consequence of these conditions was that surveyingin large,contiguous blocks of tractsbecame difficult, and at times impossible.

43. See note 40.

THE

ARCHAEOLOGICAL

SURVEY

29

,7-vL

,..,,k,

Louros River

Acheron River

Parga
:...,/,-,,

astEphyra(Kastri

PhanariBay:.: , (Ammoudia).Lower Acheron Valley

*- t.

*':Grammeno

Arachthos Rive er

Ionian Sea
Nikopolis

a:

*.d,.

Ayios Thom 'a

'"

.?. OrmosVathy

..-. .: ."...-Peninsula :..

<K!..\
.

'_.Ambracian

Gulf

Actium

Figure2.1. Map of southwestern Epirus,showinglocationsof places mentionedin the text

10

15

20

25 5KM

SAMPLING

STRATEGIES

44. Hammond 1967; Dakaris 1971, 1972; Dakaris,Higgs, and Hey 1964; Higgs and Vita-Finzi 1966; Higgs et al. 1967. 45. Redman 1973.

The sampling strategies adopted for the Nikopolis Project survey were shaped by certain environmental and operational constraints specific to the region and to the project. The most salient of these constraints were the enormity of the study area;a lack of previous systematic exploration; characteristicsof the local terrainand vegetation cover;limits on time and availablemanpower; and specific researchinterests. The Nikopolis Project study area comprised some 1200 km2, corresponding roughly to the nomos of Preveza.This large expanse was chosen with a broad research agenda in mind, and not to be manageable for an intensive survey (see Chapters 1 and 7). While portions of the nomos (notably the environs of the ancient city of Nikopolis) have long attracted the attention of archaeologists and other scholars, no systematic exploration of the countryside had ever been undertaken.A scattering of sites was known as a result of chance finds by locals and the explorations of individuals and small research teams.44Most of the countryside in this relatively large nomos was, however, archaeologicalterraincognita in 1991. It was thus a daunting challenge to devise a scheme by which a meaningful sample could be taken in the space of a few field seasons. Our initial response was to develop, prior to the actual fieldwork, a multistage sampling strategyin which each phase of field researchinformed the direction and focus of subsequent phases.45 The entire study areawas first stratified according to general environmental and cultural criteria.In

30

THOMAS

F. TARTARON

keepingwith the broadaim of the Nikopolis Project-to explain"thechanging relationships, from prehistoric through mediaeval times, between humans and the land and resourcesthey exploited"46-the first stratum consisted of preliminary estimates of environmental parameters,using topographic and geological information. In practice, this process involved the classification of the surface of the study area by features of topography or terrain (e.g., floodplains, low foothills, high elevations, swamplands, etc.), as hypothetical correlates of distinct environmental zones. It was noted that these categories also encompassed considerable variation in soil type and modern land use. The theoretical underpinning of such a strategythat the full range of human activity in, and exploitation of, the environment can only be captured by taking a meaningful sample of a variety of environmental zones and landscape settings-is widely recognized and The classifications were generapplied in modern survey archaeology.47 ated from topographic and geological maps and in the course of groundtruthing satellite imagery in the summer of 1991, and served as general guidelines to the range of environmental zones availablefor study. A second stratum comprised the many specific researchobjectives related to cultural and historical phenomena (see pp. 8-9). One example is the basic inquiry concerning the relationship over time of the city of Nikopolis to the surroundingcountryside. In order to test the full range of core-periphery interactions, it was deemed essential to sample locations both close to and at a distance from the limits of the urban area.For lack of information about the city's hinterland, the locations of known sites and the results of previous researchwere incorporated into the sampling design, which also included survey tracts chosen at random. The integration of environmental variables, prior research, and specific researchobjectives in the sampling design created a mosaic of potential survey areas.Locations identified by the different methods frequently coincided in terrain suitable for tractwalking, providing a starting point for placing survey units. A season of nonarchaeological reconnaissance in 1991, aimed at studying environmental parametersand modern land use, permitted an initial familiarity with the landscape and the varied terrains that survey teams would encounter. During the first season of archaeological survey (1992), diverse environmental zones and landscapes were investigated, field methods were adapted and refined, and promising locations requiringfurther attention were identified. On the basis of this experience, detailed sampling plans were drawn up for intensive work in 1993 and 1994 in specific areasof interest, e.g., the lower valley of the Acheron River (Table 2.1)48and the Ayios Thomas peninsula. Once the surveywork began for a given season, information gathered on scouting trips and by other components of the projectallowed us to add suitablelocations, eliminate those that offered little hope for results (for reasons of geomorphology, access, or ground cover), and set priorities for the fieldwork. Sampling designs were subject to daily modification based on current information regarding vegetation cover and accessibility, geomorphology, new discoveries, and a host of other factors. Thus, within each environmental stratum, some survey units were placed using judgmental criteria, while others were positioned where terrainpermitted, without prejudiceof prior knowledge.

46. Wiseman 1991, p. 1. 47. Schiffer,Sullivan,and Klinger 1978, p. 12;Wright et al. 1990, p. 604; Barker1991, p. 3. 48. Tartaron1996, pp. 384-390.

THE

ARCHAEOLOGICAL

SURVEY

31

TABLE 2.1. STRATIFIED SAMPLE AND SYSTEMATIC SURVEY COVERAGE, LOWER ACHERON VALLEY, 1992-1994
TotalArea Class Topographic Floodplain Coast Swamp/Bay Foothills (<100 masl) High hills (>100 masl) Total (sqkm) 53.40 6.85 8.55 17.64 22.56 109.00 ModifiedArea (sqkm)* 0 6.85 0 17.64 22.56 47.05 Systematic Coverage (sqkm) 0.014 0.631 0.035 0.315 0.155 1.150 Percentof ModifiedArea 9.21 1.79 0.69 2.44

*This figurereflectsthe subtractionof areasdeterminedby geomorphologicalanalysisto preserveno premodernmaterials becauseof deep burialby recent alluvium,a finding confirmedby archaeological testing.

49. As explainedin Chapter3, the Palaeolithicsurvey,directedby Curtis Runnels,was a separateentity of the Nikopolis Project,with a sampling design and field methods quite distinct from those of the intensive survey. 50. See Barker1991; Cherry,Davis, and Mantzourani1991; Barkerand Mattingly 1999. Modern landscape archaeologydrawsupon severalolder traditions,includinghuman geography, the "New Archaeology," landscape and studies such as those of the British Royal Commission on Historical Monuments (Keayand Millett 1991, pp. 129-131). A sharpdistinctionis sometimes made between "landscape approach," reflectingthe concernsof with ecological processualarchaeology and geological system variables (Rossignol 1992, pp. 4-5; Rossignol andWandsnider1992), and "landscape often associatedrather archaeology," with the postmodernist's narrowly historicaland contextualfocus (Roberts 1987;Yaminand Metheny 1996; Ashmore and Knapp1999). But the distinction appearsartificial,as both perspectivesmay be profitablyapplied to the study of regions,and in our usage both may be subsumedunder the term "landscape archaeology." 51. Plog, Plog, andWait 1978; Cherry 1983, p. 387.

Close consultation with geologists,geomorphologists, members and of the separate Palaeolithic surfacesurveyoptimizedthe ongoing selection of samplelocations.Geologicalcoringandgeomorphological observationidentifiedareas whicherosionandredeposition in wereso extensive thatthe discovery remains humanactivity primary of of in contextwasnot thoselandscapes wererelegated testingto confirmor refute to anticipated; the geomorphological was findings.Information sharedto greatmutual benefitwith the Palaeolithic whosesmallsize andextensive team,49 survey methodsallowedit to rangeoverlargeportionsof the studyarea.Indeed, and sites manylateprehistoric historical werefirstdiscovered the Palaeoby lithic surveyteam,while important evidencefor Palaeolithic activitywas detectedduringintensivearchaeological survey.

SURVEY INTENSITY

AND COVERAGE

In recentyears,the practiceof surfacesurveyin the Mediterranean has been integratedinto the broader a pursuitof landscapearchaeology, diversearray approaches has cometo embrace of that both tangible(topogartifacts and features)and intangible(socialaction, environment, raphy, of symbolism, perceptions spaceandplace)aspectsof livingin the world.50 concernthemselves with the Surveysadoptingthis perspective naturally in its entiretyand in everysense of the term;thus, all tangible landscape manifestations humanactivity, of frommajorsettlementsto tiny scatters of pottery or lithic debrisin the countryside, form part of the greater and of Intensivefield techniques, fealandscape(s) aredeserving analysis. of the surface fieldwalkers close spacing at turingsystematic coverage by as intervals, emergedparticularly a way to recovermore of the archaeoextensionsof logicalrecord,and haveflourishedas ideal methodological the landscape It has been demonstrated concluarchaeology perspective. in increase the numsivelythatincreasing intensity yieldsa corresponding ber of sites and scattersdiscovered,51 for this reasonalone intensive and methodsare consistentwith the goals of landscape But archaeology. the enhancedpowerof resolution not arrived has withouta concomitant cost in termsof overallsurfacecoverage.Surveyintensityand arealcoverage are inversely and methodological related,a fact that presentstheoretical

32

THOMAS

F. TARTARON

difficulties that must be confronted in the planning of any regional project. The choice of optimal levels of intensity and coverage involves compromise between legitimate desires for a representative sample and survey precision(i.e., the power of resolution).52 This conflict is particularlyacute in the case of unusually large, poorly known regions like southern Epirus. The response formulated by the Nikopolis Project integrated extensive and intensive survey methods. This path was chosen in recognition of the need for basic information, on the one hand, and a desire to initiate an intensive study of the human landscape of southwestern Epirus, on the other. Three modes of survey were applied to the surface archaeological record:extensive nonsystematic, extensive systematic, and intensive (systematic). Extensive nonsystematic survey involved scouting, geomorphological evaluation coupled with archaeologicaltesting, and the work of the Palaeolithic survey team, which employed a judgmental sampling design based on a predictive site model. Extensive systematic mode refers to systematic search, carried out in most respects like intensive survey, except with fieldwalkersarrayedat intervals of greaterthan 15 m. The two modes of extensive survey were used to reveal the overall characteristics of the region: the number of sites, their distribution, chronology, function, and relationship to the environmental context.53Simultaneously,a program of off-site, intensive surveywas carriedout in locations of particularresearch interest and across environmental zones in order to reveal patterns of human activity of every description over the entire landscape.A primaryconcern in designing the intensive surveywas the acquisition of data comparableto those collected elsewherein the Mediterraneanarea.Quite different approacheshave been taken by other projectsfacing similarcircumstances.54 In the neighboring province of Aetolia, a recent survey employed extensive techniques that allowed for the plotting and dating of relatively obtrusive sites over a large area,but systematic collection of information on site densities and off-site phenomena was beyond the scope of the method.55 For that project, extensive methods were consistent with the stated research aim of gaining an overall impression of human settlement in the region. In an intensive survey,total arealcoverage is determined by a series of The principal determivariables,only some of which can be controlled.56 nants are labor input, measured by crew hours expended, and survey intensity, measured by spacing between crew members. Other variables include crew experience and the complexity of field operations. If the study areais relativelysmall, or if a large crew can be assembled and maintained, the pressureto attain adequate surface coverage is eased considerably.But the amount of labor that can be dedicated to surveyis often fixed by financial constraints that cannot be closely controlled, with the result that survey intensity is frequently the critical consideration in decisions concerning areal coverage. In the case of the Nikopolis Project, neither the size of the study area nor the labor contribution could be dictated exclusively by the interests of the survey. Although surface survey was the activity to which the greatest commitment of resources was made, crew members were shared with teams engaged in geology, geomorphology, aerial pho-

52. Wandsniderand Camilli 1992, p. 170. 53. Cherry 1983, p. 393. 54. Rutter 1993, table 1, with references. 55. Bommelje and Doom 1987. 56. Many of these variablesare the same ones regardedas factorsinfluencing site discovery;see, e.g., Schiffer, Sullivan,and Klinger1978, p. 4; Cherry 1983, p. 397; Barker1991, pp. 4-5.

THE

ARCHAEOLOGICAL

SURVEY

33

TABLE 2.2. TYPICAL

VALLEY Tom Tartaron (team leader) Brenda Cullen (assistant)

DAILY WORK ASSIGNMENT,

GEOLOGICAL

JUNE 28, 1994

CORING: ACHERON VALLEY

SURVEY TEAM I: LOWER ACHERON

JenniferMurray Yaz Shimizu Anne Maxson Aviva Figler

SURVEY TEAM 2: KASTRI, LOWER TOWN

ZhichunJing Mark Besonen Stephan Papageorgiou Amy Graves

ARTIFACT PROCESSING AND INVENTORY TEAM

Carol Stein (team leader) Alan Kaiser (assistant)

Lucy Wiseman Betty Banks Mike Gaddis Lisa Davis Leslie Harlacker
LITHICS:
COMPUTER

Alesia Alphin KarlaManternach KaterinaDakari Alison Spear T.J. Reed

TOPOGRAPHIC SURVEY: KASTRI

Lia

Karimali
AND DIGITIZING

DATABASES

Rudi Perkins (studentsfrom artifactprocessingteam)

PHOTOGRAPHY: ARTIFACTS AND KASTRI

Theo Chatzitheodoros Joseph Nigro

PALAEOLITHIC SURVEY TEAM

Michael Hamilton
GEOPHYSICAL PROSPECTION: KASTRI

Curtis Runnels PriscillaMurray Tjeerd van Andel

Apostolos Sarris Dimitra Papagianni Mely Do KathyMontgomery

57. This figure has been arrivedat in Wandsnider experimentalsituations: and Camilli 1992; Bintliff and Snodgrass(1988b, p. 58) use a figure of 2.5 m for the Boeotia survey. 58. Whether such scattersare actuallydiscovereddepends on a host of variablesthat relateto intrinsic record propertiesof the archaeological itself, surveystrategies,and human factors:Wandsniderand Camilli 1992.

tography, artifact and data processing, and other operations (Table 2.2). The total labor pool from season to season depended largely on the number of undergraduate students in the Nikopolis Project field school (1992: 12; 1993: 24; 1994: 16), along with a few graduate students each year from several different institutions. Ultimately, therefore, the balance of intensity and coverage could best be influenced by field procedures,and most directly by the spacing interval chosen and the complexity of field methods. Variabilityin spatial coverageis a systematic and implicit consequence of the spacing interval that must be addressed in the framework of objectives for data collection and analysis. It is expected that each surveyorwill have an effectivevisualrange of 1-2 m on either side of the surveytransect.57 At 15-m spacing, therefore, a swath of 11-13 m is unexamined, with the result that scatters or other findspots with axes of 10 m or less perpendicular to the survey transect may not be discovered;at this spacing interval, an average ground coverage of 17% (assuming teams of 4-6 surveyors) is achieved. An 8-m interval may preclude the loss of any manifestation with a perpendicularaxis of 5 m or more, and yields an averageground coverage of 32%.58 view of the availabilityof extensive reconnaissanceas a compleIn ment to systematic, intensive survey,it was decided that a spacing interval of 5 m would be adequate to provide precise and statistically quantifiable data concerning the location, density (both on- and off-site), chronology, and range of human activity within the study area.A suite of investigative

34

THOMAS

F. TARTARON

procedureswas then developed to measure these target variables.In three seasons, averaging 25 field days and 15 crew members, approximately 2 km2were investigated intensively, and an additional 3 km2received extensive systematic coverage.The area covered by extensive nonsystematic investigation is estimated to be just over 100 km2.

FIELD METHODS
The surface archaeological record of southern Epirus preserves evidence of a wide range of human behavior,from minute, single-event loci to large, fortified cities and towns. To facilitate the effective treatment of such diverse phenomena, procedureswere requiredthat would permit surveyteams to observe and record information at very different scales of complexity.59 Three separate arealunits of surveywere developed, each reflecting a different purpose and a different level of intensity: the tract, the site/scatter, and the walkover.A unique field form was created for each to ensure the full recording of data, and to be compatible with the project's computer database.
TRACTS

The tract was the basic unit of investigation for off-site, intensive survey, in which landscapes of unknown, but presumably relatively high, site potential were explored.The tract was defined as a parcel of land of varying size, the parameters of which were determined by physical boundaries, such as field borders,fences, or roads;naturalfeatures,such as topographic contours; or by some maximum size guideline.60The terrain of the tract normally exhibited relative uniformity of vegetation, visibility, and modern land use.61In open terrain, tracts were normally of rectilinear shape, but features of terrain and topography often imposed unusual outlines, as for example a doughnut-shaped tract wrapping around the lower slopes of a hill. Tracts were walked by small teams, typically of five or six members. Experience has shown that teams of smaller or larger size tend to be less efficient in gathering and reporting data. Each team included a team leader, a graduate student assistant, and three to five undergraduate field school students. The level of experience among the students varied widely; some had participatedin severalarchaeological projects,while others had no experience at all. Team leaders were charged with overseeing the survey process and gathering critical environmental and archaeological information, and therefore rarely walked survey lines themselves. For purposes of record-keeping, each tract was designated with a T for tract, followed by an accession number,sequential by the year in which the tract was walked (e.g., T93-20 refers to the twentieth tract of the 1993 season). A tract was begun by lining up the crew at the desired spacing interval-the standard interval was 5 m, although owing to diversity of local conditions, intervals as small as 3 m and as large as 8 m were observedand proceeding through the tract in paralleltransects.Team members were

trial 59. Considerable anderror wererequired before fieldmethods weresolidified. procedures The described thissection in wereessentiallyin placeby the middleof the second season: Wiseman1993a,1993b;
Wiseman, Zachos, and Kephallonitou 1998. 60. Tractsize guidelinesfor recent intensive surveysinclude the following: Boeotia Survey, Cambridge/Bradford 0.6-0.9 ha (Bintliff 1985, p. 201); Nemea ValleyArchaeologicalProject, no more than 1 or 2 ha (Wright et al. 1990, p. 604); Berbati-Limnes ArchaeologicalSurvey,average3.75 ha (Wells, Runnels,and Zangger 1990, p. 214). In the surveyof northernKea, tractsize was determinedby field walls terraces enclosing groupsof agricultural (Cherryet al. 1991, p. 22). A size limit of 2 ha was observedin our survey,but the averageunit size was ca. 0.50 ha. 61. It is recognized,however,that tractsdefined by modern featuresmay comprisemultipleunits of geomorphological deposition.The definition of surveytractsas discreteunits of deposition is an innovationof the EasternKorinthiaArchaeological Survey;see descriptionat http:// eleftheria.stcloudstate.edu/eks/ methodol.htm.

THE

ARCHAEOLOGICAL

SURVEY

35

aboutartifacts entrainedto call out to each other simple observations or counteredin their transects; calls such as "Sherd!," "Tile!," "I have a of were typical.This information concentration lithic debrisover here!" to alertedthe teamleaderandothercrewmembers the densityanddistriand fellow surveyors maintaina high bution of artifacts, encouraged to level of concentration. Each surveyor a carried tallycounteror "clicker" which to record on of the quantity artifacts, once for eachartifact noted.Whereasall clicking artifactswere counted,only those considereddiagnosticwere collected from the surface. this Generally, meantall potteryfragments exceptundecorated sherds(thoughfabricsamplesof the latterwerecollected body if not represented amongthe diagnostic pieces),a samplingof bricksand Items were retainedin plasticsamplebags tiles, and all other artifacts. until surveyof the tractwas completed; subsetof the objectswas then a contextual collection. takenfor inclusionin the project's To provide of in closecontroloverthe spatial distribution artifacts the for of artifact counts wererecorded eachcategory item(pottery sherds, tract, studentassisbricks, tiles,lithics,other)by the teamleaderor the graduate tantat 30-m intervals. This simpleexpedient yieldedgridsshowingvariaa tionsin artifact type,providing wealthof comparative densityby artifact for The taskof keepingseparate countsof information off-site locations. was differentclassesof artifact complicated the fact that tallycounters by displaya single total, and was especiallydifficultwhen a wide rangeof material represented the tract.A few simpleinnovations was in were dememory veloped to preservethe integrityof the data,while decreasing of of demandson the crew.In the overwhelming majority tracts,artifacts or flakedstone oba single material, pottery,architectural i.e., ceramics, In jects, dominatedthe surfaceassemblage. those cases,the usualprocedurewas to use the tallycounterfor a total artifact countthroughout the a mentalcount,for each30-m block, tract,while simultaneously keeping of all objectsnot fallinginto the dominantclass. A simple examplewill illustratethe implementation this proceof dure.At the 90-m markof a given tractwith a preponderance pottery of sherdsrelative otherobjects,the teamleaderfirstrequests each surto of count,which is readoff the tallycounter. veyora totalartifact Hypothetical surveyor whose total count at the previouscheckpoint (60 m) was a, or in 30-m segment.Next, 45, reports65 artifacts, 20 artifacts the current the teamleaderasksfor a countby typeof all objectsotherthanpotteryin the current 30-m segmentonly.Surveyor reports5 tile fragments 2 a and lithicflakes.Simplesubtraction reveals 13 potterysherds that werecounted a by surveyor betweenthe 60- and 90-m points.The tallycounteris not resetuntil the tractis completed, thosewalkingsurveytransects and need tractswere enonly retainmentalcountsfor 30 m at a time. Inevitably, in countered which two or even threeartifact typeswere abundant; typithesewerelocations whichhistorical-period in sherds, tiles,andbricks cally, litteredthe surface. those instances, secondmethodwas employed In a in which mentalestimatesof the relativepercentages abundantartifact of To types were maintainedand reportedat 30-m intervals. continuethe a at example,surveyor reports120 total artifacts 90 m, 40 morethan the

36
Rev.June30, 1993

THOMAS

F. TARTARON

The Nikopolis Project

Archaeological Survey Tract Form

Tract# Date I Recorded by I

1/50,000Map I TractSize (meters)

I 1/5000sheet I

I Elevation I

I Visibility(1-10) SpacingInterval l=poor;10=excellent andDirection

I Order Walking

IGPS Reading

AssociatedSite/Scatter #s ARTIFACT COUNTS: Sherds Rooftiles FlakedStone Ground Stone Metal NotebookRefs Inventoried Artifacts

Sample#s Other(specify):

Photographs

Sketchmapof tractlocation. Northshouldalwaysbe up. Includelocationsof adjacent roads, villages,knownsites, contiguous tracts,etc. as reference points.

Figure2.2. Archaeological survey tractform 80 reported at 60 m, and estimates that, over the past 30 m, 60% of the artifactswere pottery sherds and 40% fragments of rooftiles. Approximate values of 24 sherds and 16 rooftile fragments are thus obtained for a 30-m segment of a single survey transect. Obviously, the first method is preferable, and was applied wherever possible. These methods of recording artifact information at close intervals, while time consuming, were simple to implement, and allowed a remarkablyfull accounting of off-site artifact distributions. Additional passes, executed in the same manner as the first pass, were requiredwhen the parcel of land selected for the tract was larger than the team could cover in a single pass. Once the walking of the tract was finished, the team assembled to perform critical documentation activities: the preprinted tract form was filled out, the tract plotted on a 1:5,000 topographic map, and the artifactsprocessed. The information requested on the tract form (Fig. 2.2) summarizes the location and size of the tract, several environmental variables, details

THE ARCHAEOLOGICAL

SURVEY

37

of the surveyprocedures used, and the resultsobtained.Visibilityof the was groundsurface,a criticalvariablein assessingsurveyresults,62 disand from1 (poor)to 10 (excellent), enteredon cussedby the crew,ranked the form.Also at this time,the teamleaderplottedthe tracton the approand map,black-and-white color priate1:5,000 Greekarmytopographic of photographs the tractwere taken,and a locationalreading(in UTM was andlatitude/longitude) takenwith a hand-held globalpositioning system (GPS).The datarecorded the GPS serveas a checkon the manual by information sysplottingof the tract,and areidealfor use in geographic the tems (GIS) applications. Finally, artifact samplesweresortedby type to and a subsetwas selected.It was often not desirable retainall artifacts rooftilefragmentsand collectedfrom the tract;for example,redundant fabricsampleswerefrequently discarded. Once the samplewas duplicate decidedupon,samplebagswith woodentags recording contextual inforfor mationwereprepared each of the variousmaterials represented (potall tery,tile, brick,etc.).The teamleadersupplemented of these activities by includingin the field notebooka multitudeof observations regarding fieldconditions, artifact and geomorphology, survey procedures, patterning and chronology, so forth.
SITE/SCATTERS

62. See discussionin Ammerman 1993, pp. 369-371. 63. Wilkinson 1982; Dunnell and Dancey 1983; Bintliff and Snodgrass 1988a;Barker1991, pp. 5-6. 64. Thomas 1975, pp. 62-63; Dunnell and Dancey 1983, pp. 271274; Wright et al. 1990, p. 603. 65. Gaffney andTingle 1985, p. 68; Cherryet al. 1991, p. 21. 66. Doelle 1977, p. 202. 67. Thomas 1975; Foley 1981. 68. Cherry1984, p. 119. 69. Hope Simpson 1984; Schofield 1991a.

In the course of walking the countryside, survey teams frequently encountered anomalously dense scatters of archaeological material, or isolated but recognizable architecturalfeatures, such as sections of wall or agriculturalinstallations.The recognition, investigation,and classificationof these concentrations reflect a project'stheoretical orientation toward the spatial aspects of human behavior,and the ways in which behavior is preservedin surface deposits. The traditional concept of the "site,"easily recognizable by a dense clustering of artifacts and definable spatial limits, has been found inadequate to encompass the full range of human activity.One consequence of the reevaluation of the site concept was the emergence and development of intensive survey techniques, which forced a rethinking of the spatial implications of human activity as ever smaller and less clustered loci came to light. Interest arose in the meaning of low density or "off-site" and ultimately the individual artifact, rather than the site, was scatters,63 Sites were defined and redefined designated as the basic unit of analysis.64 in relative terms (e.g., as density peaks against a background of artifacts in spread across the landscape);65 absolute terms (e.g., number of artifacts or rejected altogether.67 great achievement of intenA per squaremeter);66 sive survey has been the development of approaches to the study of the kinds of activities that took place largely outside the confines of traditional settlements, among them hunting, pastoralism,agriculture,and tool manufacture.It remains true, however, that identification of off-site scatters on the landscape can be as much an interpretive as an observational and exercise,68 that the inference of specific behaviors from them is usually problematic.69 Ultimately, each survey develops its own approach to the conceptualization, recognition, and treatment of surface concentrations. Follow-

38

THOMAS

F. TARTARON

we record ing Cherryand colleagues, perceivethe surfacearchaeological of a givenlandscape "avariable as distribution residue of frompastcultural in While allowingwide activities, some placesdense,in othersless so."70 berthto evidencefor nonsettlement this does not esactivity, perspective chewthe importance the traditional of settlement. conceptof the site/ The scatter formulated the NikopolisProjectsurvey for (read"siteor scatter") wassimilarly intendedas a nonjudgmental of referring a widerange to way of material and in using it we soughtto avoidrigid,usually phenomena, unworkable definitionsof the "site." our system,the term site/scatter In referred a spatially to definable locusof pasthumanactivitycharacterized materialdistribution, by high artifactdensityrelativeto the background and embodiedalmostany identifiable, concentrated evidenceof cultural as a activity, the followingexamplesdemonstrate: smallscatterof flaked stone;a single,isolatedolive mill;the entireancientcity of Nikopolis;or an exposedstratigraphic profile. out of Picking meaningfulartifactconcentrations of a background scattered material be difficultindeed.It requires experienced can an team leaderand crewattentiveto subtlechangesin the type, chronology, and At density of materialrelativeto the surrounding landscape. times the described abovealertedthe teamleaderto the presrecording procedures ence of potentiallysignificant but patterns, often the decidingfactorwas the abilityof the teamleaderto monitorcontinuously finds and their the contextsof discovery. Fromtime to time, potentiallysignificantconcentrationsbecameapparent only afterdensitydatawere assessedand finds were inspected.In such instances, relevanttractswere earmarked the for revisits. of subsequent Althoughconcentrations veryhigh densityin absolute termswere almostalwaysrecognizedas sites, lower-density scatters wereevaluated relative the surrounding to This evaluation was landscape. done in recognitionthat certainbehaviors producefewer and less clusteredremains(or remains that arenot expectedto survive the archaeoin andthatthe quantity durable of artifacts available presfor logicalrecord), ervation variessignificantly different for periodsof the past.7" of was that Investigation sites and scatters guidedby the observation differentscales and naturesof surfaceconcentrations different require methodsof documentation; is, surveyteamsmust find ways to deal that and with sitesandscatters rangewidelyin size that consistently effectively and complexity. small concentrations, examplelocalized for Extremely of artifacts the bulldozingof a dirtroad,maybe meapatches exposedby and On sured,documented, collectedin theirentirety. the otherextreme, sites, such as the fortifiedcitadelsof southernEpirus, large,multiperiod present very different challengesof scale and complexity;clearly,the of and application total measurement collectionto largesettlementcontextsis not practical, the examination a sampleof the totalsite area and of be necessary. Becausemost site/scatters occupiedpoints on a conmay tinuumbetweenthesetwo extremes, methodsfallingbetweentotalcollection andsamplingportionsof sitesweredesirable the vast majority for of occurrences.

70. Cherryet al. 1988, p. 159. 71. Millett 1991.

THE

ARCHAEOLOGICAL

SURVEY

39
AND COLLECTION

SITE/SCATTER PROCEDURES

DOCUMENTATION

72. See Shott 1995.

Site/scattersof moderatesize (up to 100 m on the shortestaxis) were or documented normally immediately upon discovery, in a matterof days for promptaction are the thereafter. The most compelling justifications and It uncertainty potentialloss of datathat delaymayintroduce. is never for prudentto assumethat a surfacescatterwill be available studyin its presentconditionnextyear,or even nextweek.A commonproblemconcernsthe growthof vegetationandcropsduringthe field season.Artifact on concentrations areobtrusive plowed,newlyplantedfieldsbecome that invisiblein a matterof a few weeks.The changesin vegetationthat occur fromone yearto the next can be even moredramatic. The thorough,intensivesurface out studycarried in 1993 at OrmosVathy(seeFig.2.1), the mainport of the city of Nikopolis,would havebeen impossible probable in the summerof 1994 as a resultof heavywinterrainsthat fosteredthe over Smaller growthof dense,impenetrable vegetation muchof the bayarea. artifactscatterscan be washedawayor otherwiselost at anytime if they areparticularly fragileor ephemeral. Humanfactorsalso arguefor promptdocumentation. Sincedesignation as a site/scatteris often a rathersubjective decision,it is preferable that the personsmakingthe judgmentalso carryout the investigation on the basisof the patternsthat areperceivedat the time. Patternsthat are clearwhen one is focusedon a given landscape becomeblurrywith the difficultif key membersof the dispassageof time;revisitsareespecially of coveryteam are not present.Furthermore, given the constraints time and the uncertainties carryingout archaeological of one research, must neverassumethat a concentration surfaceartifacts be available of will for as obtainedat the studyin the future; a result,the amountof information time of discovery shouldbe maximized. This policyin no waydiminished the importance plannedsite revisits,which were employedto review of and expandthe originaldocumentation. as the Typically, soon as the teamleadersuspected presenceof a site/ was scatter,the tractwas stopped,and preliminary investigation begun. The teamfirstdetermined approximate, notional, the or boundaries the of artifactscatter(bearingin mind that the limits of a surfacescattermay This accomchangewith each new documentproduced).72 was normally walkoutward differentdirections in from plishedby havingall surveyors an estimated"center" the scatter, placea flag in the groundat the of and to point wherethe anomalous densityof material appeared cease.As this task sometimesrequired ratherdifficultjudgments,the team leaderand studentassistant scrutinized workclosely. a generalrule, the As graduate we felt it wasbetterto overestimate dimensions the slightlythanto underestimatethem;quantitative datacollectedas partof the site/scatter may whereasdata that are not point to the need to rein in the boundaries, documented partof the site/scatter as tend to blendwith the off-site mawiderboundaries a latertime. at terial,makingit difficultto establish

40

THOMAS

F. TARTARON

Once the site/scatter was outlined, the tract procedure was modified somewhat. The tract was continued, including that part of it falling within the site/scatter,but within the areaof the site/scatter,artifactswere counted as normal but not collected. The reasons for this modification are as follows. Artifact counts were continued at the normal intervals in order that the density data would be comparableto those of all other tracts.Artifacts were not collected within the confines of the site/scatter during the walking of the tract, however,because if they were, our ability to exercise spatial control over the site/scatter as a separate entity (in some cases, extending beyond the present tract) would be compromised. In essence, then, two entities-separate but intimately associated-were created.The Nikopolis Project recording system ensured that associated tracts and site/scatters remained linked in the project'sdatabase. Upon completion, the tract was processed as usual. If the site/scatter fell completely within the tract, it was then investigated in detail before other tracts were begun. If the site/scatter extended beyond the present tract, additional passes were normallywalked until the site/scatter was circumscribed. For very large sites, on which it was not practical or desirable to envelop the entire dense scatter of material in a single tract, several tracts were placed to encompass the site and the off-site territory surrounding it. After finishing the documentation of the tract, the team turned to the investigation of the site or scatter. Very small scatters were subjected to total measurement and collection, and the largest sites, usually known previously, were deferred for investigation using "urban survey"techniques (see below). For all but the smallest and largest concentrations, a method very much like tractwalking, but more intensive, was applied. Surveyors were arrayedat closer intervals, usually 3 m, and walked survey transects through the site/scatter. Artifact counts were recorded (by type) at 10-m instead of 30-m intervals, and diagnostic artifactswere now collected. In certain experimental cases in which spatial control was crucial, separate samples were taken for each 10-m (or even smaller) block, by team or even by individual surveyor.In all cases, descriptive information about patterns in the artifact distribution was recorded by the team leader in the field notebook. Detailed information was entered on a field form similar to that used for the tract, and the site/scatter was designated in a like manner (e.g., SS93-5 refers to the fifth site or scatter discovered in the 1993 season). The samples taken from the site/scatter were associated with, but also fully distinct from, the tract(s) from which they issued.
SPATIAL SCATTERS RELATIONSHIPS OF TRACTS, SITES, AND

For several reasons, the spatial relationships between tracts and site/scatters may be simple or they may be quite complex: tract limits were arbitrarily chosen, usually on the basis of modern features; the potential for sites in a given tract is usually unknown; and evidence for many forms and levels of concentrated human activity may be present in a small area.Hypothetical example A (Fig. 2.3) is the simplest relationship: a single site/ scatterappearscompletelywithin the confines of a tract.Moderately greater

::.

THE

ARCHAEOLOGICAL

SURVEY

4I

A
/ SS93-2 S .SS 93-3

Site / Scatter
. .

SS 93-7
? ..

..

Tract1
.. . .

.
.

SS93-17

B
ss93-8.

.3 :., .-;:.-^.-.;. SS^ ^

. : ::

:
...
. ,.. . : :0 :

Site/ Scatter Tract1 Tract2

Roman

-... ..15. '

harbor town

200

400 r

..

. .

Figure2.3. Examplesof spatial betweentractsand site/ relationships scatters: site/scatterfallscomA) pletelywithin one tract;B) site/ scatterextendsovermorethan one tract;C) locationsof surveyunits at OrmosVathy, showingseveraltracts and site/scatters within SS93-8, the harborsettlement Roman-period

complexity is introduced if, as in example B, one or more site/scatters extends over the surface of several tracts. Example C, showing the actual survey locations at Ormos Vathy, illustrates the spatial complexity that frequently exists between concentrations of cultural debris and the survey units that are superimposed on them. Our investigations at Ormos Vathy in 1993 confirmed the suspicion that this was the main port for the city of Nikopolis in the Roman period. Through scouting and tractwalking, the approximate limits of the main part of the port town were determined, and the entire enclosed areawas designated a site (SS93-8). In the numerous tracts that were walked within the site, several smaller loci of important, concentrated activity were discovered, including collapsed architecture from buildings that stood on the water's edge (SS93-5), domestic assemblages, a probable purple dye factory (SS93-23), and an inscribed tombstone (SS93-35). As a result, several site/scatters were designated within the larger site/scatter. We found this system of designation (i.e., site/scatters within site/scatters) a convenient way to portray the spatial complexity of human behavior, without having to create new terms or procedures.
URBAN SURVEY

73. Bintliff and Snodgrass1988b; Alcock 1991.

In recent years, a suite of methods based on the tract concept has been developed for surveying the surface of large, complex sites such as the urban areas of classical Greekpoleis.73It was realized that a failure to survey the large settlements that often anchored regional systems amounts to an inversion of the tunnel vision for which traditional excavation has been justly criticized, and leads to regional studies that are similarly incomplete. The need for surface survey on large settlements that have been previously excavated is no less urgent; even the most extensive excavations usually

42

THOMAS

F. TARTARON

Intensivesurvey of examineonly a smallpercentage the total site area.74 similarto those developedfor off-site contextsoffer a means techniques by which the completeextent of complex,multiperiodsettlementsites informationand detailedpatmay be examinedfor overallquantitative combenefitis the potential An additional ternsof internal periodization.75 the datato those obtainedfromoff-site contexts. of parability Previousurbansurveysin Greece have relied on the replicationof with the additionof supplemental tractprocedures, grabsamplesand/or artifactsamples controlledcollections,as a way to effect representative The andnear-total groundcoverage.76 methodsadoptedby the Nikopolis collecProjectdifferfrom these in that insteadof addingsupplemental were simplyintensified,creatingsmallersample tractprocedures tions, and in contextsandresulting enhancedcontroloverartifact densityinforwere armation.Tractswere laid out in the samefashion,and surveyors werecountedandcollectedaccording of at aninterval 5 m. Artifacts rayed at with countsrecorded 30to the sameguidelinesusedfor off-sitetracts, a m intervals. Unlike off-site tracts,however, samplewas routinelytaken createdsamplecells of approxiThis modification at each30-m interval. Documentamately30 x 25 m (dependingon the numberof surveyors). The dataobwerethe sameas those observedfor tracts. tion procedures of tainedfromurbansurvey permitthe construction densitymapsforeach localizedfuncof and periodof occupation, offerthe possibility discerning of tionalattributes cells or groupsof cells. the to innovations wereintroduced accommodate overwhelmCertain in of materialoften encountered urbancontexts.Surveyors ing density of for wereprovided with two tallycounters recording quantities the most of commonclassesof artifact: potterysherdsandfragments tile andbrick. were collected,they were Since all examplesof other classesof material archiUrbansites often preserve countedwhen sampleswereassembled. tecturalfeatures, includingclearremainsof houses and other buildings, team leaderswere wells and cisterns,and towers.In anotherinnovation, of allowedto treatsuch featuresas uniquetracts,regardless their size or with the same methodsused on shape.They could then be investigated to otherurbantractsor,if deemedappropriate, subjected total collection such contexts,the integrity or samplingat closerintervals. separating By couldbe of artifacts mayformcoherentandmeaningful that assemblages preserved. of Urbansurveys wereinitiatedat a smallnumber fortifiedtown sites in of in 1994.The most comprehensive thesewas at Kastri, the floodplain of the lowerAcheronRiver(see Fig. 2.1). DakarisidentifiedKastrias the site of the Elean colony of Pandosia,mentionedby Strabo(6.1.5) and out was others.77 Intensivesurvey carried on 11 ha of this 33-ha site,repMore than 85,000 and resentingmost of the accessible walkableterrain.
artifactswere counted, including over 15,000 pottery sherdsand over 70,000 tile and brick fragments, yielding an average figure of more than 7,700 artifactsper hectare. A topographic survey carriedout at the same time as the urban surveygenerated a new site map, incorporating several architectural features that were discovered in survey tracts.The results of the urban surveys at Kastri and elsewhere will be presented in volume 2.

74. Alcock 1991, pp. 422, 443. 75. Bintliff and Snodgrass1988b,

p. 58.
76. Bintliff and Snodgrass1988b, pp. 58-59; Alcock 1991, p. 448. 77. Dakaris 1971, p. 164. For furtherdiscussionof the identification of Kastri,see Chapter6.

THE

ARCHAEOLOGICAL

SURVEY

43

WALKOVERS

In our terminology, walkover was a multipurpose designation referring generally to any systematic reconnaissance that was conducted differently from regulartract or site/scatter practices.The main function of the walkover designation was to provide documentation in a standard format for activities that lay outside the methodology of systematic, intensive survey, in recognition of circumstances in which treatment in the manner of a tract was impracticalor undesirable.Walkovers documented the following kinds of investigative activity: extensive systematic off-site survey,extensive nonsystematic survey (scouting), site revisits, and resurvey of tracts. Extensive systematic survey,defined as any investigation carried out at a systematic spacing interval of greater than 15 meters, and exploration of territories by nonsystematic means (scouting), formed integral elements of the Nikopolis Project sampling design that allowed the rough estimation of local environmental and culturalfeatures,often in advance of placing intensive survey units (see the discussion of sampling above). Systematic walkoverswere mapped, photographed,and documented on a walkover field form similar to the tract form; nonsystematic walkovers were documented by narrativenotes, with maps and photographs where appropriate. Additional visits to previously investigated units (e.g., site revisits and tract resurveys) were also designated walkovers. Tracts that were walked for a second or third time were documented as walkovers, even if tract procedures were used; the artifacts and documents associated with the walkover were linked to all previous investigations of the tract.
SITE REVISITS AND RESURVEY

78.Theseteamsareoftenknownas
"verification" teams, a term I find becauseof the implicadissatisfactory tion of a group of expertsthat functions in part to correcterroneousobservations made in the discoveryphase (though this is certainlynot out of the question).In many cases,valid observationsmade duringthe initial visit will no longer be justified at the time of the revisit.An importantrole of the revisitteam is thereforeto measure the changes that have taken place in the surfacescatter. 79. Ammerman 1981, 1985, 1993.

The information acquiredin the fieldwalking phase was supplemented by comprehensive follow-up studies of selected tracts, site/scatters, and other locations of interest on the landscape. This second phase was normally initiated by revisit teams composed of archaeologists,geologists, and geomorphologists.78The initial purposes of the revisit were to check the information supplied by the discovery team (e.g., surface visibility, dimensions, chronologicalperiods represented);to describethe geomorphological setting; and, if necessary,to make additional observations and collections. Revisit teams also visited locations not designated as sites or scatters by the discovery team, but deemed worthy of reconsiderationbased on analysis of the density data and the artifactsamples.These activities were documented as walkovers,and were accompanied by specialist reports as necessary. Additions and alterations were also made to the original tract and site/scatter documents. Based on the findings of the discovery and revisit teams, certain units were selected for furtheranalysisinvolving resurveyor geophysicalprospection. Resurvey of a small percentage of tracts was a planned activity that allowed archaeologists to measure the effects of changing conditions of discoverybrought about by naturalprocesses (erosion, deposition, and vegetation growth) and human actions (such as agriculture and land clearance).79Since conditions of access and surface visibility may exercise a

44

THOMAS

F. TARTARON

X,

4.,

'
" ' :'"':* -- . '':': . ^ :':.. :: * ?* ; *

Figure2.4. Generalview of the site at Grammeno(SS92-6).View is fromthe east end of tract93-1, the designationin 1993 for the samearea the surveyed previousyearas tract 92-39.

profound influence on the archaeological document that is produced for a given landscape, resurvey results served to qualify and inform inferences drawn from surface data. Geophysical prospection was another integral component of comprehensive site investigation. Remote-sensing techniques, including resistivity, electromagnetic conductivity, magnetometry, and ground-penetrating radar,were applied to a number of sites, some previously known and others discovered by the surface survey.Geophysical surveys were often performed on sites with high artifact densities but no visible architectural remains as a means to test for subsurface structural features. In several cases, the likely existence of buildings, roads, and other featureswas established. An example from the locality "Grammeno,"near the modern village of Archangelos, illustrates the typical manner in which such detailed studies came together. In the course of walking tracts in the Grammeno areain 1992, a fallow field walked as tract 92-39 was found to contain massive amounts of pottery sherds and brick and tile fragments, as well as other materials, including ground stone, flaked stone, and bits of glass, limestone, and metal slag (Fig. 2.4). Over 17,000 artifactswere counted in the field, which measures ca. 1 ha in area; this quantity contrasted sharply with that encountered in the surrounding tracts. The field was immediately designated a site (SS92-6). Later in the 1992 season, SS92-6 was resurveyed,resulting in an additional collection of diagnostic artifacts.Preliminary analysis suggested a domestic assemblage of Roman times, indicating the presence of perhaps a farmstead or small villa. In the winter of 1992-1993, it was discovered that the fallow field had been freshly plowed, bringing to the surfaceconsiderablequantities of new material. Consequently, the site was again surveyed at the beginning of the 1993 season, this time yielding a remarkabletotal of over 36,000 artifacts, among them many diagnostic types not previously collected there.80 It was noted that the dimensions of the definable scatterhad not increased,81 but that the chronology and function of the site were clarified. Later in the

80. Although a portion of this twofold increasein artifactsis accounted for by the furtherfragmentation of surfaceobjectsduringplowing, the recoveryof numerousnew diagnosticceramic,glass, and stone artifactsindicatedclearlythe richness of the subsurfacedeposits. 81. We cannot be certainof this finding, however,becausethe adjacent field most likely to harboran extension of the site remainedheavilyovergrown throughoutthe period of our investigations.

THE

ARCHAEOLOGICAL

SURVEY

45

field season, a geomorphological profile was made of the environsof SS92-6. The site was also selectedfor geophysicalsurvey, owing to the the dense,well-delimitedscatter; suspicionthat brokenbits of limestone of derivedfromarchitectural blocks;and the generalamenability the site conducElectricalresistivity, to geophysical electromagnetic techniques. on tivity,and magnetometry surveyswere performed gridslaid out over resultsindicatethe existenceof one portionsof the site.The preliminary well with the and or morestructures at leastone linearfeature, accord and remains.82 madefromthe surface inferences of at conditions prevailed SS92-6 in the summer 1994. Verydifferent of winterrainsandthe return the fieldto fallowcausedthe ground Heavy to be obscured scrubby by vegetation.It was also noted that a substantial of had takenplace,probably a consequence earth as reburial artifacts of of movementduringpluvialrunoff.A resurvey the tractin 1994, in circumstances dramatically of reducedvisibility, produceda count of 3,000 far two artifacts, fewerthanin eitherof the previous years. at on The historyof investigation SS92-6 is instructive manylevels. it to Fromone perspective, is a testament the amountandquality inforof mationthat can be recovered intensive,multidisciplinary by studyof surfacephenomena. anotherpoint of view,it muststandas a cautionthat To the vagaries time andnature havea profound of can effecton ourinterpreand of tationsof surfacepatterns, as an affirmation the need for replication studies.Had we firstdiscovered site in 1994, ourperception it this of wouldundoubtedly quitedifferent. be

CONCLUSION
In responseto a unique set of research conditions,we chose a strategy for and reconnaissance methods. former The calling bothextensive intensive us to acquire, a coarsescale,information on aboutthe diverse permitted of landscapesof southernEpirus,and the distribution culturalremains of uponthem.The latterwereundeniably high intensityin termsof spacand ing intervaland the numberof datameasurements, wereintendedto furnishhigh-resolution data about a smallersampleof locationswithin the surveyarea.We willinglysacrificed broadarealcoverage the belief in that the total spectrum approaches of wouldprovidea good initialunderof of standing the region's throughanexceptionally expanse time. past long The successor failureof the intensivesurveywill becomeevidentas continueand the resultsareheld up to the scrutiny the scholof analyses andthosewho followus in the studyof the Epirotelandarlycommunity studies,makinguse of the surfacedata obscape.Detailed interpretive tainedso meticulously, havebegun to appearandwill continueto do so, both in volume 2 of this series and elsewhere.83 are remindedthat We decisionsin survey are designandexecution verymuchaboutcompromise among competingpriorities,and balancingopportunitiesagainstconstraints. The final evaluation must consider, amongotherthings,the deto which the intensivesurveyservedthe overallaimsof the project. gree

82. Weymouth 1993. 83. E.g., Moore 2000; Tartaron 1996, 2001; Tartaron,Runnels,and Karimali1999.

CHAPTER

THE EARLY STONE AGE OF THE NOMOS OF PREVEZA:

LANDSCAPE AND SETTLEMENT

H. by CurtisN. Runnelsand Tjeerd van Andel

INTRODUCTION
The first extensive and methodical search for Palaeolithic sites in Greece was begun under the direction of Eric Higgs of the University of Cambridge in 1962.' Palaeolithic finds had been made there before as the result of other activities, but Higgs's survey was aimed specifically at identifying Palaeolithicsites and cast a wide net over nearlyall of northernGreece from Thrace through Macedonia to Epirus. It was in Epirus that Higgs's team concentrated their efforts because of the large number of sites suitable for excavation (for place-names, see Fig. 3.1). Higgs tested three sites by excavation, the cave sites of Asprochaliko and Kastritsaand the openair site at Kokkinopilos, all in the 1960s.2 For various reasons these excavations were not published and researchlanguished after Higgs's death.Work was resumed in 1979 by another Cambridge University team under the direction of G. N. Bailey (now at the University of Newcastle-upon-Tyne), with additional surveys, geological research, and an excavation of Klithi Cave in the Zagori, northern Epirus.3 These two projects established Epirus as the center for sustained Palaeolithic research and attracted other scholars to the field. A Greek team is now excavating the cave of Boila near Klithi, and the work reported here is partof an internationalGreek-American effort,the Nikopolis Project, centered on the nomos of Preveza.4 The integrated evidence from geoarchaeology and archaeology relating to the long history of human settlement and land use is the unifying theme that connects the scholars working on the many different periods covered by the Nikopolis Project. In this context it was our aim to deepen and extend the work of our predecessors in order to obtain a detailed picture of the early prehistory of the region. By far the longest episode of human occupation of the region, the Palaeolithic and Mesolithic span the later Pleistocene and earliest Holocene and so provide a foundation for the study of later prehistoric and historical periods. To accomplish this goal we concentrated our efforts on the investigation of the geological history of the red Pleistocene sediments that in this region are so closely associated with human artifacts, and extended the

1. Dakaris,Higgs, and Hey 1964. 2. Higgs and Vita-Finzi 1966; Higgs et al. 1967. 3. Bailey 1992. 4. For Boila and Klithi:Bailey et al. 1999; Kotzambopoulou, Panagopoulou, and Adam 1996.

48

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

searchfor new sites to those parts of the Preveza nomos that lie west of the Louros valley toward the sea. In order to obtain the most complete picture possible of land and resource use, we paid special attention to the search for very small findspots marked by only a few artifacts in a limited area. When we accepted the invitation of its directors to join the Nikopolis Project,we were especially interested in three topics. Could we extend our methods of studying Palaeolithic sites in their landscape contexts-methods developed in the course of fifteen years of surveys in the Argolid and Thessaly-to the quite differentlandscapesof coastalEpirus?Second, could we, in the face of strong skepticism regarding the value of studying Palaeolithic open-air sites,5demonstrate their essential role in developing hypotheses of Middle Palaeolithic land use? Finally,we wanted to test the utility of our paleosol stratigraphyfor the exploration for Palaeolithic sites and their relative dating.6
THE IMPORTANCE OF PALEOENVIRONMENTS

In prehistoric research it is essential to place all traces of past human activities in a framework of detailed, dated paleoenvironmental reconstructions. Humans have made an impact on the natural environment, chiefly through predation and the use of fire, from the time they arrived on the scene until they became a major determinant of the environment with the introduction of agriculture.In Palaeolithic and Mesolithic landscapestheir impact was still small, and it was the environment that played a central role in structuring human activity.To understand their movements and activities we must take into consideration all of the evidence for tectonic activity, geomorphological processes, climate, sea-level changes, hydrology, and flora and fauna. Like Pleistocene landscapeseverywhere,those of southern Epiruswere very different from today. At times, the climate was considerably colder and especially drier;glaciers capped the high ranges of the Pindos and the now submerged continental shelf was a large coastal plain exposed by lowered sea levels. Reduced runoff and, as the level of the sea fell, steepened river gradients greatly altered the cycles of aggradation and incision. The flora was drasticallyreduced to a shrub and sagebrush (Artemisia)steppe and most tree species were isolated in sheltered refugia in lowland valleys. Large herds of herbivores browsed the lower slopes, valley bottoms, and coastal plains, where bison, wild ass, antelope, and aurochs were vastly more numerous than their human predators. At other times, the climate would swing to a warmerinterstadialphase or even a full interglacial.Melting ice caps caused the sea to rise, drowning the prime grazing land of the coastal plain and reshaping the coastline. Trees recolonized valleys and plains, and deer, elk, pig, and other forest species flourished. Rivers carriedglacial outwash, aggrading valley floors, building deltas, and burying or destroying sites of previous human occupation. Throughout, the cycles of cold and warm climates were accompanied by tectonic activity that might change drainagepatterns, raise an area and so expose it to erosion, or cause another to sink and be turned into a lake. The nature and continuity of human exploitation were continuously

5. Bailey,Papaconstantinou, and Sturdy1992. 6. van Andel 1998a.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

49

Figure3.1. Map of Epirusand areas surrounding

20km

LEUCAS

affected by such environmental changes and can only be properly interpreted when they can be related to them. Among Greek landscapes, the nomos of Preveza is distinctive for the widespread dominance of limestone landforms (karst), including numerous basins of internal drainage controlled by bedrock patterns and tectonics. Because they supply reliable sources of fresh water and associated resources, these basins have always been major factors in determining the pattern of human occupation in coastal Epirus. The explanation of the relations between karst features and Pleistocene archaeology is our most important finding, with the widest implications for understanding past human behavior.

GEOLOGICAL

FEATURES

AND

PAST

HUMAN

BEHAVIOR

Consciously or subconsciously,the images that come to mind when contemplating Late Pleistocene landscapes inhabited by human beings are dominated by the action of glaciers, rivers, and the rise and fall of the sea. It was a landscape of floodplains and river terraces with a backdrop of

50

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

mountains and distant glaciers that informed much of the thinking of Eric Higgs and his coworkers.7Yet the landscape west of the Pindos from the Gulf of Corinth to the Albanian border and beyond is, with few exceptions, not like that at all. Here the main influence on the landscape of the past five million years has been the dissolution of its limestone bedrock during uplift and subsequent peneplanation and the more recent creation of basins of internal drainage through renewed uplift accompanied by extensive faulting.The resultis a classic karstlandscape that contrastssharply with the riverlandscapes ofThessaly, Macedonia, both sides of the Gulf of Corinth, and parts of the Peloponnese. Being mountainous and tectonically active, the Epirus karst landscape is dominated by erosion. More than three-fourths of its surface is being worn down continuously, mainly by dissolution, and only a few localized areas receive and preserve the sediments that record the years since human beings first walked on Greek soil. These sediments, produced by the dissolution of limestone and widely but loosely known as terra rossa, were deposited in depressions of the rugged karst plains and in the numerous closed basins createdby recent tectonics. They are closely and fundamentally associated with the Palaeolithic inhabitation of the region. Here and there are river-dominated landscapes, some as extensive as the tectonic graben that underlies the Ambracian Gulf and has caused the development of coastal Epirus's main rivers, the Louros and Arachthos, others minor, such as the lower Acheron valley.Thus riverine landscapes, always valued by ancient humans, are the exception not only in coastal Epirus but also farther north along the Adriatic and in the Dinarides. What attracted Palaeolithic hunter/gatherers to these often-barren karst regions and governed the pattern of their movements there?We shall show below that the swamps and lakes of the many basins scattered among the precipitous, barren slopes of the mountains were the main attraction, together with the resourcesof the coastal plains when exposed by low glacial sea levels. The resources of the karst and lake environments were so important that the rationale for the distribution pattern of Palaeolithic sites in coastal Epirus can only be understood in the context of their paleoenvironmental reconstruction.

PREVIOUS RESEARCH
The first phase of Epirote Palaeolithic research(1962-1967) began under the direction of Eric Higgs and combined both survey and excavation. The survey involved the inspection of all caves, rockshelters, and redbeds to identify likely places for more detailed investigation and excavation. Only preliminary reports on this research have been published and it is sometimes difficult to know exactly where in Epirus Higgs and his team went.8 It is nevertheless clear that Higgs identified two broad Palaeolithic phases, an earlierMousterian chiefly representedby open-air sites associated with redbeds (e.g., Kokkinopilos, Louros, and Stephani), and a later Palaeolithic presence in small rocksheltersand caves such as Asprochaliko and Kastritsa.9

7. Higgs and Vita-Finzi 1966; Higgs et al. 1967; Harrisand VitaFinzi 1968; MacLeod and Vita-Finzi 1982. 8. E.g., Dakaris,Higgs, and Hey 1964; Higgs and Vita-Finzi 1966; Higgs et al. 1967. 9. Bailey 1992.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

5I

10. Dakaris,Higgs, and Hey 1964, pp. 215-221. 11. Dakaris,Higgs, and Hey 1964. 12. The discrepancy between radiocarbon yearsand calendaryears, nearlyalwayssignificant,is particularly largebetween ca. 10 and 50 kyr B.P., when 14Cyearslag up to 4,000 years behind the calendar(Laj,Mazaud, and Duplessy 1996). Given the rapidly oscillatingpaleoclimateof the interval, calibrationis essential(van Andel 1998b). Therefore,we have calibrated all radiocarbon dates cited in the literaturein calendaryearsas thousands Each of yearsbeforepresent (kyr B.P.). calibrateddate is followed by the radiocarbon date, cited as b.p. and without the publishedconfidence limits. 13. Higgs and Vita-Finzi 1966, p. 21.

At this early stage of research Higgs made a detailed study of one of the richest of the open-air sites at Kokkinopilos ("redclay"in Greek). The study pinpointed thirteen locations where Higgs thought he could identify "chipping floors,"by which he appears to have meant assemblages of Two of these chipping floors were tested by excaflintknapping debris.10 vation (sites a and P, hereafter Alpha and Beta) with somewhat ambiguous results. Large numbers of worked lithic artifacts (ca. 800 in site Beta) were recovered and, although the flints were not associated with features or organic remains, Higgs regarded them as in situ occupation surfaces. Site Alpha produced a typical late Palaeolithic assemblage with backed blades, along with typical Middle Palaeolithic artifacts, and site Beta an early Palaeolithic (Mousterian) industry. It is unclear what stratigraphic relationship the two sites with their different industries might have, especially because they were not placed on a plan or map and are difficult to identify today. Nevertheless, the artifacts from the excavations complemented the unprovenanced surface materials, and it is reasonable to conclude that there are two periods of occupation at Kokkinopilos, early and late Palaeolithic. A typological analysis of the early Palaeolithic artifacts by Mellars described the Kokkinopilos assemblage as a variant of the Mousterian with many varieties of side scrapers,Mousterian points, and He bifacial foliates ("leafpoints")."1 considered the Kokkinopilos assemblage to be a mixture of different industries. A metrical analysis of the flints allowed Mellars to compare the industry (or industries) with the European Mousterian and he concluded that the Greek industry was distinct enough to be considered a separate entity, although it shares some characteristicswith the typical Mousterian. To clarify the stratigraphicposition of the Mousterian, Higgs excavated a rockshelter at Asprochaliko, approximately 4 km northeast of Kokkinopilos in the Louros River valley. The stratified deposits of the rockshelter are approximately 2-4 m deep. The Mousterian is found in layers 14-18. After a stratigraphichiatus, the Mousterian layers are overlain by layersrich in Upper Palaeolithic artifacts,dated by a series of radioThe carbon assays to 17-29 kyr B.P. (thousands of years before present).12 majorityof the Mousterian levels were too old to be dated by the radiocarbon method, but were assumed by Higgs to be older than ca. 39 kyr B.P. (at that time, the effective upper limit for detecting radioactive carbon).The age of the earliest deposits of the Mousterian is unknown. Higgs divided the Middle Palaeolithicinto two units, the earlier"basal" Mousterian and a later Mousterian called "micromousterian" because of the small size of the flints. He related the basal Mousterian to the Kokkinopilos industry on the basis of the use of similar fine retouch, but he also noted many dissimilarities between the two assemblages,particularly the absence of leafpoints at Asprochaliko.13 The later Upper Palaeolithic layers at Asprochaliko have Gravettian and Epigravettian backed-blade industries. The industrial succession seen in the flints is reflected also in the animal bones. The Mousterian was found with an extinct Pleistocene megafauna, including Merck's rhinoceros, aurochs, bison, buffalo, antelope, and wild horse. The Upper Palaeolithic shows a marked change in the fauna, with a great emphasis on the hunting of horse, red deer, ibex, and chamois.

52

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

at south of Higgs conducteda thirdexcavation the site of Kastritsa, Ioannina.Here he found a long sequenceof UpperPalaeolithic backedblade industries, dating to ca. 13-24 kyr B.P. and similarto those from but differences the lithicsandfauin Asprochaliko, therewereimportant nal remains.14 Becausethe sites were largelycontemporary, these differences encouraged and Vita-Finzi to postulatethat the sites were Higgs seasonalcampsor specializedhunting standsalong the route of faunal migrationfrom coastallowlands,used in the winter,to uplandsummer Their model was based on the differentlocationsand elevapastures.15 tions of the two caves,the differencesin fauna and artifacts,and on a with transhumance activities modern of Sarakatsani comparison thepastoral shepherds. In sum,Higgs established evidencefor two phasesof humanoccupation in the Middle and UpperPalaeolithic. Middle Palaeolithic The was foundchieflyon the surface undatedredbedsiteswhereit was not assoat ciatedwith animalbones.The Upper Palaeolithic from the rockshelters was interpreted with an innovativemodel of seasonallogisticsbasedon This considerable in advance prefollowingherdsof migrantmegafauna. historicknowledgeput Epiruson the map, as it were, and the state of knowledgestood herefor morethan a decade. BecauseCorfu was connectedwith the mainlandof Epirusduring much of the last glacialperiod,the two areascan be regarded a single as culturalregion, and a researchproject undertakenin 1964-1966 by revealed similarsequenceof cultures.16 a Sordinas loAugustusSordinas cated and sampleda large numberof Middle Palaeolithicsurfacesites associated with redbeds.He classifieda Middle Palaeolithic industryas with the sameregional features typicalMousterian, (e.g., leafpoints)as in at Epirus.His excavation GravaCavebroughtto light an undatedUpper Palaeolithic similarto those fromAsprochaliko Kastritsa.17 and industry His mostimportant was discovery a low tell at Sidarion the northcoastof the island,whichwasexcavated the 1960s.18 in Sidariproduced sequence a of threemajor datedto the Mesolithic,the EarlyNeolithic,andthe layers EarlyBronzeAge. After the deathof Eric Higgs, prehistoric research Epiruswas inin until 1979, when a new projectwas initiatedby Bailey.This terrupted the and projecthad several goals.One was to reexamine excavations surveys of Higgs with a view to updatinghis conclusionsaboutsettlement andlanduse, specifically modelof transhumance.19 the Anothergoalwas a campaignof research northernEpiruscenteredon the excavation in of the rockshelter Klithi in the Zagorinear Konitsa.The resultsof this of
project have now been published in full.20

The excavation records Asprochaliko, of and Kastritsa, Kokkinopilos 14. Galanidouet al. 2000. wereinspectedand checkedagainstthe remaining sectionsand the finds 15. Higgs and Vita-Finzi 1966. storedin the magazines the Ioannina of 16. Sordinas1968. Museum.Further Archaeological 17. Sordinas1969. work was undertaken Asprochaliko clean the section and extract at to 18. Sordinas1970. and burnedflints for dating analysis.21 study of the Middle A samples 19. Bailey et al. 1983a. Palaeolithic artifacts usefulin correcting was some of the earlier viewson 20. Bailey 1997. the Mousterian. was shownthat the earliest"basal" It Mousterian(layers 21. Huxtableet al. 1992; Bailey, 16 and 18) madegreater of the Levalloistechniquefor the production Papaconstantinou, Sturdy1992. use and

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

53

of largebladesor bladelikeflakes,while the laterMousterianof layer14 madeless use of the Levalloistechniqueandwas characterized the use by of Mousterianpoints and small,atypicalpseudo-Levallois points called called"micromousterian" by points.The laterMousterian, Asprochaliko to was shown by reanalysis differlittle from"basal" Mousterian, Higgs, are as and the two assemblages now regarded similarin compositionand is The at typology.22 conclusion thattherearenot two typesof Mousterian andthe different labelshavebeendropped. New datesindiAsprochaliko, cate that the Mousterianat Asprochaliko spansa considerable periodof time.The Mousterian the lowestlayersmaybe as muchas 90-100 kyr of of while not precisely B.P., while the Mousterian the upperlayers, dated,
may date to 39 kyr B.P. or later.23

22. Bailey,Papaconstantinou, and 1992. Sturdy 23. Huxtable et al. 1992; Bailey, and Papaconstantinou, Sturdy1992. and 24. Bailey,Papaconstantinou, Sturdy1992. 25. Bailey,Papaconstantinou, and Sturdy1992, p. 140. 26. Bailey,Papaconstantinou, and Sturdy1992, p. 142; cf. pp. 91-95, below. 27. Bailey et al. 1983a, pp. 76-77. 28. Bailey 1992, 1997.

reexamination the excavation at Kokkinopilos acof sites was Bailey's an attemptto date the redbedsediments. The conclusions by companied reachedby Bailey'steam differfrom those proposedearlierby Higgs.24 The artifacts fromsitesAlphaandBetahadbeenconsidered Higgs and by with Pleistocene his colleagues partof in situchippingfloorsassociated as red clay depositsthat had accumulated relatively rapidlyas the resultof in erosionandaeoliandustdeposition. Baileyandhis colleagues, contrast, concludedthat the redearthdepositsin the regionweremucholderthan In Pliocenein age.25 their the Middle-LatePleistoceneandwereperhaps view, the depositswere formedas a resultof dissolutionof limestone,a that They conprocess theybelievedto haveceasedbeforethe Pleistocene. were muchyoungerthan the redbedsand cludedthat the lithic artifacts into wereincorporated accidentally the depositsat a latertime ("reworked" in their terminology), perhapswashedfrom the surfaceinto gullies as a resultof tectonicuplift,slumping, ponding,andgullyerosion. As a resultof this analysisBaileyconcludedthat the open-airsites of often of greatly were essentiallyfortuitousadmixtures lithic artifacts, with the redbeds. fromthe fact that the lithicswere differingages, Apart in associated with the redbeds spatialterms,Baileyattached little analyticalvalueto thembecausetheywereunstratified couldbe datedonlyin and units of 100,000yearsor more.26 Much attentionwas given to the new excavations Klithi,which at of Baileyand his colleaguesplacedat the centerof a new interpretation the Higgs model of Upper Palaeolithictranshumance. first Bailey At a or proposedthat Klithiwas probably "homebase" basecampat the center of a hierarchical settlementpatternconsistingof smallerseasonally occupiedsheltersand specialized activitycampsspreadacrossa largeand diverse This hypothesis to be abandoned had when geographically region.27 the excavations revealedKlithito be a smallspecialized huntingcamp.28 Klithiis locatedin a rivergorgein the Zagori,at the headof the Konitsa plain, and was occupiedonly duringthe summerover a periodof a few thousand In yearsfollowingthe lastglacialmaximum. thattime the nearby and peakswererelatively unglaciated the campwas usedfor huntingibex. Highly fragmentedibex bones and horn cores,along with many thousands of backedblades (from projectiles) and end scrapers (for cutting meatandprocessing indicate shortbutintensespecialized the hunthides),
ing activities.

54

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

This evidence, when combined with a reanalysis of the fauna at Asprochaliko and Kastritsa,allowed Bailey's team to conclude that there was a degree of specialization in hunting and that the three excavated In rockshelterswere hunting stands.29 recent years, noting that the exploitation territoriesof the late glacialwere very large and embracedthe coastal plains, Bailey's team proposed a modified version of Higgs's model to explain the data.The new model relies heavily on the study of bedrock, flora, The and the distributionof animalpopulations.30 model posits "topographic from active tectonics and composed of rock supporting barriers," resulting a flora of unpreferredspecies, that bounded discrete zones where the bedrock supplied preferredbrowse for horse, deer, chamois, and ibex. Human predators did not "follow"the herds, as in Higgs's formulation, but pursued a round of residential mobility, shifting periodically from one rockshelter camp to another. Each camp was strategically located on the edge of one of the zones of preferredbrowse, taking advantage of the tendency of animals to concentrate in these zones. The caves gave the hunters access to water and shelter just out of sight of the herds. Following the seasons, animals were drawn to the mountains in the spring and summer, and the hunters shifted their camps accordinglyin a periodic seasonalfashion, only to return to the coastal plain in autumn and winter.This logistical strategy of land use took full advantage of the available resources and sustained hunters over a period of 10-20,000 years.The maximum extent of the activity,however, appearsto have been in the millennia immediately following the last great glacial advance between 20 and 25 kyr B.P.These modifications to the Higgs model advance our understanding of the period considerably,and the new model incorporates a rich body of paleoenvironmental data derived from the study of the geomorphology of the Klithi region and two major pollen cores from northern Epirus.3

THE LATE QUATERNARY WESTERN EPIRUS

GEOLOGICAL HISTORY

LANDSCAPE

OF

Epirus is located at the meeting point of three tectonic plates (Fig. 3.2) whose rapid convergence (10-15 mm/year) is causing widespread deformation of the whole region west of the Pindos, including subsidence of The deformationis regardedby Bailey and coworkers the AmbracianGulf.32 as a major force which affected the Palaeolithic landscape by creating, at a rate perceived on a human timescale, "topographic barriers"that had a large impact on Palaeolithic resource availability and use.33The required rates of uplift, however, seem excessive and to evaluate this proposition we examine the tectonic state of the region in some detail. From the Early Mesozoic to the Late Eocene, Epirus formed part of a vast midocean plateau covered by shallow marine Jurassic limestones In overlain by deepwaterlimestones of Cretaceous-Eocene age.34 the Oligocene the eastern portion of the Pindos range began to rise and sand and silt were shed westward, covering the limestones with flysch deposits. Intense deformation with a northwest trend, still visible in today'slandscape,

29. Bailey et al. 1983b. 30. Bailey,King, and Sturdy1993. 31. Bailey et al. 1990; Bailey 1997; Willis 1994. 32. Kahleet al. 1993; Le Pichon et al. 1995. 33. Bailey,King, and Sturdy1993, fig. 5; King, Sturdy,and Whitney 1993; King and Bailey 1985. 34. Aubouin 1959;Jacobshagen 1986.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

55

After Le Pichon et al. 1995, fig. 1

Figure3.2. Tectonicsof northwestern Greeceandthe IonianSea. Offshorefromthe Ambracian Gulf the subduction the Mediterranean of plateunderthe Peloponnesian continentalmarginchangesto a collisionbetweenthe Italian/Apulian and Greekcontinentalblocks.The changeis marked a strike-slip by fault(opposingarrows), the transform fault. Cephalonian ended this phase. In the Miocene, the emergence of western Epirus created a land floored predominantly with limestone that was eroded down to a broad peneplain (a rough planarsurface)during the Pliocene.35Raised remnants of this peneplain can still be discerned from the widespread uniform elevations of high plateaus, for instance around Loutsa southeast of the lower Acheron valley. Late in the Pliocene, the collision with the Apulian block reinitiated the deformation of the region west of the Pindos front from the Gulf of Corinth to Albania.36 The compression, continuing today, resulted in reof the Pliocene peneplain, while cross-faulting createdyoung gional uplift grabens in the Ambracian Gulf,37Kalamas delta, lower Acheron valley, Doliana basin, and elsewhere (Fig. 3.3). The region is thus seismically active and geological maps show many strike-slip, thrust, and normal faults (Fig. 3.3).38 For many of those faults present activity has been assumed, but hard evidence is sparse and many faults are relics of the mid-Cenozoic mountain-building phase and now inactive.39 The best evidence for present tectonic activity arethe fresh striae on fault planes, identified with detailed fieldwork by Waters (Fig. 3.4). The high uplift and subsidence rates (up to 100 m in historical times) proposed by Bailey and his group in support of their paleoenvironmental resource models for Late Quaternary Epirus are not locally documented; they rest mainly on an assumedsimilarityin seismic activitybetween Epirus and other active but actually tectonically quite different regions, such as California, New Zealand, Japan,and the Middle East.40How valid are the proposed magnitudes?

35. E.g., Etudegeologique; Schrbder 1986. 36. Clews 1989; Sorel 1989; Underhill 1989. 37. Piper,Kontopoulos,and Panagos 1988. 38. Etudegeologique; Waters 1994, fig. 1.4:b,plans 1,2. 39. Sorel 1989; Underhill 1989; Waters 1994, pp. 168-211. 40. Bailey,King, and Sturdy1993.

56

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

Figure3.3. Possiblyactive(Late tectonicfeaturesof Quaternary) of westernEpirus.The structure the andlower Ambracian, Kalamas, Acherongrabensis muchsimplified. Majorgorgeswereincisedby rivers duringPlioceneupliftof the
peneplain. After Waters 1994, fig. 5.10

Figure 3.4. Present tectonic activity in western Epirus as indicated by fresh striae on fault planes. Compare with presumed active faults shown in
Figure 3.3. After Waters 1994, fig. 5.9

Waters, assuming that deformation began about 3.5 million years ago, showed that prominent fluvial surfaces now raised above the present river level in the Botzara and Epirus synclines indicate that the synclinal axes are rising at 10 m/100 kyr.41The crestal elevations of the Pantokratorand Parga anticlines suggest uplift at 25 m/100 kyr, and regional uplift of 1535 m/100 kyr is implied by remnants of the Pliocene peneplain. During the same interval, the Ambracian Gulf has widened at ca. 2 cm/yr.42 Values of 0-16 m/100 kyr,indicated by coastal deposits of the last interglacial, fall within the same range (see below, pp. 76-83). These rates do not suffice to alterthe Epirote landscapeperceptiblyin the last 100-200,000 years and consequently Bailey's models lack a credible foundation.43

41. Waters 1994, pp. 208, 213. 42. Kahle et al. 1993. 43. Cf. Bailey,Papaconstantinou, and Sturdy1992.

EARLY STONE

AGE OF THE NOMOS OF PREVEZA

57

" Q -!..... : y ....... '""

Figure3.5. Simplifiedbedrockmap of westernEpirusshowingthe of prevalence Mesozoicto Eocene limestonewest of the Pindosfront (PF). AfterEtudeg6ologique

I
"

Quaternary Miocene-Pliocene

**

..

Oligocene |[|| Mesozoic-Eocene limestone

LANDSCAPES

KARST

44. Ford and Williams 1989. 45. Ford and Williams 1989, pp. 396-412. 46. Fordand Williams 1989, pp. 31-34, 96-114.

Landscapes are shaped principallyby two forces, the internal dynamics of the earth that deform its crust and the external forces of climate, water, wind, and vegetation (and recently also human beings) that modify the relief generated by tectonics. The impact of the external forces is strongly influenced by the kind of bedrock on which they operate. Because nearly all large continental regions display a mosaic of diverse rock types, most landscapes have been created by the work of rivers, ice, and wind, modified by zonal climates. In contrast, areaswhere limestones dominate, such as Epirus (Fig. 3.5) and much of former Yugoslavia,are shaped mainly by the dissolution of limestone by CO2-charged rain, runoff, or groundwater.44 A weatheredlimestone land surface,called a karst,drainsmainly downward through cracks and fissures into extensive subterranean conduits, rather than horizontally at the surface by rivers. Characteristic for karst landscapes are thus a lack of rivers, and numerous sinkholes or dolines (Fig. 3.6), round, steep-walled pits, usually flat-floored and up to a few hundred meters across, that derive from collapse of subterraneanchannel roofs.45 Typical dolines can be seen along the road from Loutsa to Strounga on the plateau south of the lower Acheron valley.Another diagnostic karst feature are blind valleys, former stream valleys deprived of surface runoff by the formation of a subterraneandrainage system. Rather than by a whole panoply of processes ranging from weathering to river incision, karst surfaces are molded almost entirely by limestone dissolution and minor slope wash. Solution of limestone removes all calcium carbonate (CaCO3), leaving behind a small, fine-grained, insoluble residue that is iron-rich and hence red. Because the rate of solution varies with small-scale variationsin the propertiesof the bedrock, a rough,jagged surface forms, which is randomly dimpled by depressions.46 There water collect, stand for a while, and gradually dissolve the rock, thereby may

58

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Cross section of a doline

doline

of Figure3.6. Formation a doline (sinkhole)by roof collapseor dissolutionin a subterranean drainage system.The dolineshown herehas a subaqueous in a lake exit or sea,but subaereal outletsarealso common. widening and deepening the ponds. On gentle slopes the residual mantle, called terra rossa in the Mediterranean, slowly thickens and, being finegrained, reduces infiltration, allowing sheet wash to transferthe weathering residue into the depressions (Fig. 3.7: Time 1). In the absence of major tectonic activity,the result is a peneplain dotted with red patches of finegrained, redeposited terrarossa. Over time, both primary and redeposited terra rossa thicken and spread, but because dissolution rates are slow and the insoluble residue constitutes a mere few percent of the pure limestone of Epirus, millions of years may pass before redbeds cover the whole region.47 Recent deformation and uplift of western Epirus have reshaped the Pliocene peneplain into many synclinal troughs, initially not connected with each other or the sea by rivers,and separatedby steep, lofty anticlinal mountain ranges.The uplift has acceleratedthe downslope transferof the terra rossa by slope wash, especially on fault scarps (Fig. 3.7: Time 2). Where exposed faults cut acrosssubterranean channels, springshave turned many closed depressions into lakes drained by swallowholes, called on "katavothres" Greek maps, which draw off excess water into subterranean channels.48 Consequently, enclosed flat-floored basins and others raised and in the process of being eroded dot the landscape of western Epirus (Fig. 3.8; Table 3.1). Common in Yugoslaviaalso, they areknown by the word "polje." The poljes of western Epirus range from ca. 1 to 6 km in length and up to 1 km in width, exluding a few large ones (Table 3.1). Until recently,many had permanent or seasonal lakes (Fig. 3.9), but today most of those have been drained for agriculture. Large, permanent lakes have accumulated thick calcareous deposits; an example is the Ioannina basin, which has preserved a record of paleoenvironmental change going back to the Early Pleistocene. Western Epirus is not the only part of Greece where large karstbasins with lakes are conspicuous landscape components.49Boeotia has its Lake Kopais and, in the Peloponnese, the Stymphalos, Orchomenos, Mantineia, and Tegea basins are of similar origin. In most of those regions deformation ceased or became insignificant some time ago, whereas the continuing tectonic activityin Epirus has lent its karstbasins a specialcharacter.

47. Spate et al. 1985. 48. Ford and Williams 1989, pp. 428-432. 49. Pfeiffer 1963; Sweeting 1985.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

59

Time1 - Peneplanation and dissolution

terra rossa LOUTSA tera
I I I I [ I I I I I I I I I I I I i J ! I I I I I 1 I

ro

redep. terra assa

I I I I I

Figure3.7. Diagramof the genesisof loutsesandpoljeson a karstic Time 1) The karstic peneplain. peneplaindevelopsa mantleof dissolutionresiduecalledterrarossa, whichis washeddownand deposited in shallow, closedkarstsurface calledloutses.Time 2) depressions basinsof internaldrainage Elongate calledpoljesdevelopas a resultof renewedtectonicactivity, which raisesridgesalongnormalor reversed faults.Accelerated weathering and slopewashfill poljeswith terrarossa.Time 3) secondary Continuedupliftraisesthe poljeand surfacestreamscut backinto its erosion,often depositsby headward alongweakenedfaultzones.

!I

I I i , I I I

I II

i [ I I

Time 2 - Faulting,uplift,polje formation

! II'
I

I I

I
i I

I i

',' '......-_::::::::::::::, - 1ww,Xwe@ |I w X . I If I I5 ~.' . @evZ*-{ I I .,:';- ], ,

I I f i Ii Ii .I IN -1I I I I I I I I 1, , II I I' i, :I :I I
.I I I I

redep. terra rossa

ACTIVE POLJE
I.
I

I. .
I I

I I

I I 1,

;/,I,I,I

,I

II I,

Time 3 - Upliftand/or headward stream erosion,

redep. terra
iLOUTSA ^LOUTSA /, /rossa \ DISSECTED y POLJ

polje dissection

50. Nicod 1992. 1993. 52. Budel 1973. See also, e.g., Gams 1973; Rathjens1960.

51. E.g.,Gams1978;Sweeting

Compared with the basins named above, most Epirote poljes are small and many have been raised recently, sometimes to great heights, and are now subjectto erosion. Examples aretwo deeply eroded,fault-bound hanging poljes perched on opposite flanks of the Thesprotiko valley, Kranea (Fig. 3.8:25, 300 masl) and Galatas (Fig. 3.8:27, 200 masl).50 Is it of the essence of poljes that they have structural origins, or are The solusimple solution basins in the karst surfacealso a class of poljes?51 tion basins tend to be shallow (Fig. 3.10) and small and are fed by winter and spring runoff rather than by springs. Their value as a resource in the Palaeolithic context is therefore seasonal and differs much from that of true poljes. For this reason we have adopted for use in this paper the separate term "loutsa"(pl. loutses), the Greek name for a seasonal pond or wet sink. Budel and others have suggested that poljes are fossil elements of the landscape that are no longer being formed.52In a tectonically active area, however,nothing is permanent.While deformation constantly createsnew poljes, continued uplift permits small streams to cut back upstream and capture former poljes, draining the lakes, drying the surfaces, and exposing the stratigraphyin stream incisions (Fig. 3.7: Time 3). Epirus west of the Pindos front contains poljes in every stage of evolution from recent birth (e.g., Valtos Kalodiki, Figs. 3.8:6, 3.9), to old age (e.g., Cheimadio, Figs. 3.8:24, 3.11), to stream incision and removal (e.g., Kokkinopilos, Fig. 3.8:30).

60

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

32

/
I

Cheimadio

%~~~~~~~~~4

/04/

o
F F

Doline Fault

[.i. , .] Coastal or river plain ... Active or drained polje Raised, dissected polje or loutsa
Figure3.8. Poljesandloutsesin westernEpirus.The deposits dominatethe Quaternary landscape exceptfor the GulfofAmbracia grabenandits feederrivervalleys. Sordinas(1983) alsoreports wideterrarossaon spreadredeposited Corfuandadjacent islands.The upperAcheronvalleymayconceala largepoljeor its surfacerunoffmay havebeenlost to underground drainage, makingit a "blindvalley." Numbersreferto Table3.1.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

6i

TABLE 3.1. DIMENSIONS AND ELEVATIONS OF POLJES AND LOUTSES IN WESTERN EPIRUS
Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Name Ayia Kokkinographos Domia Gouri ValtosKalodiki Mavradis Mazaramia Mavrikambos Arapanitos Saita Nerotopos Paramithia Length (km) 1.2 2.0 0.7 2.2 1.0 3.8 10.5 3.0 3.5 1.8 2.0 5.0 18.0 2.0 2.4 3.2 3.5 2.5 1.6 1.3 2.6 1.6 2.7 1.0 6.0 0.6 3.0 4.0 1.6 1.3 2.8 1.1 20.0 Width (km) 0.5 0.4 0.4 0.8 0.5 1.3 1.7 Elevation (masl)
ca. 300 ca. 520 Type

400 380 440

110 110

Raisedloutsa Raisedloutsa Raised loutsa Raised loutsa Active polje Active polje, drained Active polje, drained Active polje Raised loutsa/polje Raisedloutsa/polje Active polje Raisedloutsa/polje Raised loutsa/polje Raised polje Raisedloutsa/polje Blind valley? Blind valley? Blind valley? Subsidedloutsa? Raised active?loutsa Raised active?loutsa Active polje, drained Hanging paleopolje? Active polje, drained Hanging paleopolje? Active polje Active polje, drained Raised dissectedpolje Raised polje/loutsa Raised polje/loutsa Raisedpolje/loutsa Active polje

1.5
1.8 0.6 0.8 2.5 6.0 1.1 0.6 1.9 2.0 2.0 1.2 1.3 1.0 0.7 0.5 0.6 1.8 0.5 1.4 1.5 1.0 0.7 0.7 0.9 6.0

ca. 130 ca. 250 160

ca. 160 100

Kalaboukia
Mesovouni Morphi Lamma Kalosykies Tsapela Kordeli Alonaki Loutsa Pyra Cheimadio Kranea Thesprotiko Galatas Kalyvia-Kraneas Lake Mavri Kokkinopilos Gymnotopos Tsoukka Ioanninabasin

80 260 290 ca. 160

100

50 30 20 10-15 ca. 200 400 400 ca. 300 40-60

ca. 200 ca. 30

20-30 120 ca. 300 ca. 500 ca. 1400

ca. 400

Number = location numberin Figure 3.8. Sites 4-15,17,18, and 31-33 were studied only on topographic(scales 1:50,000 and 1:5,000) and geological (scale 1:50,000) maps. TERRA ROSSA: DERIVATIVES AND LOOK-ALIKES

Terrarossa,once commonin westernEpirusand elsewhere the limein stoneregionsof Greeceandthe Mediterranean, beendefinedformally has
as the weathering residue of limestone in a warm climate that is commonly assumed to have been that of the Late Miocene and Pliocene;53 strictly speaking, terra rossa is a Pliocene paleosol (Fig. 3.12). Other red Mediterranean sediments of different origin exist and mature Mediterraneanpaleosols areusually also red (see below). This diversityof"redbeds"

53. Foucault Raoult1980. and

62

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

and has been insufficiently appreciated the resultingloose usage of the rossa" robbedit of most of its meaning. has term "terra initiallyprecise as This is regrettable because,althoughredbedsarewidely regarded inand in tractable termsof theirgenesis,stratigraphy, geochronology,54 they with Palaeolithic with their vivid colors and their close association are, field a artifacts, welcomeguidein Palaeolithic studiesin Greece,andsurely as in elsewhere the Mediterranean well.55 In practicethere are three main types of red deposits,each with a site role different 1) genesisanda different in Palaeolithic preference: primaryterrarossa,the in situ limestoneweatheringresiduecoveringthe 2) originalruggedkarstsurface; polje and loutsa redbedsthat are terra and redbeds formed in lowplaces(Fig.3.12:a); 3) colluvial rossaredeposited streams alluon by slopecreep,debrisflows,andsmall,braided ephemeral vial fans.56 thirdgroupdiffersfrom the othersbecausethe currents The of coarse material to sandandchertgravel involvedarecapable carrying up size (Fig. 3.13), so makingthem lithologically verydistinct. Paleosolsoften are as red as terrarossa(Fig. 3.12:b) and to confuse bear the two is easy.Red sediments, however, the imprintof theirenvironwhereaspaleosolsrecordthe durationand ment and time of deposition, a of between intensityof the weathering a stablesurface; cleardistinction is for and redsedimentsandredpaleosols thuscritical stratigraphic paleoenvironmental investigations.

In Greecethe Quaternary mainlya time of erosion,assistedsince is the Neolithicby humandeforestation landuse.57 and terrarossa Primary is therefore rare,althoughsmallpocketsremainin the rugositiesof karst surfaces. fromsteeplimestoneslopesby sheet erosionandredeStripped in wet poljesandloutses,terra rossaretains diagnostic grain its fine posited but the typicalredcoloris lost by reduction wet environments. in In size, alluvial terra rossaretainsits color, fans,whichareusually redeposited dry, but is mixedwith coarsecomponents.
STRATIGRAPHY TERRA ROSSA AND SEDIMENTOLOGY OF REDEPOSITED

Almost all terra rossa in western Epirus is now on secondary location, some of it in colluvium and alluvialfans, but most in loutses or poljes. The color of primaryterrarossa ranges from red to darkred (10R 4/6 to 2.5YR 3/6 and 4/6) because of abundant hematite.58Because groundwater preferentially dissolves hematite59and because active loutses and poljes are seasonally or permanentlywet, the colors of redeposited terrarossa tend to be paler,more yellowish in hue (5YR to 7.5YR).6? The reducing effect of a varying groundwater level is also evident in discoloration (gley) to gray and yellowish gray (10YR to 2.5Y 4/6-7/0) as horizontal bands, or mottling (Fig. 3.12:c).The same process is responsible for manganese coatings on fractures and bedding planes and the formation of manganese-iron nodules.61Discoloration of subvertical stripes, probably due to the decay of root systems, is common. The reduction process can be reversed as the sediment dries during long intervals of drought or by uplift.

54. E.g., Schneider1977. 55. Dakaris,Higgs, and Hey 1964; Higgs and Vita-Finzi 1966; Runnels and van Andel 1993a, 1993b. 56. Bull 1972, 1977; Coussot and Meunier 1996; Innes 1983. 57. van Andel, Zangger,and Demitrack 1990. 58. Boero and Schwertmann1989; Mirabellaand Carnicelli1992. 59. Goethite is more stablein wet conditions;see Schwertmann1971. 60. Yassoglou,Kosmas,and Moustakas1997. 61. Boero and Schwertmann1987, 1989; Mirabellaand Carnicelli1992.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

63

Figure3.9. View of ValtosKalodiki, an activepoljepartlyusedfor farmlandandpartlystill in its original the restate,illustrating natural available prehistoric to sources hunter-gatherers

loutsa Figure3.10. The eponymous on the raisedpeneplainsouth of the lowerAcheronvalley(Fig. 3.8:22). Althoughcurrently being dissected headward erosionfromsmall by streamson its northernedge, Loutsa still pondswaterin winterand vegetationand spring,attracting wildlife.

Figure3.11. The poljeof Cheimadio, at an earlystageof incisionand drainage streamsat eachend by

64

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure 3.12. Red sediments and paleosols: a) redeposited terra rossa in a small raised polje at Kranea; b) mature paleosol (Bt horizon) in the raised polje of Kokkinopilos; c) banded terra rossa, alternating between reduced (pale) and oxidized (red), near the base of the redeposited terra rossa in the loutsa of Ayia; d) red oxidized terra rossa overlying (with a sharp boundary) yellow reduced terra rossa redeposited in lacustrine conditions in the raised polje of Kokkinopilos

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

65

66

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

deposit

terra rossa primary

I I

slump/landslide
terrarossahas a characteristic bimodalgrain-sizedistriRedeposited butionwith a sharpupperlimit at ca. 0.04 mm (Fig. 3.14) and,exceptfor than0.064 somelimonitemicronodules (Table3.2), it lacksparticles larger for mm.Together with poordisaggregation duringthe preparation grainthe 20%sandnotedbyDakaris sizeanalysis, thesenodules explain probably
and his colleagues.62

Figure3.13. Terrarossamaybe in that redeposited fan complexes consistof colluvium,debrisflows, or slumps,andlandslides, on alluvial fansdepositedby smallephemeral streams.Becausefan formationis intermittent, paleosolsmayformon stablesurfaces. temporarily

Terra or two rossa, modes, displays sizefrequency primary redeposited, a silt modebetween0.010 and0.040 mm makingup 5-30%of the whole and a clay mode below 0.002 mm rangingfrom 50%to more than 90% (Fig. 3.14;Table3.3). In a largenumberof casesthe distinctionis sharp, butthe 0.002 to 0.010 fraction variable mayobscure fine limbof is and the the silt mode. the Mineralogically, claymodevariessomewhatfromone site to anbut is consistentwithin sites (Tables3.4 and 3.5). The mode conother, sists mainlyof illite,exceptat Kokkinopilos wherekaoliniteandillite occurin roughly Smectite(chlorite vermiculite) a minor and is equalamounts. andquartz rare. is The claymineral wellwith those accessory spectra agree of Yassoglou of Barbaroux Bousquet, and and who concluded that the influenceof Mesozoic and Cenozoic sourcerockswas limited and attributed the variations in mainlyto differences the weathering historyof the In sourcedeposits.63 contrast, composition the silt modeis uniform the of to the point of monotony; consistsalmostentirelyof quartz it with rarely a little feldspar, latterexhibitinga variable the ratio orthoclase/plagioclase An exceptionis the Rodakistonyred deposit,which is (Tables3.4, 3.5). and, as will be discussedbelow,should not be classifiedas feldspar-rich terrarossa. The bimodal of distribution redeposited terra rossa first was grain-size noted by Tippett and Hey,who suggesteda long-distanceaeolianorigin

62. Dakaris, Higgs,andHey 1964, fig. 15. 63.Yassoglou, and Kosmas, Moustakas and 1997;Barbaroux 1976,fig.3. Bousquet

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

67

TABLE 3.2. COMPOSITION OF THE FRACTION TERRA ROSSA >0.064 MM IN REDEPOSITED

Sample VA93-01 VA93-05 VA93-02 VA93-04 VA93-03 VA93-06 VA93-07 VA93-08 VA93-09 VA93-11 VA93-12 VA93-13 VA93-16 Description sediment Poorly disaggregated Poorly disaggregatedsediment Limonite concretionsto 3 mm and poorly disaggregatedsediment Poorly disaggregatedsediment Small (0.5-3 mm) limonite nodules and poorly disaggregatedsediment As above As above Limonite nodules,2-10 mm in diameter Limonite nodules,2-6 mm in diameter Limonite nodules,2-6 mm in diameterand heavily Mn-stained, plus poorly disaggregatedsediment As above sediment Poorly disaggregated As above Remarks Paleosol Bt Paleosol Bt

Paleosol Bt Paleosol Bt Paleosol Bt

20-, .r?" -, A

VA93-04

Kokkinopilos
Go CD

VA94-19 Galatas

F F

20-

VA93-08

Kokkinopilos

VA94-23

Galatas

10I! 0
._

........

a
-

20- o

VA94-04

Loutsa

VA94-14 Galatas

10 -

Figure3.14. Typicalgrain-size of diagrams terrarossa frequency in poljesandloutses. redeposited Materialcoarser than silt size is veryrare.Percentage (>0.064mm) of clay(<0.002mm) is shownon the area left. Approximate of the silt mode is shaded.

20- Cs (D

a,"P

VA94-12 Alonaki
L0

VA94-07

Cheimadio

. . ..... ...............:...................

ioj,,''''',"'"'

,F

.
I

.......... I

10 t % 2-4 microns

100

10

100

size (log microns)

68

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

TABLE 3.3. GRAIN-SIZE DISTRIBUTION TERRA ROSSA REDEPOSITED

Sample Silt Size (mm) %

OF

Clay(<0.002 mm) Silt/Clay Depth below % Ratio Surface(m) 62 90 73 0.274 -0.50

ALONAKI (#21: subsided loutsa?) 17 0.010-0.035 VA94-12

AYIA-I

VA94-11 VA94-26
AYIA-2

(#1: loutsa) 0.008-0.035 0.010-0.040

(#1: loutsa)

5 10

0.055 0.137

-0.50 -0.80 -1.20 -2.50 -4.00 -7.00

VA94-27 VA94-29 VA94-30 VA94-31

CHEIMADIO

0.010-0.040 0.008-0.035 0.008-0.035

14 27

74 68 51 96

0.122 0.206 0.529

(#24:

polje)

VA94-07
GALATAS

0.010-0.030

14

65

0.215

-0.50

VA94-14 VA94-15 VA94-16 VA94-16 VA94-17 VA94-18 VA94-19 VA94-20 VA94-21

(#27, south section: polje) 0.009-0.035 12 0.010-0.035 9 0.008-0.040 16 0.010-0.040 11 0.010-0.040 13 0.009-0.040 13 0.009-0.040 15 0.008-0.035 10 0.010-0.035 18 polje) 5 13 14 12 9 11 9 11 8 13 19 14 23 22 15 13 11 11

72 74 65 71 67 73 68 73 65

0.167 0.122 0.246 0.155 0.194 0.178 0.220 0.137 0.276

-0.15 -1.00
-1.90

-1.90 -2.85 -3.20 -3.60 -4.70 -5.30

GALATAS (#27, west section: VA94-22 0.010-0.040 VA94-23 0.009-0.040 VA94-24 0.006-0.030
KOKKINOPILOS (#30: polje)

79 67 74

0.063 0.194 0.189

-0.60 -1.10 -1.60

VA93-01 VA93-05 VA93-02 VA93-03 VA93-04 VA93-06 VA93-07 VA93-08 VA93-09 VA93-11 VA93-12 VA93-13 VA93-16

0.010-0.035 0.010-0.035 0.010-0.035 0.008-0.035 0.008-0.040 0.009-0.040 0.010-0.035 0.008-0.040 0.010-0.035 0.010-0.035 0.010-0.040 0.008-0.035 0.009-0.040

72 78 72 73 78 65 64 56 65 57 60 68 73 70 73

0.167 0.115 0.153 0.123 0.141 0.123 0.203 0.339 0.215 0.403 0.367 0.220 0.178 0.157 0.151

-0.50 -0.50 -2.50 -2.50 -2.50 -4.00 -6.00 -7.00 -10.50 -14.00 -17.00 -22.00 -30.00 surface -0.50

LOUTSA (#22: loutsa) VA94-04 0.010-0.040 VA94-36 0.009-0.030

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

69

TABLE 3.3-Continued
Sample
MORPHI (#16,

Silt Size (mm)

section

% 17 7 17 18 13 14

Clay(<0.002 mm) Silt/Clay Depth below % Ratio Surface(m) 52 73 51 50 65 62 0.327 0.096 0.333 0.360 0.200 0.226 -3.00 -4.00 -11.50 -16.80 -17.00 -19.50

1: polje)

PP1-19 PP1-17 PP1-15 PP1-02 PP1-03 PP1-01

RODAKI

0.008-0.040 0.010-0.030 0.008-0.030 0.008-0.040 0.008-0.040 0.009-0.040

(coastal marsh)

VA94-06 VA94-35

0.010-0.045 0.010-0.040

12 15

66 59

0.182 0.254

surface surface

Numbers in parenthesesreferto locations in Table 3.1 and Figure 3.8. Grain-size analysisof fraction <0.064 mm (64 m) with a MicromeriticsSedigraph 5000ET and 5100 V3.07 with computerinterface(Jones, McCave, and Patel with 0.2% Calgon. Fraction>0.064 mm (not always 1988) after disaggregation fine present)consists of limonite micronodulesand poorly disaggregated sediment, except for samplesVA94-06 and VA94-35, which contain detritalsand and gravel.

for the silt of the Kokkinopilos redbedsas did MacLeod and Yaalon.64 streamgravelsat the peripheryof the KokOthers,citing interbedded kinopilos deposit, which we have been unable to confirm, regarded
it as an alluvial fan complex.65Following MacLeod's brief but percep-

tive study,we considerthe high sortingand fine gradeof the silt mode as conclusive evidence for a long-distance windblown origin of this componentof the sediments.During glacial conditions,strong southwesterlyto southerlywinds duringwinter and spring and easterliesin The the summerwere probablyat least as common as they are today.66 sortingof modernNorth Africandust, collectedat many sites in Crete and elsewhere,resemblesclosely our own and a North African source
seems possible, but the Epirus dust is slightly coarser and may instead belong to the attenuated southern fringe of the central European loess
belt.67

64. Dakaris,Higgs, and Hey 1964; MacLeod 1980; Yaalon1987. 65. Higgs and Vita-Finzi 1966; Harris and Vita-Finzi 1968; MacLeod and Vita-Finzi 1982. 66. Rossignol-Strick1983. 67. Kubilayet al. 1997; Pye 1992; Yaalon1997.

Whatever the sources of terra rossa silt may have been, its distant origin and complete bleaching during aerial transport make the material suitable for luminescence dating. Moreover, aeolian dust fall is likely to have been more constant than the local flux of weathering residue. The variations in the ratio of the two components might thus be useful to correct bulk sedimentation rates. Uniform as polje and loutsa sediments are, stratigraphic sections of raised and dissected poljes suggest depositional histories that differ from place to place. Best surveyed is Kokkinopilos (Fig. 3.15), studied and restudied by the Higgs and Bailey teams and by ourselves in collaboration with Panayiotis Paschos from the Institute of Geology and Mineral Exploration (Preveza branch).

7?

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

TABLE 3.4. MINERAL COMPOSITION OF REDEPOSITED TERRA ROSSA AT KOKKINOPILOS

kaol Clay(<0.002 mm) ill sm + ve
Silt (0.010-0.064 mm)

ve/ch

feld

qtz

qtz abund

Sample
VA93-01 VA93-02 VA93-03 VA93-04 VA93-05 VA93-06 VA93-07 VA93-08 VA93-09 VA93-11 VA93-12 VA93-13 VA93-16

(%)
38 48 46 53 48 43 46 45 48 47 42 40 40

(%)
47 31 52 40 52 49 50 41 39 48 50 60 46

(%)
15 21 2 7 8 4 15 13 5 7 14

ratio
xx xxxx xxxx xxxx xx xxx xxx xx xx xx ? x xx

(%)
5 -

(%)
95 100 100 100 100 100 100 100 100 100 100 100 100

(mm)
65 30 34 80 32 47 85 80 59 46 60 47 7

kaol = kaolinite;ill = illite; sm + ve = smectite and vermiculite; ve/ch ratio = vermiculite/chlorite from low (x) to high (xxxx), based on higher of two ratio, 14.2 A peaks in XRD tracerelativeto 10A peak;feld = feldspar,orthoclaseand plagioclasecombined;qtz = quartz;qtz abund = quartzabundanceas peak height of 4.23 A quartzline.

The Kokkinopilos fromthe Lourosvalleyby a limepoljeis separated stone faultblock.A paleosol,now strippeddownto a matureBt horizon, is preserved alongits edgesandspottilyin the slightlybowl-shaped center, whereit formsthe foundationof the vents for a subterranean sectionof the Romanaqueduct leadingto Nikopolis.68 Locallywithin the paleosol
numerous flint artifacts occur in situ.69

is whichare Kokkinopilos beingerodedrapidly ephemeral streams, by for its spectacular badlandtopography haveexposeddeep and responsible sections the throughout sediment body(Figs.3.16,3.17).A mature paleosol which locallypreservesfaint, thin inter-bedding tops a red zone (C),70 with subhorizontal, bleached In subvertical gray, layers. places, gray, stripes, root probably channelsof the existingpinewoodland,areseen.A thin but desiccation zone andan immature conspicuous paleosol,indicatinga brief hiatusin deposition,separate fromthe moreyellow(5YR-7.5YR)zone it B. Both layerscan be tracedthroughout centralarea. the In zone B, careful inlaminations fine,subhorizontal cleaningreveals almosteverywhere, diffusegraygley and dicatingsubaqueous deposition zones markfluctuations groundwater duringandafterdeposition. in level The depositiontook place mainlyunderwater,but two moderately maindicatebreaks the depositionanddrysurfaces in ture,truncated paleosols thousand Both paleosols, locatedat 10 and 14 m exposedfor several years. below the top of the sequence,areassociated with thin (10-30 cm), discontinuousgravel lenses rich in small flint fragments,many of them Palaeolithic artifacts.

68. For furtherdiscussionof paleosols,see below.The well-preserved bridgesfor the Roman aqueduct(see Fig. 1.7) lie just east of Kokkinopilos. 69. Runnels and van Andel 1993b. 70. Dakaris,Higgs, and Hey 1964.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

71

OF TABLE 3.5. MINERAL COMPOSITION TERRA ROSSA FROM POLJES AND REDEPOSITED LOUTSES IN WESTERN EPIRUS
mm) Clay(<0.002

Silt (0.010-0.064 mm)

Sample
ALONAKI

kaol (%)

ill (%)

smec (%)

qtz

qtz (%) 93 93

feld (%) 7 7

or/pl ratio 1.1 1.1

VA94-12 VA94-32 VA94-11 VA94-26 VA94-27 VA94-29 VA94-31 VA94-15 VA94-17 VA94-19 VA94-21 VA94-04 VA94-36 VA94-06 VA94-35

(#21: subsided loutsa?) 7 66 28 66 6 28 19 24 22 20 25 12 14 25 32 37 63 67 62 67 74 68 54 56 49 60 18 9 16 6 5 20 32 19 19 3

AYIA (#1: loutsa)

+ + -

95 95 95 100

5 5 5 -

0.3 0.6 -

GALATAS (#27: polje)

+ + +

100

LOUTSA (#22: loutsa)

93 65 58

7 35 42

1.7 2.8 2.2

RODAKI (coastal marsh)

22

73

Numbersin parenthesesreferto locations in Table 3.1 and Figure 3.8. kaol = kaolinite;ill = illite; smec = smectite (chlorite and vermiculite);qtz = quartz (+ = trace);feld = feldspar,orthoclaseand plagioclasecombined;or/pl = orthoclase/plagioclaseratio (height of 3.19 A peak over height of 3.24 A peak).

Figure3.15. The raisedpoljeof Kokkinopilos

72

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure3.16. Badlanderosionat Kokkinopilos

Figure 3.17 (opposite).Cross section

Zone B overlies the deep red (10R-2.5YR) zone A, again with signs of desiccation and, here and there, minor erosion at the boundary.The zone itself is in part mottled with gray and rests on a karst surface.Just upstream from Tsiropolis, the basal portion is interbedded with layers of very fine, white sand, up to 30 cm thick, that consist of clear,well-formed calcite crystalswith a few percent quartz,indicating precipitation as a playa evaporite combined with dust fall. Notwithstanding the evidence for breaks in deposition and for many alternations between wet and dry conditions, the uniform grain-size distribution convincingly argues for the same restricted source of sediments throughout. The raised, dissected polje at Morphi (Fig. 3.18) resembles that at Kokkinopilos. The sequence begins with a thin modern soil resting on a truncated mature paleosol that grades into a yellow zone similar to zone B at Kokkinopilos.71 The sequence contains a few thin, distal debris flows and four moderately mature truncated Bt horizons that mark hiatuses of significant duration. The yellow zone rests on a thick (ca. 2 m) tephra deposit derived from an eruption in centralItaly and dated by Ar/Ar methods to 374 + 7 kyr B.P.Underneath, several meters of leached, gray (7.5YR 6/8) polje deposits rest on yellowish red (5YR) silty clays that are separated from the underlying karst surface by a very mature truncated Bt horizon.

throughthe incisedpoljedepositsof Left:Grain-size Kokkinopilos. for frequency diagrams the size range 0.000-0.070 mm (silt mode shaded); the percentage clay(<0.002mm) is of indicatedin the upperleft cornerof eachdiagram. Center:Silt/clayratio variation with time;datafromTable 3.3. Right:Geologicalsectionbased on RunnelsandvanAndel 1993b, fig. 6; depthsin metersabovesea level (masl)by altimetersurvey. Zone labelsfromTippett (in Dakaris, Higgs, and Hey 1964, pp. 221-225).

71. Pyle et al. 1998.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

7 73

%/Phiunit

paleosol
f.0

-140
(0

paleosol -paleosol paleosoll -:130 fine flint gravel

"handaxe"

I'a

Nc%

c~o

(0

I0

Nz >

-120 'paleosol m -pale gray silt 11paleosol

Ico

lw0 0

- 110 masl karst surface

10

100

log microns

74

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure3.18. Morphipoljeoutcrop with paleosolsforminghard,

protruding benches

AYIAcomposite
339 -

profile

Bt (M2)
*) *? **

- VA94-27 - VA94-28

mottled terra rossa Bt(M5) gravel (debris

flow)

55 5 5 - VA94-29 55 51 $5

mottled

terra rossa

335-

- VA94-30 stratified
-

(lowest

stone

mottled terra rossa tools)

- VA94-31

massive

dark red terra rossa

Bt (>M5?) 330 masi limestone karst

Figure3.19. Compositeprofileof Ayialoutsashowingthe lithological sequenceandpaleosols.Palaeolithic

stone tools occur throughout the section from ca. 333 to 339 masl. Paleosol maturity codes (MS 2, MS 5) from Table 3.8.

Occasional bands of fine to medium stream gravel were laid down by small ephemeral streams during a brief period of flooding of the polje floor or as thin debris flows produced by catastrophic failure of the slope mantle or a fan (Fig. 3.18). These bands testify to brief invasions of a highenergy regime, probably during times of exceptionally high rainfall because they are too thin and sparse to indicate major climate changes of stadial/interstadialor glacial/interglacial rank.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

75

lower Stratified Figure3.20. (above) sectionof the Ayialoutsalooking west. Mousterianartifacts are embeddedin the exposedsection wherethe figure poin ting; (below) is detailof Mousterian artifacts the in situ. The history of the deeply dissected loutsa at Ayia is simpler (Fig. 3.19). A rugged karst surfacewith a relief of 50-200 cm locally retains in depressions truncated, very mature Bt horizon as at Morphi. Bt the sionsa truncated, mature horizon at Morphi. Elsewhere the as Elsewhere very oldest deposit is a pale (5YR 6/8), finely crystalline dolomite sand analogous to the basal calcite sand at Kokkinopilos. The loutsa fill itself is a well-bedded, red and gray mottled deposit (Fig. 3.20), interbedded with layers bearing stone tools and near the top a debris flow. A little higher, a mature, truncated paleosol is overlain by a modern soil. In summary,even without its telltale red color, redeposited terra rossa unfailingly discloses its origin by its bimodal grain-size distribution. Temporary or permanent wet conditions of deposition are indicated by soil features (mottling, gley). Dry conditions are revealed by color-banding, desiccation zones, and thin beds anomalously rich in fine quartz that represent periods when mainly dust was being deposited. The fine mmscale horizontal stratification seen associated with artifact scatters at

76

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

| mcoas:t.al;pa-:n_. coastal

plain

^^HE^y

^^::^

v^^^RFigure

Fi last glacialmaximum, the 20-18 kyr

B.P., when sea level was over 100 m

3.21. The AdriaticSea during

,-.

present coast

'

lonian
J____________________________________________________________________

Sea

lowerthan today.Rivercourses acrossthe emergedcoastalplainare

extrapolations. AftervanAndeland
Shackleton1982, fig. 4

Kokkinopilos and Ayia, and common elsewhere, is a result of deposition in very low energy conditions, far too low to entrain even the smallest flint debitage. As the fine stratificationis not easily seen, failure to observe it in the past has led to faulty stratigraphic interpretations, such as the view that at Kokkinopilos, and by implication in other redbed sequences, the artifact assemblages are on secondary location. Occasionally, artifacts are associated with debris flows or small-scale stream gravels, but those are rarebrief incidents in the history of poljes. Ultimately, the relevance of our knowledge of the genesis and history of Epirote poljes to our understanding of its Palaeolithic inhabitants depends on our ability to fit a time dimension to them. The debate about the age of Kokkinopilos, so far the only polje viewed with an age perspective, still includes those who regard all redbeds as Pliocene in age (except for portions reworkedby recent erosion) and others who see them as belonging to the Late Quaternary.We shall return to this subject below.
SEA LEVELS AND COASTAL PLAINS

All but the narrowest Mediterranean shelves are the flat surfaces of sediment wedges, which, when exposed at lowered sea level, may form wide coastal plains (Fig. 3.21). Often well watered and bordering today's rugged coastlines over long distances, they offered major wildlife resources and convenient migration paths for early humans.72

72.vanAndel1989;vanAndeland
Shackleton1982.

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

77

E I X, w

50-

100-

20

80 AGE (kyr)
LATE QUATERNARY PALEOSHORELINES

Figure 3.22. Global sea-level variations for the past 140,000 years, reflected by two oxygen isotope records based on bottom-dwelling deep-sea foraminifera (Shackleton 1987; Labeyrie, Duplessy, and Blanc 1987) and calibrated with raised coral reef data. Dots: U/Th dates on corals (Bard, Hamelin, and Fairbanks 1990; Stirling et al. 1995); lozenges: recent U/Th dates from Huon Peninsula coral terraces in New Guinea (Chappell et al. 1996). Numbers at the top indicate oxygen isotope stages.

MIDDLE

AND

Traditonal names,suchastheWirm orWeichsel Quaternary stratigraphic have no meaningin the Mediterranean. we Therefore, use here glacials, based the globalchronostratigraphy on oceanicoxygenisotopestages(OIS) of ImbrieandMartinson.73 During the last 140,000years,the intervalof interestto us,globalsealevelwastwiceat a low glacialstand(ca.-120 m in OIS 6 and2) andtwiceat an interglacial levelslightlyabove(OIS 5e) or at at value.It remained eitherextreme onlyfiveto ten for (OIS 1) its present but intermediate levelsforroughly100,000years, millennia, occupied from the climaticdeclinefollowingthe last interglacial (OIS 5d-e) throughout most of the subsequent pleniglacial. Globalglacialandinterglacial sea-level can from positions be estimated
oxygen isotope ratios (180/160) of bottom-dwelling microfossilsthat record

the volumeof seawater storedin ice caps.To obtaina truepictureof sea

level against time, the 180/160 curve must be calibratedwith past sea-level

73. Imbrie et al. 1984; Martinson et al. 1987. 74. van Andel, Zangger,and Perissoratis1990.

or positionsdeducedfromraisedreefsandcoastalterraces fromshorefeaturessubmerged continental on shelves(Fig. 3.22). In the absenceof offshoreseismicreflection datafor the Epirusshelf,pastshorescanbe determined only by appropriate contours(Table3.6), but presentbathymetric Late Quaternary sedimentstend to be thin on Greekshelvesanderrors of of positionarewithin the limits of precision the isobaths.74 the Overthe past140,000years, widthandareaof the emerged coastal plainin Epirushavevarieda greatdeal(Fig. 3.23;Table3.6). Exceptdura ing the two briefglacialmaxima, total of some 20,000 years,the coastal If continuous,was narrow. the resourcepotentialof an plain, although to environmental zone is assumed be roughlyequalto its area,most of the time the coastalplainswereat best equalin potentialto the combinedarea of all poljes.

78

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

PALEOSHORELINE TABLE 3.6. APPROXIMATE DEPTHS AND COASTAL PLAIN WIDTHS, 140 KYR B.P. TO PRESENT
Interval OIS 6 5e 5d-a 4 3 2 2 1/2 1 1 (kyrB..) >135 130-117 117-74 74-59 59-24 24-20 18-15 15-8 11 9 Event Glacial maximum Shoreline depth(m) Coastalplain width (km) 10-20 0 1-4 5-15 5-7 10-20

-130 0 to +10 Interglacialpeak -20 Run-up to glacial First glacial maximum -80 to -90 Mild phase -60 to -70 -120 Main glacial maximum -110 to -100 Earlydeglaciation -90 to -20 Main deglaciation -40 Mesolithic starts Mesolithic ends -20

1-5 1-2

OIS = Oxygen isotope stage. Shore depth from Figures3.22 and 3.24. Coastal plain width is a representative range,indicatingdistancebeyond the present shoreline.(Note: at times the plain between Corfu and the mainlandwas wider.) considerably

Whenever the sea stood above -80 m, the shelf between Corfu and the mainland was largely flooded. Given a present least depth between -45 and -50 m, however, the two were joined by a land bridge during all of OIS 6 and from 90 kyr B.P. to 10 kyr B.P. This persistent connection between Corfu and the mainland may have been a key point in strategies for hunting migrating herds of large herbivores. In the Ambracian Gulf, which has a shallow sill, the -20 and -50 m isobaths show that between 10 and 105 kyr B.P. it was occupied by a lake. The glacial sediment load of the Louros and Arachthos Rivers, the only major sediment-carrying rivers in the area,was dumped there in the form of a delta complex very similar to the present one.75 Because of the rapidity of climate change and sea-level rise during the decline of the last glacial maximum, and its importance for the latest Palaeolithic and Mesolithic occupation in western Epirus, we need a more precise sea-level curve for that interval. This requires compensation for glacio- and hydro-isostatic effects, for which we may use Lambeck's corrected curve for Kavallabecause that area is at the same distance from the northern European ice edge as Epirus.76 The corrected curve (Fig. 3.24) shows that the sea began to rise slowly some 18,000 years ago, accelerated rapidly around 14 kyr B.P. and continued through the Mesolithic to reach about -10 m 6,000 years ago. Lambecks isostatically compensated curve reads time in radiocarbon years. If we convert the deglaciation sea-level history to calendaryears by using the U/Th-dated curve of Bard, the deglaciation rise begins earlier, the Mesolithic shorelines are shallower (-30 m at the start and -15 m at the end of the period), and the coastal plain is proportionally narrower.77

75. Piper,Kontopoulos,and Panagos 1988. 76. Lambeck 1995, fig. 6:e; 1996. 77. Bard,Hamelin, and Fairbanks 1990.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

79

The Figure3.23 (above). emerged coastalplainoff Epirusat six key of momentsbetweenthe maximum the OIS 6 glacialto the present to (OIS 1), according interglacial Table3.6 and Figure3.22. Isobaths are representing paleoshores based on nauticalandtopographic charts and arehighlygeneralized. The Mesolithicshorecorresponds with the Mesolithicintervalin calendar years(Fig. 3.24). Two sea-levelrise Figure3.24 (right). curvesfor the deglaciation intervalof late OIS 2. Bard,Hamelin,and Fairbanks 1990,in calendar years,is
based on U/Th dates of submerged

' 0 present sea level _ /

E
E/ )

Bard et a. (1990)
CZ-500, c) _-'

/
/ / Lambeck (1995)

0o a)-100
_ /

/ / Mesolithic (cal BP) --

coralterraces Barbados; in Lambeck1995, in radiocarbon years

, (Fairbanks 1989), is based on the' same samples and has been used to . . , , , ,I 20,000
I

' 5000

15,000

10,000

datethe isostatically compensated localsea-levelhistory.

Age in years before present

80
THE COAST

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

OF EPIRUS

IN THE

LAST

INTERGLACIAL

(OIS

5)

For all but one of the periods in the interval OIS 1-6, the associated paleoshores are now below sea level and, in the absence of high-resolution seismic reflection studies, their nature and true position cannot be known The exception is the 10,000-year long peak of the last interglaaccurately. cial (OIS 5e), with observed sea-level positions that in stable areas range from 0 to +10 masl,78although most of those elevations do not differ significantly from the present level after a correction for glacio- and hydroisostatic effects has been applied.79 In Greece the paleoshores of the last (Tyrrhenian)interglacialhave in many places been raised above their original levels by coastal tectonics. They are marked by a distinctive warm fauna of corals and large robust mollusks, including the warmwatergastropod Strombus.80 In coastal Epirus raised shore deposits are in evidence at severalpoints (Fig. 3.25). They have generally been regardedas Late Pleistocene or earliest Holocene in age, notwithstanding the high uplift rates that the low sea levels of that intervalwould imply.A large outcrop exposed at Anavatis, on the other hand, has been mapped as Pliocene on the basis of shallowwater agglutinating foraminiferaof little stratigraphicvalue.81 The Anavatis complex, now at approximately 40 masl, is exposed in the south wall of a sand and gravel pit where it consists of thick, massive to thin-bedded layersof unconsolidated,white to pale yellow, fine, well-sorted sand (Fig. 3.26:1, 3). The sands are interbedded with thin (2-20 cm) layers of gray-black, finely (1-10 mm) laminated silt deposits with the characteristicgrain-size distributionsof marshor tidal flat silty clays (Fig. 3.26:2, 8). Locally, sand-filled channels are cut into underlying beds. Some of the moderately calcareous sand beds contain abundant coastal marine mollusks such as Cerastoderma, while shell debris is common in burrowsin the laminated silts. There are also a few lenses of rounded, well-sorted fine (15 cm) gravel. There can be no doubt that this is a coastal or very shallow marine deposit. In the opposite north wall of the pit, a thick series of unconsolidated medium-coarse sands and fine-coarse gravelsis exposed, traversedby many faults of small displacement. The coarse strata are interbedded with thick (5-20 cm) lenses of fine sand or grayish green silty clay,perhaps formed in pools on a braided, low-angle fan. Grain size (Fig. 3.26:4, 5) and chaotic bedding point to braided or torrentialstreams.At ca. 48 masl the sequence is topped by a paleosol. The maturity level of this paleosol (MS 4/5)82and extensive frost-shattering of the finer gravels indicate deposition during the cold Late Pleistocene pleniglacial (OIS 4-3). The paleosol contains an early Mousterian industry (see below). The torrentialunit, although apparentlydeposited in a low-lying area, is too close to the coastal unit of the opposite scarp to be contemporaneous. Moreover, if the coastal sediments were late glacial in age, the low sea level of the time would require an uplift rate of some 4 m/kyr, quite in excess of other tectonic rates in the region (as discussed above).More probably, they underlie the torrential unit and so are of interglacial age. Sea

78. Bardet al. 1993;Chenet al.

1991; Edwardset al. 1987; Ku, Ivanovich,and Luo 1990; Stirling et al. 1995. 79. Lambeckand Nakada 1992. 80. E.g., Kelletat 1974; Kelletatet al. 1976; Keraudren Sorel 1987; and Schr6derand Kelletat1976. 81. Etudegeologique. 82. For a discussionof paleosol maturitystages (MS), see below, pp. 86-89.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

8i

Figure3.25. Locationsof raised paleoshore depositsof the last (OIS 5e and OIS 5c) in interglacial coastalEpirus

Cumulative Figure3.26 (below). of distributions coastal grain-size the last interglacial and sedimentsof earlyHolocenefromthe Anavatis sandpit (1-5, 8) andAlonakiBeach (6, 7): 1) VA94-2,shallowmarine or dunesand;2) VA94-3,silty laminatedmarshclay;3) VA94-37, shallowmarineor dune sand; streamsand; 4) VA94-39a,torrential 5) VA94-39b,same;6) VA95-2, Holocenedunesand;7) VA95-3, same;8) VA94-38,marshor tidal siltyclay.Sedigraph analysesof the fraction4.000-0.002 mm. 100

--

0
Q _

LU D 0-

50-

.5

.25

.125 .062

.031

.016 .008 .004

.002

.001

GRAINSIZE (mm)

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure3.27. The raisedTyrrhenian visibleas a low, beachatTsarlambas, rockycoastaldepositon the right level for that time ranged from about +10 to -20 m, yielding a reasonable uplift rate of 0.4-0.6 m/kyr for the Anavatis area, as luminescence dates confirm.83 Other paleoshore deposits occur at severalpoints along the coast west of Preveza. At Alonaki Beach and Tsarlambas(Fig. 3.27), a few meters of well-consolidated, horizontally bedded, low-angle cross-bedded series of fine to medium, well-sorted sands topped by a truncated very mature paleosol (MS 5) are exposed at the base of the coastal cliffs. The sand is moderately calcareous and contains many small, thick-shelled gastropods and a few corals.We regardthese bench-forming deposits, now located ca. 3 masl, as a Tyrrhenian paleoshore which can be traced intermittently as far as the cape at Mytikas. Thick-shelled gastropod fragments, probably Strombus,also occur in boreholes in a down-faulted sequence at the entrance to the Pantokrator suburb of Preveza.84 A similar sand complex crops out east of Preveza at Ayios Thomas on the flank of a coastal hill. Topped by a red (10R 6/6) mature paleosol, the well-sorted, subrounded medium sands with lenses of rounded gravel may represent a raised Tyrrhenian coastal fan. At Rodaki, south of the mouth of the Paliourias River,a raised coastal complex is exposed consisting of weakly consolidated, low-angle, crossbedded, fine, pale yellow (10YR 8/2) sand with thin layers of coarse sand and stringers of small pebbles. It is topped by a dark red (2.5YR 4/4-3/4), mature (MS 5) truncated Bt horizon. Because the complex is located 8-20 masl, we regard it as another Tyrrhenian beach and coastal dune deposit. Like all other coastal deposits, and in stark contrast to the terrarossa, the sand contains abundant feldspar (Table 3.7). Nearby is an important site of consolidated red, thin-bedded sand and gravel containing a Middle Palaeolithic industry (see below), but because of complex active faulting and poor outcrop conditions its relation to the assumed interglacial shore deposits is unclear. Similar coastal deposits with characteristic Tyrrhenian fauna occur at 30 and 10-12 masl on
Corfu.85

lumines83.An infrared stimulated 23 cence(IRSL)dateof ca. 128?+ kyr B.P. withinthe main the places deposit (OIS phaseof the lastinterglacial 5dIRSLdateof ca. 188? 30 e). Another for kyr B.P. is questionable technical it the reasons because places and in interbedded marsh deposit the OIS whensealevelwas 6 glacial maximum low.SeebelowandTable3.10. 84. P.Paschos (pers. comm.). 85. Sordinas 1983,p. 343, table1.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

83

TABLE 3.7. MINERAL COMPOSITION OF MODERN AND LAST INTERGLACIAL COASTAL SANDS IN WESTERN EPIRUS
Sample
ALONAKI

Quartz (%)
BEACH

Feldspar (%)
dunes)

Or/PIRatio

Remarks

(Holocene

VA95-02 VA95-03

53 93

47 7

3.2 1.2

fair amount of calcite

ANAVATIS (coastal deposits) VA94-02 57 43 44 VA94-37 56 RODAKI (coastal deposits) VA94-05 54 VA94-06 65 VA94-35 58

2.6 3.0

46 35 42

2.1 2.8 2.2

fair amount of calcite

Or/Pl Ratio = orthoclase/plagioclase ratio.

At Ormos Odysseos, on the south side of the Acheron River valley, the remains of a thin alluvialfan sequence, its top at 4 masl, occur on a low, north-dipping karst surface. It consists of strongly consolidated, dark red (2.5YR 3/4), coarsesand and red clay,overlainby a definitely mature,Pleistocene coastal dune paleosol, now a little above present sea level and of last interglacial age. Associated Middle Palaeolithic findspots are discussed below.

VEGETATION

HISTORY

AND

CLIMATE,

140-10

KYR B.P.

Our understanding of the Late Quaternaryvegetation and climate history of western Epirus rests mainly on long cores from Lake Ioannina, first studied by Bottema and more recentlybyTzedakis (Fig. 3.28).86The cores, supplemented with other data from Greece, Italy, and the Balkans, reasonably reflect the long-term climatic history of northern Greece, but afford little insight regarding the diverse local conditions of the mountainous terrain of western Epirus with its largely orographic climate conditions.87Cores in lowland and highland lakes are beginning to provide some detail for local areas and for a range of elevations, but because they cover only later phases of the deglaciation period and the Holocene, the results are not directly applicable to the long interval from 60 to
25 kyr B.P.88 86. Bottema 1974, 1994;Tzedakis 1993. 87. Willis 1994; Culibergand Sercelj1996; van Andel andTzedakis 1996. 88. Willis 1992, 1994;Turnerand Sanchez-Gofii 1997. 89. Smit and Wijmstra 1970. 90. Bennett, Tzedakis,andWillis 1991;Tzedakis 1993.

During the penultimate glacial of OIS 6, a discontinuous steppe vegetation of sagebrush (Artemisia),chenopod species (indicative of aridity), and grasses predominated in southern Europe. Cold-stage pollen from and Tenaghi Philippon in Macedonia contained Eurotia ceratoides Kochia species found today in the central Asian steppe and semidesert laniflora, that point to a cold, arid climate.89 sheltered spots of the western Balkans In and mountains of Italy,however, scattered temperate tree populations survived in refugia where temperaturevariations were not extreme and precipitation was sufficient, thus enabling a swift returnof the woodland when the climate improved.90

84

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

LATE

DEGLACIATION

(I2-IO

KYR B.P.)

Climatewarming, moister;oakwoods mixedwith warmthloving speciesdevelop;coastalregionshave open Mediterraneanwoodland of pine, evergreenoak, wild olive, and pistachio.
L- A' LAST GLACIAL MAXIMUM (OIS 2)

Dry, cold climate;sagebrushand chenopod steppe, open deciduous woodland on south-facing slopes or middle rains. elevations benefitingfromorographic 3) woodMilder,moisterclimate;steppewith open deciduous landin favorable places.
MID-GLACIAL MILD INTERVAL

(OIS

c 40.

(OIS 4) treesconCold,dryclimate; steppegainson openwoodland; \tract into refugiatowardfinalphase.

FIRST GLACIAL EXPANSION TRANSITION TO GLACIAL (OIS 5D-A)

Cool and warm periods alternatebetween Mediterranean and woodland cold,drysteppe. and mixedevergreen deciduous
LAST INTERGLACIAL

(OIS

5E)

oak/elmforest than now;deciduous Climatea little warmer wild olive and evergreenoak followed by Mediterranean maximum summerinsolation. woodland;
PENULTIMATE

(OIS 6) and sagebrush Cold, aridclimate;chenopod steppe;refuigia for temperate trees.

GLACIAL

80-

0
Tree pollen (%)

100 Figure3.28. Climateandvegetation changesduringthe last two glacialcycles(OIS 1 through interglacial OIS 6), illustrated the variationof by the arboreal on pollensum. Based
Tzedakis 1993, 1994

When the OIS 5 interglacial began, trees spread outward from the refugia in a vegetation succession beginning with deciduous oak (Quercus) and elm (Ulmus), followed in southern Europe by a major expansion of Mediterranean forest characterized by evergreen oak and values of wild olive (Olea) even higher than in the Holocene.91 This was the time of maximum summer insolation (12-13% above present value) of the last interglacial and indeed the last 150,000 years. In the eastern Mediterranean, the climatic oscillations that led from the end of the full interglacial (OIS 5e) to the first large ice advance in OIS 4 produced alternations between the cold, dry chenopod and sagebrush steppe and returns of the Mediterranean mixed evergreen and deciduous woodland.92These interstadial landscapes were more open than in OIS 5e, however, and semidesert plant communities were present even Because their refugiawere close, tree populations during warmer phases.93 expanded rapidly in each interstadial, but the gradual climatic deteriora-

91.Tzedakis 1994. 92. Bottema Tzedakis 1994; 1994; and 1969;Wijmstra Smit Wijmstra and 1976;Wijmstra, Young, Witte 1990. 93. Cheddadi Rossignol-Strick and 1995.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

85

tion and increased aridity are evident in ever larger expansions of chenopods and sagebrush. Still, the cold, arid steppe took over only toward the end of OIS 4, driving the warmth-loving tree populations into refugia even in northwest Greece.94 OIS 3 is marked by several warmer intervals during which a mixed deciduous woodland with beech (Fagus), oak, elm, hazel (Corylus), and lime (Tilia) partly covered northern Greece; southern Greece was sparsely repopulated by deciduous and evergreen oak, pine (Pinus), and juniper (Juniperus)woodland.95Although OIS 3 was a good deal milder than is usually assumed, the Mediterranean woodland at the time was open in character; highest tree densities are recorded in only a few places with optimal soil conditions and sufficient moisture, such as northwestern Greece. During the latest Pleistocene a chenopod and sagebrush steppe typical of a dry,cold climate covered most of the Balkans and Greece.96A low but persistent level of tree pollen suggests, however, that the monotony of the steppe may have been relievedby patches of very open deciduouswoodland. This woodland would have been concentrated on favored south-facing slopes in middle elevations that benefited from orographic rains, precipitation being a more important limitation than temperature.97 This vegetation type vanished around 11 kyr B.P.and was replaced in northern Greece by a deciduous oak forest mixed with more warmth-loving species such as hop hornbeam (Ostrya) and pistachio (Pistacia).98 In coastal regions the Mediterranean woodland of evergreen oak (Quercus ilex), pine (Pinus halepensis),Phyllyrea,wild olive, and pistachio took over slightly later. The impact on animal populations was considerable:wandering herds of herbivores, such as wild ass (Equus hydruntinus),bison, and perhaps Saiga antelope, vanished from the cold coastal and inland plains and were replaced by the more diverse but far more dispersed wildlife of the forest, dominated by red deer and wild boar.99 In northern and western Europe, sharp oscillations between warm and cold climates, of which the Younger Dryas (12.9-12.5 to 11.6-11 kyr Whether such oscillaB.P.) was the last, marked the deglaciation period.100 tions had any real impact in southeasternEurope and the Near and Middle East is in doubt; neither Bottema nor Willis find convincing evidence for them in southeastern Europe during the deglaciation.'10
94. Tzedakis 1993.

Tzedakis 1994. 95.Wijmstra 1969;

96. van Zeist and Bottema 1982; Willis 1994; Willis et al. 1995. 97. Willis 1994. 98. Bottema 1974, 1978. 99. Jameson,Runnels, and van Andel 1994, pp. 331-338; Miracle 1995. 100. Bard and Kromer1995; Kromeret al. 1995. 101. Bottema 1995;Willis 1994. and 102. Bailey,Papaconstantinou, Sturdy1992. 103. van Andel 1998a.

CHRONOLOGY OF THE LATE QUATERNARY WESTERN EPIRUS

OF

Open-air sites are notoriously difficult to place in a chronological context. Bailey, in expressing doubt regarding the utility of Palaeolithic open-air We have approached the sites, had this difficulty very much in mind.102 dating problem in two ways: 1) by paleosol stratigraphy,designed to arrange sites in stratigraphicorder by means of paleosol maturity levels;103 and 2) by the use of thermal luminescence (TL) and infrared stimulated luminescence (IRSL) to obtain calendricalages for the aeolian silt fraction in redeposited terra rossa.

86
PALEOSOL

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

STRATIGRAPHY

IN GREECE

Mediterranean soils evolve in a summer-dry, winter-wet climate that is relatively uniform over large areas. In time the soils mature, forming chronosequences with time-dependent characteristics.104 They are therefore valuable for the identification of Palaeolithic surfaces, stratigraphic correlation across diverse bedrock lithologies, and temporal sequencing of The paleosols discussed here are the alfisols typical for the findspots.105 extensive regions of Mesozoic and Paleogene limestone and flysch and of the Quaternary alluvium derived from those terranes. On different substrates,other kinds of paleosols are found that may also be red or brownish red, such as the rendzinas on Late Tertiary marls in the Peloponnese, but they are not considered here. In using paleosol chronosequences we have limited ourselves to traditional descriptions of soil horizons based on field characteristicsthat allow the assignment of paleosols to six maturity stages (Table 3.8).10?6 Chemical methods can refine the definitions of the stages, but have not yet been used widely in Greece.107 In a typical Mediterranean soil profile, winter rains percolating down from a dark organic A horizon leach a pale E horizon and precipitate solutes in a yellow-brown to red Bt horizon, which becomes progressively enriched in iron oxides that intensify the color with time (Table 3.8). The Bt horizon has an internal structure evolving from small granular aggregates to ever larger blocks and prisms called peds; accumulating illuvial Wherever CaCO3 is clay particles form shiny clay films on ped surfaces.108 present in the substrateor the groundwater,dissolved CaCO3 precipitates to form a calcareousBk horizon below the Bt. Underneath, the soil grades into the unaltered or only little altered substrate,the C horizon. The Bt horizon expresses its increasing maturity by means of changes in color, structure, and the thickness and abundance of clay films.109 The Bk horizon similarlydevelops as a sequence of precipitatedCaCO3 stages.110 Both horizons ultimately reach a point where no further maturation can be detected, unless erosion or renewed deposition terminates the process and sets a new sequence in motion. 104. Birkeland1984; Vreeken1975. The sequence of maturity stages is shown in Figure 3.29, using dates 105. Holliday 1989; Morrison 1976. based on superimposed archaeological sites, 14C-dating of organic sedi106. Birkeland1984, app. 1; ment particles, and luminescence dates of silt grains. All of these dating Retallack1988. methods estimate the time of deposition of the substrate and hence the 107. Fitzgerald1996; Harden 1982; onset of soil formation. The time needed to form a given paleosol can be Harden andTaylor1983; McFadden, determined from U-series dates of calcareouspaleosol nodules.111 Ritter,and Wells 1989; Smith, Nance, Digests and Genes 1997. of all described paleosol Bt horizons in western Epirus are listed in Table 108. Birkeland1984, p. 16; 3.9 with their maturity stages. Retallack1990, p. 40. Paleosol stratigraphyhas worked well in the Peloponnese, Thessaly 109. Birkeland1984, figs. 1-6, 8-10, and Macedonia, and in the Pindos region of Epirus,112 western Epirus but tables 1-4, 8-2. raises problems of its own. Because the redeposited terra rossa is often 110. Birkeland1984, fig. A-2; Machette 1985. initially red, clay-rich, and CaCO3-free, the abundance and thickness of 111. Ku et al. 1979; Ku and Liang clay films on ped surfaces,the remaining Bt diagnostics, are the only ma1983. turity criteria. Color is of no value except where reduction in a water112. Pope and van Andel 1984; logged depositional environment has bleached the sediment and started Runnels and van Andel 1993a; the process of soil formation. Woodward,Macklin, and Lewin 1994.

...

..

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

87

TABLE 3.8. MATURITY INDICATORS QUATERNARY PALEOSOLS

Stage Color

OF THE B HORIZON
BtHorizon Structure granular

OF GREEK
BkHorizon CaCO3 none

ClayFilms none thin, few thin, common thin, many thick to continuous

>2 kyr B.P MS 1 10YR, medium gray-yellowishbrown >4 kyrB.P MS 2 10YR-7.5YR, yellowish to reddishbrown
ca. 10-15 kyr B.P

blocky subangular blocky subangular angularblocky angularblocky to small prismatic

I-II II

MS 3 MS 4 MS 5

7.5YR, reddishto darkbrown 5YR, yellow red to reddishbrown 2.5YR, reddishbrown to red

ca. 40 kyr B.P.

II-III III-IV IV

ca. 80 kyr B.P

ca. 110-200 kyrB.P 2.5YR to 10R, red-brownto red MS 6

medium to largeprismaticor platy thick, pervasive

MS = maturitystage;Color = Munsell color chart.For Bt horizon diagnostics,see Birkeland1984, app. 1. Bk horizon characteristics afterBirkeland1984, fig. A-2. Boundaryages (from Fig. 3.29) are approximations.

. I I

*f

I I

II

MS6
Dating method + o archaeological calibrated radiocarbon
.... :: ::

MS5
MS4

* U/Th disequilibrium * TL/IRSL X stratigraphic

1S3 M

I-

MS2
.....

Figure 3.29. Maturity stages and approximate ages of the Mediterranean paleosol chronosequence. Note that paleosol maturity asymptotically approaches a final stage beyond which no change can be observed. Age scale is logarithmic; no vertical scale. Dates are from Demitrack 1986; Pope and van Andel 1984; Pope, Runnels, and Ku 1984; Runnels and van Andel 1993b; Zangger 1993; and Table 3.9 (below).
After van Andel 1998a, fig. 5

~~~~~-. ~~~~~~MSI . ii! ......:: |~~~~~~~~~~~~~~~~~~~~~~~~~~~i' io i

::.:~~~

~~~~~~~~~~~~~...:. +S *iii

~~~~ ~

....

~,-.. ..
..+.... :' r,. +e

~ ~~ ~ ~~~
is ';

ii

*+*
+ ,I 0 .

1,000

10,000

100,000 years BP

88

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

TABLE 3.9. SHORT DESCRIPTIONS AND MATURITY HORIZONS AT KEY SITES IN COASTAL EPIRUS
Site ALONAKI SS92-23.2 SS92-22.7 SS92-22.7 SS92-22.7 SS92-22.7 SS92-22.7 SS92-22.1 Sample P2 P3topBt P3baseBt P4 VA94-12a VA94-13 P5 Maturity Stage MS 4/5 MS 4 MS 5 MS 5 MS 5 MS 4 MS 5 Color 2.5YR 5YR 2.5YR 2.5YR 2.5YR 5YR 2.5YR

STAGES OF PALEOSOL
ClayFilms

BT

Structure

ang blocky ang blocky ang blocky ang blocky ang blocky ang blocky ang blocky

thick, many thick, many thick, abundant thick, abundant massive thick, many thick, abundant

ALONAKI BEACH SS94-23 Pltop, VA95-01 SS94-23 Plmid, VA95-02 to 4 SS94-22 Plbase AMMOUDIA SS92-21 SS92-21 P1 P2top P2base

MS 1 MS 2 MS 5 MS 3/4 MS 3/4 MS 5

5YR 5YR 10YR

Bt on sandb Holocene dune sandb Bk on sandb (CaCO3 stage II) medium medium thick, abundant thick, abundant (no information) few thick, abundant
massive, pervasive

5YR 5YR 2.5YR

ang blocky ang blocky ang blocky med ang blocky ang blocky med granular ang blocky platy med granular

ANAVATI S

SS94-16
AYIA

Pltop

MS 5

2.5YR

SS93-9.1 SS93-9.2 SS93-9.2 SS93-9.2 AYIOS THOMAS

MS 4 MS 2 top, VA94-27 base,VA94-28 to 31 MS 4/5 MS 6 base

2.5YR 10YR 2.5YR 2.5YR

MS 2 MS 5 MS 3/4 MS 3

7.5YR 2.5YR 5YR 7.5YR

thin, few thick, abundant thin, few thin, pervasive medium, abundant thick, many thin, pervasive thick, abundant

CHEIMADIO

SS94-2 SS94-18

VA94-01 VA94-07

blocky,prism f ang blocky ang blocky

GALATAS SS92-13
KOKKINOPILOS

Ptop, VA94-14 to 24 VA93-05 VA94-04 P4 Top, VA94-05, 34 Base, VA94-06, 35

MS 4 MS 5 MS 4 MS 5 MS 4 MS 5

2.5YR 2.5YR 2.5YR 2.5YR 5YR 2.5YR

ang blocky med ang blocky med ang blocky ang blocky Bt on sandb Bt on sandb

SS91-3
LOUTSA

SS94-12
RODAKI

Above E55c SS92-15 SS92-15

Sample = Sample number(prefaceVA) or paleosol profile number(prefaceP); Color = Munsell soil color chart;Structure= f(ine), med(ium) ang(ular)blocky (Birkeland1984, app. 1). aFromsame stratumas "chippingfloor"industry. bColorson sand or sandstonetend to be 1-2 hue valueslighter than on clay-richsediments. cAlongmain coastalhighwayfrom Prevezato Albania.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

89

Erosion followed by deposition produces complex soil sequences. An example can be seen on the Tyrrhenianbeaches west of Preveza at Alonaki where the interglacial beach sand is topped by a truncated, light red (10R 6/6), very mature Bt horizon overlain by 2-4 m of similar but unconsolidated dune sand (Fig. 3.26:6,7) containing two paleosols. A lower immature (MS 2) Bt horizon of early Holocene age has Mesolithic finds on top of and within its upper 50 cm. It was later truncated and covered by wellsorted dune sands, which locally show an uppermost, very immature paleosol (MS 1). Dune migration and local deflation are continuing today.
DATING OF EPIRUS SEDIMENTS AND FINDSPOTS

113.Dakaris, Higgs,andHey 1964;

Higgs and Vita-Finzi 1966; Higgs et al. 1967; Higgs and Webley 1971. 114. Huxtableet al. 1992. 115. Bailey 1992.

The first attempt to obtain a chronology of the Palaeolithic and Mesolithic in Epirus was undertaken by Cambridge University teams beginning in 1962. Higgs obtained a series of radiocarbonassayson bone, charcoal,and other materials from Asprochaliko and Kastritsa.113 Kastritsawas dated to approximately10-23 kyr B.P.(10,000-20,000 b.p.), and the Upper Palaeolithic deposits at Asprochaliko evidently began to accumulate somewhat earlier,ca. 29 kyr B.P.(26,000 b.p.), but otherwise overlappedthe Kastritsa The Middle Palaeolithic levels at Asprochaliko proved deposits in time.114 to be beyond the effective range of the radiocarbon technique (at that time, ca. 39 kyr B.P.)and Higgs found nothing datable at Kokkinopilos. The later work of Cambridge University teams has added to the chronology. Radiocarbon dates from Late Upper Palaeolithic Klithi (10,420 b.p.-16,490 b.p.)15 fall between the glacial maximum (18-20 kyr B.P.)and the last cold event before the onset of the Holocene, the Younger Dryas of 11-12 kyr B.P.(10-11 kyr b.p.). New dates are also availablefrom Asprochaliko, based on 14C assays and TL analyses of sediments and burned flint.116 The new dates place layers 16 and 18 (basal Mousterian) at ca. 98.5 kyr B.P.(TL), and layer 14 (upperMousterian) at ca. 39 kyrB.P.(37,000 b.p.). Uncertainty remains, however, in part because the TL dates lack the detail necessary to evaluate them. An attempt to date sites Alpha and Beta at Kokkinopilos with optically stimulated luminescence was inconclusive, suggesting only that sediments at the test sites might be older than 150 kyr B.P.117 New dates from Kastritsa,placing the beginning of occupation somewhat earlier,range from 27 to 16 kyr B.P.(24-13 kyr b.p.).118 This program, although adequate for the study of the stratified deposits in the rockshelters, is of little use for dating open-air sites. Most open-air sites are too old to be dated by 14C, even though reliable dates are now being obtained up to 45,000 B.P.,and substances suitable for K/Ar or U/Th dating, such as tephra or flowstone, are lacking. Relative dating is difficult in the absence of floral or faunal remains, and comparisons of lithic industries are useless in the absence of stratified deposits with a succession of lithic types.
LUMINESCENCE DATING OF SEDIMENTS

116.Bailey, and Papaconstantinou,

Sturdy1992; Huxtableet al. 1992. 117. Huxtableet al. 1992. 118. Galanidouet al. 2000. 119. See, e.g., Wintle 1996.

Thermoluminescence dating of sediments has been practiced with varying success since 1979.119Since 1985 it has been possible to date sediments using optical dating methods in which a light-sensitive lumines-

9o

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

While severallight-sensitive signals have been cence signal is measured.120 infrared stimulated luminescence (IRSL) is the preferred method used, for dating loess and colluvial sediments derived from loess.121 Our confirmation of a suggestion byTippett and Heythat the silt in the Kokkinopilos redbeds might be the result of long-distance wind transport,122 a suggestion rejectedby Bailey, made this component an attractivetarget for luminescence dating, notwithstanding an earlier, inconclusive attempt by Debenham.123 this purpose, a suite of samples was collected in sealed, For TL foil-wrapped plastic tubes under conditions of total darkness. and IRSL dating were carried out by Li-Ping Zhou in the Godwin Laboratory at Cambridge University and IRSL dating by Andreas Lang at the Forschungsstelle fur Archaometrie, Max Planck Institut fiir Kernphysik, in Heidelberg, Germany (Table 3.10).124 During long-distance aeolian transport, silt-sized quartz grains will have been fully bleached before deposition. After deposition the grains become covered with more grains and are exposed to radiation from natural sources of radioactivity in their environment. Dose rate estimates depend on uranium, thorium, and potassium concentrations determined by alpha counting, X-ray fluorescence, and neutron activation analysis.In the current study, two methods were used to determine the radiation received by the samples since deposition, known as the equivalent dose (DE). In the additive method (a) a series of laboratorydoses is given to sample disks in order to increase the luminescence signal. This produces a luminescence growth curve that, when extrapolated back to a base level provided by a bleached sample (bleaching: 180 minutes forTL, 60 for IRSL), allows the naturalsignal accumulated since the last exposure to light to be converted to a measure of the equivalent dose. With the regeneration method (r), f3doses are given to sample disks after exposure to light. A match of the naturalluminescence signal with the regenerated one then allows the determination of the DE. The application of this method is ultimately limited by the long-term stability of the signal and by reaching a dose level at which the luminescence signal no longer increases with further applied doses. Thermal instability will result in an underestimation of the true age, whereas saturation of the luminescence signal will permit estimation of a "greaterthan"age. For samples of nonwindborne sediments, bleaching of the earliergeological signal may be incomplete. This will result in an overestimation of the TL age, and possibly the IRSL age, if the laboratorybleaching is more effective at reducing the signal than the original light exposure. For IRSL, the signal can be reduced to 3% of its initial value by exposure to one minute of bright sunlight, whereas 1,000 minutes are requiredfor the TL Therefore, for nonwindborne deposits, ages signal from the same grains.125 obtained using the IRSL data sets are preferred. With the TL and IRSL dates (Table 3.10) and the paleosol maturity stages described above (Table 3.9), we compiled a chronological diagram of the last two glacial/interglacial cycles in the Preveza region that for the first time seriates many open-air sites (Table 3.11). Its "golden spikes"are the confirmation of the existence of an older Middle Palaeolithic between 60 kyr B.P.and the end of the last interglacial,the identification and dating

120. Huntley,Godfrey Smith, and Thewalt 1985. 121. Lang and Wagner 1996. 122. Dakaris,Higgs, and Hey 1964. 123. Bailey,Papaconstantinou, and Sturdy1992. 124. Zhou, van Andel, and Lang 2000. 125. Wintle 1997.

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

9I

AND INFRARED STIMULATED TABLE 3.10. THERMOLUMINESCENCE SEDIMENT DATES FOR WESTERN EPIRUS LUMINESCENCE
Sample VA93-05 VA94-27 VA94-29 VA94-30 VA94-32 Method TLr IRSLa IRSLr IRSLa IRSLr IRSLr TLa
DE

D/R 4.22 4.47 4.16 4.52 5.57

Age (kyrB.P.) 91+?14 6.1?0.6 7?1 65.5 ? 6.8 84?11 83.1? 12

10?2 10?2

Remarks Kokkinopilos,paleosol (MS 5) Ayia, upperpaleosol (MS 2) Ayia, lower paleosol (MS 4) Ayia, lower paleosol (MS 4/5) Alonaki, redepositedterrarossa (MS 3)

385 ? 79 27?+2

31?4 275 ?22 354?40

376 ?49 56?+3 56?+2

TLr IRSLa
VA94-36 TLa

51?2
320 ?24 278 ?16 294 ?41 289?49 443 + 34 445 ?+ 65

5.45

9?2 59 ? 9
51?+8 52?+8

Loutsa, surfacepaleosol (MS 4)

TLr IRSLr
VA94-37 VA94-38 VA95-01 VA95-02 VA95-03 IRSLr

2.18 2.23 2.03 1.82 3.32 2.82 1.57 1.42 1.97 1.71

<128 ?23 <185?28

Anavatis,coastalsand Anavatis,coastalmarsh Alonaki Beach, earlyHolocene paleosol (MS 2) Alonaki Beach, earlyHolocene paleosol (MS 2) Alonaki Beach, Holocene dune Alonaki Beach, Holocene paleosol (MS 1)

TLa
IRSLr TLa IRSLa TLa IRSLa TLa IRSLa

14.8 ?0.7 18.3 ?1.4 34.4+ 2.6 31.3 ? 8.8

6.2 ?0.7

<188?30 7?1 11.1?+1 10.4? 1.6 10.5 ? 3.0

3.9 ?0.6 4.6 ?0.4

7.1 ?0.7
9.3 ?+0.9

VA95-04

TLa IRSLa

4.7?0.7
5.8 ?0.6

11.0?1.8

TL a IRSL = infraredstimulatedluminescence; = thermoluminescence; = additivemethod;r = regenerationmethod; DE = equivalentdose; D/R = dose rate.

of coastal deposits of the last interglacial,and the presence of Palaeolithic industries belonging to or predating the interglacial.The many sites not dated byTL or IRSL have been arrangedby the maturity of the paleosols with which they are associated. As usual, the dates are not as numerous as one might wish and the chronostratigraphicunits are therefore long. The findspots included in each unit are neither necessarily synchronous nor can they be placed in chronological order.Except for sites on alluvialfans, whose lithic assemblages may be in secondary location, all Palaeolithic and Mesolithic findspots are associated with former, level depositional surfaces of poljes, loutses, and, in a few cases, coastal plains.
LATE QUATERNARY AND PALAEOLITHIC CHRONOLOGY

Paleosols are common in stratigraphic sections of the raised poljes and loutses of western Epirus, and many of these landforms are capped with very maturepaleosols that allow the sequencing of Palaeolithicfinds (Table 3.9). The approximate paleosol-based dating has been augmented with thermal and optical luminescence dates (Table 3.10) to construct a provisional chronology of the Palaeolithic and Mesolithic periods (Table 3.11). The Palaeolithic sequence in Epirus begins approximately 200 kyr B.P., perhaps even earlier.The Mousterian is underway before or during

92

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

TABLE 3.11. CHRONOSTRATIGRAPHIC

SITES, SEDIMENTS, AND PALEOSOLS
Age (kyrB.P) OIS or Site,Sediment, Paleosol Holocene Alonaki Beach:dunes Alonaki Beach:uppersoil Ayia:upperpaleosol Alonaki Beach:site Alonaki:young paleosol site Tsarlambas:
Glacial maximum

DIAGRAM FOR ARCHAEOLOGICAL

IN THE PREVEZA
SedimentAge (calkyrB.p)
4.6 ?0.4 5.8 ?0.6

REGION
Paleosol MaturityStage (estimated range) age

6.1 ? 0.6-7? 1 9.4+ 1-10.5 + 3.0

10?2-9?+2

MS 1 (1-3 kyr B.P.) MS 2 (4-10 kyr B.P.) MS 2 (4-10 kyr B.P.) MS 2/3 (10-30 kyr B.P.)

12 2 25 Kastritsa Asprochaliko Klithi 3 13-25a 13-25a 16-10a MS 4 (30-70 kyr B.P.) 39-25a 51 ?8-59?9 MS 4 (30-70 kyrB.P.)

Galatas paleosol Asprochaliko

Loutsa paleosol 59 4 74 Ayia 5 d-a 115 Tyrrhenian/Eemian interglacial 5e 130 6 190
First major ice advance

Kokkinopilos Asprochaliko

65-85 91 ?14
96-98a,b

MS 4/5 (60->100 kyr B.P.) MS 5 (70->100 kyr B.P.)

Anavatis OrmosOdysseos Alonaki: site handaxe? Kokkinopilos

128 23 MS 5? MS 5 ca. 200c

and aBailey,Papaconstantinou, Sturdy1992

bHuxtable et al. 1992

cRunnels and van Andel 1993b

the last interglacial (115-130 kyr B.P.) and continues into the Late Pleistocene (10-29 kyr B.P.). The Upper Palaeolithic (ca. 13-34 kyr B.P.) terminates the sequence at the end of the Pleistocene. The Mesolithic is dated to the early Holocene (7-10.5 kyr B.P.). This chronology can be supplemented with dates for the Palaeolithic-Mesolithic sequences from elsewhere in Greece, including the Southern Argolid, the Argive Plain, Thessaly, Corfu, and Franchthi Cave.126 In an archaeological context what do the ages and maturity values of Table 3.11 really mean?The TL and IRSL dates tell a fairly simple story: they give the ages of the sediments upon which Palaeolithic or Mesolithic tool assemblages were left. But the meaning of the paleosol maturity values, which estimate the length of time a deposit was exposed at the sur-

126. SouthernArgolid:Jameson, Runnels, and van Andel 1994, pp. 325340; Pope, Runnels,and Ku 1984. Argive Plain: Reisch 1980. Thessaly: Runnels 1988; Runnels and van Andel 1993a. Corfu: Sordinas1969. Franchthi Cave:Perles 1987.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

93

and 127. Bailey,Papaconstantinou,

1992. Sturdy and 128.Runnels vanAndel1993a.

face, is less clear.The many findspots buried deep in loutses and poljes are either associated with paleosols of low maturity or not associated with paleosols at all (Figs. 3.16, 3.18), indicating that the floors of poljes and loutses were occupied during brief dry periods or even while they were active, in keeping with the exploitation of their resources. Most findspots, however, especially those at or near the present land surface, are associated with Bt horizons of great maturity that originally formed some meters below the original surface.Allowing for the stripping of a few meters of fragile A and E horizons, these truncated Bt horizons imply that old polje and loutsa surfaceswere stable for long periods. Since it seems unlikely that Palaeolithic humans preferred eroded Bt horizons for their campsites, are their tools all on secondary locations, removed by erosion and redeposited somewhere else and hence of no value at all, as has been claimed by Bailey?127And perhapsmost importantly,how did so many lithic suites become incorporated within the Bt horizon rather than resting upon it? We may dispose of the pessimistic view first.The tool assemblages are always embedded in extremely fine-grained sediments (see above, pp. 6376), easily mobilized by weak currents or winds powerless to move even small flint chips. The artifactstend to be matrix-supported, implying that flint and matrixwere not deposited simultaneouslyby the same agent. The fine stratification that is in most places visible to the careful observer confirms that only very low-energy transporting agents were involved. Debris flows may incorporate stone artifacts, or they may be left on gravel beds deposited by flash floods or small ephemeral streams in distal alluvial fans at polje margins, but those are lithologically distinct events (Figs. 3.13, 3.19) that are rare in the very low-energy environments of poljes and We loutses.128 have been unable to confirm claims by Bailey that the Mousterian at Kokkinopilos was redeposited by late streams in gullies incised in pre-Middle Pleistocene or Pliocene redbeds. Because the strong currentsrequiredto move objects of gravel size are rarein polje and loutsa basins, the stone tool assemblages must have been lowered onto, or more often worked into, the Bt horizon by gentle removal of the overlying A-E horizons, or the Bt horizon must have crept upward over time to engulf and protect them. Gentle erosion is possible even in the low-energy polje and loutsa environments, but its universal action to explain the position of all stone tool suites on top of mature Bt paleosols beggars belief and it is incapable of inserting them within the Bt. Artifacts may work their way down during the dry season in clays that alternately swell and crack, but this would disperse the assemblage vertically,perhaps sending the smaller ones deeper down. This is not what we observed in Epirus where the artifacts,small and large, tended to stay together in clusters or thin lenses, often giving the impression of being in situ. Another process seems required, for which we present the following working hypothesis. If the land surfaceis raised at a very slow rate, the top of the Bt horizon itself would move gradually upward as a result of slow deposition, so engulfing any artifacts laid down on former land surfaces above it. This process depends on the relative rates of sedimentation and

94

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

> DEPOSITION SOILFORMATION

debris flow

> SOILFORMATION DEPOSITION

EXAMPLE: KOKKINOPILOS
ni','

II II 1 11

Bt(M5)

r '"li,A"l",

,
Bt(M1)

4
,iii,,,, Bt(M2)

4
Bt(M4)
. Bt(M1)

AIc.4944^ debris flow

.A I

Bt(M1) gravel

3
A A

.,,,.,*flint

Bt(M1)
1811111

AA

"handaxe"

li lPI

Bt(M3)

2
111111111

Bt(M2) Bt(M1)

"""""""""' Bt(M1)

||l||Is1|||||| Bt(M2)

kst
karst

1
karst
AA

karst
Palaeolithic stone tools

mIIII paleosol (Bt) or desiccation zone

between Figure3.30. Relationship terrarossa paleosolmaturity, depositionrate,and Palaeolithic stone tool age in poljesandloutses. Stone tools aredepositedon old the surfaces, age of whichis defined byTL or IRSL dates.Left, sequence 1-4: Depositionrateexceedsrateof soil formation; immature paleosols areassociated with old Palaeolithic material. Center,sequence1-4: Rate of soil formationequalsor exceeds Bt depositionrate;maturing horizon until it engulfsstone growsupward tool assemblage. Right:Profilefrom both Kokkinopilos, incorporating phenomena.

paleosol formation. In a landscape where the slope mantle has not yet been destroyed, terra rossa deposition rates in an active polje may be as In high as 10-15 cm/kyr.129 contrast, soil formation is slow.130 If deposition is significantly faster than soil formation, either no soil or an immature one will form when deposition is temporarily interrupted by a period of drought. Stone tools may then be left on desiccation surfaces, on thin and immature paleosols, or in and on marginal fan deposits (Fig. 3.30, left). If, on the other hand, deposition raises the land surface more slowly than the Bt and Bk horizons form, the horizons will thicken upward into the overlying sediment, a process that is common in the lower floodplains of small riversin the semiarid climate of the Peloponnese and Thessaly.131 When the polje approaches old age, the rate of soil formation begins to equal or exceed the rate of deposition. In the now raised polje at Morphi, for instance, a volcanic ash dated at 374?+ kyr B.P.and located 12 m below 7 a very mature paleosol estimated to be ca. 100,000 years old implies an average deposition rate of a mere 4 mm/kyr.132 The maturing Bt horizon

129. Kukal1990, pp. 101-103; Runnels and van Andel 1993b. 130. Spaargaren 1979; Williams and Polach 1971; Magaritz,Kaufman,and Yaalon1981; Demitrack 1986; McFadden and Weldon 1987; Harden et al. 1991; Bockheim,Marshall,and Kelsey 1996. 131. Pope and van Andel 1984; Jameson,Runnels,and van Andel 1994. See also Birkeland1984, figs. 8-10. 132. Pyle et al. 1998.

EARLY STONE

AGE OF THE NOMOS

OF PREVEZA

95

until it engulfsanystone tools, protecting them from then growsupward soil andwind erosion(Fig. 3.30, center). The Kokkinopilos both prosection (Fig. 3.30, right) demonstrates a cesses.Duringits youth,the poljeaccumulated thicksedimentsequence with scatteredimmature paleosolsand widely spacedstone tool assemin its old age,however, of was whenthe dissolution limestone slower blages; than its removal,the depletionof the slope mantlesharplyreducedthe sediment supply.Moreover,uplift eliminatedmost of the runoff from streamincision.This droppedthe rateof springsand initiatedheadward depositionbelow that of soil formationand produceda thick, consolidatedBt thatincorporated stonetools andkeptthem safeuntilNeolithic, BronzeAge, or in this case,post-Romansoil erosionexhumedthem.133 A differentexampleis Ayia (Fig. 3.20), a shallowloutsawith only 810 m of fill, now almost entirely removed by recent erosion. A composite

mottledsequence with disprofile(Fig.3.19) showsa lower,red-and-gray tinctcentimeter-scale stratification formedin an alternatingly andwet dry environment. Palaeolithicstone tools are intercalated this sequence, in once by a debrisflow underneath a which, nearthe top, was interrupted mature(MS 5) paleosol. The thin, muchyoungerupperzone has an immature(MS 2) Bt horizon.

THE ARCHAEOLOGICAL
GOALS AND PROCEDURES

SURVEY

133. Bailey,Papaconstantinou, and Sturdy1992; Runnels and van Andel 1993b;van Andel, Zangger,and Demitrack 1990.

The survey of prehistoric sites took place between 1991 and 1995. Special attention was given to the red sediments (paleosols and redeposited terra rossa) because of their known association here and elsewhere with Palaeolithic artifacts, and the structure and characteristicsof paleosol horizons were investigated to establish a rough chronology of the archaeological finds. In practical terms we used the availablegeological and topographic maps as a rough guide. Our goal was to produce a complete picture of Palaeolithic activity,in as wide a variety of geographic contexts as possible, within the time availablefor searching. Fifty-seven days of fieldwork were undertakenby a specialistteam devoted entirelyto the searchfor Palaeolithic and Mesolithic sites and consisting of three to four persons at all times, with the addition of student volunteers to assist.Two strategies were pursued. The first strategy was to locate and search all occurrences of Pleistocene soils and sediments in the study area (Fig. 3.8). The second strategy was intended to increase the coverage of the surface by inspecting nonredbed surfaces, such as dunes, alluvial fans, bare hillslopes, and remnants of the old peneplain (e.g., in the vicinity of the village of Loutsa). The search for Palaeolithic and Mesolithic materials was also part of the general diachronicsurvey. The general surveyteams, consisting of three or four field school students and two experienced graduate students as leaders, were trained to identify and collect all lithic artifactsbefore walking survey tracts.This point should be emphasized: in order to minimize selection bias, fieldwalkers were taught to recognize and collect all lithic

96

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

artifactsregardless size, raw material,type, or date. The quantityof of lithic artifacts collected(ca.13,000 pieces)andthe rangeof periodsrepresented (Lower Palaeolithic moderngunflints)are evidencethat this to This procedure usefulin two ways.Largetracts was was training effective. of presumed surface wereinspected the usually and post-Pleistocene negative resultshelped to confirmour assumption that we were not missing The teams also walkedPleistocenedepositsnot any sites in these areas. team(e.g.,the AyiosThomaspeninsula), disinspected the specialized by which materially Palaeolithic increased our coveringimportant findspots confidencethat a reasonably arcomprehensive pictureof the preserved recordhad been obtained. The tractfinds collectedby these chaeological teams were inspectedon a daily basis by one of us (CR), and in cases wheregeneralsurveyteamsbroughtin lithics from tracts,walkovers, or site/scattersthat were of interest,the specialistprehistoric surveyteam revisitedthe areato make a separateinspectionand collect and record additional samplesas walkovers. At the timewhen Palaeolithic Mesolithicartifacts or werediscovered were team, the followingprocedures by the specialistprehistoric survey more detaileddescription collectionprocedures of used by employed(a generalsurveyteamsis givenin Chapter2). The firstconcernat all times was to determinethe sourceof flints found on the surface. Our working modelof site formation wasbasedon the assumption flints that processes were associated with redeposited terrarossa,and to test this hypothesis eachfindspotwas searched for carefully a sourceof the lithics.At a numberof sites (e.g.,Alonaki,Galatas,Kranea, Kokkinopilos), and flints embeddedin the sedimentswere associated with paleosols(Bt horizons)of variousmaturitystages,formedwhen the originalsurfacehad been exposed for a sufficientquantityof time. At Ayia, on the otherhand,fresh unweathered flintswerefoundin unbroken, "mint" conditionwithin unbeddeddepositsat a depthof 3 m consolidated, interrupted, horizontally below the modernsurface, wherethey must havebeen depositedduring There canbe no questionthat these seasonal, intervals. dry brief,perhaps flints arepart of the redbeddepositand must be considered situ in a in The geologicalsenseandnot laterintrusions. evidenceforlaterreworking of redbeddepositsat Kokkinopilos, citedby Baileyas proofthat the flints are accidental is situatedin gullies in intrusions,134 basedon excavations the northeastmarginsof the site; these gullieswere probably subjectto local reworking did not affectthe entirepoljefill. that The numberof flints exposedin a paleosolhorizonwas limited,and the samplefor each findspotwas supplemented collectingflints from by the surface weredeemedto be derived, highlylikelyto be derived, that or fromthe paleosol. This elementof subjective judgmentwas basedon extensive experience,and was justified by the close spatialassociationof materials within the outcropandjust belowits weathered face.The flints areoften stainedwith redclayfromthe paleosolsor havefragments Bt of material to includeall retouched artifacts, adhering them.The collections with typological charcores,completeflakes,blades,andflakingdebitage acteristics (e.g., corerejuvenation pieces).Only incompleteandtypologiwerediscarded on-site. callyunclassifiable fragments

134. Bailey,Papaconstantinou, and Sturdy1992.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

97

Our purpose was not to collect samples for analyses that requirelarge numbers of pieces for detailed typology or spatial analysis of activity areas. With the exception of one site (Spilaion; see Chapter 4), we did not attempt to map the spatial distribution of materials. Recent erosion of the paleosols was responsible for extracting flints and scattering them on the surface and in erosional gullies that dissected the redbed sequence. In our view only excavations of the paleosols would reveal culturallymeaningful spatial patterns. The samples collected from the surface were intended only to provide sufficient information to assign the findspot to a cultural period and to compare it with other findspots in the region. Papagianni has undertaken a more detailed technological analysis of the Mousterian from our collection, utilizing essentially all Mousterian artifacts found in Epirus since 1962.135 The treatment of artifacts collected on the surface was simple. The lithics were soaked in water to clean them before they were bagged for storage in PVC bags labeled with provenience data. All samples were recorded in field notebooks and on printed recording forms, which permitted the samples to be tracked through cleaning and storage, and the information transferredto the project'scomputerized database. Once cleaned, the lithics were described and assigned to typological categories according to the system of classification originally developed by Fran9ois Bordes, with certain modifications that have become accepted in recentyears.136 Selected specimenswere pulled from the samplesfor drawing and photography.These selected specimens were given separateinventory numbers, in addition to their sample numbers, to aid retrieval.
ARCHAEOLOGICAL SETTING SITES IN THEIR GEOLOGICAL

135. Papagianni1999. 136. Bordes 1992; Debenath and Dibble 1994; Mellars 1996, pp. 169192.

Our surveywas carriedout in the territorywest of the Louros Rivervalley, with an emphasis on the coast from Parga to Preveza, and produced evidence for human activityfrom the earlyPalaeolithicthrough the Mesolithic. The majority of sites are coastal with the exception of those in the Thesprotiko and Cheimadio valleys. An interesting finding was a number of smaller, perhaps specialized, sites that may include quarry sites and flintknapping areas,which supplement our picture of the regional settlement pattern. Our programdiscoveredor confirmed forty-four prehistoricfindspots called "Site/Scatters"and designated "SS,"followed by the year and number of the site recorded in that season (e.g., SS92-22 for Alonaki in the Acheron valley, the twenty-second findspot recorded in the 1992 season; see Appendix). Approximately 4,600 lithic artifacts were collected from these findspots and were used to assign them to general periods. Of these findspots, four produced Lower Palaeolithic materials, thirty produced Middle Palaeolithic, six produced Upper Palaeolithic, and six Mesolithic. We supplemented our surveywith data from an extensive program of augering in the Acheron River valley and the Louros delta (see Chapters 5 and 6); a geochronological program of radiocarbon(14C),thermolumines-

98

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

TABLE 3.12. EARLY STONE AGE CHRONOLOGY

Period Calendar Years(kyrB.P.)

Mesolithic Upper Palaeolithic Middle Palaeolithic(EarlyPalaeolithic) Lower Palaeolithic(EarlyPalaeolithic)

7 to 10.5 13 to 34 >31 >100

cence (TL), and infrared stimulated luminescence (IRSL) dating of alluvial sediments and sand dunes; and laboratory sedimentological analyses. A chronological summary is given in Table 3.12.
THE EARLY PALAEOLITHIC

The traditional terms of "Lower"and "Middle" Palaeolithic have been questioned in recentyearsprimarilybecause they referboth to chronostratigraphic units and lithic typology, which overlap and are not congruent. The Lower Palaeolithic was once regarded as a Middle Pleistocene sequence of Acheulean industries with the handaxe-cleavercomplex as type fossils. The Middle Palaeolithic was a Late Pleistocene flake industry (the Mousterian) with Levallois technology. It was also thought that the Acheulean was associated with Homo erectusand the Mousterian with Neanderthals or other archaic Homo sapiens.137 of these assumptions All have proved to be unreliable. The Acheulean, with handaxes, continues until the last interglacial (ca. 115-130 kyr B.P.) in many places, while new finds have placed the beginning of the Mousterian at more than 100,000 years before the last interglacial (ca. 200-250 kyr B.P.).There are significant overlapsin chronostratigraphicterms, and the Acheulean and Mousterian also sharethe use of the Levallois technique,flake tools, and handaxes in (or "bifaces" formal typology). It is thus no longer possible to correlate lithic technocomplexes and hominid grades. Some authorities question whether Homo erectus was responsible for the European Acheulean, which might also be attributed to archaic Homo sapiens, and both Neanderthals and anatomically modern Homo sapiensare associated with classic Middle Palaeolithic in the Near East. Neanderthals in western Europe and perhaps the Balkans are responsible for industries that are similar to and contemporarywith industries ascribed to the Early Upper Palaeolithic (EUP). It is unlikely that traditional lithic industrial identifications can be other than labels of convenience, permitting us to discuss problems and describe newly discovered materials but which in no way imply either chronological position or cultural affinities. In these cases, Rolland recommends calling the traditional Lower and Middle Palaeolithic "EarlyPalaeolithic"to avoid the problems inherent in the earlier classification.138 We will follow that suggestion in this report, although we also use the older terminology when greaterchronological or typological precision is required. Localities with the earliest materials (on stratigraphic and chronometric grounds) are found at Kokkinopilos, Alonaki, and Ayios Thomas. Kokkinopilos (SS91-3) is the most important of these, and has been de-

137. Mellars 1996, pp. 2-4. 138. Rolland 1986.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

99

site/scatters Figure3.31. Palaeolithic in the Thesprotiko valley,showing the locationof findspotsassociated with redbeds.Kokkinopilos a is site. Galatasand Kranea are major to be specialized presumed activity sites. Smallfindspotsarefoundon the marginsof the poljedepositsand at the entrances the valley. to Hatchedareasaremodernsettlements. A scribedin detail (Fig. 3.31).139 pointed biface (handaxe)of late Acheulean was found stratified within a zone of interbedded subaeriallyweathtype ered but mainly subaqueous polje or loutsa deposits (Fig. 3.17) ca. 17 m below a paleosol containing a later Middle Palaeolithic industry dated to ca. 90 kyr B.P. (Table 3.11), close to the present center of the polje and near its thickest deposits. Three undisturbed immature paleosols mark the interval between the Middle Palaeolithic paleosol and the handaxe zone, which is almost entirely sterile except for a thin (ca. 50-cm) bed of matrixsupported fine flint gravel a few meters above the handaxe. At about the same or slightly higher stratigraphiclevels, other localities roughly to the south and southwest of the findspot produced heavily patinated artifacts of large size. In 1991 we observed numerous artifacts eroding from the sediments in the northwest part of the deposit and perhaps similar to the floors described by Higgs in the northeast part of the site some "chipping" 300 m away.140 artifactsconsisted of large flake tools, non-Levallois in The with denticulate and notched edges. We were prevented from technique, making a collection of these artifacts and from sampling this layer for geochronological dating by a post-issue alterationto the Nikopolis Project's research permit; we are thus unable to give details about the lithics from the handaxe layer or to directly date the handaxe. Our best estimate of date (150-200 kyr B.P.) is derived from extrapolation of plausible sedimentation ratesbased on the dating of the very matureMiddle Palaeolithic paleosol at the top of the sequence, obtained on a sample taken before our researchwas stopped (Table 3.10; VA93-05).

139. Runnels and van Andel 1993b. 140. Dakaris,Higgs, and Hey 1964.

IO0

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

and Figure3.32. Palaeolithic Mesolithicsite/scatters the in Acheronvalley.Alonakiis a major site. Smallerfindspots(e.g., Skepare asto,Loutsa,andValanidorrachi) sites. Other specialized activity Tsoukfindspots(e.g., Ayia Kyriaki, nida,Ammoudia),perhapstemporarycamps,arelocatedon the edge of the valley. Two Mesolithicsites (Ammoudia,Loutsa)arefound near the coast,andTsouknida located was on the edge of an ancientlakeor Hatchedareasare embayment. modernsettlements. A second area of interest that produced material of Early Palaeolithic characteris Alonaki in the Acheron valley (Fig. 3.32). Alonaki (SS92-22, SS92-23) appears to be a loutsa-type karst depression filled with alluvial/ colluvial, redeposited terra rossa. An extensive outcrop was inspected and found to have at least two distinct Bt horizons. The upper Bt has a maturity of MS 4 and the lower Bt MS 4/5 or 5 (Table 3.9). The sequence has a total depth of more than 3 m below the surface, and appears to contain more than one Palaeolithic industry. Large flake artifacts were found throughout the deposit, including the lowest part, exposed in a clay extraction pit. Although our ability to correlatethe industries with outcrops of different depths is limited, it appears that a conventional Middle Palaeolithic Mousterian is found on or near the upper Bt horizon and a significantly earlier large flake industry in the lower Bt horizon. A curious feature of the Alonaki deposits is the presence of dense concentrations of angular stones mixed with lithics, which occur in discrete units 1-3 m in diameter and ca. 0.30 m thick (Fig. 3.33). They are unsorted and matrix-supported, but their sharp boundaries (at top, bottom, and laterally)against redeposited terrarossa argue strongly against an origin as a stream channel or debris-flow deposit. These features resemble in some ways the "stoneclusters"identified at Early Palaeolithic sites as far afield as Hoxne in England, which are sometimes described as artificial in origin.141 The Alonaki "stone clusters"are associated with the artifacts of the lower Bt horizon, large flakes with wide, thick platforms and large bulbs of percussion and equally large cores of non-Levallois type (Fig. 3.34). These materials were recovered from the bottom of a shallow ero-

141. E.g., Singer,Gladfelter,and Wymer 1993, p. 124.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

IOI

Figure3.33. View of a stone cluster: at Alonakiillustrating sharply the definededgesof the feature

,i

-i

sional gully that cuts through the site as well as in the modern clay extraction pit. In a few cases they were prized from within the structures found

in the lower Bt horizon, and must clearly be regarded as in situ within the paleosol. In the surface collections some mixing with later materials is unavoidable, but in the lower levels of the deposit, wherever in situ artifacts were observed, they were always of the non-Levallois big flake type. These large artifacts differ in raw material, technique, and retouched tool typology from the Mousterian and consist chiefly of core-choppers and flakes (Fig. 3.35). The raw material is a dull dark brown, fossiliferous
chert, derived from Eocene limestone, that contrasts with the glassy bluish-gray nodular flint without macroscopic fossils that is derived from the

Mesozoic Pantokrator limestone and was widely used to manufacture Mousterian artifactsin Epirus.The Eocene chert has been worked by hardhammer direct percussion. Flakes have large broad platforms and welldefined, swelling bulbs of percussion. The size of the platform and the pronounced swelling of the bulb are indications that considerable force was used to detach each flake from its core. Cores include core-choppers (Fig. 3.36) and large cobbles from which flakes were removed from one face using a broad plain striking surface (Fig. 3.37). The resulting flake scars are wide and deep. There is not enough material for a metrical analysis or a study of the complete reduction sequence, but the characteristicswe can observe seem to point to the production of broad flakes from boulders and large cobbles as the chief goal. There are other characteristicsthat separatethe lower Bt industry from the Mousterian. Retouch is confined to direct, invasive retouch, and large notches were created by a single inverse or direct blow (Clactonian technique). Notched and denticulated edges are common, and typical Mousterian forms, such as points and side scrapers,are lacking in this material.

102

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure 3.34. Early Palaeolithic artifacts from Alonaki: 1) double convergent side scraper; 2-4) notched pieces/denticulates; 5) convex side scraper.Scale1:2

Figure 3.35. Early Palaeolithic choppers from Alonaki. Scale1:4

EARLY

STONE

AGE

OF THE

NOMOS

OF

PREVEZA

Io3

Figure 3.36 (above,left). Early

from Palaeolithic core-choppers

Alonaki. Scale1:2 Figure 3.37 (above, right). Early

corefromAlonaki. Palaeolithic
Scale 1:2

Also from this area is a small biface (handaxe), found near Ormos Odysseos (W94-20), about 500 m to the west of Alonaki (Fig. 3.38). Here a thick mantle of Pleistocene red clay, sand, and gravel (and possibly involving a paleosol as well) covers a limestone karst surface inland from a coastal paleosol superimposed on a sand dune of probable interglacial age (SS92-25) which also contains Palaeolithic artifacts (Fig. 3.39). The inland deposit is today overgrown with bushes, but goats have worn trails through them and the tracks have eroded down to bedrock exposing outcrops up to 3 m thick. The handaxe was found in one such ravine (Fig. 3.40). The area is essentially level and the handaxe could not have been transported very far. Other artifacts were observed in the deposit, which may be of the same general age as the lower Bt horizon at Alonaki. The sand dune, which overlies the Bt deposit at its northwestern corner, is nearly at present sea level, but is definitely of Pleistocene age and hence can only belong to the high sea level of the last interglacial period (or an even earlier interglacial). In our estimation the preponderance of the evidence the high maturity of the Alonaki paleosols and the overlying interglacialcoastal suiteplaces the lower Bt paleosol with its associated artifacts before the last interglacial, or more than 130,000 years ago, close to the lower age suggested for the Kokkinopilos handaxe (150-200 kyr B.P.). Other Early Palaeolithic materials are found in the southern part of the survey areanear the town of Preveza.Tracts walked on the Ayios Thomas peninsula at the northern end of the Ormos Vathy (T93-17, T93-3, T93-4, T93-5) recovered large numbers of Mousterian pieces from a

I04

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure 3.38. Early Palaeolithic biface (handaxe) from Ormos Odysseos.

Scale 1:2

Figure 3.39. Interglacial sand dune (SS92-25) at Ormos Odysseos, looking southwest

Figure 3.40. Ormos Odysseos, biface findspot (W94-20), looking south

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

Io5

biface Figure3.41. EarlyPalaeolithic or bifacialcorefromAyiosThomas.

Scale 1:2

'

paleosol associated with marine deposits of Eemian age. Among these materials are large flakes of Eocene chert similar to the Alonaki lower Bt artifacts,including a rough amygdaloidalbiface or bifacial core (Fig. 3.41). Very few outcrops exist in this area and the exact source of these materials could not be pinpointed. The deposits are bedded horizontally and the material cannot have been transportedfar.Although no age assignment is possible, we suspect that these materials are of the same general age and type as those from Alonaki. Except for Kokkinopilos, Early Palaeolithic materials are found only on the present coastline, and it is for this reason that they have not been noticed before. An inspection of Higgs's collections in the Ioannina Archaeological Museum showed that they contain no artifacts similar to the lower Bt materials at Alonaki.
THE MOUSTERIAN (MIDDLE PALAEOLITHIC)

142.Dakaris, Higgs,andHey 1964.

Most Early Palaeolithic artifacts in the Preveza nomos are Mousterian in type. In his pioneering survey,Higgs found large numbers of Mousterian artifacts on surface sites in Epirus, including thousands from Morphi, Karvounari,and Kokkinopilos.142 Specialist prehistoric survey identified thirty findspots (site/scatters in databaseterminology); although this number could be easily multiplied by additional fieldwork, we believe it includes a representativerange of site types and habitats. Our study collection from Mousterian findspots includes more than 1,500 artifacts, and about 10% of the 13,000 lithics collected by the general survey teams are also Mousterian. The abundance of the Mousterian may be attributed to several factors. Geological contexts of the appropriate age are more common than those of earlier periods or those immediately following. Mousterian sites are also more conspicuous ("obtrusive"in survey terminology), and the preferentialselection of redbeds for their camps makes them easy to find: the large, often heavily patinated, artifacts stand out as white spots on a

io6

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

has The redbackground. erosionaffectingmanyof the redbeds no doubt of sitesby deeplyincisinggulliesanddeto contributed the obtrusiveness nudingthe depositsof vegetation. is BesidesEpirus,the Mousterian foundin the Argolid,Corfu,Elis, andmanyotherlocalities.143 preferThe Messenia, Kephallinia, Thessaly, is ence for open-airsettlements one reasonfor this ubiquitousness. While of is UpperPalaeolithic occupation cavesandrockshelters commonin the lastglacial Mousterian of (OIS2, ca.25-12 kyrB.P.), occupation rockshelters or cavesis rare: and are Asprochaliko, Kephalari, Franchthi at presentthe Site faconlypublished examples. selectionstrategy be another may major and climatealso playeda role.The Mousterianis found in the early tor, in glacialperiod(OIS 4-3), a time marked westernEuropeby numerous climaticoscillations from nearlyfull glacialto warmconditions,some of which (e.g., the Hengelo interstadial ca.40 kyrB.P. the Denekamp at and interstadial ca. 36-32 kyr B.P.) were quite mild.144 these conditions at If in alsoprevailed southeastern this Europe, wouldsuggestthatMousterian settlementwas encouraged or limitedto, the warmphases. by, The identityof the makersof the Mousterian a contentious is probevidenceindicates thatNeanderthals werethe prolem,but the European ducersof this industry, workinghypothesis accept.145 the Euroa we Like the Mousterianin Epirus similarlyshows relativelylittle pean finds, variation. That of the basallayers(16 and18) atAsprochaliko, datedto ca. 98.5 kyr B.P., is characterized the frequentuse of the Levalloistechby coresto producenumerous flakes.146 adIn niqueto prepare largelamellar and dition,Levalloispoints,largeconvexside scrapers (racloirs), otherelements are typical.This Mousterianbelongs to OIS 5 (ca. 115-74 kyr continues into OIS 4 and3 (ca.74-59 kyrB.P.),the early B.P.) andperhaps The Mousterianof Asprochaliko's glacial. upperlevel (layer14), poorly dated by radiocarbon assaysrangingfrom 29 kyr B.P. (26,000 b.p.) to but >39,900b.p., 47is quite similar, makesless use of the Levalloistechand is rich in Mousterian in nique points and smallside scrapers a wide range of types. It was once describedas a diminutivefacies called the but Micromousterian, this designationhas been questionedfor Asprochalikobecause difference sizebetweenthe basalandupper(orlate) the in Mousterianis not greatenough to warrant qualifier the "micro" the for latter.148 There is no question,however,that the late Mousteriandiffers fromthe precedingLevallois-Mousterian some typological, in technical, andmetricalcharacteristics that it is younger. and Mousterianartifacts found on the surfacecan be placedin chronologicalorderonly with greatdifficulty(Table3.11). There are abundant surface with Mousterian sites finds:Kokkinopilos, (SS93-9),Alonaki Ayia (SS92-22 and SS92-23), Kranea (SS92-14), the Anavatisquarry (SS9413 andSS94-16),Skepasto (SS92-20),andValanidorrachi (SS91-4),among others.Late Mousterianartifacts found at Kokkinopilos, are Ayia (in its upperlevels),Alonaki (SS92-22), Galatas(SS92-13), Loutsa (SS93-31, sites.A possiblesourceof chronology the for SS94-12),andsome smaller laterMousteriancomes from the SouthernArgolidandThessalywhere similar Mousterian artifacts datedbyradiocarbon U/Th seriesfrom are and
55 to 30 kyr B.P.149

143. Bailey et al. 1999; Runnels 1995. 144. van Andel andTzedakis 1998. 145. Mellars 1996, pp. 1-8.

146.Bailey, and Papaconstantinou,

Sturdy1992; Huxtable et al. 1992. 147. Bailey,Papaconstantinou, and Sturdy1992, p. 138 148. Bailey,Papaconstantinou, and Sturdy1992; Huxtable et al. 1992. 149. Pope, Runnels,and Ku 1984; Runnels 1988; Runnels and van Andel 1993a.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

IO7

150. Dakaris,Higgs, and Hey 1964.

A major interest of our survey was the reconstruction of Mousterian paleoenvironments, settlement, and land use. Our discussion of this topic is divided into two parts in accordwith the twofold division of the Mousterian, although the many elements of continuity should be stressed.The decisive featuregoverning land use and settlement patterns is the distribution of karst features such as poljes, loutses, and dolines that served to attract and concentrate animal, plant, and mineral resources and permitted and encouraged a seasonally scheduled, partially logistical strategy of land use. Throughout this discussion we referto strategiesthat make structured, planned, and repeated use of a landscape as logistical or partially logistical land-use strategies. The earlierMousterian is found in abundancein the redbedsof Epirus, particularlyat Kokkinopilos.The variety of locations showing evidence of early Mousterian activity is perhaps the best picture of partial logistical land use. The largest concentrations of artifacts are found at Alonaki, Kokkinopilos, and Ayios Thomas (Ormos Vathy). A fourth findspot is in the vicinity of Morphi in Thesprotia, where large numbers of Mousterian artifactswere collected by Higgs.150 These larger sites are supplemented by a series of small sites at Ayia, Kranea, Loutsa, and Anavatis. Still smaller sites, possibly specialized in located near flint outcharacter,are found at Skepasto and Valanidorrachi, cropswhere quarrying,flintknapping,and testing of nodules were the main activities. The site of Rodaki may have been occupied by Neanderthals utilizing coastal resources, although the lack of faunal remains here and elsewhere makes this hypothesis difficult to evaluate.The principal characteristics of the known smaller sites are these. All presented easily available surfacewater,which ponded on the clay surfacesof the loutses in late winter and spring and slowly evaporated in summer. We found standing water and evidence of recent wet conditions on these sites to the end of June and into earlyJuly.Loutses mainly depend on winter rain ratherthan on major springs and are found as shallow depressions in exposed localities where they dry out early.They are more exposed to the elements than poljes, which are located in deep basins that offer more sheltered conditions. Large poljes, like the modern Valtos Kalodiki, retain water much longer, or permanently in the form of shallow lakes, swamps, ol marshes. There were found reeds, aquatic plants, willows, and stands of trees in well-watered side valleystogether with a variedwildlife. Camps were placed along the margins of the poljes-partly to be on well-drained ground and partly to avoid scaring off the game-but near springs or the inlets of winter or spring streams. Locations were probably shifted often. If groups returned on a seasonal basis over a long time, the spread of artifacts from overlapping camps would make spatial analysis difficult and may account for the large quantities of artifacts. Loutses and poljes were magnets for animals and humans in this glacial landscape. Rivers in the summer carriedsome meltwater,but they had incised their channels to reach lower shorelines. Away from the rivers, springs were sources,but in the karstlandscape the limestone bedrock has no surface water in the dry season. The poljes preserved water when it would be least available,in the late summer and autumn months, and they

Io8

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

of offeredpredictable placesto find food as well.The predictability these resources explainthe partially concentrated spatially may logisticalsettlementpattern seenin Mousterian times.In a fullylogisticland-usepattern, the smaller areas wouldhaveserved funcoutlyingactivity veryspecialized tions,as is thoughtto be the casewith the ibexandchamoishuntingcamp at Klithiin the late Palaeolithic,151 thereis no evidencefor this degree but in of specialization the Mousterian. In contrastto the modifiedlogisticalpatternpostulatedhere, some scholars havedevelopeda pictureof Neanderthals opportunistic as foragin erswho movedaboutthe landscape searchof food.152 Such a modelof residential mobilityalso postulatesa repeateduse of scheduledseasonal at particular sites,andthis patternof landuse mayhavegradedinto stops variouskinds of logisticalforagingthat dependedto a greateror lesser degreeon a few basecamps. In our model of modifiedlogisticalland use, the differenttypes of sitesoffereddifferent-sized differently-timed and of packages water, plants There was also a (for food, handlesfor tools, shelter,fuel), and animals. good chanceof findingusefulquantitiesof flint for toolmakingin most locations.Ayia is a typicalexampleof the smaller "loutsa" consisting site, of sometimes of associated with smallreddeposits, largenumbers artifacts no and typically morethan300 or 400 m in diameter locatedat somedistancefromlarger sitessuchasMorphiandKokkinopilos (Fig.3.42). polje The findspotsnearLoutsaand Kranea also examples. are These sites are found in remotemountainous locationsat elevationsup to 400 masl or more.The lithics fromAyia includeflintknapping debrisand retouched tools indicatinga wide rangeof activitiesat the site (Figs.3.43, 3.44). Even smallersites show more specializedactivities.Skepastoand in Valanidorrachi the Acheronvalleyappear be flintknapping to sites.At there are manyworkedand unworkednodules of flint, some Skepasto associated with thesenodweighingup to 15 kg, erodingfromlimestone; ules are numeroustest-cores,Levalloisand other cores,Levalloisblades and flakes,and rarefinished artifacts(e.g., two Levalloispoints).The Anavatisquarry sites,with paleosolscontainingboth coresand finished on streamfan (Figs. tools, appearto be small encampments a torrential 3.45, 3.46). An interesting exampleof the morespecialized type of site is Rodaki, a red depositlocatedat the presentmouth of the Paliourias River(Fig. were found stratified a complexsein 3.45). A largenumberof artifacts quenceof paleosols,separated a normalfaultfroma thick sequenceof by marinedepositsof probable are interglacial (Fig. 3.47). The artifacts age in situ in a stonyred paleosolthat is cappedby a layer(ca.2 m thick) of stone-freeduneor coastalsand.The lower,stonypaleosolwith artifacts is buried,but the upper30 cm of its thicknessis exposed. nearlycompletely This partis rich in artifacts that aredifficultto classify(Fig. 3.48). They resemblematerials from the islandof Zakynthosat the site of Vassiliko where Sordinasfound them in red sedimentsinterstratified with marine The are to deposits.153 Zakynthosartifacts undated,but appear be a specializedvariantof the Mousterian.154 this respectthey bear a resemIn

151. Bailey 1997. 152. Mellars 1996, pp. 356-365. 153. Sordinas1968. 154. Kourtesi-Philippakis 1996.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

Io9

site Figure3.42. The Palaeolithic of andits setting.The site is Ayia locatedin a smallloutsaat an elevationbetween300 and400 masl. Hatchedareasaremodernsettlements. blance to the Pontinian of Italy, a littoral Mousterian rich in small side It scrapersfound both in caves and on open-air sites.155 should be noted that these rather simple tools made on small pebbles are not very diagnostic and may reflect similar choices of raw materials for toolmaking rather than similar cultural traditions. It is notable, however, that this type of Mousterian is found only in coastal localities, where larger sizes of raw materials are also available,suggesting that the similarity of the industries may in fact be significant. The Rodaki artifacts are small in size and made from pebbles collected from the nearby riverbed.The most characteristic transverseconvex scrapers, bladelikeflakes, types aresmall core-choppers,'56 and rare end scrapersand notched pieces. This industry does not use the Levallois technique and we regard it as a specialized coastal facies of the Mousterian.

155. Kuhn 1995, pp. 46-72. The Italian sites are dated to the earlyto mid-glacial(ca. 110-35 kyr B.P.). 156. Cf. Kuhn 1995, pp. 95-97.

IIO

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

5 4 6

Figure 3.43. Middle Palaeolithic (Mousterian) artifacts from Ayia: 1) double convergent side scraper; 2) Levallois flake; 3) end scraper; 4, 6) blades; 5) core on a flake.
Scale 1:2

Figure 3.44. Middle Palaeolithic (Mousterian) artifacts from Ayia.

Scale 1:2

EARLY

STONE

AGE

OF THEOF NOMOS THE

OF PREVEZA

III

Figure 3.45. Palaeolithic findspots in the vicinity of Kastrosykia. In this area of heavy vegetation, paleosols and redbeds are exposed only sporadically. Findspots represent small concentrations of lithics, probably remnants of ephemeral campsites. Individual Palaeolithic artifacts were noted in tracts and walkovers throughout the area, indicating that many more findspots exist. Hatched areas are modern settlements.

Figure 3.46. Anavatis site/scatter 94-13, located in the middle of the photograph on the leveled area, looking northeast

II2

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Other veM
-ou'/ta:s

smasites are__
.-__I
a

:u

ibot.

Figure 3.47. View of Rodaki (SS92redbed (left) and

Ui
. ; .

Ammoud!a,!l sites ~_, or. :raw

";ak~-'~5.Pleistocene Kyr!i~'~i '' '"

: numbers _smfall! and~~~~_~ . *..i = ;i

3-ot

i,~~ .~

-,

marine deposits (fight) are separated ;by a vertical normal fault. The Palaeolithic artifacts were found in a Z * | Pleistocene deposit in the foreground, which is overlain by the

redbed.1:2 Scaleistocene

side scraper;below, bladelike flake.

Ot

very small sites arefound her

both in the Aheron (e.g., Tsouknida,

Thesprotiko (Iliovouni, Romia, Mesaria, and Galatas) valleys that are candidates for specialized sites, perhaps hunting stands, seasonal camps, kill
f numbers o Mous imall artifacts, typically fewer than 20 terian sp ecimens,

a all within few hours walking distance from major sites such as Alonaki, Kokkinopilos. Morphi, and people. Scattered flakeswere found on former movement of Mousterian islands in Lake Mavri (Thesprotiko) and in the remote mountainous polje of Cheimadio. It is difficult to interpret these small findspots, which may be the disturbed remnants of now vanished sites, but it is reasonable to suppose that most of them represent ephemeral episodes of activity in the

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

II3

Mousterianpointswere collectedas strayfinds by Individual landscape. the generalsurveyteams at Eli, at severalplaceson the Ayios Thomas peninsula(up to five pointswere noted),at the mouth of the Ambracian These pointsare Gulf, andat KastroRizovouniin the Thesprotiko valley. clearevidenceof off-site humanactivity, hunting representing probably
losses.157

of The distribution existingsites mustbe usedwith caution,as most stablepoints in the landscape, are found in redbeddepositsat relatively Inthe only placeswheresiteswouldbe preserved. and areconsequently This and of areas showmuchevidence erosion disturbance. probtervening is with modwherethe surface covered lem is most acutetowardPreveza, the for and ernvegetation offersverylimitedopportunities observing Late Pleistocenesurface. show much activityin this period,and Sites such as Kokkinopilos wereprobably others(e.g.,the two sitesat Loutsa,Galatas) occupiedonly controlis not sufficiently at this time. Our chronological preciseto deterexwent out of use. Stratigraphic minewhetherothersites (e.g., Kranea) wherethe MoustfromAsprochaliko is available cavationevidence only ca. to eriancontinuesperhaps the beginningof the UpperPalaeolithic, 29 foundin layer14, perhaps The late Mousterian, B.P. (26,000 b.p.).158 kyr It a represents shorterperiod of occupationor less intensiveactivity. is the basedon a differentpatternof coreworking,one that deemphasizes use Levalloistechniqueandmakesgreater of diskcoresto produceshort, pointedflakes.Tool types changealso,with Mousterianpoints and side and transverse convextypes,becomingthe mostcomscrapers, particularly in mon. Although the size differenceshave been exaggerated the past, thereis a smallshift in the directionof smallertools.A similarchangeis after60 kyr B.P. Kuhnhas shown noted elsewherein Europe,occurring Mousterian Italywith a transof that this changeoccursin the Pontinian land It in andpresumably use.159 seems formation settlement lithics, pattern, would to be a reasonable foragingstrategies hypothesisthat Neanderthal changein the faceof globalclimatechange. In Epirus,the less abundant evidencefor the laterMousterian, especially on the smaller,more dispersedspecializedsites, may reflecta reare on sponseto climatechange.LateMousterians concentrated the larger, of Kokkinopilos Morphi,andthe Alonaki and morereliable, perhaps poljes well loutsa.All threewere also particularly positionednearrivervalleys of with access larger to The huntingequipvariety specialized plains. greater ment seen in the Mousterianpoints and leafpointsmay hint at an increased reliance on hunting.160
157. Cf. Runnels 1996 for a similar pattern. and 158. Bailey,Papaconstantinou, 1992. Sturdy 159. Kuhn 1995. 160. Mellars 1996, pp. 193-244. 161. Runnels 1988; Runnels and van Andel 1993a. 162. Mellars 1992.

The available evidencesuggeststhat the last Mousterian widewas in GreeceduringOIS 3.Thereis another of feature the noteworthy spread In sitesthisindustry showssigns laterMousterian. the stratified Thessalian of contactwith and borrowingfrom the EUP (Aurignacian) tradition, presentin the neighboringregionsof the Balkansca. 30-45 kyr B.P.161 Suchmixedindustries, traclearlyderivedfromlocalMiddle Palaeolithic in with the Aurignacian the Balkansand ditions,overlapchronologically where they are westwardinto Franceand Spain (e.g., Chatelperronian)
It regarded as the work of late Neanderthals.162 is particularlydifficult to

II4

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

sort out these industries in the absence of controlled, well-dated, stratigraphic excavations,but the Greek version, as seen in Thessaly, is clearly of It Middle Palaeolithic character.163 has Mousterian points, leafpoints, and side scrapers,sometimes made on Levallois flakes and worked with typical Mousterian oblique, scalar,steep retouch.The EUP elements areend scrapers, carinated burins, and marginally retouched blades, all made from the same raw materials as the associated Mousterian pieces. It is noteworthy that this industry is still found at the top of the sequence of river deposits where typical Mousterian artifacts were discovered in a paleosol with an associated date of 31 kyr B.P.(28,000 b.p.).'64If this late Mousterian was a product of Neanderthals, it is an indication that they continued in existence, in Greece at least, for some time after they had been replaced by anatomically modern humans elsewhere in the Balkans and centralwestern Europe. As in Thessaly,the Mousterian continued at Asprochaliko It until quite late.165 is possible that the splintering of Neanderthal populations into isolated refugia by the intrusion of modern humans through the heart of Europe may have contributed to their eventual demise by preventing interbreeding and disrupting ancient patterns of migration and communication. There is another possibility for European Neanderthals. If the dates for the earliest Aurignacian in the Balkans get pushed back fartherin time to ca. 45 kyr B.P.or more, the overlap with the late Mousterian peoples becomes greater.Movements of modern humans, the presumed makers of the Aurignacianindustry,into territoriesonce occupied exclusivelyby Neanderthals could have caused the displacement of the latter.166 NeanderThe thals may have been confined to less favored reaches of Greece as a consequence of finding more northerly parts of the Balkans too cold or already occupied by anatomically modern Homo sapiens. Sometime after 31 kyr B.P.,the Mousterian and thus the Neanderthals were gone from Greece.
THE UPPER PALAEOLITHIC

A small number of sites belonging to the Upper Palaeolithic are known in Epirus, primarilyfrom excavations by Higgs, Bailey, and a team from the Ephoreia of Caves and Paleoanthropology. Stratified Upper Palaeolithic (UP) sequences in Epirus, including Asprochaliko, Kastritsa, Klithi, and Boila, and one site in Corfu (Grava Cave) have sequences of UP layers with Gravettian and Epigravettian industries.167 Radiocarbon dates indicate that the Upper Palaeolithic began before ca. 34 kyr B.p.168 Curiously,evidence for the initial stages of the EarlyUpper Palaeolithic is very rarein Greece.169 her review of the evidence Perles noted only a In small sample of EUP artifacts from the basal layer at Franchthi Cave that may date to more than 30 kyr B.P.,and she drew attention to possible Aurignacian elements in the unpublished sites of Arvenitsa, Kephalari, and Ulbrich in the Argolid.170 recent years, new EUP finds have been In forthcoming: in Thessaly, Aurignacian artifactsoccur in a late Mousterian industry found in sites in the Peneios River valley west of Larisa;71a similar industry was investigated at two open-air sites in the northwestern and Peloponnese near Patras;172 Koumouzelis has reported a late Aurignacian level with an estimated age range from ca. 34-24 kyr B.P. from a

163. Runnels 1988. 164. Runnels and van Andel 1993a. 165. Bailey,Papaconstantinou, and Sturdy1992. 166. Mellars 1992. 167. Bailey 1992, 1997; Bailey et al. 1999; Kotzambopoulou, Panagopoulou, and Adam 1996; Sordinas1969. 168. Bailey et al. 1983b; Bailey, and Papaconstantinou, Sturdy1992. 169. Runnels 1995. 170. Perles 1987. 171. Runnels 1988; Runnels and van Andel 1993a. 172. Darlas 1989.

EARLY

STONE

AGE

OF THE

NOMOS

OF

PREVEZA

II5

Figure3.49. EarlyUpper from end Palaeolithic scrapers

Spilaion. Scale1:1

In rockshelter site in the Kleisoura Gorge in the Argolid.173 the nomos of EUP artifacts are extremely uncommon. They are lacking in Preveza, but a few artifacts of EUP type (chiefly end scrapers) Asprochaliko,174 were collected at findspots in the survey area (e.g., Galatas and Vouvopotamos). Definite EUP artifactsof Aurignacian type were found in abundance at only one site, Spilaion, which is located on a limestone ridge between the Early Palaeolithic site of Alonaki and a former channel of the Acheron River. Spilaion has a large and dense accumulation of lithics on its southeastern slope. The extraordinary abundance of lithics on the surface (ca. 150,000 pieces) permitted a detailed spatial analysis, including a controlled collection from gridded sample sites and computer-assisted analysis of the distribution and association of the lithics (see Chapter 4). The finds are typical Upper Palaeolithic, including carinated and nosed end scrapers, burins, and retouched blades (Fig. 3.49). Spilaion is undated, but the site was occupied long enough to accumulate numerous concentrations of flintknapping debris marking positions of prehistoric activity. The site is strategically located at a point where routes to the north (via the parallelvalleys from Preveza to Parga) cross those running east-west, from the coastal plain to the interior.The EUP people seem to have had little interest in the poljes and loutses that determined Middle Palaeolithic settlement. The concentration of activity at Spilaion suggests that we are dealing with a base camp. The sheer density of artifacts,including concentrations of debitage suggesting episodes of flintknapping, and the scarcity of retouched tools are the best evidence for a degree of sustained and repeated activity.The analysis of the retouched tools and the flintknapping concentrations suggests a range of activities that are expected in a base camp, shown by the more or less complete reduction sequence of stone debitage (cores, cortical pieces, blanks, tools, and debris). In the Early Upper Palaeolithic there seems to have been little interest in caves or rockshelters, here or elsewhere in Greece, and the absence of Aurignacian deposits in the stratified and excavated sites in Epirus complicates the task of interpreting settlement patterns. It is nevertheless clear that the Aurignacian site at Spilaion represents a complete breakwith the Early Palaeolithic pattern of dispersed settlement and land use based on the exploitation of loutses and poljes.

173. Koumouzeliset al. 1996; Kozlowski1999. 174. Adam 1989, p. 253.

II6

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Figure3.50. LateUpperPalaeolithic 1:1 backedblades.Scale With increasing frequency Late Upper Palaeolithic (LUP) backedblade industries(Gravettianand Epigravettian)dating to 29 kyr B.P. (26,000 b.p.) and after are found in Greece. Backed-bladelet industries are found principally in caves or rockshelters, several of which have been tested by excavations.Asprochaliko,Kastritsa,Klithi, Boila, Grava,Franchthi,Kephalari, Kleisoura, Seidi, Theopetra, Ulbrich, and Zaimis are the chief examples, and other sites have been tested in Boeotia, Elis, and Thessaly.175 Intensive systematicsurveysin the Argolid, Berbati,Nemea, and Messenia, however,have produced surprisinglylittle evidence for LUP open-air sites. To take but one example, the surface survey in the vicinity of Franchthi Cave brought to light only a handful of LUP artifacts;the three or four sites with small geometric tools and backed blades were all small, seriously Site E81, for instance, disturbed by subsequent erosion, and undated.176 had only a single backed blade and other, isolated finds of backed blades were made in tracts, perhaps lost as the result of Upper Palaeolithic hunting activity.A similar situation was noted in the Berbati-Limnes survey, where the only LUP materials were scattered backed blades or end scrapers, found in the course of tractwalking and no doubt left by the hunters who occupied the Kleisoura shelters.177 In the Nikopolis survey,small numbersof LUP materialsof Gravettian or Epigravettian type (Fig. 3.50) were noted at two sites near the village of A Loutsa, at GalatasinThesprotiko, and at Kokkinopilos.178 typicalfindspot of this period is located near Lake Pogonitsa on the Ayios Thomas peninsula, where a scatter of half-a-dozen artifacts was found in small pockets of sediment in cracks in the karst limestone. The difficulty in relating these scattered finds to a pattern of land use and settlement is compounded by two factors. The lower sea level at the time of the last glacial maximum created an extensive coastal plain that greatly enlarged the useful territory.It was also a habitat supporting biota not found in the highland interior or represented by any existing habitats A on the mainland today.179 second problem, noted by Bailey, is that late glacial foragers required large exploitation territories,while today we see only a small portion of this territory in the small areas covered by surface The LUP settlement pattern was probably hierarchical,with a surveys.'80 network of sites serving as home bases and special activity sites. This hierarchy could extend from an ibex hunting camp in the mountains (e.g., Klithi), to seasonal bases in the upland basins (e.g., Kastritsa), to winter

175. Bailey 1992, 1997; Bailey et al. 1999; Kotzambopoulou, Panagopoulou, and Adam 1996; Koumouzeliset al. 1996; 1996; Kyparissi-Apostolika Perles 1987; Runnels 1995; Sordinas 1969. 176. Jameson,Runnels,and van Andel 1994, pp. 335-340. 177. Runnels 1996. 178. Dakaris,Higgs, and Hey 1964. 179. van Andel 1989. 180. Bailey et al. 1983a.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

II7

Without takbase campsat lowerelevations(e.g., Grava, Asprochaliko). in the into consideration, smallscatters ing the largesize of this territory anyone positionareunintelligible.'81 in for The only candidate a LUP site of anyconsequence the nomos where faunalremainsof ibex, deer,elk, and of Prevezais Asprochaliko, But are aurochs evidenceof the chiefpreyof localforagers. the smallnumwith and less dense depositsat this site when compared ber of artifacts was basinareevidencethatAsprochaliko neverIoannina in Kastritsa the Whether hierarchical site thelessa highlyspecialized in a larger, system.'82 basecampor a way stationbetween was or not Asprochaliko a residential it the plainsand the mountains, is still the most likelyfocus for the scatof whichmaybe remnants small,specialized teredsurvey camps materials, The of or huntingstandsutilizedin the exploitation the territory. populalimitedto a tion in the entirenomosof Prevezaat this time was probably basis.The singlesmallbandof ca.25 to 75 personsresidenton a seasonal in of nearinvisibility the Late UpperPalaeolithic the surveyareacan be the by explained the smallsize of the humanpopulation, limitedterritory and pattern archaeologically, the evidentshift to a settlement investigated someof which base on centered residential campsin cavesandrockshelters, werelocatedoutsidethe limits of the studyarea.
POST-PLEISTOCENE SETTLEMENT HISTORY

181. Bailey 1992; Bailey et al. 1983a. 182. Bailey et al. 1983b. 183. Bailey 1992. 184. Higgs and Vita-Finzi 1966. 185. Sordinas1969. 186. Sordinas1970. 187. Petrusoet al. 1994. 188. F. HarroldandJ. Wickens (pers.comm.).

The evidence suggests that the principal LUP sites in Epirus were abandoned at the end of the last glacial (ca. 10-13 kyr B.P.). Klithi, Grava, and Kastritsa have no record of occupation in the immediate post-glacial period.183 Higgs noted a disturbed and mixed upper layer at Asprochaliko in that is sometimes called "epipalaeolithic" the literature,but this undated level, described as having "backedblades and geometric microliths,"is just as likely to be Late Upper Palaeolithic as Epipalaeolithic.184 There are,however,two excavatedMesolithic sites near Epirus. Sordinas excavated a Mesolithic coastal site at Sidari (Corfu) dated to ca. 8.5 kyr B.P. (7.8 kyr b.p.), and he noted the many differences in lithic technology, raw material,and subsistence activity at that site when comparedwith the backed-blade industries of Upper Palaeolithic Grava Cave on the same Sidari is an open-air coastal midden site characterizedby extenisland.185 sive use of marine resources and by a microlithic industry based on atypiSordinas regarded Sidari as a new cal trapezoidal fragments of flakes.186 settlement by people arriving on the island by sea. The site of Konispol Cave (near the southern border of Albania) is close to the Epirote and Corfiote sites, none of which is more than 70 km from another.The excavatorsof Konispol found tracesof LUP occupation, abovewhich is a Mesolithic deposit up to 0.90 m thick dated from 8-8.4 kyr B.P. (7-7.6 kyr The Mesolithic industry consists of small flakes and blades with b.p.).187 intensive minute retouch. There are many composite tools and also denticulates, notches, pergoirs, and end scrapers.Microlithic trapezes are present, made by retouching segments of flakes and blades. The fauna include ibex and lesser quantities of aurochs, elk, and pig. The site was apparently a temporary shelter used by seasonal hunters on an episodic It basis.188 is unclear how the Konispol Mesolithic industry compareswith

II8

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

'Mesolithic <ii .'";:,. !^j^^^-,

. 20a ?o 2).
? O.Vathy ,), .Vt

-^.
^7

and Figure3.51. Palaeolithic in site/scatters the materials Prevezaarea.Palaeolithic

are found throughout this area, but

recenterosionmakesit difficultto
20

identify specific sites. The Middle

". '

^?)fi||^^^^^^ : / PREVEZA

DPalaeolithic

symbolsmarkareas wherematerials most abundant. are

One small findspot near Lake Pogonitsa (T93-12; W93-2) may be Upper Palaeolithic in age. Three Mesolithic sites (Tsarlambas and two Alonaki Beach sites) are located on the coast west of Preveza. Hatched

Site/scatter Beachc .t

Site s cat Thomasi

d _i/ .A) . '*l1-7Alonaki

jTsariambas

.* . ..* Meso..t..ic .

Lake Pogonitsa

'

x U.?Palaeolithic * M.Palaeolithic Coastalplain

2km

areasaremodernsettlements.

that of Sidari,where a somewhat different lithic technology is found. The chief difference is that the Mesolithic industry of Sidari makes greateruse of flakes than blades, a characteristicalso of the contemporaryupper levels at Franchthi Cave. Six sites in the Preveza nomos are possibly Mesolithic (Figs. 3.32, 3.51). One site may be earlierthan the others, correspondingto the Lower the Mesolithic at Franchthi Cave;189 other five are later.An IRSL date of ca. 7-10.5 kyr B.P. for Alonaki Beach (Table 3.10) is in general agreement with the dates for the other Mesolithic sites in Greece. A total of 731 lithics were collected from these sites by the same method employed for sampling Palaeolithic sites, namely retrievalof all cores, blades, retouched tools, and complete flakes, discarding only fragments of debris without recognizable features (namely, platforms, bulbs of percussion, or retouch). Three sites are located at the western end of the Acheron valley (Fig. 3.32). The site of Ammoudia (SS92-21) is on a low limestone ridge north of Ammoudia Bay, directly on the present shoreline. Tsouknida (SS92-8) is located in a Pleistocene alluvial fan (ca. 50 masl) on the edge of the valley overlooking what was in antiquity a marsh or lake.The site of Loutsa (SS93-32) is on a limestone bluff overlooking Tsouknida and the Acheron shoreline north to Parga. All three lie directly on thin Pleistocene soils, and the artifacts on the surface show evidence of burning and weathering suggesting that they have been exposed on stable surfaces for some time. Artifacts are abundant on all three sites. The chief reasons for regarding them as Mesolithic are the characteristic typology of the lithics (e.g., bladelets, end scrapers,bifacially-retouchedpieces, pieces with silica gloss, and trapezes) and the lack of sherds. The site of Tsouknida produced a few Middle Palaeolithic flakes in the upper part of the alluvial fan where it is dissected by ravines. The Mesolithic artifactsare small in size and, while patinated, are not as heavily weathered as the Palaeolithic pieces (Fig. 3.52:1-6). The assemblage of

189. Perles 1990.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

II9

190. Perles 1990. 191. Koumouzeliset al. 1996; Runnels 1996.

nearly200 pieces consistsof small flakes,often retouched,struckfrom cores(ca.20%of all pieces).Bladeletsarepresentin smallnumglobular bers.The predominant typeis a short,steepconvexend scraper, tool sometimes createdopportunistically retouching a partof one edge of a by only flake.There aremanytruncated notchedpieces and one smallovate and retouchedpiece.Many of the artifacts have one or more edges bifacially modifiedby minutenibblingretouch. In generalthe industry resembles LowerMesolithicat Franchthi the is Cave,where the transitionfrom the Upper Palaeolithic markedby a in of largereduction the percentage backedblades(fromas muchas 60% to about 3%) and the disappearance microlithsmade by microburin of which arereplaced flakeswith minuteretouch.190 There are technique, by alsoend scrapers notchedpiecesin the Franchthi and Mesolithic.A similarindustry datedto ca. 10 kyrB.P. beenpublished has fromrockshelters in the ArgiveKleisoura.'91 The two otherAcheronsites,Ammoudiaand Loutsa,arequite differentfromTsouknida. Ammoudiaartifacts, At many of which are less than 5 mm in size, arefound in a small area,ca. 50 x 25 m. Retouched tools (ca. 10%of the total)aremadeon snappedandbrokenfragments of bladelets flakesthathavebeen shapedby fine,nibblingretouch(Figs. and 3.52:7-11, 3.53). The flintknapping technologyis similarto Tsouknida, but with few end scrapers. Retouchedtools at Ammoudiainclude trabackedblades,and smallnumbers of pezes, a singlepossiblemicroburin, and retouched notches,denticulates, were pieces.A total of 173 artifacts foundin a paleosolca.50 cm thickoverlying bedrock cutby coastal the and erosion.A curiousstone structure foundon the site (Fig. 3.54). It conis sists of a semicircular 8 with a wall ca. 1-2 m foundation, m in diameter, thick.The structure associated is with piecesof burneddaubwith impressionsof caneorwood,butwithoutexcavation is difficultto determine it if it is an ancientor modernfeature. The Mesolithicsite of Loutsais foundin a small(70 x 50 m) areaof residual terrarossa,less thanone meterthick,restingon a karstlimestone surface an elevationof 200 masl.The terrarossahasbeen erodeddown at to bedrockon its edgesbut hasbeenprotected untilrecently a coverof by scrubvegetation. Erosionwas accelerated the last fiveyearsby the conin structionof a roadon its southernedge. Artifactsare erodingfrom the sedimentsand the assemblage 135 pieceshas a largenumberof cores of The numberof retouched (16.8%). piecesis high (ca.21%),with notched anddenticulated and tools,endscrapers, pieces,composite trapezes, per,oirs in the collection(Fig. 3.55). There are four retouchedflakeswith silica of gloss on their edges. Pieces of burneddaubwith impressions cane or wood werefoundin the sameareaas the lithics. The remaining Mesolithicsites areon the coastwest of Prevezaand can be consideredtogether(Fig. 3.51). They are found directlyon the in mature presentshoreline rubefied sands,the Bt horizonof a moderately paleosol(MS 2) overlainby youngeractivedunes (Fig. 3.56). Development of summerhousingand densevegetationlimited the searchto the exposedsea scarpand areasbetween construction zones; sites were detectedby lookingfor lithic artifacts embeddedin the top 0.20-0.80 m of redsandypaleosol(Fig. 3.57).The underlying fossil duneswereonlyvis-

I20

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

I?

goS
7 8 9

10

11

Figure 3.52. Mesolithic artifacts from Tsouknida (1-6) and Ammoudia (7-11): 1,2,4,5) end scrapers;3) end scraper and perfoir on a retouched flake; 6) bifacially retouched piece; 7) bifacial straight truncation and left oblique becon a flake; 8) truncated flake with small burin (microburin?); 9) flake with abrupt truncation; 10) backed bladelet with double oblique truncation; 11) trapeze. Scale1:1

Figure 3.53. Mesolithic trapeze from

Ammoudia. Scale 1:1

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

I2I

Figure3.54. View of Ammoudia,

feature visibleat left (middle ground)

go
.... .....

ible wherethe overlyingsandhad been removed. The paleosolcrumbles

quickly when it is attacked by wave action and rain, exposing the flints embedded in the upper part of the soil horizon. It should be emphasized that flints were seen at other localities in the sea scarp where the surface was not obscured by dunes or buildings. As dunes are found from Mytikas to the Paliourias River, that area is likely to contain additional sites. Tsarlambas (SS94-19) is the most disturbed of the sites and consists of a small wave-cut paleosol scarp.The other two sites, at Alonaki Beach (SS94-22, SS94-23), are richer. One of them (SS94-22) produced two bifacially retouched projectile points in addition to a good trapeze with a retouched truncation (Fig. 3.58:7-9). The artifacts are found in a semiconsolidated Bt horizon of a paleosol (MS 2). At least five sherds were noted at SS94-22 that may be prehistoric, and it is likely that there is some mixing of periods on the surfaceof the old Bt horizon. TL and IRSL dates of 7-10.5 kyr B.P. for this site were obtained from samples taken from within the Bt horizon. The second site at Alonaki Beach (SS94-23) had only recently been exposed by deflation, and samples for TL and IRSL dating were taken from the paleosol that contains the lithics (Fig. 3.58). One trapeze and one piece with silica gloss were collected from the site and two other geometric pieces were noted in a subsequentvisit. Although this was the smallest findspot we noted (ca. 25 x 25 m), the materials are very similar in typology and technique to those from Tsarlambas and the other Alonaki Beach site, and the three sites may be contemporary. One other small accumulation of flakes was detected by a general survey team on the summit ofTourkovouni, at the southeast extremity of the Ayios Thomas peninsula. A large sample (65 artifacts)was collected from the bare karst surface, but the artifacts are difficult to classify with any confidence. All are small plain flakes of light brown flint. One has been retouched to form a minute end scraper.No other prehistoric materials were noted on this peak during this or subsequent visits and the curious

I22

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

7
Figure 3.55. Mesolithic artifacts

fromLoutsa:1, 2) cores;3-5) core andflakeswith silicagloss;6) trapeze;7) smallpoint on basally

retouched flake. Scale1:1

EARLY

STONE

AGE

OF

THE

NOMOS

OF

PREVEZA

I23

lookingsouthwest,with rounded moderndunesoverlyingBt horizon exposedin the foreground

'

. t..

. .. ............ ,.

: ;

materials and very small .andage. The raw mind..... ........................ size of the flakes bring to
. ... . .... .

artifactscatter nearPreveza (SS94the identical specimens from Mesolithic layer D at Sidari92 and the end exposed by removal of modem sand
dunes
?

. . .. ......... . . .

. a Mt

..... ..... .

those coastal sites west those found the coastal sites west of Preveza. The The scraper is similar to Preveza. e scraper similar at absence of cores and paucity of retouched tools, however, prevent us from

Sordinas ~.. 1970. ~~~~192~

192.Sordinas 1970.

his site is not included in our final t definitiveand making a attribution or Early Bronze Age brought new settlement-the long prehistoric record The transi tion an end. transition from without th e s of .occupation was atfrom the Mesolithic (post-Pleistocene sites ites evidenceof domesticated plants or animals) to theNe olithic(permanent with use domesticated villages with use of domesticated species) cannot be documented with our species) d ata. Of unknown in the Gre ece duration, M esolithic northwest survey have continued down to 8 t B.P. or after, overlapping chronologive may ha con tinued kyr o the cally with Early Neolithic settlements of eastern Greece, withwhich the f they appearto have had no contact. At some point the areaof the Nikopolis or at least the traces of such survey appears to have become uninhabited, habitationarearchaeologically invisible,and-until the very latest Neolithic or Early Bronze Age brought new settlement-the long prehistoric record of occupation was at an end.

I24

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

I ,

10

11

12

J-

13

14

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

I25

CONCLUSIONS
The physiographic characteristicsof Epirus, particularlythe wide variety and quantity of karst features such as poljes, dolines, and loutses, have created conditions that made the landscape particularly useful to early humans and served to preserve their relics to the present day. The tendency of karstdepressions to collect sediments from the surroundinglandscape and act as traps for aeolian dust is the key to understanding the fossil cultural landscape investigated in this survey. Ever since the survey of northern Greece by Eric Higgs, Epirus has been recognized as being unusually rich in Palaeolithic remains. It is no longer possible to attribute the Palaeolithic abundance here to a lack of systematic investigations elsewhere. A number of topographic and site surveys have been conducted since the 1970s, several of which targeted early prehistoric periods, and those efforts leave no doubt that some portions of Greece have little preserved evidence for Palaeolithic activity.193 An example of the disparity in numbers can illustrate this phenomenon. The Larisa district of Thessaly, roughly the size of the nomos of Preveza, has been investigated periodically for Palaeolithic materials by German, Greek, and American teams from 1959 to 1991.194 Despite the intensity of the survey methods, particularlyin the 1987-1991 survey,the total number of findspots is only around thirty and fewer than 1,000 lithic artifactswere collected. A more dramatic comparison with the Nikopolis survey,where a similar number of sites produced lithics 100 times more numerous, can hardly be imagined. Surveys of smaller areas in southern Greece (Argolid, Nemea, Berbati, Pylos) produced equally small findings, typically fewer than five sites per region, each producing fewer than 250 artifacts.195 disparity in numbers is surely to be explained by the variThe ety of geological contexts present in each region that have affected the preservationof sites and artifactsand perhaps a priori the density of settlement too. The karst depressions in the mountainous tracts of Epirus have concealedand preservedlargenumbersof Palaeolithicmaterials,while active erosion in the Peloponnese and central Greece, at times caused or accelerated by human activity,has destroyedmany sites.196 northeasternGreece In the cycles of aggradation and incision of the great rivers complicates the picture by reworking, removing, and burying older sites.197 The karstfeaturesof Epirus createdan attractiveenvironment for early humans. The poljes and loutses filled with sediment and those depressions supported marshes, swamps, and lakes. The swamps and lakes may have been permanent, seasonal, or episodic, and sometimes they disappearedas uplift forced new stream systems that drained them temporarilyand eventually for good. The Epirote system of poljes and loutses is dynamic, creating a mosaic of small environments that concentrated important resources at precise and predictablelocations. The most important of these resources was water,which in turn supported lush vegetation and wildlife. In the dry season, which extended for six months from spring to autumn, these reservoirs of water attracted birds, terrestrialanimals, and humans. Lastly, as a consequence of the dissolution of the surroundinglimestone, larger quan-

Mesolithic Figure3.58 (opposite). fromthe Prevezaarea. artifacts AlonakiBeach,SS94-23:

1, 2) truncated flakes forming

piercingtools;3) flakewith minute nibblingretouch.AlonakiBeach, SS94-22:4) flakewith minute

nibbling retouch; 5-6)perfoirs;

7) tranchetarrowhead; tanged 8) arrowhead; trapeze;10) truncated 9) flake;11) bladelet;12, 14) microliths; with silicagloss 13) bladefragment on rightedge.Tsarlambas, SS94-19:
15) obsidian core. Scale1:1

193. Runnels 1995. 194. Runnels 1988; Runnels and van Andel 1993a. 195. Runnels 1995, 1996. 196. van Andel, Zangger,and Demitrack 1990. 197. Runnels and van Andel 1993a.

I26

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

tities of flint, often in the form of high-quality nodules, were concentrated near the red deposits or in the streams leading to and from them. The evidence from Greece so far tends to resemble the "late entry" A model of European colonization.198 fossil cranium of an early sapient and from Petralona may date to ca. 200-400 kyr B.p.199 a lithic industry, of the same general age, has been found at sites near Larisa in perhaps The Thessalian lithics, and artifacts discovered recently in Thessaly.200 Macedonia by a team from the University ofThessaloniki,201are similar to the industry excavated in YarimburgazCave near Istanbul, which probWe ably dates to ca. 300-350 kyr B.P.202 can reasonablyconclude that humans were present in Greece by 300-400 kyr B.P. in sufficient numbers to leave a detectable archaeological signature. The occurrence of corechopper industries (often called Clactonian or Tayacian)with the Acheulean industry, which differs principally in having handaxes (bifaces), is typical of western Asia and Europe where they are stratigraphicallyinterspersed.203 The Epirote discoveries agree generally with this picture of late colonization, yet there are many unanswered questions that will require additional research. Are the heavy flake tools from Alonaki part of the late Acheulean technocomplex or are they a variant of early Mousterian? The question of the co-occurrence, here and much farther afield, of flake tool industries and industries with handaxes has still not been satisfactorily answered,and the presence of handaxes in Epirus serves only to remind us that we are dealing with a very small set of data. We cannot determine finally whether the differences in typology reflect temporal, environmental, or functional differences. Our conclusion with regard to the earliest Palaeolithic is that these hunter-gatherers appear to have been the first humans to appreciate the rich environmentalpossibilities offered by the poljes and loutses of Epirus. The archaeologicalevidence for human activity in the Middle Palaeolithic period is very rich. Our survey added as many as 20 new findspots to the list compiled in the 1960s by Higgs and his students. As we have shown, these findspots reflect specific activities that appear to have been carried out repeatedly in the same place in the landscape. Many, but not all, of these findspots are associated with fossil Bt horizons (paleosols) of considerable maturity and hence age. Luminescence dates obtained from aeolian silt grains in these soils indicate that they may be on the order of 85-90,000 years old and the Palaeolithic artifacts that are contained in them may be older still. This pattern of repeated logistical use of the same locations in the landscape is of considerable interest. A similar pattern of partiallylogistical land use in the Middle Palaeolithic was detected in the Argolid, but the small number of sites and artifacts hampered interpretation.204 The land-use evidence in Epirus is of much higher quality, partly because of the better preservation and partly because the fixed positions and long duration of the karstfeatureshelped shape the structureof human activity. Moreover, our search methods-designed to detect very small scatters of materials so that no findspot, however small, would be overlooked-allow us to reconstruct the use of the landscape in considerable detail. In our view the Mousterian sites appear to conform to a modified or partial lo-

198. Roebroeksand van Kolfschoten 1995. 199. Darlas 1995. 200. Runnels and van Andel 1993a. 201. K. Kotsakisand S. Andreou (pers.comm.). 202. Arsebuk1993, 1996; Kuhn, Arsebuk,and Howell 1996; Stiner, Arsebiik,and Howell 1996. 203. Bar-Yosef1998, p. 268. 204. Jameson,Runnels,and van Andel 1994, pp. 325-335; Runnels 1996.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

I27

and sites quarrying, butchgistical pattern. Specialized (e.g.,fortoolmaking, revisited and across interior the ering)weredistributed uplands repeatedly who or on a seasonal basisby foragers, wereNeanderthals, moregenerally archaic Homosapiens. speaking, behaviorhas been hotly debated,and The questionof Neanderthal Neanderthal bescholarsare dividedin theirviews abouthow "modern" believethat theywereessenhaviorshouldbe regarded.205 scholars Some who had little abilityto plan futureactivities foragers tiallyopportunistic in to of or to structure theirmovements the landscape takeadvantage prefor dictable resources. Otherscholars willingto admita limitedfaculty are behavior. This issue is a largeone and here is not the place to logistical issuesinvolved,but we believethat we can discussthe largertheoretical detectan essentially "modern" to the Mousterian cast of pattern landuse.206 The basis for this conclusionis the repeateduse of the poljes and loutsesoverlong periodsof time.At Kokkinopilos, Alonaki,Ayia,andthe Anavatisquarry sites (aswell as on Corfu),the largenumber artifacts of is If an indicationof the sustainedlevel of past activity. Neanderthals had acrossthe landscape, would expectto find we foragedopportunistically with specificlocations. materials spreadfar andwide and not associated Thereis additional evidencein the existenceof verysmallsitesthatappear to havehad specialized In functions. the Acheronvalleya findspotat the easternend (Skepasto) a specialized is site, quarry/flintknapping an interthatmayalsoapplyto two othersites(Valanidorrachi Ormos and pretation An of of in variability the Mousterian Epirus Vathy). analysis the intersite
by Papagianni supports this view.207

205. Mellars 1996, pp. 366-391. 206. Mellars 1996, pp. 245-268. 207. Papagianni1999. 208. Bailey et al. 1983b. 209. van Andel andTzedakis 1996; van Andel 1989. 210. Bailey,King, and Sturdy1993.

In a strictlylogisticalland-usepattern, base campswouldbe used as activities locatedacrossthe landscape, and but stagingsitesfor specialized no basecampshaveyet been identifiedin Epirus.Asprochaliko, only the excavated Mousterian as site, is best regarded a seasonalcampor hunting standratherthan a home base.208 high incidenceof retouched The pieces at almost everysite indicatesspecialization. lack of preservation The of faunalandfloralremains the absenceof recognizable and suchas features, burialsor fireplaces, remainsa problem. shorelineswere displaced During much of the Middle Palaeolithic, seaward the coastalplainsprovidedadditional and spacefor winterbase At times, however, coastalplainswere greatlyreducedin exthe camps. and tent,209 it would be unwiseto try to explainthe lack of base camps to of solelyby reference the submergence the continentalshelf.It is also thatbasecampsareunrecognizable did not exist.If the Mousor possible terianland-usepatternwas a modifiedlogisticalone, the Neanderthals that engagedin "residential mobility," is, they movedaboutthe landscape to maximizeforagingopportunities, wereforward-looking but enoughto theirmovesto take advantage watersources, of flint, and game that plan fromexperience locatedatcertain were siteswhich theyhadlearned specific they visited on a seasonalschedule.This patternof residential mobility resembles that documented the LateUpperPalaeolithic,210 differs for but fromthe LUP modelby beingcentered a set of predictable on waterholes in the loutsesandpoljes,rather thanon vegetationcommunities governed The morespecialized LUP modelwasperhaps by bedrock. typicalof Late Pleistoceneforagers.

I28

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

Kuhn has developed a model of Mousterian foraging for Latium (Italy) In that may also be applicable to Epirus.211 Kuhn's model the different provisioning of raw materials,which is reflected in the Mousterian industries in a number of ways, can be used to reconstructthe degree of residential mobility of Mousterian foragers. Kuhn argues for rational logistical behavior on the part of Neanderthals,212 describes a pair of strategies and that may have been in use at different times during the Middle Palaeolithic.213 One strategy emphasized scavenging over hunting and required greatermobility of individuals over their territory.Mobile individualswere provisionedwith retouched tools and Levallois flakes as supportsfor transported artifacts.In this mode, cores areefficiently used and tools areheavily retouched and often recycled. In the second strategy,hunting of live game permitted groups to maintain longer residences in base camps where raw materials could be stockpiled in the form of nodules and cores and worked on site as needed. The resultwould be large concentrations of flintworking debris at relativelyfew sites, with the artifacts and cores showing little use of the Levallois technique and less intensively worked. These different strategies of foraging are interchangeable. Kuhn nevertheless detects a definite chronological pattern in the Italian data, with a tendency for the "hunting strategy"of intensive use of a smaller number of residential sites to predominate after ca. 55 kyr B.P.214and both strategies to be tied to fluctuations in sea level and the resulting availabilityof resources. Kuhn stresses that this pattern,while evident in the data from Latium, need not apply to the rest of Italy, much less the Mediterranean. But we believe that his general conclusion-that the Mousterian people "behaved over the long term in an economically rational manner, adjusting patterns of tool manufactureand use to fit problems inherent in different patterns of land-use"-has wider relevance.215 The Mousterian period in Epirus is of unknown duration but appears to extend from at least the last interglacial (ca. 115-130 kyr B.P.) to a point late in the late glacial (ca. 31 kyr B.P.). Although climatically it spans a warm interglacial,with a long slow decline to a first, rather modest glacial maximum (between ca. 70 and 60 kyr B.P.), followed by an equally long intermediate interstadialclimate, it was evidently a time of cultural stability and continuity that may be divided roughly into two distinct phases, similar to those defined by Kuhn. In the first phase, the Mousterian was based on the production of large flakes and blades, often by means of the Levallois technique,which were modified into points, scrapers, denticulates, and other tools, and would correspond to Kuhn's scavenging strategy.In the second phase, which may have begun around 60 kyr B.P., there was a shift to smallerless carefullyworked flake blanks struckfrom non-Levallois cores (disk and Mousterian cores), corresponding to Kuhn'shunting strategy. The flakes were used to produce a wide range of side scrapers and points. Stiner and her colleagues have noted that the evidence for population density in the early Middle Palaeolithic suggests that Neanderthals and archaichuman populations were very small and dispersed,216 making them hardto detect archaeologically. The same authorsalso observed abrupt very population density increases in the late Middle Palaeolithic and in the Late Upper Palaeolithic-Epipalaeolithic (Mesolithic), patterns which we

211. Kuhn 1995. 212. Kuhn 1995, pp. 174-180. 213. Kuhn 1995, pp. 36-37. 214. Kuhn 1995, pp. 157-183. 215. Kuhn 1995, p. 182. 216. Stiner et al. 1999.

EARLY

STONE

AGE

OF THE

NOMOS

OF PREVEZA

I29

have observed also in Epirus.The evidence for these increases comes from the analysis of faunal remains from excavated sites that point to growing populations which were compelled to intensify food collection and to pursue lower-ranked prey and other food sources. We cannot test their hypothesis with faunal data from Epirus, but demographic pressuremight be one explanation for the concentration in Epirus on the karst featureswith their dependable (if low-ranked) resources such as fish, birds, mollusks, turtles, and reptiles. Our data from surfacesites cannot be comparedwith that from a large number of excavated stratified sites, and chronologically meaningful patterns of land-use strategy are thus difficult to support. The incidence of Levallois Mousterian and large numbers of retouched tools at some small outlying loutses (e.g., Ayia, Kranea, and the Loutsa sites), however, may reflect periods of high mobility, perhaps linked with scavenging and the use of scattered small-scale resourcessuch as shellfish, aquatic animals and On birds, turtles, and the like.217 the other hand, sites like Kokkinopilos, Alonaki, and Morphi have enormous quantities of flintknapping debris that may reflectlonger-term residence at preferredcentralsites with a lesser degree of mobility and greateremphasis on hunting. No doubt there was a mixture of these patterns that resulted from the adjustment of the Mousterian people to the large, often sharp oscillations of climate experienced over the 70,000 years or more of the Middle Palaeolithic period.218 The principal results of our analysis are confined to the elucidation of the spatial land-use patterns that appearto have been broadly synchronic. The lack of a detailed chronology presents important difficulties, in particular the inability to determine whether any sites aretruly contemporary.
THE TRANSITION UPPER HUMANS FROM THE MOUSTERIAN AND THE TO THE ORIGINS OF

EARLY MODERN

PALAEOLITHIC

217. See Kuhn 1995, pp. 150-151,

168.

218. van Andel andTzedakis 1998. 219. Runnels and van Andel 1993a. 220. Harrold1993; Mellars 1996, pp. 392-419.

Asprochaliko Cave is the only site in Epirus where the transition from the Mousterian (Middle Palaeolithic) to the Upper Palaeolithic can be seen. There is a stratigraphichiatus between the Middle and Upper Palaeolithic levels in this site. A similar disconformity is seen in Thessaly where the Peneios River sites have a recognizable Mousterian that persists until ca. 31 kyr B.P., after which time it abruptly ceases.219 An interesting feature of the Thessalian Mousterian is the admixture of Mousterian tool types (side scrapers, points) and techniques (Levallois) with EUP types (end scrapers, burins, and retouched blades). This kind of mixed industry is found widely in the Balkans and westward in Europe (e.g., the Chatelperronian)where it is interpreted as evidence of the incorporation by Neanderthals of tool types and techniques as the result of contact with anatomically modern Homo sapiens.The existence of this late Mousterian has many implications for the debate surrounding the origins of modern humans.220 abruptdisappearance the Mousterian throughout Greece The of is evidence for their ultimate extinction because this disappearancecomes after a period of overlap of the Mousterian peoples with anatomical moderns. Greece was evidently one of several isolated geographic refuges for late Neanderthals. Here, as in Iberia (Zafaryya Cave; Lapedo) and

I30

CURTIS

N. RUNNELS

AND TJEERD

H. VAN ANDEL

Italy (Ulluzian), Neanderthal sites continue as late as 31 kyr B.P.or even

later, indicating the survivalof Neanderthalslong after anatomically modernhumanshad come to be the only human speciesin the rest of
Europe.221

The evidence fromEpirusmaysupport replacement the of hypothesis modernhumanorigins.Spilaion,at the mouthof the AcheronRiver, a is credible candidate a homebasewithits abundant for flintknapping debitage. The site is a largeandveryrich EUP site providing evidencefor a totally different are patternof landuse. Outsideof this site,EUP artifacts found at thatthesekarstfeaonly as smallscatters poljesandloutses,suggesting tureswerelittle used.
THE LATE UPPER PALAEOLITHIC AND THE MESOLITHIC

The uncertainty dating the transitionfrom the Earlyto Late Upper in before the establishment the stratifiedcave sites of the in Palaeolithic, better-knownGravettianand Epigravettianindustriesrich in backed The occurrence Gravettian Epiof and blades,is an unsolvedproblem. in industries Epiruscoincideswith a movefromopen-airsites gravettian to rockshelters cavesand a concomitant and changein land-usestrategy. The changein resource has exploitation been studiedin detailby Bailey andhis colleagues who havedeveloped modelof seasonal a of exploitation In and big game animals.222 their model, the cave sites of Asprochaliko Kastritsa functioned hunting as locatedstrategically control to camps, points of accessto limestoneplateaus wherethe largest numbers animals of would feed. This patternis fully logistical,a fact that can be supported the by compositionof faunalremainsand lithicsin the cavedepositsthat demonstratethe specialized activitiestakingplaceat each site. Baileyand his colleaguesare surelycorrectwhen they arguethat the LUP exploitation territories wereverylargeand includedcoastalplainsexposedduringthe lastglacialmaximum. shelf They locatethe homebaseson the continental and relegatethe known mainlandsites to the statusof subordinate seasonalspecial-activity in a regionalsettlementhierarchy, Higgs had sites as concludedearlier.223 changein exploitation The to strategy the pursuitof close following,and ambushtactics)and the largegame (by monitoring, inclusionof coastalplainsin the territory covered togetherexplainthe can marked differences betweenthe Middle andUpperPalaeolithic. sharply The Late Upper Palaeolithic a very short phase in Epirusand it is to The only appears haveended sometimebetween10 and 13 kyrB.P.224 firmevidenceforthe latestPalaeolithic the Nikopolissurvey comes in area fromAsprochaliko wherethe upperlayersare much disturbed have and not been fullypublished. The lithicsfromthe uppermost excavated layer datesandindicatean aban(asreported) agreewith the latestradiocarbon donmentof the caveat the end of the Pleistocene, conclusionsupported a the most recentexcavations Klithiin the Zagori. at by The transitionfromthe Pleistoceneto the post-Pleistoceneis difficult to interpret Greecebecauseof the smalldatabase. in The Nikopolis six surveyhas contributed new sites to those known for that period in and Greece,alllocatedcloseto the earlyHoloceneshoreline, it is probable

221. D'Errico et al. 1998. 222. Bailey,King, and Sturdy1993. 223. Bailey,King, and Sturdy1993; Higgs and Vita-Finzi 1966; Higgs et al. 1967. 224. Bailey 1992.

EARLY

STONE

AGE

OF THE

NOMOS

OF

PREVEZA

I3

thatadditional sitesarestillto be foundalongthe westerncoastof Epirus. The new sitesarefoundin two areas, AcheronRivervalleyandamong the the coastalduneswest of Preveza. of Despiteextensive searching the poljes and loutses,and other areas,no Mesolithicartifacts havebeen identified in these contexts. The Acheronsites havethe samelithic industryas the Prevezadune sites andresemble coastalmiddensite of Sidarion Corfu.Fourof the the to Ammoudia Epirotesitesarein geographic settingssimilar thatof Sidari. sits on a low hill overlooking sea at the mouth of the AcheronRiver the and the three sites west of Prevezaaredirectlyon the modernshoreline amongdunessituatedin an ancientcomplexmosaicof streams, swamps, coastallagoons,and dunes.Tsouknidaand Loutsa are in quite different the settings.Tsouknida, probably earliestof the sites,is locatedon a ridge that protrudes into the rivervalleyand may havelooked over an areaof marshesand swamps,while Loutsasits on a limestoneridgehigh above with a commanding Tsouknida view of the westernend of the valley. The lithicspoint to huntingas the mainactivityat all sites:the trapezes could havebeen used to tip arrows, while smallmultipurpose tools wereuseful formaintaining repairing and or reedsto build huntingequipment working huts or "papyrella"-type boats. simple A smallcore,evidentlyof obsidian, fromTsarlambas (Fig. 3.58:15)is a reminder that these smallbandsof huntersand fisherswere awareof a world.225 this maybe addedthe coastalorientation the Epirote To of larger andCorfusites,suggesting seafaring partof the Mesolithicadapthat was
tation.

THE CONTRIBUTION OF THE PREHISTORIC SURVEY TO THE NIKOPOLIS PROJECT

Southern record its karstmorphology and Epirushasa richarchaeological the lineaments a fossilcultural of in preserves landscape the sensethatthe hastracesof humanactivityassociated surface with still extantportionsof the landscape. The patternof ancientland use is seen at more than one scale,from the small but rich site of Spilaion,where the flintknapping is in activityof the EarlyUpperPaleolithic registered the staticstructure of the lithic scatter, the scaleof the entireprovince, to with its networkof loutses,andsites. poljes, In the end, how do we accountfor the richnessof Epirus? There are threemajorfactors. The firstis the prevailing weatherpatternthatbrings lush flora.To this westerlyrainsto the provinceand supportsa relatively factorwe can add the patchwork runoff-collecting of karstdepressions that owe theirexistenceto the presentactivetectonicsof the region. Thus the Epirotelandscape dottedwith marshy is basinsofferinga widevariety of essentialresources. fromeasternGreeceby Lastly,Epirusis separated the PindosRangeandfromcentral southern and Greecebythe Ambracian andCorinthian Gulfs.As a result,the inhabitants the regionhaveoften of been relatively independentand in culturalterms orientedmore to the Balkans the Italianpeninsula and thanto the restof Greeceor the Aegean.

225. The sourceof the obsidianis of unknown,but the visual appearance the materialresemblesthe "snowflake" obsidianfrom Italianvolcanic fields such as Lipari.

I32

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

The Balkans easilyreached following northwest-southeast the trendare by ing valleysof the centralmassif,and Italywas connectedby wide coastal plainsduringcold periodsandby the sea at othertimes.A sea passageto been saferthanthe moreperilous southern Italyvia Corfuhasalways voyto aroundthe Peloponnese the Aegean. age These factorshaveactedtogetherto drawhumanmigrants Epirus to in manyperiods,by land and sea, from Italy and southeastEurope.The isolationand rich ecosystemof the regionencourrelativegeographical aged stablehuman adaptations duringthe Palaeolithic period.We may addto this the redbeds thathavepreserved record earlycultural a of activand the archaeological richesof Epiruscan be explained without inity, or vokingdifferentintensitiesof research surveymethods. of How doesthe prehistory Epiruscontribute the understanding to of the largerissues confronted the Nikopolis Project? the most by Perhaps is declineof the hustrikingfeatureof the earlyprehistory the apparent man presencein the Upper Palaeolithic and Mesolithic.This tendency continuesinto latertimes,with the possibility that the nomosof Preveza was largelyuninhabitedduringpart of the Neolithic.The breakin the chainof humanpresencemayhavehad consequences laterperiodsby for the regionof nativeinhabitants possessionof long-standing in depriving traditions. This hypothesis too uncertain pursuefurther, another is to but maybe moreto the point.The recordis clearlyan indicationof the tenof dency in all periodsfor the inhabitants this provinceto look first to themselvesfor supportand inspiration, and secondlytowardsoutheast This lastfeature loomslargein the subsequent EuropeandItaly. historyof the regionand is returning prominence the presentage. to in One last point can be mentioned. Almost all the laterhistoricalsites thatriseto prominence afterthe end of the BronzeAge andthe beginning of the Iron Age, to say nothingof moderntimes, are situatedin or near thoseareas in waterandsoil andwhicharealsothosethathaveyielded rich the richestprehistoric finds.The emphasisfound in this reporton the of that is properties the landscape helpedto shapehumanbehavior surely usefulfor studentsof these historical periods.

EARLY

STONE

AGE

OF THEOF NOMOS THE

OF PREVEZA

I33

ACKNOWLEDGMENTS
A great many people participated in the Nikopolis Project and made contributions to our fieldwork and analysis, and others contributed to the preparationof this report.To all of them we are deeply grateful.We wish to thank especiallyJames Wiseman and Konstantinos Zachos for inviting us to participate in the project and for giving us aid and encouragement. We wish to thank Panayiotis Paschos of the Preveza IGME office, who was always generous and helpful in the field. We also thank Gillian Foreman, Chris Jeans, and Richard Powys at the University of Cambridge for the sedimentological analyses, and we owe a special debt of gratitude to Li-Ping Zhou, Andreas Lang, and Ann Wintle, who gave us much valuable advice in the preparationof the section on Late Quaternary chronology. Our thanks are due to the many Boston University field school students who joined us from time to time in the field. James Wiseman and Dimitra Papagianni read early drafts of this paper and their thoughtful comments were very helpful. Our greatest debt, and hence our deepest appreciation, is due to Priscilla Murray,who labored with us in the field from first to last and who contributed useful advice, penetrating insight, and some of the best finds.

I34

CURTIS

N.

RUNNELS

AND

TJEERD

H.

VAN

ANDEL

AND MESOLITHIC APPENDIX: PALAEOLITHIC IN THE SURVEY AREA* SITE/SCATTERS

Name Alonaki Lower Middle Upper SS/T/WNo. Palaeolithic Palaeolithic Palaeolithic Mesolithic x
x x x x

SS92-22 SS92-23 Alonaki Beach SS94-22 SS94-23 Ammoudia SS92-21 Anavatisquarry SS94-13 SS94-16 SS93-9 Ayia n/a Ayia Kyriaki T93-3-5,17 Ayios Thomas Cheimadio SS94-1 SS94-2 SS94-18 Eli SS92-19 Galatas SS92-13 SS92-17 Gymnon Iliovouni SS92-9 SS91-3 Kokkinopilos Koumasaki SS92-24 Kranea SS92-14 Lake Pogonitsa T93-12, W93-2 Loutsa SS93-31 SS94-12 SS93-32 Mesaria SS92-16 Ormos Odysseos SS92-25 Rizovouni SS92-11 Rodaki SS92-15 Romia SS92-10 SS92-20 Skepasto SS92-37 Spilaion SS92-18 Stephani Tsarlambas SS94-19 Tsouknida SS92-8 Valanidorrachi SS91-4 SS92-12 Vouvopotamos

x x x x x x x x x x x x x x x x
x x

x x x
x

x x x x x x x

x
x

x
x

x
x

*Only datablesite/scattersarelisted here.

CHAPTER

EARLY

UPPER

PALAEOLITHIC

SPILAION: SURFACE

AN ARTIFACT-RICH SITE

by Curtis N. Runnels, Evangelia Karimali, and Brenda Cullen

The site of Spilaion, a high-density lithic scatter located near the mouth of the Acheron River (Fig. 4.1), is a good example of a surface site that archaeologists often find difficult to interpret.1Spilaion does not have a cultural context in the usual sense and belongs instead to a class of sites that exist entirely on the surface with no extant stratification or cultural deposits. While it has long been recognized that some sites may be strictly confined to the present surfacewith no relationship to a stratified deposit, such sites are often dismissed as of relatively little value for interpreting the past.We believe that it is possible to extractvaluableinformation about past culturalactivities from two-dimensional spatial associations preserved on such sites. Our analysis of Spilaion'shigh-density artifact distribution demonstrates that these surface sites may retain substantial evidence of spatial distributions created by past cultural activities. The discovery and study of surface sites of all kinds is at the heart of intensive regional survey, regardless of whether sites are the manifestations of buried strata or phenomena existing only on the present-day surface. Careful study of high-density and low-density artifact scatters has demonstrated that they may preserve information about human activities in the past.2 As archaeologistshave come to appreciatethe richness of the archaeological record, various problems associated with explaining how concentrations of artifactson the surfacewere formed have been recognized. One major issue is that archaeological materials are not always distributed on the surface in well-bounded clusters, but may exist in patchy and largely
1. We want to thank the codirectors of the Nikopolis Project,JamesWiseman and KonstantinosZachos, for their supportand encouragement. We benefited also from discussionswith Kenneth Kvammein connectionwith the spatialanalysisand with Janusz and Paul Kozlowski,Ofer Bar-Yosef, Mellarsin connectionwith the lithic We industry. wish especiallyto thank Kael Alford, who oversawthe collecting of the randomsamplesin 1992. Tjeerd van Andel andJamesWiseman kindly read an earlydraft of this chapterand made many useful suggestionsthat we have endeavoredto address. Verylarge concentrationsof artifactsare found directlyon the surfacein many parts of the Mediterranean world, where they are sometimes called scatters,findspots,or highWe density distributions. preferto use the traditionalterm of "site" its sense in of "place" "location" a generalterm or as for any surfacescatterof artifactsthat has a well-defined spatialconcentration of culturalmaterials,but without the unnecessaryconnotationof "settlement"or "habitation" this term that sometimescarries. 2. Alcock, Cherry,and Davis 1994; Cherryet al. 1988.

I36
M /+I

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

Louros River

Acheron River

rp

.1.

v Spilaion

Arachthos _ River

Ionian Sea
I JR

3
~7,': ^ 4;:

e.
Ambracian Gulf -^^I

*f\

10

15

20

25 KM
??? ?

-Y

Figure4.1. Map showingthe locationof Spilaionat the mouthof the AcheronRiver

discontinuous spatial patterns.3This observation raises serious questions and underscores the need to regarding the traditional concept of "site,"4 refine field techniques and explanatory hypotheses that account for the The number of artifactsor the relative denformation of artifact scatters.5 sity of artifact concentrations is usually employed to define intra-site density thresholds and to convey the degree of a site's discreteness or boundedness,6but there is a consensus that the explanation of any surface site is a complicated issue that demands an understanding of the geological and cultural processes active at the site. This problem has been faced in various ways. One approach is to define new sampling and quantitative methods that will increase the accuracyof the recoverytechniques.7Another is to drawon ethnographicanalogies to explain how artifacts are distributed on the surface (e.g., field manuring).8Yet another attributes surface scatters to short-term cultural episodes, such as flintknapping,9the off-site storage of equipment by agriculturalists or pastoralists, or isolated activities associated with animal folds, milking pens, or dumps.10Other forces that are invoked to account for low-density distributions of artifacts include natural or anthropogenically induced processes such as soil erosion, deflation, or downslope movement.1l In one case study carried out in the Southern Argolid in the early 1980s, soil erosion was documented in every valley investigated and was cited as one agent for moving artifactsfrom higher elevations to low-lying parts of valleys and spreading them over the plains.'2 Evidence for the

3. Cherryet al. 1991;Wright et al. 1990;Wells, Runnels,and Zangger 1990. 4. Fotiadis 1992. 5. Cherryet al. 1991; Bintliff and Snodgrass1988b. 6. E.g., Cherryet al. 1988, 1991. 7. Kvamme1996. 8. Alcock, Cherry,and Davis 1994. 9. Kvamme1996. 10. Murrayand Kardulias1986. 11. Wells, Runnels, and Zangger 1990; Whitelaw 1991. 12. Jameson,Runnels,and van Andel 1994, pp. 172-194, 325-414; Pope and van Andel 1984.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I37

rearrangementof valley watersheds by anthropogenic erosion is also found in the Berbati and Limnes valleys in the northern Argolid, where it is cited as a major agent in shaping the archaeological record.13In the Corinthia, tectonics and manuring are cited as of equal importance to anthropogenic erosion in the formation of off-site scatters.14 From these examples, and the others cited above, it is evident that a diverse arrayof factors is responsiblefor scattering culturalmaterialsacross the surface of Greece and creating both low-density and high-density distributions. Although some surface scatters of artifacts are explained as the result of physical processes, we take the position that some surface sites, especially high-density concentrations of artifacts,preserve culturalinformation in the form of spatial patterning, typically of a kind that has only recently engaged the full attention of archaeologists. On the most basic level, if artifactswere once part of a culturalmatrixthat has been destroyed, their location in the landscape is nevertheless the result of cultural choice, and is of some use in reconstructing prehistoric settlement and land-use patterns.

TAPHONOMY

OF SURFACE

SITES

13. Wells 1994; Wells, Runnels,and Zangger 1990; Zangger 1992. 14. Wright et al. 1990. 15. Schiffer 1987. 16. Rick 1976. 17. Efstratiou1985.

The study of site formation processes has received serious and widespread attention since the 1970s and covers those natural and culturaloperations that are responsible for the patterning of materials found in the archaeological record.Although a systematic and comprehensive surveyof all processes involved in the formation of sites is far from complete,15much has been learned that is useful for interpreting high-density artifact concentrations on the surface. The downslope shifting of artifactsis one factor affecting surface sites on steep terrain. This process was studied in the Andes by Rick,16who found that the absence of topographic barriersor heavy vegetation on one such site permitted heavier artifactsto move downhill through time as the result of gravity and slope wash, leaving only small artifacts in situ at the top of the slope. Rick concluded that in places where slopes are steep the operation of gravity will tend to sort artifacts by size and mass, and that only smaller artifactswill preserve cultural patterning. Another factor affecting the spatial distribution of artifacts on the surface is deflation, a physical process by which wind removes the finegrained materials from a site, leaving the larger,heavier artifacts exposed on the surface.In a related process, wind and water can move surface artifacts from one place to another horizontally through sheet erosion. Coastal sites may be affected by rising sea or lake levels. Recent investigations of submerged sites have shown that the horizontal patterning of culturalmaterials may be retained even after the stratification of the site has been destroyed. At Ayios Petros, for example, a Neolithic site in the Aegean Sporades, Efstratiou demonstrated that the spatial distribution of potsherds and other artifacts in the settlement was preserved, despite exposure to the full force of winter waves as the sea passed over the site with the eustatic rise in sea level in the early Holocene.17

I38

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

for that scatters preserve evidence Support the hypothesis surface may of spatialassociations comesfroma long-termstudyof an Archaicperiod site in the desertsouthwestof the United States,a site similarto Spilaion in terms of its high-densitydistribution lithics.18 of Assumingthat the of the lithicswas at leastpartlydetermined cultural spatialdistribution by Kvamme the subjected site to detailedmapping,employinga patterning, Information the Geographic System(GIS)to analyze distribution. using By GIS in such an analysis,Kvamme was able to assessthe distribution of individual typologicalclassesof lithicsin relationto changesin topografeatures. combinaThe phy,vegetationcover,slope,andothergeographic tion of detailedmappingand spatialanalysisdetecteda well-preserved with numerousconcentrations "hotspots," or patternof culturalactivity, each consistingof a ring of largepieces of debitageon the marginsand small-sizedflakesin the center.Kvammeexplainsthis structure the as resultof the sortingof debitageby size that occursduringflintknapping, whichhe regards a fundamental as formation in lithic process high-density distributions. supports theorywith computer He his andobsermodeling vationsderived fromhard-hammer The flintknapping experiments. structure of the lithic distribution Kvammeto concludethat the artifacts led were depositedby humanswho repeatedly visited the ridge,wherethey andengagedin flintknapping toolmaking.19 and camped It is thus probable that the natural cultural and forcesresponsible for the formationof sites can be identifiedand distinguished from one anotheron the basisof geomorphological observations artifact and analysis. Even in those caseswhere a site has been seriouslyaffectedby erosion, there is reasonto bedownslopemovement,deflation,or submergence, lieve that the horizontalpatterning the artifactdistribution of maypreservesome culturally Our spatialanalysisof the significantassociations. lithicsite of Spilaionis intendedto demonstrate valuethat the prehistoric artifact-rich surface sites mayhavefor archaeology.

THE SITE OF SPILAION

The site of Spilaion(SS92-37) is a large,densescatterof artifacts distributed overthe surfaceof a low hillockat the westernend of the Acheron Rivervalley(Figs.4.2, 4.3). The site was discovered 1992 in the course in of the studyof Palaeolithic Mesolithicsites(seeChapter The name and 3). in side Spilaion(or"cavern" Greek)refersto a sinkholeon the northern of the hillock, whichis in facta limestoneoutcrop madeup of a highlyweatheredkarstsurface. The site is now situatedapproximately kilometer one fromthe presentcoast.In the Pleistocene, at timeswhen the sea however, levelwas lower,the site was muchfurtherinlandand overlooked gently a Riverflowed down to a coastal slopingvalleywhere the paleo-Acheron plain.At times of highersea level, in the Pleistoceneand the Holocene, the shoreline often closer. was The elevationof the site ensured that at all timesit had a commanding view of the entirearea. The presentkarstsurfaceof Spilaionhasa thin covering scrub of oak vegetation(evergreen and in olives)rootedin crevices the rockandsmallpatchesof sediments, chiefly relictPleistocenedepositsor terrarossa(Fig. 4.4).

18. Kvamme 1996;K. Kvamme

(pers. comm.). 19.Kvamme 1996.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I39

Figure4.2. Map of Spilaionshowing contours. The artifact topographic scatteris found on the southsoutheastslope.The squaremarks
the location of the 60 x 50 m sample

grid,whichis locatedin the approximatecenterof the scatter.

Figure4.3. View of Spilaion,looking southwest. The artifacts found in are the open areasvisiblein the foreground. .. J -

The site consists of a dense, continuous, and extensive scatter oflithic implements on the south-southeast slope of the hillock. We estimate that ca. 150,000 artifacts or fragments of artifacts are present within an area more than one hectare in size, with a density that averagedca. 15 artifacts/ m2.We measured the number of lithics on the site by counting the lithics present in 100 sample units (5 x 5 m each) and using this count to estimate the density of artifacts on the entire site. The limits of the site and the variable density of lithics were investigated by repeated walkovers to examine the surface visually. During the walkovers, the number of artifacts

was tabulatedby three persons with handheld counters, stopping to record counts at one-meter intervals. The majority of the Spilaion lithics are uniform in typology, technology, and raw material, and appear to belong to a single cultural component. A small number of Late Bronze Age sherds were found in some of the sample units, but there are severalreasons for believing that the major-

I40

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

Figure4.4. View of the ruggedkarst surfaceon the southeastslopeof Spilaionat the time of collection ity of the lithics are not of the same date as the sherds. An inspection of lithics from Bronze Age sites at other locations in the Acheron valley permits us to distinguish lithics typical of the Bronze Age from those of the Palaeolithic.20Bronze Age lithics are unpatinated, smaller in scale, made of different raw materials, and include retouched tool types that are different from those of the Palaeolithic in general and from those found at Spilaion in particular.Small numbers of possible Bronze Age lithics were collected at Spilaion (much less than 1%of the total) and were easily distinguished from the earlier materials on the basis of their morphological characteristics. The Spilaion lithics were manufactured from a good quality, finegrained, blue-gray flint availablein local deposits of Cretaceous limestone. The artifacts exhibit a considerable degree of weathering manifested in the form of a thick white patina, an indication that the artifacts have been exposed on the surface for some time. At present we have no means of ascertaining how long artifacts must be exposed to acquire such a patina. There has been little study of the phenomenon of patina formation with regardto lithic artifacts in Greece, but our observations of materials from other sites may provide a few guidelines. Flints found on the surface and exposed to weathering are usually patinated. This weathering process can begin at two points: either at the time of deposition, before the artifact is buried in a deposit of some kind, or once the artifact has been removed from whatever matrix it was in. Middle Palaeolithic artifactsfrom the nearby site of Alonaki have an estimated age greater than 50 kyr B.P. (thousands of years before present) and are sometimes weathered completely through, rendering the original flint into a material resembling chalk. Other sites in the Acheron valley have artifactsof Bronze Age type (e.g., tanged and barbed arrowheads)that are unlikely to be more than 6,000 years in age and are not deeply patinated. These facts provide reasonable criteria for estimating the amount of time necessary for accumulating a patina on flint artifacts, although we must allow for much variabilityand uncertainty.The Spilaion lithic artifactsare rather heavily patinated when compared with Bronze Age lithics, and are

20.Tartaron 1996; Tartaron,

Runnels, and Karimali1999.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I4I

Figure4.5. Samplegridon the

southeast slope of Spilaion during
collection

generally patinated to the same degree or slightly less than the Middle Palaeolithic ones. Thus we conclude that they have been exposed to weathering for much more than 6,000 and somewhat less than 50,000 years. The numerous lithic artifacts and the technological and morphological uniformity exhibited by the tools and the other debitage categories justifies the use of the term "site"for this scatter.The initial survey convinced us that this site required detailed study to determine if cultural patterning was present in the spatial distribution of the artifacts.The large number of lithics, which may have been exposed directly on the surfacefor many thousands of years, offered an unusual opportunity to test the efficacy of spatial analysis for investigating the forces responsible for the site's formation. To document the spatial distribution of the Spilaion lithic scatter, a systematic program of sampling was undertaken in 1993. Although the site is more than one hectare in extent, the slope of the ridge is irregular and has steeper patches of nearlybare rock interspersedwith flat areasthat trap fine sediment and artifacts.The patches of sediment (Pleistocene terra rossa) are no more than 0.15 m in thickness. Taking into account the large size of the site and the irregularityof the terrain,a small sample grid (60 x 50 m) was established on the part of the slope where the surface was not obstructedby vegetation or interruptedby outcrops of limestone (Fig. 4.5). The grid was subdivided into four large squares (T93-100, T93-126, T93-127, and T93-128), each 25 m on a side, with a 10-m wide strip separating the northern and southern squares.Each squarewas then subdivided into twenty-five smaller squares measuring 5 m on a side. All of the lithics on the surface were collected from eighty of the 5 x 5 m sample units (one large grid square,T93-126, was subsampled with a random sample of five squares).The density of artifacts in the sample area was low, and there were many small, undiagnostic fragments. A total of 3,218 identifiable artifacts were selected for the analyses reported here. Portions of the site not covered by the sample grid were investigated by walkovers to verify that the lithics collected are representative of the site as a whole.

I42

C. N. RUNNELS,

E. KARIMALI,

AND B. CULLEN

THE LITHIC ASSEMBLAGE

On the basis of a ratherlimited range of tool types-chiefly carinated and nosed end scraperson retouchedblades and flakes,notches and denticulates, and rare burins-and a lack of typical Aurignacian bladelets (Dufour bladelets) and microretouched points, the lithic assemblage from Spilaion most closely resembles the Typical Balkan Aurignacian.21 Detailed study of the lithic artifactswas carriedout in 1994 and 1995 with three aims: to identify diagnostic tool types; to understand the technical and morphological characteristicsof the assemblage in terms of the recognized lithic reduction stages (chaineoperatoire); to assess the spatial distribution of and the by-products of the different stages in the reduction sequence, especially cores, cortical flakes, plain flakes, and retouched tools. The Spilaion lithics have suffered from their long exposure on the surface, and heavy patination, breakage, and scarring make many of the pieces difficult to classify. Although the sample of retouched artifacts is relatively large (n = 131), the absence of stratigraphiccontext, the possibility of a mixture of different phases of activity, and the difficulties of classifying the damaged surface materials make our conclusions regardingthe cultural affinities of the assemblage somewhat tentative. The lithics were sorted on the basis of five technical categories: cores, cortical flakes, plain flakes, blades, and retouched tools (Table 4.1). Different attributeswere recorded for each category, including size, raw material, technology, and presence of cortex. Maximum length, width, and thickness were measured for each artifact.The maximum widths of platforms and flake scars on the faces of cores were recorded when possible. Other characteristicsthat were noted include patina, the state of preservation of the artifacts,and the identification of the flintknapping techniques employed. The classification of flakes was based on the following categories:primary(>75%cortical), secondary (25-75% cortical), and plain ("tertiary")flakes (<25% cortical). The majorityof the artifactswere manufacturedfrom bluish grayflint, but flint of other varieties is also present, ranging in color from grayish orange to moderate reddish brown. Pieces in other colors, however, are usually found with no patina and exhibit different technological characteristics,thus we attributethem to a later period (possibly the Bronze Age). Many artifactsshow evidence of burning, such as crazing, potlid fractures, and a reddish color. Cores comprise 6.5% of sample (Table 4.1) and are derived primarily from cobbles gathered from the river rather than nodules extracted directly from outcrops. These cores have traces of a thick cortex, which was usually removed with the first series of flakes (25.7% of the cores have cortex remaining on their surface). The flake cores are variable in size and shape, and they include globular or polyhedral (13.8%), conical (6.7%), flat (2.8%), spherical (1%), and rectilinear (0.5%) types. The majority of cores are irregular in shape (37%) or are fragmentary (32%). Two Middle Palaeolithic Levallois cores were also noted in the sample.

21. Kozlowski1999.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I43

TABLE 4.1. CATEGORIES DEBITAGE

Type Cores Cortical flakes Plain flakes Blades Retouched tools Total n 209 399 2,419 60 131 3,218

OF FLINTKNAPPING
% 6.5 12.4 75.2 1.8 4.1 100

22. Kozlowski 1999, p. 106.

Given the great variety of forms, the cores may be divided into two broad categories:rounded cobbles (core-choppers) and unshaped nodules. Core-choppers preserve vestiges of unidirectional and, more rarely,bidirectional flaking. In the case of bidirectional flaking, the blows were directed to platforms positioned at right angles on the core surface. Unshaped nodules were heavilyflaked in different directionsand usuallyretain little cortex. Some of these specimens are small discoids with irregularly placed striking platforms resulting from multidirectional flaking. In many cases severalplatforms were created on a single core, although only a small number of flakes were struck from them. The same platform was commonly used for detaching flakes from more than one face of the core.These strategies resulted in continuous flaking along the periphery of the core. Core rejuvenationwas accomplished by flaking off a piece of the core carrying the old platform. Four conical cores collected from one tract (T93-128) range from 16 to 56 mm in length and from 15.7 to 37.7 mm in thickness. These cores have a single plain flat platform from which semiparallelor irregularflakes were detached by direct percussion. Blade cores are represented by two complete and eight fragmentary specimens, and were used for striking elongated and irregularblades/flakes by direct percussion. The best example of a complete blade core (65.6 mm in length and 34 mm in thickness) is patinated and has traces of large, irregularblade scars originating from one plain platform. Core-processing activities at Spilaion show that lithic strategies were expedient and opportunistic in terms of goals, techniques, and the quality of the final products. The knappers employed simple direct percussion to remove blanks, and there is no evidence for the use of preparatorytechniques such as cresting or platform faceting. The relativelyhigh frequency of plain noncortical flakes with little evidence of use indicates that core testing and preparationtook place repeatedly at the site. Given the expedient flaking procedures at Spilaion, the goal of the chaineoperatoire followed at the site is unclear.It appearsthat flaking was intended to extract flakes of different sizes suitable for deliberate retouch. Negative flake scarspreservedon the cores indicate that flakes ranged from 8 to 15 mm in length. Blades were evidently produced in much the same manner as flakes. The rarity of typical Aurignacian bladelets is a characteristic of the Typical Balkan Aurignacian (and the earlier Bachokirian)22

I44

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

TABLE 4.2. TYPES OF RETOUCHED

Type End scrapers Simple end scrapers Simple end scraperson blades Carinatedend scrapers Nosed end scrapers Denticulated end scrapers Atypical end scrapers Side scrapers Denticulates Simple denticulates Denticulatesforming a bec Denticulates forming a perfoir Denticulate,backed Notched pieces Retouched flakes Burins Piecesesquillees Arrowheads(FN type) Total n

TOOLS
%

7 6 3 8 13 5 3 29 12 5 1 8 25 3 2 1 131

5.3 4.6 2.3 6.1 9.9 3.8 2.3 22.1 9.2 3.8 0.8 6.1 19.1 2.3 1.5 0.8 100

and probablyof the Aurignacian in Greece as well.23 The eastern MediterraneanEarlyUpper Palaeolithic(EUP) assemblagesfrom the Balkanseastward to Turkey24 and the Near East are similar in this respect.25 Another feature of the knapping techniques employed at Spilaion is the retouching of core fragments to manufactureend scrapersand denticulates, often on the same blank. A total of 131 retouched tools (4.1% of the sample) were identified and classified into sixteen types (Figs. 4.6-4.11; Table 4.2). Denticulated and notched pieces (42%) and end scrapers (32%) dominate the assemblage (see, e.g., Figs. 4.6, 4.7). Simple flakes with irregularlyretouched margins (e.g., Fig. 4.10:6) comprise the next largest group of tools (19%), and there are small numbers of side scrapers(e.g., Fig. 4.7:6), burins, and pieces esquillees, as well as an arrowheadof a later (Final Neolithic or Bronze Age) date. The end scrapers include simple, carinated, nosed, and denticulated types (Figs. 4.10:3-5, 8 and 4.11). The majority were made on plain flakes (85.7%) ranging in size from 22 to 70 mm in length and from 6 to 30.5 mm in thickness. Only six end scraperswere made on blades. The nosed end scraperswere made on both thick and thin flakes. Notched pieces and denticulates (e.g., Fig. 4.10:7) were made on flakes of all sizes, sometimes by the Clactonian technique (i.e., the notches were created by the removal of a single flake) and sometimes by retouch. The denticulates were often placed on blanks with other tools, such as becs or perfoirs. Large primarycortical flakes appearto have been preferredblanks for denticulates. There are two carinated burins on flakes. There is also one small fragment of a bifacial foliate.

23. Darlas 1989, p. 157; Koumouzeliset al. 1996; Perles 1987. 24. Kozlowski1992, 1999; Kuhn, Stiner,and Gule9 1999. 25. Clark 1994; Olszewski and Dibble 1994.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I45

from Figure4.6. Lithicartifacts dentic2-3) Spilaion:1) raclette; All ulates;4-6) end scrapers. artifacts areflint. Scale 1:2

1
1

'

2
2

Figure 4.7. Lithic artifacts from Spilaion: 1-5) end scrapers (2 and 4 are on blades); 6) convex side scraper on a primary cortical flake. All artifacts are flint. Scale1:2

-: i

5 4 6

I46

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

-:::':~~~ --xtoM ]

from Figure4.8 (left).Lithic artifacts Spilaion:1) largebladewith an end

^-' /''''* scraper and abrupt lateral retouch;

2-3) end scrapers. artifacts All are

flint. Scale 1:2

Figure 4.9 (right). Lithic artifacts

from Spilaion: 1-2) end scrapers;

3)perfoir; 4) fragment of a blade core; 5-6) notched pieces; 7) typical flake. All artifacts are flint. Scale1:2

^6 ?^ j )<

Figure 4.10 (below,left). Lithic artifacts from Spilaion: 1-2) notched and pointed pieces (becs/perfoirs); 3end scrapers;6) retouched flake 5, 8) with small end scraper;7) denticulate. All artifacts are flint. Scale1:2 Figure 4.11 (below,right). End scrapers from Spilaion. All artifacts
are flint. Scale 1:2

4 3

61~~

~ ~~~~2

f 5

EARLY

UPPER

PALAEOLITHIC

SPILAION

I47

GIS MODELING AND THE INTERPRETATION OF SURFACE SITES

The special taphonomic conditions pertaining to extensive open-air sites such as Spilaion shaped the objectives of our spatial analysis. The processes that affect open-air sites are quite different from those associated with enclosed habitation sites (e.g., caves and rockshelters). In addition, the absence of stratified deposits at Spilaion prohibited three-dimensional analysis and limited the investigation to issues of horizontal patterning and site-formation. One goal of our study was to determine if the lithics were distributed over the surfaceby naturalforces or culturalactivities.We considered three hypotheses of site formation that may have played a role in shaping the site. Two of these hypotheses focus on cultural activities and the third presumes naturalprocesses.We began the analysisby considering the possibility that artifactswere deposited directly on the surfacewhere they are found today (i.e., directly on the bedrock) as the result of short-term cultural activities unassociated with the deposition of sediments. The cultural activities may include flintknapping (e.g., core testing and flaking) or the secondary disposal of artifactsthat have been removed from their use locations (i.e., materials that were dumped). As an alternative hypothesis, we considered the possibility that human occupation of Spilaion over a considerable period of time resulted in the accumulation of sediments that formed the matrix of a stratified deposit. In this hypothesis, the artifacts ended up on the bedrock after the sediments containing them were removed by erosion or deflation (the "concentrationeffect").26 these two In we assume that culturalpatterning is preserved in the associahypotheses tions of artifacts of different type, and that size or weight has relatively little part in shaping the distribution.Thus, associations between different artifact classes (e.g., cores and flakes, flakes and tools) will be statistically significant despite differences in the dimensions and masses of the artifacts in question. In other words, if cultural processes are major factors shaping the spatial distribution, artifactsof different sizes and weights will have no statistically significant association with the slope of the site as they would if some natural force such as downslope movement was at work. A third hypothesis for the artifactdistribution at Spilaion holds natural forces rather than cultural activities as being chiefly responsible for shaping the lithic scatter. According to this hypothesis, artifacts would have been transportedfrom another location by a physical process such as erosion or downslope movement and redeposited on the surface where they are found today. If this hypothesis is correct, surface artifacts should cluster on the basis of their size and weight (and, perhaps to a lesser degree, their shape) rather than by their type. The key factor in shaping the distribution is the gradient of the slope. In cases of steep gradients with low vegetation cover,artifactsare expected to scatter in a predictablemanner,with heavier artifactssuch as cores working their way downslope leaving smaller artifacts behind.27We assume that the association, or lack thereof, between the frequency and size of artifactsand the topography of

26. Kirkbyand Kirkby1976. 27. Rick 1976.

I48

C. N.

RUNNELS,

E.

KARIMALI,

AND

B. CULLEN

it Of this the sitewill be decisiveforevaluating hypothesis. course, is probforcesareat workon the site thusblurand ablethatboth cultural natural ring the originalpattern,but it is importantto determinefirst if natural forcesarethe most important agentsof accumulation. In a case where naturalforcesare determinedto be negligible,eviin distribution be sought. denceof horizontal may patterning the artifact and in for The needto search patterning the debitage tool typesat Spilaion is corroborated two additionalfactors:1) the remarkable typological by and of the lithic assemblage the high degree andtechnological uniformity and in of consistency flakingtechniques; 2) the high degreeof integrityin of material the assemblage's (i.e.,the dominance stonetools), composition of to the occurrence a singlecultural activity(i.e.,flintknapping), pointing of on ratherthanto a mixture activities the site. betweencomponentsof Emphasisis given here to the relationships of the reduction or interrelated madepossible sequence, stages succeeding which assignslithic of approach, by the application the chaine operatoire use, or specimensto differentstages in their life cycle (i.e., production, the derivefrom We considered possibilitythat artifacts may deposition). (e.g., core testing and flakingtaking primarydepositionof by-products disposalof artifactsthat have been place on the site), or the secondary that were dumped).In removedfrom their use locations(i.e., materials horizontal refersto the caseof primary meaningful patterning deposition, the degreeand natureof association clusters,statistical (e.g., overlapping of correlation) artifactgroupslinkedin the productionchain (e.g., cores andcortical flakeslinkedin the decortication stage,blanksandtoolslinked in the retouchingstage).In the case of secondary deposition,redundant disposalpatternsare expectedto resultin mixed depositsof artifactsin of different stagesof theiruse-cycle(e.g.,products the earlier stages,nonretouched with postdepositional scarflakes,tools,recycled tools,artifacts ring). Unfortunately,due to the considerable degree of weathering of assignedto the UpperPalaeolithic period,any (patination) the artifacts on information tool recycling, or that use-wear, postdepositional scarring couldbe usedto distinguish has groupsof artifacts been lost. Information common morphological characteristics regarding (e.g., color,texture)of flint categories alsolost, preventing refittingof pieces.Neverthewas the less, basedon the obvioustechnological uniformityand integrityof the we to material, endeavored discernanytype of meaningful spatialassociation betweenclassesof artifacts related the samesequence, in despitetheir in differences dimensions massesandtheirrelation the slopeof the and to
site.

SPATIAL

ANALYSIS

OF THE

HIGH-DENSITY

ARTIFACT

DISTRIBUTION

To testthe threehypotheses, undertook spatial we a of analysis the Spilaion lithicsthat comprised rangeof datasets and analytical a Betechniques. forediscussing eachtechniquein detail,we providea few notes regarding the variables the unitsof analysis and chosen.Becauseof the technological exhibited the lithicassemblage the lackof a stratigraphic and uniformity by

EARLY UPPER

PALAEOLITHIC

SPILAION

I49

dimension,the whole surveyedgrid surface(60 x 50 m) was treatedas a single unit with grid cells of 5 x 5 m. In addition,datasummary plots were calculatedon the basis of counts of artifacts grid cell. Finally, per artifact was calculated a productof length andwidth. size as
DOWNSLOPE MOVEMENT

In orderto discernthe roleof natural distriagentsin shapingthe artifact butionat Spilaion, in movement larger of particularly detectingdownslope we for pieces, usedtechniques developed usein spatial mapping/GIS analysis. These techniquesoffer the possibilityof manipulating different the variables detectsignificant to in associations spatial As patterns. a foundation for this analysis, data(e.g., slope and configuration of topographical wereenteredandprocessed our GIS database, in GRASS. bedrock) As the firststepin ouranalysis, attempted evaluate effectsof we to the as either25, 50, 75, or 100%)on the accuracy our artiof visibility(rated fact counts.Calculations basedon artifactcountsand visibilitymeasurements from all cells pointed to a moderatepositive correlation (0.430)
between these two variables,indicating that vegetation and soil cover may have affected visibility somewhat in the sample units at Spilaion. A test of the strength of the correlation between artifact size and the gradient of the slope showed a small negative correlation (-0.173), indicating a very weak tendency for artifacts to be sorted by size. The largest artifactswere not concentrated at the foot of the hill, as one would expect if there was significant downslope sorting by gravity.A subsequent analysis, however, revealed a slightly positive correlation (0.229) between slope and the total number of lithics, suggesting that the gradient of the slope did play a limited role in the distribution of artifactsat the site. This observation is confirmed by spatial mapping, which revealed that clusters of plain and cortical flakes tend to concentrate at the southern end of the grid (see below). The results of these analyses seemed to indicate that, while the gradient of the slope was a factor in shaping the lithic distribution at Spilaion, it was a small factor and did not fully explain the distribution of artifacts across the site. The low gradient of the slope, which did not exceed the "angleof repose"for large artifacts,and the distribution of artifactsof different dimensions and masses across the slope suggest that the materials were not carried by erosion from some higher, more distant source. Evidently, the gentle slope of the site, the irregularand highly weathered karst surface, and the thin covering of scrub vegetation prevented the continuous shifting of artifactsby erosion or downslope creep.
SPATIAL CORRELATION

Spatial correlationbetween different classes of artifacts,irrespectiveof their relative masses, followed. The classification of artifacts into meaningful groups is a necessaryprerequisiteof any correlationstudy,and artifactclasses were defined on the basis of the typology of the presumed reduction sequence (i.e., core decortication,blade and flakeproduction,retouchof blanks to create predetermined tool types).

I50

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

TABLE 4.3. DEGREE OF ASSOCIATION BETWEEN PAIRS OF CLASSES OF FLINTKNAPPING DEBITAGE

DebitageClasses cores/blades retouchedtools/blades corticalflakes/blades cores/retouchedtools plain flakes/blades cores/corticalflakes corticalflakes/retouched tools cores/plainflakes tools plain flakes/retouched corticalflakes/plainflakes Correlation Coefficient 0.254 0.312 0.414 0.456 0.518 0.578 0.594 0.642 0.650 0.836

Comments:there arerelativelystrong correlations between the spatialpatterningof corticalflakes and plain flakeswith flakes and tools, but the correlationis weaker between cores and retouchedtools or blades,suggestingthat these elements were The relativelystrong correlations between cores and flakes is spatiallysegregated. evidencethat the artifactsof differentmasseshave not been sortedby natural processessuch as downslopemovement.The weakest correlationsarefound between the blades and all other classesof debitageexceptplain flakes,suggesting that the bladeswere spatiallysegregated.

Spatial associationsbetween artifactclasseswere assessedvisuallyfrom distributionplots producedby the mapping programSURFER (Figs. 4.12, 4.13). This visual analysis was supplemented with a correlation study, the results of which are summarized in Table 4.3. Interestingly, there is a significant correlation between artifacts that belong to successive stages of the reduction sequence (e.g., cortical and plain flakes, plain flakes and flake tools). The most prominent correlation revealed by the study, and confirmed by visual inspection of the distribution plots, is that between cortical and plain flakes (0.836). Overlapping clusters of these two artifactgroups can be seen in the southwest corner of the southeast quadrant. No doubt, the overlapping concentration in the partiallymapped southwest quadrantis part of the same tendency.Thus, it can be concluded that cortical and plain flakes tend to form overlapping clusters of relatively high concentrations (n = >30 and n = >80, respectively) in the southern part of the grid. A spatialoverlapbetween cortical/plainflakes and tools was also noted (0.594 and 0.650, respectively). As apparent from the distribution plots, tools tend to overlap with the hot spots of cortical flakes and plain flakes in the southeast and northwest parts of the grid (Fig. 4.12). Numerically speaking,however,these clusters representrelativelylow concentrations of tools (n = <6), showing that the latter are rather dispersed across the site. In contrast, cores show only modest spatial correlationwith these categories (i.e., cortical flakes, plain flakes, and tools). Cores tend to form distinct clusters in the northern and the southeastern parts of the grid (Fig. 4.12). Only in the southeastern portion of the grid do cores, and in particularcortical cores, overlap with cortical and plain flakes. Although this is the least intensively surveyed area of the grid, the overlapping cluster of cores observed here is spatially segregated, and does not continue

'5'

0 0o "t

'. o,

r\.

_>

U4

C)

cC 0 ,-.
bl-I I

fL4
:^

,-1.

~ L~
4

C)

U 0 a

--

k,.;
^3 <
*
1-

I t
0

en

rli *r . 0_
.._ ,-a

j,

II
_

cr, :ri

,mWTV,,.... .- 7 - .

a wl,.- :

'
,,,... Ct

U
Ct

C)

0 U OCt
"l

_G
1-

_
--~_.. '--~U :~:.d

,...._
4 4 4 4

*
-

M ,

v\.-

'.

.. F...

,-

U 0I-

--';s 021

,...,

ac

w1 i~,

_.,w

_-

E'

_e

II C

is? ;

:?'

1K
q

,flcC U U

*-

ot

4-'

i1a

.
'4-4

o
I_I
_,

~rj
C: -.1. I.I--

E
wN _r rs

r
U,
-I 0 H1UO

4J

U cC0 4U U

If
',r'

bca

._

ms
i
I I f 48 i Pf f11 , *

n A U^

*144808o do 0 -

EN1) SCRAPERIS (A I.,I, TY1'E1S)(n=361

EN1) SCRAP'ERS (ON B1I,ADES) (n n=6}

-..,

I) I' NI

..

........

k...~.........

z,

X i

.'!;>.1
;

JNMI
w

&
ii .. ..
1-

A.,

ml

0
.1 > * \ r f LA

-?.^

"
a

t ! *

'

T
:: A
AI I

,4
fT j,i

iA

,.7

iI f-T-~

I N) TC HED) FL AKES (n=8)

!E) RETO CIIE FLAKES (n=25)

3>

s, .

1 1
.

3,

00 4.

'909
1 >

f t
0

4
tj30$ 20 1

r,,
LASTr1^

9 IC-0
0 o

'

-t

tf
4

-J.0 :

I0

*
I ;

....
3S 4*0 J3 40

As T r

Figure 4.13. Spatial distribution of individual categories of retouched tools. Each box represents the total area covered by the areas were not collected. Densities are indicated by the colors shown in the keys.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I53

sincethey are bladescomprisea separate towardthe east.Lastly, category, Bladesarerathermoredispersed, limited in number. althoughthey tend to overlapspatially with bladecores. is or of as Characterization these concentrations primary secondary of the most difficulttask.As already noted,recording the stateof preservation of artifactssharingcomparable (i.e., technologicalcharacteristics was impossible Spilaiondue to the obliteration the paat Palaeolithic) by In tina of all vestiges of use-wear,recycling,and resharpening. cases of one of flakes, products an earsecondary deposition, wouldexpectcortical of lierphase,to be foundmixedwith tools,the finalproducts the sequence. at by Althoughthis patternis discerned Spilaion,it maybe explained the character blankproduction the selectionof anytype of and opportunistic of blank(corticalflakes,exhaustedcores,etc.) for retouch.There are no distinctiveor clear-cutstagesof reduction be found at this site;reducto tion evidently in eachotheron stages,succeeding proceeded intermingled the basisof immediateneedsand material restrictions. If the lithicscatterat Spilaionis a primary deposit,the spatialclustering betweencorticalandplainflakesmayindicatethat coredecortication andflakeproduction werenevertwo distinctstagesof the sequence this at site.Giventhe rather of decortication minimally cores, prepared incomplete these two stepswere interrelated phasecomprising partsof a continuous partialcleaningof the core and immediateblankdetachment(i.e., flakes were detachedfrom the cleanedsurfaceimmediately afterits decorticaflakesandtools is betterunbetweencortical/plain tion).The correlation derstoodif we takeinto accountthe fact that most of the retouched tools weremadeon corticalandplainflakes.In contrast, dispersal cores of the and theirgroupingin distinctclusterssuggestthat they were transferred andflakedat anyspot. to of tools Turning the distribution plotsof singlecategories retouched betweenindividual tool (Fig.4.13), ouraimwas to detectanyassociations types and theirblanks(e.g., flakesand corticalflakes).Of majorinterest is the correlation end scrapers of with the hot spots of corticaland plain the weremade. flakes,sincethe lattercomprise blanksfromwhichscrapers Anothersignificant association thatbetweenplainflakesandretouched is flakes. othertooltypes(i.e.,endscrapers blades, All on notched denticulates, tendto formpartially in clusters the centerof the northflakes) overlapping ern grid. Generally, tool concentrations not consistof largenumbers do of artifacts, therearesome significant but correlation patterns arisingbetween some tool types and the debitagecategorieson which they were formed. In sum,ouranalysis that indicates natural werenot the most processes factorsin shapingthe spatialpatterning Spilaion. at significant Although natural are in processes often significant casesof denseopen-airdistributionsof lithics,suchas at Spilaion, analysis our showsno significant correlationsbetweenartifact andslope,a relationship size necessarily presentif disturbance natural had significant by processes takenplace.The only indicationof naturalprocessesshapingthe distribution the slight posiis tiverelationship foundbetweenthe number artifacts the slope.The of and ratherstrongtendencyof corticalandplainflakesto clusterat the southern side of the grid maybe partially resultof downslopemovement. the

I54

C. N. RUNNELS,

E. KARIMALI,

AND B. CULLEN

results analysisyielded comparable Spatialmappingand correlation linked association different of asto the degreeof spatial categories debitage of blankproduction in an operational chain(e.g., the reduction sequence The strongestassociations and toolmaking). producedby both analyses flakesand some arebetween1) corticaland plainflakes; cortical/plain 2) retouched and tool groupsmadeon theseblanks(i.e.,end scrapers, flakes); 3) bladesandbladecores. The smallsize of the sample(ca.2.1%of the totallithicson the site) of disand the considerable degreeof patinaon the majority the artifacts moredetailedconclusions fromthe shapeof the us courage fromdrawing artifactdistribution. The most difficultproblemposed by the analysisis at whetherthe largenumberof artifacts Spilaionwere the resultof priare maryor secondary deposition.If these associations the resultof prithe character flintknapping of marydeposition, highlight opportunistic they aim at Spilaion. The primary of flakedetachment this sitewasto create on immediate blanksfortool production. Thus,therewereno clear-cut stages of production, corescouldbe partlydecorticated reusedat a differas and In ent spot for flake detachment. spatialterms,this resultedin the disof clusters persalof flakecoresin all areasand the formation overlapping of production. with debitagetypeslinkedto succeeding stages It is alsodifficultto determine duration the activityrequired the of to the accumulate largenumber artifacts, of whethertheseactivities occurred overa long periodof time or consistedof a few,shortintensiveepisodesof in We if flintknapping. were equallyunsuccessful determining stratified the sedimentsonce existedat Spilaion, removal whichwouldhaveconof centrated artifacts the bedrock. the on Yet, our technological studysugin gests that the site was utilizedprimarily one period,the EarlyUpper a of Palaeolithic, conclusionsupported the uniformity types,materials, by and techniques.

CONCLUSIONS
Spilaionis a high-densityscatterof lithicswith prodigious quantitiesof in in flintknapping debitageorganized discreteactivityareas, presumably determined associations. artifact The culturally spatial typologypointsto the EarlyUpperPalaeolithic as (Aurignacian) the main periodof use of the site, and the "hotspots" thus be as much as 30,000 yearsold or may even more.The site was evidentlynot used extensively other periods. in Scatteredand highly erodedartifactsof Middle Palaeolithic, Neolithic, andBronzeAge typeaccountforless thanone percentof the totalsample, and canbe discounted the analysis. in The Spilaionassemblage classifiedas Aurignacian the basisof is on tool typologyand flintknapping The rarityof typicalAurigtechnology. nacianretouched bladesand the absenceof Dufourbladeletsand microretouchedpoints,types typicalof the Italianand west European Aurigbut nacian, are notable featuresof the Typical Balkan Aurignacian,28 otherwisethe assemblage conforms the generalpatternof Aurignacian to
assemblages in Greece.29

28. Kozlowski 1999, p. 106. 29. Darlas 1989; Koumouzelis et al. 1996; Kozlowski (pers. comm.); Perles 1987.

EARLY

UPPER

PALAEOLITHIC

SPILAION

I55

30. Zilhao and D'Errico 1999, p. 43. 31. Kozlowski1999. 32. Kozlowski1999. 33. Zilhao and D'Errico 1999, p. 43. 34. Darlas 1989; Koumouzeliset al. 1996; Perles 1987; Runnels 1988, 1995. 35. Perles 1987, phaselithiqueI. 36. Perles 1987, p. 96. 37. Koumouzeliset al. 1996. 38. Darlas 1989. 39. Runnels 1988; Runnels and van Andel 1993a. 40. Allsworth-Jones1986; Runnels 1995. 41. Runnels 1988; Runnels and van Andel 1993a. 42. Koumouzeliset al. 1996; Kozlowski1999, p. 114. 43. Kozlowski1999, p. 108. 44. Kozlowski1999; Kuhn, Stiner, and Giile9 1999; Olszewski and Dibble 1994, p. 70.

An attempt has been made recently to deny that the Early Upper Palaeolithic of the Balkans (termed "Bachokirian" found at Bacho Kiro and and Temnata Caves) is in fact Aurignacian,30 this view is not accepted but by those most familiar with the assemblages in question.31The issue is partly one of nomenclature.The Bachokirian is unrelated to the underlying Middle Palaeolithic industries at Bacho Kiro and Temnata and is unbut doubtedly Early Upper Palaeolithic in character,32 Zilhao and D'Errico wish to reserve the use of the label Aurignacian strictly for those EUP industries having Dufour bladelets, numerous burins, and bone and ivory points.33By their definition the Spilaion assemblage is not Aurignacian but Bachokirian. We believe that this distinction does not help to clarify matters and serves only to confuse the reader.For the present, we shall continue to refer to the EUP industry in Greece as Aurignacian. Aurignacian sites similar to Spilaion are rare in Greece. Surface sites are found in Elis and Thessaly, and the cave sites of Kephalari,Kleisoura, and Franchthi in the Argolid also contain Aurignacian materials.34 The lithics from the earliest Upper Palaeolithic layer at Franchthi Cave35exhibit typological traits of the Aurignacian (carinatedand nosed end scrapers) but they were found in extremely small numbers and therefore cannot be taken as certainly Aurignacian.36 Finds from a rockshelter in the Kleisoura Gorge near Argos exhibit a similar preference for end scrapers on flakes and short blades.37The surface sites in Elis38and Thessaly39produced industries of mixed character,combining Mousterian and Aurignacian elements (e.g., carinated and nosed end scrapers, marginally retouched blades and burins, along with Levallois flakes and Mousterian side scrapers),and similarMiddle Palaeolithic or Early Upper Palaeolithic industries with this mixed characterare known in the Balkans.40 The age of the Greek Aurignacian has not been precisely determined. It was apparentlypresent at sites exposed in the banks of the Peneios River in Thessaly between 45 and 30 kyrB.P., determinedby radiometricdates.41 as The recently excavated Kleisoura shelter has a rather late Aurignacian, dated to ca. 34-22 kyr B.P. (uncalibrated).42 cannot say where in this We is to be placed, and can only give a rough chronologilong period Spilaion cal range of ca. 45-22 kyr B.P. for the cultural activity at the site. Outside of Greece, the Spilaion assemblage can be comparedwith the assemblages from Bacho Kiro (layers9-11) and Temnata Cave (layers3-4) in Bulgaria, where the Aurignacian layers have been dated from 45 to 28 kyr B.P. The (uncalibrated).43 Spilaion assemblageis thus similarto the Aurignacian and other EUP assemblages of the eastern Mediterranean sensu lato.44 If we are correctin assigning the majorityof the Spilaion lithics to the Early Upper Palaeolithic, this one site has more than 250 times as many artifacts as are found on the other EUP sites in Greece. Thus Spilaion is perhaps the largest lithic site in Greece. It is extraordinaryeven by local The entire lithic collection from the rest of the Nikopolis Epirote standards. survey,which is based on the total collection of lithics from all tracts, is less than 15,000 pieces. The largestMiddle Palaeolithic sites in the Preveza region (e.g., Kokkinopilos), which are certainly among the richest lithic sites in the country, have less than one-tenth the number of lithics visible on the surface at Spilaion. The size and preservation of the EUP lithic

I56

C. N.

RUNNELS,

E. KARIMALI,

AND

B. CULLEN

scatter at Spilaion, therefore, presents a rare but important opportunity to study a site of this period, despite the complete absence of stratified deposits. Artifact-rich surface sites are common in Greece, but there has been We some debate about their value for archaeology.45 acknowledge that such sites cannot be studied by means of traditional excavation techniques, but we believe that the study of spatial patterning permits archaeologists to make greater use of them. If the quantities of artifacts preserved are large enough, spatial analyses can be useful in interpreting past culturalactivity, even where stratigraphicassociations have been lost or were never present. The number of these artifact-rich sites has increased greatly as the result of intensive surface reconnaissance on a regional scale. Such sites are not exclusively prehistoric or marked only by scatters of lithics. We believe that the methods detailed in this report can be applied successfully to historical sites and to sites with rich concentrations of sherds, rooftiles, and other cultural materials.The identification of patterns in the artifact distribution at Spilaionshouldserve as an incentive for the continued study of surface sites in Greece and throughout the Mediterranean.

45. This debate is summarizedin

et Cherry al. 1988andAlcock,Cherry, andDavis1994.

CHAPTER

THE THE AND

COASTAL AMBRACIAN ITS

EVOLUTION

OF

EMBAYMENT TO

RELATIONSHIP

ARCHAEOLOGICAL

SETTINGS

by ZhichunJing and George (Rip) Rapp Coastal landscapes are a sensitive interface for environmental change. In the past 10,000 years, the Ambracian embayment and its vicinity have witnessed dramatic landscape changes in response to Holocene eustatic sea-level rise, sediment infill, erosion, and tectonic movement (Fig. 5.1). The changing landscape in this area,utilized since the Lower Palaeolithic period,1has affected both the spatial and temporal distribution of archaeological remains.Thus, the pattern of prehistoric and historical settlement must be understood in the context of the evolving coastal landscape. Paleoenvironmentalreconstruction associatedwith archaeologicalinvestigation in Epirus has focused on the Palaeolithic period,2and no investigation has been conducted to examine the Holocene environmental context of settlements based on the subsurface stratigraphy.A limited number of studies based on geologic or sedimentary perspectives have been undertaken to address the evolution of the coastal landscape during the Holocene. Although these studies revealedsea-level and coastline changes, they provide no essential data for the interpretation of archaeological settings in terms of either the temporal or spatial scales dealt with in the investigation of settlements in the embayment of Ambracia and its vicinity.3 In this chapter,we describe the changing landscape in the Ambracian embayment during the Holocene based on an analysis of the subsurface and stratigraphy, we establish the environmental context of various prehistoric and historical settlements. In order to reveal the subsurfacestratigra1. Hammond 1967; Dakaris 1971; Runnels and van Andel 1993b; Bailey 1997. 2. Bailey,King, and Sturdy1993; Dakaris,Higgs, and Hey 1964; King and Bailey 1985; SturdyandWebley 1988; Sturdy, Webley,and Bailey 1997; Turnerand Sanchez-Gofii 1997; VitaFinzi 1978, pp. 139-158. 3. Piper,Panagos,and Kontopoulos 1982; Piper,Kontopoulos,and Panagos 1988; Poulos, Lykousis,and Collins 1995;Tziavos 1997. Both Poulos, Lykousis,and Collins (1995) and Tziavos (1997) studied the Quaternary subsurfacestratigraphy throughthe analysisof 3.5-kHz seismic reflection profilesacrossthe AmbracianGulf, providingsome criticalinformationon the formationand developmentof the basin duringthe Late Pleistocene and earlyHolocene. Tziavos (1997) carried out some drillingin the coastalplain north of the gulf aimed at studyingthe evolution of the basin paleogeographic duringthe Quaternary period.

158
_ _ _,_

ZHICHUN

JING

AND

GEORGE

(RIP)
P7I I

RAPP A

-?

0. <

^v-^
',;'
3

i i~~~
4 .'" "

Li (7' S^'-^~~~~~~~"''i I I ..m:%,'

/,

~..r..

....
'? / ,'/

I. I-^^
(. ._j .....L~ ..

m.

'
.

m ~],,

/ T^Nj

..&~.mV' .--V~.~ . . m',,qm

':1i'Ii ...

..

',' Acheron

R.

......

... /

. ,

'

~ ..

N
V%
1 ------;

26

'2 --/t/. /***'--/ 2) ,2' L2S / \-^ Stavros \.' '2

-/-!

I v-'... I~~~~~~~~~~~~~~~~~-. .

n-

Rokia

....

'.

:.-^
\ i

"
K

"

'
, 7,?

:; '

<aSI>,

"t i r

4^\'

-<.

... ? ...

r,

**:-^

;-\

..:

,. '-

...

'i ',-

) /21

..

; . v....

Vi

v\

{-

<

Ion 'Sea
co,~(-C. . .

\ii,
.

.7:
. -/1 . . ?'?..

...Salaora
" .,,.v -

Al;
-

u ~1. II.M a r.n en o ~m (in z~o I.

1. P.A ..A.

T o a iV.

V. \l

che

'

.theAmbracian Embayment and its Vicinity ,.

. /2___
0T^-I

.
//d4 ,.. /
. . /N
'.

.
;/.
. / '. 4,.--.'

a.Anbracian
Gulf-....
:k'15-=[/ '

1__

6,

._

1.... KM
IIHoloc'

.7
Alluvi."

, IV .K
'

Nlesoioic

',;KJ,, "I

m..

Fvsch

rtiarx ["*,,'" ',

.lisocL..

.-

Hoi,cene Swamp

, \.

,-" , ? ". ' { ..

'v":

1. Actium 6. Comaruis I I. Cassope 16. Strongyli 21. Kastri

2. Anactoriurn 7. Ormos Vathy 12. Riza

3. Pantokrator 8. Michalisi

4. Margarona 9. Archangelos 14. Cheimadio

5. Nikopolis 10. Kastrosykia 15. Kastro Rogon

1 /0'
II,

,; . - .._2_ ......

/I
6 1

.
" ' "

:....
" "

13. Palaiorophoros

17. Vigla 18. Kastro Rizovouni 22. Kanalaki

19. Nckyomanteion

20. Koumasaki Ancient Hlarbor

fly

"

'

'.1.*

23. Ephyvra 24. IThesproliko

---~~~~~~~~~~~~~~~~~~~
Figure5.1. Geologyandgeomorphologyof the Ambracian and embayment its vicinity Locationsof Figure5.2 (opposite). coresandcrosssections. geologic Forlegend,see Figure5.1. For unnumbered cores,see Figures5.3, 5.4, and5.14.

phy, we took a series of geologic cores using a hand-operated auger. Our drilling was limited to the western part of the embayment. A total of 35 cores were drilled with a maximum depth of 13 m in different parts of the lagoon-swamp-coastal delta plain along the Ambracian Gulf (Fig. 5.2). In addition to drilling in areas of geologic importance, we drilled a large number of cores around historical settlement sites to better understand their paleogeographic setting, particularlytheir relation to shorelines and possible harbors. These sites include Nikopolis on the Preveza peninsula, the Roman harbor town site surrounding Ormos Vathy on the Ayios

COASTAL

EVOLUTION

OF THE AMBRACIAN

EMBAYMENT

I59

Thomas peninsula, and Kastro Rogon on the limestone hill south of Mt. Rokia.4
4. Each core was describedand logged in terms of lithology,color, structure, consistency,plant and faunal debris,culturalinclusions,stratigraphic position (depth and thickness),and other observablesoil and sediment properties such as the presenceof calcium carbonateand pebbles.The terminology used follows Folk 1980 for sediments, and Soil SurveyStaff 1975 and Birkeland1999 for soils. Sediment and soil samplesfrom the coreswere taken for laboratoryanalyses,including grain size, microfossils(particularly ostracoda and foraminifera), organicmatter,and calciumcarbonatecontent. A total of seventeenAMS radiocarbon dates were determinedon core samplesat the Universityof California,Riverside,and Peking University(see Table 5.1). Six of these dated samplesaremarshgrass and wood debrisassumedto have seagrown in or nearcontemporary level positions.

I6o

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

GEOLOGY AND GEOMORPHOLOGY AMBRACIAN EMBAYMENT

OF THE

The Ambracian embayment is a major tectonic depression (post-orogenic graben) in southwestern Epirus, situated in the so-called Ionian zone between the Hellenide mountain chain (Pindos Mountains) and the Ionian coast (Fig. 5.1).5 The embayment includes the marine gulf itself and a low-lying lagoon-coastal delta plain to the north. At present over half of the embayment, including the Ambracian Gulf and three lagoons, is under water.The Ambracian Gulf measures ca. 35 km from east to west and 10 km from north to south. The lagoon-coastal delta plain to the north is ca. 10 km from north to south. The gulf itself is relatively deep with a maximum water depth of ca. 65 m in its southern part. The average tidal range is only 5 cm, with a maximum recorded range in a single tidal cycle of 25 cm. The gulf floor shows a gentle gradient on the north due to sediment deposition from the Arachthos6 and Louros Rivers, while it drops off quite rapidly (to 40-65 m) on the other sides of the gulf. The gulf is sheltered from wave processes of the Ionian Sea by both the Preveza peninsula and the sandy spit at Actium; thus, secondarywaves formed within the gulf create a train of littoral transportthat is responsible for the formation of sandy barriersalong the north shore. Most of the gulf is floored with rather uniform olive-gray silty sediments composed of 65-75% clay. The gulf is connected to the Ionian Sea by a narrow channel (Preveza Strait; 600 m wide) north of Actium.7 The Ambracian embayment is bounded by bedrock to the east, south, and discontinuously to the west. The southern flank of the gulf is bordered by a sharp cliff incised into Mesozoic limestones. A Tertiaryflysch fringes the embayment in the east. To the west and southwest of the embayment is the Preveza peninsula, cut off from the mountainous limestone platform and flysch basin in the north by Mt. Zalongo. The low, hilly peninsula is composed mainly of interbedded mudstone, sandstone, marlstone, and pebbly conglomerate that formed in shallow marine, alluvial delta plain, and fan environments during the Pliocene and Pleistocene.8 In contrast to Mesozoic limestones, these Pliocene and Pleistocene sediments are easily eroded. As a result, the littoral transport of sandy sediments has created a relativelybroad sandy beach along the peninsula'sIonian coast. To the north, the embayment is borderedby a series of bedrock mountains that consist of alternating Mesozoic limestone and Tertiary flysch formations. The limestone ridges have elevations of more than 600-1000 m, while flysch mountains have relativelylow elevations (150-600 m) and usuallyconstitute the basins between Mesozoic limestone platforms.These alternating limestone platforms and flysch basins strike N25W; and they disappearunderneath the Pliocene to Quaternary sediments in the south along an east-west structuralfeaturerepresentedby severaleast-west striking limestone mountains including Zalongo, Stavros, and Rokia. Along the Ionian Sea, the Mesozoic limestone ranges have relatively low elevations (500-700 m) and drop abruptlyto the sea, forming a steep and nearly harborlesscoast in the northwest part of the study area.The steep coast is broken only by two small bays:Phanari Bay,into which the Acheron River flows, and a second small bay located slightly to the south (Fig. 5.1).

5. Monopolis Bruneton and 1982. 6. The Arachthos River eastof is

the areashown in Figure 5.1; see Figure 6.1 for its location. 7. Piper,Panagos,and Kontopoulos 1982; Piper,Kontopoulos,and Panagos 1988. 8. Doutsos, Kontopoulos,and Frydas1987.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

9. Piper,Kontopoulos,and Panagos 1988, p. 285. 10. Currently, Louros River the emergesfrom the deeply incised valley in the north and unexpectedlyturnsto the west along the foot of Mt. Rokia. The riverthen turnsback towardthe south and enters the gulf nearMichalitsi. Aerial photos show a seriesof abandonedchannelsin the swampy areato the north of Rodia Lagoon, suggestingthat the Louros Rivermight have enteredthe lagoon in the past. The currentflow patternof the Louros Riverindicatesthat there might have been a majorchannel diversionin This channel diversionwas antiquity. most likely made to drainthe western part of the low-lying lagoon-coastal plain and expandthe farmablearea. 11. King, Sturdy,and Whitney 1993. 12. Clews 1989; Etudegeologique; King, Sturdy,and Whitney 1993. 13. King, Sturdy,andWhitney 1993; Papazachosand Comminakis 1971. 14. King, Sturdy,and Whitney 1993, p. 157; Pirazzoliet al. 1994; Underhill 1989; Stiros et al. 1994.

The Ambracian Gulf is separated from the northern delta plain by a and three lagoons (Rodia,Tsoukalio, and Logaswamp, the Salaorabarrier, rou), and from the lowland north of Nikopolis by Mazoma Lagoon. The Salaorabarrieris relatively narrow,with a large projection to the north. It is approximately8 km long, and is believed to have been formed by littoral transport of sand and gravel sediments eroded from the Preveza peninsula and from those derived from the mouth of the Louros River.The sandy barrierbordering Logarou Lagoon may be associatedwith the main abandoned channels of the Arachthos River. Both the Arachthos and Louros Rivers enter the embayment from the north and they have provided the majorityof sediments to the Ambracian Gulf and its coastal plain. The Arachthos River, with an estimated average annual discharge of 80 m3/s, is one of the largest rivers draining the high Pindos mountains of Epirus. It has been dammed and regulated since 1980. The Louros River, dammed since 1963, drains the mountains to the west of the Arachthos River.In terms of water discharge,the Louros River is much smaller, with an estimated average annual discharge of 30 m3/s.9The Arachthos River is responsible for the formation of the eastern portion of the coastal plain, and the Louros River is the primary contributor to the development of the western portion. The eastern part of the coastal plain shows a much more developed alluvial morphology than the western part.This circumstance may be attributed to the much larger dischargeof the Arachthos in the east.The Louros River and other streams emerging from mountains to the north and west have relativelysmall discharges.As a result, the three lagoons occur only in this part of the embayment, and much of the areasurroundingthe lagoons is swampy (Fig. 5.1). Part of the swampy area,particularlynorth and west of Mavrovouni ridge, has been drained and the land reclaimed for agriculturalpurposes. Bathymetric contours along the north shore of the Ambracian Gulf indicate several southward extending protrusions representing prodelta platforms that have developed near the mouths of the rivers (Fig. 5.1). There are two southward extending prodelta platforms in the western part of the gulf. The first one, projecting southeast, is relatively small and is associated with the current mouth of the Louros River. Another, near Salaora, protrudes southward. A series of abandoned channels exists to the east of the current Louros River channel, from Kastro Rogon southward to Tsoukalio Lagoon, suggesting that the Louros River may have flowed southward directly into the Ambracian Gulf after emerging from the mountain valley.10 Tectonically the Ambracian embayment and its surrounding area are a triple junction between the Ionian, Aegean, and European plates, showing a relatively complex pattern of local tectonism.11The present morphology of the embayment was shaped by Oligocene-Miocene compressional folding and faulting (north-northwest to north-northeast) followed by extensional faulting (west-northwest to east-southeast) during the Late Pliocene and Quaternary.12 Continuing tectonic activity makes this region one of the most active seismic areasin the world.13The embayment itself has been subject to continuous tectonic subsidence since the PliocenePleistocene, but the Preveza peninsula to the west has been uplifted as indicated by anticlines.14

I62

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

SUBSURFACE STRATIGRAPHY AND PALEOGEOGRAPHIC RECONSTRUCTION

Our approach to reconstructing coastal landscape change is to determine the vertical and lateral sequences of the subsurfacestratigraphythat record past geographic change from various processes such as eustatic sea-level Walther'sLaw of correchange, deposition, erosion, and local tectonism.15 lation of sedimentary facies constitutes the conceptual framework for our The single most important analysis of changing coastal environments.16 of coastal landscapereconstructionis to determine "paleo-time component The depositional surface may be a deltaic flooddepositional surfaces."17 plain above or near sea level, a coastal barrier above sea level, a lagoonal deposition surface below sea level, or a coastal marsh or swamp near sea level. Using paleo-time depositional surfaces,one can draw the shorelines and the spatial patterns of various lithosomes for different periods in the past. These sedimentary concepts have been successfully applied to the study of coastal environmental change in archaeological contexts in Greece.18
PREVEZA PENINSULA

Nikopolis is located on the middle portion of the Preveza peninsula that separates the Ambracian embayment from the Ionian Sea. Specifically it is located on a low Pliocene-Pleistocene ridge near the southwest coast of Mazoma Lagoon (Fig. 5.1). Between Mazoma Lagoon and the Ionian Sea is a lowland that dissects the Preveza peninsula composed of interbedded mudstone, sandstone, and pebbly conglomerate of the Pliocene to Pleistocene periods. The lowland, here referred to as the Nikopolis isthmus, is ca. 2.3 km long and 250-1000 m wide and is covered with alluvial and slope-wash sediments of the Holocene period. The highest portion, in the middle of the lowland, is ca. 16 m above current sea level (Fig. 5.3). According to Strabo, there were two harbors near Nikopolis during the Roman period. Comarus, the nearer and smaller of the two, which forms an isthmus of sixty stadia with the Ambracian Gulf, and Nikopolis ..., and the other, the more distant and larger and better of the two, which is near the mouth of the gulf and is about twelve stadia distant from Nikopolis.19 The smaller harbor, Comarus, is situated on the Ionian Sea (Fig. 5.1). Both Hammond20and Dakaris21interpreted Strabo'smeasurement of 60 stadia (12 km) as the length of the isthmus from the smaller harbor, Comarus, on the Ionian Sea, to Nikopolis, lying on the Ambracian Gulf. However, the length of the isthmus is only ca. 2.3 km (11.5 stadia) at present and it was even shorter during the Roman period due to marine transgression.If the location of the Comarus harboris correctlyidentified, Strabo'smeasurement of 60 stadia must be wrong. Both Hammond22and Leake23placed the second harbor at Ormos Vathy, about 12 stadia (2.4 km) from Nikopolis. Ormos Vathy is situated at the junction between the Preveza peninsula and the Ayios Thomas peninsula (Figs. 5.1, 5.4).

15. Kraft Chrzastowski and 1985; and Kraft, 1985; Kayan, Aschenbrenner
Rapp and Kraft1994. 16. With Walther'sLaw one is able to reconstructancient sedimentary landscapesthrough time and spaceby establishingthe three-dimensional stratigraphic shapes of coastalsedimentarylithosomes or the shapes of sedimentarybodies depositedin discretecoastalsedimentaryenvironments (Middleton 1973). 17. Kraft1985. 18. Kraftand Aschenbrenner1977; and Kraft,Aschenbrenner, Rapp 1977; Kraftet al. 1987; Niemi 1990; Zangger 1991, 1993, 1994. 19. Strab.7.7.5 (C 324), trans.H. L. Jones, Cambridge,Mass., [1924] 1954. 20. Hammond 1967, p. 48. 21. Dakaris 1971, p. 6. 22. Hammond 1967, p. 48. 23. Leake 1835, I1, 195-196. pp.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I63

Figure5.3. Map of the Nikopolis isthmusshowingthe locationof geologiccoresandcrosssections

Figure5.4. Map of OrmosVathy showingthe locationof geologic coresandcrosssection

I64

ZHICHUN

JING AND GEORGE (RIP)

RAPP

Archaeological survey conducted by the project in 1993 and 1994 shows that Roman and Late Antique sites are scattered along the flanks of the ancient bay.Among these the most important is a Roman site, ca. 250 m wide and 900 m long, along the western shore of the bay.It is believed that this site must have served Nikopolis as a harbor town. Leake24presents a sketch map showing that the bay extended furtherto the north in the early 19th century than it does today. Geologic cores drilled on the Nikopolis isthmus and in the area north of Ormos Vathy provide the data for determining the paleogeographic setting of the city and its associated harbor town.
NIKOPOLIS ISTHMUS

On the Nikopolis isthmus, eight cores, with a maximum depth of 13 m, were drilled from the present shore of Mazoma Lagoon to ca. 1.2 km inland (Fig. 5.3).25 Five stratigraphiccross sections based on these cores show the relationshipsof the marine and alluvialdeposits acrossthe Nikopolis isthmus. Two cross sections are parallel to the axis of the isthmus (Figs. 5.5, 5.6), and the other three are perpendicular to the axis (Figs.
5.7-5.9).26

Cross section D-D' is based on three cores (92-03, 93-03, and 92-04) and extends ca. 1.2 km from Mazoma Lagoon (Fig. 5.5). The lowest sedimentary unit found along this traverse consists of deposits in a marine embayment or lagoon. The unit consists mainly of bluish gray (5BG 4/1)27 reduced mud containing marine and brackish gastropoda (Monodonta, Basslerites, ostracoda(Loxoconcha, bivalves(Cerastodema), Cyprideis, Cyclope), and foraminifera (Ammonia beccarii,Elphidium, TrochamLeptocythere), mina). The marine or paralicunit is covered by olive (5Y 5/6, 5Y 4/3) and light olive-brown (2.5Y 5/4,2.5Y 5/6) silt, sandy silt, mud, and sandy mud of alluvial or slope-wash origin. Their contact is sharp and not transitional. Inland (in core 92-04) the lower boundary of the alluvialunit lies at a depth of 7.0 m while near the shore (in core 93-03) it is found at a depth of ca. 4.85 m, showing an increasing thickness away from the shore. The overlying alluvial or slope-wash deposits are separatedinto two parts by a relativelywell-developed paleosol, which is characterizedby its olive (5Y 5/4) and olive-yellow (5Y 6/6) colors, a carbonate-enriched layer (Bk), crumb and blocky structure, and carbonate-enriched remains. Paleosols represent periods of landscape stability because they only form on stable surfaces over a relatively long span of time.28A few ceramic fragments of red color were found in the top part of the paleosol in core 93-03. A very thin (0.35 m) massive darkgray (2.5Y 4/1 to 5Y 4/1) mud is found in core 93-03 overlying the paleosol, most likely deposited in a mudflat environment as seen near the shore today. Cross section E-E' (Fig. 5.6) is located northeast of cross section DIt exhibits generally the same vertical stratigraphicsequence but with D'. some variation.The top alluvialor slope-wash sediments aresuperimposed on a bluish gray mud deposited in a lagoon or marine embayment environment. In the alluvialor slope-wash deposit there is a buried paleosol as

24. Leake 1835, I, p. 187.

25.Twoadditional C92-05 cores, andC92-06,wereattempted the on barrier separates that Mazoma sandy fromtheAmbracian Gulf,but Lagoon
they penetratedonly 1.35 and 1.75 m,

due respectively, to theveryloose

natureof the barriersand. 26. Elevationsarebased on the 1:5,000 Greek Army topographicmaps and measurements using an electronic total station. 27. The color index is based on the Munsell Soil Color Chart. 28. Birkeland1999.

m 12
10

8 6 4 2
0

-2 -4 -6 -8
-10

200

400

600

800

1000

1200

14

clay

sand

"' ii sand-"~~shell I^? ceramic fragments

A ?

carbonate s nodule plant remains

soil

bivalve gastropod

t - - Imud

I l.?:S gravel " 11~0 ?-

radiocarbonte

Figure 5.5. Stratigraphic cross section D-D', parallel to the axis of the Nikopolis isthmus. For core locations,

m 12 93-04

:
10

'
:-'-4-'-.

Tectonic Uplift

8i?
0 - x

a'd

.
0 10 10

,|u d L 1 2
onI/
?

..0. 0

...''

" , allu'.

-0 4.'0

'

\ ""
"

\* N- _ _ O_ _ ^ ^^^ /////.

-\ \ // \/ \ /-was //\ \- \ \ \

* * * - _- ..-. _ %\'x d ~ la^nr^^^^^ 'x *

~-

"02

^T^^^flat^^90

93-0

.~ ~..x

///\/,////baym/ ////^ ^ // 7 \

o /t\

'

- ':' -0 ." " . -..-^ .alluvium ..

2 Lagoon'e ~ -~ ~

|'// ////X4-/////////////..-

wsh ' slope/

'

-10

- 293

200

400 200 400

600 -600

800 800

1000 1000

1200 ~~~~~1200

Figure 5.6. Stratigraphic cross section E-E', parallel to the axis of the Nikopolis isthmus. For core locations, see Figure 5.3; f

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

i67

describedin cross section D-D'. The upperboundaryof the marineor unit paralic lies at a depthof 9.1 m in core93-04,5.3 m in core92-02, and unit 3.8 m in core93-01.The marineorparalic is morethan6.2 m thickin core93-01, and ca. 1.5-2.0 m thickin both cores92-02 and 93-04. In core 93-04 the marine or paralicunit overlies a layer of black (N2/) andverydarkgray(N3/) silt andmudthatis veryfirmandcontains oxide).The blackor very dark manyblacknodules(possiblymanganese graysilt and mud,0.8 m thick,is underlain a veryfirmolive-gray (5Y by nodules(0.1-1.5 cm in 4/2) mud containingcommoncalciumcarbonate a Both of these firm depositsarebelievedto represent Pleisdiameter). tocene paleosol.The top blackand very darkgraypartis an A horizon, andthe lowerolive-gray partis a B horizon. unit in core92-02 restson an olive In contrast, marineor paralic the 5/4 or2.5Y 6/4) silt andsandymudcontaining manycalciumcarbon(5Y ate nodules,some pebbles,and some terrestrial (snails).This gastropoda but olive layershowsa moderately developedblockystructure it is much morefriablethan the bottom sedimentsseen in core 93-04. Basedon its the inclusions,and consistency, low-lyingolive silt is incolor,structure, as terpreted a paleosoldevelopedon slope-washor alluvialsedimentsof the earlyor middleHolocene. unit dateswereobtainedfromthe marineor paralic Fourradiocarbon dates are on samplesfrom core 92-03, lo(Table5.1). Two radiocarbon shoreof MazomaLagoon due east of Nikopolis. cated on the landward The whole columnof the coreconsistsof interbedded verydarkgray(5Y 3/1 or N3/) andbluishgray(5BG 4/1) sandymud,gravelly, muddysand, marineandbrackish mollusks(gastropoda andmudcontainingabundant The and andbivalves) microfauna and (ostracoda foraminifera). layerat a depth of 6.0-7.55 m yielded many culturalremainsincludingcharcoal, bone,burnedwood, ceramicsherds,and plasterfragments. Among these Romansherdsfroma depthof 7.1-7.3 remains some diagnostic are Early m. A charcoal samplefrom6.0 to 6.2 m (5.9 m below modernsea level) radiocarbon date of 650-440 B.P., much youngerthan gives a calibrated the diagnosticsherdsfrom the layer,which suggeststhat the culturally rather than a primary altered deposit.Giventhe geolayeris a secondary remains mighthavebeenwashed positionof the core,the cultural morphic old awayfromthe hill slope at Nikopolis.A relatively date of 5320-4840 B.P. was obtainedfroma shell samplefroma depthof 5.8-5.95 m (5.7 m belowmodernsealevel).Inversion youngerandolderradiocarbon of dates of be attributed the redeposition shells derivedfrom preexisting to may sediments. The top 6 m of core 92-03 must haveformedin the past 650 Thus the average years. depositionratefor the past 1,000 yearsis greater than 9 mm/yearnearthe presentshoreof MazomaLagoon.Such a large to of sedimentation maybe attributed the rapidtectonicsubsidence rate the easternside of the isthmus. A calibrated radiocarbon of 3350-2930 B.P. comesfroma chardate coalsamplefoundat a depthof 5.3-5.4 m (2.4 m belowmodernsealevel) in core 92-02. The sampleis from the top of the marineunit.This date indicatesthat the shoreof the lagoonor shallowmarineembayment was beyondthe locationof core 92-02, 550 m inlandfromthe presentshore,

I68

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

TABLE 5.1. RADIOCARBON DATES FROM THE AMBRACIAN EMBAYMENT

Depth below MSL Core (m) Material NC92-02 2.4 NC92-03 5.7 NC92-03 5.9 NC92-04 +0.8 NC92-07 6.6 NC92-08 3.7 NC92-09 4.2 NC92-09 6.3 NC92-09 7.1 NC92-09 7.5 NC92-10 4.5 NC93-07 1.0 NC93-09 3.6 NC93-11 5.8 NC93-16 4.4 NC94-19 5.1 NC94-19 1.2 charcoal shell charcoal charcoal charredroots charredroots charcoal charcoal wood wood charredroots peat peat peat peat peat peat Sedimentary Facies estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) estuary(mud basin) swamp (marsh) swamp (marsh) swamp (marsh) swamp (marsh) swamp (marsh) swamp (marsh) Conventional Radiocarbon Age (B.P) 3430 ? 70 4940 ? 80 1060 ? 70 6430 ? 70 4290 ? 70 4010 ? 70 970 + 60 840 ? 60 880 ? 70 1250 ? 80 1600 ? 70 4810?60 1670 ?60 5900?70 2510 ? 70 4090 ? 80 1510? 80 28 Max. Calibrated Age (CaL Age Intercepts) Min. CalibratedAge (B.P.)b 3350 (3160) 2930* 5320 (5070) 4840 * 650 (530) 440* 6950 (6750) 6600* 4440 (4240) 4010 * 4080 (3850) 3640 * 560 (490) 330 * 490 (370) 260 * 520 (420) 270 * 880 (670) 540* 1210 (1030) 890 * 5650 (5590) 5330 1710 (1550) 1410 6890 (6720, 6700, 6690) 6500 2760 (2710,2630,2620,2560,2550) 2350 4830 (4570,4560,4550, 4540, 4530) 4410 1560 (1410, 1400, 1390) 1290

Lab No.a UCR-3201 UCR-3202 UCR-3219 UCR-3218 UCR-3220 UCR-3221 UCR-3203 UCR-3204 UCR-3205 UCR-3206 UCR-3222 UCR-2691 UCR-2692 UCR-2693 UCR-2694 BK-94168 BK-94169

UCR = RadiocarbonLaboratory, BK aDatinglaboratory: Universityof California,Riverside; = RadiocarbonDating Laboratory, Peking University. bCalibrated ages obtainedusing CALIB 4.3, developedby Quaternary Isotope Laboratory, Universityof Washington. *Marinereservoircorrectionmade for the estuarysamples,using AR = 118 + 35, a regionalaveragefor the easternMediterranean (Siani et al. 2000).

around 3000

B.P.

Further inland, in core 92-04, a date of 6950-6600

B.P.

was obtained from a charcoal sample at a depth of 7.15-7.25 m (0.8 m above modern sea level)-near the top of the marine or paralic unit in which some microfauna (Loxoconcha, Elphidium) were found in core 9304. This circumstance suggests that the area of both cores 92-04 and 9304 was then covered by seawater. From cross sections D-D' and E-E' we can see that the upper boundary of the marine or paralic unit is sloping upward away from the shore. The upper boundary in both cores 92-04 and 93-04 is ca. 1-2.5 m higher than modern mean sea level. The radiocarbon date from the top of the unit in inland core 92-04 is much older than that in core 92-02; moreover, a much younger date is situated at a much lower elevation in core 92-03 on the shore. Based on the radiocarbon dates on peat samples from buried coastal marsh and swamp in the coastal delta plain north of the Ambracian Gulf, we know that relative sea level was much lower prior to 6,000 or 7,000 years ago (see below). Therefore the relatively high elevation of the marine or paralicunit in cores 92-04 and 93-04 does not indicate that sea level was higher during the formation of the unit than exists today. A reasonableexplanationwould be that the marineor paralicdeposits, formed in the Nikopolis isthmus during the early phase of marine transgression, were elevated by tectonic uplift. As mentioned earlier,the Preveza peninsula- in contrast to the subsiding Ambracian embayment-has been subjected to tectonic uplift since the Pleistocene.29

and 29. King,Sturdy, Whitney 1993.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I69

CrosssectionsA-A', B-B', and C-C' (Figs.5.7-5.9) areperpendicuin lar to the axisof the Nikopolisisthmusand show the variation stratiand The top alluvial shoreto 1.2 km inland. current unitsfromthe graphic slope-washsedimentsincreasein thicknessawayfrom the shore.Cross sectionB-B', locatedca.500-550 m eastof MazomaLagoon,is basedon Both coresshowthe sameupper cores92-02 and93-03, only 150 m apart. with interbedded sediments and basedon alluvial slope-wash stratigraphy a 0.3-0.5 m thick mudflatmud,but they havecompletelydifferentsediand In in ment assemblages the lowerstratigraphy. core93-03 the alluvial by slope-washunit is underlain morethan 7 m of marineor paralicsedimarineor paralicunit is only ments,while in core 92-02 the underlying or on 1.55 m thickandlies on a paleosoldeveloped relatively earlyalluvial in crosssecThis stratigraphic patternis not seen slope-washsediments. end tions A-A' and C-C'. At the northern of crosssectionC-C' (in core unit restson a Pleistocenedeposit. 93-04) the marineor paralic from the top of the marineor paralicunit in date The radiocarbon evidenceand radiocore92-02 is 3350-2930 B.P. Basedon stratigraphic a carbondates,the low-lyingslope-washdepositmay represent fan that fromthe hill slopein the northsometimein the earlyor middle protruded Holocenebut before3,000 yearsago.This fan would havedammedpart The of the lagoonor shallowmarineembayment. fan was subjectto mabetween3000 and3500 rinetransgression fora shortperiod,probably only The B.P. deepestpartof this earlier probably lagoonor marineembayment lay on the south side of the isthmus,as indicatedin crosssectionsB-B' further inlandalongthe andC-C'. Eventhe moderncontourlinesproject southside of the isthmus. of Fromthe subsurface stratigraphy the Nikopolisisthmus,a paleocan be made showingthe changingshorelineover the geographicmap past severalthousandyears(Fig. 5.10). The morphologyof the isthmus might have been createdby tectonic faultingduringthe Pleistoceneor Plioceneperiod.In the middleHolocene,the easternportionof the isththat muswas submerged belowa largeembayment extendedmuchfurther inlandthanthe presentMazomaLagoon.The furthestinlandcoreyielding Holocene marineor paralicsedimentsis core 93-04, 1.2 km west of Mazoma Lagoon,in the middle of the 2.3-km long isthmus.Here the of isthmushas an elevation 12 m, andthe highestportionhas an elevation has of only 16 m. If the tectonicupliftof the Preveza peninsula beenrapid, the isthmusmight havebeen low enoughto be an open channelbetween the IonianSeaandthe Ambracian Gulf duringthe earlyHolocene(before of 6500 or 7000 B.P.) . When marineregression the occurred, shoreline the on the Ambracian eastward. 3000 B.P. Gulf sideprograded By embayment the shoreline was ca. 1 km inlandof the current shore,and by 500 B.P. it was likely less than 400 m inland.The date of formationof the sandy on as barrier the eastside of MazomaLagoonremains unknown, we were for unableto core along its length. On analogywith the Salaora barrier, which we do have some evidence(see below),the Mazomabarrier may at until1000 B.P. orlater. barrier Mazoma The not haveformedcompletely is believedto haveformedby littoraltransport the sandandgravelsediof mentserodedfromthe Pliocene-Pleistocene rockyshorealongthe north side of AyiosThomaspeninsula.

m 12-

-XT'I--

N 1 lKOpOllS

-1r

-10

///

I2^vyyy^

93-02

:1

i^^

':

^ ~w

*-10:':' -:*;"':

-4alluvium 311"^""1"'^

-^

--92-01

*mudflat -_ -4-,

200

400

600

800

Figure 5.7. Stratigraphic cross section A-A', perpendicular to the axis of the Nikopolis isthmus. For core locations, see Figure 5

12

10

-2

-4

-6

-8

-10
0 100

200

300

400

500

600

Figure 5.8. Stratigraphic cross section B-B', perpendicular to the axis of the Nikopolis isthmus. For core locations, see Figure 5.

20

18

16

14

12

10

-2
0

100

200

300

400

500

600

Figure 5.9. Stratigraphic cross section C-C', perpendicular to the axis of the Nikopolis isthmus. For core locations, see Figure 5.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I73

Figure5.10. Paleogeographic reconstruction the easternside of of the Nikopolisisthmusshowingthe shorelinesat differentperiods: 6500/ 7000 B.P., 3000 B.P., and500 B.P.

30. Rapp 1986.

In the past 5,000 years or so, at least two distinct phases of alluvial or slope-wash sediments and their associated soil development occurred on the isthmus. Natural factors, such as tectonic movement and climatic fluctuation, as well as human impact might have caused hill erosion on both sides of the isthmus. To determine the exact timing of hill erosion and associated deposition would require more data. These slope-wash and alluvial sediments covered the preexisting marine or paralic unit, elevating the surface of the isthmus by ca. 5-10 m. During the Roman period, the city of Nikopolis might have been closer to the shore than today, facing a larger area of water.The barrierof Mazoma Lagoon might not have been fully formed and thus the lagoon could have been a very well sheltered harborfor the city. During the occupation of Nikopolis, the surfaceof the isthmus northwest of the city might have been 3-6 m lower than today.Continued uplift of the Preveza peninsula may have led to frequent earthquakes,which could have destroyed many structuresin Nikopolis as well as in neighboring towns.30 From the above discussion, we can see that shoreline changes on the Nikopolis isthmus are the result of dynamic interactions among a variety of factors, including a rise in relative sea level, tectonic movement and subsidence, and hill erosion and associated deposition. The rapid tectonic uplift of the Preveza peninsula constitutes the dominant factor leading to the gradual shoreline progradation during prehistoric and historical periods.

I74

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

ORMOS

VATHY

from the Ambracian OrmosVathyis a narrow protruding Gulf into bay the junction between the Preveza and the Ayios Thomas peninsulas (Figs. 5.1:7, 5.4). The bay is ca. 1 km long and 120-450 m wide, and is narrowest its middle.It is surrounded hills of relatively relief at low by composedof Pliocene and Pleistocenedeposits.A low ridgewith an elevationof 12 m projects into the northend of the bay,creating extensions (arms)of the inlet on either side of the projectingridge.The northern The eastarmhas a larger edgesof both armsarefringedby coastalmarsh. areaof fringe marshthan the west arm as it has a muchwider areaenclosedwithin contourlines of low elevation(below4 m). Given its location and size, the bay must have been a well-protected harborbeforea modern andcauseway builtacross openingto theAmbracian were its bridge Gulf. Fourdrillcoresweretakenin the two armsof the bay. Threeof them weredrilledalongthe west arm,and one on the east arm(Fig. 5.4). Core 93-07 was locatedin the marsh,360 m north of the currentshore.The corepenetrated only 5.17 m, and consistsmainlyof bluishgray(5B 5/1) andvery darkgray(N3/) organicmud and muddypeat that formedin a coastalmarshenvironment. decayed The dominantly grassfroma muddy peat layerat a depth of 1.88-2.00 m (ca. 1.0 m below modernsea level) radiocarbon of 5650-5330 B.P. The relatively date yieldeda calibrated high fora dateof this agemaybe attributed the tectonicupliftof the to position Prevezaand Ayios Thomas peninsulas. Partof the coastalmarshareais morethan2 m higherthanthe current levelwhilethe Ambracian sea Gulf has a very low tidal range(5 cm on average). Tectonicuplift must have been a significantfactorin the retrogradation progradation the and of marine duringprehistoric historical and times.Assumingthatthe two bay armsof the bayhad the samerateof upliftas the surrounding ridge,relative sea level around5500 B.P. was ca. 1 m lowerthan today. Figure5.11 is a crosssectionbasedon threecores:94-10, 94-06, and 93-08. The stratigraphic sequence beginswith a Pleistocenedeposit,seen in the furthermost inlandcore, 94-10, ca. 600 m from the current only shoreof the bay. The depositis composedof olive(5Y 5/4) sandymudand siltcontaining manycalcium carbonate nodules. the Overlying Pleistocene is greenishgray (5G 5/1 and 5BG 5/1) and very greenishgray deposit (5GY 4/1) soft muddysand,sandymud,and mudrepresenting shallow a marineembayment to depositioncomparable thatseenin the modernbay to the south.Few microfauna were found in the samplescollectedfrom this marineunit,but someforaminifera Ammonia becspecies(Elphidium, carii)were identifiedin core 94-10. The upperboundaryof the marine unit is 1-1.5 m higher than currentsea level as a resultof continuous tectonicupliftduringthe Holoceneperiod. The marine unit is overlain a 1-m thickcoastalmarshdepositconby sisting of bluishgray(5B 6/1) and darkgreenishgray(5G 4/1) organic mud and humifiedgrass.The final unit of the sequenceis a slope-wash depositconsistingof yellowishbrown(10YR 5/6), darkyellowishbrown (10YR 4/4), and olive brown (2.5Y 4/4) silt and mud with common

12

41

200

300 500-

400 slope wash600

of swamp

700

Figure5.11. Stratigraphic crosssection along the

Ormos V

- .:-

_---

I76

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

Neolithic Period (ca. 5500 B.P.)

Roman Period (ca. 2000 B.P.)

t/ SK^^y^^ =
|>^ I'V~^:^^%/ 0 : 400 M

Figure5.12. Paleogeographic
reconstructions of Ormos Vathy indicating shoreline changes from

............... .^ ......the

Neolithicthroughmodern
periods. Tectonic activity was a

Modern Period
calcium carbonate nodules and some pebbles. These calcium carbonate nodules are not in primary pedogenic context and thus do not indicate a well-developed soil associated with the slope-wash deposit. The nodules were derived from a Pleistocene soil on the surrounding ridges and were deposited in the bay as a result of hill erosion. A few small red ceramic fragments were found in the middle of the unit. A Turkish limestone struc-

dominantfactorin shoreline progradation.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I77

wereencountered a depthof 0.5 m in at tureand some redtile fragments core93-08. fromthe stratigraphic date secAlthoughno radiocarbon is available estimateon the timingof the Holocenedepositioncan tion, a preliminary be made.As discussedabove,sea level relativeto the surrounding ridge This level is a little was ca. 1 m lower around5500 B.P.than at present. higherthan the bottomof the marineunit in core94-10, suggestingthat the embayment extended beyondthe corealongthe west armbefore5500 After the sea-levelrise slowed,the rate of tectonicuplift apparently B.P. exceededthe rise in relativesea level. As a result,the shore of the bay movedgradually and seaward, a coastalmarshstartedto form along the of fringeof the bay.The majority the slope-washdepositseems to have formedafterthe extensiveoccupation the Romanperiod. in 5.12 is a paleogeographic reconstruction OrmosVathy, of showFigure shorelines fringemarshes and fromthe Neolithicthrough ing the changing
the present. By 5500 B.P.the marine transgressionhad extended more than

750 m inlandof the current shorealongthe west armof the bay,but the eastarmwasoccupied fringemarsh insteadof baywater. otherwords, In by the baydid not extendveryfarinlandalongits easternarm.As the risein sea-levelslowed,the continuous tectonicupliftmovedthe shoregradually seaward. the Romanperiod,the shoreof the bay mayhaveprograded By to townwasbuilt halfway the gulf alongthe west arm.The Romanharbor the alongthe west shoreof the bay.Becauseit waswell protected, baymay havebeenone of the mostimportant harbors the cityof Nikopolis. serving In additionto the majorportionof the bay,the extendedwest armmight havebeen wide enoughto providegood anchorage. Extensivehabitation sincethe Romanperiodon the hills surrounding OrmosVathyhasled to erosion.As a consequence, to increasing slope-washsedimentsstarted fill in the west armof the bay,accelerating progradation the shoreline. the of Much of the east armhas remained coastalmarsh,havingreceivedmuch less sedimentfrom slope-washprocesses. This may be attributed the to gentle slope of the surrounding ridge and to the possibilitythat sparser on occupation this side of the bayresultedin less erosion.
GRAMMENO PLAIN

31. Archaeologicalsurveyundertaken by the Nikopolis Projectin 1992 and 1993 suggests that sites of various periods,including a large Roman site nearthe modernvillage of Archangelos and two ByzantineandTurkishsites, aredistributedmainly along the hilly edge of the plain.

The Grammeno plain is a tectonic lowland that cuts through the Preveza peninsula near its northern end. It is covered with Holocene alluvial and slope-wash sediments. To the east it merges with the floodplain of the Louros River.Two cores were drilled on the east side of the plain to determine how far westward the maximum marine transgressionextended and to show the changing landscape of the plain (Fig. 5.2).31 Figure 5.13 is a stratigraphiccross section (II-II') based on cores 9312 and 93-13. The marine estuarine or lagoonal unit occurs only in the lower part of core 93-13, ca. 800 m east of the Louros River.The estuarine or lagoonal unit is composed of greenish gray (5GY 4/1) soft mud with some olive (5Y 5/3) oxidized mottles. It contains many microfaunal remains including ostracods (Cyprideis, and Loxoconcha) foraminifera(Elphidium,Ammonia).The upper boundary of the unit is clear but more or less transitional.

I78 m

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

12
10

- E
8
6 4 2
0

-2 -4

-6
-8
-10

0 500 1000

1500 2000 2500

m

The marine estuarine or lagoonal unit is capped by 3.75 m of alluvium showing a moderately developed soil profile. The top 1.2 m of the profile is light brownish gray (2.5Y 6/2) to dark gray (5Y 4/1) silt and sandy silt with fine blocky structure.The silt contains decayed rootlets and many iron oxide mottles. The top layer is an A horizon. It is underlain by a 1.5m thick weak B horizon that is an olive gray (5Y 4/2) blocky silty clay containing calcium carbonate nodules. Underlying the B horizon is olive gray (5Y 5/2) massive clay and sandy mud with some iron oxide mottles. The sediments of this alluvialunit could be derived from both the Louros River and small streams flowing on the Grammeno plain. Core 93-12 is located ca. 870 m southwest of core 93-13. No estuarine or lagoonal deposit occurs in this core. The sequence begins with pale olive (5Y 6/4) and yellow (5Y 7/4) firm to very firm clay and clayey silt with a 0.15-m thick very darkgray (N3/) clay on top. This unit occurs at a depth of 5.95 m, and may represent a soil formed in a Pleistocene alluvial deposit. The basal unit is covered by two phases of Holocene alluvial deposits, both of which show moderate soil development characterized by carbonate-enriched Bk horizons. These two alluvial deposits and associated soil profiles may correspond to those top slope-wash units observed on the Nikopolis isthmus in terms of the timing of the hill erosion that provided the sediments.

cross Figure5.13. Stratigraphic sectionnearthe Grammenoplain. Forcorelocations,see Figure5.2 (sectionis labeledII-II');for legend, see Figure5.5.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I79

in observed thesetwo coresindicatesthat The stratigraphic sequence did the Holocenemarinetransgression not extendveryfarinto the Gramformedby the premeno plainbecauseof the relatively high topography somewhat reached The beyond transgression existingalluvium. maximum channel as the locationof core93-13, perhaps faras 1 km fromthe current As of the LourosRiver. in the areato the northof the Ambracian Gulf, in the marinetransgression the Grammeno plainmight havereachedits after1500 B.P. as maximumaround4500 B.P., with gradual progradation alluvialand slope-washsedimentsfilled in the estuaryat an increasing hill sedimentsupplymight be due to accelerated erorate.The increased sion causedby humanimpact.
KASTRO ROGON AND STRONGYLI

32. Hammond 1967, p. 427. 33. Dakaris 1971, p. 42. 34. Dakaris 1971, p. 178; Hammond 1967, p. 427. 35. The KastroRogon site and its surrounding floodplainwere surveyed by the Nikopolis Projectin 1992 and 1994. Archaeologicalfinds rangefrom the Classicalthroughpost-medieval
periods (early 5th century B.c.-19th

centuryA.C.).

Kastro Rogon is ancient Bouchetion, an urban site situated on the top of a Jurassiclimestone hill near the northern edge of the coastal delta plain of the Ambracian Gulf (Figs. 5.1:15, 5.14). It was firstbuilt as a colony around 700 B.C.,32but the earliest wall might not have existed until the early 5th The site was occupied until the medieval period, and addicentury B.c.33 tions and repairs to the circuit walls were made throughout its history. Bouchetion was one of four important walled settlements of the colonists from Elis during the Classical and Hellenistic periods (late 5th century to 168/7 B.c.), and it remained an important site after the Roman conquest because of its strategic position.34 The hill of Kastro Rogon, ca. 65 masl, is located south of Mt. Rokia and stands isolated from other hills to the east and north. The Louros River flows along the southwest side of the hill and then turns to the north at the hill's northwest corner. Across the Louros River to the south and west are a floodplain and reclaimed swamp (previouslybrackish estuarine marsh)with a relativelylow elevation (0-4 masl). To the east of Mt. Rokia, the delta floodplain rises gradually northward toward the deeply incised valley of the Louros River.35 Strongyli, a Roman-period villa rustica, is situated on a small ridge north of Koryphi, the northernmost protuberanceof Mt. Mavrovouni located in the low-lying estuarine swamp (Figs. 5.1:16, 5.14). The Strongyli site is ca. 3.2 km southwest of Kastro Rogon. In addition to Roman remains, there are remains of earlier and later times, including the Hellenistic and Late Byzantine or Turkish periods. A total of seven cores were drilled in the area around Kastro Rogon and Strongyli. Based on these cores, three stratigraphiccross sections were constructed to interpret the evolving sedimentary environments in this historically strategic location. Cross section C-C' (Fig. 5.15) consists of two cores, 94-15 and 94-19, near the hilltop site of Kastro Rogon. Both cores are on the right bank of the Louros River (see Fig. 5.14). The bottom unit of the stratigraphicsequence shown in the cross section is a very darkgray (N3/) muddy peat consisting mainly of humified grass.The peat layer is found at a depth of 7.5 m (about 5 m below current sea level) in core 94-19, located only 130 m from the 10-m contour on the hill to the northeast. The layer has been radiocarbondated to 4830-4410 B.P. and is believed to representcoastal fringe swamp before the maximum transgres-

I80

ZHICHUN JING AND GEORGE (RIP) RAPP

I |I

Mesozoic limestone v vV Neogene-Pleistocene1 9211 v hill at 20 m contour v v v M hill at 20 m contour ? 0

_T

geologic core 2
_

f
3 km

swamp
(reclaimed) [

cross sectio

sion around 4500 B.P.This peat layer is overlain by greenish gray (5GY 6/ 1, 5G 5/1, and 5BG 5/1) soft estuarine or lagoonal mud containing some brackish ostracoda species (mainly Cyprideistorosa)and a few freshwater species (Candona) in its upper portion. Within the estuarine or lagoonal unit there are two intercalated layers of dark brown (7.5YR 3/2) muddy peat and peaty mud seen in core 94-15, located 150 m north of the Kastro Rogon hill. These two intercalated peat layers indicate that part of the foothill area of Mt. Rokia could have been intermittently swampy during the period of maximum marine transgression,ca. 4500-1500 B.P. In core 94-19 the estuarine or lagoonal unit is overlain by 1.4 m of olive (5Y 5/4) gravelly sandy mud containing many pebbles and secondary calcium carbonate nodules along with some plaster fragments. The gravelly unit has very sharp boundaries with both underlying and overlying sediments, and it may represent a small colluvial deposit derived from slope washes. Alternatively, this unit may form part of a causeway built to connect the Kastro Rogon site to the mainland. If this was the case, the stratigraphy indicates that the causeway was built before 1500 B.P., most likely during the Roman period. More evidence would be needed to substantiate this hypothesis. On top of the estuarine and lagoonal unit lies a greenish gray (5G 6/1 and 5GY 6/1) muddy peat and peaty mud containing abundant

Figure5.14. Map of KastroRogon andvicinityshowingthe locationof geologiccoresandcrosssections

4I'-

3
*2.'

9415 <94-1519
-

*.. ., ? .

slope wash

< ...

.-

--

''.

.".

r,~

_.......

ai__

alluvium

. ?'

'.

':-

, '_|

-2

\^, 2 1 I

'-v '-'_ _.:__ X\ \ \ \\ [....i-.c.\lt\ura\l<..\ -

_--. '.swamp',' , ~

..--

"

1 560- 1,290 B.P.

',estuary /lagoon\ \ \ \\ .X

- "'---.
. '-f

'

\ \sop" wx l \ ''. ...?..\\....

o

3 '~%~~~~7

& '...... \estuary/lagoon'

H .

, --|..-5 ._ _._^ -.....-

..

'

. --

.-,

'\ ~'~

' , . ..-. a, 4,830 -4,410 B.P.

Figure 5.15. Stratigraphic cross section C-C' at Kastro Rogon. For core locations, see Figure 5.14; for legend

I82

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

decayed grass debris (very pale brown, 10YR 7/4). The unit is ca. 20-60 cm thick and is found in both cores. Ostracoda from this unit include both brackishand freshwaterspecies (Cyprideis, Candona).Based on a calibrated radiocarbon date of 1560-1290 B.P., we interpret this layer as a brackish swamp deposit formed along the northern fringe of the ancestral lagoon associated with the Ambracian Gulf after the end of maximum marine transgression, probably around 1500 B.P. Clearly, the Kastro Rogon hill was an island in a marine estuary during the period of maximum transgression, from 4500 B.P. to 1500 B.P. The swamp unit is capped by a 2.5-m thick floodplain alluvium, consisting of light olive brown (2.5Y 5/4) and olive (2.5Y 5/4) silt and silty clay.The lower boundary of the unit is gradual,as indicated by the downward increase of light greenish gray (5GY 7/1) mottles. The lower part of the alluvialunit yielded some freshwatergastropoda and ostracoda species (Candona).The top part of the unit shows weak blocky structureand contains some fine calcium carbonate nodules, representing a weak cumulative soil profile. Overlying the floodplain alluvium is 0.75-0.9 m of strong brown (7.5YR 5/6) silt and silty clay with rootlets and some limestone pebbles. This top strong brown deposit occurs only on the Rogon side of the Louros River; the top deposit on the floodplain side of the river is olive (2.5Y 5/4) silty clay and clay (see Fig. 5.17). The strong contrast in color suggests different origins for these deposits.The olive deposit on the floodplain side must have developed from overbank deposition of the Louros River, while the strong brown silt and silty clay might derive from the weathered hill slopes by sheet erosion. The strong brown deposit must have formed after the Louros River flowed in its current channel because it occurs only on the Rogon side of the river.Thus, we are able to determine when the Louros River started flowing in the current channel. We know that the floodplain alluviumoverlying the post-transgression swamp (dated to 1560-1290 B.P.) formed after 1500 B.P. A certain amount of time was needed to form the 2.5-m thick alluvium underlying the top strong brown deposit. It seems reasonable to suggest that the diversion of the Louros River into the currentchannel did not occur until the 10th century A.C. or later. On the floodplain side of the river,many culturalremains of the postmedieval period (middle 15th-19th century A.C.) were found in a plowed field 100 m from the river.These post-medieval remains enclosed in the top floodplain deposit suggest that the Louros River was diverted into the current channel before the 15th century A.C. On the Rogon side of the river,many artifactswere found in the low-lying area surroundingthe fortified site. Although most of these artifactsarepost-medieval in date, some belong to the Classical through medieval periods. We believe these older artifacts were eroded from Kastro Rogon after the Louros River was diverted. A channel diversion sometime between the 10th and 15th centuries might explain the distribution of archaeological remains on the two sides of the river. Figure 5.16 is a stratigraphiccross section (B-B') located in the central part of the deltaic floodplain formed by the Louros River as it enters into the Ambracian embayment (see Fig. 5.14). Along the central part of the floodplain, the surface rises rapidly toward the northeast. Core 94-16

< 't 4'

;
COASTAL EVOLUTION OF THE AMBRACIAN

EMBAYMENT

I83

12
10 8

,-

SW

6 4 2
0 -2 -4

:alluvi*um .y ^1-N J
J^^ (deltaic).i-|:<

I I

.:>

<. .

? .

1 - ; /.aluiu . -4 j -^ <,,4,,-~~~~~~~~~~~~~~~.
--, . W__,_ A_r<^
5

' _W_,

4' -

<

. 4

p
_sf
i

j
, , -

<

swamp (reclaimed)
*^ k *
AI

>-

' -- 1 .;i.<. .4._-. .,-;.-..v..

s~~~~~~~~~~- -" ^ .^

11IN,s

ll

^ ^1-, ~

^--

lagoon/ estuary
_ -n

,,

,,

j ie

A.

s A

,k

as? ',
IP -IL_

_,V_ -Al

S-

A -!1k.

-6
*v > w > 4 S q 4 ^ e .
41,

*
AA

nearshore 'J . .^' n , *A

. 4A ..

-Lr

_ v

-,4

-8
A

k swampI v
11

^~ ~
-W

~~~.. W..

-Al- - st
sbw Al

"if A _W_
1

Av A

l:->t^

-10

~.-

800 1600

u 3200

llll

llll 4800

ll

2400

4000

5600 m

cross Figure5.16. Stratigraphic sectionB-B' nearKastroRogon.For corelocations,see Figure5.14;for legend,see Figure5.5.

was drilled in reclaimed swamp with an elevation of 1 masl. The top part of the core consists of a 1.9-m thick light yellowish brown (2.5Y 5/4) massive silt, sandy silt, and silty sand with some mud laminations. Within the top unit, gray (N6/) mottles increase downward. The unit is interpreted as a swamp deposit formed at the front of the deltaic floodplain. In core 94-18 the top unit is a deltaic floodplain deposit composed of 3.4-m thick olive yellow (2.5Y 6/6) and dark grayish brown (2.5Y 4/2) silt, clayey silt, sandy silt, and silty sand containing many decayed grass rootlets and freshwaterostracodaspecies (Candona).Underlying the swamp or deltaic floodplain deposits in both cores is a 4-5 m thick estuarine or lagoonal unit that consists of greenish gray (5BG 5/1) interbedded sandy mud, mud, and muddy sand. The unit contains both brackish and freshwater ostracoda species (Cyprideis,Candona),but more freshwaterspecies appearin the upper part of the unit. Based on its stratigraphiccontext, the estuarine or lagoonal unit must be the product of maximum marine transgression (4500-1500 B.P.). The estuarine or lagoonal unit rests on a swamp deposit composed of greenish gray (5G 6/1) and bluish gray (5B 5/1) peaty mud and darkbrown (7.5YR 3/2) peat and peaty mud. This swampy unit is believed to have formed before maximum marine transgressionand it may extend quite far into the previous valley of the Louros River. A 0.7-m thick gray (N6/0)

I84

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

within the swampyunit in core 94-16. The gravellysandis interlayered microfauna gravellysand layer contains many brackishand freshwater and foraminifera ostracoda(Cyprideis, Candona), includinggastropoda, a It (Elphidium). may represent nearshore depositformedduringa relativelyshortperiodof sea transgression. cross Stratigraphic sectionA-A' (Fig.5.17) is basedon fourcores,9419, 92-11, 92-07, and 92-08, betweenMt. Rokiain the north and Mt. in Mavrovouni the south.The surfacealongthis traverse dips gently towith floodplain half wardthe south.The northern of the sectionis covered and alluvium, the southernhalf is occupiedby reclaimed swamp.Kastro end Rogon is locatedat the northern of the section,and Strongylilies at the southernend. The stratigraphic sequence beginswith a basalswampunitin core9419 at the northernend of the section (see above).Corresponding the to end The swampunit is a nearshore depositat the southern nearStrongyli. lies at a depthof 3.6 m in core92-08 and4.15 m in core nearshore deposit 92-07 and it consistsof darkgray(N4/0), gray(N5/0), andgreenishgray (5GY 5/1) interbedded sandymud,muddysand,silt, andmudwith some unit is very rich in marineand thin shelly sand lenses.This nearshore brackish foraminifera bivalves, (Monodonta, (Trochamgastropoda Cyclope), and mina, torosa, Ammonia), ostracoda Protelphidium, (Cyprideis Elphidium, dates on Two Loxoconcha). radiocarbon weremeasured charred grass samples fromthe nearshore unit.The samplefroma depthof 6.68-6.75 m in core 92-07 gavea calibrated of 4440-4010 B.P. The samplefrom3.8 to 3.9 date m in core 92-08 dated to 4080-3640 B.P. Both dates suggest that the nearshore depositformedduringmaximummarinetransgression beginning around4500 B.P. As statedearlier, estuarineor lagoonaldepositrests on the basal an unit in core94-19. This estuarine lagoonalunit is seenin all the or swamp coresacrossthe section.It consistsof greenishgray(5GY 5/1 and 5BG somebrack5/1), gray(N5/0), anddark gray(N4/0) softmudthatcontains ish ostracoda foraminifera and (Elphidium), veryfew species(Cyprideis), freshwater ostracoda(Candona). The majorportion of the estuarineor problagoonalunit formedduringthe periodof maximum transgression, or ablybetween4500 B.P. and 1500 B.P. In core 92-08 the estuarine launit is only 0.6 m thick andis covered a 1.8-m thicknearshore goonal by of mudwithabundant brackdepositcomposed sand,shellysand,andsandy ish shells.The dominance nearshore of faciesin core92-08 maybe attributed to its locationon the edge of Mt. Koryphi. During the Romanpenearshore environment could have provided riod, the seashell-enriched food resources the inhabitants Strongyli. for of important Overlyingthe estuarineor lagoonalunit is a swampdeposit that is buriedby floodplainalluviumin cores 94-19 and 92-11 and crops out southward both cores92-07 and92-08. In core94-19 the swampy in unit is muddypeatandpeatymuddatedto 1560-1290 B.P. The swampformation startedat the northern of the sectionafterthe end of maximum end marinetransgression 1500 B.P.) and movedgulfward fluvialsedias (ca. ments fromthe LourosRiverfilled in the estuaryand gradually covered the swamp.

-0
m

Rokia

Mav

12
10

8 6 4 2
0

-2 -4 -6 -8
-10 0

400 800 1200

m
1600 2000 2400 2800 3200

Figure 5.17. Stratigraphic cross section A-A' near Kastro Rogon. For core locations, see Figure 5.14; for legend, se

i86

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

of evolution the Kastro the Beforediscussing paleogeographic Rogonwe firstneed to examareaand its archaeological implications, Strongyli crosssectionacrossthe whole coastalplain-lagoonine the stratigraphic Gulf so thatwe canplaceour barrier systemto the northof the Ambracian in the contextof the whole embayment Fig. 5.2). Fig(see interpretation ure5.18 is a crosssectionbasedon five cores:93-11, 93-09,92-10, 92-09, end and93-16. The northern of the sectionis at the foothillof Mt. Rokia This traverse showsa very barrier. and the southernend is at the Salaora from the foothill in the north to the edge of Rodia gentle topography Lagoon. sandwith The stratigraphic beginswith a 2-m thickgravelly sequence to manyangular subangular pebblesseen only at the base of core 93-11, The basalgravelly located350 m southof the 10-m contourof Mt. Rokia. the sandis of fluvialor colluvialoriginand constituted pre-transgression This sandlayeris oversurface alongthe edge of the tectonicembayment. lain by a 0.9-m thick swamplayercomprisedof darkyellowishbrown (10YR 3/4) peat and peaty mud.The peat deposit dates to 6890-6500 thanthe dateobtainedfromthe basalpeatlayer B.P., ca.2,000 yearsearlier in core94-19 (4830-4410 B.P.). age difference The maybe due to a lower elevationassociated with the formerdate (see Table 5.1). Moreover, the of datesfor this basalpeat layerimpliesthat relativesea level largerange roseveryslowlyfrom7000/6500 B.P. to 4500 B.P., allowing peatto develop in the coastalfringeswamp. On the basalpeatlayerlies an estuarine lagoonalunit composedof or a darkgreenishgray(5BG 4/1 and 5G 4/1) soft mud interbedded with or varisandymudandmuddysand.The estuarine lagoonalunit contains able amountsof decayedplant remainsand marineand brackishfauna suchasforaminifera and Xesto(Ammonia, Elphidium) ostracoda (Cyprideis, Fromthis unit moremicrofauna foundin landward are leberis, Basslerites). in cores,particularly cores93-11 and 93-10 (not shownin crosssection), and moreplantremainsareseen in coreson the lagoonside,especially in core92-09. A dateof 1210-890 B.P. was measured a wood sampleat a on depth of 5.15-5.25 in core 92-10. Core 92-09 yieldedfour radiocarbon dates.The corewas drilledin the swampon the edge of RodiaLagoon. The top 1.75 m of this core is darkreddishbrown(5YR 3/2) peat and muddypeat. Underlyingthe peat unit are darkgray (N4/), gray (N5/), and greenishgray(5GY 5/1) interbedded sandymud, muddysand,and sandcontaining All datesareyounger than plantremains. fourradiocarbon 900 B.P., indicatingincreasing sedimentation ratesfrom the foothills to the lagoonwith gradual infillingof the lagoon. The estuarine lagoonalunit is overlain a swampunit.The lower or by of unit a southward, boundary thisswamp risesgradually suggesting gradual of post-transgression withincreased estuarine progradation swamp infilling. At the northern of the section,a layerof peatymud(0.6-0.9 m thick) end seen in both cores 93-09 and 93-11 constitutesthe bottom part of the ostracoda swampunit.The peatylayeryieldedmanybrackish (Cyprideis) andforaminifera A radiocarTrochammina, (Elphidium, Cribroelphidium). bon samplefrom core 93-09 datesthe peatylayerto 1710-1410 B.P. As discussedabove,the top peat deposit in core 94-19 yielded a calibrated

18 16 14 12 10 8 6 4 2
92-10

1' S

Rodia
_-2"

92-09 JA
,19 _ e

L Lagoon
--

Tsouk Lagoo

0 -2

-=

swamp
,(partially reclaimed)
v "

back bar . ,710-1,410 B..P '

swam

-4

^^^:$. \\
-6 -8 -10 -12 -14
0 0 2 ? I I ? I ? 10 i

12

Figure 5.18. Stratigraphic cross section north of the Ambracian Gulf showing sedimentary sequences and environment coastal plain-lagoon-barrier system. For core locations, see Figure 5.2 (section is labeled I-I'); for legend, see Figure 5.5

I88

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

radiocarbondate of 1560-1290 B.P. (see Fig. 5.15). Given their sedimentary and stratigraphiccontext, these dates suggest that maximum marine transgressionended around 1500 B.P. In both cores 93-11 and 93-09, the upper part of the swamp unit is composed of greenish gray mud (5BG 5/1) with some muddy sand and sandy mud laminations. It is overlainby a top floodplain alluvium.Toward the south the swamp unit crops out and extends to the edge of Rodia Lagoon. Top alluvium, 4.0-5.5 m thick, is found only in cores 93-11 and 9309 on either side of the Louros River.It thins seawardand merges into the swamp in the south. The upper part of the alluvial unit is olive brown (2.5Y 4/4) and light olive brown (2.5Y 5/4) silt and silty clay with a very weakly developed soil profile on the top. The lower part consists of olive (5Y 5/6) and olive yellow (5Y 6/6 and 2.5Y 6/6) silt, sandy silt, and silty clay with gleying mottles increasing downward (5GY 4/1 and 5Y 6/1). The alluvialunit started forming after the end of maximum marine transgression, probably around 1500 B.P. The top part of the unit likely formed from overbank sedimentation of the Louros River that started flowing along the northern edge of the embayment after the 10th century A.c. Other streams emerging from mountain valleys to the north and northwest might also have contributed a significant amount of sediment to the formation of the lower part of the alluvial unit. Core 93-16 was taken on the lagoon side of the Salaora barrier.The barrierprojectslandward.At the east end it is connected to SalaoraIsland; at the west end it is attached to the Preveza peninsula. The top deposit in the core is composed of 2.2 m of shelly sand, sand, and silty sand containing abundant shells. The next 2.1 m is dark greenish gray (5GY 4/1) silty mud and mud with common shells and some decayed plant remains. At a depth of 4.3-5.0 m is a back barrier swamp deposit consisting of dark brown (7.5YR 3/3) peat layers interbedded with dark greenish gray (5GY 4/1) muddy sand and sandy mud. The back barrierswamp unit is superimposed on an estuarine or lagoonal unit composed of dark greenish gray (5G 4/1) interbedded sandy mud and muddy sand with common thin sand laminations. Marine and brackishfauna are common in the estuarine or lagoonal unit, including ostracoda (Basslerites, Loxoconcha, Xestoleberis, Ammonia,Elphidium). Cyprideis)and foraminifera (Trochammina, A radiocarbondate of 2760-2350 B.P. was determined on a peat sample in the back barrierswamp unit at a depth of 4.5-4.7 m, suggesting that the overlying barrierunit started forming after 2500 B.P. Alongshore deposition rather than offshore deposition is most likely responsible for the barrier's formation.Thus the barrierstarted developing from either or both ends by alongshore transportof the sand and gravel sediments eroded from the Preveza peninsula and SalaoraIsland, probablyaround 4500 B.P. when maximum marine transgressionwas reached. The barriermigrated laterally as the sea level gradually rose. The radiocarbon date from the back barrierswamp unit in core 93-16 suggests that the barriermight not have migrated to the location of core 93-16 until 2500 B.P.36

36. Core 93-16 was drilledin the middle of the centralbarrierisland.A Turkishmilitarymap publishedin 1900 shows that at that date therewas still a large opening in the western part of the barrier. believe that the previous We or estuarywas open to the lagoon AmbracianGulf.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

189

37. Strab.7.7.5 (C 324), trans. H. L. Jones, Cambridge,Mass. [1924] 1954. 38. Hammond 1967, pp. 61-63; Dakaris 1971, pp. 57, 178, 180. 39. Hammond 1967. 40. Dakaris 1971. 41. Dakaris 1971, p. 6: "Pseudo32) Scylax(Periplous wrote in 380-360 BC that the shorebetween the mouths of Louros and Arachthoswas 40 stadia 8 wide (approximately km)."See also Dakaris 1971, fig. 9.

With all available subsurfacedata from the coastal plain-lagoon area north of the Ambracian Gulf, we can reconstructthe paleogeographic setting of both Kastro Rogon and Strongyli. Strabo describes Kastro Rogon (Bouchetion) as follows: "NearCichyrus is Buchetium, a small town of the Cassopaeans, which is only a short distance above the sea; also Elatria, Pandosia, and Batiae, which are in It the interior."37 is easy to understand that Elatria (Palaiorophoros),Pandosia (Kastri), and Batiae (Kastro Rizovouni) are "in the interior":Palaiorophoros and Kastro Rizovouni are situated in mountainous highlands and Kastri is a hilltop site located well inside the Acheron valley (see Fig. 5.1). But it is harder to reconcile the description of Bouchetion as lying "only a short distance above the sea."The hilltop site of Kastro Rogon is currentlylocated well inland. The direct distance between Kastro Rogon and Salaorais ca. 13 km, and the distance along the Louros River is more than 20 km. Historically, Kastro Rogon was believed to be a port serving two urban settlements-Batiae and Elatria-during the Classical and Hellenistic periods.38Neither Hammond39 nor Dakaris40suggests that the port was located on the sea coast. Instead, both scholars believe that the port was linked to the Ambracian Gulf by the Louros River, and that the lower portion of the river was navigable.This belief is based on the assumption that currentgeomorphic elements existed in antiquity as well, an assumption we have shown to be incorrect. Figure 5.19 shows the evolution of the paleogeographic setting near Kastro Rogon and Strongyli based on the subsurface stratigraphic data discussed above.During maximum marine transgression,ca. 4500 B.P. (Fig. 5.19:b), the shoreline was at the foot of Mts. Stavros and Rokia, at the northern edge of the Ambracian embayment, thus making islands of previously inland hills. Mt. Mavrovouniwas the biggest of these islands.Kastro Rogon also became an island during this period, but it was very close to the mountainous mainland. The town of Bouchetion was situated on the top of the island, 65-75 masl, during the Classical, Hellenistic, and Roman periods.This geographic setting fits well with Strabo'sstatement that "Buchetium... is only a short distance above the sea."Thus, Kastro Rogon was a logical site for a seaport and it held a strategic position within the embayment. Our analysis also revealed evidence for the changing course of the Louros River.During maximum marine transgression,the marine embayment probably extended inland along the river channel after it emerged from the deeply incised valley in the north (Fig 5.19:b). During the historical periods, however, the position of the channel was in some dispute. In an attempt to reconcile ancient sources about the Louros River,including an account by Pseudo-Scylax, Dakaris proposed that the river flowed to the east of Mt. Mavrovouni for the Classical through Roman periods.41 This placement may be appropriatefor the period around 1500 B.P. and later but not for the Classical, Hellenistic, and Roman periods. According to our paleogeographic reconstruction, the shoreline was well north of

I90o

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

4500 B.P. 2 1 b

3 km

Figure 5.19. Paleogeographic reconstructions of Kastro Rogon and vicinity showing the changing coastlines and environments from 7000/6500 B.P. through 1000/500 B.P.: a) 7000/6500 B.P.; b) 4500 B.P.; c) 1500 B.P.; d) 1000/500 B.P.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

191

I92

ZHICHUN

JING AND GEORGE (RIP)

RAPP

Mavrovouni during these periods and Pseudo-Scylax's measurement of the distance between the mouths of the Louros and Arachthos Rivers (40 stadia or 8 km) was likely correct.After the end of maximum marine transgression around 1500 B.P.,the deltaic floodplain began to develop toward the south and southwest as more and more sediments entered the estuary (Fig. 5.19:c).The Louros River flowed in a relativelystable channel at this time. Based on the trend of the contour lines in the deltaic floodplain, the river likely flowed south or southwest directly into the lagoon or the Ambracian Gulf during the early phase of estuarine infilling. The river was not diverted into the current channel until sometime between the 10th and 15th centuries A.C. (Fig. 5.19:d). This channel diversionwas cultural rather than natural.

COASTAL LANDSCAPE CHANGE AMBRACIAN EMBAYMENT

OF THE

Major environmentalchanges have occurredin the Ambracianembayment. On the basis of subsurfacestratigraphyand its implied sedimentary environments in archaeologicallyand geologically important locations, we can reconstruct the changing coastal landscape of the Ambracian embayment during the Holocene epoch (10,000 B.P.to present).
RELATIVE SEA LEVEL AND LOCAL TECTONISM

Change in relativesea level during the Holocene and the precedingWiirm glaciation was the single most important element in shaping the morphology of the coastal landscape. Many studies have shown that there was a rapidrise in eustatic sea level from the end of the Wiirm glaciation (15,00020,000 B.P.) to 6000-7000
B.p.42 However, the change in eustatic sea level

over the past 6,000-7,000 yearshas remained in dispute. It has been shown that relative sea level is more useful and appropriatethan eustatic sea level for paleogeographicreconstructionwith archaeologicalimplications.43 The in relative sea level is controlled mostly by eustatic level, tectonic change movement, sedimentation, and compaction of the preexisting sediment column. Local tectonic subsidence or uplift has been widely considered more critical than eustatic effects to the development of the coastal landscape in Greece over the past 6,000-7,000 years.44During the evolution of the Ambracian coast, both tectonic uplift and subsidence have played a significant role in shaping the configuration of the embayment. The Preveza peninsula has been subjected to continuous tectonic uplift, as clearly indicated by the subsurfacestratigraphicsequence.Thus the small embayments projecting into the Preveza peninsula, such as Ormos Vathy, have witnessed shoreline progradationfor 6,000-7,000 years. As a result, much of the previouslydeposited marine or estuarinestratahave been elevated above sea level. The Ambracian embayment itself has a different history of marine transgression and regression due to tectonic subsidence. Here the maxi-

42. E.g., Fairbanks1989. 43. Kraft,Aschenbrenner, Rapp and 1977; Kraft,Rapp,and Aschenbrenner 1980; Kraft,Kayan,and Aschenbrenner 1985; Rapp and Kraft1994. 44. Flemming 1968,1972; Flemming andWebb 1986; Kraft, and Aschenbrenner, Rapp 1977.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I93

-2

o - -4

7 5
-6 [/-'.
--

004------------400300
maximum
transgression

00

1oo

dated peat samples from swamp

deposits north of the Ambracian

7000

6000

5000

4000

3000

2000

1000

Gulf

Calibrated Radiocarbon Age (B.P.) mum marine transgression lasted from 4500 to 1500 B.P., with a subse-

from of sediment infill over the resulting quentregression dominance the rise in relative sea level tectonic Radiocarbon or subsidence. dates on peat from deposits in the northernpartof the Ambrasamples swamp the coastal cian embayment indicate a gentle rise in relative sea level over the past the theAmbracian fact that 7,000 years (Fig. 5.20). Recalling embayment is a tectonic graben that been subsiding since the Plioceneand Early has rise in relative sea level is most likely attributed to conPleistocene, this the itself tMarin t e ransgression tinuous tectonic subsi dence of embaymen in andregression are dictatedby thechange relative sea level. Thus, relative sea level change should be used in the interpretation of subsurface in the Ambracian strat igraphy embayment.such a generalized Obviously relative thatrend be applied to an area to tectonic sea-leveltonic cannot graben subject use Any of a curve uplift, such asthe Nikopolis isthmus. relative sea-level for paleogeographic reconstruction must be made in an appropriatetectonicsediment ary and context. It is important to know that the rise in relative sea level is with refera ence to geodetic datum. A does change in relative sea level not mean the sea in same change in level, whichis eustatic measured reference to the center of the earth. Based on observations of submerged Hamremains, mond states: e T hereareindications in thecoast of of Epirus that the level th sea e it was at least three or four feet lower in th fourth century than is in tod ay...s The lower sea-level antiquity affected, for instance, the entry to the Gulf of Arta, and it may have reduced the area of swamps which are found today near the mouths of the Louros....

I94

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

The fertile plain on the north shore of the Gulf of Arta may have been more extensive in antiquity.45 This is an example of how a misunderstandingof relative sea level can lead to an inappropriate interpretation of paleogeographic change. It is true that relative sea level was more than three or four feet below current sea level due to tectonic movement, but this does not imply that absolute sea level was necessarilylower in the 4th century than today.Furthermore,the extent of transgressionwas not determined solely by the absolute sea level. Instead, as mentioned previously,it was a result of the combination of sea level, tectonics, sediment supply, and compaction of the preexisting sediment column. Contra Hammond, the embayment saw its maximum sea transgressionduring the 4th century,with the sea reaching the foothills of the mountains in the north.

PALEOGEOGRAPHIC

DEVELOPMENT

The Ambracian embayment is a shallow backarcbasin, initially shaped by Oligocene-Miocene compressional folding and faulting followed by Pliocene-Quaternary extensional faulting. At the end of the Wiirm glaciation, ca. 15,000 B.P., sea level was 100-120 m below its present level.46 The shoreline of the Ionian Sea lay about 5 km west of the Preveza peninsula. Isolated from the Ionian Sea, the Ambracian embayment was mostly exposed subaerially;only a small portion might have been under water, forming small isolated lakes, particularlyin the southern part of the basin.47 As a large volume of glacial ice melted, the sea level started rising very rapidlyaround13,000 B.P. By 10,000 B.P., sea level had risento approximately 45 m below current sea level and the Ionian Sea began to intrude into the Ambracian embayment through the narrowchannel at the south end of the Preveza peninsula.48Sea level continued to rise, and the water body in the embayment graduallyexpanded. Previously inland hills, such as Mt. Mavrovouni and Salaora,were left in the embayment as islands. Apparently, eustatic sea level played a dominant role in the development of coastline change and geomorphic configurations from 13,000 to 6500 B.P. After 6500 B.P. or the beginning of the Neolithic period, the rise in eustatic sea level diminished greatly or ceased. As a result, the shoreline migrated at a much slower rate, creating a favorablecondition for the formation of coastal fringe swamp (Fig. 5.21:a). From 6500 B.P. onward,local tectonic movement became the primary element in the further evolution of the embayment. Relative sea level continued to rise because of tectonic subsidence, and the embayment migrated landward as transgression proceeded. By 4500 B.P. or the beginning of the Bronze Age, the embayment had gained maximum marine transgression, and the sea had extended to the northern edge of the embayment leaving no or a very narrowpassage along the foothills of Mts. Rokia and Stavros (Fig. 5.21:b). As tectonic subsidence was still proceeding at a rate greater than sediment infill from the rivers and streams in the north and northwest, relative sea level continued to rise until 1500 B.P., about the end of the Roman period.

45. Hammond 1967, pp. 42-43.

46. Chappell Shackleton and 1986; Nakada Lambeck and 1988;Fairbanks

1989.

47.The analysis 3.5-kHzseismic of reflection that profiles suggests small

water bodies existed in the south of the AmbracianGulf, particularly within the easternpart,duringthe late Wurm glaciation;see Poulos, Lykousis,and Collins 1995;Tziavos 1997. 48. Tziavos 1997, p. 428.

COASTAL

EVOLUTION

OF THE

AMBRACIAN

EMBAYMENT

I95

1967,p. 19.

49. Dakaris 1971, p. 5; Hammond

the During the period of maximummarinetransgression, entire embayment was flooded by the sea. The Louros River flowed directly into the embayment and formed a subaqueous delta near the mouth of the deeply incised valley of the Louros River.The Salaorabarrierstarted developing as relative sea level rose, but the full barrierdid not form until the postmedieval period. In other words, the entire embayment was basically open water.KastroRogon hill, previouslyinland,was an island in the embayment, but it was separatedfrom the mainland by a very narrow stretch of water. Structures could have been built to connect the hill to the mainland. In addition, the hill was very close to the mouth of the Louros River.This environmentally advantageous location gave Kastro Rogon strategic significance during the Classical, Hellenistic, and Roman periods. During Classical and Hellenistic times, Bouchetion, one of four important walled towns of the Elean colonists, was located on an island hill. A seaport could have been associated with the settlement that served other towns-including Batiae (Kastro Rizovouni) and Elatria (Palaiorophoros)-in the mountainous hinterland. Because of its strategic position, Bouchetion remained an important urban site during the Roman period when other Elean settlements were destroyed and abandoned. Owing to a different tectonic context, the areasurroundingNikopolis has witnessed marine regressioninstead of marine transgressionsince 6000 B.P. The Preveza peninsula has been subjected to tectonic uplift since the Pleistocene. From 13,000 to 6000 B.P. the rapid rise of eustatic sea level was greater than the tectonic uplift of the peninsula. As a result, the tectonic lowlands projecting into the uplifting peninsula were graduallysubmerged as the sea level rose. By 7000-6500 B.P. these lowland areas, including the Nikopolis isthmus and Ormos Vathy,had witnessed maximum transgression.The west arm of Ormos Vathy extended 750 m inland of the currentshore and the Nikopolis isthmus may have been an open channel between the Ambracian embayment and the Ionian Sea. After 6000 B.P., eustatic sea level rise ceased or greatly slowed, and tectonic uplift became the primary factor controlling shoreline change. With continued uplift, the shorelines in the small embayments migrated seaward. Continuous tectonic uplift also led to increased slope erosion. Deposition of slope-wash sediments affiliated with increased slope erosion accelerated marine regression. Sometime after 6000 B.P. the Nikopolis isthmus had been elevated to a level so that no possibility of a channel remained. By 3000 B.P. the shoreline of the Mazoma embayment had migrated seaward to within ca. 1 km of the current shore. During the Roman period, however, both the Mazoma embayment and Ormos Vathy were still wellsheltered harbors serving the city of Nikopolis and other towns on the Preveza peninsula. From late antiquity onward, beginning ca. 1500 B.P., the rate of sediment supply from the riversexceeded the rate of relative sea-level rise and the estuarine embayment began to fill in, moving the shoreline seaward. The increased rate of sediment supply is likely related to human-induced erosion since the Roman period.49The Louros River continued to enter directly into the estuarine embayment with a delta developing at its front.

I96

Figure 5.21. Paleogeographic reconstructions of the Ambracian embayment showing the shoreline changes from 7000/6500 a) 7000/6500 B.P.; b) 4500 B.P.; c) 1500 B.P.; d) 1000/500 B.P.

I98

ZHICHUN

JING

AND

GEORGE

(RIP)

RAPP

the As the delta advanced, alluvialplain aggradedand the riverflowed the acrossthe plain(Fig. 5.21:c).In additionto the LourosRiver, streams in flowingout fromthe mountains the northand northeastalso contribhad for utedsediments estuarine By infilling. 1000 B.P. the shoreline moved Becausethe sedimentsupplyfromthe to the vicinityof Mt. Mavrovouni. muchof the area was rivers not enoughto developan extensive floodplain, was northof the embayment left as swamp(Fig. 5.21:d). Sometimebetween 1000 and 500 B.P., duringthe medievalperiod, the Louros Riverwas divertednear KastroRogon and startedflowing west along the foothills and then south along the west flank of the had infill and sea regression left the (Fig. 5.21:d).Estuarine embayment by seaportfar inland,separated a wide swampyzone from the lagoon or embaymentto the south.The channeldiversioncould have servedtwo a use, (1) purposes: drainingswampsfor agricultural and (2) establishing Gulf. connection betweenKastro RogonandtheAmbracian transportation was The areaalongthe northernandwesternflanksof the embayment a by logical route to dredgea channelas it was coveredprimarily alluvial Kastro in parta resultof this diversion, sediments. doubtatleast No Rogon remainedan importanttown during the medievaland post-medieval periods.

CONCLUSION
and reconstruction provided have Subsurface stratigraphy paleogeographic sites contextof archaeological of the a pictureof the changinglandscape Around10,000 B.P., the sea coastalzone of the Ambracian embayment. sea level had risento about45 m below current level and the Ionian Sea After6000 B.P., into Ambracian hadintruded the graben-like embayment. the rateof eustaticsea-levelrisegreatlyslowedor ceased,but relativesea had level continuedto rise.By 4500 B.P. maximummarinetransgression occurred and the shorelinestood more than 12 km north of its current was position.The entireembayment floodedby the sea.This geomorphic did until the end of the Roman configuration not change significantly erosionincreased sedimentsupplyforestuaperiodwhenhuman-induced was rineinfilling.By 1500 B.P. much of the fringeareain the embayment The changinggeomorphicconfiguration exposedbut remainedswampy. of of the Ambracian was embayment criticalto humanexploitation this
region.

CHAPTER

THE

LOWER

ACHERON

RIVER

VALLEY:
AND THE

ANCIENT
CHANGING

ACCOUNTS
LANDSCAPE

byMarkR. Besonen,George (Rip)Rapp,and ZhichunJing

INTRODUCTION
Recognizing that the earth's coastal systems have undergone profound change since the end of the Pleistocene (about 10,000 years ago), the Nikopolis Project set as one of its objectives the interpretation and understanding of the changing geomorphology, topography,and paleoenvironments in the lower Acheron Rivervalleyfrom the middle Holocene through the present (Fig. 6.1).1 Archaeological remains in the valley are abundant, and literary and historical references go back at least to the 8th century B.C.,when Homer and his contemporaries considered the Acheron to be an infernal river and held that the valley was an entrance to the Under-

world(Od.10.508-515).
Various other ancient literary and historical sources also make reference to the valley, and provide details of a landscape configuration that is inconsistent with the current physiography.The inconsistencies pose a problem for archaeologists trying to equate ruins in the valley with particular settlements mentioned in ancient accounts. Are these ancient authors mistaken in their descriptions of the valley,or can a naturalsequence of landscape evolution account for these discrepancies?There are three conspicuous inconsistencies whose explanation and resolution have provided a focus for this component of the Nikopolis Project: 1) the size of the Glykys Limen (modern Phanari Bay); 2) the nature, geometry, and evolution of the Acherousian lake; and 3) the course of the Acheron River with respect to Kastri during the classical period.
1. This chapteris summarizedand updatedfrom Besonen 1997, a Masters thesis completedby the senior authorat the Universityof Minnesota, Duluth. An electronicversion of Besonen 1997 in Adobe AcrobatPDF formatis freely availableover the Internet at http:// or www.paleoenvironment.org, by requestinga copy from the authorvia e-mail (besonen@geo.umass.edu).
THE BAY) SIZE OF THE GLYKYS LIMEN PHANARI

(MODERN

The smallmarine harbor located the mouthof the Acheron at River is known as of today Phanari (Fig.6.2).Wellprotected a series high Bay by limestone and flushed bythehighdischarge the out of cliffs, continuously Acheron River itstributaries, bayhascharacteristics make and the that for anideal marine harbor. it is Unfortunately, verysmall, measuring 700 only x 350m,witha depth lessthan10 m.In ancient of the times, embayment

200

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

0
I.....

30

60

90km
I

was known as the Glykys Limen ("SweetHarbor").According to the Greek geographer and historian Strabo (7.7.5 [C 324]), who lived through A.D. 21, this was because the influx of fresh water from the Acheron and its tributariescaused a dilution of the marine water filling the bay. Strabo'saccount is not singular;many other ancient authors also mention the Glykys Limen, indicating that it was a well-known feature along the Epirote coastline. Three of these authors provide evidence for a discrepancy between the ancient and modern landscape: while the modern harbor is quite small, the ancient harbor was apparently quite large. The late-5th-century B.C. Greek historian Thucydides (1.46.1-5) wrote in his

Figure6.1. Areamapof Epirus

LOWER

ACHERON

RIVER

VALLEY

201

. .

Area Map of the Lower Acheron Valley

Kicisoura (I __1 2 I
; X

3 kim ....... ,

i .' ;

Mot.zakaiika

X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
9 '''--..l ... , Stav, rochorion " ohl. of i. NEK()OMNTI[ON' i 'I NDOSIA Kastri 1 .. ,,% :. , Nii: ,Xirolophos ' \ ^,;Acherousia . ,Narkissos I &:

/ ! "/

,'

IF,,^igure 6.3 Observation Point J l l, ; !.,' /....-----. wEP ,v \ lThen.elon

} .

. c...

EPHYRA 'N x \ \^ FI \'. Y R A! >_ 'c. . .? '~ CICHYRUSV X ,.,v -" ^.." ..........%Xylokastro r'--/

GI/.}ALSIIiJJ A*.\ '>r.A " nl l//.t. \W'w

Xi^^A1nioudia

D'N

iVel'

I)rontos

; -\^?^^
,,

Acheron
,E, v
v.<,,
Vilanidorraehii ..

\ Skalaiatos

Kanallakion

/i v

, Pountas

lonian
Sea
<
Cd.s'Ksc'i.s.t

X
.. . ; ..

(I iTsotknida -I....1' . . .

,,

i/

l,-/
V

,Bie

... ..
A BC iancienlt
nalmli

abc
_ ^ ^ ^ ^_ _ .. . . . .

illoce' Id
. . . . .

almCe
.... ..... ................. ........... ............... .. ..

Figure6.2. Areamapof the lower Acheronvalley

2. Hammond 1967, p. 69.

3. Dakaris 1971,p. 5. 4. Dakaris 1971,p. 5.

history of the Peloponnesian War that the Corinthians and their allies anchored 150 of their ships in the Glykys Limen before the Battle of Sybota in 433 B.C. Dio Cassius (50.12.2), another Greek historian and Roman official of the 2nd and 3rd centuries A.C., reported that in the summer of 31 B.C., Octavian moored 250 of his ships in the harbor a few days before his confrontation with Mark Anthony and Cleopatra in the Battle of Actium. Finally, Anna Komnena (Alexiad 4.3) recorded in the 12th century A.C. that in A.D. 1081/1082, nearly 1,100 years after the Battle of Actium, the Norman Robert Guiscard and his large fleet wintered over at the Acheron delta. Modern Phanari Bay could not possibly accommodate such large naval fleets. In his account of his travels through the region, the British historian Nicholas Hammond briefly suggested that the bay had silted up since ancient times.2 Sotirios Dakaris, an archaeologist who did extensive work in the area, addressed the topic more thoroughly. Motivated by the accounts ofThucydides, Dio Cassius, and Anna Komnena, he supplied two further lines of geologic evidence that definitively indicate the harbor was once much larger.Dakaris noted the existence of a strip of ancient beach sand, similar to the white sand beach that surroundsPhanari Bay today, ca. 1.5 km east (inland) of the village of Ammoudia (Fig. 6.3).3This strip of sand, in conjunction with "aboring near the confluence of the Cocytus and the Acheron [that] brought to light a layer of sand with sea shells at a depth of 17.5 m from the present surface,"4 provides unequivocal geologic evidence

202

M.

R.

BESONEN,

G.

RAPP,

AND

Z.

JING

Mesopotainon

Tsouknida

Valanidorrachi

:; .. ..

ia A2mrnoid

I'lanari Bav

,. ..

that the Glykys Limen formerly extended furtherinland at some unknown point time. point intgme. in Dakaris'sobservations are significant, but they lack chronological control and thus cannot be used to verify the accuracyof the ancient literary and historical accounts. They provide only a snapshot of the landscape configuration at an unknown moment in time, and do not afford the archaeologist an understanding of the changing landscape. Therefore, our first objective was to develop a detailed picture and absolute chronology for the evolution of the Glykys Limen.
THE AND EVOLUTION OF THE

Figure6.3. View of concentric beachridgessurroundaccretionary PhanariBay,lookingsouth.This photograph takenfromthe was bedrockhighlandson the northside of the valley;see Figure6.2 for the locationof the observation point. The AcheronRiveris delineatedby the faint darkbandof treesvisiblein the background. M. Photo Besonen

NATURE, ACHEROUSIAN

GEOMETRY, LAKE

A second significant discrepancybetween ancient references to the valley and the observable modern landscape concerns the nature, geometry, and evolution of the extinct Acherousian lake (Fig. 6.4). The existence of the lake is not in question, for its final swampy remnants persisted until just after the First World War, at which time they were drained and backfilled for agriculture.5During Greek and Roman times, the lake was apparently a conspicuous feature given that many authors make reference to it (Thuc. 1.46.3-4; Pseudo-Scylax 30; Strab. 7.7.5 [C 324]; Plin., HN4.1.4; Livy 8.24; Paus. 1.17.5). By medieval times, it was referredto as the Acherousian swamp, apparently reflecting a natural infiUing.6Though the number of referencesto the lake-swamp is significant, few provide any detailed topographic information that is useful in determining its location and nature. Several modern authors have considered the existence of the lake in the valley.William Leake, who traveled through the region in 1809, left a fairly detailed description of the marshy valley bottom with its few, shallow, isolated pools.7 He concluded that the marsh-lake present below the hill of Kastri was the Acherousian swamp known from antiquity, seemingly not considering the possibility that it might previously have had a different nature or proportions (Fig. 6.4, upper left). Alfred Philippson and Ernst Kirsten presented a different scenario in their survey of the Greek landscape, suggesting that the swampy, marshy ground which rep-

5. Hammond 1967, p. 68. 6. Hammond 1967; Dakaris 1971. 7. Leake 1835, I, p. 232; IV, pp. 5154.

203

o,

0
CD 0
.i

-4
o

204

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

but since anresentedthe lake had expandedareally, become shallower, One of theirmapsshowsa dottedoutlineof whatis presumcienttimes.8 lake ablythe Acherousian (theAcheronRiverentersone sideandexitsthe other).This lake stretchesnorthfrom Kastriup the valleyalmostto the that pointwherethe AcheronRiverexitsfromthe bedrock uplands bound the valleyto the east (Fig. 6.4, upperright). By the time Hammondpassedthroughthe valleyin the middleof the the of 20th century, final remnants the lake had been filled in. He indicated more definitiveboundaries the Acherousian for lake basedon anand cientliterary historical the of references, descriptions Leake,andsome The boundaries indicatedarethe Mesopoearlierworkby Dakaris.9 he tamon/Tsouknida to valleyconstriction the west, the bedrockhighlands to the south,andthe Pountasridgeand Kastrito the east (Fig. 6.4, lower left). Dakarispresented most careful the consideration the subject, of basing his theory on ancient literaryand historicalreferencesas well as his own observations. reconstruction the lake'ssize andlocationis His of similar thatgivenby Hammond,but he extendedthe eastern to boundary of the lakepastthe Pountas towardKanallakion ridgeandKastri (Fig.6.4, lowerright).10 basisfor this eastward The extensionwas the chancefind of ten wooden beams duringthe excavation a drainagecanaleast of of Pountasridge and southwestof Kanallakion. Dakarisinterpreted these beamsas partof the keelof an ancientboatthathadoncepliedthe lake;he also noted that a spot on the easternside of Pountasridgeis still referred to as"DromosSkalamatos," whichmeans"port" "place embarkation" or of (Fig. 6.2).11 Dakaris, Hammond,andothersbasedtheirreconstructions primarily on indirect but influenced theirobservations evidence, werealsogreatly by of the modernlandscape the valley. in Their reconstructions overestimate the size of the lake at least as an open body of water,and lack definitive control. aboutthe lakeduring chronological They provideno information times, an item of interestto the NikopolisProject.A compre-classical of and based pleteanddetailed chronology the lake's development evolution on geologicevidencehas neverbeen prepared. isParticularly important suesto resolveincludewhen the lakecameinto existence, mechanism the the by which this occurred, natureof the lake,and its geometryand dimensionsthroughtime.These questionsframedoursecondobjective.
THE COURSE OF THE DURING ACHERON THE RIVER WITH RESPECT

TO KASTRI

CLASSICAL

PERIOD

The courseof the AcheronRiver,like that of most riversin their lower is was stretches, constantly shifting.Our thirdobjective to determinethe locationof the courseof the Acheron Riverwith respectto the hillock
Kastri during the first millennium B.C.(Fig. 6.2). This is particularlyim-

of the portantto helpresolvethe long-standing problem identifying ruins

on that hillock with those of Pandosia, a fortified urban settlement often referencedin ancient literaryand historical sources (Dem. 7.32;Justin 12.2; Livy 8.24; Plin. HN 4.1.4; Strab. 7.7.5 [C 324]). The major discrepancy

and 1956,II, 8. Philippson Kirsten 9 Hammond 1967,p. 69 10. Dakaris 1971, pp. 4-5 andfig. 7. 11. Dakaris 1971,p. 57.

LOWER

ACHERON

RIVER

VALLEY

205

frustrating this identification has been that Kastri is located to the north of the Acheron River,but ancient sources indicate Pandosia was located to its south (Dem. 7.32; Strab. 7.7.5 [C 324]); hence the difficulty in equating the two. In ancient times, the Acheron River served as a political boundary dividing the territory of Thesprotia to the north from the territory of Cassopaia to the south (Fig. 6.2). Ancient sources indicate that, in addition to Pandosia, a fortified urban settlement named Ephyra also existed in the valley. Geographic references situate Ephyra north of the Acheron River in the territory of Thesprotia, close to the sea, and near the Acherousianlake (Paus. 1.17.4-5; Strab.7.7.5 [C 324];Thuc. 1.46.4). Pandosia, in turn, was located further inland, and within the territory of Cassopaia (Dem. 7.32; Strab. 7.7.5 [C 324]). Besides the ruins at Kastri, the remains of a second fortified urban settlement can be found in the valley today.This second site is located just north of Mesopotamon on the ridge known as Xylokastro (Fig. 6.2). Since the Xylokastro site is closer to the sea, and the Kastri site is further inland, one might immediately suggest the first site should be identified with Ephyra, and the second site with Pandosia. Combined with simple differences of opinion, however, the issue of the rivercourse has prevented consensus about the identification of the ruins in the valley. For example, Hammond did not consider the Kastri/Pandosia identification as appropriate, suggesting instead that the ruins at Gourana, much furtherupvalley, were actually those of Pandosia.12 Dakaris did prefer the Kastri/Pandosia identification, and he reconciled the issue of the rivercourse by suggesting that the river had shifted since ancient times. Wherever the river banks are not supported, or when the river overflows, it could result in a change in course.... The slight inclination of the Acheron plain, the swamps, and the lake, formed by the river to the south of Kastri hill, contributed to the change in the river bed, which, in ancient times, had the hill with the ruins to its south, at [sic] Cassopaia.13 While Dakaris's suggestions concerning the dynamic nature of the river course are correct, he did not provide any geologic evidence to show that the river had indeed shifted its course from the northern to the southern side of Kastri since the first millennium B.C.Hence, our third objective was to examine the changing course of the Acheron River with respect to Kastri during the past 2,000 years, and either confirm or deny the shift proposed by Dakaris.

GEOLOGY ACHERON
12. Hammond 1967, p. 478. 13. Dakaris 1971, pp. 136-137.

AND THE NEOTECTONICS VALLEY AREA

OF THE

14.Aubouin 1959and1965. 15.Etude geologique.

Jean Aubouin presented two detailed studies interpreting the stratigraphy and tectonics of Epirus which have served as the foundation for subseThese studies were followed by Etude g6ologique,15 another quent work.14 major monograph on the geology of Epirus, which resulted from petro-

206

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

l--IC

- I

leum exploration work. The major features of the landscape of Epirus are structural in origin. The region consists of a series of north-northwest/ south-southeasttrendingfolds and thrustfaultblocks that have been formed in a sequence of compressional orogenic events since the Late Jurassic period.16 These folds and fault blocks form a series of parallel limestone mountain ranges with intervening flysch basins that can be delineated in a satellite image of the region (Fig. 6.5). Some of the ranges reach over 2,000 m in elevation, but on average range from 1,200 to 1,700 m.17 The marked and varied relief noted between the ranges and basins is a direct function of the structureand contrasting lithologic properties of the limestone and flysch.'8 Relief is even more spectacularalong the coasts, where bedrock cliffs rise directly from the sea, or very flat coastal riverplains give way in abrupttopographic discontinuity to carbonatebedrock valley walls. A map of the simplified geology of the lower Acheron valley is shown in Figure 6.6. Recent alluviumfloors the flat valleybottom, which is flanked by the steep, carbonate bedrock valley walls. These valley walls are composed, for the most part, of Mesozoic and some Eocene limestones. The limestones are cherty,range from fine-grained to sublithographic,are usually fossiliferous (with the remains of calcareous algae, radiolarians,rudist clams, ammonite cephalopods, and globigerinid and other foraminifera), and in places are dolomitized and/or brecciated. Upper Eocene to Lower Miocene (Aquitanien) flysch crops out at the base of the eastern valley wall. The flysch is composed primarilyof alternating soft micaceous sandstones and shales with intercalated,thinly bedded biogenic limestones and

Figure6.5. Satelliteimageof Epirus. North of the Acheronvalley,the structural of configuration the region is delineatedby a seriesof parallel limestonemountainranges(trending north-northwest south-southeast) to with interveningflyschbasins.Note the elongategeometryof the ThyamisandArachthosriverdeltas with fringingdeltatop/front "barrier" beaches.

16. Etude g6ologique.

17. King, Sturdy,and Whitney 1993.

18. Etude g6ologique.

LOWER

ACHERON

RIVER

VALLEY

207

Qal

all Quaternary uviuLtl

'X

Nt-4

thrustfault--teeth on ovcrthrustblock

SC

%% scree slopes and talus cones

nornmal faull--U (Up) and D (down) show relative motion of blocks

(:i;iii ii; lignilic ii!iii:iii

Plioccnc ArkhangclosFormation-m-ixcd marine and continentalconglomerates,muddy sands, and

and marine shalcs

Lower Miocene (Aquitanien)flysch--soft, micaccous sandstonesand shales with intercalatcd thinly-beddcd near biogenic limestones and zmails top Mesozoic and 1locene carbonates--cherty. fossilifierous, fine-grainedto sublithographic limestones;occasionally dolomitic or brccciatcd

microcrystallincgypsum--age uncertain

Figure 6.6. Simplified geology of the lower Acheron valley

208

M.

R.

BESONEN,

G.

RAPP,

AND

Z.

JING

marls near the top. The top of this flysch unit effectively marks a large shallow thrust fault over which the more competent Mesozoic limestone has ridden to create one of the limestone ranges seen in Figure 6.5. Recent talus and scree slopes cover the contact and most of the flysch unit. A small strip of the Pliocene Arkhangelos Formation crops out in the southern valley wall to the east of Pountas ridge. This formation is a mixed marine and continental unit that consists of conglomerates, muddy sands, and lignitic and marine shales. Finally, an inferred active, east/west-trending normal fault exists along the south valley wall,19though this fault was not recognized by earlierwork.20 While Etude geologiqueis comprehensive through the Pliocene, the tectonic history of the Pleistocene and more recent periods was not included. Fortunately,a recent dissertation by David Waters fills the gap.21 Waters providesan inventoryof geologic evidence (e.g., incised rivergorges, wave-cut notches, and raised shell burrows)that suggests mainland Epirus and much of the coast has been undergoing uplift since the Pliocene. At the same time, the evidence suggests that certain areas,such as the Ambracian Gulf, lower Acheron valley,and lower Thyamis valley (Figs. 6.1, 6.5), are subsiding; very thick deposits of Quaternary sediments are found at these locations. Subsidence also seems to be occurring along the northwest coast of the mainland opposite Corfu, a process indicated by the steep, rocky shorelines (with numerous small coves and islets) and the lack of beach platforms.22 Waters attributesmodern subsidence of the lower Acheron valley bottom to movement on the inferred active normal fault along the southern valley wall (hanging wall to the north) (Fig. 6.6).23The existence of this fault would make the valley configuration that of a half-graben, though Waters never explicitly describes it as such. While the alluvialvalley bottom appears to be subsiding, there is some evidence to indicate uplift of the carbonatevalley walls;Waters identified a wave-cut platform, 1.7 masl, on the north side of Phanari Bay.

HOLOCENE RELATIVE REGION OF EPIRUS

SEA LEVEL IN THE

Coastal evolution is intimately tied to relative sea level, itself determined by eustatic sea-level changes, isostasy,and vertical tectonic movements. A record of relative sea-level change for a particularregion must be compiled from local evidence. A relativesea-level curve for the southwestern Epirote coast is shown in Figure 3.24. Particularlyimportant to note is that, during the last 5,000 years,relativesea-level rise along the southwesternEpirote coast has been less than 2 m. The rate of sedimentation at river mouths, however, has been much greater.A significant implication of this relationship is that the physical sedimentology and microfossil assemblages contained in the stratigraphyare more important indicators for reconstructing shoreline position than the local sea-level curve.

19. Waters 1994, figs. 5.7, 5.10. 20. Cf. Etudegeologique. 21. Waters 1994. 22. Waters 1994, p. 197. 23. Waters 1994, figs. 5.7, 5.10.

LOWER

ACHERON

RIVER

VALLEY

209

FIELD AND LABORATORY METHODS

Twenty-eight sediment cores were retrieved from various points in the lower Acheron valley during summer field seasons from 1992 to 1994 (Fig. 6.7; Appendix). All cores were retrieved by a hand-operated, 3 cm diameter Eijkelkampgouge augerwith the exception of cores 94-02 and 94-03, which were taken with a 7 cm diameter Edelman auger bit. Cores were described and logged on site using terminology following Folk.24Fieldobservable parameters that were recorded include lithology and approximate grain-size distribution, color when wet,25sediment consistency,plant and animal macrofossils, pedogenic characteristics (structure, sesquisoxide/reduction mottling, and calcium carbonate filaments or nodules), and chance finds such as pottery fragments. Sediment samples were collected for laboratory analysis in the U.S. with approximately 300 taken during the 1994 field season, and a much smaller number during the 1992 and 1993 seasons. Laboratory analyses were focused mostly on sediment samples from the 1994 season. All 1994 samples were analyzed for dual-frequencymagnetic susceptibility and anhysteretic magnetization along their length using facilities at the Limnologic Research Center and Institute for Rock Magnetism at the University of Minnesota,Twin Cities. Microfossil assemblages were determined in fifty samples, from ten different cores, all but one of which was collected during the 1994 season. In most cases, the total microfossil population, including ostracods,foraminifera,gastropods, pelecypods, and charophyte oogonia, was picked and identified. Relative percentages of fresh and brackishwater microfossils were calculated from species counts to provide an approximation of the salinity of the environment of deposition. Eight cores from the 1994 season were analyzed along their length for organic carbon and carbonateusing the method of Dean.26 Grain-size distribution by pipette analysis was determined for twentythree samples using the method of Folk27to supplement the field-based approximationof grain size. Eight samples of organic materialwere radiocarbon dated by the acceleratormass spectrometer (AMS) method at either the Radiocarbon Laboratory at the University of California, Riverside, or Beta Analytic LaboratoriesInc. of Miami, Florida (Table 6.1). Results from field observations and laboratory analyses were examined together to determine the probable environments of deposition for each lithostratigraphic unit. A summary of these data for each sediment core can be found in the appendix to this chapter,with the full complement of primary data availablein Besonen 1997. A study of early maps of the area, as well as literary and historical references by both ancient and more recent authors (in particular, Homer, Thucydides, Strabo, Anna Komnena, and Leake), was undertaken to supplement and provide a context for the geological data. Finally, three cross sections through the valley were drawn (see Figs. 6.9-6.11), and reconstructions showing the evolution of the landscape in the valley at eight points in time during the past 5,000 years were constructed (see Figs. 6.12-6.15).

24. Folk 1980. 25. Colors were describedusing the Munsell Soil Color Chart (rev.ed. 1994). 26. Dean 1974. 27. Folk 1980.

210

M.

R.

BESONEN,

G.

RAPP,

AND

Z.

JING

Figure 6.7. Core locations in the lower Acheron valley

TABLE 6.1. RADIOCARBON

LabNo.a UCR-3217 UCR-2695 UCR-2696 UCR-2697 Beta-80531 Beta-80532 Beta-80533 Beta-80534 Core NC92-20 NC93-18 NC93-19 NC93-21 NC94-04 NC94-13 NC94-20 NC94-23 Depth below Surface(m) 5.30-5.40 0.70-0.75 7.00-7.20 5.80-6.10 2.95-3.00 5.35-5.40 6.05-6.20 10.35-10.55

DATES FROM THE ACHERON

13C/2C

RIVER VALLEY
Conventional 14C Age (B.) 2470 ?60 2890 +40 4520 +?60 3460 ? 60 340 ?50 950 ? 50 1740+60 3700 +?60 CalibratedAge
(s.e.J)b

Material charcoal,root fragments peat with organicmaterial peat with wood fragments peat with organicmaterial wood plant material wood plant material

Ratio n/a -27.18%o -20.36%oo -28.24%o -26.00oo -23.3%o -39.9%oo -27.0%o

2650 +70/-290 2980 +90/-30 5140 +160/-100 3690 +140/-60 380 +90/-70 850 +80/-60 1670 +40/-120 4030 ?100

bCalibration

= aDating laboratory: UCR = University of California, Riverside; Beta Beta Analytic Laboratories Inc. of Miami, Florida. from conventional 14C age to calendar years was performed using the CALIB Rev. 3.0.3c computer program available from M. Stuiver and P. Reimer of the Quaternary Research Center at the University of Washington, Seattle. All options were set at their default values. The data set used to make the calibrations was the INT93CAL bidecadal

but calibrationcurve.A decadalcalibrationis also available, is meant for use with high-precisiondates dendrochronologic (a<40years).

LOWER

ACHERON

RIVER

VALLEY

211

mapof the Figure6.8. Topographic lowerAcheronvalleybottom. Contourlines maynot be continuous at valleyedgeswherethey are Sedimentcorelocations compressed. aremarked blackdots;see Figure by 6.7 for labels.Heavyblacklines show the locationsof crosssections in illustrated Figures6.9, 6.10, and 6.11.

Topographic control for all elevations cited in this study was provided detailed 1:5,000 topographic maps produced by IGME (the Greek Inby stitute of Geology and Mineral Exploration) in 1981. These maps are contoured at 1-m intervals over the flat valley bottom, but shift to 4-m intervals for the steep bedrock valley walls. The maps also record hundreds of individually surveyedpoint elevations throughout the valley bottom where topography is slight. Figure 6.8 presents a very reduced set of this topographic data, contoured at 2-m intervals to simplify presentation.

MICROFOSSIL ECOLOGY

ASSEMBLAGES

AND RELATED

28. Neale 1964; Phleger 1960.

Ostracoda and foraminifera have been used with great success as indicators of paleoenvironments in marginal marine systems because they are We extremely sensitive to salinity and temperature,among other factors.28 examined the microfossil assemblages in fifty sediment samples, paying particularattention to ostracods and foraminifera for paleosalinity information. Identification of the ostracods was achieved down to the species level for twenty-four forms, down to the genus level for one form, and left undetermined for one form. Identification of the foraminifera was less rigorous: down to the species level for three forms, to the genus level for four forms, and to the family level for one larger,well-known group.

212

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

Reference works based generally on Mediterranean localities or similar marginal marine settings were used to identify the microfossils and to gather information about their ecologic and environmental preferences.29 This allowed us to define two microfossil assemblages indicative of paleoenvironments with differing salinities. The first microfossil assemblage is characteristicof shallow, freshwater environments, while the second assemblage is indicative of shallow, nearshore, brackish to marine waterconditions. Microfossil identifications and paleoecological interpretations based on the assemblages were later confirmed and corrected by micropaleontologistFrederickSwain of the University of Minnesota, Twin Cities.30 Whereas microfossilswere sparseor absent in some freshwatersamples, all shallow,nearshorebrackishto marinesediments showed high total abundances. Certain forms were present in nearly all samples from brackish to marine deposits, and in some cases occurred in extreme abundance. The shallow,freshwaterassemblage is comprised of fourteen ostracod Candona species: Candonaalbicans,Candonacf. caudata,Candonacompressa, cf. lactea, Candona neglecta, Candona truncata, Cyclocypriscf. laevis, Darwinula stevensoni, cf. Herpetocypris reptans, Ilyocypris gibba,Limnocythere cf. inopinata,Limnocythere Potamocypris villosa, and Ostracod sp. A cf. sp., (possibly Prionocypris zenkeri). Twelve ostracod species and a series of foraminifera characterize the elonshallow,nearshore,brackishto marinewater assemblage:Cushmanidea cf. torosa,Cytheridea gata, Cyprideis neapolitana,Cytheridea sorbyana,Cythcf. Loxoconcha bacescoi, Leptocythere castanea, eromorphafuscata, Leptocythere Loxoconchacf. granulata, Loxoconchaovulata, Paracytheroiscf. elliptica, amnicola.The foraminifera in the assemacuminata,and Tyrrhenocythere blage include Ammonia beccarii,Bolivina sp., Bulimina sp., Cribrononion Fursenkoina and members of the family translucens, sp., Elphidiumcrispum, Miliolidae (including several Quinqueloculina spp. and Triloculinasp.). While mixing of fresh and brackish to marine assemblages may be it significant, especially in regions with large tidal fluxes,31 was minimal in the sediment samples from the Acheron valley. This is not unexpected given the small tidal variation (20 cm) in the region. Cathleen Villas noted some mixing of marine microfossils in the freshwaterenvironments of the Acheloos delta, just 150 km to the south of the Acheron valley,32 the but Acheloos delta plain is totally unprotected and experiences the unbuffered assault of storm waves. This is not an issue in the Acheron valley where Phanari Bay is enclosed and well sheltered by the large carbonate cliffs.

SEDIMENTOLOGY DEPOSITION

AND ENVIRONMENTS

OF

The modern sedimentary environments in the lower Acheron River valley are very similar to those found at other spots along the Greek coast.33For simplicity, we divide them into two broad depositional systems. The first system, herein referredto as the fluvial depositional system, consists of all the sedimentary environments landward of the shoreline. Six distinct environments can be identified: river channel, natural levee, crevasse splay,

29. Ascoli 1964; Bhatia 1968; Devoto 1965; Ellis and Messina 19522000; Puri, Bonaduce,and Gervasio 1969; Puri, Bonaduce,and Malloy 1964; Sars 1928;Tassos 1975;Tziavos 1977; Villas 1983;Wagner 1957;Yang 1982. 30. A complete summaryof this work, including scanningelectron microscopeplates of the microfossils encounteredin the Acheronvalley,is freelyavailableonline;see note 1 of regardingthe availability Besonen 1997. 31. Kilenyi 1969. 32. Villas 1983, p. 54. 33. Tziavos 1977; Villas 1983.

LOWER ACHERON

RIVER VALLEY

2I3

34. Middleton 1973. 35. See Chapter5 for a short discussionof Walther'sLaw.

lake.The seconddeposiand floodplain, backswamp, shallowfreshwater tionalsystemis a deltaicnearshore association, composedof the environof the shoreline withinthe marineembayment and mentslocatedseaward of the GlykysLimen.Eight distinctenvironments be identifiedin this can freshto brackish waterdeltatop marsh, deltadistributary channel, system: channel mouth bar, subaqueouslevee, lower delta front, distributary beach. interdistributary and accretionary prodelta, bay, On the modernlandscape, many of these environments gradeinto it difficultto placeexactboundaries one another between laterally, making them.This difficulty further is when attempting reconstruct to magnified basedon a finite numberof 3-cm diametersediment paleoenvironments cores.However,Walther'sLaw of the correlation facies34 of providesa the This powerfultool to interpret subsurface stratigraphy.35 is especially truein a marginal marine environment the lowerAcheron like when valley, workis grounded physical in and of sedimentology the analysis microfossil assemblages. What followsis a briefsedimentological geomorphic and for environments deposition of mentioned description eachof the fourteen above. beginwith the environments the fluvialdepositional We of system, and then discussthose of the deltaicnearshore association. Riverchanneldepositsarecomposedof the coarsest sedimentsfound in the fluvialdepositional system,and includelag depositsand barsthat formdirectly the riverchannel.Depositsareusually or buffin color, in tan but mayexhibita reducedcolorif trappedin an environment such as an oxbowlake.In the lowestreaches the valley, wherethe riverchanneland of barsystemgradesinto the deltaicenvironment, sandsandgravels mayalso have a graycolor.Ostracodsand other microorganisms not generally do inhabitsuchenvironments, the occasional and of carapaces detritalorigin that do make it to the riverchannelare quicklydestroyedin the highor clasticmaterial. Reenergyenvironment, aredilutedin the abundant workedmicrofossils fromthe localbedrock occurin extreme abundance in since it is the maintransport depositsof this environment agentfor such material. Subaerial natural leveesarewedge-shaped ridgesof sandand muddy sand that are depositeddirectlyadjacent a riveralong its length, and to thin awayfromthe river. These depositsmayformsignificant topographic when coarsesediments carried overbank a floodhighs,andarecreated by ing riveraredroppedout of suspension. They aregenerally finer-grained thanchanneldeposits,becomeincreasingly awayfromthe riverchanfine Becausethese nel, and eventuallygrade into floodplainor backswamp. depositsareexposedsubaerially, tend to be tan to orangeto brownin they and weakpedogenic suchassesquisoxide features mottles color, mayexhibit and nodules,and carbonate filamentsand smallnodules.Occasionalostracodcarapaces detrital of reworked microfossils liboriginandabundant eratedfromthe localbedrockarefoundin these deposits.Naturallevees have playeda significantrole as agentsof geomorphicevolutionin the lower Acheronvalleythroughthe middle and late Holocene, and their will below. significance be discussed Crevasse splaydepositsformduringperiodsof exceptional flooding, when channelsare cut throughthe naturallevee systemallowingwater and bedload sediment to escape onto the adjacentfloodplain or into

214

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

backswamp or interdistributarybay environments. They occur as lobeshaped wedges of sand- to mud-sized sediment that thin away from the river channel. A modern lobe-shaped crevassesplay deposit, delineated by the 2-m contour line, can be seen in the floodplain to the southwest of ancient Ephyra (Fig. 6.8). Although these deposits are similar in composition to naturallevee deposits, they can be distinguished by their geometry and by the fact that they appearas abruptpulses of coarsersediment within the mud and silt of floodplain, backswamps,or interdistributary bays. Such deposits contain abundant reworked microfossils from the local bedrock, and relatively small amounts of organic matter. The floodplain environment consists of the flat, low ground adjacent to a river channel and natural levee system that acts as a settling basin for fine-grained suspended sediment carried over the river'sbank during flooding. Floodplain deposits consist mostly of silt and clay with occasional fine sand laminae. This environment is exposed subaerially, so its sediments tend to be tan to orange to brown in color, exhibit slight to moderate pedogenic development, and tend to be more compact and stiffer than sediments from other environments. Occasional modern ostracod carapaces of detrital origin, common fragments of terrestrialgastropod shells, and abundant reworked microfossils from the local bedrock are found in such deposits. These deposits contain moderate amounts of organic carbon. The backswamp environment represents a transitional step between floodplain and shallow lake environments. It commonly occurs in low, poorly drained areasadjacentto the river channel or valley walls, and consists of nearly perennially saturated swampy and marshy ground. In the Acheron valley, it also occurs in the low swales between the spectacular accretionary beach ridges east and northeast of Phanari Bay (Fig. 6.3). Backswamp deposits are composed of dark gray to brown, organic-rich mud and clays, though sandy intervals may be present depending on the proximity of the river channel. In some cases, vegetation is so abundant that the backswamp is essentially a freshwatermarsh, and deposits consist of peat and peaty mud. Such deposits are composed of up to 25% (by weight) organiccarbon.Members of the freshwaterostracodgenus Candona occur in common to abundant quantities in backswamp deposits, while other freshwater forms occur in lesser quantities. Shallow freshwaterlakes and pools are no longer present in the lower Acheron valley because they have been filled in for agriculture,but they occupied a significant portion of the valley bottom in the past. These lakes are commonly transitional with backswamp and marsh environments. Deposits from such lakes are generally gray in color, and extremely rich in clay-sized particles;they have a moderate organic content, ranging from 3 to 8% by weight. Microfossils present in these deposits include relatively sparse numbers of freshwaterostracods and gastropods. Microfossil abundance increaseswhen the deposit is transitionalwith backswampand marsh deposits, and this is probablythe result of the greaterorganic content (food supply) of shallower environments. The most significant mechanism for the creation of these shallow lakes in the lower Acheron valley involves the impingement of a river channel and levee system against the bedrock valley walls, as described below. Oxbow lakes, which are very common in

LOWER

ACHERON

RIVER

VALLEY

2I5

36. Russell 1954;Villas1983. 37. Villas1983,p. 75.

other coastal river plain localities,36are infrequent in the lower Acheron valley at the present day.The only example that currentlyexists is the horseshoe-shaped loop immediately north of the Acheron River,ca. 1.25 km to the east-southeast of Phanari Bay (Figs. 6.2, 6.8). The fresh to brackishwater delta top marsh is a thick accumulation of reeds and marsh grasses fringing the shoreline on the delta top, such as that which exists at present on top of the Acheron delta to the south and southeast of Phanari Bay.The marsh is situated at approximatelysea level and receives input of water and sediment from the fluvial and marine systems. Deposits consist of peat and peaty mud with occasional sand layers, and are composed of up to 25% (by weight) organic matter.The microfossil assemblages in delta top marsh deposits grade upward from extremely abundant shallow brackishwater forms (especially Cyprideis torosa,Leptocf. castanea, Loxoconcha Ammoniabeccarii, and Cribrononion cythere elliptica, to translucens) abundant freshwater forms. This distribution of microfossils reflects its location at the transition from the marine to freshwater system during an overall regressive sedimentary regime. The delta distributarychannel, distributarymouth bar, and subaqueous levee are active delta front environmentswithin the marine embayment where the majority of deposition and delta progradation occurs. All three environments are essentially subaqueouscontinuations of the subaerialfluvial channel and natural levee environments. The coarsest sediments in the system are generally sands and sandy gravels that floor the delta distributarychannel. Subaqueous levees border the distributarychannel and are composed mostly of sand and silt. As currentsin the distributarychannel lose competence, sediment is dropped out of suspension and forms a broad sandy apronaroundthe distributaryknown as the distributarymouth bar. All recognized active delta front deposits from the lower Acheron valley are gray to darkgray in color; however, Villas reports that both gray and tan components exist in the subaqueouslevee deposits of the Acheloos River.37 Microfossils present in such deposits consist primarily of abundant numbers of brackish to marine water organisms, and abundant reworked microfossils from the local bedrock that were carried to the delta by the fluvial system. Deposits from these environments grade basinward into the laminated clays, muds, and fine sands of the lower delta front and prodelta, and laterally into the interdistributarybay environment. The lower delta front and prodeltaenvironments arelocated basinward of the active delta front, and act as a settling basin for suspended sediment. Deposits from both environments consist of gray to dark gray laminated clays, muds, and fine sands, but the sediments of the lower delta front are noticeably coarsersince they are a distal extension of the active delta front. Deposits from these environments have a low to moderate organic carbon content (3-4% by weight), and their microfossil assemblagesconsist strictly of abundant brackish to marine water organisms without any freshwater forms. The interdistributarybay is a shallow open body of water located to the side or partiallybehind the active delta front. At present, there are no interdistributarybays on top of the Acheron delta because Phanari Bay is almost entirely filled in, but such bays did exist in the past. Deposits from this environment are composed of gray to dark gray silts and clays that

216

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

washedin over the natural settle out of suspension,and sandymaterial the channelsof the deltatop. Creleveessurrounding fluvialdistributary in vassesplaydepositsarealso commonlyfound interbedded depositsof thesebays,conwhichsurround interdistributary Delta top marshes, bays. to carbon content(4-8%byweight). tribute theirmoderate highorganic to abundant brackish also Depositsfromthis environment containextremely microfossils as to marinewatermicrofossil assemblages, well as reworked fromthe localbedrockin commonto abundant quantities. beachridgesandinterA spectacular seriesof concentric accretionary modernPhanariBay (Fig. 6.3). The Acheron vening swalessurrounds delta top and front providea constantsourceof sandysedimentthat is and waveactivity, then gentlypiledup overthe regureworked normal by that larwavebaseby springandwinterstormwaves.Longshorecurrents excesssanddo not exist couldkeepthe systemin equilibrium removing by As these Phanari is so well sheltered. a result, or areveryweakbecause Bay ancientbeach ridgeshave accretedone by one, continuously decreasing theseridgesaregenerally the size of Phanari The sandsthatcomprise Bay. with smallpebbles,andaretan to buff in color. coarse-grained occasional swalesarefloodedseasonally becauseof theirlow elevaThe intervening and tion, and often accumulate backswamp marshydeposits.This beach is transitional with the deltatop and ridgeandswaleenvironment laterally deltafrontenvironments.

MIDDLE AND LATE HOLOCENE GEOMORPHIC EVOLUTION OF THE GLYKYS LIMEN

We havedocumented relative the of evolutionindisequence geomorphic catedby subsurface and this stratigraphy, supplemented with eight radiocarbondateswhich provideabsolutechronological control.Overall,the middleandlate Holocenesedimentary recordin the valleyis regressive in naturereflectingalluviation duringa periodof very slowlyrisingrelative sea level. Dakarissuggestedthat the GlykysLimen was formerlymuch larger, extendingback to nearthe Mesopotamon/Tsouknida valleyconstriction "acertain at The was geological period."38 suggestion basedon his observation a fossilbeachridge1.5 km eastof the villageof Ammoudia of encoun(on PhanariBay),and the presenceof fossil marinemacrofauna teredin a boringnearthe confluence the Acheronand VouvosRivers. of The Mesopotamon/Tsouknida is obstruction valleyconstriction a natural in the valleyboth areally in the subsurface and becauseof shallowbedrock to (Figs.6.9, 6.10), andlogicallymighthaveservedas a natural boundary Holocene seas. Our resultsindicate,however,that marine transgressing influencereached evenfurther inlandthanDakarissuggested; risingHolocene seasstretched leastto the locationof core94-17 (Fig. 6.7), sevat eralhundredmeterseast of the valleyconstriction, around2100 B.C. (see radiocarbon datesprovideabsolutechronological conFig. 6.12). Several trolfor this and othershoreline the past4,000 years. positionsduring The reconstructed shorelines presentshouldbe takenonly as genwe eralized locationsof the shoreline position.Becausewaveandtidalenergy

38. Dakaris 1971, p. 5.

LOWER

ACHERON

RIVER

VALLEY

2I7

Phanari arelow alongthe Epirotecoast,andevenlowerin well-protected the Acherondelta is dominatedby fluvialprocesses.Such fluvially Bay, Unfordominateddeltasdisplayan elongategeometry, spacepermitting. filledin, this elongategeomsincePhanari is almostentirely tunately, Bay in cross at It etryis not apparent present. canbe seen,however, subsurface sectionC-C' (Fig. 6.11), aswell as in otherriverdeltasin Epirus,suchas the those of the Thyamisand ArachthosRivers(Fig. 6.5). Additionally, These ambichannels. Acherondeltamayhavehad multipledistributary the guitiesprecludethe possibilityof reconstructing exactshorelineconat figuration anymomentin time. havea similar straCores94-17 and94-23 (Figs.6.7,6.10;Appendix) of nature the sediments illustrate overall the andclearly tigraphy regressive laid down in the lowerAcheronvalleyduringthe middle and late Holocene. Both cores consist of depositsfrom the followingenvironments water order: deltatop to front,2) brackish 1) givenin normalstratigraphic into freshwater deltatop marshgradingupward marsh,3) shallowfreshA date waterlake,and4) floodplain. radiocarbon on peatfromthe bottom of the brackish waterdeltatop marshof core94-23 returns calibrated a lo
This peat belongs to B.C. B.P., or 2080 +?100 range of ages from 4030 +?100

the extensive subsurface deltatop marshdepositseen in the A-A' andBB' cross sectionsthroughthe Mesopotamon/Tsouknida valleyconstriction (Figs.6.9, 6.10).The radiocarbon fromcore94-23 indicates date that the interface betweenthe deltatop to frontandbrackish waterdeltatop marshenvironments found today at PhanariBay has migratedat least 5.3 km seaward the expenseof the GlykysLimen since approximately at 2100 B.C. (Fig. 6.12). Brackish waterconditionsalso existedat the localityof core 94-17, which is locatedfurther inland,ca. 5.7 km fromPhanari Bay.A radiocarbon datewas not obtainedfrom this core,but it is reasonable assume to that the base of the delta top marshis approximately same age or the extentof slightlyolderthanthatof core94-23.The maximum post-glacial the marineembayment not knownbecauseonly a few relatively is shallow coresareavailable of cores94-17 and 94-23. east Core93-21 (Figs.6.7, 6.9;Appendix)showsa basalstratigraphy that is similarto cores 94-17 and 94-23, and providesanotherradiocarbon date that furtherhelps to constrain positionof the ancientshoreline. the Delta top and front sedimentsdirectlyoverliebedrock,and are in turn succeeded a deltatop marshenvironment. However,since93-21 is loby catedca. 0.5 km to the west of the othertwo cores,within the Mesopotamon/Tsouknida where severalfluvialsystemscoavalley constriction fluvialdepositsdominatethe stratigraphy abovethe delta lesce, subaerial A radiocarbon date from the delta top marshpeat returnsa top marsh. la calibrated rangeof agesfrom3690 +140/-60 B.P. (1740 +140/-60 B.C.). This age is approximately yearsyoungerthan the 14C 350 date fromcore which is appropriate that core93-21 is closerto the modern 94-23, given shoreline. West of the Mesopotamon/Tsouknida both geovalleyconstriction, logic evidence and historical documents provide information about the changing size of the Glykys Limen. Core 94-13 (Figs. 6.7, 6.10;

North
1200

A
-j

1100 II00 _ I000 -

Mesopotamon ridge o Acheron

Tsukid ridge Tsouknida ridge

900 -

800 8 -700 600 --

/
,

River

/_.

500 -

3 M
a
t K

N;C-94-20

NC-94-12

NC-(93-21

.........00 /

c
;'

400 4
300 -

|
fluvial channel and subacrial natural Icvcc
4: 1670 Cal. ('e-.1.C-14: 1670

floodplai -.

--

fluv and na

>

o C)

200

al,? lu , 100
E

_n<*:$backswamp -

? ............... . -.
""""".' -140 -601BP 1 ~'":"~-=-

/-

noo_~ 0
O^

120 -40/-120BP ?:-40 liP \

.>'

1c4

iuvia

~ ~ ~ ------ ~
[

-100 ~ ~

<

-200 -300 --400

-

fluvio-deltaicfill | I j / delta ti-ont distributary channel,channel mouth bar.and subaqueouslevee

200 300 111 .

7arsh \ \

T IT 0 . 100

\ and delta top and friont interdistributlar bay

50(x vertical exaggeration(

delta distributary channel

Figure6.9. North-south crosssectionthroughthe Mesopotamon/Tsouknida valleyconstrictio

2I9

Elevationabove sea level in centimeters

C5
IC

"0

K,
! c; ;> sW \' ^ ." - .

i >,

. Cit:::
R <

:Ei
0a-

(
,d

':::

1
C D\

Et
,,;
:.........

CD

o
""

mi :

-j

----- B~

= ::'

tri :

r I_
SD s

i I^R'> 2 ? ?C

Ill

c d

CCo

*C C

00

0 st

0 C l

-' 00 rCt1

00 0 C) 0 oc vC

SJOa9LUHUU3LUl[AOl B1S 3AOqb UOl}3A31j

Northeast
1200 1100 1000
cfl

900 800 700 600 500 400 300 NCNC-93-17

I-.

C) C)

P.
C) C)

P.

C)

-0
0 C)

200 100 0
-100 -200-

... -floodplain
-

.-

NC-92-16
.....

accretionar beach

I..

. . .............. - -cNC.'-93-19v_ '------.--. . ..

floodplain

el I erial' J-. g__,~~,~~

_~

-300-400- -:

Ishallow marineembayment (Glykys Limen)

'

''

300

600

900m

500x verticalexaggeration

Figure 6.11. Northeast-southwest cross section through the valley bottom (area of former marine emba

LOWER

ACHERON

RIVER

VALLEY

22I

~ Figure6.12. Paleogeographic reconstructions the lowerAcheron of

valley for 2100 B.C.and the 8th

.
\?

< ^

i \

V\

centuryB.C. Smallblacksquares markcorelocations;see Figure6.7 for labels. Appendix), ca. 3.5 km from modern Phanari Bay, is composed from the base upward of shallow marine deposits of the Glykys Limen which are overlain by delta front sediments. The delta front sediments grade upward into deposits of a distributary mouth bar, and then an interdistributary bay. The sequence is capped by subaerial fluvial sediments. A marsh reed retrieved from the distributary mouth bar deposit was radiocarbondated and returns a calibrated la range of ages from 850 +80/ -60 B.P. (A.D. 1100 -80/+60). The vertical sequence in this core indicates that it is not directly in front of the prograding delta, but on its flank.

222

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

Figure6.13. Paleogeographic of reconstructions the lowerAcheron

valley for 433 B.C.and 1 B.C.The

dashedline in the 433 B.C. panel course indicatesa possiblealternative for the VouvosRiver.Smallblack markcorelocations;see squares Figure6.7 for labels. Therefore, it is not appropriate to use this as an indicator of the actual delta front position, which would have been somewhat seaward of this location. A hypothetical delta front position for this time is illustrated in Figure 6.14. Several historical documents, in particularearly maps of the region, provide information that helps to reconstruct the evolution of the Glykys Limen since A.D. 1100.39The maps are clearly not geographically accurate, but they do indicate that the Glykys Limen was still of significant size through the 15th and 16th centuries A.C. Leake's description of the valley as he passed through the region in 1809 provides important infor-

39. Besonen 1997 shows 16 maps; of see note 1 regardingthe availability Besonen 1997.

LOWER

ACHERON

RIVER

VALLEY

223

Figure 6.14. Paleogeographic

\ K>j '
>

reconstructions the lowerAcheron of valleyfor A.D.1100 andA.D.1500. Smallblacksquares markcore locations;see Figure6.7 for labels.

v
Acherousian Swamp

mation about the landscape configuration east of the Mesopotamon/ Tsouknida constriction, but there are few details about the coastline and actual delta front.40 The modern village of Ammoudia, which surrounds present day Phanari Bay, did not come into existence until after Leake's time, in the early part of the 20th century.Therefore, the position of the shoreline in 1809 must have been a bit further to the east (Fig. 6.15). There is one radiocarbondate from the areaof the Glykys Limen that seems anomalously old, given its location and the type of deposit from which it was obtained. Core 92-20 (Figs. 6.7, 6.11; Appendix) is situated
40. Leake 1835. in the middle of the area of the Glykys Limen, ca. 1.6 km from Phanari

224

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

* ..1 ^"/. * ,~ ^X, jf-Ai)^ ' 7 "^/^^ C f/ xS^L^ d~~ c

"

-f, '

~y^^

f,.

y _

->

Figure6.15. Paleogeographic reconstruction the lowerAcheron of forA.D. 1809 anda mapof the valley modernlandscape. Smallblack markcorelocations;see squares Figure6.7 for labels.

Bay. It consists of inferred floodplain and natural levee deposits that directly overlie either bedrock or gravel. A radiocarbondate on organic material retrieved 50 cm above the base of the core returns a calibrated la range of ages from 2650 +70/-290 B.P., or 700 +70/-290 B.C. Such a date would suggest that the delta top was located here as early as 700 B.C., forcing the delta front position even further basinward.This is problematic since it is in gross contrast with the coherent sequence of coastal evolution documented by the rest of this study. The anomalously old radiocarbon date from core 92-20 is likely due to reworking of older deposits. The stratigraphyin the core is ratherpecu-

LOWER

ACHERON

RIVER

VALLEY

225

liar, and is only similar to that seen in core 92-16, less than 600 m away. Though the deposits in these cores are apparentlysubaerial(according to their color), they occur up to 5 m below sea level. Bedrock in the area is very shallow, as indicated by limestone knobs that stick up through the alluviumjust 500 m to the south, and 700 m to the west (Fig. 6.8). These bedrock knobs are covered with red sediment and vegetation at present, and would have been small islands before the infilling of the Glykys Limen. Consequently,it seems probablethat the deposits aroundthe bedrockknobs, such as retrieved in core 92-20, may represent reworked older sediment and material shed off of the islands.

CONTROLS ON SHORELINE IN THE ACHERON VALLEY

PROGRADATION

41. Waters 1994, p. 197. 42. Tziavos 1977. 43. Besonen 1997.

Our data indicate that the rate of shoreline progradation in the lower Acheron valley varied significantly through time; it was slow earlier on, but then much more rapid over the last millennium. In the 3,200 years from 2100 B.C. to A.D. 1100, the shoreline position progradedjust 2 km (Figs. 6.12-6.14). In the 850 years from A.D. 1100 to the present, however, almost 3.5 km of shoreline progradation has occurred (Figs. 6.14, 6.15). Rapid recent progradation is also supported by a detailed consideration of the 3-km wide system of beach ridges noted east of modern Phanari Bay (Fig. 6.3). As Waters has shown, the valley bottom is subsiding;41 if these ridgeswere accretingslowly over time during subsidence,one would expect to find them at progressivelylower elevations moving inland. This is not the case, however. Careful examination of surveyed point elevations on the 1:5,000 topographic maps shows that all the ridges are no higher than 1 masl, and no particularprogression of ridge elevations can be noted moving inland. This suggests that the beach ridges have accreted rapidly. What were the controls on the rates of shoreline progradation?The simple dynamics of basin infilling were probably important factors. Following stabilization of sea level in the middle to late Holocene, sediment deposition would have been directed toward filling the deeper parts of the Glykys Limen. As the basin grew continuously shallower,an increasingly largerproportion of the sediment load could be dedicated to the shoreline, we leading to the increasedrateof progradation have documented.A similar has been noted for the Spercheios delta on the eastern coast phenomenon of Greece, where delta growth seems to be occurring at a continuously increasing rate.42 One significant local geomorphic control that moderated sediment delivery to the coastline was the formation of the Acherousian lake. As will be discussed below, the lake probably did not come into existence until sometime between the 8th century B.C.and 433 B.C. (Figs. 6.12, 6.13). It then served as an efficient sediment trap,capturing materialtransported by the Acheron River that would otherwise have been carriedto the coast. The lake'sability to trap sediment was further enhanced by a spillway that was built increasingly higher, and subsidence of the lake floor.43These

226

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

factorsallowedthe lake to accommodate nearly9 m of sedimentinfill

before being breached, probably after A.D. 1100 but before Turkish times

the (Fig. 6.14). Once this occurred, AcheronRiverwas againableto deliverits sedimentload directlyto the shoreline. Several over controls, operating morethanjusttheAcheron larger-scale alterthe quantityof valley,may also havehad the abilityto significantly sediment deliveredto the coast, therebymoderatingrates of shoreline In influencehas been implicated anthropogenic progradation. particular, for as responsible a profound changein landscape stabilityandassociated eventsoverGreeceas a whole beginningaround 4500 erosion/alluviation B.P.44 In Epirus,pollen studiesfrom two sites located ca. 80 km to the erosiveeventsduring northof the Acheronvalleyalso recognizedseveral lakeand the middleto late Holocene.Using pollendatafromGramousti Willis recognized erosiveeventsfrom Rezinamarsh(Fig. 6.1), Katherine
ca. 6300-5000 B.P., 4300-3500 B.P., and finally at 2500 and 2000 B.P.45

influencewere cited as Though both climaticshifts and anthropogenic causesfor these periodsof increased inerosion,anthropogenic possible for fluencewas the morefavored explanation, especially the event dating eventsin the to 4300-3500 B.P.We do not recognize of theseerosional any evolution of the Acheron valley,despite its proximityto geomorphic but erosional Willis'sstudyarea, evenadjacent regionsmayhavedifferent
histories.46

A secondlarge-scale controlthat mayhavemoderated sedimentdea to the coastis a changein climate,in particular,changein moisture livery and balance. Though the systemof responses feedbacks maybe complex, balanceaffectvegetationcover,andthus couldeasily changesin moisture thereis little paleoclimatic alterthe effectivityof erosion.Unfortunately, available Greece,and that which does exist is predomifor information Someeffort,however, beenfocusedon interprethas nantlypollenwork.47 A in moisture balance basedon lake-levelfluctuations.48 lowing changes recordof interpretedlake-level fluctuationsexists for Lake resolution Ioannina, 55 km to the northeastof the Acheronvalley(Fig. 6.1),49 just but the datafromthe last 5,000 yearsaretoo sparseto relateto geomorA existsfor phic changesin ourarea. new recordof lake-levelfluctuations
44. Davidson 1980;van Andel, located on the other side of the Pindos Mountains,a majororographic Runnels,and Pope 1986;van Andel, to and Further- Zangger,and Demitrack 1990. boundary, thus a comparison our areais not appropriate. more,the datafromLakeXinias-like thatfromLakeIoannina-are very 45. Willis 1992. 46. In particular, the comparison see sparsefor the middleandlate Holocene. While middle and late Holocene paleoclimaticinformationfrom of the SouthernArgolid and Argive theredoes appear be increas- Plain regionsexaminedin van Andel, to Greecemaynot be the most impressive, and Demitrack 1990. event around Zangger, reviewsin Robertsand abruptaridification ingly robustevidencefor a significant, 47. See half and 4200 B.P. overthe eastern of the Mediterranean WestAsia.51 Pre- Wright 1993, and Willis 1994. this eventwouldhaveaffectedGreeceas well, and mayhaveled 48. Harrisonand Digerfeldt 1993; sumably in thusincreasing effectivity erosion Digerfeldt, Olsson, and Sandgren2000. the of to a reduction vegetation cover, 49. Harrisonand Digerfeldt 1993. and resultingin a higher flux of sedimentbeing deliveredto the coast. 50. Digerfeldt, Olsson, and addressed with the presentbody this However, issuecannotbe adequately Sandgren2000. information(e.g., mostly pollen analyses).Furof Greek paleoclimatic 51. Weiss et al. 1993; Dalfes, Kukla, to with a proxylikepolthis thermore, eventmaybe impossible recognize and Weiss 1997; Cullen et al. 2000; of influencethatbegins Weiss 2000. len becauseof the strongoverprint anthropogenic

Lake Xinias,50 just 160 km to the east of the Acheron valley,but the lake is

LOWER

ACHERON

RIVER

VALLEY

227

for of To at this time.52 resolvethe issue,development proxyrecords moisturebalancethat areunaffected humanactivity(e.g., an oxygenstable by wouldbe moresuitable. isotoperecord) of localfactorssuch as the dynamics basininfillingand In summary, lake of the formation the Acherousian certainly playeda rolein moderatof ing the progradation the shorelinein the Acheronvalley.Larger-scale influenceand climatechangewerecapable factorssuchas anthropogenic to of affectingthe amountof sedimentdelivered the coast,but the data to are available not yet sufficient link eitherfactorto the changcurrently ing rateof shoreline progradation.

MIDDLE AND LATE HOLOCENE EVOLUTION OF THE ACHEROUSIAN LAKE

the and We havedocumented development evolutionof the Acherousian lake(Fig.6.4),whichuntilnowwasmostthoroughly considered Dakaris by The followedby Hammond.53 absolutechronology our studyis based for of and dates,andpartlyon an analysis literary hispartlyon radiocarbon torical references.Unfortunately,the reconstructionsof Dakaris and Hammondwerebasedprimarily indirectevidence,the modernlandon in and that scapeconfiguration the valley, the assumption the lakefilledin shallower areally expansive and less overtime. Howbecoming gradually, for ever,the mechanismresponsible the impoundmentof the lake was and a infillsequence dynamic, thus it did not experience typicallacustrine andevolution. a the lakemaintained shallow Instead, profilebutgrewcontinuouslylarger,spreading upvalleythroughtime. As a result,Dakaris, the Hammond,and othersoverestimated size of the lake, at least as an of water. open body The developmentand evolutionof the lake is best recorded the by aroundand to the east of the Mesopotamon/Tsouknida valstratigraphy constriction(Fig. 6.2). Sediment cores 94-23 and 94-17 (Fig. 6.7; ley natureof the middleand late Appendix)documentthe overallregressive Holocene stratigraphy the valley,and the entirehistoryof the Achein rousianlake.As described above,the coresconsistfromthe baseupward of deposits fromthe following environments: deltatop to front,2) brack1) ish waterdeltatop marshgradingupward into freshwater marsh,3) shallow freshwater The shallowfreshwater lake delake, and 4) floodplain. is fromthe Acherousian lake.A radiocarbon on peat fromthe date posit bottom of the freshto brackish waterdelta top marshof core 94-23 rela turnsa calibrated rangeof agesfrom4030 ? 100 B.P., or 2080 + 100 B.C. We therefore concludethat the Acherousian lake came into existenceat some point afterca. 2100 B.C. The stratigraphy cores94-23 and 94-17 indicatesthat the marsh in was essentially drownedas the lakecameinto existencedirectly top of on it. Some mechanism the west of these coreswas therefore to responsible for the impoundmentof the lake. Analysisof the stratigraphy cores in 94-20, 94-12, and 93-21 (Fig. 6.7; Appendix),locatedca. 600 m to the west in the Mesopotamon/Tsouknida showswhatthis valleyconstriction, mechanism havebeen. might

52. See note 44 above. 53. Dakaris 1971; Hammond 1967.

228

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

The A-A' crosssectionbasedon these cores(Fig. 6.9) showsa massiveplug of fluvialsedimentsfillingthe valleyat this point.The stratigrathatareimmephyin thesecoresconsistsof deltatop andfrontsediments natural levee,and floodplain diatelyoverlain fluvialchannel,subaerial by In core94-23 consistsof the samedeltatop andfront sediments. contrast, sediments overlain 7.5 m of sedimentfromthe Acherousian lake.From by this relationship, is clearthat the lakewas impounded the eastof the it to becauseof fluvialsediments Mesopotamon/Tsouknida valleyconstriction that essentially pluggedthe constriction. of This fluvialplug records migration the channelandlevee systhe tem of the AcheronRiverand/orone of its tributaries. the channel/ As side leveesystembuiltsouth-southwestward the eastern of the Mesofrom near potamonridge,it eventually impingedonto the bedrock promontory Tsouknida closeddepression was (Figs.6.12, 6.13). As a result,a shallow, off to the eastbehindthis channel/levee/proximal pinched sysfloodplain tem and water ponded up, drowningthe delta top marshto form the Acherousian (Fig. 6.13). lake The stabilityand longevityof this fluvialplug systemareimportant of pointsto emphasize. Followingthe initialimpoundment the lake,fluvial sedimentation dominatedin the areaof the valley constriction has untilthe presentday,as illustrated crosssectionA-A' (Fig. 6.9).Thus, by as the channel/levee/proximal floodplain systemslowlyaggraded through rise time, it causeda progressive in the surfaceelevationof the lake.Becausethe lakewas alsoreceiving sedimentinput,this processallowedit to 9 accommodate m of sedimentinfillwhile simultaneously a maintaining shallowprofile.54 Moreinformation the regarding progressively risingsurface elevationof the lakeand its arealexpansion be discussed will below. This mechanism riverchannelandlevee migration an extremely of is evolutionin the valley,andits effectscan important agentof geomorphic be seen in the topography otherpointsin the valleytoday. at Three excellent examplesincludethe topographic to depression the west of Koroni, the depression betweenKastriand Kanallakion, the smalldepression and to the east-southeast Ephyra(Fig. 6.8). In these cases,the migrating of courseof the Kokytos AcheronRivers and impingedonto a bedrock highland and pinchedoff a shallow, closedbasinupvalley the constriction. of Richard Russellnoteda similar processin his studyof the MeanderRiver in westernAnatolia.55 this case,a rapidly In deltafront/coastal prograding plainbuilt acrossthe entranceto a marineembayment, essentiallytrapHe ping a standingpool of waterwithin the embayment. alsorecognized shallowlakes("levee-flank that depressions") hadformedon the deltatop in the areabehind/between intersection two streamchannel/levee the of
systems.56

date Though the radiocarbon fromcore94-23 indicatesthat the imof the Acherousian lake must have occurredafter ca. 2100 poundment B.C., a more tightly constrained chronologycould be determined anby other14C date at the top of the freshwater marshdeposit.Unfortunately, limitedresources not providethis option.Consequently, closerdatdid a of the lake'sinceptionwill be basedon an analysis literary hisof and ing toricalreferences ancientauthors. by Differingopinionsaboutthe accu-

54. Besonen 1997. 55. Russell 1954. 56. Russell 1967, p. 17.

LOWER

ACHERON

RIVER

VALLEY

229

racy and validity of topographic references made by ancient authors are certain. However, if such references,taken in chronological order,present a coherent and logical sequence of events, they may be useful. On the contrary,if they present a sequence of events that is clearly impossible, or if various references contradict one another, one may be inclined to question their validity.This, however, is not the case in the Acheron valley. A detailed analysis of ancient literary and historical references in chronological order presents a logical and coherent picture of the probable evolution and development of the Acherousian lake. That these references fit nicely within the story developed by geological and sedimentological evidence lends them some credence. The earliest reference to the valley comes from the Odyssey Homer. of Current thought suggests that the Odyssey may have been written in the 8th century, but describes some events and settings that go back to the 12th century B.C. in the Late Bronze Age. Homer writes: And when in your ship you have traversedOceanos, Where the scrubby strand and groves of Persephone are, Both tall poplars and willows that lose their fruit, Beach your ship there by deep-whirling Oceanos; But go on yourself to the moldy hall of Hades. There into Acheron flow Puriphlegethon And Cocytus, which is a branch of the Styx'swater, And a rock and a concourse of the two resounding rivers.57 Homer makes no mention of the lake; in fact, he strictly describes a scene in which several tributariesfeed into the Acheron River.The adage "lack of evidence doesn't constitute evidence for a lack"is applicable here, but it may be suggested that Homer does not mention the lake because it did not exist at the time a contemporarywitnessed the topography in the valley. The lake then probably formed at some point between the writing of the Odyssey the 8th century B.C.)and the time of Thucydides' account (in of the valley (about 400 years later), when the lake is mentioned for the first time. Thucydides, who wrote contemporaryhistory,gives a description of a recently nascent Acherousian lake in his account of the Battle of Sybota in
433 B.C.

It is a harbour,and above it lies a city away from the sea in the Eleatic district of Thesprotia, Ephyra by name. Near it is the outlet into the sea of the Acherusian lake; and the river Acheron runs through Thesprotia and empties into the lake, to which it gives its
name.58

Of interest here is the fact that Thucydides strictly states "nearit is the outlet into the sea of the Acherousian lake,"as if the lake empties directly into the sea. This seems to imply that the Acherousian lake and the sea
Cook, ' NewYork, 1967. 58.Thuc.1.46.4,trans. F. Smith, C. Mass.,[1928]1956. Cambridge, 57. Od 10.508-515, trans. A.

(actually the Glykys Limen) are very close-the two are split by only a very narrow barrier of land on which is situated the lake spillway (Fig. 6.13). This narrow barrier of land is the channel and levee system of the Acheron (or one of its tributaries) that caused the impoundment of

230

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

the lake, as explainedabove.Thucydidesclearlyidentifiesthe Acheron Riveras flowing into the lake,but says nothing about its exit from the lake. in His accountis distinctfromall laterreferences that it suggeststhe lake and the sea. Later accounts extremeproximityof the Acherousian and is suggestthatmorethanjust a lakespillway present, thatthe channel waterfromthe laketo the sea is long enoughto be identifiedas carrying For that of the AcheronRiver. example,Strabowrites: and Then comesCape Cheimerium, also GlycysLimen,into which the RiverAcheronempties.The Acheronflowsfromthe riversas tributaries, that so Acherusian Lakeandreceivesseveral it sweetensthe watersof the gulf.59 the that It appears the stripof landseparating lakefromthe seahadgrown and of wide in the 400 yearsbetweenthe accounts Thucydides sufficiently the lake Strabothatthe channeldraining Acherousian couldbe identified fed several tributaries into the as that of the AcheronRiver. Furthermore, Acheronafterit exitedfromthe lake. lake date Basedon the radiocarbon fromcore94-23, the Acherousian must have formed after2100 B.c. And, since Homer,Thucydides,and coherentpictureof the development Straboall presenta chronologically it that of the lakein the valley, is probable the lakeformedsometimeafter bit B.C. the 8th century butbefore433 B.c.An additional of circumstantial evidencesupportsthe notion that the lake did not come into existence recordfromthe valuntilthis time.Dakarisnotedthatthe archaeological in duringthe Archaicperiod(ca.700ley indicateda decrease population
500 B.c.),60and the data from the diachronicsurveyof the Nikopolis Project

tend to supportthis conclusion.Dakarissuggestedthat the population to whichhas always been a probdeclinemighthavebeenrelated malaria, lem in the low-lying coastalareasof Epirus.Why malaria would have time was unknown,sinceDakarisprobably asflaredup at this particular sumedthe lake had been presentin the valleyfollowingthe post-glacial rise of sea level. Our analysisseems to indicate,however, that the Acherousianlake came into existenceat the same time as the Archaic-period decline.While the timingof theseeventsmaybe coincidental, population the birthof the lake and associated swampyareasmayhavegiven rise to the malarial by epidemicpostulated Dakaris. BecauseDakarisandothersdid not recognize mechanism the responsiblefor the lake'simpoundment, they assumedthat it followeda typical lacustrineinfill sequenceand becameincreasingly shallowerand areally less expansive throughtime.In contrast, suggested PhilippsonandKirsten that the lake had become largersince ancienttimes,61 they did not but this explainwhy they considered to be the case nor did they provideevidenceto support theirconclusion. They alsoplacedthe laketoo farupvalley 59. Strab.7.7.5 [C 324], trans.H. L. (Fig. 6.4, upper right). The results from our study suggest that their Jones, Cambridge,Mass., 1960. the assertion but docuregarding size of the lakeis correct, we furthermore 60. Dakaris 1971, p. 12. ment the mechanismand detailsof the lake'sevolutionas well as its true 61. Philippson and Kirsten1956, II, location. p. 105.

LOWER

ACHERON

RIVER

VALLEY

231

Soon afterits formation,the lake existedas a shallowbody of open watersurrounded a fringeof marshy by ground(Fig. 6.13). Sedimentcarried by the Acheronwould have quicklyfilled it in were it not for the slowlyaggrading spillwaymechanismdiscussedabove.This allowedthe laketo accommodate increasingly an volumeof sedimentvertically larger and areally, the lake expanded as becauseof its slowlyincreasing upvalley
surface elevation.62

62. Besonen 1997. 63. Dakaris1971; Hammond 1967; Leake 1835; Philippson and Kirsten 1956.

Evidencefor the initialsmallsize of the lake,andfor the subsequent of can expansion marshy, swampygroundupvalley, be seen by comparing the stratigraphy cores94-23 and 94-17 with that of core 93-22 (Fig. in 6.10;Appendix).Cores94-23 and94-17 arelocatedjust to the eastof the fluvialplug in the valleyconstriction contain7.5 and 5.9 m, respecand of lacustrine mud and clayfromthe Acherousian lake.These lake tively, depositsbegin at 3.1 and 1.7 m below sea level, and run to 4.4 and 4.2 Core 93-22 is locatedca. 1 km east of cores94-23 and masl,respectively. 94-17, in the areaconsidered Dakarisandothersto be the ancientlake. by At this locality, a however, muchthinnersequence (3.5 m) of mixedlacustrine and marshdepositsoccursbetween1.2 and 4.7 masl.The lake dewith posits in the core are underlainby a very stiff floodplainalluvium somepedogenicdevelopment. of and Thus,thispackage lacustrine marshy and nature depositsshows stratigraphic onlap upvalley, its transgressive confirms gradual the increase the lake's of surface levelandits areal expansionthrough time.The lakeprobably neverextended muchfurther upvalley than the locationof core 93-22 becausethe mixedlacustrine marsh and of depositin this coreis indicative the lakeedge and shore. This information helpsconstrain size andlocationof the lake, also the at leastas an open bodyof water.Mixedlacustrine marshsedimentaand tion at the locationof core93-22 could not havebegununtil the surface level of the lake had reachedat least 1.2 masl (i.e., the base of the lacustrine materialin that core).When did the lake surfacelevel reachthis elevation? ignoringfactorssuch as subsidence changesin the rate and By of sedimentation spillwayaggradation, can looselybase it on the or we fromcore94-23. Elevationally, maslcorresponds 1.2 chronology approxisedimentation matelywith the middleof the lacustrine sequenceof core 94-23. Fromourpreceding of analysis core94-23, we concludedthat the lake probably came into existenceafterthe 8th centuryB.C., but before 433 B.C.Continuous,uninterrupted thereuntil after depositionoccurred the FirstWorldWar,at whichtime the finalremnants the swampwere of backfilled. Assumingthat the surfacelevel of the lake rose at a constant rate,it would have reached1.2 masl in the middle of this time span,or lake and marsh roughlyA.D. 850. Thus, we estimatethat the expanding reachedthe localityof core93-22 aroundthe 9th century A.C. ground This evidencesuggeststhat Dakaris,Hammond,and othersgreatly overestimated size of the lake (Fig. 6.4), especiallyconsidering the that theirreconstructions supposedto show the extentof the lake during are the classical In the period.63 some of the reconstructions, shapeandlocation of the lake contradict moderntopography. example, the For Dakaris andHammondsuggestthat the lakehad a northeast/southwest-trending

232

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

the shorebetweenMesopotamon Kastri. and lines However, topographic thatwouldhavedefinedthe lakeshorein this areahavea northwest/southeast trend,exclusiveof the elevatedsubaerial naturalleveeswhich flank the AcheronRiver(Fig. 6.8). also Dakaris's reconstruction suggeststhat a branchof the lake extendedto the eastbetweenPountas Kanalridgeandthe villagesof Kastri, but This areais a closeddepreslakion,andAcherousia, this is not correct. sion (Fig. 6.8) that came into existenceby the same mechanismwhich the of lake. caused impoundment theAcherousian In thiscase,theAcheron riverchannelandlevee systempinchedoff the depression againstthe tip of the Pountasridge,which projects fromthe south.This depression, up therefore,would not have come into existenceuntil the course of the Acheron shifted to the south of Kastri.As we discussbelow,this shift occurred veryrecently, probably perhapsaroundthe end of the 16th century A.C.

There is additional geologic evidence to suggest that the main body of the Acherousian lake to the west of Pountas ridge was not confluent with

lacustrine andclays silts the waterbodyto the eastof the ridge.Laminated

do indeed occur in this small basin, but they form a relatively thin layer and are too high topographicallyto have been deposited by the Acherousian lake. Core 94-03 (Fig. 6.7; Appendix), taken from the center of this small depression, is composed of a backswamp deposit overlain by a freshwater marsh deposit, which is in turn succeeded by floodplain deposits. Core 94-21 (Fig. 6.7; Appendix), located just 450 m to the west, exhibits identical stratigraphybut bottoms out with a floodplain deposit as well. Though core 94-03 did not penetrate these lower floodplain sediments, its proximity to core 94-21 and the fact that it is shorter support the inference that further penetration of core 94-03 would have encountered the same floodplain deposit. The Acherousian lake would have necessarily had a surface elevation at or below the elevation of the fluvial plug sediments that impounded it. This fluvial plug was continuously aggrading,but never reached more than 5.0 masl, the present elevation at the Mesopotamon/Tsouknida valley constriction.Thus, sediments from the Acherousian lake could only have been deposited up to this height. But the backswamp and freshwater marsh deposits in cores 94-03 and 94-21 occur between 5.1 and 7.7 masl. Therefore, the body of standing water in which these sediments were deposited could not possibly have been confluent with the Acherousian lake as the standing water had a significantly higher surface elevation. This conclusively proves that the body of ponded water that once existed here was not a branch of the larger lake as Dakaris indicated. By Turkish times, the Acherousian lake had become a swamp with a few isolated pools of water (Fig. 6.14).64 Continued growth of the Acheron riverchannel and levee system split the remains of this swamp.This interpretation is supported by the broad topographic high of the river channel and levee system to the east of Mesopotamon, and by the closed depression directly to the east of Ephyra, created when the channel and levee system impinged against the bedrock ridge (Fig. 6.8). Leake provided an

64. Hammond 1967, p. 39.

LOWER

ACHERON

RIVER

VALLEY

233

excellentdescription the marshy of valleybottomfromhis travels through the regionin the springof 1809, and he noted that several pools of open waterstillexisted(Fig.6.15).65 Afterthe FirstWorldWar,the finalmarshy remnants the formerAcherousian werefilledin for agriculture.66 of lake

THE CHANGING COURSE OF THE ACHERON WITH RESPECT TO KASTRI

In order to reconcile the archaeological remains in the valley with the accounts of ancient authors, Dakaris suggested that the Acheron River had shifted its course to the south of Kastri since classical times.67Unfortunately, he could not provide geologic evidence with chronological control to supporthis theory.Cores 94-02 and 94-04 providethe evidence to document this shift. Core 94-02 (Fig. 6.7; Appendix) was retrieved north of Kastri, between it and the larger of the two hillocks named Xirolophos (Fig. 6.2). The core consists from the base upward of deposits from the following environments: 1) floodplain, 2) backswamp,3) floodplain, 4) fluvial channel, and 5) floodplain. At the interface between the lowest floodplain unit and the backswamp, a small reddish pottery fragment was encountered. The fragment is abraded and lacks diagnostic features, but ceramic specialists on the project have suggested that the texture of the sherd should place it some time in the classical period. Since this pottery fragment occurs below the deposits of a fluvial channel, it provides a terminus post quem for the existence of the river channel at that location. Therefore, at some point past the beginning of the classical period, a fluvial channel existed north of Kastri. Core 94-04 (Fig. 6.7; Appendix) was also retrieved north of Kastri, between the hillock of Koronopoulos and the largerof the two Xirolophos hillocks (Fig. 6.2). From the base upward, deposits from the following environments occur in succession: 1) floodplain, 2) backswamp, 3) fluvial channel, 4) backswamp,and 5) floodplain. The fluvial channel sediment is over 1.5 m thick, and contains gravel clasts up to 1 cm in diameter.This deposit is from a significant river channel, like that of the Acheron, and not from a smaller stream.A radiocarbondate on a piece of wood from the base of the fluvial channel deposit returns a calibrated la range of ages from 380 +90/-70 B.P., or A.D. 1570 +70/-90. For radiocarbon dates this young, however, the calibration curve is relatively irregularand the specimen could date to almost any time during the last 500 years.Nevertheless, the radiocarbon date shows that a river channel, probably that of the Acheron River, was operating to the north of Kastri within the last 500 years. When Leake passed through the region in 1809, he recorded that the Acheron River followed a course to the south of Kastri, as it does today.Therefore, if the fluvial channel sediments in core 94-04 are indeed from the Acheron River, it would suggest that the course of the Acheron shifted from the north of Kastri to its south sometime between ca. 1500 and 1809.

65. Leake 1835, I1, 232; IV, p. pp. 51-54. 66. Hammond 1967, p. 68. 67. Dakaris 1971, pp. 136-137.

234

M.

R. BESONEN,

G. RAPP,

AND

Z. JING

CONCLUSIONS
Numerous ancient authors,beginning with Homer in the 8th century B.C., make reference to the lower Acheron valley and indicate a landscape configuration that is significantly different from at present.Three notable discrepanciesbetween the ancient and modern landscapeexist.The first problem concerns the size of the Glykys Limen (modern Phanari Bay), which at present is very small, but was much larger in ancient times. The second significant discrepancyconcerns the evolution of the extinct Acherousian lake, which ancient sources indicate was a conspicuous feature in the valley.The final discrepancyconcerns the course of the Acheron River,which today flows to the south of Kastribut was once located to the north of that site. Geologic evidence based on twenty-eight gouge auger sediment cores taken at various locations in the valley indicates that significant geomorphic change has occurred in the valley during the last 4,000 years. The shoreline of the Glykys Limen has prograded nearly 6 km in that time, doing so at varying rates. The Acherousian lake developed relatively late in the Holocene probablybetween the 8th century B.C.and 433 B.C. Since that time it has been filled in by natural alluvial processes, modified by a constantly aggrading spillway.Finally, the Acheron River appearsto have occupied a channel to the north of Kastri, and has only shifted to the south of that hillock in the last 500 years. It appears that the discrepancies between the ancient accounts and the modern landscape are not due to errorsin the ancient sources, but are instead the result of a naturalsequence of landscape evolution in the valley. Furthermore, careful examination of the ancient accounts may in some cases provide details and information for paleogeographic and paleoenvifrom the geologic record. ronmentalreconstructionsthat arenot recoverable The disciplines of geology and archaeology find a natural interface here, both contributing to, and benefiting from, one another. Indeed, the dynamic geomorphic evolution seen in the Acheron valley during the last 4,000 years reaffirmsthe need for multidisciplinaryarchaeologicalinvestigations that strive for a broad understanding of the dynamics of environmental change.

LOWER

ACHERON

RIVER

VALLEY

235

APPENDIX: CORE STRATIGRAPHY AND LITHOLOGY

from all twentyThis appendixcontainsthe sedimentcore stratigraphy Width of the core,lithologicpatcorestakenin the Acheronvalley. eight reflectthe grainsize andtype of terns,anda "Sediment description Type" matter observations. sediment basedon fieldandlaboratory Organic present is in the stratigraphy indicated one of the symbolsin the legendbelow. by "Color" 14C Locationsof calibrated AMS dates areindicatedby arrows. to (according the Munsell Soil Color Chart),weight percentof organic matterdeterminedby loss on ignition analysis("%OrganicContent"), are and resultsof the microfossilanalyses("Microfossils") also included our of field represents (see legendbelow).The "Environment Deposition" data.All primary basedon all available of the stratigraphy interpretation data,includingresultsfrommagneticanalyses, pipettegrain-sizeanalysis, microfossil platesandcounts,andanydatanot includedhere,canbe found in in Besonen1997,which is freelyavailable AdobeAcrobatPDF format note 1 for details). (see Symbol
w\il

Explanation
common matter coarse-grained organic abundant matter organic coarse-grained few to tracecoarse-grained matter organic common matter fine-grained organic abundant matter organic fine-grained few to tracefine-grained matter organic calibrated C-14AMSdatein yearsB.P. brackish marine to qty. XXX = quantity/totalnumber of freshwater, and in microfossils the sample water, reworked of formsin quantity XXX 1.3%F = percentage freshwater = percentage brackish marine to of waterformsin 91.4%B XXX quantity of microfauna quantity in XXX 7.3%R = percentage reworked

(1J_) _~
(_i)
C-100 44BP F, qty.615: 1.3% 91.4%B, 7.3%R

Sediment Type

Color

% Organic Content
0
I I I

Microfo

a- -L C- a- CD 190
slightly sandy silt
(KJL/) ('JJ)- - - - -

0j
I

50 /140 100 / 90
cr,

not recorded 5Y5/2

mud 5Y4/2

150 /40 200 /-10

0-

2501/-60 300 /-110

slightly silty clay

5Y5/2

Cr

350 /-160 400 /-210 4501/-260 500 / 310 fine sand 5Y5/1

0a) 0

0z
01

5501/-360 600 / -410 650 /-460 700 /-510 7501/-560 a-= a 0 a a-L
Zs a

clay 2.5Y; N5/ 7.5YR;_N-4f-

NC-92-16

Sediment Type

Color

% Organic Content

Microfo

0 /100

50 /50 100 /0 150/ -O 200 /-100 250 /-150

a) a)a)

mixed beach sand

not recorded

300 / 200 350 /-250

cj -

~0
C
C.) a1)

C)
0

C)

NC-92-1 7

Sediment Type

Color

% Organic Content

Microfo

0/100

_ _ _

mud mixed beach sand

not recorded 2.5Y5/2

2.5Y4/4

150/ -50
200 /-100 ''^'
i

()

mud -

CD

C)

~0
110

0
0

NC-92-18

Sediment Type
CD Ct C

Color

% Organic Content
) 0)
t-i

Microfo

0)

-100 /40 -50 /-10 0 / -60

1n

sea level - - - - - - - - - - - - - - not recorded-5Y3/1

7.5YR5/0 with 2.5Y5/6 mottles 4E-25Y6/6with 5Y3/2 mottles 2.5Y4/4 with 41-2.5Y4/0 mottles

50 /-110 100 /-160

fine sandy mud

0-

150 /-210 200 / -260

..... . .. . ..

~ ~

(\ lL /

muddy fine sand fine sand with clay interlayers

tf

250 /-310 0) 300 /-360 350 /-410 400 /-460

5Y4/1 clay grading upwardto slightly clayey silt

-F :

3~~~~~C
CD/

2.5Y3/0

0 C.

4501/-510 500 /-560 5501/-610 600 / -660

0D

5Y3/1

NC-92-19

Sediment Type

Color

% Organic Content
C)C)
tj wJ

Microfo

o 90
50 /40 100 /-10 150 /-60 200 /-110
CA 0a) 04-

a,-~3

n-7 not recorded

-

Ji 4 - (\JL/)

- clay 2.5Y6/5 with IOYR5/6 mottles

E-5Y5/ Iwith -- 5Y6/6 mottles 2,5Y4/4 mottles

250 /-160
300 /-210

350 /-260 400 /-310

\jIL

fine sand to muddy fine sand interbeddedclay, mud, silt, and muddy fine sand BEDROCK OR GRAVEL

2.5Y3/0

0~
0) Q 0--

450 /-360 500 /-410 550 /-460 600 /-510 2 0 P := a l0 a C,C a

\IIC-14: 2650 = +70/ 290 BP

2.5YR3/0

01
0)

NC-92-20

241

a o

OS-

JS

30 20 10 V 20-

p e- 0

pebble

c. sand -c-

m. sand -

- m. sand

f. sandsilt clay 1 I

,- f. sand - silt

o
? )

o
?

o
m

Cl o)
m

o
t

oC

o C
t

o
^

o O o
e n

C o

C>
_

(sl3mj3

luo33) JlA3l 3oS OAOqt UOlthA3jp /

JlOO Ul

qldaI

Sediment Type

Color

% Organic Content
o0
_

Microfo

o
I

N?

(-

0o
I

0
I

0/300 50/250 100/200 ,

--C~~~~~~~~C ~~~~~~~~~.._
~~ e level

silt mud slightly silty clay mud

2.5Y6/6

150/ 150 200 /100 250/50 300/0

_/3

2.5Y5/4

level ---sea ~.-. _::_-.._...k,

(_I/

N3/0 and N5/0 clay 5G5/1 5BG5/1 slightly sandy silt fining upwardto mud
5G5/1 to 5G4/1 with 20-50% 2.5Y5/6 mottles

4-

5Y5/2

350 /-50 400 / -100

-?

450 /-150 500/-200 -

r(

_: :/:_ : :::
'.." _'fl (_ _ .'. v.
_.. : '.

- __ _- _ _
_'.'.n '.m 4 : 7

clay

5G4/1

0
0
cn

550 /-250 600 /-300 _) B..".: - J

. . _ .. _ .. /::'.o _ .. ......-650 /-350 - :::...'-':':. _io. ':?*'rf'.' 700/-400 ~.' ' :' '-/: ".: .-: _ .' '

interbeddedgravelly sands to fine sands N5/0 to N4/0

?_,.-_--L

o
.-

. -se levelo

750 /-450 800 /-500 850 / -550 900 /-600 950 /-650 -

...

...

Q2

i'a'|Xt; ?.

X'Xa

(\JI)

~~~~~(iJ/ ^lll?~~(1~
*;: *--:*.:
; ,;* * ;

fine sandy mud grading upwardto fine sand

N3/0

*-

1000 / -700 1050/ -750 1100 / -800 1150 / -850 1200 / -900 1250/ -950

.:' ':g------- .
-

laminatedclays, muds, and silts with several cm-thick muddy fine sand layers 5G4/1

5GY4/1

NC-93-14

Sediment Type

Color

% Organic Content
o0
s _I

Microfo

0o

0/50 50/0 100/ -50 150 /-100

0
... :.-'

0 I

sea level
.-. : .. . .,.,

gravelly sand fining upward to sandy mud

-1a-YR33:: 5Y5/6 5Y6/6 5G4/1

200 / -150 250 / -200 300/ -250 350 / -300

'

. -_...:._: ...

.i._.

' interbeddedfine sands and muddy sands with some mud layers

N4/0
N4/0 interbedded with 5Y4/1

13

0
ra (D
-

400/ -350 450/ -400 500/ -450 550/ -500 600 / -550 650/ -600 700/ -650 mud ' ' ',':',., :
._.?1 . -.'.--.'' '.-_.-". -. '~..

5BG4/1

0
o0 {D 0D

5G4/1 5GY4/1

'.
i

v fine sand

0 0

-_c. 0

CL

NC-93-15

Sediment Type a a-Q. a0/50 50/0100 / -50

Color

% Organic Content
o0
I

Microf

o
I

0o
I

0
I

..............
.. _ _ . . . .. ... -. ..

sea level
.. . .. .

xl/
... -

slightly clayey silt grading upwardto fine sandy silt mud

muddy fine sand

10OYR4/4 5Y6/4 and

5Y6/6 5G4/1 5Y5/1 5G5/1

150 / -100-a a)

200 / -150250 /-200? . :_ .. -..

gravelly sand muddy fine sand

300 / -250

a , a
Da)

*-4

c~
c)

NC-93-17

Sediment Type
.. _ _
Ii 1

Color Color

% Organic Content % Organic Content

I I I I

Microfo Microfo

- 0- sr CL
1 1 1

0/50 50/0

.*-*

-..

..- ....

- .**-** --

sea level
- -----I --

\jL/

= _==

i
^

Cr

z ---r~ 100 /-50 *| ";- D;.Jl " f.:.: .@ ~?.'i-: e. B'~.ri 150 /-100 - '*;g *- :.C' ;-.'- .'-, '.- '..,'.'- ?/ ,,.-.i****^'

g +90/-30BP

slightly clayey silt mud and peaty mud

iinteri6edded-N2/70 --. -andlQOYRl5/6 5Y4/2 and 5Y7/3

10Y3/2

200 / -150 250 /-200 300 /-250

+-,

;-i-i-I li

t- r. _.****

.n

interbeddedclay, mud, and gravelly sand

interbedded 5Y8/3 and 5G4/1 5Y8/2 5GY4/1

350 /-300 400 /-350 450 /-400 500 /-450 * ** 3 2

mud coarsening upwardto fine sand

5BG5/1
5BG5/1 with some 10YR5/3 laminae

0 _0

550 /-500 600 /-550 650 /-600 700 /-650 750 / -700 -

o--__((i-i _ _ _ 0 _ _ _ _

laminatedclays, muds, and silts

5G4/1

800 /-750 850/ -800

p.g....
? .

-~

NC-93-18

.-

r
00 CD

Sediment Type

Color

% Organic Content

Microfo

0 0/90 50/40 100 / -10

sH

Pz 0
C.c

--

5Y6/4 _ _ _ _ _ fine to coarse sand with some muddy layers 5GY4/1 5B4/1

..-

N2/0

---

5Y5/4

150/ -60 200 / -110 250 /-160 300 /-210 350 / -260

4--4

0
a)

5BG4/1 to 5GY4/1 laminatedmuds, silts and fine sands

co 0 u 0

400 / -310 450/ -360 500 / -410

i0
0
W-

550 / -460 600 / -510 650/ -560 \L

\AZ/ C-14: 5140
+160/-100 BP

5GY4/1 muddy fine sand mud and peaty mud clay N4/0

g
0
*-

700 / -610 750 / -660

<u 0)

0u

. V.

NC-93-19

" 0

p -

ac a

CD

Sediment Type Type

Color Color

% Organic Content
0 0 0

Microf

0 /500

50 /450 100 / 400 150 /350

IOYR6/6 slightly clayey silt 2.5Y5/6

5GY4/ land 2.5Y6/6

a-) a-)
0

200 /300 250 /250 250 cm above sea level

mud and peaty mud BEDROCK

N3/0

aa-) a-) a-)

0

a-)4 a-) a-)

0
0

a-)

NC-93-20

Sediment Type
ca0--& a

Color

% Organic Content

Microfo

0/500 50 / 450 100 / 400 150/350 200 / 300 250/250 300 / 200
0 (_L) . ' ... .. -. ... ? .

slightly clayey silt with some sand muddy fine sand with interbeddedsilty layers slightly clayey silt and mud with some interbeddedsilty and sandy laminae sand and gravelly sand with some interbedded muddy layers

10YR6/6

.... . .. .... .
.... ......... .

2.5Y5/6

2.5Y5/8 5GY4/1
interbedded N4/0 and N3/0 5GY4/1 with some N4/0 layers 5Y5/2 and 5Y5/4 5GY4/1 and 5Y5/4

350 / 150 400 / 100 450 / 50 500/0 550 / -50

'-.' ,' e-'.? (

0D 0

00
a)

? '-'

.40.60 BP

mud peat and peaty mud interbeddedmuds and silts with some fine to gravelly sand layers

600 / -100 650 / -150 700 / -200 750 / -250

-. .C'. _ -

7.5Y3/2
10YR4/1 ---:N30Oand_N4ZG0:::

800/ -300

gravel BEDROCK

5GY4/1 with some 5Y4/1 layers

NC-93-21

Sediment Type
:0 C L- ?

Color

% Organic Content
o0 o 0o 0

Microfo

0 / 520 50 / 470 100 / 420

rA

slightly clayey silt

2.5Y5/2 and 5Y4/3

5Y4/2 to 5Y4/4 slightly silty clay 5GY5/1 5GY4/1 5BG4/1 5GY4/1 with 10OYR5/6 mottles 10 OYR5/6with 5BG5/1 mottles

150 / 370 200 / 320

a)

250/270 300 / 220

c-i-

350/ 170 400 / 120 450 / 70 500 / 20 sea level _ _a ||

0 0
a1)

slightly clayey silt slightly silty clay

qty. 116: 3.5% F, 1

550 / -30
a)
ac', o<

a) 5-4

a)

NC-93-22

Sediment Type

Color

% Organic Content
0 0 _I I 0
I

Microfo

0 / 1280 50 / 1230 100/ 1180

o-o

2.5Y4/3 slightly clayey silt 2.5Y5/3 peat and peaty mud :5BG6/1-5BG6/1 5B5/1 mud 5GY5/1 5Y4/2 slightly clayey silt 10OYR4/6 mud with some fine sand interlayers poorly sorted silty fine sand 800 cmw above sea level
_i i . ?_ ' .i
_

150/ 1130 200 / 1080 250/ 1030 -

+j 0 ? 0
0 I

300/980 350/930

qty. 41: 97.6% F, 2

0 0
0) C0 0 (D

400/ 880 450 /830 500/780 -

qty. 7: 100.0% F, 0

CA 550/730 -? 600 /680 ?

650/630 700/580 -

qty. 132: 1.5% F, 0.

5B5/1 5Y5/3
5Y4/3 & 10Y5/4

qty. 210: 0.0% F, 0.0

qty. 220: 0.0% F, 0.0

1.
V-4 u
0
0

750/530 800/480
i
CD

NC-94-01

Sediment Type
CD

Color

% Organic Content

Microfo

0/ 1580 50 /1530 100/ 1480 150/ 1430 200 /1380 250 /1330 300/ 1280
....... ... ... . .
. ... .... ... ... ..._...._ ... .. . .... .. .. ._. . .

slightly clayey silt with some fine sand interlayers -sand fini-/ig gravelly ... upwardto fine sand slightly clayey silt with some muddy, fine sand interlayers mud

2.5Y5/3

.... . . _ . .. .

. _ -: .

2.5Y7/4 2.5Y6/6 5BG4/1 5GY4/1

350/ 1230 400 /1180

c)
0-

450/1130 500/ 1080 550/ 1030 600/980

_ _ _ _ .. .. ... ... ...

slightly clayey silt _~

5Y5/4

650/930 700 /880 750/830

f2

13
*=-

. .

n-

a) c)

NC-94-02

Sediment Type
c AD Z CD

Color

% Organic Content
o>
w ON

Microfo

o/ 970
50 /920 100 /870 150 /820 200 /770 250 /720 300 /670 350 /620 400 /570 0 0
0 0.

. . . ... ... . ..

~lc

2.5Y4/3 slightly clayey silt 5Y5/2 i.

qty. 12 1:I00.0% F

peat and peaty mud

5G4/1

qty. 76: 100.0% F,

mud 500 cm above sea level

5BG4/1 5BG5/1

450 /520 500 /470

qty. 121: 89.3% F, 0

0 -CD 0- c

0
0 0

NC-94-03

Sediment Type

Color

% Organic Content
I I I I

Microfo

0/1780 50/1730100/1680 150/1630

rJ

2.5Y5/3 slightly clayey silt with some fine sand interlayers

- --

2.5Y4/3 2.5Y5/4
5Y5/2

200/1580C-14:

J/

380

250/1530a

300/1480350/1430 400/1380-

." : . '." . . ..

.._.

. .

5G5/1 very poorly sorted muddy, fine to gravelly sand mud slightly clayey silt

.. .. -..-.. .. _.

._ .

5G4/1

c)

450/1330500/1280550/1230600/1180650/11301150 cm above sea level 5GY5/1 5Y4/2

_ __ _2
I i _^

i *

*-

ct

___C

NC-94-04

.oo
y-

.e B p r

Sediment Type

Color

% Organic Content
-o o

Microfo

,n vL0-

0/ 2210 50/2160 100/2110 '.'- ':, ::. .':._. .:: '.:_--.:;. ' .-::-." (_~z) slightly sandy silt 2.5Y4/3

150 /2060 ,Ia

slightly clayey silt

200/2010 250/ 1960 300/1910 350/1860

_._<. _ ._...(_ _ .

5Y4/3 transitional

mud

5GY6/1

au
rA

400/1810 450/1760

500/ 1710 550/1660 600/1610 -

*, . . . :-'p. . :--*.......

i-^ i

slightly silty clay

5GY5/1

o 0

, ?

?~

650/1560 700/1510 750/1460 800/1410 850/1360

mud with a few silt and fine sand interlayers

_ ._._

5BG4/1

poorly sorted muddy sand slightly clayey silt with some sand near top 1300 cm above sea level 1300 .eve._ . above sea

5Y4/2

C)

900/1310 950/1260

Q..

NC-94-05

Sediment Type

Color

% Organic Content
0 ) -P < C 00

Microfos

0/425 50/375 100/325

a)

, _ .. .. _ .. .. _

(WI)

5Y5/3

150 /275 200/225 -

2.5Y4/4 slightly clayey silt 2.5Y3/2

qty. 800: 0.0% F, 0.0%

E
0

au

250/ 175 300/ 125 350/75 400/25 sea level

_

5Y4/2

c)

450 /-25 500 / -75 550 /-125 5Y4/4

0D

(D

a)

600 /-175 650 /-225 700 /-275 -

... .._
( I)

muddy fine sand mud

2.5Y4/4

qty. 559: 1.6% F, 0.0

5G4/1 poorly sorted muddy, fine to coarse sand interbeddedmuds and silts with some fine sand and clay interlayers 5BG4/1

a) 0

750 /-325 800 /-375 -

0 0
* _

850 /-425 900 / -475 o

_ .. ?____ ..

0 CD.

NC-94-08

Sediment Type
CD

Color

% Organic Content
o=
FJ
I I

Microfo

4~- O~
I I

00
I

0 /500

50/450
100 / 400
00

kL/)0

2.5Y4/3 2.5Y4/4 IOYR3/2 slightly clayey silt

150 /350 200 /300 250 /250

qty. 276: 0.4% F, 0

300 /200 350 / 150 400 /100 450 /50 500/ 0

- - - - - - - - .. . .. . . ..

qty. 90- 21.1% F, 0

C.)

5Y4/3

qty. 192: 3. 1% F, 0 qty.58.

O0% F,0.O

550/ -o
C) a) a)

muddy fine sand

5Y5/4

600 / -100 650 / -iO 700 / -200 .... ..... s ale e coarse, gravelly sand gradingupwardto silt 5G4/1 mud 5BG5/1

qty. 46: 2.2% F,O0

0
10

qty 50: 56.2% F, 13

750/-250
800 /-300 850 /-350

0)

~ -CD

NC-94-09

Sediment Type

Color

% Organic Content

Microf

0/ 600 50/550 100/500

I . _.._. ... . ..._ . .._. .._. . . ... . .._. . .

2.5Y6/4 slightly clayey silt

.. ... .........

2
e.
..

150/450 200 /400 250 /350 300 / 300 350 /250 400 / 200 450/ 150
...........

2.5Y5/3 2.5Y4/3

mud

5Y8/3

0g

>

slightly clayey silt

5Y4.5/2

500/100 550 / 50 600/0 650/ -50 700/-100 750 /-150 800/ -200
...... ..-.... ... ... :. .?':..' . . ..
K._. . _ .. _ . . . _ . . . .:_z.:.. _ . , .

>

fine sandy silt

sea level

2.5Y5/4

'*3

O '
,

slightly clayey silt

?

2.5Y4/4

850/-250 900 /-300

. .....

... _ ..

interbeddedmuds and silts with some fine to medium sandy layers

5BG5/1

,
;

950/ -350 1000 /-400 1050/-450

very poorly sorted, slightly muddy, fine to gravelly sand

5B4/1

NC-94-11

Sediment Type
.. . 0.

Color

% Organic Content
(= "

Microfo

~- Cl~ 00

0>

500

. ..... .. . ..
..............n

50/450100/40000-1

2.5Y5/4 5Y4/3 2.5Y4.5/4 ..level. sea. slightly clayey silt

$a)

a)
a)
0D a)

150/350-

200 300250/250300/200-

15

350 150400/100-

a)
00 a)-

2.5Y5/4

0U 0D

450 /50500 /0550 / -o

600/-100... . . .. . .. 0

... . . .. . .. .
... . <. . .

peat and peaty mud interbeddedmuds, silts, and muddy sands with some peaty horizons laminated clays, muds, and silts BEDROCK

transitional N4/ 5G5/1

qty. It: 90.9% F, 9

qty. 39: 7.7% F, 71.

650 -1500
a) 700/-200-

qty. 146. 1.4% F, 39

qty. 113: 1.8%F, 98

750 /-250800 /-300850 /-350-

5G4/1

qty. 111: 2.7% F, 90

qty. 143: 0.0% F, 10

a)

NC-94-12

Sediment Type

Color

% Organic Content
oI
ON 00

Microfo

0/390
..... ......

50/340 100/290

slightly clayey silt

2.5Y5/3

150 / 240 200 / 190 250/ 140 300/90

s'-'
>

2.5Y4/4

qty. 21: 42.9% F, 0.

qty. 48: 43.8% F, 0.

..-

slightly silty clay

._ _

2.5Y4/3 transitional

qty. 11: 90.9% F, 0.

350/40

400 /-10 450 /-60 500 /-110 550 /-160 600 /-210 650/-260 700 / -310 -

_ sea level
(

mud with silty laminae towardbottom fine sandy mud grading upwardto mud

qty. 0: 0.0% F, 0.0

5GY4.5/1 5BG5/1 5GY3/1

qty. 10: 0.0% F, 0.0

m >
'

-14: 850 ',c -+80/-60 BP

qty. 423: 4.0% F, 13

mud with interbedded clay laminae interbeddedmuds, silts, and muddy sands
*_ .:_::=:: 3
CD

750 / -360

U x,C-4

800/-410 850 /-460 900/-510 950/ -560 1000 /-610 1 1050 /-660 -

5GY4/1
5

. .'.-._.._. .. . , . ,

laminatedclays, muds, and silts

i

qty. 119: 0.0% F, 10

qty. 101: 0.0% F, 10 _ ~ ? _-I

NC-94-13

Sediment Type

Color

% Organic Content
'I

Microfos

I I I

I I

I I

0/ 480
--4

50 /430 100/380 a) 150/330 200/280 250/230

-

slightly clayey silt (_/)

2.5Y4/2

mud

2.5Y5/2

/
( )
clay
- sea level
_ _ _ _

300 / 180 (D 350 / 130 400 / 80 -

o
* 0

5BG4/1

450/30

________

500 /-20 550 / -70 -0

. Pg

_~ _~

~~~~C
~~~~C
'IL,

_ _

600 /-120 650 / -170 700 /-220 750 /-270 peat and peaty mud 5GY4/1

5~
0

800 /-320 850 /-370 900 / -420

I I I I I I

qty. 292: 0.0% F, 60.6

poorly sorted sand

,. C--~._ _

NC-94-17

Sediment Type
o C Cn C)D

Color

% Organic Content
o Ki
I I

Microfo

.
I

oC
I

oo0
I

0/ 470 50 /420 100/370 150/320 200/270 250/220

au ca

. _. .. _. .. . ..

.... ... ... ...

.. .

)
slightly clayey silt

2.5Y5/3

._. - _. ... . _. . ... a

2.5Y4/3

300/ 170 350/ 120 400 / 70 - _ ._-._ .500 /-30 450/20 550 /-80 600 /-130 sea level - -

silty fine sand gradingupwardto sandy silt -mud coarse, gravelly sand gradingupwardto silty fine sand
C-14. 1670 40 /-120 BP

2.5Y5/4

qty. 54: 0.0% F, 0.0

r0

5BG5/1

0
a)

qty. 2: 0.0% F, 100

650 /-180 -

BEDROCK

(D

0 o -a

NC-94-20

Sediment Type

Color

% Organic Content
oI I o I I I

Microfo

0 / 1000 50 / 950 100 / 900

._. ._ .. ._. .. . .._.
._._

2.5Y5/4
.. ... .

150/850 200 / 800

CA

slightly clayey silt

2.5Y5/3

250 / 750 300 / 700 350 / 650 400 / 600 450 / 550

transitional

peat and peaty mud

5B5/1 5G5/1

slightly silty clay 5BG5/1 slightly clayey silt

leve 400 cm ibv sea level cm above se i0 __ _ _, ,, , _

r4I

500 / 500

01 0

5Y6/4

550/450 600 / 400 _ _ _ _

_. . . CL

NC-94-21

Sediment Type Type

U) CD

Color

% Organic Content % Organic Content

I.C) I 0>
I

Microfo

0
I

wJ

0/520
50/470100/420150/370(\iI/)

slightly clayey silt

2.5Y4/2

qty. 16: 6.3% F, 0.

mud
-

5Y4/3

qty. 229: 10.50 F, 0

200/320
a) 0 15 a) a)

250/270 300 220

350/170400/120-

slightly silty clay

5Y5/2

qty.

24: 12.50 F, 0.

qty. 53. 3.8% F, 0.0

qty. 11: 36.4% F, 0.

450 /70500 /20550 /-30600 /-80a1) ra

5Y4/3 sea level

- - - - - - - -

qty. 6: 100.000 F, 0

clay

qty. 7: 14.300 F, 0.0

qty.21: 71.4% F, 0.

650 /-130700 /-180 750 -2305BG65/1

0 \LI C-1: 03 +10 10 B

0)
0

qty. 122: 100.000 F,

la

800/-280-

qty.5: 100.000F, 0
qty. 96: 100.000

850 -330900/-380
950 -430-

F, 0

peat and peaty mud

(-ML.

l000/-480 1050 -530 1100 -580

1150/-630-

5G5/1 for sediment; and 2.5YR2.5/1 for peat

qty. 33: 100.000

F, 0

qty. 68: 95.600F, 4

88 qty.879: 010I%F,

laminatedclays, silts, and fine sands

qty. 615: 1.30oF, 9

5G5/1

0
qty. 579: 1.4%

F, 92

1200 -680
.- CD0 O

NC-94-23

CHAPTER

SUMMARY

OBSERVATIONS

byJames Wiseman and Konstantinos Zachos

1. Including all the questionslisted in Chapter 1, pp. 8-9.

This first volume of the results of the Nikopolis Project provides the theoretical and methodological underpinnings of the research and describes the changes in the landscape of southern Epirus from the time of the earliesthuman inhabitants(more than 250,000 yearsago) up to the present. These reports constitute the frameworkinto which will be set the remaining results of the project (to be presented in volume 2), including discussions of the changing patterns of settlement and land use revealed by the diachronic survey.At the same time, the reports in this volume provide in themselves contributions both in substantive results and in the severalcritiques of methodologies. The authors have endeavored in all cases to be explicit about the aims of the research,how the investigations were carried out (what worked and what did not), the constraints on the fieldwork and analyses (however imposed), and the significance of the results. A final assessment of the significance of the project'sresults, however, must await the publication of volume 2. The broad aim of the Nikopolis Project-to explain the changing relationships between humans and the landscape they inhabited in southern Epirus-required an intensely interdisciplinaryapproach.In orderto study humans in their landscape, it was essential to determine early in the investigation just what that landscape was, and to develop parameters of as many other environmental factors as possible. The collaboration of archaeologists and geologists was vital not only to the general aim of the project, but also to many of the research questions concerned with problems or issues belonging to specific time periods.1It is important to stress this close collaborationbecause it affected almost all aspects of the project, from conception to publication. Geologic and geographic/political considerations, for example, played a greater role in determining the boundaries and size of the zone to be investigated than the likely areathat could be walked by archaeological survey teams. We were well aware from the beginning that 1,200 km2constituted too large an areafor intensive survey over even most of it, much less all of it, and such a survey was never intended. It was our aim instead to test by survey all the different kinds of environmental zones, and eventually to focus the diachronic survey on a few regions of particularinterest, as determined both by culturaland envi-

266

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

factors.As discussedin Chapter2, two of those regionswere ronmental the AyiosThomaspeninsula the lowerAcheronvalley.Still, the diaand chronicsurveywas a component the project,not the definitionof the of itself.In geologicterms,the size of the project was no impediarea project encloseda reasonably coherentareabased on the ment;the boundaries lower coursesof the two principalrivers,the Acheronand the Louros, while still providingdiverseenvironmental zones (a desideratum the of research design),rangingfrom coastlinesand marshesto inlandvalleys, The surveyzone, essentiallythe uplandplains,and ruggedmountains.2 modernnomos of Preveza,includesmost, perhapsall, of the territorium controlled a directly Nikopolisin Romantimes,3 usefulunit of analysis by for a time periodof particular interestto the project. minorconsiderA ationwas thatwe werealsoableto test the applicability remote-sensing of to a largeregionalstudy. imagery The papersin this volumeprovidesubstantial evidenceof the utility of this combinedgeologic-archaeological Tartaron's approach. reporton detailsseveral the waysthat the operation of of 2) methodology(Chapter the surveybenefitedfromthe geologiccomponents, rangingfrom selection of areasfor surveyto the interpretation certainphenomenaobof servedin the field.The theoretical methodological and discussions that in samechapter context,and placethe conductof the surveyin its historical of to providesufficientdetail,we hope,for readers assessthe significance the survey both in its relationships othercomponents the project to of and in the factorsaf(ultimately) understanding culturaland environmental of acrossthe landscape. fectingthe distribution artifacts The discovery earlyin the 1991fieldseasonby RunnelsandvanAndel of the first Lower Palaeolithictool-an Acheuleanhandaxe-found in southeast to Europein a secure geologiccontextwas the precursor a series of discoveries analyses and that madepossibletheirpresentation a coof herentevolutionof earlyhumanhabitation Epirus(Chapter3). Their in of explanation the creationand evolutionof poljesand loutsesin the dynamickarstlandscape Epirusis both an important of contribution geoto the of morphologyand a basis for understanding attractions the region were throughtime. Smallbandsof LowerPalaeolithic hunter-gatherers drawnto the lakesand pondsthat accumulated the karstdepressions, in not only as a sourceof waterbut alsofor the birdsandotheranimalsthat there.Erosionof the surrounding limestoneand alongthe assogathered ciated streamsalso made available concentrations flint. The Middle of Palaeolithic more numerous than theirpredecessors milleninhabitants, nia before,wereequallyattracted the lakes,marshes, swampsscatto and teredacrossthe landscape fromthe Lourosgorgeto the ridgesthat (now) overlookthe Ionian Sea.The specializednatureof manyof the Middle Palaeolithic sites,whichwereseasonally revisited, suggeststhatthe foraging groups-Neanderthalsor archaicHomosapiens-were followinglogisticalpatterns. The EarlyUpperPalaeolithic (EUP), the periodduringwhich modern HomosapiensreplacedNeanderthals,is represented more sparsely. RunnelsandvanAndel point out that the smallscattersof EUP artifacts atpoljeandloutsasitesindicatea patternof landuse different fromthatof

2. See the discussionin Chapter 1, pp. 2-3, and in Chapter2.

3. The principal concerns question the eastern On boundary. the northeast it terminated the territory at of
Photike;the ancient town has been identified as the archaeological site some 3 km south of the town of in Paramythia the plain of Chrysauge, but its territorial limits areuncertain (see Samsaris1988). The territoryof Nikopolis may have includedpart of the deltaicplain of the ArachthosRiver west of Ambracia,the ancient Corinthiancolony,which was evidently in ruins at the time of the founding of Nikopolis; see Doukellis 1990.

SUMMARY

OBSERVATIONS

267

earlierperiods. Anomalous for this time period is the site of Spilaion near the (present)mouth of the Acheron, perhapsthe largestlithic site in Greece. It is an open-air Aurignacian site, rare in Greece, and extraordinarilyrich in lithic artifacts,which are estimated to total some 150,000. In Chapter 4, Runnels, Karimali, and Cullen report on their spatial analysis of the material,which demonstratesthe existenceof specificareasof activitywithin the site. The success of the spatial analysis is testimony to its utility in the study of artifact-rich sites. No new sites of the Late Upper Palaeolithic, which was of short duration in Epirus, were discovered by the project,but six Mesolithic sites were added to the small number previously known in Greece. All these new sites, discussed by Runnels and van Andel, were along the new Holocene coast of the Ionian Sea. Chapter 5 is a detailed account by Jing and Rapp of the geomorphologic changes over the past 10,000 yearsof the north coast of the Ambracian Gulf and the Nikopolis peninsula. Their extensive program of geologic cores, laboratoryanalyses, and repeated examinations of the landscape has resulted in an understanding of the dramatic changes over time in these regions, which are displayed in a series of maps showing paleogeographic reconstructions of the Nikopolis peninsula and the northern coast of the Ambracian Gulf from ca. 6500 to 500 B.P. They have also determined the principalgeologic and environmentalforcesthat brought about the changes, as well as some of the cultural influences that were also at work. They demonstrated that eustatic sea-level rise from the melting of glaciers was the dominant force in determining coastal change from ca. 13,000 to ca. 6500 B.P., after which tectonic subsidence caused the sea level to continue to rise, but more slowly, until ca. 4500 B.P. when maximum marine transgression was reached; at that time, the Ambracian embayment extended north to Mts. Rokia and Stavros,leaving at most a narrowpassage at their base. Relative sea level continued to rise for the next 3,000 years, until about A.D. 500, because tectonic subsidence proceeded at a rate greater than the accumulation of sediment from rivers entering from the north. At that time, the amount of sediment from rivers and streams, and from erosion, exceeded the relative sea-level rise, and the northern shoreline moved graduallyinto the gulf, incorporating islands and creating lagoons and swamps. Jing and Rapp correlatethese dramatic changes with some of the notable archaeologicalsites in the region, especially Nikopolis and its harbor on the peninsulaand KastroRogon, the ancientBouchetion, nearthe mouth of the Louros River gorge. They show, for example, that the latter was originally a town on a small island near the coast (the gap was bridgeable), and so could have served as a regional port town itself. In late antiquity it became an inland town, and was only reconnected to the sea when the Louros River was diverted by human means in the medieval period; as a result of this diversion, the river flowed beneath the town walls in a new channel that led along the mountains and the Grammeno plain to enter the gulf near Nikopolis. Besonen, Rapp, and Jing in Chapter 6 document the equally dramatic changes during the Holocene in the lower Acheron River valley and in Phanari Bay,the ancient Glykys Limen (or "Sweet Harbor")at the present

268

JAMES

WISEMAN

AND

KONSTANTINOS

ZACHOS

to mouthof the river. Geologiccores,again,provedindispensable the paof involved andto the understanding factors reconstructions, leogeographic both of the bayandthe lowerAcheron in coastalchangeandthe evolution The bay valley. muchlarger thatexistedduringthe BronzeAge andclassifromthe IonianSea,now makesunderwith two entrances cal antiquity, in of the standable accounts ancientliterature the greatfleetsthatcouldbe there.The size of the earlierbay enhancedthe strategic accommodated on locationof the BronzeAge site identifiedas Ephyra, the hill of Xylomainportionof the bay. whichclosesthe easternsideof the earlier kastro, of Other resultsincludethe resolution the long-standing problemof the mentionedby ancientwriters,and its locationof the Acherousian lake, urban settleand over relationship timewiththeAcheronRiver the fortified ment upstream, known locally as Kastriand which may be the ancient Pandosia. as We close this chapterwith a modest disclaimer: editors,we have in on observations the reports volume1, not a set herepresented summary whichwill followin volume2. That volumewill also of finalconclusions, containreportson the studyof the cultural remains, alongwith the intesurveywith the geomorphologic grationof the resultsof the diachronic here. that investigations havebeen presented

REFERENCES

and Adam, E. 1989. A Technological of TypologicalAnalysis Upper Palaeolithic StoneIndustries Epirus, of Greece Northwestern (BAR-IS 512), Oxford. Alcock, S. E. 1991. "UrbanSurveyand the Polisof Phlius,"Hesperia 60, pp. 421-463. .1993. Graecia Capta: LandThe RomanGreece, Cambridge. of scapes Alcock, S. E.,J. F. Cherry,andJ. L. Davis. 1994. "IntensiveSurvey, AgriculturalPractice,and the ClassicalLandscapeof Greece,"in Classical AncientHistories and Greece: Modern ed. Archaeologies, I. Morris, Cambridge,pp. 137-170. Allsworth-Jones,P. 1986. TheSzeletian and theTransition from Middleto in Paleolithic Central Upper Europe, Oxford. Ammerman,A.J. 1981. "Surveysand Annual ArchaeologicalResearch," Review ofAnthropology pp. 6310, 88. . 1985. "Plow-Zone Experiments in Calabria,Italy,"JFA 12, pp. 33-40. . 1993. Rev.of Cherry,Davis, and Mantzourani1991, inJFA 20, pp.367-372. a Arsebiik,G. 1993. "Yarimburgaz, Lower PaleolithicCave Site near in Istanbul," BetweentheRiversand OvertheMountains: Archaeologica anatolicaet mesopotamica Alba Palmieridedicata, M. Frangied. pane, H. Hauptmann,M. Liverani, P. Matthiae, and M. Mellink, Rome, pp. 23-36.

. 1996. "The Cave of Yarimburgaz(the Oldest Stratified Site Yet Known in Turkey)," in Vom Halyszum Euphrat(Altertumskundedes vorderenOrients 7), ed. T. Beran,Miinster,pp. 113. EcoAscoli, P. 1964. "Preliminary logical Study on Ostracodafrom Bottom Cores of the Adriatic dellaStazione Sea,"Pubblicazioni di Napoli 33 (suppl.), Zoologica pp.213-246. Ashmore,W., and A. B. Knapp,eds. 1999. Archaeologies Landscape: of Malden, Contemporary Perspectives, Mass. a Aubouin,J. 1959. "Contribution l'etude geologique de la Grece Les septentrionale: confins de Annales l'Epireet de la Thessalie," des 10, geologiques pays helleniques pp. 1-525. . 1965. Geosynclines, Amsterdam. Bailey,G. 1992. "The Paleolithicof Klithi in Its Wider Context,"BSA 87, pp. 1-28. , ed. 1997. Klithi:Palaeolithic Settlement Quaternary and Landin 2 scapes NorthwestGreece, vols., Cambridge. Bailey,G. N., E. Adam, E. Panagopoulou, C. Perles,and K. Zachos, eds. 1999. ThePalaeoand lithicArchaeology Greece of AdjacentAreas: Proceedings the of ICOPAGConference, Ioannina, 1994 (BSA Studies 3), September London.

270

REFERENCES

Bailey,G. N., P. L. Carter,C. S. Gamble, and H. P. Higgs. 1983a. "Epirus Revisited:Seasonalityand Inter-Site Variationin the Upper Palaeolithic of North-West Greece,"in Hunterin A Gatherer Economy Prehistory: ed. G. Bailey, European Perspective, Cambridge,pp. 64-78. . 1983b. "Asprochaliko and Kastritsa: FurtherInvestigations of PalaeolithicSettlement and Economy in Epirus (North-West Greece),"PPS 49, pp. 14-42. Bailey,G., G. King, and D. Sturdy. 1993. "Active Tectonics and LandUse Strategies:A Palaeolithic Examplefrom Northwest Greece," Antiquity67, pp. 292-312. Bailey,G. N.,J. Lewin, M. G. Macklin, andJ. C. Woodward.1990. "The 'Older Fill' of the Voidomatis Valley,Northwest Greece, and Its Relationshipto the Palaeolithic Archaeologyand Glacial History of the Region,"JAS17, pp. 145150. and Bailey,G., V. Papaconstantinou, D. Sturdy.1992. "Asprochaliko and Kokkinopilos: Dating TL and Reinterpretation Middle of PalaeolithicSites in Epirus, Northwest Greece,"CAJ2, pp. 136144. Barbaroux, and B. Bousquet.1976. L., "Contribution l'etude des sols a Les mediterraneens: sols rouges du domaine helleniquede Grece occimdditerraneenne dentale,"Geologie 3, pp.237-250. Bard,E., M. Arnold, R. G. Fairbanks, and B. Hamelin. 1993. "23?Th-234U and 14CAges Obtained by Mass RadioSpectrometryon Corals," carbon pp. 191-199. 35, Bard,E., B. Hamelin, and R. G. Fairbanks. 1990. "U-Th Ages Obtained by Mass Spectrometryin Corals from Barbados: Level during Sea the Past 130,000 Years," Nature346, pp. 456-459. Bard,E., and B. Kromer.1995. "The YoungerDryas:Absolute and RadiocarbonChronology," in TheYounger Dryas:Proceedings of a Workshop theRoyalNetherlands at on Academy Arts and Sciences of 11-13 April 1994 (Verhandelingen der KoninklijkeNederlandse

Akademievan Wetenschappen, Afdeling Natuurkunde,Eerste Reeks 44), ed. S. R. Troelstra, J. E. van Hinte, and G. M. Ganssen, Amsterdam,pp. 161-166. to Barker,G. 1991. "Approaches in Roman ArchaeologicalSurvey," Landscapes: Archaeological Surveyin theMediterranean Region,ed. G. BarkerandJ. Lloyd, London, pp. 1-9. Barker, and D. Mattingly.1999. G., "GeneralEditors'Introduction: The POPULUS Project," in Past Reconstructing Population Trends Mediterranean in Europe (3000 B.C-A.D. 1800) (The Archaeology of MediterraneanLandscapes 1), ed. J. Bintliff and K. Sbonias, Oxford,pp. iii-ix. 0. Bar-Yosef, 1998. "EarlyColonizations and CulturalContinuities in the Lower Palaeolithicof WesternAsia,"in EarlyHuman in Behaviour GlobalContext: The Rise and Diversityof theLower Paleolithic ed. Record, M. D. Petragliaand R. Korisettar, London, pp. 221-279. Bennett, K., P. C. Tzedakis,and K. Willis. 1991. "Quaternary Refugiaof North EuropeanTrees," Journalof Biogeography pp. 10318, 115. Besonen, M. 1997. "The Middle and Late Holocene Geology and LandscapeEvolution of the Lower Acheron RiverValley,Epirus, Greece"(M.S. thesis, Universityof Minnesota). Bhatia, S. B. 1968. "PleistoceneOstracodes from the Upper Karewasof Kashmir,India,"Micropaleontology 14, pp. 465-483. Binford,L. R. 1964. "AConsideration of ArchaeologicalResearchDesign," AmerAnt29, pp. 425-441. Bintliff,J. 1985. "The Boeotia Survey," in Archaeological Field Surveyin BritainandAbroad, S. Macready ed. and F. H. Thompson, London, pp.196-216. . 1997. "RegionalSurvey, Demography,and the Rise of Complex Societies in the Ancient NeoAegean: Core-Periphery, Malthusian,and Other Interpretive Models,"JFA24, pp. 1-38.

REFERENCES

271 .1983. "FrogsRound the Pond: Perspectiveson CurrentArchaeological SurveyProjectsin the MediterraneanRegion,"in Keller and Rupp 1983, pp. 375-415. . 1984. "Common Sense in MediterraneanSurvey?"JFA 11, pp.117-120. . 1994. "RegionalSurveyin the Aegean:The 'New Wave'(and After),"in Kardulias1994b, pp. 91112. J. Cherry, F.,J. L. Davis, A. Demitrack, E. Mantzourani,T. F. Strasser, and L. E. Talalay.1988. "Archaeological Surveyin an Artifact-Rich Landscape:A Middle Neolithic Example from Nemea, Greece," AJA92, pp.159-176. J. Cherry, F.,J. L. Davis, and E. Mantzourani.1991. Landscape Archaeology as Long-Term Keos Northern History: in the Cycladic Earliest Islandsfrom Settlement until ModernTimes (Monumentaarchaeologica 16), Los Angeles. J. Cherry, F.,J. L. Davis, E. Mantzourani, andT. M. Whitelaw. 1991. "The SurveyMethods,"in Cherry, Davis, and Mantzourani1991, pp. 13-35. I: Chrysos,E., ed. 1987. Nicopolis Proceedings theFirstInternational of on Symposium Nicopolis(23-29 1984), Preveza. September Clark,G. A. 1994. "Migrationsas an ExplanatoryConcept in Paleolithic Archaeology,"Journal Archaeologiof calMethodand Theory pp. 3051, 343. Controls on Clews,J. 1989. "Structural Basin Evolution:Neogene to Quaternaryof the Ionian Zone, Western of Greece,"Journal the Geological Society London146, pp. 447-457. of Coussot, P., and M. Meunier.1996. "Recognition,Classification,and MechanicalDescription of Debris Reviews37, Flows,"Earth Science pp.225-252. Culiberg,M., and A. Sercelj.1996. in Events "Slovenia," Palaeoecological duringtheLast 15,000 Years: RegionalSyntheses Palaeoecological of Studies Lakesand Miresin Europe, of ed. B. E. Berglund,H.J. B. Birks, M. Ralska-Jasiewicowa, H. E. and Wright, London, pp. 687-700.

Bintliff,J., and A. Snodgrass.1985. "The Cambridge/Bradford Boeotian Expedition:The First FourYears," 12, JFA pp.123-161. . 1988a. "Off-Site Pottery Distributions:A Regional and Perspective," Interregional CurrAnthr pp. 506-513. 29, . 1988b. "Mediterranean Survey and the City," Antiquity62, pp. 57-71. Birkeland,P.W. 1984. Soilsand New Geomorphology, York. . 1999. Soilsand Geomorphology, 3rd ed., New York. Bockheim,J. G., J. G. Marshall,and H. M. Kelsey.1996. "Soil-Forming Processesand Rates on Uplifted MarineTerracesin Southwestern 72, Oregon,USA," Geoderma pp.39-62. Boero, V., and U. Schwertmann.1987. "Occurrence Transformation and of Iron and Manganesein a ColluvialTerraRossaToposequence of NorthernItaly,"Catena14, pp.519-531. . 1989. "Iron-OxideMineralogy of TerraRossa and Its Genetic Geoderma Implications," 44, pp. 319-327. Bommelj6,S., and P. K. Doom. 1987. Aetoliaand theAetolians: Towards theInterdisciplinary Studyof a Greek Region,Utrecht. Bordes,F. 1992. Lefonssurle Paleolithen ique2: Le Pal6olithique Europe, Paris. Bottema, S. 1974. "LateQuaternary VegetationHistory of Northwestern Greece"(diss. Universityof Groningen). . 1978. "The Late Glacial in the EasternMediterraneanand the Near East,"in TheEnvironmentalHistoryof theNear and MiddleEast sincetheLastIceAge, ed. W. C. Brice,New York, pp. 15-28. . 1994. "The Prehistoric Environmentof Greece:A Review of the PalynologicalRecord," in Kardulias1994b, pp. 45-68. .1995. "The YoungerDryas in the EasternMediterranean," Science Reviews14, Quaternary pp. 883-892.

Bradley,R., T. Durden, and N. Spencer. 1994. "The CreativeUse of Bias in Field Survey," Antiquity68, pp. 343-346. Buidel, 1973. "Reliefgenerationen J. der Poljebildungim Dinarischen Raum,"Geographische 32, Zeitschrift pp.134-142. Bull, W. B. 1972. "Recognitionof Alluvial Fan Deposits in the StratigraphicRecord,"in Recognition of AncientSedimentary Environments (Society of Economic Paleontologists and MineralogistsSpecial Publication16), ed. J. K. Rigby and W. K. Hamblin,Tulsa,pp. 68-83. . 1977. "The Alluvial Fan in Environment," Progress Physical 1, Geography pp. 223-270. Cavanagh,W., R. Jones, and A. Sarris. 1996. "The Phosphate and Geoin physicalSurveys," TheLaconia and Change in Survey:Continuity a Greek RuralLandscape Archaeo2: Data (BSASuppl.27), logical ed. W. Cavanaghet al., London, pp.235-261. Chappell,J., A. Omura,T. Esat, M. McCulloch,J. Pandolfi,Y. Ota, and B. Pillans. 1996. "Reconciliation of Late QuaternarySea Levels Derived from CoralTerracesat Huon Peninsula with Deep-Sea Oxygen Isotope Earthand PlanetaryScience Records," Letters141, pp. 227-236. Chappell,J., and N.J. Shackleton.1986. "OxygenIsotopes and Sea Level," Nature324, pp. 137-140. Cheddadi, R., and M. RossignolStrick. 1995. "EasternMediterranean QuaternaryPaleoclimates from Pollen and Isotope Records of Marine Cores in the Nile Cone Area,"Paleoceanography pp. 29110, 306. Chen, J. H., H. A. Curran,B. White, and G. J. Wasserburg. 1991. "Precise Chronology of the Last Interglacial Period:234U/230Th Data from Fossil Coral Reefs in the Bahamas," Bulletin Geological Society America of 103, pp. 82-97. J. Cherry, F. 1982. "APreliminary Definition of Site Distribution on Melos,"in An IslandPolity:The Archaeology Exploitationin Melos, of ed. C. Renfrewand M. Wagstaff, Cambridge,pp. 10-23.

272

REFERENCES

Cullen, H. M., P. B. deMenocal, S. Hemming, G. Hemming, F. H. Brown,T. Guilderson,and F. Sirocko. 2000. "ClimateChange and the Collapse of the Akkadian Empire:Evidence from the Deep Sea,"Geology pp. 379-382. 28, Dakaris,S. I. 1958. "AvocoxapLtxal

Dalfes, H. N., G. Kukla,and H. Weiss, eds., 1997. ThirdMillennium B.C.Climate Change and Old World (NATO ASI Collapse Series 1, Global Environmental Change 49), Berlin. a6oDarlas,A. 1989. "HOptv6xota too Axxociat," T?rXvca EXLtoxoptoo TYv ArchEph128, pp. 137-159. ?p?ovoa ?tL; OuYqp(LXqV E(p6pav xao TON?xooyaovT?sovrq{ Apxocaa; . 1995. "The EarliestOccupation of Europe:The Balkans," in OerpwOrxToc;,"Prakt, pp. 107-113. . 1960. "Avacxawpy N?XOoToo TheEarliestOccupation Europe: of Too Ax?povTO;xac OoXtoScience acvT?-ro Proceedings theEuropean of TO6TacDpoo FoundationWorkshop Tautavel at 7tXfOLOv IIapyao," Tj; Prakt,pp. 114-127. (France), 1993, ed. W. Roebroeks . 1961. "AvaoxacpYi TO andT. van Kolfschoten,Leiden, ect; Nexoo[avrTsLOv Ax?povTO;," TOO pp. 51-59. Prakt,pp. 108-119. Davidson, D. A. 1980. "Erosionin . 1962. "The Dark Palaceof Greece duringthe First and Second Hades,"Archaeology Millennia B.C.," Timescales in in 15:2, pp. 85-93. . 1963. "AvaaoxaqocpTO ed. etg Geomorphology, R. A. CullingToo N?XuoOiaVT?lOV AxpoVTO;," ford, D. A. Davidson, andJ. Lewin, Prakt,pp. 89-92. Chichester,N.Y., pp. 143-158. . 1964. "Avaoxacpy TO Davis,J. L., S. E. Alcock,J. Bennet, ?L<; Y. G. Lolos, and C. W. ShelmerToo N?xooV[acVT6oV Ax?pOVTO;," dine. 1997. "The Pylos Regional Prakt,pp. 44-53. . 1971. Cassopaia theElean and ArchaeologicalProject,PartI: Colonies Overview and the Archaeological (Ancient Greek Cities 4), Athens. 66, Survey," Hesperia pp. 391-494. . 1972. O?TrpcoTLr'a (Ancient Dean, W. E., Jr. 1974. "DeterminaGreek Cities 15), Athens. tions of Carbonateand Organic . 1975a. "Avaoxaocpyl TO E Matter in CalcareousSediments Too and SedimentaryRocks by Loss N?xooiaVT?Lrov AXE?pOVTOg," on Ignition: Comparisonwith Prakt,146-152. . 1975b. TheAntiquity Other Methods,"Journal Sediof of Necromanteion, 44, mentary Petrology pp. 242-248. Epirus:TheAcheron, Athens. Debenath, A., and H. L. Dibble. 1994. Ephyra,Pandosia,Cassope, . 1977. "N?xooatxvt-r?Lov. Handbook Paleolithic 1: of Typology E6popa. Lowerand MiddlePaleolithic Prakt, Kaooo&TrC," pp. 140-148. of . 1978. "Avacxacpq KocooaTrv Europe,Philadelphia. aoTrctH7?ipoo," Prakt, pp. 99-106. Demitrack,A. A. 1986. "The Late .1979. "Avaoxawp oTYv KaoQuaternaryHistory of the Larissa XctW7n Plain,Thessaly,Greece:Tectonic, H7?rtpoo,"Prakt, pp. 114-118. .1980. "AvaoxaxpY Kacoaqv Climatic, and Human Impact on cOytrn the Landscape" (diss. Stanford Htr?(poo," Prakt, pp. 21-32. .1981. "Avacxawp] KaoOTYMv University). Prakt,pp. 72-77. D'Errico,F.,J. Zilhao, M. Julien, ornqT," .1982. "Avao7xaoqp7 D. Baffier,andJ. Pelegrin.1998. Kaoc7trqTM," "Neanderthal Prakt,pp. 79-84. Acculturationin .1983. "Avaooxaocp) Western Europe?A Critical oTMV Review of the Evidence and Its Prakt,pp. 69-77. Ka7aon7t," . 1993. TheNekyomanteion the CurrAnthr 39, Interpretation," of Athens. Acheron, pp. 1-44. PleisDakaris,S., E. S. Higgs, and Devoto, G. 1965. "Lacustrine R. W. Hey. 1964. "The Climate, tocene in the Lower Liri Valley," romana4, pp. 291-368. Environment,and Industriesof Geologica Stone Age Greece:PartI,"PPS 30, Digerfeldt, G., S. Olsson, and P. Sandof pp. 199-244. gren. 2000. "Reconstruction

REFERENCES

273 Gaffney,V., and M. Tingle. 1985. "The Maddle Farm (Berks.)Projectand Micro-RegionalAnalysis,"in ArchField Surveyin Britain and aeological Abroad,ed. S. Macreadyand F. H. Thompson, London, pp. 67-73. Galanidou,N., P. C. Tzedakis,I. T. Lawson, and M. R. Frogley.2000. "ARevised Chronologicaland Palaeoenvironmental Framework for the KastritsaRockshelter, Northwest Greece," Antiquity74, pp. 349355. Gams, I. 1973. "Die zweiphasische quartarzeitliche Flachenbildungin Polje und Blindtalerndes nordwestlichen Dinarischen Karstes," Geo14332, graphische Zeitschrift pp. 149. .1978. "The Polje:The Problem of Its Definition,"Zeitschriftfur 22, Geomorphologie pp. 170-181. Hammond, N. G. L. 1967. Epirus: TheGeography, theAncientRemains, theHistory,and the Topography of Areas, Oxford. EpirusandAdjacent Harden,J. W. 1982. "AQuantitative Index of Soil Development from Field Descriptions:Examplesfrom a Chronosequencein Central Geoderma pp. 1-28. California," 28, Harden,J. W., and E. M. Taylor.1983. "AQuantitativeComparisonof Soil Development in Four Climatic Research Regimes,"Quaternary 20, pp. 342-359. Harden,J. W., E. M. Taylor,C. Hill, R. K. Marl, L. D. McFadden, M. C. Reheis,J. M. Sowers,and S. G. Wells. 1991. "Ratesof Soil Development from Four Soil Chronosequencesin Southern Great Britain,"Quaternary Research 35, pp. 383-399. Harris,D. R., and C. Vita-Finzi. 1968. "Kokkinopilos-A Greek Badland," Journal134, pp. 537Geographical 546. Harrison,S. P., and G. Digerfeldt. 1993. "European Lakes as Palaeohydrologicaland Palaeoclimatic Science Indicators," Quaternary Reviews 12, pp. 233-248. ChaHarrold,F. B. 1993. "Mousterian, and telperronian, EarlyAurignacian in Western Europe:Continuity or in Discontinuity?" TheHumanEvolution Source Book,ed. R. L. Ciochon

Lake-Level Changes in Lake Xinias, CentralGreece, duringthe Last 40,000 Years," Palaeogeography, and Palaeoclimatology, Palaeoecology 158, pp. 65-82. Doelle, W. H. 1977. "AMultiple Survey Strategyfor CulturalResource Management Studies,"in ConservationArchaeology:A GuideforCultural Resource ed. Studies, Management M. B. Schifferand G.J. Gumerman, New York,pp. 201-209. Doukellis, P. 1990. "EvoaOx-ruo aypoTLXW)v opLtovoJmv Tst&8ioaa -M

x 'Ap-rac,"MeA&E,yaTaroo KE.R.A.

10, pp. 269-286. Doutsos, T., N. Kontopoulos,and D. Frydas.1987. "Neotectonic Evolution of NorthwesternContinental Greece,"Geologische Rundschau pp. 433-450. 76, Douzougli, A. 1998a. "Kotv6orrra ArchDelt48, B' EzpoyyuoX;,"
(1993), pp. 282-285. .1998b. "Mta p(oPalixYl aoypotxoc c-ri;acxxr; T-ooAli3paxxtxo6 x6X7ou," ApXatoaoyia xcatTXvye,

68, pp. 74-78.

Downum, C. E., and G. B. Brown.

1998. "The Reliabilityof Surface ArtifactAssemblagesas Predictors of SubsurfaceRemains," Surface in ed. Archaeology, A. P. SullivanIII, pp. Albuquerque, 111-123. Dunnell, R. C., andW. S. Dancey. 1983. "The Siteless Survey:A Regional Scale Data Collection Stratin egy,"in Advances Archaeological Methodand Theory ed. M. B. 6, Schiffer,New York,pp. 267-287. Dunnell, R. C., andJ. F. Simek. 1995. "Artifact Size and Plowzone Processes,"JFA pp. 305-319. 22, Edwards,L. R.,J. H. Chen,T. L. Ku, and G. J. Wasserburg. 1987. "Precise Timing of the Last Interglacial Period from Mass-Spectrometric Determinationof Thorium-230 in Corals,"Science 236, pp. 15471553. Efstratiou,N. 1985. AgiosPetros, A NeolithicSite in theNorthern Sporades: AegeanRelationships during theNeolithicof the5th Millennium (BAR-IS 241), Oxford. Ellis, B. F., and A. R. Messina. 19522000. Catalogue Ostracoda, New of York.

= Etudegeologique Etudeg6ologique de nord-occidentale) (Edil'Epire(Grece tions Technip 2), 2 vols., Paris 1966. R. Fairbanks, G. 1989. "A17,000 Year Glacio-EustaticSea Level Record: Influence of Glacial Melting Rates on the YoungerDryas Event and Nature Deep Ocean Circulation," 342, pp. 637-642. Fish, S. K., and S. A. Kowalewski,eds. 1990. TheArchaeology Regions: of A CaseforFull-Coverage Survey, Washington,D.C. Fitzgerald,J. 1996. "Soil Profile Development Indices for Soils and Paleosolsof the Mount Olympus Piedmont"(M.S. thesis, Ohio University). Flemming, N. C. 1968. "Holocene EustaticMovements and Eustatic Sea Level Change in the Peloponnese,"Nature217, pp. 1031-1032. .1972. "EustaticandTectonic Factorsin RelativeVertical Displacementof the Aegean Sea: Coast,"in TheMediterranean A NaturalSedimentation Laboratory, ed. D. J. Stanley,Stroudsburg, Pa., pp. 189-201. Flemming, N. C., and C. D. Webb. 1986. "Tectonicand Eustatic Coastal Changes duringthe Last 10,000 YearsDerived from ArchaeologicalData,"in Dating Mediterranean Shorelines (Zeitschrift fur Geomorphologie Suppl. 62), ed. A. Ozer and C. Vita-Finzi, Berlin,pp. 1-29. Foley,R. 1981. "Off-Site Archaeology: An AlternativeApproachfor the Short-Sited,"in Patternof thePast: Studiesin HonourofDavid Clarke, ed. I. Hodder, G. Isaac,and N. Hammond, Cambridge,pp. 157183. Folk, R. L. 1980. Petrology Sedimenof Austin. taryRocks, Ford,D., and P.Williams. 1989. Karst and Geomorphology Hydrology, London. Fotiadis,M. 1992. "Units of Data as Deployment of Disciplinary in Codes,"in Representations ed. Archaeology, J.-C. Gardin and C. S. Peebles,Bloomington, pp. 132-148. Foucault,A., andJ.-F. Raoult. 1980. Dictionnaire geologie,Paris. de

274 andJ. G. Fleagle, Englewood Cliffs, NJ., pp. 585-603. Hawkins, R. 1998. "ComingFull Circle:Plowzone Assemblagesand of the Interpretation Fort Ancient in Settlement Structure," Surface ed. Archaeology, A. P. SullivanIII, Albuquerque, 91-109. pp. Hemans, F. P.,J. W. Myers, and J. Wiseman. 1987. Remote Sensing from a Tethered Blimpin Greece (Center for Remote SensingTechnical Paper2, Boston University), Boston. Higgs, E. S., and C. Vita-Finzi. 1966. "The Climate, Environment,and Industriesof Stone Age Greece: Part II,"PPS 32, pp. 1-29. Higgs, E. S., C. Vita-Finzi, D. A. Harris,and A. E. Fagg. 1967. "The Climate, Environment,and Industriesof Stone Age Greece: Part III,"PPS 33, pp. 1-29. Higgs, E. S., and D. Webley.1971. "Further InformationConcerning the Environmentof Palaeolithic Man in Epirus,"PPS 37, pp. 367380. Holliday,V. T. 1989. "Paleopedology in Archeology," Paleopedology: in NatureandApplication Paleosols of (CatenaSuppl. 16), ed. A. Bronger, K. Catt, andJ. A. Catt, pp. 187206. Hope Simpson, R. 1984. "The Analysis of Data from SurfaceSurveys," JFA 11, pp. 115-117. Huntley,D.J., D. I. Godfrey Smith, and M. L. W. Thewalt. 1985."Optical Dating of Sediments," Nature 313, pp.105-107.

REFERENCES

Series C, Mathematicaland Physical Sciences 126), ed. A. L. Berger,

J. Imbrie, D. Hays,G.J. Kukla, J.

and B. Saltzman,Dordrecht, pp.269-306. Innes,J. L. 1983. "DebrisFlows," in 7, Progress Physical Geography pp.469-501. and Isager,J.,ed. 2001. Foundation and Destruction, Nikopolis NorthwesternGreece: Archaeological The the Evidenceforthe CityDestructions, Foundation Nikopolis, the and of (Monographsof the Synoecism Danish Institute at Athens 3), Athens. von V., Jacobshagen, ed. 1986. Geologie Berlin. Griechenland, Jameson,M. H., C. N. Runnels,and T. H. van Andel. 1994. A Greek The Countryside: Southern Argolid to from Prehistory thePresent Day, Stanford. Jones, K. P N., I. N. McCave, and P. D. Patel 1988. "AComputerInterfacedSedigraphfor Modal Size Analysis of Fine-Grained Sediment,"Sedimentology 35, pp.163-172. Kahle,H. V., M. V. Muller, S. Miller, and G. Veis. 1993. "The Kephalonia TransformFault and the Rotation of the Apulian Platform: Evidence from Satellite Geodesy," Letters Research 20, Geophysical pp.651-654. P. of Kardulias, N. 1994a. "Paradigms the Past in Greek Archaeology," in Kardulias1994b, pp. 1-23. the , ed. 1994b. Beyond Site: Area, RegionalStudiesin theAegean Lanham. Keay,S.J., and M. Millett. 1991. "SurfaceSurveyand Site Recognition in Spain:The Ager Tarraconensis Surveyand Its Background," in Schofield 1991b, pp. 129-139. Keller,D. R., and D. W. Rupp,eds. 1983. Archaeological Surveyin the Mediterranean Area(BAR-IS 155), Oxford. Kelletat,D. 1974. Beitragezur des regionalen Ku'stenmorphologie Mittelmeerraumes (Zeitschriftfur Suppl. 19), Berlin. Geomorphologie Kelletat,D., G. Kowalczyk,B. Schr6der,and K. P.Winter. 1976. "A SynopticView on the

Huxtable, A.,J.A. J. Gowlett, J.

G. N. Bailey,P. L. Carter,and V. Papaconstantinou. 1992. "ThermoluminescenceDates and a New Analysis of the EarlyMousterian from Asprochaliko," CurrAnthr 33, pp.109-114. Imbrie,J.,J. D. Hays, D. G. Martinson, A. C. McIntyre,A. C. Mix, J.J. Morley,N. G. Pisias,W. L. Prell,N.J. Shackleton.1984. "The OrbitalTheory of Pleistocene Climate:Supportfrom a Revised Chronology of the Marine 180O in and Record," Milankovitch Climate.Understanding Response the toAstronomical Forcing(NATO ASI

REFERENCES

275 uurkunde,Eerste Reeks 44), ed. S. R. Troelstra, E. van Hinte, J. and G. M. Ganssen,Amsterdam, pp. 167-172. Ku,T. L., W. B. Bull, S. Freeman, and K. G. Knauss.1979. "Th23?-U234 Dating of Pedogenic Carbonatesin GravellyDesert Soils of Vidal Valley,SoutheasternCalifornia," Bulletin of Society America Geological 90, pp. 1063-1073. Ku,T. L., M. Ivanovich,and S. Luo. 1990. "U-SeriesDating of Last InterglacialHigh Sea Stands: BarbadosRevisited,"Quaternary Research pp. 129-147. 33, Ku,T. L., and Z. C. Liang. 1983. "The Dating of ImpureCarbonates with Decay-Series Isotopes," Seminar Alpha on Proceedings, ParticleSpectrometry Low Level and Measurement, Harwell,pp. 563-571. N. Kubilay, N., A. C. Saydam,S. Yemenicioglu,G. Kelling, S. Kapur, C. Karaman, E. Ak;a. 1997. and "SeasonalChemical and Mineralof ogical Variability Atmospheric Particlesin the Coastal Region of the Northeast Mediterranean," Catena28, pp. 313-328. Lithic Kuhn, S. L. 1995. Mousterian Technology:An Ecological Perspective, Princeton. Kuhn, S. L., G. Arsebiik,and F. C. Howell. 1996. "The Middle Pleistocene Lithic Assemblage fromYarimburgaz Cave,Turkey," Paleorient pp. 31-49. 22, Kuhn, S. L., M. C. Stiner,and E. Giileg. 1999. "InitialUpper Palaeolithic in South-CentralTurkeyand Its Regional Context:A Preliminary Report," Antiquity73, pp. 505-517. Kukal,Z. 1990. "Ratesof Geological Earth Sciences Reviews Processes," 28,pp. 7-284. Kvamme,K. L. 1996. "Investigating Chipping Debris Scatters:GIS as an AnalyticalEngine,"in New Old Methods, Problems: Geographic in Information Systems Modern ed. ArchaeologicalResearch, H. D. G. Maschner,Carbondale, Ill., pp. 38-71. N. Kyparissi-Apostolika, 1996.
"EtcYjXLo

Neotectonic Development of the PeloponnesianCoastal Regions," der Geologischen Zeitschrift Deutschen 127, Gesellschaft pp. 447-465. Keraudren, and D. Sorel. 1987. B., "TheTerracesof Corinth (Greece): A Detailed Recordof Eustatic SeaLevel Variationsduringthe Last Marine Geology 500,000 Years," 77, pp. 99-107. Kilenyi,T. I. 1969. "The Problemsof OstracodEcology in the Thames in Estuary," TheTaxonomy, and Morphology, Ecology Recent of ed. Ostracoda, J. W. Neale, Edinburgh,pp. 251-267. King, G., and G. Bailey.1985. "The Palaeoenvironment Some Archof aeologicalSites in Greece:The Influence of AccumulatedUplift in a SeismicallyActive Region,"PPS 51, pp. 273-282. King, G., D. Sturdy,andJ. Whitney. 1993. "The LandscapeGeometry and Active Tectonics of Northwest Greece,"Geological Society America of Bulletin150, pp. 137-161. Kintigh, K. W. 1990. "Commentson the Case for Full-CoverageSurvey," in Fish and Kowalewski1990, pp. 237-242. 1976. A., Kirkby, and M. J. Kirkby. "GeomorphicProcessesand the SurfaceSurveyof Archaeological Sites in Semi-Arid Areas,"in Geoand Earth Science the archaeology: Past, ed. D. A. Davidson and M. L. Shackley,London, pp. 229-253. Kirsten,E. 1987. "The Origins of the First Inhabitantsof Nikopolis,"in Chrysos 1987, pp. 91-98. E., Kotzambopoulou, E. Panagopoulou, and E. Adam. 1996. "H raXaitoXLO0txr Epeova car-v Mrt6'Xa,"
ApxatoAoyax xatcTiXveq 60, pp. 30-35.

Koumouzelis,M., J. K. Kozlowski, M. Nowak, K. Sobczyk,M. Kaczanowska,M. Pawlikowski,and A. Pazdur.1996. "Prehistoric Settlement in the KlisouraGorge, Argolid, Greece (Excavations1993, 1994),"Prehistoire 8, europeenne pp. 143-173. G. Kourtesi-Philippakis, 1996. "Nea
raX rTotLXsca atoXtotAXY; xactotxqcOTg ouT Zocxovo," ApXaioAoy6a xaL

Tixveq60, pp. 77-80.

Kozlowski,J. K. 1992. "The Balkansin the Middle and Upper Paleolithic: The Gate to Europe or a Cul-deSac?" PPS 58, pp. 1-20. . 1999. "The Evolution of the BalkanAurignacian," Dorothy in Garrod theProgress thePalaeoand of lithic:Studiesin thePrehistoricArchaeology theNear East and Europe, of ed. W. Davies and R. Charles, Oxford, pp. 97-117. Kraft,J.C. 1985. "MarineEnvironments:Paleogeographic Reconstructions in the LittoralRegion,"in Sediments Context, in Archaeological ed. J. K. Stein andW. R. Farrand, Orono, Maine, pp. 111-125. Kraft, C., and S. E. Aschenbrenner. J. 1977. "Paleogeographic Reconstructions in the Methoni Embaymentin Greece,"JFA pp. 19-44. 4, Kraft,J. C., S. E. Aschenbrenner, and G. RappJr. 1977. "Paleogeographic Reconstructionsof CoastalAegean 195, ArchaeologicalSites,"Science pp. 941-947. Kraft,J. C., and M. J. Chrzastowski. 1985. "CoastalStratigraphic in Sequences," Coastal Sedimentary ed. Environments, R. A. Davis Jr., 2nd ed., New York,pp. 625-663. Kraft,J.C., I. Kayan,and S. E. Aschenbrenner.1985. "Geological Studies of Coastal Plains Applied to ArchaeologicalSettings,"in ed. Archaeological Geology, G. Rapp Jr.,andJ. A. Gifford, New Haven, pp. 57-84. Kraft, C., G. RappJr.,and S. E. J. Aschenbrenner. 1980. "LateHolocene Paleogeographic Reconstructions in the Area of the Bay of Navarino:SandyPylos,"JAS7, pp. 187-210. Kraft,J.C., G. RappJr.,G.J. Szemler, C. Tziavos,and E. W. Kase. 1987. "The Pass at Thermopylae,Greece," JFA 14, pp. 181-198. Kromer,B., B. Becker,P.Trimborn,and M. Spurk.1995. "Tree-ringBased 14C Calibrationand StableIsotope Series at the YD/BP Boundary," in TheYounger Dryas:Proceedings a of at Workshop theRoyalNetherlands on Academy Arts and Sciences 11of 13 April 1994 (Verhandelingen der KoninklijkeNederlandseAkademie van Wetenschappen,Afdeling Nat-

OT6c-zpa;:

OL7CaXXaLO-

XtL0LX?; srcLtX)cL;g," ApXatoAoy(a

xat Td?ve;60, pp. 37-41.

276 Labeyrie,L. D.,J.-C. Duplessy,and P. L. Blanc. 1987. "Variations in Mode of FormationandTemperature of Oceanic Deep Watersover the Past 125,000 Years," Nature 327, pp. 477-482. Laj, C., A. Mazaud, andJ. C. Duplessy. 1996. "GeomagneticIntensity and 14C Abundancein the Atmosphere and Ocean duringthe Past 50 kyr," Letters Research 23, Geophysical pp. 2045-2048. Lambeck,K. 1995. "LatePleistocene and Holocene Sea-Level Change in Greece and SouthernTurkey:A Separationof Eustatic,Isostatic,and Tectonic Contributions," Geophysical JournalInternational 122, pp. 10221044. . 1996. "Sea-LevelChange and Shore-Line Evolution:A General of Framework Modelling and Its Application to Aegean Greece since Upper PalaeolithicTime,"Antiquity 70, pp.588-611. Lambeck,K., and M. Nakada.1992. "Constraints the Age and on Duration of the Last Interglacial Period and on Sea-Level Variations,"Nature357, pp. 125-128. Lang, A., and G. A. Wagner.1996. StimulatedLuminescence "Infrared Dating of Archaeosediments," 38, Archaeometry pp. 129-141. Leake,W. M. 1835. Travelsin Northern 4 Greece, vols., London. Le Pichon, X., N. Chamot-Rooke, S. Lallemant,R. Noomen, and G. Veis. 1995. "GeodeticDetermination of the Kinematicsof Central Greecewith Respect to Europe: Implicationsfor EasternMediterranean Tectonics,"Journal Geophysiof calResearch 100, pp. 12675-12690. Lloyd,J. A., and G. Barker.1981. "RuralSettlement in Roman Molise: Problemsof Archaeological in and Survey," Archaeology Italian Prehistoric, Roman,and MedSociety: ieval Studies(BAR-IS 102), ed. G. Barkerand R. Hodges, Oxford, pp.375-416. Machette, M. N. 1985. "CalcicSoils of the SouthwesternUnited States,"in Soilsand Quaternary Geology the of Southwestern UnitedStates(Geological Society of America Special Paper203), ed. D. L. Weide, Boulder,pp. 1-21.

REFERENCES

MacLeod, D. A. 1980. "The Origin of Red MediterraneanSoils in Epirus, 269, of Greece,"Journal Soil Science pp.573-577. MacLeod, D. A., and C. Vita-Finzi. 1982. "Environment and Provenance in the Development of Recent Alluvial Deposits in Epirus,Greece,"Earth Surface Processes Landforms and 7, pp. 29-43. Magaritz,M., A. Kaufman,and D. H. Yaalon.1981. "Calcium CarbonateNodules in Soils:
180/160 and 13C/12CRatios and 14CContents,"Geoderma 25,

pp.157-172. Martinson,D., N. G. Pisias,J. D. Hays,J. Imbrie,T. C. Moore Jr., and N.J. Shackleton.1987. "Age Dating and the OrbitalTheory of the Ice Ages: Development of a High-Resolution 0-300,000 Year Chronostratigraphy," Quaternary Research pp. 1-29. 27, McFadden, L. D.,J. B. Ritter,and S. G. Wells. 1989. "Use of MultiparameterRelative-AgeMethods for Age Estimation and Correlation of Alluvial Surfaceson a Desert Piedmont, EasternMojave Desert, Research California," 32, Quaternary pp.291-293. McFadden, L. D., and R. J. Weldon. 1987. "Ratesand Processesof Soil Development in Quaternary Terracesin Cajon Pass,"Geological Bulletin98, Society America of pp.280-293. Mee, C., and H. Forbes.1997. "Survey Methodology,"in A Place:TheLandRoughand Rocky and scape Settlement Historyof the MethanaPeninsula, ed. Greece, C. Mee and H. Forbes,Liverpool, pp. 33-41. and Mellars,P. 1992. "Archaeology the Population-Dispersal Hypothesis of Modern Human Origins in Transactions Europe," Philosophical of theRoyalSociety London337, of pp.225-234. .1996. TheNeanderthalLegacy: An Archaeological Perspectivefrom Western Europe,Princeton. Middleton, G. V. 1973. "Johannes Walther'sLaw of the Correlation of Facies,"Geological Society of America Bulletin84, pp. 979-988.

REFERENCES

277 Sequencesin Post-Orogenic Basins, Western Greece,"Sedimentary 55, Geology pp. 283-294. Piper,D. J. W., A. G. Panagos,and N. Kontopoulos.1982. "Some Observationson SurficialSediments and Physical Oceanography of the Gulf of Amvrakia," Thalasso5, pp. 63-80. graphica Pirazzoli,P. A., S. C. Stiros,J. Laborel, F. Laborel-Deguen,M. Arnold, S. Papageorgiou, C. Morhange. and 1994. "Late-Holocene Shoreline Changes Relatedto Palaeoseismic Events in the Ionian Islands, Greece,"TheHolocene pp. 3974, 405. Plog, F. 1990. "SomeThoughts on in Full-CoverageSurveys," Fish and Kowalewski1990, pp. 243-248. Plog, S., F. Plog, and W. Wait. 1978. "DecisionMaking in Modern in in Surveys," Advances ArchaeologicalMethodand Theory1, ed. M. B. Schiffer,New York,pp. 383-421. Pope, K. O., C. N. Runnels,and T. L. Ku. 1984. "DatingMiddle PalaeolithicRed Beds in Southern Greece,"Nature312, pp. 264-266. Pope, K. O., andT. H. van Andel. 1984. "LateQuaternary Alluviationand Soil Formationin the Southern Argolid:Its History,Causes, and ArchaeologicalImplications,"JAS 11, pp. 281-306. Poulos, S. E., V. Lykousis,and M. B. Collins. 1995. "LateQuaternaryEvolution of Amvrakikos Gulf, Western Greece,"GeoMarineLetters15, pp. 9-16. Purcell,N. 1987. "The Nicopolitan Synoecismand Roman Urban in Policy," Chrysos 1987, pp. 71-90. Puri, H. S., G. Bonaduce,and A. M. Gervasio.1969. "Distributionof Ostracodain the Mediterranean," in TheTaxonomy, and Morphology, ed. Ecology RecentOstracoda, J. W. of Neale, Edinburgh,pp. 356-411. Puri, H. S., G. Bonaduce,andJ. Malloy. 1964. "Ecologyof the Gulf of dellaStazione Naples,"Pubblicazioni di Zoologica Napoli 33 (suppl.), pp. 87-199. Pye, K. 1992. "AeolianDust Transport and Deposition over Crete and AdjacentPartsof the Mediterranean Sea,"Earth Surface Processes and Landforms pp. 271-288. 17,

Millett, M. 1991. "Pottery: Population in or SupplyPatterns?" Roman Landscapes: Archaeological Surveyin theMediterranean Region,ed. G. BarkerandJ. Lloyd, London, pp. 18-26. Mirabella,A., and M. S. Carnicelli. 1992. "IronOxide Mineralogyin Red and Brown Soils Developed on CalcareousRocks in CentralItaly," Geoderma pp. 93-110. 55, Miracle,P.T. 1995. "Broad-Spectrum AdaptationsRe-Examined:HunterGathererResponsesto Late Glacial EnvironmentChanges in the EasternAdriatic"(diss. University of Michigan). Monopolis, D., and A. Bruneton.1982. "IonianSea (Western Greece):Its Structural Outline Deduced from Drilling and GeophysicalData," 83, Tectonophysics pp. 227-242. Moore, M. 2000. "Surveying Epirote Pottery:Ceramics,Cuisine, and Social History in SouthernEpirus, Greece,300 B.C.-A.D. 500" (diss. Boston University). Morrison,R. B. 1976. "Quaternary Soil Stratigraphy: Concepts, in Methods, and Problems," Soils:Papers the3rd of Quaternary on Research, Symposium Quaternary ed. W. C. Mahaney,Norwich, pp. 77-108. P. Murray, M., and P. N. Kardulias. 1986. "AModern-Site Surveyin the SouthernArgolid,"JFA13, pp. 21-41. W. Murray, M., and Ph. M. Petsas. 1989. OctaviansCampsite Memorial theActian War(TAPS79), for Philadelphia. Nakada,M., and K. Lambeck.1988. "The Melting History of the Late Pleistocene AntarcticIce Sheet," Nature333, pp. 36-40. Neale, J. W. 1964. "SomeFactors Influencingthe Distributionof Recent British Ostracoda," PubblicazionidellaStazioneZoologica di Napoli 33 (suppl.),pp. 247-307. Nicod, J. 1992. "Recherches sur l'6volutiondu karstdu Massif de Montrieux (Basse Provence)," in Geomorphology Geoecology: and Karst(Zeitschriftfur Geomorphologie Suppl. 85), ed.J. Nicod, K. H. Pfeffer,and M. Sweeting, Berlin, pp. 39-57.

Niemi, T. M. 1990. "Paleoenvironmental History of SubmergedRuins on the Northern Euboean Gulf Coastal Plain, CentralGreece," 5, Geoarchaeology pp. 323-347. Odell, G. H., and F. Cowan. 1987. "EstimatingTillage Effects on AmerAnt52, Artifact Distributions," pp.456-484. Olszewski, D. I., and H. L. Dibble. 1994. "The ZagrosAurignacian," CurrAnthr pp. 68-75. 35, Papagianni,D. 1999. "MiddlePalaeolithic OccupationandTechnology in Northwest Greece:Evidence from Open-Air Sites"(diss. University of Cambridge). .2000. MiddlePalaeolithic and in Occupation Technology Greece Northwestern (BAR-IS 882), Oxford. Papazachos,B. C., and P. E. Comminakis. 1971. "Geophysicaland Tectonic Featuresof the Aegean Research Arc," Journalof Geophysical 76, pp.8517-8533. J. Parsons, R. 1990. "CriticalReflections on a Decade of Full-Coverage Regional Surveyin the Valleyof Mexico,"in Fish and Kowalewski 1990, pp. 7-31. Perles,C. 1987. Les industries lithiques tailleesdeFranchthi Grece) (Argolide, 1: Presentation et generale industries (Franchthi 3), pal6olithiques Bloomington. .1990. Les industries lithiques tailleesde Franchthi Grece) (Argolide, 2: Les industries mdsolithique du du et initial (Franchthi 5), neolithique Bloomington. Petruso,K. M., B. B. Ellwood, F. B. Harrold,and M. Korkuti.1994. "Radiocarbon Archaeomagand netic Dates from KonispolCave, Albania," Antiquity68, pp. 335-339. Pfeiffer,D. 1963. "Die geschichtliche Entwicklungder Anschaungentiber das Karstgrundwasser," Geologisches 57, Jahrbuch pp. 1-111. Philippson,A., and E. Kirsten.1956. Die griechischen 2 Landschaften, vols., Frankfurt. and Phleger,F. B. 1960. Ecology Distributionof RecentForaminifera, Baltimore. Piper,D. J. W., N. Kontopoulos,and A. G. Panagos.1988. "Deltaic Sedimentationand Stratigraphic

278
Pyle, D. M., T. H. van Andel, P. Paschos, and P.van den Boogaard. 1998. "AnExceptionallyThick Middle PleistoceneTephraLayer from Epirus,Greece,"Quaternary Research pp. 280-286. 49, ArchaeoRapp,G., Jr. 1986. "Assessing logical Evidence for Seismic 4, Catastrophes," Geoarchaeology pp.365-379. Rapp,G.,Jr., and Z. Jing. 1994. "Reconstruction the Changing of Holocene Coastal Landscapein the Nikopolis Region, Northwestern Greece"(paper, Washington,D.C. 1993), abstractin AJA98, p. 316. Rapp,G.,Jr., andJ. C. Kraft.1994. "Holocene Coastal Change in in Greece and Aegean Turkey," Kardulias1994b, pp. 23-35. an Rathjens,C. 1960. "Beobachtungen hochgelegenen Poljen im stidlichen dinarischenKarst(ein Beitragzur Frageder Entstehung and Datierung der Poljen),"Zeitschriftfur 4, Geomorphologiepp. 141-151. Redman,C. 1973. "MultistageFieldwork and AnalyticalTechniques," AmerAnt39, pp. 61-79. Reisch, L. 1980. "Pleistozanund Urgeschichteder Peloponnes" Friedrich(habilitationsschrift, AlexanderUniversitat). Retallack,G. J. 1988. "FieldRecogniin and tion of Paleosols," Paleosols Time: through Geologic Weathering (GeoandApplications Principles logical Society of America Special Paper216), ed. J. Reinhardtand W. R. Sigleo, Boulder,pp. 1-20. . 1990. Soilsof thePast:.An Introduction Paleopedology, to London. Rick,J. W. 1976. "DownslopeMovement and ArchaeologicalIntrasite AmerAnt41, SpatialAnalysis," pp. 133-144. Roberts,B. K. 1987. "Landscape in and Archaeology," Landscape Culture: andArchaeoGeographical ed. logicalPerspectives, J. M. Wagstaff,Oxford, pp. 77-95. Roberts,N., and H. E. Wright Jr. 1993. Lake-Level, and "Vegetational, Climatic History of the Near East and SouthwestAsia,"in Global sincetheLast Glacial Climates Maximum,ed. H. E. Wright Jr.,

REFERENCES

T J. E. Kutzbach, Webb III, W. F. Ruddiman,F. A. StreetPerrott,and P.J. Bartlein,Minneapolis,pp. 194-220. Roebroeks, W., andT. van Kolfschoten. 1995. "The EarliestOccupationof of Europe:A Reappraisal Artefactual and ChronologicalEvidence," in TheEarliestOccupation Europe: of Science of Proceedings theEuropean FoundationWorkshop Tautavel at 1993, ed. W. Roebroeks (France), andT. van Kolfschoten,Leiden, pp.297-315. Rolland,N. 1986. "RecentFindings from La Micoque and Other Sites in South-Westernand Mediterranean France: Their Bearingon the Problemand Middle 'Tayacian' in PalaeolithicEmergence," Stone Studiesin Memoryof Age Prehistory: ed. Charles McBurney, G. N. Bailey and P. Callow,Cambridge,pp. 121151. Rossignol,J. 1992. "Concepts, Methods, andTheory Building,"in Rossignol andWandsnider1992, pp. 3-16. Rossignol,J., and L. Wandsnider,eds. 1992. Space, Time,andArchaeological New York. Landscapes, Rossignol-Strick,M. 1983. "African Monsoons: An Immediate Climate Responseto OrbitalInsolation," Nature304, pp. 46-49. Runnels,C. 1981. "On the Destruction of ArchaeologicalSites,"JFA8, pp. 91-93. . 1988. "APrehistoricSurvey of Thessaly:New Light on the Greek Paleolithic,"JFA 15, pp.277-290. . 1994. "Pleistoceneand Early Holocene ArchaeologicalSites in SouthernEpirus,Greece"(paper, Washington,D.C. 1993), abstractin AJA98, pp. 315-316. . 1995. "The Stone Age of Greece from the Palaeolithicto the Advent of the Neolithic,"AJA99, pp.699-728. . 1996. "The Palaeolithicand Mesolithic Remains," Wells and in Runnels 1996, pp. 23-35. Runnels,C., andT. H. van Andel. 1993a. "The Lower and Middle Palaeolithicof Thessaly (Greece)," JFA 20, pp. 299-318.

REFERENCES

279 0. Spaargaren, C. 1979. Weathering and SoilFormation a Limestone in AreanearPastena, Italy (Publicaties van het Fysisch Geografischen BodemkundigLaboratorium, Universityof Amsterdam30), Amsterdam. Spate,A. P.,J. N. Jennings,D. I. Smith, and M. A. Greenaway. 1985. "The Micro-Erosion Meter:Use and Earth Surface Processes Limitations," andLandforms pp. 427-440. 10, Stein, C., and B. Cullen. 1994. "SatelliteImageryand Archaeology: A Case Study from Nikopolis" (paper, Washington,D.C. 1993), abstractin AJA98, p. 316. Stiner,M. C., G. Arsebiik,and F. C. Howell. 1996. "CaveBears and PaleolithicArtifactsin Cave,Turkey: Yarimburgaz Dissecting a Palimpsest," 11, Geoarchaeology pp. 279-327. Stiner,M. C., N. D. Munro, T. A. Surovell,E. Tchernov,and 0. Bar-Yosef.1999. "Paleolithic PopulationGrowth Pulses Evidenced by Small Animal Exploitation,"Science 283, pp. 190-194. Stirling,C. H., T. M. Esat, M. T. McCulloch, and K. Lambeck.1995. "High-PrecisionU-Series Dating of Corals fromWestern Australiaand Implicationsfor the Timing and Duration of the Last Interglacial," Earth and PlanetaryScience Letters 135, pp. 115-130. Stiros, S. C., M. Arnold, P. A. Pirazzoli,J. Laborel,and F. LaborelDeguen. 1994. "The 1953 Earthquakein Cephalonia(Western Hellenic Arc): CoastalUplift and Halotectonic Faulting,"Geophysical JournalInternational 117, pp. 837849. Stoddart,S. K. F., and N. Whitehead. 1991. "Cleaningthe Iguvine Stables: Site and Off-Site Analysis from a CentralMediterranean Perspective," in Schofield 1991b, pp. 141-148. Sturdy,D. A., and D. P.Webley.1988. "Palaeolithic Geography:or Where arethe Deer?"WorldArch 19, pp. 262-280. Sturdy,D. A., D. P.Webley,and G. Bailey.1997. "The Palaeolithic in Geographyof Epirus," Bailey 1997, pp. 2:587-614.

. 1993b. "AHandaxe from Kokkinopilos,Epirus,and Its Implicationsfor the Paleolithicof Greece,"JFA pp. 191-203. 20, Russell,R. J. 1954. "Alluvial MorpholAnnalsof ogy of Anatolian Rivers," theAssociation AmericanGeograof phers44, pp. 363-391. . 1967. River Plainsand Sea Coasts, Berkeleyand Los Angeles. Rutter,J. B. 1993. "The Prepalatial Bronze Age of the Southernand CentralGreek Mainland," AJA97, pp. 745-797. aTotLxia Samsaris,D. 1988. "HpLtOCLXYJ
rt tI)TcLXYlq," IzToptxoye?oypacxpoxd

2,pp.23-31. Sarris,A., and R. E. Jones. 2000. and RelatedTech"Geophysical niques Applied to Archaeological A Surveyin the Mediterranaean: 13, Review,"JMA pp. 3-75.
Sars, G. 0. 1928. An Account of the

Crustacea Norway9: Ostracoda, of Bergen. Schiffer,M. B. 1987. Formation Processes theArchaeological Record, of Albuquerque. Schiffer,M. B., A. P. Sullivan,andT. C. Klinger.1978. "The Design of ArchaeologicalSurveys,"WorldArch 10, pp. 1-28. Schneider,H. E. 1977. "Problemsof Terrestrial Red Beds in the Neogene and Quaternaryof Greece,"
in Proceedingsof the VIth Colloquium

on the Geology theAegeanRegion, of ed. G. Kallergis,Athens, pp. 477493. Schofield,A. J. 1991a. "Interpreting Artefact Scatters: Introduction," An in Schofield 1991b, pp. 3-8.
, ed. 1991b. InterpretingArtefact

Scatters: Contributions Ploughzone to (Oxbow Monograph4), Archaeology Oxford. Schroder,B. 1986. "Das postorogene Kanozoikumin Griechenland/ von Agais,"in Geologie Griechenland, ed. V.Jacobshagen, Berlin, pp. 208240. Schroder, and D. Kelletat.1976. B., Conclusions from "Geodynamical VerticalDisplacement of Quaternary Shorelinesin the Peloponnesos, Greece," Neues Jahrbuchfur Geologie und Paltontologie, Monatshefte,

pp. 174-186.

Schwertmann,U. 1971. "Transformation of Hematite to Goethite in Soils,"Nature232, pp. 624-625. Sever,T., andJ. Wiseman. 1985. on Conference RemoteSensing: PotentialfortheFuture,NASA, Earth ResourcesLaboratory. Shackleton,N.J. 1987. "OxygenIsotopes, Ice Volume,and Sea Level," Science Reviews6, Quaternary pp.183-190. of Shott, M. J. 1995. "Reliability ArchaeologicalRecordson Cultivated Surfaces: Michigan Case A 22, Study,"JFA pp. 475-490. Siani, G., M. Paterne,M. Arnold, E. Bard,B. Metivier,N. Tisnerat, and F. Bassinot.2000. "Radiocarbon ReservoirAges in the Mediterranean Sea and Black Sea,"Radiocarbon 42, pp.271-280. Singer,R., B. G. Gladfelter,andJ. J. Wymer. 1993. TheLowerPaleolithic Siteat Hoxne,England,Chicago. Smit, A., andT. A. Wijmstra.1970. of "Application TransmissionElectron MicroscopeAnalysis to the Reconstructionof FormerVegetation," Pollenet spores pp. 615-621. 13, Smith, G. W., R. D. Nance, and A. N. Genes. 1997. "Quaternary Geological History of Mount Olympus, Greece,"Geological Society America of Bulletin109, pp. 809-824. Soil SurveyStaff. 1975. Soil Taxonomy (U.S. Department of Agriculture Handbook 436), Washington,D.C. Sordinas,A. 1968. "The Prehistoryof the Ionian Islands:The Flints and (diss. HarvardUniversity). Pottery" . 1969. "Investigations the of Prehistoryof Corfu during 19641966,"BalkSt10, pp. 393-424. . 1970. StoneImplements from Northwestern MemCorfu,Greece, phis. . 1983. "Quaternary Shorelines in the Region of Corfu and Adjacent Islets," in Quaternary Coastlines and MarineArchaeology: Towards Prethe and Continenhistoryof LandBridges tal Shelves, P.M. Masters and N. ed. C. Flemming, London, pp. 335-344. Sorel, D. 1989. "L'evolution structurale de la Gr&ece nord-occidentaledepuis le Miocene dans le cadregeodynamique de l'arcEgeen"(diss. Universite de Paris-SudXI).

280

REFERENCES

Sullivan,A. P., III, ed. 1998. Surface Archaeology, Albuquerque. Sumner,W. M. 1990. "Full-Coverage RegionalArchaeologicalSurveyin the Near East:An Examplefrom in Iran," Fish and Kowalewski1990, pp. 87-115. Sweeting, M. M., ed. 1985. Karst Pa. Stroudsburg, Geomorphology, Sweeting,M. M. 1993. "Reflectionson the Development of Karstgeomorphology in Europe and a Comparison with Its Development in GeomorChina,"in Klimagenetische Geomorphphologie(Zeitschriftfuir ologieSuppl. 93), ed. R. W. Fairbridge and K.-H. Pfeffer,Berlin, pp. 127-136. T. SettleTartaron, 1994. "Prehistoric ment in SouthernEpirus:Preliminary Resultsfrom Survey" (paper, Washington,D.C. 1993), abstractin AJA98, p. 316. . 1996. "BronzeAge Settlement and Subsistencein Southwestern Epirus,Greece"(diss. Boston University). . 2001. "GlykysLimin: A MycenaeanPort of Tradein in SouthernEpirus?" Prehistory and History: Ethnicity,Class,and ed. PoliticalEconomy, D. W. Tandy, Montreal,pp. 1-40. Tartaron, and C. Runnels. 1992. T., "The Discovery of a Palaeolithic Handaxe at Kokkinopilosin Epirus,"Context9:3-4, pp. 1, 4-7. Tartaron, C. Runnels,and E. KariT., mali. 1999. "Prolegomena the to Study of Bronze Age Flaked Stone in SouthernEpirus,"in Meletemata:Studiesin Aegean Archaeology Presented Malcolm Wiener He to H. as EntersHis 65th Year (Aegaeum III 20), ed. P. P. Betancourt,V. Karageorghis, R. Laffineur,and W.-D. Niemeier, Liege, pp. 819-825. Tartaron, and K. Zachos. 1999. T., "The Mycenaeansand Epirus," in H wrcptipeta ro .uVox7vdixo6 A' xoayovo: AelOvgAi?tcoaTcLovtxoSvoyrr6oto Aai(a, 25-29 1994, Lamia, e7rerftSppioV pp. 57-76. Tassos, S. T. 1975. "Holocene Sediments and Some Oceanographic Parameters KalloniBay,Lesvos of Island,Greece"(M.S. thesis, Universityof Minnesota).

Thomas, D. H. 1975. "Non-Site Samplingin Archaeology:Up the Creekwithout a Site?"in Sampling in Archaeology, J. W. Mueller, ed. Tucson,pp. 61-81. Trigger,B. G. 1989. A HistoryofArchaeological Cambridge. Thought, Turner,C., and M. Sanchez-Gofii. 1997. "LateGlacial Landscapeand in Vegetationin Epirus," Bailey 1997, pp. 2:559-585. Tree Tzedakis,P. C. 1993. "Long-Term Populationsin Northwest Greece throughMultiple Quaternary Climate Cycles,"Nature364, pp. 437-440. . 1994. "VegetationChange throughGlacial-Interglacial Cycles: A Long Pollen Sequence PerspecTransactions tive,"Philosophical of theRoyalSociety LondonB 345, of pp. 403-432. Tziavos,C. C. 1977. "Sedimentology, of Ecology, and Paleogeography the SperchiosValleyand Maliakos Gulf, Greece"(M.S. thesis, Universityof Delaware). . 1997. "Paleogeographic Evolution of the AmvrakikosGulf, Western Greece,"in TheEngineerMonuing Geology AncientWorks, of Sites:Proceedments,and Historical I, ingsof an International Symposium ed. P. G. Marinos and G. C. Koukis, Athens, pp. 425-430. Underhill,J. 1989. "LateCenozoic Deformation of the Hellenide Western Greece," Foreland, Bulletin Society America Geological of 101, pp. 613-634. van Andel, T. H. 1989. "LateQuaternary Sea-Level Changes and Archaeology," Antiquity63, pp. 733-745. . 1998a. "Paleosols,Red Beds, and the Old Stone Age in Greece," 13, Geoarchaeology pp. 361-390. . 1998b. "Middleand Upper PalaeolithicEnvironmentsand the Calibrationof 14CDates beyond
10,000 B.P.,"Antiquity 72, pp. 26-

33. van Andel, T. H., and C. Runnels. 1987. BeyondtheAcropolis:A Rural Greek Past, Stanford. van Andel, T. H., C. N. Runnels, and K. 0. Pope. 1986. "FiveThousand Yearsof Land Use and Abuse in the SouthernArgolid, Greece," 55, Hesperia pp. 103-128.

REFERENCES

28I

van Andel, T. H., andJ. C. Shackleton. 1982. "LatePaleolithicand Mesolithic Coastlines of Greece and the Aegean,"JFA9, pp. 445-454. van Andel, T. H., and P. C. Tzedakis. 1996. "Palaeolithic Landscapesof Europe and Environs,150,00025,000 YearsAgo," Quaternary
ScienceReviews 15, pp. 481-500.

. 1998. "Priority and Opportunity: Reconstructingthe European Middle PalaeolithicClimate and in in Landscape," Science Archaeology:AnAgendafortheFuture,ed. J. Bayley,London, pp. 37-46. van Andel, T. H., E. Zangger,and A. Demitrack.1990. "LandUse and Soil Erosion in Prehistoricand Historical Greece,"JFA17, pp. 379396. van Andel, T. H., E. Zangger,and C. Perissoratis. 1990. "Quaternary Transgressive/Regressive Cycles in the Gulf of Argos, Greece,"Quater34, naryResearch pp. 317-329. van Zeist, W., and S. Bottema. 1982. "Vegetational History of the EasternMediterranean and the Near East duringthe Last 20,000 in PalaeoYears," Palaeoclimates, and environments, Human Communitiesin theEasternMediterranean (BAR-IS Regionin LaterPrehistory 133), ed. J. L. Bintliff and W. van Zeist, Oxford, pp. 277-321. Villas, C. A. 1983. "The Holocene Evolution of the Acheloos River Delta, NorthwesternGreece: Associated Environments,Geomorphology,and Microfossils" (M.S. thesis, Universityof Delaware). Vita-Finzi, C. 1978. Archaeological Sites in TheirSetting,London. Vreeken,W. J. 1975. "Principal Kinds of ChronosequencesandTheir Significancein Soil History," Journalof Soil Science pp. 37826, 394. Wagner,C. W. 1957. Surles Ostracodes quaternaire du recent Paysdes Bas et leurutilisationdansl'etude des The holocenes, geologique depots Hague. Wandsnider,L., and E. L. Camilli. 1992. "The Characterof Surface ArchaeologicalDeposits and Its Influenceon SurveyAccuracy," JFA 19, pp. 169-188.

Waters,D. W. 1994. "TheTectonic Evolution of Epirus,Northwest Greece"(diss. Universityof Cambridge). Weiss, H. 2000. "Beyondthe Younger Dryas:Collapse as Adaptationto Abrupt Climate Change in Ancient West Asia and the EasternMediin Disasterranean," Environmental terand theArchaeology Human of ed. Response, G. Bawden and R. Reycraft,Albuquerque, 75-98. pp. Weiss, H., M. A. Courty,W. Wetterstrom,F. Guichard,L. Senior, R. Meadow, and A. Curnow.1993. "The Genesis and Collapse of Third Millennium North Mesopotamian Civilization,"Science 261, pp. 9951004. Wells, B. 1994. "APrehistoricEnvironmental Catastrophe: The Case of Berbatiand Limnes,"in Archaeology in thePeloponnese: Excavations New and Research (Oxbow Monograph 48), ed. K. A. Sheedy,Oxford, pp. 65-76. Wells, B., and C. Runnels,eds. 1996. TheBerbati-Limnes Archaeological Survey,1988-1990, Jonsered. Wells, B., C. Runnels,and E. Zangger. 1990. "The Berbati-Limnes ArchaeologicalSurvey:The 1988 Season,"OpAth 18, pp. 207-238. Weymouth,J. W. 1993. "Preliminary Report of GeophysicalSurveys, Nikopolis Project1993,"reporton file at the Departmentof Archaeology, Boston University. Weymouth,J. W., and R. Huggins. 1985. "GeophysicalSurveyingand ArchaeologicalSites,"in Archaeological Geology, G. RappJr.,and ed. J. A. Gifford, New Haven, pp. 191235. Whitelaw,T. M. 1991. "Investigations at the Neolithic Sites of Kephala and Paoura," Cherry,Davis, and in Mantzourani1991, pp. 199-215. of Wijmstra,T. A. 1969. "Palynology the First 30 Metres of a 120m-Deep Section in Northern Greece (MaceActaBotanica Neerlandica donia)," 18, pp.511-557. Wijmstra,T. A., and A. Smit. 1976. "Palynologyof the Middle Part of the 120m-Deep Section in Northern Greece (Macedonia)," ActaBotanica Neerlandica 25, pp.297-312.

Wijmstra,T. A., R. Young,and H. J. L. Witte. 1990. "AnEvaluation of the Climatic Conditions duringthe Late Quaternaryin Northern Greece by Means of MultivariateAnalysis of Palynological Data and Comparison with Recent Phytosociological and Climatic Data," Geologie en Mi/nbouw 69, pp. 243-251. Wilkinson, T. J. 1982. "The Definition of Ancient ManuredZones by Means of Extensive Sherd9, SamplingTechniques,"JFA pp.223-233. Williams, G. E., and H. A. Polach. 1971. "Radiocarbon Dating of Arid Zone CalcareousPalaeosols," Bulletin Geological of Society America 83, pp.3069-3086. Willis, K.J. 1992. "The Late QuaternaryVegetationalHistory of Northwest Greece:I. Lake Gramousti;II. Rezina Marsh; III. A ComparativeStudy of Two ContrastingSites,"New Phytology 121, pp.101-117, 119-138, 139155. . 1994. "The Vegetational History of the Balkans," Quaternary Science Reviews 13, p. 769-788. Willis, K.J., P. Siimeg,M. Braun,and A. Toth. 1995. "The Late QuaternaryEnvironmentalHistory of PalaeoBatorliget,N.E. Hungary," Palaeogeography, Palaeoclimatology, 118, pp. 25-47. ecology Wintle, A. G. 1996. "ArchaeologicallyRelevantDating Techniquesfor the Next Century:Small, Hot, and Identified by Acronyms,"JAS 23, pp. 123-138. . 1997. "Luminescence Dating: Proceduresand ProtoLaboratory cols,"RadiationMeasurements 27, pp. 769-817. Wiseman, J. 1987. "Suggestionsfor the Study of Nikopolis in Its Changing Context,"in Chrysos 1987, pp. 411416. .1991. "FirstSeason'sResults: SatelliteImagery,Survey,and the EarliestHumans in Greece," Newsletter Boston, 1, Nikopolis pp. 1-12. . 1992a. "Archaeology and Remote Sensing in the Region of 9:3-4, Nikopolis, Greece,"Context pp. 1-4.

282

REFERENCES

. 1992b. "The Boston University Nikopolis Project,1992: Surveys and Investigationsby Land, Sea, Newsletter and Air,"Nikopolis 2, Boston, pp. 1-12. . 1993a. "Land,Sea, and Aerial Surveysin NorthwesternGreece," Context10:3-4, pp. 11-15. . 1993b. "The Nikopolis Project: Interdisciplinary Archaeologyin SouthernEpirus,Greece," Nikopolis Newsletter Boston, 3, pp. 1-12. . 1994. "Harbors, Towns, and Prehistory:Surveyand Field School in Greece,"Context11:1-2, pp. 1-4. Archaeol. 1995a. "Landscape ogy in NorthwesternGreece:The Nikopolis Project,1994,"Nikopolis Newsletter Boston, pp. 1-12. 4, . 1995b. "Peopleand the Changing Landscapein Southern Epirus,Greece,"Context12:1-2, pp. 18-21. . 1996a. "SpaceMissions and GroundTruth," 49:4, Archaeology pp. 11-13. . 1996b. "Wondersof Radar Archaeology 49:5, pp. 14Imagery," 18. . 1997a. "StudySeasonsin Ioanninaand the SurveyZone in Epirus,Greece,"Context13:1-2, pp. 10-11. . 1997b. "RemoteSensing and Archaeologyin Northwestern Greece"(paper,Gainesville,Floron ida 1996), abstractin Workshop RadarImagingand Cultural at Resource Management theAngkor on Eco-Siteand Symposium New and Technologies GlobalCultural New York, Resource Management, pp. 32-33. . 1998. "Rethinkingthe 'Halls of Hades,'" 51:3, pp. 12Archaeology 16, 18.

Wiseman,J., and A. DouzougliYamin,R., and K. B. Metheny,eds. Zachos. 1994. "The Nikopolis Proj1996. Landscape Archaeology: the ect, 1991-1993: Overview of the ReadingandInterpreting American Historical MultidisciplinaryStudy of Southern Landscape, Knoxville. Epirus"(paper, Washington,D.C. 1993), abstractin AJA98, p. 315. Yang,L. C. 1982. "The Distribution andTaxonomyof Ostracodesand Wiseman, J., C. Robinson,and C. Stein. 1999. "Archaeological Benthic Foraminifers Late Pleistin of Radarand SPOT ocene and Holocene Sediments of Applications the Troad(Biga Peninsula),Turkey" SatelliteImageryin Northwestern Greece"(paper,Chicago 1999), (M.S. thesis, Universityof Delaabstractin Abstracts the SAA64th ware). of AnnualMeeting,Washington,D.C., Yassoglou,N., C. Kosmas,and N. Moustakas.1997. "The Red Soils, p. 309. Their Origin, Properties,Use, and Wiseman,J., K. Zachos, and F. Kephallonitou. 1996. "EXX-qvoocxpepLxccvLxo Managementin Greece,"Catena28, eocovt)V," pp.261-278. rrpoypocqa ?e7tL(pq0av?LaxCov ArchDelt46, B' (1991), pp. 247-251. Zachos, K. 1998. "0(IY OpacyxoxxXy]-

. 1997.
aqeptLxavLXq

"ZovepyocTLnx]zXXqvoEr?ctaVSLax-x epeovaX

Pitb,v (aypoTSiaixto aia xoLv6'orqYt A0avawoio)," ArchDelt 48, B' acpxov

aoTo (1993), pp. 301-302. 47, vopt6 Ilp?p3?a;,"ArchDelt B' Coastal (1992), pp. 293-298. Zangger,E. 1991. "Prehistoric Environmentsin Greece:The 8ts. 1998. "EX?XYvoacpi?pxocvLx6 VanishedLandscapeof Dimini Bay TLto7TYrLOVLx6tcp6ypaxuxla sTL(pqaVetand Lake Lerna,"JFA pp. 1-15. axxcv epeovC0v N6oa 'HTcsLpo," 18, oTq] . 1992. "Prehistoric Hisand ArchDelt48, B' (1993), pp. 309-314. toric Soils in Greece:Assessing Woodward,J.C., M. G. Macklin, and the NaturalResourcesfor AgriculWeathJ. Lewin. 1994. "Pedogenic in ture,"in Agriculture Ancient ering and RelativeAge Dating of ed. Alluvial Sediments in Greece, B. Wells, Stockholm, Quaternary the Pindus Mountains of Northwest pp. 13-18. and . 1993. The Geoarchaeologythe Greece,"in RockWeathering of Evolution,ed. D. A. RobLandform Argolid,Berlin. . 1994. "Landscape inson and R. G. B. Williams, New Changes aroundTiryns duringthe Bronze York,pp. 259-283. Age,"AJA98, pp. 189-212. Wright,J. C.,J. F. Cherry,J.L. Davis, E. Mantzourani,S. B. Sutton, and Zhou, L. P.,T. H. van Andel, and R. F. SuttonJr. 1990. "The Nemea A. Lang. 2000. "A Luminescence Dating Study of Open-Air PalaeoValleyArchaeologicalProject:A lithic Sites in Western Epirus, 59, Preliminary Report," Hesperia 27, pp.579-659. Greece,"JAS pp. 609-620. Dust and Yaalon,D. H. 1987. "Saharan Zilhao,J., and F. D'Errico.1999. Desert Loess: Effect on Surround"The Chronology andTaphonomy of the EarliestAurignacianand ing Soils,"Journal AfricanEarth of Sciences pp. 569-571. Its Implicationsfor the Understand6, . 1997. "Soilsin the Mediterraing of NeandertalExtinction," nean Region:What Makes Them 13, Journalof World Prehistory Different?"Catena28, pp. 157-169. pp. 1-68.

INDEX

ACARNANIA, 3

Acheloos delta,212 Acheloos River,215 Acheron River,4, 5, 8, 160,199,204205,233,234,266 Acheron Rivervalley:and Acherousian lake, 202-204, (Fig. 6.4) 203,205, 225,227-233,234,268; and Bronze Age, 140; and coastalgeomorphological studies,27; and dolines, 57; and field and laboratory methods, 209-211; and geological coring, 209-211, (Fig. 6.7) 210,235-263, 268; geology and neotectonicsof, 205-208; and Glykys Limen, 199202,212,213,214,215,216-225, 229,234,267-268; and Kastri,204205,233,234; and loutses, (Fig. 3.10) 63; map of, (Fig. 6.2) 201; and Mesolithic, 118, 119; and microfossil analyses,209,211-212,213,214, 215,235-263; and Mousterian,112; and Palaeolithicsurvey,97, (Fig. 3.32) 100, 131; paleogeographic reconstructions (Fig. 6.12) 221, of, 6.13) 222, (Fig. 6.14) 223, (Fig. (Fig. 6.15) 224, and radiocarbon dating,209, (Table6.1) 210,217, 233; sea-level changes in, 208,216217; and sediments,209-216, 225226,231; and shorelineprogradation, 225-227; simplifiedgeology of, 206, (Fig. 6.6) 207, 208; stratigraphicanalysisof, 213,216-217, 227-233,235-263; and surface 28, survey, 30, (Table2.1) 31,266; and tectonic activity,50, 55; topographicmap of, (Fig. 6.8) 211 Acherousianlake, 22, 199,202-204, (Fig. 6.4) 203,205,225,227-233,

recon234,268; paleogeographic structionsof, (Fig. 6.13) 222, (Fig. 6.14) 223 Acheulean,98, 99, 126, 266 Actium, 160; Battle of, 3,201; Straits of, 3,4 Adriatic Sea, 50, (Fig. 3.21) 76 aerialphotographyby tetheredblimp, 7, 15, 16, 17, (Fig. 1.6) 18, (Fig. 1.7) 19,28 Aetolia, 3, 32 and agriculture: Acherousianlake, 202, 233; and artifactdistribution,136; and human-landrelationships, 48; and Louros River,198; and poljes, 58; and resurvey, 44; and site/ 43, scatters,39; and southernEpirus, 4-5,8 Alonaki:and land use, 129; and Mesolithic, 118, 121, (Fig. 3.58) 124; and Mousterian,100, 106, 107, 112,113, 126; and Neanderthals, 127; and Palaeolithicsurvey,96, 98, 100-101, (Fig. 3.32) 100, (Fig. 3.33) 101, (Figs. 3.34, 3.35) 102, (Figs. 3.36, 3.37) 103,140; and paleosol,82, 89, 103 AmbracianGulf: and agriculture, 4; and coastallandscapechange, 192198; and geological coring, 12, 13, 158, (Fig. 5.2) 159, 1594,267; geomorphologicstudies of, 27, (Fig. 5.1) 158, 160-161,267; and Holocene, 157, 169, 192; and land use, 131; and Mousterian,113; and Nikopolis isthmus, 162, (Fig. 5.3) 163,164-173; and Ormos Vathy, 158, 162, 174-177; and paleogeographicdevelopment,194-198, (Fig. 5.21) 196-197; and Preveza

284 peninsula,160,162-177; and radiocarbon dating, 167, (Table5.1) 168; and rivers,50; and sea-level changes,78,157, 162, 192-195, (Fig. 5.20) 193, 198,267; subsidence of, 208; and subsurface 157-158, 162-192, stratigraphy, (Fig. 5.18) 187; and surveyzone, 4; and tectonic activity,54, 55, (Fig. 3.2) 55, 56, 157, 161, 162, 167, 169, 192-194; wetlands exploitationin, 9 American School of ClassicalStudies, 6 Ammerman,A. J., 24 Ammoudia, 112,118, 119, (Figs. 3.52, 3.53) 120, (Fig. 3.54) 121,131, 201, 223 Ammoudia Bay,4, 9. SeealsoGlykys Limen; PhanariBay Amphilochia,3 Anatolia,228 Anavatis,80, 82, 106, 107,108, (Fig. 3.46) 111, 127 aqueduct,17,18, (Fig. 1.7) 19, 70. See alsoAyios Georgios Arachthos River,4, 50, 78, 160, 161, 217 Archaic,4, 230 architectural features,42 Argive Plain, 92 Argolid, 48, 92,106, 114,116, 125, 126,136 Argolid ExplorationProject,27 Arizona, 25 Arta (= Ambracia),4, 2663;nomos of, 3 artifacts: density of, 38, 39, 40, 42, 43, 44, 115, 135,136, 137, 139, 141, 148-154,156; distributionof, 35, 37, 93, 136, 137, 139, 147,149, 153-154, 156; and KastroRogon, 182 artifacts,stone: and deposition,95; Early Stone Age chronology,(Table 3.12) 98; and Palaeolithicchronology, 91-95, (Fig. 3.30) 94; and Palaeolithicsurvey,95-124; and patination,118, 140-141, 142,143, 148, 153; and tracts,35, 95-96. See alsochipped stone; stone tools Arvenitsa,114 Asia, 126 Asprochaliko:and backed-blade industries,116; excavationrecords of, 52; and hunting, 54, 127, 130; and later Palaeolithic,114, 115, 117, 130; and Mousterian,51, 53, 106, 113, 114, 127; and Mousterian/

INDEX

EarlyUpper Palaeolithictransition, 129; and Palaeolithicresearch, 47, 50; and radiocarbon assays,89; and Upper Palaeolithic,51, 52 Aubouin,Jean,205 Augustus,2-3, 9. SeealsoOctavian Aurignacian,113, 114, 115, 142, 143144,154-155,267 of, Ayia:and flints, 96; stratigraphy (Fig. 3.12) 65, (Fig. 3.19) 74, 75, (Fig. 3.20) 75, 76, 95; and Mousterian,(Fig. 3.20) 75, 106, 107,108, (Figs. 3.43, 3.44) 110, 129; and Neanderthals,127; site of, (Fig. 3.42) 109 Ayia Kyriaki,(Fig. 3.32) 100,112 Ayios Georgios, 17. Seealsoaqueduct Ayios Petros,137 Ayios Thomas peninsula:and later Palaeolithic,116; and Mesolithic, 121; and Mousterian,107, 113; and Ormos Vathy,174; and Palaeolithic survey,98, 103,105, (Fig. 3.41) 105; paleosol of, 82; samplingstrategy for, 8; and surfacesurvey,28, 30, 266
BACHOKIRIAN, 143,155

Bailey,G. N., 47, 52-56, 69, 85, 90, 93, 96, 114,116,130 Balkans,83, 85, 98, 113,114,129,131132, 142,143-144, 155 Batiae, 189, 195. SeealsoKastro Rizovouni Baugh,Timothy G., 16 Berbati,116, 125, 137 Berbati-LimnesArchaeologicalSurvey, 27,116 Besonen, Mark,21, 22,209, 267 bivalves,164, 167, 184. Seealso microfossilanalyses blind valleys,57 Boeotia, 58, 116 Boila, 47, 114, 116 Bordes,Francois,97 Boston University,6 Bottema, S., 83, 85 Bouchetion, 179, 189,267. Seealso KastroRogon bricks,35, 42 Bronze Age, 5, 8, 52, 95, 123,132, 139, 140,142,144,154,194, 268 Buiidel, 59 J., Bulgaria,155 179 Byzantine,6, 17731,

INDEX

285 of, 9; erosionallandscapesof, 27; fortifiedcitadelsof, 38; maps of, (Fig. 1.2) 3, (Fig. 2.1) 29, (Fig. 4.1) 136; and Pleistocene, 48; and surfacesurvey,27, 28, 32; and survey area,3-5 erosion:and AmbracianGulf, 157, 162; and artifactdistribution,136-137, 147; and Holocene, 226; and limestone landforms,50; and Nikopolis isthmus, 173; and poljes,59, 70, (Fig. 3.16) 72, 93; and Prevezaarea, 62; 113; and Quaternary, and 43; redbeds,53, 97; and resurvey, and southernEpirus,27; and surface 27, survey, 31; and tectonic activity, 48; and terrarossa,119 Etudeg6ologique, 205-206,208 Europe,51, 98, 106, 114, 126, 129, 132 of excavation: sites in Epirus,5, 47, 50, 52, 53, 89; and permit regulations, 18, 25; surfacesurveycomparedto, 26, 41-42; targeted,24
and Ambracian Gulf, 186, 188; FAUNA:

55 CALIFORNIA, Boeotian Cambridge/Bradford Expedition,25, 27 Camilli, E. L., 25, 27 Cassopaia,205 Cassope,4, 5 chaineoperatoire, 143, 148 142, 129 Chatelperronian, Cheimadio,59, (Fig. 3.11) 63, 112 J. Cherry, F., 38 chipped stone: chert, 101, 105; flint, 51, 52, 70, 93, 96, 97, 101,107, 108, 115,119, 121,126,127, 128, 129, 131,138, 140, 142-146, 147; obsidian,(Fig. 3.58) 124,131,131225 Chrysos, Evangelos,6 Clactoniantechnique,101,144 189, 195 Classical,4, 8, 17, 179, 17935, climate:and glacial-interglacial cycles, 74, 78, (Fig. 3.28) 84; and Holocene, 226; and landscapedynamics, 57; and Late Quaternary, 83-85; and Middle Palaeolithic,129; and Mousterian,106, 107,113,128; and 48, paleoenvironments, 131; and sea-level changes,78; and surface survey,28 coastalplains:and Adriatic Sea, (Fig. 3.21) 76; and AmbracianGulf, 27, 157, 162, 192-198; and Arachthos River,161; Epirus coast in last interglacial,80-83; and findspots, 91; and human occupation,50; and Late Quaternary, 76-83; and Middle and Late Quaternary 77-78, (Table 3.6) paleoshorelines, 78, (Fig. 3.23) 79; and Middle Palaeolithic,127, 130; and Mousterian,80, 109; and Palaeolithic survey,97; and spatialdistribution, 137; surfacesurveyof, 9; in western Epirus, (Fig. 3.8) 60 colonial activities,8, 42, 179, 2663 Comarus,162 computer-aidedanalysis,6, 115 controlledcollections,24-25, 115 Corfu, 52, 78, 82, 92, 106, 114, 127, 132,208 Corinth, Gulf of, 50, 131 Corinthia,137 Cullen, Brenda,22,267 culturalresourcemanagement(CRM), 25
DAKARIS, SOTIRIOS, 5,42,66, 162, 189,

231-232,233 data comparison,26, 27, 32 databases,19-20, 34 dating. Seeluminescencedating; radiocarbon(14C)dating Dean, W. E., Jr.,209 Debenham, N., 90 deflation, 136, 137 deposition:and Acheron Rivervalley, 209,212-216, 225; and Nikopolis isthmus, 173; and paleosol,92-95, 43; (Fig. 3.30) 94; and resurvey, and terrarossaredeposition,62, (Fig. 3.12) 64-65, 66, (Fig. 3.13) 66, 6970,72,74-76,96 D'Errico,F., 155 Dio Cassius,201 28 documentaryresearch, Dodona valley,4 dolines, 57, (Fig. 3.6) 58, 107 Douzougli, Angelika, 6 downslope movement, 149,153
EFSTRATIOU, 137 N.,

201,202,204,205,216,227,230,

Elatria,189, 195. Seealso Palaiorophoros electricalresistivity, 45. Seealso 44, geophysicalprospection 44, electromagneticconductivity, 45. Seealsogeophysicalprospection Eli, 113 Elis, 106, 116, 155, 179 England, 100 environmentalzones, 30, 32,265,266 Ephoreiaof Caves and 114 Paleoanthropology, 205,214, 228,232,268 Ephyra, 51, Epigravettian, 114, 116, 130 survey,50; Epirus:and archaeological cities of, 3; geological history of, 5456; Late Quaternarychronology, 85-95; Late Quaternarylandscape, 54-85; maps of, (Fig. 1.1) 2, (Fig. 3.1) 49, (Fig. 6.1) 200; and Mousterian,91-92, 97, 101,127, 128; and Palaeolithicchronology, 91-95; and Palaeolithicsites, 47-50, 125-126; Pleistocene,47-48; previousPalaeolithicresearch,5054; Roman intrusioninto, 9; satellite imageryof, (Fig. 6.5) 206; surface surveyof, 7, 9, 28; and tectonic activity,54-56, (Fig. 3.2) 55, (Figs. 3.3, 3.4) 56, 205,208 Epirus,southern:choice of, 1, 2-3; and culturalremains,8; economic basis

and artifactdistribution,136; and Asprochaliko,127; and climate history,85; and coastalplains,76; and Grammenoplain, 177; and human populationdensity,129; and hunting specialization,54; and karst landscapes,129; and KastroRogon, 182, 183-184; and Klithi,53; and land use, 107, 130; and last interglacial,80, 82; and later Palaeolithic,116, 117; and Mesolithic, 117,123; and microfossil analyses,209,211-212; and Mousterian,51, 107, 108, 128; and Neolithic, 123; and Nikopolis isthmus, 164, 167, 168; and Ormos Vathy,174; and paleoenvironments, 48; and poljes, 125, 266; and sealevel changes,78; and Upper Palaeolithic,51, 52 field methods:and site revisits,39, 4345; and site/scatters,37-42; and surfacesurveys,27, 30, 34-45; and tracts,34-37 fieldwalkingteams, 9, 31, 32 First InternationalSymposiumon Nikopolis, 5-6 Fish, S. K., 23 flora:and Asprochaliko,127; and Late 83-85; and Mesolithic, Quaternary, 123; and paleoenvironments, 48,

286 54. 131; and transhumance, Seealso vegetation Folk, R. L., 209 foragers,108, 113,127,128, 129 foraminifera, (Fig. 3.22) 77, 80, 1594, 164,167,174,177,184,186,188, 206, 209, 211-212. Seealso microfossilanalyses Fort Ancient, Ohio, 24 France,113 FranchthiCave, 92,106, 114, 116, 118, 119,155 Frangoklisia,18 GALATAS, 96, 106, 112, 113, 115, 59, 116 gastropoda,80, 82, 164, 167, 182, 184, 209, 214. Seealsomicrofossil analyses geographicinformationsystems (GIS), 13, 37, 138, 147-154 geological coring:and Acheron River valley,209-211, (Fig. 6.7) 210,235263,268; and AmbracianGulf, 12, 13, 158, (Fig. 5.2) 159, 1594,267; and archaeological survey,13; and Grammenoplain, 177-179; and KastroRogon, 179, (Fig. 5.14) 180; and Nikopolis isthmus, (Fig. 5.3) 163,164,16425, 167; and Ormos Vathy,(Fig. 5.4) 163,164, 174-177; and shorelinechanges, 12-13; and Strongyli,179; and surfacesurvey, 31 geomorphologicstudies:of Ambracian Gulf, 27, (Fig. 5.1) 158, 160-161, 267; of Epirus, 7, 12, 28; and extensivenonsystematicsurvey,32; of Glykys Limen, 216-225,226; and paleoenvironments, and 48; selection of fields and tracts,9; and site revisits,43, 45; and surface scatters,25; and surfacesurvey, 27, 31 geophysicalprospection:and Nikopolis Project,6, 7; and site revisits,44, 45; and subsurface features,17-18 global positioning systems (GPS), 16, 37 Glykys Limen, 199-202,212,213,214, 215,216-225, (Fig. 6.12) 221, (Fig. 6.13) 222, (Fig. 6.14) 223, (Fig. 6.15) 224, 229, 234, 267-268. See alsoAmmoudia Bay;PhanariBay Gourana,205 grain-size analysis:and Acheron River valley,209; and coastalsediments,

INDEX

80, (Fig. 3.26) 81; and Kokkinopilos polje, 72, (Fig. 3.17) 73; and terra rossa,66, (Fig. 3.14) 67, (Table3.2) 67, (Table3.3) 68-69, 69, 75 Grammeno,24,44-45, (Fig. 2.4) 44; plain of, 177-179,1773 , (Fig. 5.13) 178 GravaCave, 52, 114, 116, 117 Gravettian,51, 114, 116, 130 Greece:and climate history,83, 85, 226; and late entry model, 126; and Mesolithic, 123; and Mousterian, 113, 129; and nonsite surveys,25; and Palaeolithic,125; paleosol in, stratigraphy 86-89; regional studies in, 26-28; and surface survey,26, 27, 28; and urban surveys,42 Greek,22, 41,200, 201,202 44. radar, Seealso ground-penetrating geophysicalprospection Form (GTF), 19 Ground-Truthing of satelliteimagery,7, ground-truthing 13, 16, 17, 19, 28, 30 Guiscard,Robert,201 Gulf of Arta. SeeAmbracianGulf
HAMILTON, MICHAEL,17

Hammond, N. G. L., 5, 162,189,193194,201,204,205,227,231-232 Hellenistic, 4, 8,17, 179, 189,195 Hey, R. W., 66,90 Higgs, Eric: and Asprochaliko,47, 50, 51, 89, 117; and Kastritsa, 50, 47, 52, 89; and Kokkinopilos,5, 47, 50, 51, 53, 69, 89, 99, 105; and later Palaeolithic,114; and Morphi, 107; and surveyof northernGreece, 125; and transhumance, 53 52, Homer, 199,209, 229, 230, 234 Homoerectus, 98 Homosapiens,98, 114, 127, 129,266 humanbehavior:and coastalplains, 76; and geological features,49-50; and landscapeproperties,132; and Mousterian,128; and 48, paleoenvironments, 49; and political leagues, 8; and regional dynamics,26, 27; and residential mobility,108,112,114,127, 128, 129; spatialaspectsof, 37; and transhumance, 53, 54. Seealso 52, agriculture; hunting;land use; settlementpatterns human-landrelationships: and Nikopolis Project,1-2, 30, 265; and 48-49, 126, 132 paleoenvironments,

INDEX

287 Koryphi,Mt., 184 Koumouzelis,M., 114 Kowalewski,S. A., 23 Kranea,59, 96, 106, 107, 108, 113, 129 Kuhn, S. L., 113,128 Kvamme,K. L., 138
LACONIA 24-25 SURVEY,

human occupation:decline in, 123, 132; and Kokkinopilos,51; and Late Pleistocene,49; and Mousterian, 113; and Palaeolithic,47, 50, 52, 117, 126,266; and paleoenvironments, 48; and sea-level changes,78, 128; of Spilaion, 147; and terra rossa,50 and hunter-gatherers: karstlandscapes, 50; and poljes, 126, 266 hunting:and Asprochaliko,54,127, 130; and Corfu, 78; and Kleisoura, 116; and Klithi, 53, 54,108; and land use, 117, 128, 129; and offsite human activity,113; and redbeds,105; and seasonalcamps, 52, 54, 107, 108, 112,116-117, 127, 130 hydrology,48
IBERIA,129

Iliovouni, 112 Imbrie,J., 77 infraredstimulatedluminescence (IRSL) dating:97; and Anavatis,82, and 8283; Mesolithic, 118, 121; and Palaeolithicopen-air sites, 85; and sediments,90, 91, (Table3.10) 91, 92. Seealsoluminescencedating Ioannina,4, 6, 7,21 Ioannina,Lake, 83,226 Ionian Sea: and AmbracianGulf, 160, 162, 169,194, 195,198; and contactsbetween peoples, 8; and geological coring, 12, 13; and Nikopolis Project,4 Iron Age, 5,132 Italy,83,109, 113, 128, 130, 131, 132
JAPAN,55

Jing, Zhichun, 21,22,267

KANALLAKION, 4,204,228

Karimali,Evangelia,22,267 karstlandscapes: land use, 49, 59, and 107, 125, 126, 129, 131; and Late 57-59, (Fig. 3.5) 57; Quaternary, and Mousterian,107; and Preveza, 49, 50; and Spilaion, 138, (Fig. 4.4) 140; and tectonic activity,50, 131 105 Karvounari, Kastri:and Acheron Rivercourse,199, 204-205,233,234; and Acherousianlake,202,204,268; and urbansurvey,12, 42. Seealso Pandosia

Kastritsa: backed-bladeindustries, and 116; excavationrecordsof, 52; and hunting, 54, 130; and later Palaeolithic,114, 116, 117; and Palaeolithicresearch, 50, 52; and 47, post-glacialperiod, 17; and radiocarbon assays,89 KastroRizovouni,4, 5, 17, 113. Seealso Batiae KastroRogon: aerialphotographyof, 17, (Fig. 1.6) 18; and Ambracian Gulf, 159,267; and geological coring, 179, (Fig. 5.14) 180; and Louros River,179, 182, 184,189, 192, 198; paleogeographic setting of, 189, (Fig. 5.19) 190-191,195; cross sections of, 179stratigraphic 180, (Fig. 5.15) 181,182-184, (Fig. 5.16) 183, (Fig. 5.17) 185. Seealso Bouchetion Kastrosykia, (Fig. 3.45) 111 Kephalari,106, 114, 116, 155 Kephallinia,106 6 Kephallonitou,Frankiska, Kintigh, K. W., 23 Kirsten,Ernst,202,230 Kleisoura(Epirus),4 Kleisoura(Argolid), 115, 116, 119, 155 Klithi:and backed-bladeindustries, 116; excavationof, 47,52; and hunting camp, 108; and later Palaeolithic,114, 116, 130; and post-glacialperiod, 117; and radiocarbon dating, 89; and 53 transhumance, excavationrecordsof, 47, Kokkinopilos: 52, 53, 89; and flints, 96, 129; and handaxe,99, 103; and Higgs, 5, 47, 50, 51, 53, 69, 89, 99, 105; and later Palaeolithic,116; and Lower Palaeolithic,98; and luminescence dating, 89,90; and Middle Palaeolithic,155; and Mousterian, 50,51,93,105,106,107,108,112, 113; and Neanderthals,127; as open-air site, 51; and Palaeolithic chronology,(Fig. 3.30) 94,95, (Fig. 3.31) 99; polje of, (Fig. 1.5) 16, (Fig. 3.12) 65, 70, (Fig. 3.15) 71, 72, (Fig. 3.16) 72, (Fig. 3.17) 73; and terra rossa,66, 69, (Table 3.4) 70, 75-76 KokytosRiver,5,228. SeealsoVouvos River Komnena,Anna, 201,209 Konispol Cave, 117-118 Kopais,Lake, 58 Koronopoulos,233

Lambeck,K., 78, (Fig. 3.24) 79 land use: and artifactdistribution,137; and erosion,62; and foragers,108, 113, 127, 128, 129; and geoarchaeology,47; and karstlandscapes, 49, 59, 107, 125, 126, 129, 131; and laterPalaeolithic,115, 116, 130; and Mesolithic, 117; and Middle Palaeolithic,48, 126; and Mousterian, 107, 108, 113, 127; and Neanderthals,127, 128; and openair sites, 48; and paleoenvironments, 48-49; and poljes, 115,125,126, 127,266-267; and surfacesurvey, 30; and tectonic activity,54; and tracts,34; and transhumance, 54 52, and landscapearchaeology: Nikopolis Project,1-2; and paleoenvironments, 48-49; and surfacesurvey, 26, 31, 3150 Lang, Andreas,90 Late Antique, 6, 9, 164, 195,267 Leake,W. M., 162, 164,202,204,209, 222,223,232-233 Leucas, 3 Levalloistechnique:and Asprochaliko, 113; and Mousterian,52-53, 98, 106, 108, 109, 114, 129; and Neanderthals,128; and Spilaion, 142 limestone. Seekarstlandscapes Limnes valley,137 LogarouLagoon, 161 long-term replicationstudies,24 Louros,50 Louros River:and AmbracianGulf, 160, 161,16110,198; delta, 97; glacialsediment load of, 78; gorge, 5,17; and KastroRogon, 179,182, 184, 189, 192, 198; and sea-level changes, 195; and surveyzone, 4, 266; and tectonic activity,50; water channel and aqueductbridges,(Fig. 1.7) 19 Loutsa, (Fig. 3.10) 63, 106, 107, 108, 113,118, 119, (Fig. 3.55) 122, 129, 131 loutses:and Acheron Rivervalley, (Fig. 3.10) 63; and findspots,91, 93;

288 and karsticpeneplain,(Fig. 3.7) 59; and land use, 115, 125, 126, 127, 129, 266-267; and Mousterian, 107; and Palaeolithicchronology, 93-95, (Fig. 3.30) 94, 99; poljes distinguishedfrom, 59; and terrarossa,62, (Fig. 3.12) 65; uniform sediments of, 69; in western Epirus,(Fig. 3.8) 60, (Table3.1) 61 luminescencedating:and coastal plains, 82; and Palaeolithic,126; and paleosol stratigraphy, of 86; sediments,89-91, (Table3.10) 91; and terrarossa,69, 85. Seealso infraredstimulatedluminescence (IRSL) dating;thermoluminescence (TL) dating
MACEDONIA, 83, 86, 126 50,

INDEX

Myers, Eleanor Emlen, 17 Myers,J. Wilson, 17 Mytikas, 82

NEANDERTHALS: climate, 113; and and

MacLeod, D. A., 69 44, magnetometry, 45. Seealso geophysicalprospection malaria,230 Mantineia,58 Martinson,D., 77 Mavri, Lake, 5, 112 Mavrovouni,Mt., 189, 190, 194, 198 Mazoma Lagoon, 161,162,164, 167, 169,173 medieval,1, 9,22, 179,182,198, 202, 267 Mellars,P., 51 Mercouri,Melina, 6 Mesaria,112 valley,216, Mesopotamon/Tsouknida 217, (Fig. 6.9) 218, (Fig. 6.10) 219, (Fig. 6.11) 220,223,227, 228,232 Messenia, 106, 116 microfossilanalyses:and Acheron River valley,209, 211-212,213,214,215, 235-263; and sea-level changes,77, 208 51, Micromousterian, 53, 106 Middle East, 55 mineralcomposition:of last interglacial coastalsands,(Table3.7) 83; of terra rossa,66, (Table3.4) 70, (Table3.5) 71 Moore, Melissa, 8,21 Morphi, 72, (Fig. 3.18) 74,105, 107, 108,112,113, 129 Mousterian,50, 51, 52, 53, (Fig. 3.20) 75, 80,89,91, 93,97, 98,100,101, 103, 105-114, (Figs. 3.43, 3.44) 110,126,127, 128, 129, 155 Murray,Priscilla,7

Homosapiens,129,266; and land use, 127,128; and Mousterian,98, 106, 107, 108, 114; and Mousterian/ EarlyUpper Palaeolithictransition, 129-130; populationdensity of, 128 Near East, 98,144 Nekyomanteion,5 Nemea, 116, 125 Nemea ValleyArchaeologicalProject, 27 Neolithic, 5, 52, 62, 95, 123, 132, 137, 144,154, (Fig. 5.12) 176, 177,194 New Zealand,55 Nikopolis:and AmbracianGulf, 158; territoryof, 2663;city plan of, 7; and interactions,30; core-periphery founding of, 9; mappingof, 6; and Ormos Vathy,39, 41,177; and Prevezapeninsula,162; regional context of, 5, 7; Roman period, 173, 266 Nikopolis isthmus:62,164-173; and geological coring, (Fig. 5.3) 163, 164,16425,167; map of, (Fig. 5.3) 163; and sea-level changes, 168, 173,193, 195; shorelinechanges of, 169, 173, (Fig. 5.10) 173; stratigraphiccross sections of, 164, (Fig. 5.5) 165, (Fig. 5.6) 166, 167-169, (Fig. 5.7) 170, (Fig. 5.8) 171, (Fig. 5.9) 172 Nikopolis Project:backgroundand organizationof, 5-8; dailywork assignments,(Table2.2) 33; documentation,19-20; field school studentsof, (Table 1.2) 12; and human-landrelationships,1-2, 30, 265; methodologies of, 9,12-13, 15-18,265; and post-fieldwork analyses,21; prehistoricsurvey contribution,131-132; and presentationof results,21-22; projectstaff of, (Table 1.1) 10-11; purposeof intensive survey,28; and regionalstudies in Greece,26-28; researchaims of, 8-9; and sampling strategies,29-31; and southern Epirus as study choice, 1, 2-3; surveyzone of, (Fig. 1.1) 2, 3-5, (Fig. 1.2) 3, 7, (Figs. 1.3, 1.4) 14, 29, 47,266

INDEX

289 167, (Table5.1) 168, 169,186,188, 193, (Fig. 5.20) 193; effective range of, 51, 89; and Glykys Limen, 216, 223,224; and KastroRogon, 179, 182, 184; of sediments,89, 98; and Upper Palaeolithic,114 Rapp,George (Rip), 7,21,22,267 redbeds:and archaeological survey,50; diversityof, 61-62; and erosion,53, 97; and hunting camps, 105; and karstlandscapes,58; and Kokkinopilos,53, 96; and Middle Palaeolithic,52; and Mousterian sites, 50, 106, 107; and open-air sites, 50, 53; and Pliocene, 53, 76 regionaldynamics,26,27 remote sensing,6, 7, 13-18, 24, 44,266 43-45 resurvey, Rick,J.W., 137 rockshelters: Argive Kleisoura,119; and and Bailey,52; and Higgs, 50,51; and hunting,54, 117; and land use, 130; and laterPalaeolithic,115,116, 117; and Mousterian,51,106; stratifieddeposits in, 89; and Upper Palaeolithic,50, 51, 52 Rodaki,66, 82, (Table3.7) 83, 107, 108, 109, (Figs. 3.47, 3.48) 112 Rodia Lagoon, 161,186, 188 Rokia,Mt., 4, 160, 180, 186, 189, 194, 267 Rolland,N., 98 Roman, 1, 9, 17, 18, 22, (Fig. 2.3) 41, 41,44, 70, 7068,158, 162, 164, 173, 179, 180, (Fig. 5.12) 176, 177, 17731, 184,189,194,195, 198, 202,266 Romia, 112 Runnels,Curtis, 7,22,266,267 Russell,Richard,228
SALAORA BARRIER, 161,169,188,195

OBSIDIAN,

(Fig. 3.58) 124, 131, 131225 Octavian,9,201. SeealsoAugustus Orchomenos,58 Ormos Odysseos, 83, 103, (Figs. 3.38, 3.39, 3.40) 104, 112 Ormos Vathy:158, 162, 174-177; and EarlyPalaeolithicmaterials,103; and geological coring, (Fig. 5.4) 163,164, 174; and Holocene, 177; map of, (Fig. 5.4) 163; and Mousterian,107, 127; and Nikopolis, 39, recon41, 177; paleogeographic structionsof, (Fig. 5.12) 176, 177; and Roman period, 162, 164, 177; and sea-level changes, 174, 177, cross sections of, 195; stratigraphic 174, (Fig. 5.11) 175, 176-177; and surfacescatters,24; surveyunits of, (Fig. 2.3) 41; and tectonic activity, 177, 192 ostracoda,1594,164,167,177,180, 182, 183, 184, 186, 188,209,211212,213, 214, 215. Seealso microfossilanalyses oxygen isotope stages (OIS), 77, (Fig. 3.22) 77, 78, (Table3.6) 78, (Fig. 3.24) 79, 80

17. See PALAIOROPHOROS, alsoElatria 77-83, (Table3.6) 78, paleoshorelines, (Fig. 3.23) 79, (Fig. 3.25) 81 paleosols:and Ayia, (Fig. 3.19) 74; of, chronosequences 86-89; and coastalplains, 80, 82; and Kokkinopilos, 70; maturitylevels of, 80, 82, 85, 86, (Fig. 3.29) 87, (Table3.8) 87, (Table3.9) 88, 89, 90, 91, 9293; and Mesolithic, 119, 121; and Morphi, 72, (Fig. 3.18) 74; and Nikopolis isthmus, 17,164; and Palaeolithicdating problem,85; and Palaeolithicsites, 48; and Palaeolithic survey,95, 96; and Rodaki, 108; and terrarossa,61, 62, (Fig. 3.12) 64, (Fig. 3.13) 66, (Fig. 3.30) 94 PaliouriasRiver,82, 108, 121 Pandosia,204-205,268. Seealso Kastri Papagianni,Dimitra, 21, 97, 127 5 Paramythia, 4, 56, 97, 115, 118 Parga, Paschos,Panayiotis,69 See pastoraltranshumance. transhumance patination,118, 140-141,142, 143, 148, 153

Peloponnese,50, 58, 86, 94, 114, 125, 131 PeloponnesianWar,201 peneplains:and Palaeolithicsurvey,95; and tectonic activity,50, 55, 56, 58, (Fig. 3.7) 59 Perles,C., 114 PhanariBay,4, 160, 199-202, (Fig. 6.3) 202,212,213,214,215,216-225, 234. SeealsoAmmoudia Bay; Glykys Limen Philippias,4 Philippson,Alfred, 202,230 phosphatestudies,24-25 Pindos: and AmbracianGulf, 160; and Arachthos River,161; and glacial periods,48; and karstlandscapes, (Fig. 3.5) 57; and land use, 131; and 86; paleosol stratigraphy, and poljes, 59; and tectonic activity,54, 55 Plog, S., 23 Pogonitsa,Lake, 116, (Fig. 3.51) 118 poljes:coastalplains comparedto, 77; and erosion,59, 70, (Fig. 3.16) 72, 93; and findspots, 91, 93; and land use, 115, 125, 126, 127,266-267; and Mousterian,107-108; and Palaeolithicchronology,93-95, (Fig. 3.30) 94, 99; and tectonic activity,58, 59, (Fig. 3.7) 59; and terrarossa,62, (Fig. 3.12) 64; in western Epirus,58, (Fig. 3.8) 60, (Table 3.1) 61 post-medieval,17935,182, 195, 198 pottery fragments:and Acheron River valley,209, 233; and tracts,35; and urbansurveys,42 Preveza,nomos of: chronostratigraphic diagramfor,90, (Table 3.11) 92; limestone landforms,49; and paleoshoredeposits, 82; and Pleistocene, 48; and surveyzone, 3, 4,29,47,266 Prevezapeninsula:and Ambracian Gulf, 160, 162-177; and Mesolithic, (Fig. 3.51) 118, 119, 121; and Nikopolis isthmus, 162, 164-173; and Ormos Vathy,174-177; and Salaorabarrier, 161; and tectonic activity,168, 169, 173, 174, 192, 195 Pseudo-Scylax,189, 192 Pylos, 125
RADIOCARBON (I4C) DATING: and

Acheron Rivervalley,209, (Table 6.1) 210,217,233; and Acherousian lake,227,228; and AmbracianGulf,

SalaoraIsland, 188 samplingstrategies,and surfacesurvey, 23-25, 26,29-31, 43 Sarris,Apostolos, 18 satelliteimagery:effectivenessof, 13, 15; of Epirus, (Fig. 6.5) 206; and fieldwalkingteams, 9; groundtruthingof, 7, 13, 16, 17, 28, 30; and landscapelocations, 16; of survey zone, (Figs. 1.3, 1.4) 14 scouting:and extensivenonsystematic survey,32, 43; and site/scatter,41; and surfacesurvey,28, 30 sea-level changes:and Acheron River valley,208, 216-217; and AmbracianGulf, 78, 157, 162, 192-

290

INDEX

195, (Fig. 5.20) 193, 198, 267; and coastalsites, 137; and Holocene, 208; and Late Quaternary, 76-83; and later Palaeolithic,116; and Nikopolis isthmus, 168, 173, 193, 195; and Ormos Vathy,174,177, 195; and oxygen isotope stages (OIS), 77, (Fig. 3.22) 77, 78, (Table 3.6) 78, (Fig. 3.24) 79, 80; and 48; paleoenvironments, and Spilaion, 138 sediments:and Acheron Rivervalley, 209-216,225-226,231; and AmbracianGulf, 157,160,162, 192, 195,198; and Grammeno plain, 177, 179; and karstlandscapes, 125; Late Quaternaryand Palaeolithicchronology,91-95; of loutses, 69; luminescencedating of, 89-91, (Table3.10) 91, 92; and Nikopolis isthmus, 164,167,173; radiocarbon dating of, 98; and Spilaion, 138, 141, 147; and stone artifacts,93; Walther'sLaw, 162, 16216, 213 Seidi, 116 settlementpatterns:and Ambracian Gulf, 157-158; and geoarchaeology, 47; and Higgs, 52; and later Palaeolithic,115, 116, 117; and Mousterian,107,108, 113; postPleistocene settlement history,117124, 130-131; in southernEpirus,8 Sidari,52, 117-118, 123, 131 simulationstudies,24 site revisits:and field methods, 39, 4345; and Palaeolithicsurvey,96; purposeof, 43, 4378 site/scatters(SS): and artifactdistribution, 136-137; and concept of site, 1351,136, 141; documentationand collection procedures, 39-40; explanationof, 20; and field methods, 37-42; and Palaeolithic survey,97, (Fig. 3.31) 99, (Fig. 3.32) 100, 105, (Fig. 3.42) 109, (Fig. 3.45) 111, (Fig. 3.51) 118, 134; and taphonomyof surfacesites, 137138; and tracts,40-41, (Fig. 2.3) 41 Skepasto,106, 107, 108 Sordinas,Augustus,52, 108, 117 Spain, 113 spatialanalysis,and Spilaion, 141, 147, 148-154, (Table4.3) 150, 156,267 spatialcoverage:and site/scatters,4041; and surfacesurvey,33-34; and tracts,34

causes spatialdistributionof artifacts: of, 136, 137-138; and Palaeolithic survey,97; and Spilaion, 135-136, 141,142, 147-148, 150, (Fig. 4.12) 151, (Fig. 4.13) 152, 153 spatialpatterns:and AmbracianGulf, 162; and artifactdensity,137; and Spilaion, 148, 153, 156 Spercheiosdelta,225 Spilaion:and Aurignacian,115, 130, 142, 143, 154-155,267; categories of flintknappingdebitage,142, (Table4.1) 143; and end scrapers, 115, (Fig. 3.50) 115, 142, 144, (Figs. 4.6, 4.7) 145, (Figs. 4.8, 4.9, 4.10, 4.11) 146, (Fig. 4.13) 152,153,154, 155; and geographicinformation systems (GIS) analyses,147-154; and land use, 131; lithic assemblage of, 142-146, 155-156,267; maps of, (Fig. 4.1) 136, (Fig. 4.2) 139; samplegrid, (Fig. 4.5) 141; site description,138-141; and spatial analysis,141,147, 148-154, (Table 4.3) 150, 156, 267; and spatial distribution,135-136,141,142, 147-148, 150, (Fig. 4.12) 151, (Fig. 4.13) 152, 153; topographicmap of, (Fig. 4.2) 139; views of, (Fig. 4.3) 139, (Fig. 4.4) 140 SPOT imagery,of surveyzone, 13, (Figs. 1.3, 1.4) 14. Seealsosatellite imagery Stavros,Mt., 4, 160, 189, 194,267 Stein, Carol, 8, 15 Stephani,50 Stiner,M. C., 128 stone tools:Acheulean,98, 99, 126, 266; backed-blades,51, 52, 53, 114, 115-116, (Fig. 3.50) 116, 117, 119, (Fig. 3.52) 120, 130; becs,(Fig. 3.52) 120, 144, (Fig. 4.10) 146; bifaces, (Fig. 3.41) 105; bifacialfoliates (leafpoints),51, 52, 113, 114, 144; bladelets,118, 119, (Fig. 3.58) 124; blades,53, 96, (Fig. 3.43) 110, 115, 118, 142, 143, (Fig. 4.12) 151, 153, 154; burins, 114, 115, (Fig. 3.52) 120, 129, 142, 144; choppers,(Fig. 3.35) 102; Clactoniantechnique, 101, 144; core-choppers,19, 101, (Fig. 3.36) 103, (Fig. 3.48) 112, 126, 143; cores,96, 101, (Fig. 3.37) 103, 106, 108, (Fig. 3.43) 110, 113, 115, 118, 119, (Fig. 3.55) 122, 123, (Fig. 3.58) 124, 128, 131,142-143, (Fig. 4.9) 146, (Fig. 4.12) 151, 153;

INDEX

29I

corticalflakes,142,144,150, (Fig. 4.12) 151, 153, 154; denticulates, 99, 101, (Fig. 3.34) 102, 117,119, 128, 142,144, (Fig. 4.6) 145, (Fig. 4.10) 146, (Fig. 4.13) 152, 153; Dufour bladelets,154; end scrapers, 53, 109, (Fig. 3.43) 110, 114, 115, (Fig. 3.49) 115,117, 118, 119, (Fig. 3.52) 120, 121, 123, 129, 142, 144, (Figs. 4.6,4.7) 145, (Figs. 4.8,4.9, 4.10, 4.11) 146, (Fig. 4.13) 152, 153,154, 155; Epigravettian, 51, 114, 116,130; Gravettian,51,114, 116, 130; handaxes,98, 99, 103, (Figs. 3.38, 3.40) 104, 126,266; Levalloistechnique,52-53, 98, 106, 108,109,113,114,128,129,142; microburins,119; microliths,117, 119, (Fig. 3.52) 120, (Fig. 3.58) 124; notched pieces, 99, (Fig. 3.34) 102, 109, 117, 119, 142, 144, (Figs. 4.9, 4.10) 146, (Fig. 4.13) 152, 153; perfoirs,117,119, (Fig. 3.52) 120, 144, (Fig. 4.10) 146;piecesesquillees, 144; plain flakes, 142, 144, 150, (Fig. 4.12) 151,153, 154; points, 53, 101,106, 108, 113,114, (Fig. 3.53) 120, 128, 129; raclette, (Fig. 4.6) 145; retouchedtools, 96,101,108, 114, 115, 117, 118, 119, (Fig. 3.52) 120, 121, (Fig. 3.55) 122, 123, (Fig. 3.58) 124, 127, 128, 129, 142, 144, (Table4.2) 144, (Figs. 4.6-4.11) 145-146, (Fig. 4.12) 151, (Fig. 4.13) 152,153, 154; side scrapers, 51,101, (Fig. 3.34) 102, 106, 109, (Fig. 3.43) 110, (Fig. 3.48) 112, 113, 114, 128, 129, 144, (Fig. 4.7) 145; and silica gloss, 118, 119, 121, (Fig. 3.55) 122, (Fig. 3.58) 124; tanged arrowhead, (Fig. 3.58) 124; tranchet arrowhead, (Fig. 3.58) 124; trapezes, 117, 118, 119, (Figs. 3.52, 3.53) 120, 121, (Fig. 3.55) 122, (Fig. 3.58) 124,131 Strabo,162, 189, 200,209,230 Strongyli,18, 179, 184, 189 Stymphalos,58 surfacescatters,24, 25, 39, 4378, 1351, 136-137 coordinationof, survey,archaeological: 13; dailywork assignments,(Table 2.2) 33; diachronic,26, 95,265,266; of Epirus,7, 9, 28; excavation comparedto, 26, 41-42; extensive, 28, 32, 43; field methods, 27, 30, 34-45; and human activity,17;

intensity and coverageof, 31-34, 265; intensive,32; as less destructive methodology of, 12, technique,2632; 25-28; and Nikopolis Project,8, 30; and Palaeolithicsurvey,95-124; purposeof, 28; and remote sensing, 13; and samplingstrategies,23-25, 26, 29-31, 43; and surfacescatters/ 24-25 subsurface relationship, survey,geological,9, 13 survey,geophysical,28 survey,Palaeolithic:conclusionsof, 125-131; Early Palaeolithic,98105; and extensivenonsystematic survey,32; geological setting of sites, of, 97-124; goals and procedures 95-97; laterPalaeolithic,114-117; Mesolithic, 117-124, 130-131; and methodology,31; Mousterian (Middle Palaeolithic),97, 103,105114; Neolithic, 132; and surface survey,28 42 survey,topographical, survey,urban,12, 40, 41-42 surveyzone: multispectral(MSS) imageryof, 13, (Figs. 1.3, 1.4) 14; of Nikopolis Project,(Fig. 1.1) 2, 3-5, (Fig. 1.2) 3, 7, (Figs. 1.3, 1.4) 14, 29, 47,266; size of, 9 212 Swain, Frederick, Sybota,Battle of, 201,229
THOMAS, 8, 12, 21,266 TARTARON,

tectonic activity:and Acheron River valley,50,55; and AmbracianGulf, 54, 55, (Fig. 3.2) 55, 56, 157, 161, 162, 167,169,192-195; and Epirus, 54-56, (Fig. 3.2) 55, (Figs. 3.3, 3.4) 56,205,208; and Ormos Vathy,177, 48, 192; and paleoenvironments, 131; and poljes,58, 59, (Fig. 3.7) 59; and Prevezanomos, 49; and Preveza peninsula,168, 169, 173,174, 192, 54 195; and transhumance, Tegea, 85 Tenaghi Philippon, 83 terrarossa:color of, 62; and grain-size analysis,66, (Fig. 3.14) 67, (Table 3.2) 67, (Table 3.3) 68-69, 69, 75; and karstlandscapes,58, (Fig. 3.7) 61-62; 59; and Late Quaternary, and luminescencedating, 69, 85; and Mesolithic, 119; mineral composition of, 66, (Table3.4) 70, (Table 3.5) 71; and Palaeolithic survey,95; and redeposition,62, (Fig. 3.12) 64-65, 66, (Fig. 3.13) 66,

69-70, 72, 74-76, 96; and Spilaion, 138, 141; and tectonic activity,50 Thematic Mapper (TM) satelliteimagery, 13. Seealsosatelliteimagery Theopetra, 116 thermoluminescence (TL) dating:97; and Asprochaliko,89; and Mesolithic, 121; and Palaeolithic open-air sites, 85; and sediments, 90, 91, (Table3.10) 91, 92. Seealso luminescencedating Thesprotia,205 Thesprotikovalley,(Fig. 3.31) 99,112113 Thessaly:and Aurignacian,113,155; and backed-bladeindustries,116; and deposition,94; landscapeof, 48, 50; and Mousterian,106; and Mousterian/EarlyUpper Palaeolithictransition,129; and Palaeolithic,114, 125, 126; and Palaeolithicchronology,92; and 86; paleosol stratigraphy, and prehistoricperiods,2 Thucydides,200-201,209, 229-230 Thyamis River,217 Thyamis valley,208 tiles, 35, 42, 177 Tippett, H., 66, 90 topographicmaps:of Acheron River valley,(Fig. 6.8) 211; as documentation, 16, 20, 36, 37, (Table3.1) 61, 16426, 211,225; of Spilaion,(Fig. 4.2) 139 and topography: Acherousianlake,202, 229; and spatialdistributionof artifacts,138, 148, 149; and surface survey,28, 30; and tectonic activity, 54; and tracts,34 Tourkovouni,121 tracts(T): archaeological surveytract form, 36-37, (Fig. 2.2) 36; database, 19; and field methods, 34-37; and Palaeolithicsurvey,95-96; and site/ scatters,40-41, (Fig. 2.3) 41; and surfacesurveys,28, 30; and urban surveys,42; and walkovers,43 30, tractwalking, 41 transhumance, 53, 54 52, Tsarlambas, (Fig. 3.27) 82, 121, 82, (Fig. 3.58) 124, 131 TsoukalioLagoon, 161 Tsouknida,112, 118, (Fig. 3.52) 120, 131,228 TypicalBalkanAurignacian,142,143, 154 Tzedakis,P. C., 83

292

INDEX

ULBRICH,114, 116

distribution,138, 147; and Spilaion,

138. See also flora

Wuiirm glaciation,192,194
XINIAS,LAKE,226

and resource UnitedStates: cultural 25; (CRM)surveys, management andsurface scatters, andsurface 24; sites,137-138;andtectonic activity, 55
VALANIDORRACHI, 107, 108,127 106,

Valtos Kalodiki, (Fig.3.9) 63,107 59, vanAndel,Tjeerd, 7,22,266,267 108 Vassiliko, River and valley, vegetation: Acheron 214;andglacial-interglacial cycles, (Fig.3.28) 84;andlandscape 57; dynamics, andLateQuaternary, and 83-85;andloutses poljes,125; andMousterian, 108;and 107, 43, resurvey, 45;andspatial

villas, 18, 44, 179 Villas, Cathleen,212 Vita-Finzi, C., 52 VoulistaPanayia,4, 17 Vouvopotamos,112,115 VouvosRiver,5, 216. SeealsoKokytos River WALKOVERS 20, 28, 43, 96, 141 (W), Walther'sLaw, 162, 16216,213 Wandsnider,L., 25,27 Waters,David, 55,56,208,225 Weymouth,John, 17-18 Willis, Katherine,85,226 Wiseman,James,6 Wiseman, Lucy,7-8

Xirolophos,233 Xylokastro,205
D. YAALON, H., 69

YoungerDryas, 85, 89 57, Yugoslavia, 58

6 KONSTANTINOS, ZACHOS,

Zaimis, 116 Zakynthos,108-109 Zalongo, Mt., 4, 160 Zhou, Li-Ping, 90 Zilhao,J., 155 Ziros, Lake, 5

0-87661-532-9