Hesperia, Volume 59, Issue 1 - Proceedings of the First International Conference on Archaic Greek Architectural Terracottas

HESPERIA AMERICAN SCHOOL OF THE JOURNAL ATHENS AT STUDIES CLASSICAL OF VOLUME 59: NUMBER 1 JANUARY-MARCH 1990 FIRS...

Author: The American School of Classical Studies at Athens; various authors

14 downloads 485 Views 58MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

HESPERIA

AMERICAN SCHOOL OF THE JOURNAL ATHENS AT STUDIES CLASSICAL OF VOLUME

59:

NUMBER

1

JANUARY-MARCH 1990

FIRST INTERNATIONAL CONFERENCE ON ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

DECORATED ARCHITECTURAL TERRACOTTAS FROM THE ATHENIAN ACROPOLIS: CATALOGUE OF EXHIBITION

AMERICAN

SCHOOL

OF CLASSICAL STUDIES

AT ATHENS

HESPERIA

(USPS

VOL. 59, No. 1.

112-590)

TABLE OF CONTENTS

PAGES

Proceedingsof the First InternationalConferenceon Archaic Greek ArchitecturalTerracottas.......

1-323

CHRISTINAVLASSOPOULOU: Decorated Architectural Terracottas from the Athenian Acropolis: Catalogue

of Exhibition

American DcDoolof Clasofcal

Dtubtec at atbenf

COMMITTEE:Leslie P. Day (Chairman), Mabel L. Lang, Steven Lattimore, James R. PUBLICATIONS

McCredie (ex officio), Susan I. Rotroff, Michael B. Walbank.

Marian Holland McAllister. EDITOROF PUBLICATIONS:

HESPERIA. The annual subscription price is $40.00 net in the United States, $46.00 net in Canada and other countries, payable in advance in dollars. Published quarterly. Single fascicules (current and back numbers when available) $11.00 in the United States, $12.50 in Canada and other countries, postpaid. Volumes I-XLI, Index Volume I (Hesperia 1-10, and Supplements I-V), and Supplements I-XI should be ordered from Swets and Zeitlinger, B. V., P.O. Box 801, 2160 SZ Lisse, The Netherlands, Volume XLII and following and Supplements XII and following from the American School of Classical Studies at Athens, c/o The Institute for Advanced Study, Princeton, New Jersey 08543-0631, U. S. A. Single fascicules of Hesperia 59.1 available at $40.00 plus postage. Index to Hesperia, Volumes 11-20, Supplements VII-IX. vi + 434 pages quarto, paper, 1968. Supplement XII: The Athenian Constitution after Sulla. By DANIEL J. GEAGAN.1967. Supplement XIII: Marcus Aurelius, Aspects of Civic and Cultural Policy in the East. By JAMES H. OLIVER.1970.

SupplementXIV: The Political Organizationof Attica. By JOHNS. TRAILL.1975. Supplement XV: The Lettering of an Athenian Mason. By STEPHENV. TRACY.1975. Supplement XVI: A Sanctuary of Zeus on Mount Hymettos. By MERLEK. LANGDON.1976. Supplement XVII: Kallias of Sphettos and the Revolt of Athens in 286 B.C. By T. LESLIESHEAR,JR. 1978. Supplement XIX: Studies in Attic Epigraphy, History and Topography. 1982. Supplement XX: Studies in Athenian Architecture, Sculpture and Topography. 1982.

1986. SupplementXXI: Excavationsat Pylos in Elis. By JOHNCOLEMAN. Supplement XXII: Attic Grave Reliefs that Represent Women in the Dress of Isis. By ELIZABETHJ. WALTERS.1988. Publication office: c/o The Institute for Advanced Study, Olden Lane, Princeton, New Jersey. Produced at 242 S. Eden Street, Baltimore, Md. 21231. All communications for the Editor should be sent to THE AMERICAN SCHOOL OF CLASSICAL STUDIES AT ATHENS c/o

THE INSTITUTE FOR ADVANCED

STUDY,Princeton, New Jersey, 08543-0631, U. S. A. Second-class postage paid at Princeton, New Jersey and at additional mailing offices. Postmaster: Send address changes to Hesperia, THE AMERICANSCHOOLOF CLASSICAL STUDIESAT ATHENSc/o THE INSTITUTEFORADVANCEDSTUDY, Princeton, New Jersey, 08543-0631, U. S. A.

ISSN 0018-098X ISBN 87661- 500-0

PROCEEDINGS OF THE

FIRST INTERNATIONAL CONFERENCE ON

ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS DECEMBER

2-4, 1988

EDITED BY

NANCY A. WINTER

n

AMERICAN

SCHOOL OF CLASSICAL STUDIES AT ATHENS

TABLE OF CONTENTS ABBREVIATIONS . GLOSSARY . WILLIAM

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

D. E. COULSON:Opening Remarks

.

.

.

.

.

.

.

.

.

.

.

.

. .

5

.

.

.

.

.

.

.

.

.

.

.

.

7

. . . . . .

. ...

.

..

11

NANCY A. WINTER: Defining Regional Styles in Archaic Greek Architectural Terracottas . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

CHRISTIANLE ROY: Les terres cuites architecturales de Delphes vingt ans apres la publication . . . . . . . . . . . . . . . . . . . . . . . . . . .

33

. . . . . . .

JOACHIMHEIDEN: Die archaischen Dacher von Olympia MARY C. ROEBUCK:Archaic Architectural

NANCY K. COOPER:Archaic Architectural MARIE-FRANyOISE

Terracottas from Corinth

41

..

. . . . . .

d'Argos et d'Epidaure.

BILLOT:Terres cuites architecturales

.

47 65

Terracottas from Halieis and Bassai .

Notes de typologie et d'histoire . . . . . . K.

.

.

. . . . . . . . ..

95

G.

KOLOKOTSAS: AppuoVLKESXapa4e?L (rra apXaLKa apXLTreKTrovKa 7r7Xlva rov . . . . . . . . .. . . ..... . .. 'Apyovs KaL Ts27 Erbapov

Antefixes from the Argive Heraion . . . .

149

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

157

CHRISTOPHERA. PFAFF: Three-peaked BERIT WELLS: The Asine Sima.

141

VERONIKA MITSOPOULOS-LEON:Ein neuer Antefixtyp

aus dem Heiligtum

der

Artemis in Lousoi . . . . . . . . . . . . . . . . . . . . . . . . .

163

HUBNER: Die Dachterrakotten

der archaischen Tempel von Kalapodi . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 . (Phokis) F. DAKORONIA: ApXaiLK' KEpa/dles, a7TrOT77VAvaToALK . . . . . . 175 r AoKpa

GERHILD

AENNE OHNESORG:Archaic Roof Tiles from the Heraion on Samos . . . . . . .

181

EVA SIMANTONI-BOURNIAS: Chian Relief Pottery and its Relationship to Chian and East Greek Architectural Terracottas. . . . . . . . . . . . . . . .

193

JACQUES Y. PERREAULT:L'atelier de potier archaique

de Phari (Thasos):

. . . . .

201

PETER SCHNEIDER:New Information from the Discovery of an Archaic Tiled Roof in Ionia . . . . . . . . . . . . . . . . . . . . . . . . . . .

211

La production de tuiles .

. . . . .

KONRADZIMMERMANN: Archaische Dachterrakotten MADELEINE MERTENS-HORN

.. .

. . .

aus Histria

AND LUISA VIOLA: Archaische

griechischer Typologie in Delphi und Olympia VOLKERKASTNER:Scheibenformige

.

. . . . . . . . 223 Tondacher

west-

. . . . . . . . . .

Akrotere in Griechenland

und Italien

. 235

. . . .

251

4

TABLE OF CONTENTS

An East Greek Master Coroplast at Late Archaic Morgantina JOHN F. KENFIELD: CHARLOTTE The Artemision Sima and its Possible Antecedents . WIKANDER:

265

. . . . Archaic Roof Tiles: The First Generations. ORJANWIKANDER: ERNST-LUDWIGSCHWANDNER: Uberlegungen zur technischen Struktur und Dachterrakotten . . . . . . . . . . ... archaischer Formentwicklung

285

RAINERC. S. FELSCH:Further Stamped Roof Tiles from Central Greece, Attica, . and the Peloponnese . . . . . . . . . . . . . . . . . . . . . . PLATES 1-56

CHRISTINA Decorated Architectural Terracottas from the Athenian VLASSOPOULOU:

Acropolis:Catalogue of Exhibition

275

291 301

ABBREVIATIONS In additionto the abbreviationslisted in the AmericanJournalof Archaeology90, 1986, pp. 384-394 and 92, 1988, pp. 629-630, the following are used: = A.Akerstr6m,Die architektonischenTerrakottenKleinasiens,Lund 1966 = E. Buschor,"AltsamischerBauschmuck,"AM 72, 1957, pp. 1-34 = N. K. Cooper, The Developmentof Roof Revetment in the Peloponnese,diss. Universityof Minnesota, 1983 (Studiesin MediterraneanArchaeology,forthcoming) CorinthIV, i = I. Thallon-Hill and L. S. King, Corinth,IV, i, DecoratedArchitecturalTerracottas, Cambridge,Mass. 1929 = NFGH G. Dontas, Dontas, "Denkmalerund Geschichte eines kerkyraischenHeiligtums,"NFGH, pp. 121-133 = E. Dyggve Dyggve, Das Laphrion.Der Tempelbezirkvon Kalydon,Copenhagen 1948 = 1979 R. C. S. Felsch, "BoiotischeZiegelwerkstattenarchaischer Zeit," AM 94, 1979, Felsch, pp. 1-40 1987 R. = C. S. Felsch, "Berichtiiber die Grabungenim Heiligtum der ArtemisElaphebolos Felsch, und des Apollon von Hyampolis 1978-1982," AA (JdI 102) 1987, pp. 1-24 Felsch and Schuler, = R. C. S. Felsch and H. Schuler, "Apollonund Artemis oder Artemis und Apollon? 1980 Bericht von den Grabungen im neu entdeckten Heiligtum bei Kalapodi 19731977,"AA (Jdl 95) 1980, pp. 38-112 = E. D. Van Buren, GreekFictile Revetmentsin the ArchaicPeriod, London 1926 GFR GM = L. Shoe, Profilesof GreekMouldings, Cambridge,Mass. 1936 = J. Heiden, KorinthischeDachziegel. Zur Entwicklung der korinthischenDacher, 1987 Heiden, Frankfurt 1987 1975 = G. "Dachterrakottenim Magazin des Museums von Nauplia," in Tiryns. Huiibner, HuAbner, Forschungenund BerichteVIII, Mainz 1975, pp. 117-136 Koch = H. Koch, "Studienzu den campanischenDachterrakotten,"RM 30, 1915, pp. 1-115 = K. A. Rhomaios, "Die Dachterrakotten,"in G. Rodenwaldt,Korkyra,I, Der ArteKorkyraI mistempel, Berlin 1940, pp. 97-162 = C. Le Roy, Fouilles de Delphes, II, Les terrescuites architecturales,Paris 1967 Le Roy, 1967 Martin = R. Martin, Manuel d'architecturegrecque I, Paris 1965 Mertens-Horn, 1978 = M. Mertens-Horn, "Beobachtungenan dadalischen Tondachern,"JdI 93, 1978, pp. 30-65 NFGH = Neue Forschungenin griechischenHeiligtiimern,Tubingen 1976 = W. Dorpfeld and R. Bormann,in E. Curtius and F. Adler, Olympia.Die Ergebnisse Olympia II der Ausgrabungen,II, Die Baudenkmiler, Berlin 1892 Orlandos = A. Orlandos,Les matefriauxde constructionet la techniquearchitecturaledes anciens grecs I, Paris 1966 = K. Rhomaios, Kcpaot rijs KaXAv8avos, Rhomaios Athens 1951 = H. S. Robinson, "Roof Tiles of the Early Seventh Century B.C.," AM 99, 1984, Robinson,AM pp. 55-66 = H. S. Robinson, "Temple Hill, Corinth,"NFGH, pp. 239-260 Robinson,NFGH = H. S. Robinson, "Excavationsat Corinth:Temple Hill: 1968-1972," Hesperia 45, Robinson,TH 1976, pp. 203-239 Schwandner,1985 = E.-L. Schwandner, Der dltere Porostempel der Aphaia auf Aigina (Denkmiiler antikerArchitektur16), Berlin 1985

ATK Buschor, 1957 N. Cooper, 1983

6

ABBREVIATIONS

TdA I, II

= E. Buschor, Die Tonddcher der Akropolis, I, Simen, II, Stirnziegel, Berlin 1929-1933

Vlassopoulou

= C. Vlassopoulou, Decorated Architectural Terracottas from the Athenian Acropolis:

C. Wikander, 1986

= C. Wikander, Sicilian Architectural Terracottas. A Reappraisal (Skrifter utgivna av SvenskaInstituteti Rom, 80, 15), Stockholm1986

Catalogueof Exhibition, Princeton 1989 (also appears in Hesperia 59, fasc. 1, 1990)

0. Wikander,1-988 = 0. Wikander,"AncientRoof-tiles-Use and Function,"OpAth17, 1988, pp.203-216 = C. K. Williams, II, "Demaratusand Early CorinthianRoofs,"in 17tj-r. TOplOSels Williams, -TT']\1 NutKoXaov KoVroXAeoVro,Athens 1978, pp. 345-350 lAvri,rlT

GLOSSARY Acrotere.See Akroterion. AyeXaL'a KepatLs.See Plain tile. See Antefix. AKpoKe'pap,o. AKROTERION (Fr. acrotere, Germ. Akroterion, Gk. aKpOTnptov). A decorative element, such as a

statue, placed at the edges of the roof, either at the cornersof the pedimentor along the ridge. See Ovolo. A1AV'yaXo. AvayAv7rra fv4XXa.See gadroon.

Antefissa.See Antefix. ANTEFIX (Fr. antefixe, It. antefissa, Germ. Stirnzie-

gel, Gk. aKpoKepapo).The plaque which closes

the lower end of the bottommostcover tile of each row of overlapping cover tiles running from the ridge down to the eaves. ANTHEMION.A floral chain pattern.

Baguette. See Torus. Bandeau.See Fascia, Taenia. Bastoncino.See Torus. Bec de corbin.See Hawksbeak. Becco di civetta.See Hawksbeak. Blattstab.See Tongue pattern. Blattstabsima.See Cavettosima. Canale di gronda.See Waterspout. Cavet. See Cavetto. CAVETTO(Fr. cavet, It. gola, Germ. Kehle, HohlA concave molding. kehle, Gk. KoLAXcoJLa). CAVETTOSIMA(Fr. sima en cavet, Germ. Blattstab-

sima). A rakingsima with cavettoprofile,generally carryingthe painted decorationof a tongue patternabovea single or doubleguilloche. Chainette. See Guilloche. Colmo. See Ridge. TILE (Germ. Flach- und Deckziegel, COMBINATION Gk. KEpap,se7rTLvyos). A tile which has the pan

and coverelements manufacturedin one piece. Corinthiansima. See Ovolo sima. Couvre-joint.See Cover tile. COVERTILE (Fr. couvre-joint, Germ. Deckziegel, Gk. KaAv7rTTpas). The tile designed to cover

the space between adjacentpan tiles, to protect the interstices and the wooden beam below; it sits on top of the two pan tiles.

Dachziegel. See Roof tile. Deckziegel. See Covertile. Doric leaf. See Tongue pattern. Doric tongue. See Tongue pattern. EAVES(Fr. egout, rive, Germ. Trauf). The bottom edge of a sloped roof. EAVESTILE (Fr. tuile d'egout or de rive, Germ. Traufziegel

or Traufrand,

Gk. o-TpwT7pas

The bottommostpan tile of each row -qyE?oWv). of overlappingpan tiles running from the ridge down to the eaves. Ecailles. See Scale pattern. Egout. See Eaves. See Volute. 'EXLKaS.

Emissaired'eau. See Waterspout. Engobe. See Slip. E7raeTs O-4l,J.See Raking sima. E7re'bvvrvo.See Revetment plaque. See Slip. E7ridXpLrua.

Faite. See Ridge. Falz. See Flange. FASCIA (Fr. bandeau, It. listello di base or di zoccolo,

Germ. Plattenborte, Sockel, Gk. aro-a).The flat, verticalpart of the plaque at the bottomof a sima. Feuille dorique.See Tongue pattern. Filet. See Taenia. First. See Ridge. Firstpalmette.See Ridge palmette. Firstziegel. See Ridge tile. Flach- und Deckziegel. See Combinationtile. Flachziegel. See Pan tile. FLANGE (Fr. recouvrement, Germ. Falz). A tongue

projecting from one edge of a piece (pan tile, raking sima) to overlapthe next lower piece on the slope. Flechtband.See Guilloche. Frontleiste.See Taenia. GADROON

(Gk. avayAv7rra fvXA\a). A tongue in

convexrelief. Gargouille. See Waterspout.

GLOSSARY

8

GEISONTILE (Germ. Geisonziegel). An eaves tile

with a curvedsoffitforminga drip. Geisonziegel.See Geison tile. GewohnlicheZiegel. See Plain tile. Giebelsima.See Raking sima. Gola. See Cavetto. GUILLOCHE (Fr. chainette, tresse, It. treccia, Germ. A cable pattern. Flechtband,Gk. 7rAoXoXM). Half-round. See Torus. HAWKSBEAK (Fr. bec de corbin, It. becco di civetta, Germ. Uberschlagskarnies, Gk. KovKovf,ayta).

A concaveprofile which resemblesthe beak of a hawk. HIP ROOF(Fr. toit a croupe or a aretier, Germ. A doubleWalmdach, Gk. o-reyry TpLp'lXrT7). roof with a third at one short end sloped slope in place of a pediment. Hohlkehle. See Cavetto. KaXov7tir.See Mold.

KaAv7rrT rpas. See Covertile. Kehle. See Cavetto. See Raking sima. E7raETrs. KepaCigs Kepa,usl7rapaerLT.See Lateral sima. See Combinationtile. KepapIlsE7rLCvyos. Kepa/LisKopvfala. See Ridge tile. See Roof tile. rs o-re-yrys. KepaLt'Is See Cavetto. KoLCXwOia. Kopv4'. See Ridge. KovKov/3ayta. See Hawksbeak. See Molding. KViALTLo. Languette.See Tongue pattern. Lastra. See Plaque. Lastre. See Plaque. LATERALSIMA(Fr. sima laterale, sima d'egout or de

rive, It. sima laterale, Germ. Traufsima, Gk. 7rapaeTist crLM7r,KepaMiL 7rapaeTIV). The sima which runs along the sides of a roof at the bottom of the slope. Linguetta.See Tongue pattern. Listel. See Taenia. Listello di base. See Fascia. Listello di coronamento.See Taenia. Listello di zoccolo.See Fascia. Listello superiore.See Taenia. Aovpita. See Taenia.

Matrix. See Mold. Matrize. See Mold. Megarian sima. See Ovolo sima. M?rpa. See Mold. MOLD(Fr. matrix, moule, Germ. Matrize, Gk. ,IjTpa, KaXovTL). The form into which clay is

pressedto manufacturemultiple copies. (Fr. moulure, It. sagoma, Germ. ProfilA contouredornamenton leiste, Gk. Kv/uaaTLo). the visible face of a piece. Moule. See Mold. Moulure. See Molding.

MOLDING

-rovAos. See Terracotta. OT7r6 Ove. See Ovolo. OVOLO (Fr. ovolo, It. ovoloor ove, Germ. Wulst, Gk. A convex molding forming almost aju1vybaXo). one quarterof a circle. OVOLOSIMA (also Corinthian or Megarian sima,

Germ. Wellensima). A raking sima with an ovolo profile above a fascia, over which is generally painted an anthemion. PAN TILE (Fr. tuile courante, Germ. Flachziegel, Gk. mrTpwTrjpaS). The tile designed to carry

rain water off a sloped roof. HIapaerisipor4j.See Lateral sima. PLAINTILE (Germ. gewohnliche Ziegel, Gk. ayeXala KEpa,isA).The ordinary tiles of the roof

slope. HXdaKa.See Plaque. PLAQUE(Fr. lastre, It. lastra, Germ. Platte, Gk. 7rXaKa).A vertical element which can be sepa-

rate or attachedto one end of a covertile (forming an antefix) or a pan tile (forminga sima). Platte. See Plaque. Plattenborte.See Fascia. o. See Guilloche. mIAoyx Profilleiste.See Molding. Pa/38oyXvfo. See Torus. RAKINGSIMA(Fr. sima de rampant, sima rampante,

It. sima di rampante, sima rampante, Germ. Giebelsima,Schragsima,Gk. E7raerT?lr4A7r,KEpadl's E7TaErTL).The sima which runs along the

edge of the pediment (the rakes). Recouvrement.See Flange.

GLOSSARY Revetement.See Revetmentplaque. REVETMENT PLAQUE(Fr. revetement, It. rivestimenGerm. to, Verkleidungsplatte, Gk. e7rEvbv0-?)).

A plaque which is generally nailed to a wooden element to protectit. RIDGE(Fr. faite, It. colmo, Germ. First, Gk. Kopvrin). The apex of a sloped roof, at the ridge beam. RIDGEPALMETTE (Germ. Firstpalmette). An updecorated with a palmetteon both right plaque to the attached faces, ridge tile. RIDGE TILE (Fr. tuile faitiere, Germ. Firstziegel, Gk. Kepacld Kopv(ala). The uppermost pan and/or cover tile which overlapsthe ridge of a sloped roof.

Rive. See Eaves. Rivestimento.See Revetmentplaque. Roll. See Torus. ROOFTILE (Fr. tuile, It. tegola, Germ. Dachziegel, Gk. K?pap,tlsr7ts orreyqy). A piece of fired clay, generally laid on the roof to overlap and be overlappedby another, to protect the building below.

Rundstab.See Torus. Sagoma. See Molding. SCALE PATTERN (Fr. ecailles, It. scaglie, Germ. Schuppenmuster, Gk. fvAX\a foXtbOdra).A painted design resembling overlapping fish scales.

Schragsima.See Raking sima. Schuppenmuster.See Scale pattern. Scaglie. See Scale pattern. SIMA. A pan tile with a vertical plaque attached

along one edge to divertrain water off the roof. Sima d'egout.See Lateral sima. Sima de rampant.See Raking sima. Sima de rive. See Lateral sima. Sima di rampante.See Raking sima. Sima en cavet. See Cavettosima. Sima laterale. See Lateral sima. Sima laterale. See Lateral sima. Sima rampant.See Raking sima. Sima rampante.See Raking sima.

Sockel.See Fascia. Erey,? TpLplXTr?.See Hip roof.

Stirnleiste.See Taenia. Stirnziegel.See Antefix. See Pan tile. -rpworTpaS. TrpwT7rpasr?ye/ucov.See Eaves tile. TAENIA(Fr. bandeau, filet, listel, tainia, It. listello di

coronamento, listello superiore, Germ. Stirnleiste, Frontleiste, Gk. 'ralvia, Xovplba).A flat band running across the top of a sima or the front edge of an eaves tile. Tainia. See Taenia. TaLvla.See Taenia. Tegola. See Roof tile, Tile. TERRACOTTA (Fr. terre cuite, It. terra cotta, Germ. Fired clay. Terrakotta,Gk. OTTOd wXAods). Terre cuite. See Terracotta. TILE (Fr. tuile, It. tegola, Germ. Ziegel). See Roof tile. Toit a croupeor a aretier.See Hip roof. Tondino. See Torus. TONGUEPATTERN (also Doric leaf or tongue, Fr. or feuille dorique, It. linguetta, languette Germ. Blattstab, Zungenmuster,Gk. fvAXxo). A row of panels with roundedends at the top. Tore. See Torus. TORUS (also half-round, roll, Fr. baguette, tore,

It. bastoncino, tondino, Germ. Rundstab, Wulst, Gk. pa/,b6'yAvfo).A convex,half-round molding. Trauf. See Eaves. Traufrand.See Eaves tile. Traufsima. See Lateral sima. Traufziegel. See Eaves tile. Treccia. See Guilloche. Tresse. See Guilloche. Tubo di gronda.See Waterspout. Tuile. See Roof tile. Tuile courante.See Pan tile. Tuile faitiere. See Ridge tile. Tuile d'egout.See Eaves tile. Tuile de rive. See Eaves tile.

SLIP (Fr. engobei Germ. Uberzug, Gr. e7rlXpl-r,uLa).

A thin layer of purified clay spread in liquified form across the surface of the piece before firing.

9

Uberschlagskarnies.See Hawksbeak. Uberzug. See Slip. Y8poppo'?.See Waterspout.

10

GLOSSARY

Verkleidungsplatte.See Revetmentplaque. Voluta. See Volute. VOLUTE (Fr. and Germ. volute, It. voluta, Gr. EAtKas).A band which coils into a spiral. Io-ra.See Fascia. 4>vAAafoXtbora. See Scale pattern. See Tongue pattern. 1X>vAAo.

Wasserspeier.See Waterspout. WATERSPOUT (Fr. emissaire d'eau,gargouille,It. canale or tubo di gronda, Germ. Wasserausguss, Wasserspeier, Gr. vbpopport).The opening through a lateral sima or akroterionbase at the eavesthroughwhich rain water is discharged. Wellensima. See Ovolo sima. Wulst. See Ovolo, Torus.

Walmdach.See Hip roof. Wasserausguss.See Waterspout.

Ziegel. See Tile. Zungenmuster.See Tongue pattern.

Ladies and Gentlemen, On behalf of the American School of Classical Studies, I would like to welcome you to the School this morning to this First International Conference on Archaic Greek Architectural Terracottas. The conference is sponsored by the American School, and I thank you all for coming.

This conferenceis important for two reasons. First of all, it is the first international conferenceon Archaic Greek architecturalterracottasto be held anywhere. As such, it brings togethera truly internationalgroup of scholars.The talks reflectthe wide variety of work that is presently being accomplishedin the study of Greek terracottas.In the past, finds of terracottashave been relativelyneglectedby excavators,but the situation is beginning to change,as this conferencereflects.Great interestis now being shown in the field, for, in the Archaic period especially,the presenceof architecturalterracottasoften providesthe only evidencefor the existence of buildings. Further, the typologies now being established can provide criteria for dating. On a broaderscale, the study of early terracottasprovides material that adds to the discussions on the developmentof the form and material of the Greek temple and its decorativeadornment. The papers of this conferencerange widely and cover most of the ancient world from mainland Greece and the islands to Ionia and the western colonies. The concentrationis, however, on mainland Greece. The first day containspapers on material from sites principally in the Corinthiaand the Argolid.The terracottasfrom Delphi and Kalapodialso play an importantpart in the program.The secondday expands the topic to include the Aegean islands and sites furtherafield, such as Didyma, Histria, and Morgantina.The topic has by no means been exhausted, but we hope that we have made a start by bringing together scholarsfrom the United States and Europe to presentand discusstheir material and ideas. Since the material presented in the papers is of the utmost importanceand urgently needs to be brought before the public in order to further scholarly research, we plan to publish the proceedingsof the conferencein a forthcomingissue of Hesperia. We hope that the discussionsscheduledafter every one or two papers will be extensive and will also contribute much to increaseour knowledgeof the developmentof Archaicterracottas. The two excursions that follow the conferencehave been planned so that sites where importantdiscoverieshave been made can be visited. The first trip goes to Corinth, Nemea, Argos, and Tiryns, and the second to Delphi and Kalapodi. You will, therefore, have a chanceto examine the physical remainsafter the presentationof the evidencein the papers. The second reason for the importanceof this conferenceis that it is the first one to be held by the AmericanSchooland thus marksan importantevent in the historyof the School. Since membersof the AmericanSchoolhave long been involvedin the study of architectural terracottas,I cannot think of a more appropriatesubjectfor the School'sfirst international conference.I believe that the School should play a more active role in bringing together

WILLIAM D. E. COULSON

12

scholars from different parts of the world, and we hope to invite you back for other conferencesin the future. We have in mind a second internationalconferenceon the Archaic architecturalterracottasof Magna Graecia. The conferencecould never have been made possible without three factors. First and foremost is the anonymous donor who has providedthe funds for the conference;to this donorwe must remain deeply grateful.Secondis the initiativeof the librarianof our Blegen Library,Dr. Nancy Winter, whose idea this conferencewas and who has taken chargeof its organization.As you can appreciate,this has entailed a lot of work, and on behalf of the School, I would like to thank Nancy Winter for making it all happen so smoothly. And third, I would like to thank the staff of the Schoolwhich has workedextremelyhard for this conference. Dr. Robert Bridges, the School Secretary, has supervised the audio-visual equipment. Mrs. loanna Driva, our business manager, has handled all the financial arrangements;the Director's secretary,Mrs. Maria Pilali, has done much of the secretarial work; and our principal receptionist,Christina Traitorou, has done the translationsof the abstractsfound in the conferencepackets.To all of these and to the other staff of the School whose names it would be impossiblefor me to mentionindividually,we are all very grateful. The staff of the Schooljoins me in wishing this conferenceevery success. WILLIAMD. E. COULSON

Director AMERICAN

SCHOOL

OF CLASSICAL

54 Souidias Street GR-106 76 Athens, Greece December 2, 1988

STUDIES

AT ATHENS

DEFINING REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS (PLATES 1, 2) T

WO MAIN SYSTEMSof tiled roofshavelong beenknownto have dominatedthe

architectureof the ancient Greek world: they are called Laconian and Corinthian, names which are documentedin epigraphical sources of Classical and Hellenistic date.1 Laconian roofs (Figs. 1, 2) are characterizedby large, concavepan tiles which drain rainwater from the slopes and narrower, convex cover tiles, so called because they cover the spaces between pan tiles. Corinthian roofs (Figs. 3, 4) have relatively flat pan tiles and pitchedcovertiles. This simple and straightforwarddistinctionbetweenthe two systemshas long been acceptedas providinga clear pictureof the tiled roofsof ancientGreece,and every roof with curvedtiles has been labeled as Laconian,while those with flat and pitched tiles are called Corinthian.I would like to suggest, however,that the picture during the Archaic period is considerablymore complex and that various regions evolved their own roofing systemswhich are characterizedboth by technicaland by decorativefeatures.The situation has been obscured by the fact that many roofs are found outside their originating city or region. In order to clarify the picture, one must determinethe features that characterize each roof within a given city and then associate the roofs of identical type which occur elsewhere. In general a regional picture emerges, except for the examples found further afield in internationalsanctuariesor in places with politicalties to the originatingcenter.In sanctuaries,small buildings such as treasuries often used roof decorationcharacteristicof the dedicatingcity, probablyso that it would be readily identifiable. The Laconiansystemis well documentedin the numerousfragmentsfound in the excavationsof the sanctuaryof ArtemisOrthia at Sparta,many of them stratigraphicallydatable thanks to the flooding of the near-by river, an event which was probably greeted by the ancient Spartans with less enthusiasm than by modern archaeologists.These fragments demonstratethat Laconian-systemroofs, as known at Sparta, have a limited range of elements. The pan tiles are large and concave,the covertiles narrowerand convex, the ridge I First identifiedby W. Dorpfeld, "Die Skeutothekdes Philon,"AM 8, 1883 (pp. 147-164), p. 162; Martin, pp. 67, 70-72; Orlandos,pp. 82-83. I would like to thank Kostis Iliakis, who made the drawingspublished here. In no case is the entire roof preserved, and so the reconstructionshave been made accordingto the available information. Most often lacking are the plain tiles and the full widths of the raking simas and eaves tiles. Frequently cited works are abbreviatedas follows: Daux = G. Daux, Guide de Thasos,Paris 1968 Dawkins = The Sanctuaryof ArtemisOrthiaat Sparta (JHS, Suppl. 5), R. M. Dawkins, ed., London 1929 Launey = M. Launey, Etudes thasiennes, I, Le Sanctuaireet le culte d'Herakles a Thasos, Paris 1944 Rhomaios, 1957 = K. A. Rhomaios, <>,'ApX'Eo 1957, pp. 114-163

14

NANCY A. WINTER

o

FIG.

~~~~~0,50

1. Roof of the Heraion at Olympia. Drawing by K. Iliakis

tiles convexwith openings for the covertiles; all are paintedwith blackor red glaze.2At the apex of the pediment, capping the end of the ridge tile, sat disk akroteriawith numerous molded elements;the earlier examples are monochromeblack, the later ones polychrome, 2 Dawkins, pp.

142-143 (pan and covertiles), no. 37, p. 141, fig. 100 (ridge tile).

REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

Filln

I

K

FIG.2. Roof of the Heraion at Olympia. Drawing by K. Iliakis

15

NANCY A. WINTER

16

0

0,50

FIG. 3. Roof of the secondTemple of Apollo at Corinth. Drawing by K. Iliakis

and they carrymany of the same patternsfound on Laconianpottery,a fact which helps in dating them (P1. 39:a).3Along the edges of the pediment ran a colorful raking sima. The bottommostrow of cover tiles along the eaves consisted of curved antefixes with painted designs of primarily geometricpatterns (P1. 1:a); moldedtoruses were eventually addedto an otherwise flat, painted face.5Occasionallyantefixes were coveredby a lateral sima, an I

Dawkins, nos. 15-19, p. 137, fig. 90 (monochromeakroteria);nos. 1-14, pp. 135-137, figs. 87, 88,

pls. XXII-XXIV (polychrome akroteria). 4

Dawkins, nos. 30-33, pp. 139-140, pl. XXVI.

5Dawkins, nos. 20-28, pp. 137-139, figs. 92-94, pl. XXV.

FIG. 4. Roof of the secondTemple of Apollo at Corinth. Drawing by K. Iliakis

18

NANCY A. WINTER

unusual concept.6Geison tiles seem neverto have been an essentialpart of the canonicalLaconianroof, as only one fragmenthas been publishedfrom the Orthia excavations.7 Sparta has producedclose to half the known examples of disk akroteriaand is thought to have manufacturedmost of the similar examples found elsewhere, at near-by Amyklai and the Menelaion, and further afield at Kynouria, Bassai, and Olympia (P1. 39:b, c).8 Antefixes with crescent patterns, the earliest style of Laconian antefix, occur only fairly close to Sparta, at the Menelaion, Amyklai, and Epidauros Limera in Lakonia, and just north of Lakonia at Kynouriaand Bigla near Tegea.9 Other types of decorationnot documentedin Sparta, however,occur on roofs with Laconian-styletiles. Should these be considereda manifestationof the same system, which an accidentof history has left unrecordedin Sparta itself? I think not. A series of small disk akroteriawith proportionatelyheavy molded decoration,often thought to be antefixes because of their scale, are found on roofs in Arkadia,for example, and may indicatea regional variation. They occur at Bigla and Methydrion, with an archaizingversion from Lusoi.10 An Arkadian regional system may also favor curved antefixes with molded figural decoration, not documentedat Sparta, such as the heraldic sphinxes from Bassai discussedby Nancy Cooper (pp. 87, 89 below, figs. 19, 20), and the Gorgoneiafrom Alipheira.IIAnother feature of the Arkadian system may be a geison tile with a cavettoprofile and painted tongue pattern, examples of which have been found at Bigla, at Alipheira along with the Gorgoneionantefixesjust mentioned,and at Olympia, where they occur on the Bouleuterion togetherwith semicircularantefixes also decoratedwith Gorgoneia.12 The same approachto the Corinthiansystem produceseven clearerresults. I will give here only the most general outline of Corinthian roofs at Corinth, as Mrs. Roebuck discusses these pieces in greater detail (pp. 47-49 below). Terracottaroofs at Corinth13down 6 Dawkins, no. 33 A, B, p. 140, fig. 98. 7 Dawkins, no. 36, p. 141, fig. 99, shown upside down. This fact might indicate,however,that geison tiles were not recognizedas such and thereforewere not saved. 8 H. Lauter-Bufe, "Entstehungund Entwicklungdes kombiniertenlakonischenAkroters,"AM 89, 1974 (pp. 205-230), pp. 213-214. 9 Menelaion: Dawkins, p. 122, fig. 95. The Amyklaion examples are more canonical,though also larger than most of the Sparta antefixes: W. von Massow, "Vom Amyklaion,"AM 52, 1927 (pp. 34-63), p. 43, fig. 22; Dawkins, no. 11, p. 122. Epidauros Limera: Koch, p. 95, fig. 46. Kynouria:K. A. Rhomaios,<JIEpLohELa Kara r712VKvvovpL'av>>, fIpaKrLKa1953 (pp. 250-257), p. 253, fig. 2. Bigla: Koch, pp. 87-88; Rhomaios, 1957, pp. 117-118, fig. 3; N. D. Papahatzis, IIavo-av'a, aAXLK KaL APKabLK& Athens 1980, p. 382, fig. 419. The same fragmentsare publishedas being from Pallantion:BCH 83, 1959, pp. 626-628, fig. 18. 10 Bigla: Rhomaios, 1957, pp. 118-119, figs. 6-7; Papahatzis (footnote 9 above), p. 381, fig. 418. Methydrion:F. F. Hiller von Gaertringenand H. Lattermann,ArkadischeForschungen,Berlin 1911, pp. 35-36, figs. 9, 11; GFR, p. 180, no. 8, fig. 107 center. Lusoi: W. Reichel and A. Wilhelm, "Das Heiligthum der Artemiszu Lusoi,"OJh4,1901 (pp. 1-89), pp. 61-62, fig. 128; Koch,pp. 90-91; GFR, p. 180, no. 6; LauterBufe (footnote8 above), p. 213, said to be too small for an akroterionand classifiedas an antefix. ' A. Orlandos, 'H 'pKa8LK % 'Aa'4ELpa KaLT% , Athens 1967-1968, pp. 78-79, 112-113, rplAC figs. 52, 85-87. 12 Bigla: Rhomaios, 1957, p. 123, fig. 14. Alipheira:Orlandos,op. cit., p. 114, fig. 88. Olympia:OlympiaII, pi. CXVI:3-4. 13 The generalizationsgiven here refer to the overall patternpresentedby the majorityof pieces at Corinth. Exceptionsto the patternare the antefixesof two roofswhich I would classifyas part of the Argive system (see

REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

19

a

0

b FIG. 5. Roof 9 from Delphi. Drawings by K. Iliakis

~~~~o,so

20

NANCY A. WINTER

o

0,50

FIG.6. Roof of the Megarian Treasury at Olympia. Drawing by K. Iliakis

to the third quarter of the 6th centuryB.C., if I am correct,are characterizedby the use of combinationtiles, that is, tiles in which the pan and coverelementsare formedas one piece. As mentionedbefore, Corinthianpan tiles are relativelyflat, while the cover tiles rise to a peak at the center. During the first half of the 6th century B.C. (Fig. 5:b), the decoration below, Fig. 9:a, b, and P1. 5, FA 547 and FA 24). The first of these antefixeswas found in the Demeter Sanctuary and so could representthe dedicationof a building by a pious Argive. The secondrtype comes from later fill in the city site and cannot be associated with a specific building or sanctuary, but virtually identical examples come fromArgos, Delphi, and the AthenianAkropolis,so that Corinthdoes not have exclusiveclaim to the type.

FIG.7a FIG. 7. Roof of the Megarian Treasury at Olympia. Drawing by K. Iliakis

22

NANCY A. WINTER

0

0,50

FIG.8. Roof from the sanctuaryof Aphaia on Aigina. Drawing by K. Iliakis

along the slopes of the pedimentconsistsof a raking sima with a cavettoprofile,on which is painted a tongue pattern abovea single or double guilloche. Along the eaves of the roof run pentagonal antefixes with floral motifs molded in relief and painted,joined to eaves tiles carryinga paintedguilloche (P1.5, FA 101). There is no decorationalong the ridge. During the second half of the 6th century (Fig. 4), a new style of raking sima consists of an ovolo moldingabovea flat fascia;the face of the sima is decoratedwith a doublechain of palmettes and lotus flowers, rendered in paint only (P1. 6, FS 101). At the lower corners of the pediment, akroterionbases are attached to the back of the sima, with drainage from the slope provided through lion's head spouts. Along the eaves are antefixes whose plaques projectabove the cover tile and take the form of a lotus and palmette (P1. 7, FA 3) or palmette and double volute (P1. 6, FA 16), molded in relief. These antefixes are attached to eaves tiles decoratedwith a painted guilloche (P1. 6, FT 231) or, later, a maeander.The

REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

23

ridge covertile is now surmountedby a plaque painted on both faces with a palmette and double volute (P1. 6, FR 3). By the Late Archaic period (Figs. 6, 7), the plain tiles of the slope are separately made pan tiles and cover tiles, but the pieces at the ridge and eaves continueto be made as combinationtiles. Therefore,these combinationtiles are a consistent featureof roofs in the pure Corinthiansystem. The Corinthian system as it is documentedat Corinth thus shows a coherent style. Enough material is preservedthere to suggest that it is representativeof the Archaic production of this very important center. Many examples of roofs consistent with the style found at Corinth are found outside the Corinthia, primarily at sanctuariessuch as Delphi and Olympia. Once again I would suggestthat the roofsfoundoutsideCorinthwhich do not conformto the style as seen in Corinth itself shouldnot be consideredpart of the Corinthian systemproper. In some cases similar clays have been used, and the roofs were possiblyeven executedby workmentrained at Corinth, but the designs and conceptsare foreign to those of the pure Corinthiansystem producedat Corinth. The roofs of Aitolia, from its famous sanctuariesat Thermon and Kalydon, have long been consideredto represent stages in the developmentof the Corinthian system, stages which, by chance, are not documentedin Corinth itself. Dr. Madeleine Mertens-Horn has disputedthis interpretation.14 I too preferto see the Aitolian roofs as examples of a regional roofing system with its own characteristics.The pan and cover tiles are in most cases separately made, even at the ridge and eaves, a completedivergencefrom the practiceat Corinth. In addition,the region shows a distinctpreferencefor figural decorationin the form of human heads applied to the front of the antefix and lion's heads used as spouts all along the eaves.15The close grouping of antefixes and waterspoutson the two successiveroofs from Thermon creates the effect of a continuous sima along the eaves, a concept which is formalized into the only known Archaic lateral sima from Greece, at Kalydon.16Not only is this conceptdifferentfrom that of roofs in the pure Corinthiansystem, but most important, no similar pieces exist at Corinth. I would also separate from the Corinthian system the roofs from the Argolid. Dr. Nancy Cooper first suggestedin her master'sthesis on the Halieis roofs that a special regional systemexisted in the Argolidand on Aigina, an island with politicalties to Argos,but later in her doctoraldissertationshe viewed the style as part of the Corinthian system.17I agree with her earlier assessment.Typical of this Argive system, as she had noted, are the separatelymade pan and covertiles, includingthe elementsof the ridge and eaves. The tiles consistof flat pan tiles and pitchedcovertiles similar to those of the Corinthiansystem, but the fact that they are always separatelymade precludestheir belonging to the Corinthian 14

Mertens-Horn, 1978, pp. 30-31. See especially G. Kawerau and G. Sotiriades,"Der Apollotempelzu Thermos,"AntDenk 2,1902-1908, pp. 1-8, pls. 49, 53, 53A. 16 Dyggve, pp. 169-190, 214-225; Rhomaios, pp. 53-80. 17 N. K. Cooper, "Three Roofs from the Sanctuary of Apollo: Halieis," M.A. thesis University of Minnesota, 1977, pp. 7-30; eadem, The Developmentof Roof Revetment in the Peloponnese,diss. University of Minnesota, 1983 (SIMA, forthcoming). 15

NANCY A. WINTER

24

LtS?2 a

b

c

d

e

FIG. 9. Typology of Argive system antefixes. Drawing by K. Iliakis

REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

FIG.

25

10. Kalapodiantefix. Drawing by K. Iliakis

system. The earliest roofs of this system, examples of which are found in the Argolid at Halieis (P1. 9:b), Mases, and the Argive Heraion (discussed by Mr. Pfaff [pp. 149-156 below, P1. 12]), and at the sanctuaryof Aphaia on Aigina (Fig. 8), with similar pieces at Delphi and Olympia,18have simple undecoratedantefixes which rise at the top to three peaks. Unlike the Corinthiansystem, a nearly contemporaryexample of which is shown in Figure 5, the antefixesare not attachedto the eaves tile but insteadsit on top of the upwardcurvingside edges of the underlyingeaves tiles and thus have a finishedbottomedge with a characteristicdouble curve. This same technical feature can be observed on a series of antefixes (Fig. 9:b) from Delphi, the AthenianAkropolis(Vlassopoulou,nos. 5, 6), Corinth (P1. 5, FA 24), and Argos (Billot, p. 105 below, fig. 2, P1. 1O:e,f), where the three peaks have now been enlargedto carry a stampeddesign. A terracottamodel of a roof of this type comes from the Athenian Akropolis (see Vlassopoulou,no. 10 bis); it has a hipped end like that of the early Temple of Zeus at Nemea which carried a similar series of antefixes (Fig. 9:c). A furtherstage in the developmentof this regionalstyle occursboth at Nemea and at Tiryns (Fig. 9:d). Dr. Billot (pp. 105-107 below) describesthese and other examples of this systemwhich preservethe three peaks at the top and the doublecurvealong the bottom, even when the decorationon the plaque face changes. Equally characteristicmay be the raking simas, which prefera flat verticalface to the cavettoof the Corinthiansima, as noted by both N. Cooper (p. 68 below) and Billot (p. 130 below, fig. 7, and P1. 10:d). For references, see the papers by C. Pfaff (pp. 149-156 below), N. Cooper (pp. 65-93 below), and J. Heiden (pp. 41-46 below). Note that some of the tiles from the Olympia roof are dated by context no later than the third quarterof the 7th centuryB.C. 18

26

NANCY A. WINTER

a

M.~~~~~~~~~~~~~~~~M

0

b FIG. 11. Roof 12 from Delphi. Drawings by K. Iliakis

0,5f0

REGIONAL STYLES IN ARCHAIC GREEK ARCHITECTURAL TERRACOTTAS

0

27

0,50 M. FIG.

12. Roof of the temple at Halai. Drawing by K. Iliakis

Anotherregional system with some of the same technicalfeaturesfound in the Argolid, but with a differentdecorativevocabulary,occurs in Central Greece in the areas of Thessaly, Phthiotis, Lokris, Boiotia, and possibly Euboia. Connectionsbetween these areas and Aigina, which shares the early Argive system, have already been shown by Dr. Felsch's publication of tile stamps,19and so this similarity in roofing systems should not come as a surprise. As in the Argive system, the plain tiles are separately made, flat pan tiles with pitched cover tiles, and the antefixes have a double curve along their bottom edge. The earliest example of this system thus far known occursat Kalapodi;it is discussedbelow by Dr. Htibner (pp. 167-174). Although the antefixeshave a decorationwhich is unparalleled in the Argolid, they rise at the top to three peaks and have the double curve along their bottomedge which allows them to straddletwo eavestiles (Fig. 10). One of the raking simas from this site (P1. 16:c) also has a flat verticalprofile like some found in the Argolid. Two nearly identical roofs from Delphi (Fig. 11) and the Athenian Akropolis20 may also belong to this Central Greek system, although they bear resemblancesto both the 19

Felsch, 1979, pp. 1-40. Delphi, Roof 12: Le Roy, 1967, pp. 45-47, pls. 7, 98. AthenianAkropolis:TdA I, pp. 8-10, Sima II and III, figs. 1, 9; II, p. 35, Antefix V, fig. 46, pl. 2; Vlassopoulou,nos. 20-23. 20

28

NANCY A. WINTER

Corinthianand Argive systems. Both have a raking sima with a cavettoprofile and painted tongue pattern (Vlassopoulou,nos. 21-23), characteristicof Corinthian-systemroofs of the first half of the 6th centuryB.C. Both also have palmetteand double-voluteantefixes (Vlassopoulou,no. 20a-b) which can be paralleledin the Corinthiansystem.The Delphi antefix, however,preservesthe double curvealong the bottomedge that indicatesthat the tiles of the slopes and eaves edge were separatelymade.A comparisonbetween the Delphi roof and the nearly contemporaryTemple of Apollo at Corinth (Figs. 3, 4) demonstratesthe differences between the two systems.I would suggestthat the same Greek city dedicatedboth Treasury XV at Delphi, aroundwhich the fragmentsof this roof were found, and the small building on the Akropolis. The suggestion that this city might be located in Central Greece is based on the existence of some very similar roofs of later date from that area, the best preservedof which comes from Halai in Lokris (Figs. 12, 13, P1. 1:b-d), dated ca. 490 B.C.21 Although the raking sima, ridge palmettes, palmette and double-voluteantefixes, and eaves tiles with a painted guilloche again resemble those of the Corinthian system, the plain tiles and the ridge and eaves tiles are separatelymade and so cannotform part of the Corinthiansystem. Two examples of the antefix, it must be admitted,are attachedto eaves tiles, which would qualify them as being part of the Corinthiansystem;22both the width of the antefix and the height of the eaves tiles differ fromthose of the other examples,and so these pieces cannotbe part of the original roof.23I would view this roof as the continuationof the system encounteredhalf a century earlier at Delphi and Athens, and even earlier at Kalapodi. Very similar antefixes have also been found at Theotokou in Thessaly and on Skyros, where separatelymade pan and covertiles were also excavated.24A ridge palmettefrom Kyparissi in Lokris25can also be consideredan element of this Late Archaicregional system. Attica, apart from the Athenian Akropolis,has producedsurprisinglyfew examples of Archaicarchitecturalterracottas.Althoughthe Akropolismaterialmay not be characteristic of Attica in general, it does display some peculiaritieswhich might indicate a local system. The earliestexamples belong to the pure Corinthianand Argive systemsrespectively(Vlassopoulou, nos. 1-4 and 5-7), after which time the roofs show a more distinctivestyle, not 21 A.

L. Walker and H. Goldman, "Reporton Excavationsat Halae of Locris,"AJA 19, 1915 (pp. 418437), p. 433, figs. 8, 9; GFR, pp. 85-86, no. 51, 148, no. 20, fig. 97; H. Goldman, "The Acropolisof Halae," Hesperia 9, 1940 (pp. 381-514), pp. 440-442, nos. 1, 4, 5, 7, figs. 100, 102, 104. 22 Goldman,Op.cit., p. 440, no. 3, fig. 103. 23 These attached antefixes are very similar to those of the stoa from Opous, discussed below by Mrs. Dakoronia (pp. 175-180); perhaps they decorateda stoa in the sanctuary at Halai, somewhere near the temple. 24 Theotokou:A. J. B. Wace and J. P. Droop, "Excavationsat Theotokou,Thessaly,"BSA 13, 1906-1907 (pp. 309-327), p. 314, fig. 5; GFR, p. 148, no. 17. The pieces describedand illustratedin GFR, pp. 147-149, nos. 16 and 21, figs. 101 and 102, are identicalto the fragmentsfromTheotokouand either are mistakenlysaid to be from Gonnoi or the wrong photographs have been substituted for the Gonnoi examples. Skyros: D. Euangelides,<<'Ava-xKaa'L KaLepevvaL Ev IKVpcp>>, AeAr 4,1918 (1921), Hapapnrila (pp. 34-45), pp. 3637, figs. 4, 5. 25 AeXT 33, 1978, B' 1 (1985), p. 140, pl. 47:e.

FIG.

13. Roof of the temple at Halai. Drawing by K. Iliakis

30

NANCY A. WINTER

0

0,50

FiG. 14. Roof of the Polygonal Building in the sanctuaryof Herakles on Thasos. Drawing by K. Iliakis

representedelsewhere. The cavettoraking simas with painted tongue and guilloche (Vlassopoulou,nos. 11-15) have long been recognizedto be under Corinthianinfluence,but they possess some individualtraits, such as the use of cornerakroteriaattachedto the top edge (Vlassopoulou, no. 13), which is a divergencefrom canonical Corinthian-systempractice but may be typical for Athens: several corner-simafragmentshave insets along the back edge for the insertionof anotherelement,now missing (Vlassopoulou,nos. 11 and 14). Note that the so-called Olive Tree Pediment from the Akropolis shows a small building with a roof hipped at both ends and cornerakroteria.More distinctlylocal are some of the antefix types. Buschor's Antefix II (Vlassopoulou, no. 8), for instance, recalls the three-peaked antefixesfrom Kalapodi,but, unlike the antefixesof the CentralGreek system,it is attached to the eaves tile ratherthan having the double curveon the bottomedge. One final regional system is thus far documentedat only a few sites of northeastern Greece. Both in forms and in decorationthis system draws mainly upon East Greece. One roof (Figs. 14, 15), from the Polygonal Building in the Herakleion on Thasos, features figural decorationunparalleled in any of the other regional styles of mainland Greece: a raking sima with horse-riders(P1.2:a) and, along the eaves, the well-known antefixeswith relief decorationof Bellerophonon Pegasos (P1. 2:b) which alternatedwith ones showing the Chimaira (P1. 2:c).26Figural terracottafriezes, such as chariot races, are common in 26 Raking sima: C. Picard, "Une cimaise thasienne archafque,"MonPiot 38, 1941, pp. 55-92, figs. 1, 2, pl. V; Daux, p. 101, fig. 48. Bellerophonantefixes:Launey, nos. 1-8, pp. 39-43, figs. 12-14, pls. VIII:1, 2 and IX:1, 3; Daux, p. 101, fig. 45; GreekArt of the Aegean Islands, New York 1979, no. 166, p. 208. Chimaira antefixes:Launey, nos. 9-11, pp. 43-44, fig. 15, pl. VIII:3; Daux, p. 101, fig. 47.

FIG.

15. Roof of the Polygonal Building in the sanctuaryof Herakles on Thasos. Drawing

32

NANCY A. WINTER

East Greece, especially at Larisa on the Hermos, while pentagonal antefixes with figural decorationoccur at several sites of Asia Minor, such as Didyma and Miletos, as well as on the island of Lesbos.27Other examples from Thasos include a marvelous fragment of a raking sima with centaursand antefixes with gorgoneia(Pl. 2:d).28The Gorgon heads are East Greek in style, their closest parallels being the gorgoneionantefixesfrom Miletos and from the near-by sanctuarydiscussedby P. Schneider(pp. 211-222 below) which have the same placementof the fangs, closeto either side of the tongue, and disk earrings.29A Gorgon antefix has also been found at Torone in the Chalkidiki.30 There may be even more regional roofingsystems,or further refinementsof the ones I have proposed,still to be recognized.The overall picture which emerges is far more complex than a simple distinctionbetween buildings which use Laconian-styleroof tiles and those which use Corinthian.Scholarshave long recognizedthat each Greek city-statehad a political identity and characterof its own, and often its own regional style of pottery and sculpture.To these I would add regional styles of architecturalterracottas. NANCY A. WINTER AMERICANSCHOOLOF CLASSICALSTUDIES

54 Souidias Street GR-106 76, Athens, Greece 27 Figural friezes from Larisa on the Hermos: ATK, pls. 19, 21-25. Pentagonal antefixes with figural decoration:ATK, pls. 12 (Mytilene), 53 (Miletos), 56, 57 (Didyma). 28 Raking sima with centaurs:B. Holtzmann, "Une nouvelle sima archaique de Thasos," BCH-Suppl. 5, Thasiaka,Paris 1979, pp. 1-9; The Human Figure in Early Greek Art, Washington 1987, no. 30, p. 110. Gorgon antefixes: C. Picard and C. Avezou, "Les fouilles de Thasos (1913)," CRAI 1914 (pp. 276-305), p. 295, fig. 6; Launey, pp. 44-45, pl. X:3; Daux, p. 101, fig. 49. 29 ATK, no. 3, p. 103, pl. 53:2. 30 "Epyov 1978 (1979), pp. 28-29, fig. 33; A. Cambitoglou, <<'AvaOcKa4p 1978 TopOvv>?,HIpaKrLKIa (1980), p. 88, pl. 75:a.

-

~~~~~~~~~~~~4

b. Raking sima from Halai, Thebes M Archive) a. Antefix from Sparta, Sparta Museum (no number). (Photograph G. Georgiadis)

c. Antefix and eaves tiles from Halai, Thebes Museum. (Photograph ASCS Archive) NANCY

A.

d. Ridge palmette fr (PhotographASC

WINTER: DEFINING REGIONAL STYLES IN ARCHAIc GREEK ARCHITECTURAL TERRAC

a. Raking sima from Thasos, Thasos Museum. (Photograph"EcoleFranpaise

b

nei

ihBleohnfo

h

(Photograph Ecole Franpaised'A

c. Antefix with Chimaira from Thasos, Athens, N.M. 19444. (PhotographEcole Franpaised'Ath'enes) NANcy

A.

d. Gorgoneionantefix from Thasos, Thaso Ecole Franpaised'Ath'enes)

WINTER: DEFINING REGIONAL STYLES IN ARCHAIc GREEK ARCHITECTURAL TERRAC

LES TERRES CUITES ARCHITECTURALES DE DELPHES VINGT ANS APRES LA PUBLICATION E BUT DE CE COLLOQUE etant de permettrela presentationet la discussiond'un llmateriel entierementou partiellementinedit, la presentecommunication,qui porte sur un materielpublie depuis vingt ans,1a quelque chose d'anachronique.La seule decouverte veritablementsignificativefaite depuis lors, deux pieces d'angle portant une sima a decor d'oveset des gargouillesa tete de lion, a ete publiee en 1978.2Les reflexionsqueje vais vous presentern'apporterontrien de nouveau, mais veulent etre un regardcritiquejete sur une publicationpassee. Comme il arrivesouvent,ce regardcritiqueincite a la modestie: livres et hommesvieillissent egalement, et je suis naturellementconscientque certains aspects de la publication delphiquesont aujourd'huidepasses,ou en tout cas doiventetre soumis a revision.Le materiel nouvellementdecouvertou etudie en Grece meme, a Corinthe, Nemee ou Egine par exemple, ou bien en Italie du Sud, a Metaponte ou Sybaris, amene a rectifier certaines attributions.ILconvientegalementde reflechirsur la taxinomie, les principesde classement des terres cuites, et aussi sur les tentatives de groupementregional et chronologiqueque j'avais faites il y a vingt ans. Au prealable,et pour expliquer certainsdes partis qui furent pris alors, il conviendrait de rappeler ce qu'etait l'etat de la recherchedans le domaine qui nous occupe. Pour ce qui est des ateliers corinthiens,qui ont fourni l'essentieldu materieldecouverta Delphes, on ne disposaitque de la publicationd'I. Thallon-Hill et L. S. King,3sans compterle cataloguede Mrs. Van Buren.4Or, si elles ont le merite de rassemblerpour la premierefois un materiel jusqu'alors neglige, ces publicationssont loin de satisfairenotre curiosite: elles ne donnent que des descriptionssommaires,depourvuesde profils, et l'organisationgenerale n'est pas satisfaisante.Les seuls modelesauxquels, a la fin des annees cinquante,je pouvais me referer se situent hors de Corinthe : il s'agissait de la publicationd'Olympie,de celle des terres cuites architecturalesde l'Acropolepar E. Buschor,et de celle des toits de Calydon par E. Dyggve et K. Rhomaios.5 Si le contextebibliographiqueetait loin d'etresatisfaisant,que dire du contextearcheoFouille>> de Delphes, comme,je le crains, a logique et museologique?Lors de la <
1Le Roy, 1967. 2 3

J.-F. Bommelaer,<Simaset gargouillesclassiquesde Delphes?, BCH 102, 1978, p. 173-197. Corinth,IV, i.

4

GFR.

5

Olympia II; TdA; Dyggve; Rhomaios.

CHRISTIAN LE ROY

34

discretet laconique,et meme trop souventsilencieux.II a fallu, dans cette situation,elaborer une presentation,une methodede classement,et tenter au moins de poser les bonnes questions, meme si l'on n'etait pas assure de pouvoiry apporterles bonnes reponses. I. PROBLEMES GENERAUX DE TAXINOMIE : LA NOTION DE TOIT

A quelques raresexceptionspres, commepour le temple en tuf d'Athenaa Marmaria et la Lesche de Cnide, il ne se degageaitdes Journaux de fouille et des inventairesaucun regroupementsignificatifpermettantd'associerun ensemblede terres cuites architecturalesa un monumentprecis. La perte d'informationqui resulte des conditionsde la fouille est certainement importante.Elle est peut-etre moins grave qu'on pourraitle craindre.Dans une stratigraphie,une tuile est tres souventun element intrusif, qui rompt la coherencedu materiel recueilli dans une couche archeologiquedonnee. Cela pour des raisons evidentes: la ?vie>d'une tuile, entendueau sens de perioded'utilisation,est normalementplus longue que celle d'un vase. Dans une couche de destruction,tuiles et vases ne sont donc pas contemporains :j'entendspar la que la tuile peut avoirete fabriqueeplusieursdizainesd'anneesavant le vase. Ainsi, a Corinthe,une antefixepentagonalearchaiquea ete decouvertedans un puits combleentre 460 et 420 av. J.-C.6 Lorsqu'onpeut, grace a la ceramique,dater la couchede destructiond'un batiment,cette date ne s'appliquepas aux tuiles trouveesdans cette couche, une grandemarged'incertipuisque ces dernieresdatentde la constructionde l'edifice: d'oCu tude. A ces difficultess'ajoutentdeux autres facteursde confusion : la dispersiondes fragments et l'habitudedes remplois : c'est, a Delphes, le cas de loin le plus frequent.II n'existe rien qui rappelle, meme de loin, le toit etudie dans ce meme volume par le Dr. Schneideret decouvertpres de Didymes: un ensembleisole, homogene,et cependantcomplet.7 II est donc extremementprobableque, a Delphes commea Corinthe,si les journaux et inventaires avaient ete tenus avec les memes scrupules qu'aujourd'hui,ils n'auraient fait qu'enregistrercette dispersion,ces remplois, et ces incoherencesstratigraphiques.Quelles doivent etre, dans ces conditions,les regles pour le classementdes terres cuites architecturales? Je les resumeraien disant que ce sont des elements d'architecturequ'il faut traiter commede la ceramique,et des elementsde ceramiquequ'il faut traitercommedel'architecture. Mais il ne faut surtout pas sacrifierl'un des deux aspects au profit del'autre. Quand on s'interesse a peu pres exclusivement au decor, on considere chaque piece isolement comme Buschor,on separe les simas des antefixes,et on etudie separementleur developpement stylistique. Certes, a la fin du volume,l'auteur a tente un regroupementrapide pour formerautant de toits. Mais chacunsent bien qu'il s'agit d'unitescomposeesde faSonartificielle, et que, pour Buschor,ce n'etaitpas la l'essentiel.8Inversement,E. Dyggve a fait pour 6

M. Z. Pease, "A Well of the Late Fifth Century at Corinth,"Hesperia 6, 1937 (p. 257-316), p. 312, n?237, fig. 43. 7 P. Schneider, "Aus einem archaischen Bezirk an der Heiligen Strasse von Milet nach Didyma: Das und Bauforschung Tonziegeldach des Ostbaues,"Bericht uiiberdie 34. TagungfifurAusgrabungswissenschaft vom 7.-11. Mai 1986 in Venedig (Koldewey-Gesellschaft),Bonn 1988, p. 16-19; K. Tuchelt, IX. Kazz SonuqlartToplantzsz1987, Ankara 1988, p. 80, fig. 6-8; et la communicationdu Dr. Schneiderdans le present volume (ci-apres). 8 est significatifque dans la publicationde Buschor (TdA) le chapitre traitant des toits ne compte que Il quatre pages (p. 69-73), et le chapitretopographiquecinq (p. 74-78).

LES TERRES CUITES ARCHITECTURALES DE DELPHES

35

le Laphrionde Calydonune etude qui releve presque exclusivementde l'histoirede l'architecture, etudiant les toits en tant que tels, comme autant d'unites organiques dont les elements sont indissociables. Peut-etre, en revanche, a-t-il quelque peu neglige l'etude du decor en tant que tel.9 C'est cependantce second parti, plus structuralque stylistique, qui me sembleetre le seul valable, et de plus correspondretout a fait aux tendancesactuellesde la taxinomie archeologique.A ne s'interesserqu'au decor,on s'expose a des interpretations lacunaireset a une veritable perte d'information.10Faire des terres cuites architecturales, c'est privilegier un ou de <<Stilgeschichte>>, exclusivement,un objet de <> classementqui n'a de sens que par rapport a la visee du classificateur.Inversement,retrouver a travers le toit et les terres cuites un ensemble architectural,c'est retrouver le classement,l'ensemble et la structurevoulus a la fois par l'auteur (l'architecteet le fabricant) et par l'usager antique. Les deux demarchesne s'excluentpas reciproquement,mais la secondedoit avoir la primaute. En theorie, une publicationde terres cuites architecturalesdevrait donc etre une publication de toits. En pratique, ce n'est pas toujours le cas, pour des raisons faciles a comprendre : des qu'on se trouve en presence d'une collection d'une certaine ampleur, la constitutiondes ensemblesdevientde plus en plus difficile,tant s'accroitle nombredes combinaisons possibles. Plutot que de procederpar associationforcee, on preferera- c'est du moinsle parti quej'ai pris pour la collectiondelphique- laisserdes elementsisoles, formant des serieshomogenesmais ou les elementsconstitutifsdu toit corinthien(sima, antefixe,tuile de toits ne soit de rive) ne sont pas tous presents.Encorefaut-il que le nombrede ces ?<
sans doute moyen de reduirele nombrede ces residus, commeje tenteraide le montrerplus loin. Cette impressionde dispersionest probablementrenforceepar le fait, releve par Heilmeyer et B. d'Agostino,12que je me suis refuse a donner des reconstructionsgraphiques completes des toits, du type de celles qu'on retrouve en abondance, par exemple dans les publications italiennes, et qui fournissent ensuite de belles illustrations aux manuels d'architecturegrecque.13II m'a semble en effet impossible,comptetenu de nos incertitudes, de mettreen circulationdes dessinsou la part de l'hypotheseserait trop grande.Meme pour la Lesche de Cnide ou le temple en tuf de Marmaria,j'aurais ete amene a juxtaposer un 9 Les reflexionsfinales de Dyggve (p. 201-212) sont en definitiveplus techniquesque stylistiques. 10Par exemple, la publication par B. Holtzmann (<
36

CHRISTIAN LE ROY

nombre arbitraire de tuiles de longueur hypothetique pour reconstituerun rampant de frontondont la longueurtotale etait elle-meme inconnue.... En revanche,on aurait pu diminuerle nombrede toits et de series, et donc donnerplus de coherencea l'ensemble,en augmentantle nornbred'associationsentre tuiles de profil et de motifs tres voisins. A partir de quel seuil les differencesentre tuiles et simas peintes justifient-elles qu'on les classe dans des toits ou des series eux-memes differents?II faut a nouveau se placer ici du point de vue de l'utilisateur. Le probleme se pose alors de deux manieres.Ou bien les differencesde decorreleventde la simple negligence,et l'on est amene a se demandera quel niveau se situe le degre d'acribieacceptablepour le commanditaire d'un toit decore.Ou bien les differencessont voulues et il faut alors decouvrircommentelles s'organisent, et a quels principes obeissent les variations de couleurs et de motifs. Pour illustrer ce propos d'exemplesempruntesa la publicationdelphique,j'ai probablementeu tort de dissocierdu toit de la Leschede Cnide (toit 56) deux fragments(toit 60) tres proches, a ceci pres que l'entraxe du meandrede base est plus long de 9 a 10 mm : cette marge me parait aujourd'huitout a fait tolerable. De meme,je pense qu'il faudraitreunir en un seul toit mes toits 14 et 15 : profils et motifs sont identiques, seul le jeu des couleurs differe. II aurait fallu reflechir sur la polychromie, l'alternancedes couleurs, soit a l'interieur d'un meme fronton,soit d'un fronton a l'autre.14 II. LES CRITERES DE CLASSEMENT : TYPOLOGIE, ATELIERS, CHRONOLOGIE

1) Typologie.II n'y a guere a revenirsur les analysesproprementstylistiques,du genre de celles qui ont ete developpeespar A. Mallwitz a proposdes simas classiquesd'Olympie.15 II est sur en revancheque le classementdes antefixes archaiques que j'avais propose doit etre revu. A la lumiere des publications d'Egine et de Nemee, la sequence ?antefixes a - antefixes hexagonales - antefixes pentagonales?>> comrnes peut etre consideree comme sure.16Les toits et series 5 a 26 de la publication delphique devraient etre reclasses en consequence.17 2) Ateliers. De tres grands progres ont ete accomplis en vingt ans dans l'analyse des argiles. L'exemple unique que j'avais donne est naturellementtres insuffisant.Non seulement parce qu'y manquent les elements-trace,dont la determinationest capitale pour distinguer les provenances,mais aussi parce qu'une analyse unique n'a en soi aucune valeur. Certes, l'archeologie de laboratoire ne peut pretendre resoudre tous les problemes. La pratiquedes ateliers itinerants,et l'existencede nombreusesreparations,amenentnormalement a repererdes pieces de style identiqueet d'argilesdifferentes.Ces techniquesdevraient neanmoinspermettred'utiles regroupements: par exemple, celui des tuiles de Grece occidentale, avec les ateliers de Thermos et de Calydon;ou encore les ateliers d'Italie du Sud, autour de villes comme Sybariset Metaponte. 14 Cf. Le Roy, 1967, p. 15 en

128-132 (toit 56, Lesche de Cnide); p. 135 (toit 60); p. 49-52 (toits 14 et 15). dans Je pense particulier aux pages consacreesa la <<sima pale>>et a la ?petite sima aux acanthes?>> A. Mallwitz et W. Schiering,OlForsch,V, Die Werkstattdes Pheidiasin Olympia, Berlin 1964, p. 110-134. 16 Cf. en particulierSchwandner,1985, p. 75-76, fig. 47, pl. 26. 17 Ainsi, il n'y a plus a hesitera attribuera Corinthele toit 5 (Le Roy, 1967, p. 28-31); d'autrepart, l'antefixe ?<> A 23 (toit 12) devrait se placer nettementen tete des series a palmette et volute (Le Roy, 1967,p.45-47).

LES TERRES CUITES ARCHITECTURALES DE DELPHES

37

Les fouilles recentes de Corinthe et de Nemee ont d'autre part apporte d'utiles precisions a la connaissancedes atelierscorinthiens: l'antefixede Delphes A.41 (serie 26) n'avait commeparallele, au moment de la publication,qu'une piece athenienne.L'apparitiond'exemplairestres semblablesa Nemee montre qu'il faut desormaisassocierce type aux ateliers corinthiens.18 3) Chronologie.Dans un domaine ou, commeje l'ai indique ci-dessus, le contexte archeologiqueest souventabsent ou trompeur,la mise en chronologierelativedes toits ou des pieces isolees prend une grandeimportance.II faut ici prendregardede ne comparerque ce qui est comparable,et par consequentde n'opererun classementqu'a l'interieurdes series homogeneset des ensemblesregionaux. Je crois avoir moi-meme commis une erreur de ce type en etudiant la sima IV de Calydon (L6wensimadach),de profil corinthien,mais decoree de languettes et d'une double tresse.19En me fondant sur le profil et sur une certaine negligencedans l'execution du decor,j'avais cru etablir que cette sima se placait vers 560550. Mais B. d'Agostinoa fait remarquera juste titre20que la techniquedecorativen'evolue pas forcementau meme rythme a Calydon et a Corinthe, et qu'on ne pouvait batir une chronologierelative en passant de l'une a l'autre. L'etude de ce meme toit a d'autrepart illustre, s'il en etait besoin, la necessitede ne pas separerl'examen des acroteresde celui des tuiles proprementdites. Dyggve associaiten effet a cette sima une tres belle tete de sphinx archaique,indiscutablementanterieurea 550, et plus probablementdes premieresdecenniesdu vIe siecle. Heilmeyer rejetaiten consequence la date de 550 pour la sima, puisque l'acrotereetait anterieure. Mais K. Rhomaios avait montre21que c'etait en fait un autre sphinx d'acrotere,nettementplus petit, de technique plus negligee et surtout nettementplus tardif, qu'il fallait associer a la ?<>, precisionne doit pas faire illusion?.De fait, si l'on admet, commej'ai cru pouvoir le demontrer,que les simas n'etaient pas moulees, mais executees a I'aide d'un gabaritreproduisanten negatif le profil, gabaritpasse a la main le long d'un profil a peine degrossien argile encoreplastique, il en resulte qu'on ne saurait avoir un profil rigoureusement constant. Le gabarit ne sera jamais exactementparallele a la sima pendant toute la 18

Cf. Stephen G. Miller, "Excavationsat Nemea, 1979," Hesperia 49, 1980 (p. 178-205), p. 185, pl. 38 et39. 19

Dyggve, p. 169-188.

20

D'Agostino (ci-dessusnote 12), p. 121. Rhomaios,p. 39. Bommelaer(ci-dessusnote 2), p. 183.

21 22

CHRISTIAN LE ROY

38

duree de l'operation,et a plus forte raison il n'aura pas une position exactementidentique d'une sima a l'autre. On aura done de legeres variationsde profil d'une sima a l'autre et a l'interieurd'une meme sima, ce qui invalideles rapportschiffres,au moins si l'on veut tirer des conclusionsde variations de faible amplitude. La date proposeepar J.-F. Bommelaer pour les deux simas a oves et fers de lance qu'il publie (premier tiers du Ive siecle) est sur des donneesobjectives.23 d'autantplus precieusequ'elle est fond&ee Toute chronologieabsolue doit en effet etre fixee, soit sur l'analyse du contexte,comme dans le cas que je viens d'analyser, soit par referencea des batiments bien dates dont on possede le toit. II faut, me semble-t-il, savoir resister a la tentation de dater les tuiles par reference au systeme ornemental des ceramiques.Certes, les motifs sont les memes (palmettes, lotus, volutes). Mais les evolutionsne sont pas paralleles. Il suffirade rappelerque le passage du <<sombre sur clair? au <
Ibid. p. 196-197.

Le fait reste douteux pour le ?<> Weinberg, "CorinthianRelief Ware," Hesperia 23, 1954, p. 129, 130, 133. 25 Sur l'atelierceramiquede Phari a Thasos, voir, pour la localisationet le contextegeologique,Y. Garlan, Vin et amphoresde Thasos,Athenes 1988, p. 8, fig. 6; et pour la productionde tuiles, la communicationde M. J. Perreaulta ce meme colloque (ci-apres). 26 0. Broneer,Corinth,I, iv, The South Stoa, Princeton, 1954, p. 94-95. 27 Sur la constructiondu gymnase de Delphes, voir J. Bousquet, Etudes sur les comptesdelphiques,Paris 1988, p. 61 et 67 (debut de la constructionen 334/33); p. 68 (fin probable de la constructionen 327/26); p. 171 (en 327/26, 'adductiond'eau montreque le gymnaseest pret a fonctionner).On aurait pu aussi songer a l'hoplotheque,desormaisidentifieeau portiqueOuest, commenceeegalementen 334/33 (ibid., p. 61) et terminee en 321/20 (p. 68). 28 R. Martin, compte rendu de la publicationd'O. Broneer,REG 69, 1956, p. 213-216; du meme auteur, ?<<Bulletin REG 72, 1959, p. 297. archeologique?>>, 29 La necessitede rajeunirla date de la constructionde la Stoa Sud de Corinthes'est affirmeetout au long de la fouille des batimentsqui ont precede la Stoa. Ch. K. Williams, II et J. E. Fisher ("Corinth1971, Forum Area,"Hesperia 41, 1972, p. 170-171) releventque, sur le sol superieurdu ?<
LES TERRES CUITES ARCHITECTURALES DE DELPHES

39

matierede stoai, J. Coulton,a pu en placerla constructionau debutdu IIe siecle.30Dans ces conditions,la date du toit de Delphes devraitelle aussi etre abaissee,et l'attributionau gymnase abandonnee,sans queje puisse du reste proposerune autre attribution. Un regard retrospectif jete sur la publication des terres cuites architecturales de Delphes montre donc que, commeon l'a vu, celle-ci est, sur plusieurs points, sujettea revision. Cela est encourageant,dans la mesureoiucela temoignedes progresde la recherche.Et c'est egalementnormal, dans la mesure oCu ce livre partait en quelque sorte en avant-garde, sans pouvoir s'appuyer sur une methodeet des resultats eprouves.Reste, outre le materiel proprementdit, une problematiquequi me sembletoujoursvalable. La demarcheconsistant a utiliser le materiel non-corinthienpour identifierles ?tresorscoloniaux?(Corfou, Sicile, Italie) a fait ses preuves, meme si les attributionsdoivent encore etre affinees. En ce qui concernela chronologie,la tendance me semble etre au rajeunissement,qu'il s'agisse des temples ?protocorinthiens?d'Isthmia, de Corinthe et de Delphes ou des toits du Ive siecle. Reste un point qui n'est pas indifferent: il est certescommodede tracerune ligne de demarsur sombre>> cation au debut du ve siecle, lorsque la technique en <>. Cette coupureest cependantloin d'etretotale. Le profil <> survit quelque temps au ve siecle,3' et les motifs decoratifsde ce que j'ai appele le <<style severe>sont tres prochesde ceux de la fin de l'archai'sme.32 L'executionreste d'une grande qualite, et ce bien avant dans le ve siecle.33Et F'interethistorique des tuiles d'epoqueclassique est reel. II ne faudraitpas que cela soit oublie. CHRISTIANLE ROY UNIVERSITE DE PARIS - PANTHEON - SORBONNE

17, rue de la Sorbonne 75231 Paris Cedex 05, France apres sa destructionpar la Stoa Sud, a ete decouvertun depot de monnaies dont l'enfouissementn'est pas anterieur a la derniere decennie du Ive siecle. Cette datation est en gros confirmeepar les memes auteurs ("Corinth1972, The Forum Area,"Hesperia 42, 1973, p. 27). Cf. a nouveau Ch. K. Williams, II, "Corinth Excavations, 1979," Hesperia 49, 1980, p. 107 : la constructionse place apres le milieu du Ive siecle et peut descendrejusqu'a 320 av. J.-C. La complexite du problemeest bien marquee par Ch. K. Williams et J. E. Fisher ("Corinth1975, Forum Southwest,"Hesperia 45, 1976, p. 116 et note 20): on a au moins une, et peutetre deux destructions,l'une vers 330, l'autre pas avant la dernieredecenniedu Ive siecle. Et les reconstructions des troisieme et dernierquarts du Ive siecle ont ete causees par quelque chose de plus importantque la simple donationmacedoniennepour la constructionde la Stoa Sud. On notera cependantque la date de Broneer (350-330) est maintenue en gros par G. R. Edwards (Corinth,VII, iii, CorinthianHellenistic Pottery, Princeton 1975, p. 197 et note 12). 30 J. J. Coulton, The ArchitecturalDevelopmentof the GreekStoa, Oxford 1976, p. 228. 31 Cf. Le Roy, 1967, toits et series 49 a 53, p. 121-127. 32 On compareraLe Roy, 1967, pl. 35 (fin du vIe siecle) et pi. 43 (<<style severe>). 33 Les analyses d'A. Mallwitz rappeleesci-dessusnote 15 ont bien marque l'interetesthetiquedes simas en terre cuite d'epoqueclassique.

DIE ARCHAISCHEN DACHER VON OLYMPIA (PLATES D

3, 4)

IE SAMMLUNGARCHITEKTONISCHERTERRAKOTTEN,diedieolympi-

schen Ausgrabungenergaben, dtirfte was Reichhaltigkeitund Mannigfaltigkeitanlangt, und mit Hinsicht auf die Erhaltung der gefundenen Stuicke,jede andere derartige Sammlungtibertreffen." Mit diesen Worten leitete R. Borrmannbereitsim Winter 1878/79 ein erstesVerzeichnis der in Olympia gefundenen Dachterrakottenein. In Olympia wurde damals erst im vierten Jahr gegraben and R. Borrmann, der unter anderem die Aufgabe uibernommen hatte, die Dachterrakottenzu bearbeiten, war in dieser Grabungskampagnezum ersten Mal vor Ort. Heute kennenwir nebender olympischenSammlungnochdie reichenBestandevon Korinth und Delphi. Bezuglich der Vielfalt und der guten Erhaltungder Dachterrakottenbehalt BorrmannsEinschatzungauch 113 Jahre nach Grabungsbeginnnoch ihre Guiltigkeit. Seit der Alten Olympiapublikationwurde durch die weiteren Grabungendie Anzahl der Dachterrakottenwesentlich erh6ht, so dass heute deren Bearbeitung,die sich auf die Ordnung von A. Mallwitz stuitzt,aufgrund der breiteren und damit aussagekraftigeren Materialbasisleichtermoglich ist. Aus archaischerbis hellenisticherZeit stammen ungefahr 4000 dekorierteZiegel der Dachrander,die sich auf uiber50 Dacher verteilenlassen. Diesen Dachern steht eine Liste mit ebenso vielen Gebaudengegenuiber,die ein Tondach getragenhaben mtissen.Auch in zeitlich aufeinanderfolgendeGruppen unterteilt,entsprichtjeweils die Anzahl der Dacher ungefahrder Anzahl ihrer urspruinglichzugehorigenGebaude. In Olympia birgt gerade die Zuweisung der Dacher grosse Schwierigkeitenin sich, da die Dachterrakottenmeist verschlepptoder in spaten Mauern verbautund nicht bei ihrem Gebaudegefundenwurden. Ist eine solche Fundsituationdatierbar,kann sie h6chstensAufschlussdaruibergeben,wann das Gebaudeabgedecktwurde und das Dach in die Erde kam. Von den ca. 50 Dachern Olympias gehoren 19 in archaischeZeit. Sie sind in die drei grossen Gruppen der korinthischen,lakonischenund westgriechischenDacher zu trennen. Anhand von ausgewahlten Beispielen soll im folgenden ein Uberblick tiber diese archaischen Dacher gebotenwerden. Die Reihe der korinthischenDacher beginnt mit zwei fruhen Walmdachern.Leider haben sich in Olympia nicht so viele Ziegel wie von den protokorinthischenDachern aus Korinth und Isthmia1 erhalten, aber die wenigen vorhandenen Stuickemuissenzu zwei Dachern gehoren, da Massunterschiedebestehen. Das kleinere Dach mochte man wegen I Mit der alteren Literatur:Robinson,AM, S. 55-66; Heiden, 1987, S. 17-27. S. M. C. Roebuck,S. 47-63 unten.

42

JOACHIM HEIDEN

seiner geringen Ziegelabmessungengerne einem fruhen Schatzhauszuschreiben,das gr6ssere entsprichtin seinen Massen etwa den erwahntenTempeldachern. Ein fast vollstandigerWalmziegel, Flachziegelfragmente,Kalyptere,und Walmkalyptere lassen die Rekonstruktioneines Walmgrateszu. Der Walmziegel (P1. 3:a) hat an seiner Oberseiteeine rechteckigeAussparung,in die der Falz des nachtshoherenZiegels eingriff. Sein untererFalz ist abgebrochen,erganztman ihn, misst die Seitenlangeknapp 65 cm. Ein in seiner gesamtenLange erhaltenerKalypter misst 63.1 cm. Die an den Walmziegel anstossendenFlachziegel waren an ihren Randern sehr stark aufgebogen,um von der seitlichen zur hinteren Dachneigung zu vermitteln. Diese durch ihre Dicke sehr schweren Flachziegel sind notwendig,da der Walmziegel an seiner Unterseite vollig plan gearbeitetist. Nur seine Oberseitezeigt den Walmgrat und damit die beiden Dachneigungen. Solche ausgleichenden Flachziegel muss es auch fur das Dach des ersten korinthischen Apollontempel gegeben haben, denn dessen Walmziegel zeigen ebenfalls die plane Unterseite. Das olympischeDach entsprichtnicht v6llig den bislang bekanntenprotokorinthischen Dachern. So sind z.B. Flach- und Deckziegelnicht als "combinationtiles"gearbeitetund an den untersten Deckziegeln scheinen bereits einfache Hornerantefixe,wie wir sie aus anderen Grabungenkennen,3gesessenzu haben. Das Hornerantefix(P1.3:b) tragt wie auch der Firstkalypterkeine Verzierung. Durch den Scheitelwinkeldes Firstkalypterslasst sich eine sehr schwache Dachneigung von 12.5 Grad errechnen. Wegen der Antefixe mochte man zeitlich etwas von den protokorinthischenDachern abrucken.Dennoch ist das Dach in das 7. Jh. v. Chr. zu datieren:ein Fragmentwurde im Fundament des alteren Schatzhausesvon Sikyon gefunden. Das Schatzhausist um 600 v. Chr. errichtet worden. Andere Flachziegelfragmentestammen aus dem spatestens im 3. Viertel des 7. Jh. v. Chr. verftilltenBrunnen 118 im Suidostendes Grabungsgelandes.Diese schon sehr frtih in die Erde gekommenenZiegel muissennicht das Ende des Daches und seines Gebaudes angeben. Sie konnen schon beim Versetzen beschadigtworden sein. Der gr6ssere Teil des Daches stammt aus Fundschichten,die in die Jahre um 500 v. Chr. zu datierensind, und uns damit vermutlichdie Zerst6rungszeitdes Gebaudesanzeigen. Welches Gebaude Olympias dieses Walmdach einst deckte ist nicht zu erschliessen. Gab es-in Analogie zu Korinth,Isthmia und Delphi-vielleicht docheinen Vorgangerdes fruhklassischenZeustempels? Erstaunlicherweisegibt es in Olympia kein einziges Fragment einer korinthischen Hohlkehlsimamit Blattstabdekor,wie sie uns in anderenHeiligtumern in der Zeit um und nach 600 v. Chr. so haufig begegnen.4 2 Robinson,AM,

S. 65, Abb. 5. Vor allem in der diagonalenSchnittzeichnungzu erkennen. I Zusammenstellungder H6rnerantefixebei: N. Cooper, 1983, S. 62-65, Taf. 20, 21. 4 Heiden, 1987, S. 29-45.

DIE ARCHAISCHEN DACHER VON OLYMPIA

43

KorinthischeDacher treten in Olympia erst wieder in der zweiten Halfte des 6. Jh. v. Chr. auf. Sechs Dacher mit abgesetzter Wellensima und Palmettenantefixen-darunter das bekannte Dach des Megarerschatzhauses-reprasentieren das normale korinthische Dach und werfen weniger in der Rekonstruktionund Datierung als in der Zuweisung Problemeauf. In der ersten Halfte des 6. Jh. v. Chr., in der korinthischeHandwerker anscheinend nicht in Olympia arbeiteten,haben vor allem lakonischeund ftir ihre Schatzhauserauch westgriechischeWerkstattenDacher hergestellt. Dennoch konnen aus der Menge der lakonischenZiegel als erkennbareDachzusammenhange nur drei Stuck angesprochenwerden. Es sind dies die Dacher des Heratempels (Winter, S. 14 oben, fig. 1) mit seinem gewaltigen Firstakroter(Kastner, P1. 40), das mit seinem Durchmesservon fast zweieinhalb Metern ein Meisterwerk der Brenntechnik darstellt, des Bouleuterions mit seinen von Gorgonen geschmuicktenScheibenantefixen (P1. 3:c), und ein Dach, dessen Stirnziegel ein sichelformiges Ornament tragen. Die Dacher des Heraions und des Bouleuterionssind durchdie Fundsituationihren Gebauden zugewiesen.

Alle ubrigen lakonischenZiegel gehbrenentweder nicht in archaischeZeit oder konnen-da sie keine Verzierungentragen-nur sehr schwer als einzelne Dacher identifiziert werden. Sie werden uns unbekannte,weniger bedeutendeGebaudegedeckthaben. Die dritte Gruppe umfasst sieben westgriechische Dacher. Ein besonders reich geschmiicktes Dach ist durch die Fundumstande-die Ziegel waren gemeinsam mit den Geisonbl6ckenin der Herulermauerverbaut-dem Geloerschatzhauszugeschrieben.Vermutlich geh6rt zu der spaterenVorhalle dieses Schatzhausesein Dach, das im Sinne einer verkleinertenund vereinfachtenWiederholungden alteren Dachrandzitiert. Als Hypothese fur die folgendenUberlegungen zu den tibrigen fuinfwestgriechischen Dachern gilt, dass fur deren Zuweisung nur die sizilischen und unteritalischenSchatzhauser in Frage kommen. Der reiche Komplex der Dacher mit verzierten Hornerantefixenumfasst knapp 350 Stiicke und kann in drei einander ahnliche Dacher, die sicher in der gleichen Werkstatt angefertigt wurden, untergliedertwerden. Die Dacher muissenmehr oder weniger zeitgleich sein, denn stilistischeUnterschiedesind nicht feststellbar. Diese Dacher weichen vom sizilischen Dachsystem-wie etwa dem Dach des Geloerschatzhauses-vor allem dadurchab, dass sie keine Sima mit R6hrenwasserspeiernan der Langseiten der Gebaude haben. Hier schliessen die Hornerdacherganz wie die korinthischen Dacher nur mit Traufziegeln und Antefixen ab. Unter den Traufziegeln verkleidet eine Terrakottaplattedie Geisonstirn. Diese Verkleidungsplattenziehen sich um den gesamten Dachrand herum und schmtickenso auch an der Giebelseite das Schrag- und das Horizontalgeison.Die Giebelsimawird wie die Traufseite von Hornerantefixenbekront. In Olympia haben sich drei verschiedene Giebelsimen mit H6rnerantefixen, drei Traufziegel, zwei Antefixartenund vier verschiedeneGeisonverkleidungsplattenerhalten.

44

JOACHIM HEIDEN

Die Giebelantefixe(P1.3:d) sassenjeweils am unterenEnde des Ziegels uber den Uberlappungsfalzen. Der an den Hornerantefixender Traufseite sitzende Kalypter weist im Querschnittdie typische korinthischeForm auf. Die Sima ist nicht wie bei den korinthischenund sizilischen Dachern eine hoch aufgebogeneRinne, in der das Wasser zu den Gebaudeseitenabgeleitetwird, sondernbestehtnur aus einem leicht uiberden Ziegel erhabenenRand. Die Stirn der Sima und die untere Zone des Antefixes sind mit einem einfachenMaanderbandverziert. Eine Eigenart der Hornerdacherbesteht darin, dass die Ornamentbandermit einem Rollstempelin den weichen Ton gedrticktsind, so dass ein sehr feines Relief entsteht.Diese Reliefierungdiente wahrend der Bemalungals Begrenzungslinien. Von den vier verschiedenenVerkleidungsplattenhaben drei eines gemeinsam:sie zitieren in Ton Elemente der dorischenOrdnung. So ist der Aufbau der ersten Platte (P1.4:a) mit den fur Dachterrakottenublichen Elementenwie Rundstaben,Flechtbandund Maanderdurcheine Reihe von Regulae mit Guttae unterbrochen.Hierzu gibt es in Delphi ein sehr ahnlichesVergleichsstuick.5 Bei der zweiten Verkleidungsplatte (P1. 4:c) ist zwischen zwei Doppelrundstabe ein Triglyphenfries gesetzt. Darunter folgen noch Regulae, diesmal ohne Guttae und ein Maanderband. Die dritte Platte zeigt unter einem gegenstandigenAnthemienbandwieder die Regula mit den Guttae. Von einer Verkleidungsplatte(P1. 4:b) sind bei weitem die meisten Fragmente gefunden worden. Auch fur ihr Spiralbandist der Rollstempel benutzt worden. Sie ist niedriger als die tibrigen und ist nicht durch Elemente der dorischen Ordnung bereichert. Daher nehme ich zur Rekonstruktionder Eindeckungenan, dass diese Platte an allen drei Dachern in gleicher Weise die Schraggeisaschmtickte,denn die Verkleidungsplattenmit den architektonischenMotiven, die alle in horizontalerRichtungden Tempel gliedern,sind nur schwer am Schraggeisonvorstellbar. Die Hornerdachertreten nicht allein in Olympia auf. Ein sehr eng verwandtes-oben schon erwahntes-Dach wurde in Delphi gefunden. Ch. Le Roy stellte bereits die beiden Dacher nebeneinanderund vermutete,dass eine der machtigenunteritalischenStadtezwei Schatzhauser-eines nach Delphi und eines nach Olympia-gestiftet hat.6 Ich mochte D. Mertens folgen, der das Produktionszentrumsolcher Hornerdacherin Kroton lokalisierte und fur das bislang unbenannte Fundament auf der olympischen Schatzhausterrasseeinen Thesauros dieser unteritalischenStadtvorschlug.7 Die Dachterrakottenaus Krotonund dem nahegelegenenApollonheiligtumvon Crimisa lassen sich sehr gut mit den delphischenund olympischenSttickenvergleichen.8 Auch hier ist die Giebelsima mit den Hornerantefixennur als schwache Leiste angelegt. Die Verkleidungsplattensind durchMotive der dorischenArchitekturwie Triglyphen I

Le Roy, 1967, S. 80 (G.19), Taf. 25:1. Le Roy, 1967, S. 81-84. 7 D. Mertens, "I santuari di Capo Colonna e Crimisa,"Atti del 23. convegnodi studi sulla Magna Grecia, Taranto 1983 (1984), S. 215-217. 8 Ebd., S. 208-223, Abb. 5, Taf. 27, 28. 6

DIE ARCHAISCHEN DACHER VON OLYMPIA

45

und Tropfenleistenzusatzlich verziertund die Ornamentbandersind mit einem Rollstempel in den Ton gedrtickt. Die drei olympischen H6rnerdacher,die sich durch ihre angeftihrtenEigenarten als Gruppe zusammenschliessenund sich von den sizilischen Dachterrakottenabheben, werden die unteritalischenSchatzhauservon Sybaris, Metapont und vermutlich Kroton gedeckthaben. Aus der Gruppe der westgriechischenDacher bleibennoch zwei aus der 1. Halfte des 6. Jh. v. Chr. tibrig, die zu den Schatzhausernvon Syrakus und Selinunt gehoren muissen. Einstweilen kann nicht entschiedenwerden, welches Dach welches der beiden Gebaude deckte.Uber das Schatzhausvon Syrakuswissen wir von Pausanias nur, dass Gelon nach der Schlacht von Himera drei Linnenpanzeraus der Punierbeute in das Schatzhaus stiftete.9Das sagt uns nichts tiberdie Entstehungszeitdes Schatzhauses. Das altere von beiden ist das sogenannteRosettendach(P1. 4:e). Eine einfache Sima, deren Hohlkehle einen Blattstab ziert und auf deren oberen Leiste die namengebenden Rosetten zu finden sind, steht uibereiner Geisonverkleidungsplattemit kraftigen Rundstaben. Fur den Wasserabflusssorgeneinfacher6hrenf6rmigeWasserspeier. Der Rand des letzten zu erwahnendensizilischen Daches wurde von einer den ganzen Bau umspannenden Blattstabsimaund Verkleidungsplattenmit einem Lotos-Palmetten Dekor geschmtickt(P1. 4:f). An jedem Traufsimablocksitzen zwei Wasserspeier.Als Besonderheitliegen uiberden Stosskantender Simablockekleine Kopfantefixe(P1.4:d), die als in die Literatureingingen. Leidergibt es bis heute keine weiteren Funde, "Gorgonenk6pfe" die uns das Untergesicht zeigen k6nnten, dennochscheint es sich mir wegen des schmalen Gesichtsschnitteseher um Frauenkopfantefixeals um Gorgonenzu handeln. Ein archaischesDach, das sowohl korinthischewie auch westgriechischeEigenarten aufweist, ist sicher dem Schatzhausvon Epidamnoszugeschrieben. Die Anlage des Daches ohne Traufsima und ohne Verkleidungsplattenist rein korinthisch, das reich gegliederteProfil der Sima und der Traufziegel erinnerteher an westgriechische Dacher. Diese Situation entsprichtder Stadt Epidamnosals korinthischeKolonie, die aber aufgrundihrer Lage starke Kontaktenach Westen unterhielt. Die Nagell6cher in den Traufziegeln haben den gleichen Abstandund die gleiche Gr6sse wie die Nagelreste in den Geisonbl6cken,die dem Schatzhausvon Epidamnoszugeschiebensind und bestatigen so die Zuweisung. Zusammenfassendsoll ein Blick auf die historischeSituationOlympiasim 7. und 6. Jh. v. Chr. das Auftretender verschiedenenDachartenerklarenhelfen. Die altesten Eindeckungen Olympias-die zwei korinthischen Walmdacher der 2. Halfte des 7. Jh. v. Chr.-fallen in die Regierungszeitder Kypseliden. Die guten Beziehungen dieser korinthischenTyrannen zu Olympia sind bekannt.Sie weihten dem olympischen Zeus die noch von Pausanias bestaunte Kypseloslade (Pausanias, V.17.5) und ein I

F. Eckstein,Hrsg.,Pausanias, Munchen1986, VI.I9.7, Anm.58.

JOACHIM HEIDEN

46

goldenes Zeusbild (Pausanias, v.2.3), das wegen seiner Pracht beruihmtwurde. Die ersten korinthischenOlympioniken sind schon im spaten 8. Jh. v. Chr. in den Siegerlisten zu finden.10 Moglicherweise stifteten die Korinther ein Schatzhaus in das Heiligtum, denn die wertvollenWeihgeschenkewird man nicht unter freiemHimmel aufbewahrthaben. Dieses friihe, uns nicht mehr bekannte Schatzhaus konnte mit dem kleinen Walmdach gedeckt gewesen sein. In der Zeit zwischen ca. 600 und der Mitte des 6. Jh. v. Chr., in der es in Olympia keine korinthischenDacher gibt, steht das Heiligtum unter starkemspartanischemEinfluss. Die beeindruckendeReihe der spartanischenOlympionikensprichtfur sich.11 So verwundertes nicht, dass der dominanteHeratempelmit einem lakonischenDach gedecktwar. Die westgriechischenDacher sind ausschliesslichfur die SchatzhauserSiziliensund Unteritaliens hergestelltworden. Die Forderungnach einem Schatzhausvon Krotonund der Anwesenheiteiner krotonischenDachwerkstattin Olympiawird mit einem Blickauf die Siegerlistenum so verstandlicher:aus archaischerZeit sinduns nichtwenigerals 20 Krotoniaten als Olympionikenuiberliefert,davonsiegten 11 allein in der2. Halfte des 6. Jh. v. Chr.12 Die Reihe derspartanischenOlympionikenbrichtimJahre 552 v. Chr.derartjahab, dass sogardie Meinung vertretenwurde, Spartahatte an den Spielennichtmehrteilgenommen.13 Die darauf folgende zweite Jahrhunderthalfteist genau die Zeit, in der die korinthischen Werkstattenmit den Dachern mit abgesetzterWellensima wieder fur Olympia zu arbeitenbegannen. Die korinthischenDachern lassen sich von diesemZeitpunktan bis in den Hellenismus verfolgen. JOACHIM HEIDEN

Deutsches Grabungshaus 27065 Olympia (Elis) Greece

1

A. Honle, Olympiain der Politik der griechischenStaatenwelt,Tiibingen 1968, S. 49-51. I Honle (Anm. 10 oben), S. 128-130. 12 Honle (Anm. 10 oben), S. 80-87. 13 Honle (Anm. 10 oben), S. 124-128.

PLATE 3

t~~~~~~~~~~~~~~~~~~~~~~~~~Ai a. Walmziegel und Flachziegel lK32, lK33, 1K34

b. Hbrnerantefix 1K29

c. Bouleuterionantefix3L34

&,~~~~~~A

d. Hdrnerantefix4W24

PLATE 4

b. Geisonverkleidungsplatte4W96

a. Geisonverkleidungsplatte4W171

d. Kopfantefix3W42

c. Geisonverkleidungsplatte4W186

c.~ Beisnverleidngspamt ~ ~ ~ ~ ~un4W186rlednspat32 ~~~~~~f

e. Rosettensima 2W7

tS"_~JACI

HIE:

DI

W

RHASHNDAHRVN

LMI

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH (PLATES

5-8)

N DISCUSSINGTHE ARCHITECTURALTERRACOTTASof the Archaicperiod at Corinth I would like to show the stylistic developmentof some types originated there and, in addition, indicate some measure of the influence which the Corinthian industry exertedon other areas of Greece. At the outset we should understandthat Corinth cannot be equated with sanctuaries like Delphi and Olympia, nor with a city such as Athens. Corinth was essentially an industrial city. It did have a focal point in the early 7th-century city on the outcropping known as Temple Hill, but apparentlyan agorawas late to crystallize,and public buildings were few and scattered. Instead the city was composedof loosely knit centers combining houses and craft industries.1While relatively little material of high quality, manufactured at Corinth, has been found, it is clear that Corinth by the end of the 8th century was a flourishingcommercialcenterwith an extensivetrade in Greece and to the East and West.2 Any discussionof Corinthiantiles must begin with the roof of the Protocorinthiantemple, the predecessorof the 6th-century Temple of Apollo.3This temple was built ca. 680 B.C.4 and remainedin use for about a century,when it was destroyedby fire. Both their early date and their unusual form make the tiles important.They were in many ways precursors of later tile systems. The 7th-centuryroof (Fig. 1) is made up of combinationtiles except for a set of single covers.Both curvedand angular elementsare used in the tiles. They are undecoratedexcept for a reddishbrown or black glaze on some. The clay is coarsewith many grits and ranges from greenish or yellowish buff to reddishbrown in color. Usually the tiles have a smooth clay slip. In spite of the variationsall evidentlybelong to a single roof, although more than one period in its lifetime, includingrepairs,is probablyrepresented. 1 C. Roebuck, "SomeAspects of Urbanization in Corinth,"Hesperia 41, 1972 (pp. 96-127), pp. 96-105, 116-127; C. K. Williams, II, "Excavationsat Corinth,"AeAr 23, 1968, A', pp. 134-136; C. K. Williams, II, "Corinth1969: Forum Area,"Hesperia 39, 1970 (pp. 1-39), pp. 35, 38; C. K. Williams, II and J. E. Fisher, "Corinth 1970: The Forum Area," Hesperia 40, 1971 (pp. 1-51), pp. 5-10; eidem, "Corinth 1972: The ForumArea,"Hesperia 42, 1973 (pp. 1-43), pp. 14-17; Robinson,NFGH, p. 240 and note 5; Robinson,TH, p. 212 and note 28; J. Salmon, Wealthy Corinth, A History of the City to 338 B.C., Oxford 1984, pp. 99, 163-164,402-403. 2 J. N. Coldstream,GeometricGreece,London [1977], pp. 186-188; Salmon, op. cit., p. 62. 3 M. C. Roebuck, "Excavationsat Corinth; 1954," Hesperia 24, 1955, pp. 147-157; Robinson, TH, pp. 205, 216-217, 224, 230-235; Williams, IrTAt, pp. 346-347; Salmon (footnote 1 above), pp. 59-61; A. Mallwitz, "Kritischeszur Architektur Griechenlands im 8. und 7. Jahrhundert,"AA (JdI 96) 1981 (pp. 599-642), pp. 637-638; Heiden, 1987, pp. 17-20, 23. 4Williams, Ir.Tfrt, p. 346; Robinson, AM, p. 57. [As pointed out by R. C. S. Felsch (footnote 40, pp. 313-314 below), the date of 680 B.C., providedby the pottery found in the working chip layers, provides only a terminuspost quem for the temple and its roof-Editor.]

48

MARY C. ROEBUCK

I

HI

P

T IETL

U P P E It fI

R U N D L

E

DCE

I

T I D E

L E u trrE

P A

Pi

A

C OV

E

T

TL

E

r;

_

FIG.

-'-

-z

T

A

1. Reconstructionof Protocorinthiantemple roof (Corinth)

The roof had a central ridge with hips at one end, probablyat both. The slopes were coveredwith combinationpan and cover tiles, both left- and right-handed.The pans are concavein section, the covers convex. Over the junction of the pan tiles at the center were single covers. Combinationtiles also coveredthe hips. These are basically pan and cover tiles bisecteddiagonallyand thenjoined at an angle to fold over the junction of the tiles from the two slopes. This arrangementresulted in a square, cap-like cover at the lower corner made up of both curvedand angular elements.The ridge covertiles consistof a square cap, convex in form and curved in all directionsto fit over the coversfrom either slope and the adjacentridge tile. The ridge tile is folded at an angle to coverthe ridge and the tops of the pans on the slopes. At the eaves the pan tiles graduallyflatten out to form a flat underside,necessaryif the tile is to rest tightly on the woodenframeworkunderneath.An offset on the undersideforms a stop against the beam. On its free side the upper surface of the eaves pan tile slopes upward to form a slight peak with the adjacentpan tile and to fit tightly under the cover

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

49

attachedto it. The cover,which graduallybecomesangular as it nears the edge of the eaves, is set back slightly from the pan sectionof the tile. There is a slight peak at the centertop. The tiles display an ingenious systemof oblique cuttings,notches,thinning, and offsets to hold them togethersecurely and also to lessen the weight. There is evidence,too, of iron pins used to hold the single coversin place. Clearly we are dealing with a fairly sophisticatedroofingsystem. How long the process of developmenttook we do not know. The temple models suggest that tile roofs were not commonfor very long beforethe Corinthroof was constructed.The Ithacamodelof the late 8th centurywith its painted black and white squareson the roof is the only one which may depict a tile roof.5But, given the long experienceof the Corinthiansin working with clay, once the idea of a tile roof was conceived,the actual implementationwould not have been difficult. The Corinth roof, although the earliest, was not the only roof of this type. Others existed at Delphi, Isthmia, and Perachora,possibly also at Olympia.6All the tiles are essentially the same. The clay of the tiles from the first three seems to be identical. It correspondsto clay from beds in the Corinthian area. Probablyall were made of Corinthian clay and either manufacturedat Corinth and exportedto the other sites or made from clay and molds shipped to the other sites, more likely the former. Thus, we find that in the first half of the 7th centuryCorinthhad alreadyestablisheda trade in architecturalterracottas.7It is interestingthat all the sites at which these tiles have been found, apart from Corinth, are sanctuaries.So, presumably,we should regard sanctuaries as significant points of diffusion in tracing stylistic influences of architecturalterracottas.Since travel to them was common,their buildingswith their new, at the time, roof styles would have become well known and would stimulate a wish to acquire such terracottasfrom the sourceor ideas of imitating and adaptingthem in other towns. It seems likely as well that both the angularCorinthianand the curvedLaconiantype of tiles were originally offshoots of this early roofing system. Obviously they developeddifferently, and some areas preferredone, others the other. As various places experimented, improved, and refined the forms, regional systems were developed.With the addition of I

M. Robertson,"Excavationsin Ithaca, V," BSA 43, 1948 (pp. 1-124), pp. 101-102, pl. 45:a-g; R. M. Cook, "The ArchetypalDoric Temple," BSA 65, 1970 (pp. 17-19), p. 17, note 1; Robinson,AM, pp. 58-59 and note 14; Salmon (footnote1 above), pp. 97-98; idem, "The Heraeum at Perachoraand Early Corinthand Megara,"BSA 67, 1972 (pp. 159-204), pp. 179-180, 185-187; Williams, I:r Wq,p. 346 and note 8; Heiden, 1987, p. 27. 6 Delphi: Le Roy, 1967, pp. 21-28; Heiden, 1987, p. 22. Isthmia: 0. Broneer, Isthmia, I, Temple of Poseidon,Princeton1971, pp. 40-55; J. J. Coulton, GreekArchitectsat Work,London[1977], p. 35; Heiden, 1987, pp. 20-21, 23. Perachora:Heiden, 1987, p. 21; Robinson,AM, p. 55, note 1. In the summer of 1937 there were many tile fragments of this type on the site, and others were reportedlyamong the Perachora material in the National Museum. At this time two pieces were broughtto Corinth from Perachoraand are now in the Corinth Museum, inventoriedas FC 102 and FC 103. 7 For a discussion of trade and the movementof workmen and materials see Salmon (footnote 1 above), pp. 120-126; Le Roy, 1967, pp. 54, 201. Heiden (1987, pp. 49-51) commentson the use of a top dressingof Corinthianclay in the tiles from Thermon and Kalydonand the use of molds in Olympia and Thermon. He concludesthat there were Corinthianworkshopsset up on the sites.

MARY C. ROEBUCK

50

decorationthe processwas acceleratedas elementsof designwere selectedand mingled.The study of forms and decorationseems to reveal the influence of the Corinthian style at the outset over much of the Peloponnesosand West Greece. Olympia, Bassai, and some other sites using the curvedtiles soon developeda system quite independentof the Corinthian. But the use of a black wash on some of the Olympia tiles and on some of those from Thermon may reflectthe influenceof the early Corinthian roof.8 Some localities, for example Sicily, Corfu, Thermon, and Kalydon, used various hybridforms on some of their roofs, combiningelementsof both systems.The flat pans and angular covers used at the eaves of the early roofs were soon adoptedfor all the pans and covertiles in the Corinthianarea. Certain featuresof the early forms continuedto be used. The undercuttingof the pan tiles at the lower end to form an offset catching on the tile below becomes common practice.9Later this offset develops into a trough, although the simpler form also is found in later tiles. The slight curveupward of the upper surfaceof the pan toward the side is also used in later tiles. The undercuttingof the free side of the cover continuesin use as well. The notch cut on the upper end of the covertile appearsin the 6thcenturyTemple of Apollo at Corinth.Here its use has been extendedto the other side of the cover.10This is a feature which also occursin Roof B2 (the blassgelbenroof) at Kalydon.11 The oblique cutting of the lower end of the pan and of the upper end of the coverappear in the tiles of the so-called Corinthian roof at Corfu and in single cover tiles at Thermon.12 The single three-peakedcoversat Halieis and Aigina which rest on eaves tiles with raised edges forminga triangle13have a surprisingresemblanceto the tiles of the eaves on the early temple at Corinth (P1. 5, FT 209). Evidently the flat and angular tiles adopted for the Corinthiansystem becamebasic for most of the areas using such tiles. Where there are differencesthey occurmainly in featuressuch as the semicircularridge covertiles and in decoratedtiles.14There varying influencesand differingtastes becomeapparent.Since elements from differentareas may be combinedin a number of ways, it is often difficultto trace the origins and disseminationof any given style. When the tiles were first decoratedthe system which was developedat Corinth consisted of a cavetto-type raking sima decorated with a painted tongue pattern and single guilloche, eaves tiles with single guilloche, and pentagonal antefixes with a design of 8

GFR, p. 50 for Olympia, p. 66 for Thermon.

9 Examples of this are to be seen in the tiles from Aigina: Schwandner,1985, pp. 73-74,

fig. 46, no. 207. The same is true at Halieis: N. Cooper, 1983, p. 29, pl. 16, no. 97. There are numerousexamples at Corinth. It is also found at Kalydon:C. A. Rhomaios, <<'OKicpaAos!TroVAaspaiov i-S KaXv8Wvos!o, 'ApX'E4 1937 (pp. 300-315), p. 309, fig. 4. 10 Corinth IV, i, p. 41, fig. 48. A number of pieces have now been inventoriedas FP 260, FC 70, FC 74, FC 75, FC 91. Rhomaios,p. 13, fig. 3:a, b, p. 15, fig. 4:a, b. 12Corfu: Korkyra I, p. 114, fig. 90, p. 135, figs. 106, 107. Thermon: G. Sotiriades, <<'Ava-Ka4a'1ev OfpIuA?, 'ApX'E4 1900 (cols. 161-212), cols. 197-198, fig. 6, but in a covertile. 13N. Cooper, 1983, pp. 31, 34; E.-L. Schwandner, "Der Altere Aphaiatempel auf Aegina," NFGH (pp. 103-120), p. 113, figs. 9, 10. 14 For example the buntesroof at Kalydonwith semicircularridge covers,Dyggve, p. 229, figs. 229, 231; at Corfu, KorkyraI, p. 114, fig. 90 for semicircularridge cover, p. 115, fig. 91 for a sample of the elaborate decoration;Thermon, GFR, p. 67; Delphi, Le Roy, 1967, pp. 65-84.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

51

palmetteleaves and tendrils ending in spirals. The decorationof the eaves tile and antefixes is an obviousapplicationof ornamentto the earliertile forms.The changeto a gable necessitated a raking sima. We do not have any examples of the ridge tiles of this type of roof at Corinth. This decorativesystem was quite widespread in the late 7th and early 6th centuries. Examples have been found at Delphi, Eleusis, Athens, Aigina, Epidauros,Tiryns, Perachora,and Troizen.15 Examples of the cavettoraking sima found at Corinth are limited and do not add significantlyto ProfessorLe Roy's study of the Delphi material.16We are more fortunatein the case of the pentagonalantefixes.The seriesbegins with two examplesof the same type,17 FA 101 (P1.5) and FA 327. At the bottomof the antefix is a triangulararea set off from the upper part by a narrow band. This triangle reflects that formed on the early tiles by the curving up of the surface of the eaves tiles at the sides. Initially the free side of the antefix seemsto have been cut on a slight diagonalat the bottomto fit on the upward sloping surface of the eaves tile below. This practicewas soon abandonedin favor of a straightedge for the bottom of the antefix and a flat surface for the top of the adjacenteaves tile. Where single tiles, rather than combinationtiles, were adopted,as for example at Aigina and Halieis,18 the upward sloping sides of the eaves tiles may be retained,and the bottomof the antefix is cut out to fit over them. Above the apex of the triangle on these two antefixes two thick tendrils rise and curve aroundtoward the sides to form upcurvingspirals. Set between the tendrils in the centeris an elongatedheart with three palmette leaves above.Others of this type have been found at Eleusis and Athens.19The coloringmay be black on creamas at Corinthand Eleusis or red on cream as at Athens. At this stage red and black seem not to have been used on the same tile. The date of ca. 610 B.C. for the pre-PeisistratidTelestereion with which the Eleusis antefixesare associatedgives us a dating for this group. A simple variation of this first type, FA 237 (P1. 5)20 places the heart of the palmette abovethe tendrilsratherthan betweenthem and addsa leaf betweenthe spiralandthe loop of the tendril. The spiral fills only a part of the side insteadof reachingto the top of the upper border.This type also occurs at Delphi.21A more squat form from the Potters' Quarter at Corinth, FA 204 (P1. 5),22 lacks the triangularspace. The heart of the palmetteis set above See GFR, pp. 75-79, 100-103, 128-133 and Le Roy, 1967, pp. 31-62 for some of these. Le Roy, 1967, pp. 44-62. 17 Corinth IV, i, FA 101, pp. 57-58 and p. 11, fig. 1; GFR,p. 129, no. 2; Williams, 1ir Wq,p. 347, note 13, pl. 154; H. Payne, Necrocorinthia,A Study of CorinthianArt in the ArchaicPeriod, Oxford 1931, pp. 252, 256, fig. 106; Heiden, 1987, pp. 29, 31-32, pl. 2:2. FA 327 is unpublished. 18 See footnote 13 above. 19 Eleusis: GFR, p. 129, no. 4; Koch, p. 79, fig. 36; Le Roy, 1967, p. 34, no. 4, pp. 36-37; Rhomaios,p. 38, note 2, p. 105; Travlos, p. 143; Heiden, 1987, p. 31. Athens: TdA II, pp. 29-32, figs. 40-42; GFR, pp. 128129, no. 1, fig. 18; Le Roy, 1967, p. 34, no. 2, p. 37; Travlos, pp. 143-144, fig. 194. 20 FA 237 is unpublished. 21 Delphi: Le Roy, 1967, p. 34, no. 3, A.5, series 7, pp. 32, 37, pls. 5, 118 (for A.9 on plates read A.5); Heiden, 1987, p. 31. 22 A. N. Stillwell, Corinth, XV, ii, The Potters' Quarter,The Terracottas,Princeton 1952, p. 281, pl. 58, no. 61; Le Roy, 1967, pp. 33-34, no. 1, p. 37; Williams, 1ir Wq,p. 347, pl. 155; Heiden, 1987, p. 32, pl. 2:1. 15 16

52

MARY C. ROEBUCK

the tendrils,which are very thick. It is difficultto tell whether it is earlier or later. Both red and blackpaint are used, perhapsan indicationof a later ratherthan an earlierdate. Further variations are found at Aigina, Delphi, Epidauros, and Tiryns.23Some are rather squat and the triangular space is missing. Others have added leaves, more palmette leaves or an arc between the palmette and its heart, or even a combinationof some of these. Clearly there is a good deal of experimentationwith the design, perhaps in an attempt to find a more satisfying form of decoration.Probablythese examples belong to the early part of the 6th century. A series of examples from Corinth shows the continuationof this process.Eventuallyit led to the development of the palmette type of antefix. Two antefixes from Corinth, FA 54324 and FA 553 (P1. 5), along with one from Aigina and one from Troizen,25demonstrate the first stage of the process.The triangularspace has disappeared,leaving room for more elaborationof the decoration.The spirals on either side have now been expanded to two on each side. They are held togetherby a band as before.The ratherlarge centerspirals come up through the band, then curve upward and over. Between them in the center is the usual heart with a three-leavedpalmette. Below the band the tendrils curve around, then upward and downward once more to form spirals at the outer sides of the tile. Under the band, between the tendrils, a three-leaved palmette hangs from a central heart. Above, between each pair of spirals is anothersuch palmette,while below is a single leaf. One of the Corinth examples, FA 543, is attachedto an eaves tile with a single guilloche pattern;the other, FA 553, is broken off at the bottom but was attached. The Aigina and Troizen antefixes are cruderin appearancebut the design is the same. A fragmentaryantefix from the Tile Works at Corinth, FA 422 (P1. 5),26 which uses the same design, shows how an attemptis being made to break free of the pentagonalform. Only the left quarterof the tile with the downcurvingspiral and the palmetteabovethe stem of the tendril is preserved,but here the straight edge across the top has been abandoned. Insteadthe line follows the contourof the palmetteleaves. A differentapproachto the problemis found in some other antefixes.The height of the antefix is now substantiallyincreased.Only the original two tendrils and spirals are used. These are held togetherby a band as before.Within each of the upcurvingspirals is an eye. In the space between the tendril and the spiral is a leaf. The palmette has five leaves and usually an arc between the rather triangular heart and palmette leaves. At the side of the band linking the tendrils is an eye. Below the band advantageis now taken of the increased height. The tendrils curve out and down to the base of the tile in the same way as on the 23 Aigina: GFR, p. 130, no. 8, fig. 5, left; Le Roy, 1967, p. 34, no. 6, p. 37. Delphi: GFR, p. 130, no. 10; Le Roy, 1967, pp. 32, 34, no. 7, p. 37, pl. 5, A.171, series 7, and p. 43, A.3, Roof 9, pl. 6. Epidauros:GFR, p. 130, no. 9. Tiryns: GFR, p. 130, no. 6, fig. 108; Hiibner, 1975, pp. 118-119, fig. 1, no. 13564, pl. 64:3 and no. 17270, pl. 64:2; Heiden, 1987, p. 31; Le Roy, 1967, p. 34, no. 8 and p. 37; Koch, p. 85, fig. 39. 24 Robinson, TH, p. 236, pl. 53:a; Williams, I:r Wq, pp. 347-348, note 13; Heiden, 1987, p. 33, pl. 3:1. FA 553 is unpublished. 25 Aigina: GFR, p. 130, no. 11, fig. 5, right. Troizen: Le Roy, 1967, pp. 35, 37, no. 16; Koch, p. 86, fig. 40; W. Vollgraff, "Antiquitesde Trezene,"BCH 29, 1905 (pp. 269-318), p. 273, fig. 2; Heiden, 1987, p. 33. 26 FA 422 from the Tile Works is unpublished.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

53

antefixesof the 6th-centuryTemple of Apollo at Corinth. Between the tendrilsis a pendent palmetteof three leaves and a heart similar to that in the upper part of the antefix. At first, as we see in one Corinth example, FA 550,27 the loop of the tendrils does not rise very far above the band. The curve is higher in one from the Gymnasium at Corinth.28This trend continuesin three other examples from Corinth,FA 559 and FA 446 (P1.5), and FA 560.29 Aside from Corinththe type is found at Perachorawithout the eye in the spirals, perhapsat Delphi and at Ptoion in a variant form with only three leaves in the palmette above and filling palmettes of three leaves between the spirals and tendrils.30As already pointed out, we have in these antefixesthe basic designof the palmetteantefixesof the Temple of Apollo, but it is still confinedto the pentagonalform. The pentagonal antefixes are found in areas which traded with Corinth or in sanctuaries. Those from Delphi are of Corinthianclay,31 and a numberof othersas well appear to be. These antefixesmust extend in date down into the secondquarterof the 6th century. The three-peakedtype of antefix found in the Argolid32has a somewhat similar decoration of tendrils ending in spirals and a central palmette. This form may also be an offshoot of the Protocorinthiantype of cover. The eaves-tile covers from the early roof do have a slight peak and a slight curve. Anotherform of antefix dating in the late 7th and first half of the 6th centuriesis found with some variationsat Corinth,Athens, Delphi, Argos, and Nemea.33It, too, is essentially a three-peakedantefix but with the addition of a small palmette on the center peak and a curvedvolute on each side peak. In most the profile is accentuatedby groovedor grooved and painted lines. It is least apparent in what is probablythe earliest piece, that from the Demeter Sanctuary at Corinth, FA 547 (P1. 5). Two others, more developed,have been found at Corinth,e.g. FA 24 (P1.5). The example fromArgos,the two from Delphi, and the 27 M. Z. Pease, "A Well of the Late Fifth Century at Corinth,"Hesperia 6, 1937 (pp. 257-316), p. 313, MF 6461, now FA 550, fig. 43; Williams, 1nrj?W,pp. 347-348, note 13; Le Roy, 1967, pp. 35, 37, no. 11; Heiden, 1987, p. 32. 28 FA 518: J. Wiseman, "Excavationsat Corinth,The GymnasiumArea, 1967-1968," Hesperia 38, 1969 (pp. 64-106), p. 99, pl. 31:e; Heiden, 1987, p. 32. 29 FA 446: H. S. Robinson, 'Excavations at Corinth, 1960," Hesperia 31, 1962 (pp. 95-133), p. 114, note 66, pl. 41:e; Le Roy, 1967, pp. 35, no. 14, 37; Williams, Dir Wq,pp. 347-348, note 13 (for "post560 B.C." read "between590 and 560"); Heiden, 1987, p. 32. This antefix cannot belong to the Temple of Apollo as Robinsonsuggested;Heiden concurs.FA 559 and FA 560 are unpublished. 30 Perachora:H. Payne, Perachora, The Sanctuary of Hera, Oxford 1940, I, pl. B:2 and pp. 113-115; Le Roy, 1967, pp. 35, 37, no. 13; Heiden, 1987, p. 32. Delphi: Le Roy, 1967, p. 35, no. 12, A.4, series 24, p. 63; Heiden, 1987, p. 32. Ptoion: Le Roy, 1967, p. 35, no. 15, pl. 118; Heiden, 1987, p. 32. 31 Le Roy, 1967, p. 220. 32 For example, GFR, p. 131, no. 13, fig. 6, from the Argive Heraion; Hiibner, 1975, p. 120, pl. 64:6, no. 17264 in the Nauplia Museum. 3 For a discussion of this type see Williams, Dr4Wq,pp. 348-349; Heiden, 1987, pp. 35-36. Corinth FA 24, FA 404: Williams, Dir Wq,p. 348, note 17, pl. 155; CorinthIV, i, p. 50 for FA 24 (A24). FA 547 from the Demeter Sanctuary:Williams, DirWq, p. 348, note 16. Athens: TdA II, pp. 26-27, figs. 35, 36. Delphi: Le Roy, 1967, pp. 64-65, pl. 19:1, A.41, series 26. Argos:I am indebtedto Mme. Billot for informationabout the example there. Nemea: S. Miller, "Excavationsat Nemea, 1979," Hesperia 49, 1980 (pp. 178-205), p. 185, AT 85, 88, 90-92, 107-112, pl. 38:e, 39:b, p. 190, AT 82, pl. 40:e. There is a possible example in the Nauplia Museum as well: Hubner, 1975, pp. 119-120, no. 17260, pl. 64:4 and fig. 2.

54

MARY C. ROEBUCK

seven from Athens are very similar to them. Of these, that from Argos is of Corinthianclay and likewise one of the two from Delphi. The other is evidentlyof local clay. I do not know whether those from Athens are Attic clay or not. The Nemea antefixes are smaller and badly burned. They may be of local clay. This is another example of the way in which various types of tiles were distributedfrom area to area and copied locally. Consideringthe probabledate of these antefixesand the sites where they have been found, Corinthmay well be their place of origin. As we have seen, various styles were evolving in the latter part of the 7th century and the early 6th century.Mixtures of differentstyles are particularlyapparentin areas such as Western Greece and the sanctuariesof Olympia and Delphi. The influences move both westwardfrom Greece and eastwardfrom Sicily and, as we shall see later, even Corinthhas tracesof some of these outside influences. It is toward the end of the first quarter of the 6th century that we have the first indicationsof the style of roof decorationwhich is to becomethe Corinthian(originallycalled Megarian) roof system. This is made up of palmette antefixes, a raking sima composedof an astragal or small torus molding combinedwith a large quarter-roundor torus molding and a straightfascia, lion's head spouts at the corners,and ridge covertiles with palmettes. We do not have examples of the palmette antefix at Corinth before those used in the Temple of Apollo. By that time the form is well developed.Probablyits absenceearlier is to be explained by the lack of evidencefor the constructionof large-scalebuildings in Corinth early in the 6th century.Two other roofs, however,do providesome earlier examplesof the palmette antefixes. One is the blassgelbenroof (B2) at Kalydon.34 This is variously dated first quarterof the 6th century,ca. 580 B.C., and ca. 570 B.c.35 Here the design of the latest pentagonalantefixes is used, but the palmette has seven leaves, is completelyfree from the frame, and rises above it. This roof also has ridge covers with palmettes of seven leaves, similar to those of the Temple of Apollo. The palmettesof both the antefixes and the ridge coversare rathersquat and not so fully developedas in the Apollo temple. The Kalydonroof seemsto be Corinthianin origin. Slightly more advancedin form is the antefix of Roof 12 at Delphi,36 also of Corinthian manufacture. It is combined with a cavetto sima. Le Roy would date it after the end of the SacredWar and before460 B.C. We seem to have in these two roofs the beginning of the new style of decorationwith palmetteantefixesand ridge covers,but, as yet, without the Corinthiantype of sima. Understandablythe palmetteantefixesevolvedbeforethe new sima form,for that would have been a majorchange. 3 Dyggve, pp. 164-167, figs. 167, 168; Heiden, 1987, pp. 37, 58; Rhomaios, p. 18, fig. 7; F. Poulsen and K. Rhomaios,Erster vorlkufigerBericht iuberdie Ddnisch-GriechischenAusgrabungenvon Kalydon,Copenhagen 1927, p. 33, pl. XL, fig. 57. 35 Poulsen and Rhomaios (op. cit., p. 33) date it ca. 570; Heiden (1987, p. 37) puts it ca. 575; Dyggve (p. 226) says middleof first half of the 6th century;W. B. Dinsmoor (AJA54,1950, review of Dyggve, p. 278) puts it ca. 580; Rhomaios (pp. 12, 98-99) dates it ca. 570. 36 Le Roy, 1967, p. 46, A.23, pl. 7, Roof 12 and p. 62.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

55

The first appearance of the Corinthian sima comes soon after the earliest palmette antefixes.ProfessorLe Roy has put the earliestof these simas toward 560-550 B.C.,37beginning with S.30 from Roof 41 at Delphi and a sima from Epidauros.He follows these with the Kalydon lion roof with the inscription,dating it ca. 550-540 B.C., and sima VII-VIII fromthe Akropolisin Athens of about the same date. Then comesthe roof of the Temple of Apollo, Delphi Roof 42, and the Thermon roof. I would like to suggest a somewhat different sequencein view of the earlier date of around 570-560 B.C. now given by Robinson38 to the Temple of Apollo. I think that 560 B.C. is probablya betterdate for the roof. It fits the sequencebetterthan Weinberg'slater date of ca. 540 B.C. We have seen that the early palmetteantefixesat Kalydonand Delphi coulddatearound 570 B.C., certainlybefore 560 B.C. The Kalydonlion sima clearly representsan early use of the Corinthiansima.39But the decoration,usual for a cavettosima, suggestseither an early experimentationwith the new form beforethe lotus and palmettepatternwas adoptedor a conservativeattitude,a reluctanceto adopt the new style wholeheartedly.Comparisonwith Roof 13 at Delphi,40which Le Roy dates not later than 560, and Series 11 at Delphi41of similardate show the Corinthiancharacterof the decoration.Even the treatmentof the locks of hair on the lion's heads is similar.42One cannot, I think, rule out any Corinthian connectionbecausethe Aitolian alphabet is used for the letters on the individualpieces.43Presumablythe workmenbuilding the temple would have been nativesof Kalydon.The letters indicatingthe orderin which the tiles were to be placedon the buildingwould have been intendedfor them and hence their alphabet.This does not mean that the design and the direction and possiblythe moldsusedwere not Corinthian.In any casethe roofis one of the earliest using the Corinthiansima, and the sima profileis verycloseto that of the Temple of Apollo. The Temple of Apollo first providesus with examples of all the forms of tiles used on the Corinthianstyle of roof. The roof is made up of combinationtiles. The eaves tile (P1.6, FT 231)44 is decoratedwith a single guillochepattern.The antefixes (P1.6, FA 16)45 have a seven-leavedpalmette. It rises abovetendrilswhich are held togetherby a horizontalband. The tendrilscurveoutwardtowardthe sides to form spirals. Below the bandthey spreadout Le Roy, 1967, pp. 111-1 12. Robinson,TH, p. 217; cf. S. S. Weinberg,"Onthe Date of the Temple of Apollo at Corinth,"Hesperia 8, 1939, pp. 191-199. 39 Rhomaios, pp. 53-80; Heiden, 1987, pp. 62-66. 40 Le Roy, 1967, pp. 47-49, 62, pl. 8:1, 2, S.10, S.11. 41 Le Roy, 1967, pp. 45, 62, pl. 7:1, S.16. 42 Le Roy, 1967, pp. 48-49, pl. 8:9, LN.5; Dyggve, fig. 171; Rhomaios,p. 61, fig. 37. 43 Le Roy, 1967, p. 108; L. H. Jeffery, "The Straight Iota in CorinthianEpichoricInscriptions,"BSA 43, 1948 (pp. 201-208), pp. 201-204; eadem, The Local Scriptsof ArchaicGreece,Oxford 1961, p. 226; Heiden, 1987, pp. 48-49. 44 CorinthIV, i, pp. 25, 104-105, T41-44, T50, T51, p. 108, T112, fig. 43:b, and others from more recent excavations,including FT 160, FT 205, FT 206, FT 229; GFR, p. 103, no. 20, fig. 71. 45 Corinth IV, i, p. 12, fig. 2, pl. I, pp. 48-49, 58, A12, A16, A20, A102, FA 364; Williams, IT?'?A, pp. 347-348, note 13, pl. 156:b;Heiden, 1987, pp. 33, 71, pl. 3:2; GFR, pp. 25, 147, no. 13, fig. 71; Koch, p. 83, fig. 38. 37

38

56

MARY C. ROEBUCK

to the lower corners. Between them is a three-leavedpalmette hanging from a fourth leaf. The rakingsima (P1.6, FS 101)46is madeup of threeparts.At the top is an astragaldecorated with verticalbandsof red and blackon the creamof the slip. Below that is a largetorusmolding restingon a straightfascia.This is decoratedwith an alternatingand reversedlotus-andseven-leaved-palmettepattern in red and black on the creamy buff of the slip. The ridge palmettes (P1. 6, FR 3)47 probably had seven leaves. Beside the band are spirals while between the tendrils which spread out below is a teardrop.The painted designs are rather large and heavy in appearance,probablybecauseof the size of the pieces. What remainsof cornerlion's heads fits a date shortlybeforethe middleof the 6th century.48 In determiningthe propersequenceof the variousearly Corinthianstyle simas both the decorationand the profiles (Fig. 2) must be taken into consideration.49In the case of the Delphi sima50the decorationis in the same heavy style, but the pattern is different.There are fewer leaves in the palmette,and the lotus calyx is cut out at the outer sides. This could be an early experiment with the design or simply a preferencefor fewer palmette leaves which is at times found at Delphi. It is difficultto say whether the decorationis earlier or later than that on the Corinth sima. They must be fairly close in date. When the profilesare placed side by side (the Delphi piece is brokenat the top) there are differencesin the curve of the large torus molding. The Delphi example has a more nearly circular line than the Corinth sima (FS 21, FS 23; Fig. 2) which flattenssomewhattoward the base. As nearly as one can tell from the small-scale picture of the Kalydon sima, it is more like the Corinth example.51The sima from Epidauros52seems to be midway between the Delphi and Corinth simas. The Akropolissima (VII-VIII)53 perhaps should be put a little later than these if Buschor'sillustrationcorrectlyrepresentsits profile. Both torus moldingsmake a perfect circle if continued,but the Athens sima extends the circle farther at the bottom. The sima from Thermon54is also close, but the decorationseems somewhatlater, more fluid, and the form of the lotus with its diamond-shapedcenter is normally found after the middle of the century. 46 Corinth IV, i, pp. 19-20, 68-69, 73, 75, fig. 17, pl. IV, S21, S22, S25, S26, S56, S101-103; Robinson, TH, p. 236, note 100, pl. 53:b, FS 1052, FS 1057; there are additional pieces from the more recent excavations;GFR, pp. 25-26, 85, no. 47, figs. 71, 72. Some other pieces of similar style and size, S23, S24, FS 818 (CorinthIV, i, pp. 20, 68-69) probablyalso belongto the temple;they are perhapsthe work of anotherperson. See also Le Roy, 1967, p. 112 and Heiden, 1987, pp. 72, 89. 47 CorinthIV, i, pp. 16-17, fig. 14:a, R3, R13, R28; GFR, pp. 25,160, no. 3; Koch, p. 83, fig. 38:3; Robinson, TH, p. 236, note 101, pl. 53:c, FR 102 and FR 101:a,b, the latter a later replacement. 48 I am indebtedto ProfessorRobinson for telling me about these unpublishedpieces from his excavations on Temple Hill. 49J. Heiden, as well as Le Roy, has studied the chronologicalsequenceof the Corinthiansimas. They are coveredin pp. 70-97 of his publication(Heiden, 1987). His work has only now becomeavailableto me, and it has not been possible to add all the pertinentreferences. 50 Le Roy, 1967, pp. 96, S.30, S.112, series 41, and 112, pls. 33, 101, 123; Heiden, 1987, p. 89. 51

Rhomaios,p. 79, fig. 37. pl. 18:6 from Epidauros;Le Roy, 1967, p. 112. Le Roy, 1967, p. 112; TdA I, pp. 16-17, sima VII-VIII. Le Roy, 1967, p. 112; Heiden, 1987, p. 90; AntDenk II, Berlin 1902-1908, p. 2, fig. 6.

52GM, 53 54

FS 1078

%,

~~~~8

FS 426

FS 44

FS 23

FS 28 FS 901

FS 921

FS 862 FS 21

FS 1010

FIG.

2. Sima profiles (Corinth). Scale 1:2

o

i1

1

58

MARY C. ROEBUCK

As we have observed,the toru-sof the Corinth sima flattens toward the base to form more of a tulip shape. But the simas of the third quarterof the 6th centurytend to lengthen the curve somewhat. Considering these facts, along with the decorationand the present evidencefor the date of the Temple of Apollo, I would suggestthat Corinthshouldbe put at the beginningof the series, not much if any later than 560 B.C. The Kalydonsima would be close in date, also the Epidauros sima and Delphi S.30, none after 550 B.C. The Akropolis sima is perhaps slightly after 550; Thermon might be put ca. 540. Similar to it in profile is one in the Nauplia Museum.55It is necessaryto adjustthe drawingof the profileof this last sima to the same plane for the resemblanceto be clear. The period from ca. 540 to about 510 B.C. providesus with a numberof examplesof the Corinthian type of sima. They come from a variety of sites, among them the Argive Heraion, the Nauplia Museum, Delphi, Olympia, and Argos, as well as Corinth itself. That from the Argive Heraion56and that from the Byzantine Treasury at Olympia57both show a full curve in the torus molding,but the circle is beginningto lengthen at the bottom. The decorationis close to that on the Thermon sima. A date around 540-530 B.C. seems possible. Very close are the one in the Nauplia Museum and Delphi Roof 42.58 The curve of the torus has lengthened a little more in them. The palmettes on the Delphi sima have only five leaves, anotherinstanceof the apparentpreferencefor fewer leaves. They might be dated around 530 B.C. Three examples from Corinth show the circleelongatingfurther.FS 44 (Fig. 2, P1.6)59 is closestto the previoussimas. It and one from Olympia60might be groupedtogether.They are very similar in design. The centerof the lotus is diamond-shaped.The calyx of the lotus has an addedshort leaf on each side in the Olympia example. On the Corinthpiece the extra leaves have merged with the rest of the calyx, but the general outline has remained. The palmette leaves are now thinner and more tapered, giving a lighter impression.The other two, FS 862 and FS 901 (Fig. 2, P1. 6),61 flatten out rather more toward the bottomof the torus molding. The palmette leaves are quite tapered and more curved, the center of the lotus again diamondshaped. A new feature appears on FS 862. The tendrils are no longer continuousbut end in upturnedspirals underthe lotus flowers of the upper register.This is also found on anotherfragment,a lower fascia section.62In additionit has dots between the palmette leaves and the palmette heart. It is a little later in date. This design of tendrils Hubner, 1975, p. 123, fig. 4:a, pl. 66:3, 4, no. 17263. Roy, 1967, p. 112; C. Waldstein, The Argive Heraeum I, Boston/New York 1902, p. 130, pl. XXIII:G; GM, pl. 18:4; Heiden, 1987, pp. 90-91, 95. 57 Le Roy, 1967, p. 112; Olympia II, p. 195, pl. 119:2; GM, pl. 18:5. 58 Nauplia: Hubner, 1975, pp. 123-125, fig. 4:b, pl. 66:2, no. 17282; Heiden, 1987, pp. 91, 94-95. Delphi: Le Roy, 1967, pp. 96-100, Roof 42, S.189 in particular,pls. 1, 34:1, 102, pp. 107, 110-112; GM, pl. 18:2; Heiden, 1987, pp. 90-91, 95. 59 CorinthIV, i, pp. 20, 32, 71-72, S44; GFR, pp. 87-88, no. 59, fig. 1; Heiden, 1987, pp. 91-92, 94-95; Le Roy, 1967, p. 112; GM, pl. 18:9. 60 Le Roy, 1967, p. 112; Olympia II, pl. 118:4;GM, pl. 18:8;Heiden, 1987, pp. 93, 95. 61 Both FS 862 from the Tile Works and FS 901 are unpublished. 62 CorinthIV, i, pp. 21, 23, fig. 19:b, p. 72, S45; GFR, p. 88, no. 61; Heiden, 1987, pp. 94-95. 55

56 Le

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

59

ending in spirals is developedfurther in the light-on-darksimas. I would suggest a date of about 520 B.C. for FS 44 and the Olympia piece, about 520-510 for FS 862 and FS 901. The profileof the marbleTelestereion sima at Eleusis63is also close to these. Two other simas from Corinth, FS 426 and FS 921,64fit into this date range as well. The upper moldingson FS 426 (Fig. 2, P1.6) are similar to those on Delphi Roof 42. Like the Delphi sima its palmetteshave only five leaves. What remainsof the lotus seems similar also, but the design is straighterand stiffer.Judging from the profile, however, I think this may be a case of poorer drawing rather than differencein date. The other, FS 921 (Fig. 2, P1. 7), is of interest for several reasons. Instead of the usual double alternatinglotus-andpalmettepattern, it has only a single band of lotus and palmettewhich coversthe torus and extendspart way down over the fascia below. The tendrils are curvedup into spirals under the palmettes.The lower portion of the fascia is decoratedwith a single stoppedmaeander. In this sima we have an early indication of the direction the ornament is to take in the Classical period, as for example in the Lesche of the Knidians at Delphi65dated ca. 475460 B.C. The calyx of the lotus on FS 921 is closed at the bottom,a form found also on the early cyma-reversa light-on-dark simas. But the profile indicates a date not far from 510 B.C. Roof 43 at Delphi66also seems to fit here. Anotherpiece from Delphi, series 46,67 has the same upcurved tendrils under the palmettes, but the lotus and the profile are different. The lower part is missing. Perhaps it too had only a single band of lotus and palmette.It must date close to 500 B.C. With the next group of simas68we find the torus molding beginning to curve in more toward the top again, but the whole curve is shallower than in earlier examples. In this period of ca. 510-500 might be placed the Heraion of Argos, the Megarian Treasury at Olympia, the marbleAlkmaionidTemple at Delphi, Roof 54 at Delphi (the Tufa Temple in Marmaria), the Poros temple at Corinth, Corinth FS 1078 and FS 1010 (Fig. 2, P1. 7), and Delphi Series 44. The Megarian Treasury and Corinth FS 1078 are, as nearly as can be determinedfrom the fragmentaryconditionof the latter,fairly similar in decoration.The palmettesof Delphi Roof 45, Corinth FS 1010, and perhaps FS 1078 have only five leaves. There are some stylistic differenceswithin this group, but all seem to belong to the end of the dark-on-lightseries. At Corinth we have several fragments of Corinthian-type simas with light-on-dark decoration.The design is usually quite neat with regularlycurvedpalmetteleaves forminga Le Roy, 1967, p. 112; GM, pl. 18:10. FS 426 and FS 921 are unpublished,but see GM, pl. 18:3 and Le Roy, 1967, p. 112 for FS 426. 65 Le Roy, 1967, pp. 128-132, 139, pls. 1, 46, 103, 110; GFR, p. 94, no. 92, fig. 90. 66 Le Roy, 1967, pp. 100-101, pls. 37, 102, S.226 and p. 112. 67 Le Roy, 1967, pp. 106-107, 112, pls. 41, 102, S.229. 68 Heraion of Argos: Le Roy, 1967, p. 112; GM, pl. 18:14. Megarian Treasury: Le Roy, 1967, p. 112; Olympia II, pl. 119:4; GM, pl. 18:11; Heiden, 1987, pp. 83-88, 91-93, 95. Marble Alkmaionidtemple at Delphi: Le Roy, 1967, p. 112; GM, pl. 18:13. Delphi Roof 45: Le Roy, 1967, pp. 101-106,112, pls. 1,38-40, 102, 109, 124; Heiden, 1987, pp. 85-88, 95. Corinth, Poros temple: Le Roy, 1967, p. 112; GM, pl. 18:16. Corinth FS 1078 and 1010 are unpublished.Delphi series 44: Le Roy, 1967, pp. 101, 112, pls. 37,102, S.227, S.228. 63

64

MARY C. ROEBUCK

60

half circle.The lotus in all the preservedexamples is dividedat the base and has a diamondshapedcenterpetal. The most completepiece is FS 28 (Fig. 2, P1.7).69 Its torus is flattening again and is elongatedtoward the fascia. Close to it is FS 858 (Fig. 2),7?only the upper part of which is preserved.Both have seven-leavedpalmettes.I would suggest a date of 500-490 for them. Another fragment,FS 886 (Fig. 2, Pi. 7),71 has an even flattercurve.It is possible that it belongsto a cyma reversa,but its curvedoes not seem quite right for that. It is close to anotherpiece from Corinth, FS 27 (P1. 7).72 This is close in decorationbut not in profile to Delphi Roof 50.73 They should date around490 B.C. This brings us to an interestingproblem.The light-on-darksima pieces from Corinth have a markedlydifferentsort of profilefromthose of the roofsof the same periodat Delphi. Professor Le Roy has already noted this fact.74With the Delphi examples can also be grouped the sima from the temple at Halai,75which is still in the dark-on-lightstyle. The excavationalevidence suggests a date of ca. 500 B.C. or a little later for the Halai temple, since the earlier building was razed and buried sometimeafter 510 B.C. The decorationof the Halai raking sima shows some affinities to those at Delphi of about 500 B.C. and also to that of Delphi Roof 50 which Le Roy puts 490-480 B.C.76 But the profile is more like that of the Byzantine Treasury at Olympia. If it were not for the excavational evidence and the design one would be inclined to place it in the same general period. Or one might argue that it is archaizing. I think, however, that there is another possibleexplanation.The profilesof the dark-on-lightsimas of the period 510-500 B.C. still have a definiteoutward curve. This lessens in those of the light-on-darktype from Corinth itself. But at Delphi, in contrast, the curve has increasedon the light-on-dark simas, although the decorationis close to that of the Corinthian simas. The latest of them are in a sense very close in profile to the early cyma-reversaform. Comparison with some of the cyma-reversasimas shows the resemblance.This is, I think, what the Delphi roofs and the Halai roof are tendingtoward.At Corinththere are many simas with variationsof the cyma reversa with a decided curve. At the same time there are many others of the same period with a much shallower curve, more like the profiles of the Corinth light-on-darksimas. I think that we are dealing with a period of experimentationagain as the tile manufacturers developthe sima of the Classical period. A brief look at some of the other tiles of the secondhalf of the 6th century.Two types of antefix occur at Corinth around the middle of the century. We have already seen one type 69 CorinthIV, i, S28, pp. 22, 69, fig. 19; GFR, p. 91, no. 77; Le Roy, 70 FS 858 from the Tile Works is unpublished.

1967, p. 126, note 3.

FS 886, also from the Tile Works, is unpublished. Corinth IV, i, S27, pp. 22, 69, fig. 21. Another is FS 52, p. 22, notes 2, 3, p. 73; GFR, p. 90, no. 76; Le Roy, 1967, p. 126 (here read S.52 for S.25). 73 Le Roy, 1967, pp. 122-123, 125-127, pls. 43, 102. 74 Le Roy, 1967, pp. 125-126. 75Le Roy, 1967, p. 112; GM, pl. 18:7; H. Goldman, "The Acropolis of Halae," Hesperia 9, 1940, (pp. 381-514), pp. 440-442, figs. 100, 102-104, p. 454; Heiden (1987, pp. 80-83, 92, 95) would date the temple around 520 B.C. 76 See footnote73 above, in particularp. 127. 71 72

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

61

used on the Temple of Apollo (above, pp. 55-56). There is another of similar size which may also have been used on the temple (P1. 7).77 Insteadof the spreadingstems or tendrils with a three-leavedpalmette in the lower part of the antefix, there is a lotus with a threepetaled flower. The two types show a similar developmentof the upper palmette,the leaves becomingless stiff, more curved. The height of the lower part increaseswith time. Above the band the tendrils rise higher. Eyes alongsidethe band, addedleaves, and a bar between the palmette heart and leaves are variationsfound. The first type is less popular, perhaps because it allows for less enhancement.It is eventually supersededby the other, which is found in much the same areas as the earlier pentagonal antefixes. The latter becomesthe type commonlyused in the light-on-darkantefixes. The two types in more developedform are seen respectivelyin FA 15 (P1.7), of the same type as FA 16 (P1.6), and FA 19 (P1.8), of the same type as FA 3 (P1.6), from Corinth.78AnotherCorinthantefix, FA 430 (P1.8), is a flatter,stiffer,but well-developedexample of the secondtype. Amongthe later examples is one from the Demeter Sanctuary,FA 452 (P1.8).8o It differssomewhatfrom the othersin having an angular, rather than rounded arc between the palmette heart and leaves and a rather pointed center palmette leaf. A combinationeaves and cover tile with a palmette antefix from the Tile Works, FA 581 (P1. 8),81 is a good example of the type which evolves into the light-on-darkform used with the cyma-reversasimas. It is very similarto a group of light-on-darkantefixeswith nine-leavedpalmettesfoundin the area to the north of Temple Hill. One of the indicationsof its late date is the height of the curveof the tendrilsunder the lowest palmette leaves. Probably this piece from the Tile Works should be dated around 490 B.C. Part of a lion's head spout fromthe Tile Works, FS 883 (Pl. 8),82is an interestingpiece. It has a very fine, smooth, cream slip and is very carefully modeled. In addition to more routine pieces the Tile Works has produceda few very fine fragments,an indicationof the quality of work of which Corinthwas capable.Only the upper part of the head is preserved, none of the face. The locks of hair of the mane are in three tiers. The front row is grooved and is painted darkpurplish red and black on cream.The two other rows are darkpurplish red. The closest analogy seems to be the Peisistratidtemple on the Akropolis.83I would put the piece from the Tile Works about the same date. CorinthIV, i, A3, pp. 12, 47, pl. I; GFR, p. 152, no. 39; Heiden, 1987, pp. 76-78, pl. 9:2. CorinthIV, i, FA 15, pp. 12, 48-49, fig. 3; FA 19, pp. 12, note 5, 49; for FA 15, see also GER, p. 148, no. 18. 79 FA 430 is unpublished. 80 R. S. Stroud, "The Sanctuaryof Demeter and Kore on AcrocorinthPreliminaryReport 1: 1961-1962," Hesperia 34, 1965 (pp. 1-24), p. 20 and note 50, pl. 9:d. 81 FA 581 is unpublished. 82 FS 883 is unpublished. 83 For the Peisistratidtemple see Travlos, pp. 143, 146, fig. 197. Anotherlion of similar type but earlier is from Perachora, now in the Boston Museum of Fine Arts: H. Gabelmann, Studien zum fruihgriechischen Lowenbild, Berlin [1965], pp. 48-51, pl. 5, dated there ca. 580 B.C.; M. B. Comstockand C. C. Vermeule, Sculpturein Stone, The Greek,Roman and Etruscan Collectionof the Museum of Fine Arts, Boston [1976], pp. 9-10, no. 15, dated ca. 550 B.C., probablythe betterdate. 77

78

62

MARY C. ROEBUCK

As is the case with the antefixes, the ridge tiles of the second half of the 6th century84 show less stiffness and a few added features such as a palmetteof three leaves between the tendrilsbelow, on one example, and addedeyes and leaves. As has already been noted, the late 7th and much of the 6th century was a period of considerableactivityin building and in the developmentof the ornamentationof the buildings. It was also a period when there was constantmovementbetween areas, the result of both tradewith and travel to and fromsanctuariesand communities.The distanceswere not great. Consequently, people were exposed to a variety of styles and ideas, which exerted influences in many ways. In conclusion I would like to show two tile fragmentsfound at Corinthwhich illustrate some of this. The first piece is a small fragment,FS 659 (P1. 8),85 with a small torus moldingbelow which is a straightfascia. The upper moldingis decoratedwith verticalbandsof alternating purplish red and cream. Below them is a narrow black horizontalline and then three wide bands,two creamand one purplish red. Finally at the bottomof the fragmentthere is just a traceof a guillochewith incisedoutlines in darkpurplishred on the creambackground.The fabric, while undoubtedlyCorinthian, has a hard, very smooth, rather polished surface. This type of surfaceis found on the terracottapedimentalsculptureof the Archaicperiodat Corinth.86It is also seen on other pieces of tile such as the lion's head spout, FS 883, from the Tile Works discussedabove and on some 6th-centurypottery.87One is struckwith the resemblanceto two pieces from Delphi which Le Roy has put with a group of tiles he connects with Sicily.88One has alternating horizontal bands of cream and red, the other similar bandsbut cream,red, and black.The firstalso has a doubletorus molding (a Sicilian feature) decoratedwith vertical cream and either red or black bands, which is not stated. The fabric of one is describedas shiny. There is no reason to suppose that these pieces are Corinthian,but the similaritieswith the piece from Corinth suggests the flow of ideas and designsback and forth. The secondpiece, FA 432 (P1. 8),89 probablyan eaves tile with part of an antefix, also shows outside connections.The lower part is made up of a fascia decoratedwith a single stopped-maeanderpattern in black and dark red on a light background.Above this and set back slightly is another section with one finished side. Its upper part is missing. On the lower portion is a design which resemblesa running spiral with palmettes set in the free spaces.It is not quite a running spiral since each of the two sectionsends in a curvearounda dot. While the parallel clearlyis not exact, the impressionof the whole design is reminiscent 84 For example, CorinthIV, i, p. 16, note 9, and p. 62, R12; GFR, p. 161, no. 4; and ridge tiles like FR 47, unpublished. 85

FS 659 is unpublished.

S, S, Weinberg, "TerracottaSculptureat Corinth,"Hesperia 26, 1957 (pp. 289-319), pp. 293-294, for the smooth polished surface. 87 M. T. Campbell, "A Well of the Black-Figured Period," Hesperia 7, 1938 (pp. 557-611), no. 151, pp. 592, fig. 18, 596. 88 Le Roy, 1967, pp. 86-87, series 32, pl. 29:4, 5, G.32, G.33. 86

89

FA 432 is unpublished.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM CORINTH

63

of that used on the lateral sima of the Olympia building now identifiedas the Treasury of EpidamnosIII and datedca. 525 B.C.90It has a running spiral with a single-maeanderfret patternbelow. The antefix assignedto this building is a typical Corinthianpalmetteantefix of that period. Obviouslyideas are being borrowed. MARY C. ROEBUCK P.O. Box 65 Eastsound,WA 98245 90A. Mallwitz, Olympiaund seine Bauten, Munich [1972], pp. 169-170, fig. 130; H. K. Siisserott,"Unteritalisch-ionischeBaukeramik,"OlForschI, Berlin 1944 (pp. 136-145), pp. 140-142, fig. 40, pls. 45, 46.

PLATE 5

FA 101

FT 209, eaves tile from Protocorinthianroof

FA 204

FA 237

FA 553

FA 446

FA 422

FA 547

FA 559

FA 24

PLATE 6

FT 231

FA 16

FS 101 FR 3

~~~~~~~I _~~~~~~~~~~~~

FS 44 FS 862

FS901

FS 901

FS426

426 ~~~~~~~~~~~FS

PLATE 7

FS1010

i

FS

28,11': i1

FS 88

FS 2078

FS921~~~~~~~~~~~~~ I

_

_

_U

U~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ FA

F_1 MAYC

OBC:ACACACIETRLTRAOTSFO

OIT

FA452~~~~~~

FA452

FA 18

S69

430

MAYC

REUK:ACAI

FA

R58 ERAOTS

RMCOIT

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI (PLATE 9)

HALIEIS The Sanctuaryof Apollo at Halieis now lies submergedin the bay of the moderntown of Porto Cheli in the Argolid, where underwaterexcavationstook place in 1970, 1971, and 1973.1Within the sanctuaryare foundationsfor severalbuildings.2The two that concernus here are the Temple of Apollo, identifiedby inscription,3and Building 2, first thoughtto be a Stoa, which I now identify as a secondtemple. The two buildings stoodside by side, with their outsidewalls less than one meter apart, and were in use during the same period. When the large quantities of terracottarevetment were analyzed, it became apparent that two nearly completeroof-tile systems could be reconstructed,and each couldbe assignedto the buildingfor which it had been designed.4The tiles were found in a layer coveringthe foundationswith almost no mixing of the two types. Temple II had a Laconian-typeroof (Fig. 1). The Temple of Apollo had a roof that I first called Corinthiantype, to distinguishit from the Laconianroof (Fig. 2). Since a roof of such a design has not been found at Corinth,the term "Corinthian"should be avoided.I refer to it here as the "Halieis-style"roof.5 'The following abbreviationwill be used in this article: Yialouris, "Problems" = N. Yialouris,"ProblemsRelating to the Temple of Apollo Epikouriosat Bassai," Greeceand Italy in the Classical World,Acta of the XI InternationalCongress of Classical Archaeology,London, 3-9 September 1978, London 1979, J. M. Coldstreamand M. A. R. Colledge, edd. (pp. 89-104) 2 For a site plan and plan of the Apollo sanctuary,see M. Jameson, "The Excavationof a Drowned Greek Temple," Scientific American, Oct. 1974 (pp. 111-119), pp. 113 and 114-115; idem, A/A?T 26, 1971 B' 1 (1974; pp. 114-119), p. 117, plan 1; idem, "Halieis at Porto Cheli," ColstonPapers, 23, Proceedingsof the Twenty-thirdSymposiumof the ColstonResearch Society Bristol 1971, London 1973, pp. 219-229; idem, "PortoCheli," BCH 89, 1966, pp. 787-791; idem, "Excavationsat Halieis (Porto Cheli) 1973," A/A?T 29, 1973-1974, B' 2 (1979), pp. 261-268. 3 M. Jameson, "The SubmergedSanctuaryof Apollo at Halieis in the Argolid of Greece,"The National GeographicSocietyResearchReports 14, 1982 (pp. 363-367), p. 364. 4 Not all tiles from the Apollo Temple examined during excavationwere inventoried.More than half of the tiles from the roof were recovered.The author'sstudy (1977) of architecturalterracottasfrom Halieis and a catalogue of the fragments will be published in the projectedvolume Halieis, II, The Apollo Sanctuary, Architecture,edited by Michael Jameson. See also N. Cooper, 1983. The author's work was supportedby grants from the Minnesota Foundation, the American Philosophical Society: Penrose Fund 1979, and the University of Pennsylvaniaand Indiana University excavations. 5See footnote3 above. N. Winter has suggested"Argive-style"roof (pp. 13-32 above).

66

NANCY K. COOPER

`11~~~~~~~ ~HALIEIS

LACONIAN

0_,

FIG.

.

5m

1. Halieis, Laconian Roof, eaves and ridge, with back of disk

FIG.2. Halieis,the Templeof Apolloro f

FIG. 2. Halieis, the Temple of Apollo roof

-p

68

NANCYK. COOPER

THETEMPLE OFAPOLLO: THEHALIEIs ROOF Long submersion in the sea has damaged the surfaces of the tiles, but the technical design of the roof of the Apollo Temple can still be reconstructedalmost in entirety. Antefixeshad flat smoothfaces which rose to three peaks, the centralpeak being higher than the two on either side (Fig. 3, P1. 9:b). The bottomsurface of the antefix face was cut away at an obtuse angle, trimmedto fit the contourof the eaves tiles on which it sat. The faces of the eaves tiles were not of equal height along the length of the eaves of the roof, but followed the contours of the pan tiles, increasing in height where two eaves tiles abutted, directlyunder the antefix (Fig. 10). The triple peaks of the antefixesraisedthe faces higher than the covertile behind. One fragmentaryantefix has a flat bottomand could have been used at the cornerof the roof (Fig. 4, bottom). The antefix coveredno seam between pan tiles. It would have sat against the back edge of a sima, which was probablyno more elaboratethan a pan tile with a turned-upedge. Only two small fragmentsof eaves tiles have been inventoriedand both are from the lower cornersof the tiles (Fig. 5, top). The eaves tiles projectedbeyondthe edge of the wall plate by 0.10 m., and the undersideswere moldedinto a sharp overhangingdrip edge. This overhangingportion of the tile could not have been doweled or rabbeted.There is no indication of the method by which they were attachedto the roof. No evidenceof dowel holes was found on any tiles from this roof. The two small fragmentsof eaves tiles indicate the methodby which the antefixeswere secured.On the top surfaceof the tile, at the edge which abuts the next eavestile, a triangularlump of clay formeda wedge against which the antefix was securelybraced.The undersideof the antefix was hollowed to fit (Fig. 5, center). This systemmakes doweling of the antefix unnecessary. All pan tiles were flat bottomed.The undersideswere cut backon the lower edge to abut and overlapthe next tile below (Fig. 5, bottom).The usual length of the overlapwas 0.08 to 0.10 m. Since the tiles were erodedfrom long submersionin the sea, there is no slip left on the surface.The edges, however, still retain their original shape. Each edge of the pan tiles was smoothly finished, and the long abutting edges rose to sharp peaks. Several complete pan tiles were joined together from fragments (Fig. 6); they measured 0.82 m. long by 0.55 m. wide. The covertiles had cut-backsidenticalto those on the regular pan tiles, and the coversalso overlappedthe tile below by about 0.08 to 0.10 m. One complete cover tile was joined together from three fragments and provides the measurementsof 0.811 m., length, and 0.165 m., width (Fig. 4, top and center). There were no bevels or moldings at the edges of any of the tiles. Saddle-shapedpan and cover tiles folded over the ridge beam; these extended 0.30 m. down the roof on each side. The slope of the roof is easily calculatedfrom these well-preservedridge tiles. The angle at the peak of the roof was 1500, and so the angle at the eaves was 150 (Fig. 7).

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

0

-.0

.175

-

6-

.034

0.25m

FIG.

3. Halieis, the Temple of Apollo roof, antefix reconstruction

69

70

NANCY K. COOPER

HALIEISCOVERTILES

HALIEIS CORNER RECONSTRUCTION FIG.

4. Halicis, the Temple of Apollo roof, covertiles and cornerreconstruction

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

HC589

HCH598

HALIEISPAN TILES FIG.5. Halicis, the Temple of Apollo roof, eaves tiles, antefix construction,and overlappingpan tiles

71

72

NANCY K. COOPER

HC539X=

HC579

FIG.

6. Halieis, the Temple of Apollo roof, pan-tile fragments

Calculation of the module used in constructionof the Apollo Temple produced the "Halieisfoot"of 0.273 m.6 The size of a pan tile expressedin "Halieis feet"is 3 feet long and 2 feet wide. The agreement between the module of the tiles and the module used in the ground plan is confirmationthat the tiles belongedto the building. The longest preserved covertile attachedto an antefix is 0.27 m. At the eaves, it is possible that the pan eaves tiles and covertiles formingthe back of the antefixeswere 1 Halieis foot long (Fig. 8, bottom).A 6 F. Cooper will publish full details in Halieis II. T. Boyd and W. Rudolph ("Excavationsat Porto Cheli and Vicinity, Preliminary Report IV: The Lower Town of Halieis, 1970-1977," Hesperia 49, 1980 [pp. 333-355], p. 340) publisheda preliminarynotice of the foot measurement.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

HC543

HC544

HALIEISRIDGE TILES

FIG.

7. Halicis, the Temple of Apollo roof, ridge-tile fragmentsand ridge reconstruction

73

74

NANCY K. COOPER

FIG. 8. Halieis, the Temple of Apollo roof, two possiblereconstructions,sections

reconstructionwith short eaves tiles gives an overhangof approximately0.50 m. Between the two temples is a drainage channel, the edge of which runs along the side of the Apollo Temple at a distanceof 0.50 m. from the wall, to catchthe run-off from the two temples. The Halieis roofis characterizedby the three-peakedmoldeddecorationon the antefixes and by pan and covertiles of distinctprofilewhich are notjoined to each other. Similartiles have been found elsewhere in the Peloponnese,at Mases,7 Nemea,8 Argos,9and Kombothekra,10at Aigina,1Iand on mainlandGreeceat Delphi (Fig. 9).12At Mases, on the Franchthi headlandin the Argolid,one part of an antefix was recovered.Nothing more is known of 7J. Dengate, "The Archaic Doric Temple at Mases,"Abstractsof Papers 76th GeneralMeeting, ArchaeologicalInstituteof America, 1974, Section IIA, p. 123. 8 S. Miller, "Excavationsat Nemea, 1979 ," Hesperia 49, 1980 (pp. 178-205), p. 178, pl. 39. 9 C. Pfaff (pp. 149-156 below). 10U. Sinn, "Das Heiligtum der Artemis Limnatis bei Kombothekra,"AM 96, 1981, pp. 25-71; K. Muller, "Artemistempelbei Kombothekra,"AM 33, 1908, pp. 323-326. II A. Furtwangler,E. Fiechter,and H. Thiersch,Aegina.Das Heiligtum derAphaia,Munich 1906, fig. 119; Schwandner,1985; E.-L. Schwandner,"Der Altere Aphaiatempelauf Aegina,"NFGH (pp. 103-120), fig. 8, p. 115; E.-L. Schwandnerand D. Ohly, "Aegina,Aphaia Tempel,"JdI 86,1971, p. 534. 12 Le Roy, 1967, pl. 5:2 and p. 28.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

DELPHI

C>1

L111QIIJ

1<

>

L~

P

Lui+1cJ

3

~

<

OLYMPIA

AEGINA

MASES

0202

LzI11II~cI1

KOMBOTHEKRA

DZO/

LiI:I12IIi

W

$

=

[K4;

}

NEMEA

4020

ARGOS

ANTEFIXES 0

FIG.

9. Three-peakedantefixes,comparison

0.5m

75

NANCY K. COOPER

76

.245 .660

CORINTH

.660

ISTHMIA

.175 -

-~~~~~.550

HALIEIS

560

AEGINA

FIG.

10. Antefix and eaves tiles, comparison

this roof;the site is unexcavated.The Mases antefix still retains traces of a dark red color, which is like the coloron the Aigina tiles. The Halieis antefixeswere probablyred also. At Aigina the eaves tiles had decorationalso. The three-peakedprojectionsat the center of the eaves tile formedpseudo-antefixesechoingthe shape of the antefixes,in smaller scale (Fig. 10). This decorationwas not found at Halieis. It would seem that this addedfeatureof the Aigina tiles would make the roof later in date than the Halieis roof, but the ridge tiles at Aigina are of a more primitive type than those at Halieis. The Aigina ridge pan tiles are very narrow, only 0.10 m. on each slope of the roof; those at Halieis, 0.30 m. The Aigina ridge is closer in design to the "Protocorinthian"roof at Corinth (Roebuck, p. 48 above, fig. 1).13 The Isthmiaversionof the Protocorinthianroof is datedslightly later than the Corinth one and has the addeddecorationof one peak on the eaves tile (Fig. 10).14 The reverse situationmay exist with the Halieis and Aigina roofs.The earlierhas the addeddecoration, the later does not. The improvementof the design of the Halieis tiles at the peak of the roof is strongevidencethat it should be dated after the Aigina roof. 13 Robinson,AM, p. 56, fig. 1. At the ridge the "Protocorinthian" roof has a strong horizontalrow of tiles whose dimensionsrepresentthe size of the beam on which they sit. 14 Q. Broneer,Isthmia, I, Temple of Poseidon,Princeton 1971, p. 121.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

77

The Halieis-style roof should representthe next stage of developmentafter the Corinth and Isthmia "Protocorinthian"roofs. The Halieis tiles are not combinationtiles; each pan and cover is a separate piece. This would seem to be a step backwards,but separate tiles would have made the actual fitting of the tiles on the roof more simple and would have eliminated most of the cutting of the upper edges of the tiles that occurs on the "Protocorinthian".There are also design advances:the Halieis tiles were flat bottomedand the ridge tiles more stable. The ridge pan and covertiles were much longer and thereforemuch roof. heavier,in fact three times the length of the 0.10 m. ridge tiles of the "Protocorinthian" The tiles were stabilizedby their own weight, and no dowels or nails were neededto secure them to the ridgepoleof the roof. Of the three-peakedantefixesfound at varioussites, only at Halieis can the foundations of the temple to which they belonged be positively identified and something said of the nature of the building.The temple was long and narrow,dividedinto severalchambers,was orientednorth-south, had an entranceon the south facing the altar and other doors on the west. There were no stone columns and no ashlar masonry.The walls may have been constructedof mud brick above a socle made of roughly worked stones. The temple was not peripteral. To walk along the east side meant passing through a meter-wide space where two roofs nearly met overhead.The landscapeof the passagewayconsistedof a drain. The foundations of the Temple of Apollo seem early, and pottery dating from the second quarter of the 7th century B.C.was found inside.15The roof system of the temple is an early one. The most straightforwardconclusionis that building and roof are part of one building period;however,the potterydoes not necessarilydate the architecture. The three-peakedantefix roof should be original to this building,whateverits date. If it is true, however, that the similar tiles at Aigina and Mases were used on fully developed Doric stone temples and should be datedto the first quarterof the 6th centuryB.C.,then the Halieis roof must have replaced an earlier roof, which was made from a perishable material, such as thatch.This could not have been a simple alterationbecausethe weight of the terracottarevetmentwould have needed a stronger wooden roof structure. In the temple, some stone bases were found that could have supporteda series of wooden columns along the interiorwalls.16 TEMPLEII: THE LACONIANROOF

The remainsof the foundationsof Temple II at Halieis give no indicationof the interior plan, of the orientationof the faade, or of the functionor date of the building. The superstructurewas completelyrobbed;only a few finds similar to those in the Temple of Apollo and a thick layer of shatteredtiles were left behind. The remains of the roof are not sufficient to allow a detailedrestorationof all aspectsof its appearance;some importantfeatures, 5 Jameson (footnote3

above), p. 367. If this is the case, and such a change seems to have taken place at Bassai and Kalapodi as well, perhaps this can be seen as the origin of the interiorcolonnade. 16

78

NANCY K. COOPER

however, are certain. The roof was a typical Laconianroof with a ridge coverand an akroterion disk (Fig. 1). One nearly complete pan tile was pieced together from many fragmentsand was approximately0.45 m. wide and 0.95 m. long (Fig. 11). No full-length covertiles were found, but the length was necessarily the same as that of the pan tile. The covers widen from 0.17 m. at the upper edge to 0.22 m. at the lower and form a flattenedcurve in profile. The dimensionsand profilesof the Halieis Laconianroof tiles are remarkablysimilar to the tiles from the early Temple of Apollo at Bassai (Fig. 17). Of the several geison tiles one slab is preservedwith its full width (Fig. 12). Two fragmentsjoin and measure0.45 m., exactly the width of the upper end of a pan tile. The length is not fully preservedbut measures0.36 m. on the best example. The overhangingportionis slightly concaveon the front face, preservedto a height of 0.15 m. which must be nearly the full height. The underside,meant to be seen from below, carriesa molding.There is a hole in one fragmentwhich could have been used to attachthe geison to the woodworkbelow. There are eleven identifiablefragmentsof the ridge cover.Only two areas of the ridgecovertiles have edgeswhich permitthem to be distinguishedfrom the mass of shatteredtiles recoveredduring excavation:those portionswhich were cut to fit over the covertiles which ended at the ridge of the roof, and the rim which was overlappedby the adjoiningsectionof the ridge cover along the ridge of the roof. The other end of each ridge cover, which overlapped the rim of the next one, may have been plain, with no distinctivemoldings.A more usual design was that with one or more tori along the overlappingedge,17but nothing with this shape was recovered. The overlappedrim of the ridge-covertile was molded in such a way as to provide a completelywaterproofseal (Fig. 13). The rim projectsfrom the body, and the part of the moldingjust at the overlapforms a tiny gutter, which carriesaway water enteringthe joint. The water is directedinto the centerof the pan tile below, and flows off down the roof. Six rim fragmentshave been catalogued,and a compositeof the diametersprovidesan estimate of the size and shape of the ridge cover.It is not semicircularin section:the sides rise straight up 0.12 m. beforethey curveover the top. The approximateheight and width are the same, 0.30 m. Five fragments from the body of the ridge cover contain some portion of an aperture which fit over the tops of the covertiles at the ridge of the roof. Althoughno fragmentof the ridge coverpreservesa completeaperture,there are enough fragmentsfrom differentpoints along the curve to allow a reconstructionof the full arc. The curve agrees exactly with the profile of the flat arc of the top of a covertile. Also, the edge of the apertureis undercutto correspondto the angle of the junction of the ridge coverand covertile, giving the approximate slope of the roof, which is slightly less than 150 (Fig. 14, top). The length of the individual ridge coverscould not be recoveredbut may be restoredas being equal to the width of the upper edge of a pan tile. The joints of the ridge coverwould then fall at the centersof the 17 For example, see W. George and A. Woodward, "Architectural Terracottas,"in The Sanctuaryof Artemis Orthiaat Sparta (JHS Suppl. 5), R. M. Dawkins, ed., London 1929, pl. 27 and p. 132, fig. 100.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

_X_____

Ca

0

FIG. 11.

Halieis, Roof II, reconstructionof pan and covertiles

0.500m.

79

NANCY K. COOPER

80

0

'11~~~~~~

~0

_

_

_

FIG..12.-Halicis,

FIG.

_

0.400Om.

_. Roof

II,

0 reconstruction

M o

12. Halieis, Roof II, reconstructionof geison tile

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

0

FIG.

13. Halieis, Roof II, reconstructionof ridge cover

0.500m.

81

82

NANCY K. COOPER

pan tiles along the ridge of the roof. An arrangementof four tiles on each side of the slope of the roof would make an overhangof approximately0.50 m., since the temple is 6.45 m. wide (Fig. 14, top). No antefixeswere found. Five fragments of an akroterion disk were found. Three fragments are shown on Plate 9:a. None of the fragmentsjoin, and all are badly preserved. No trace of painted decorationremains, and the surfaceof each piece is so damagedby water that only tracesof the decorativemoldings on the face of the disk remain. The back surface of one fragment preservesa trace of a connectingstrut,which probablyconnecteddirectlyto the ridge cover. The face of the disk was decoratedin concentriczones, consistingof tori, flat areas, and moldings.All the zones were probablypainted. Those fragmentswhich have some tracesof the decorativemoldingspreservedgive an approximatediameterfor each zone of molding. A fragment from the outer edge of the disk seems to have broken saw-toothedprojections. The diameterof the disk can be restoredto about one meter (Fig. 15). The preservedmoldings are typical of those which occur on other similar disks, such as that of the Heraion at Olympia, or those from Bassai.18Since these disks have two separatezones of three or four concentrictori, it is possiblethat some zones are not representedamongthe preservedfragments from the Halieis. Mallwitz has divided akroteriondisks into two classes:the Heraion type and a coarse type, which instead of a saw-toothed outer zone has alternating rounded and pointed leaves.19The outer edge of the disk of Temple II at Halieis is not well enough preservedto indicateto which type it belonged.The absenceof any trace of painted decorationis also an obstacle to arriving at a more specific date for the Laconian roof at Halieis. Generally, however,such akroteriondisks are dated to the 6th centuryB.C. The fact that Temple II at Halieis was decoratedby an akroteriondisk is the proofthat it was a temple. Treasuries and possiblyfountainhouses are the only otherbuildingsknown to have been decoratedwith disks, and the foundationsat Halieis are too large to belong to either of these. The Archaic sanctuaryat Halieis thus included two temples, side by side, facing an altar. The design of one roof was in directcontrastwith the other. The roof of the Temple of Apollo was plain, in an early stage of developmentof this type. The Laconian roof was more elaborate.The buildings were in use simultaneously. BASSAI During the Early Archaicperiod,there were three temples in the sanctuaryat Bassai. Two stood together in a narrow valley near the summit of Mt. Kotilon. A little down the slope was the ArchaicTemple of Apollo, known first from the recoveryof decoratedarchitectural terracottasin the area aroundthe Classicaltemple. The exact locationof the Archaictemple was not known until 1970, when the foundationswere recognizeddirectlyto the south of the 18 N.

Yialouris,oTo aKpco-r-pLoToV ' Hpatov 'OXvqi7ras-, in Xapr-r'ptov bOovIV, 1967-1968, pp. 57-65; Yialouris, "Problems,"pl. 34:c.

'ts 'Avao-racratov K. 'OpiaVv-

19A. Mallwitz, "Ein Scheibenakroteraus Olympia,"AM 83, 1968 (pp. 125-146), p. 133.

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

FIG.

FIG.

14. Halieis, Roof II (top) and Bassai (bottom),sections

15. Halieis, Roof II, reconstructionof akroteriondisk

83

84

NANCY K. COOPER

Classicaltemple, duringthe excavationsconductedby Nicholas Yialouris.20Its outlines had always been visible at the site and had been sketchedon Kavvadias'site plan in 1903,21 even though they were not recognized(Fig. 16). The location,orientation,and groundplan of the Early Archaictemple were copied by the Classical temple and providea rare confirmation of the theory that Archaic temple planning had a profoundinfluence on later building. A study of the designof the Early Archaictemple explains many of the anomaliesof the design of the Classical temple, includingthe roof.22 THE TEMPLE OF APOLLO: THE LACONIAN ROOF

The Archaic terracottasconsistedof fragmentsof two sets of Laconiantiles, including antefixes decoratedwith heraldic sphinxes, and akroteriondisks.23More fragmentsof the Archaic roof were found by Yialouris in his excavationsin the 1950's and 1970's.24Yialouris has published a photographof the reconstructedakroteriondisk "A"25which differs substantially from the earlier descriptionand drawings. The "A"disk has the so-called "hallmark"of Laconianpottery,a pomegranatefrieze, which can be datedto the end of the first quarterof the 6th centuryB.C.26 Rhomaiosdistinguishedtwo sets of architecturalterracottasin his study of the disk and antefix fragments,based upon the two differentclays used.27He describedone clay as being fine and dense with a yellow color (type "A"),the other more red in color, rough and gritty ("B").The clay of the tiles, both undecoratedand decorated,found in the later excavationsis easily recognizedas fitting Rhomaios' two categories. But although two types of clay are used, the fragmentsof the two sets were almostidentical,as if one set servedas full or partial replacementfor the other. The fragmentsof "B"tiles greatlyoutnumber"A"tiles at the site. 20 N. Yialouris,<, AAA 6, 1973, part 1, pp. 39-49; idem, "Epyov 1959 (1960), pp. 106-109; Yialouris, "Problems,"pp. 89-104; L. Parlama,<, ALAr26,1971, B' 1 (1974), pp. 142-146. F. Cooper ("Bassai,"in PECS, pp. 145-146) has postulated four phases for the Temple of Apollo at Bassai. Phases I and II on the foundationsdiscussedhere (the Early ArchaicTemple); Phase III directly under the Classical temple (the Late ArchaicTemple), consistingof ashlar limestoneblockscut and reused in the Classical temple and a wall and floor found by Yialouris ("Problems,"p. 95 and fig. 2); Phase IV, the Classical temple. 21 M. Kavvadias,"Der Apollotempelvon Phigaleia,"Comptes Rendus du Congres International d'Arche'ologie I, Athens 1905 (pp. 171-179), p. 175, and K. Kourouniotis,<, 22 "Laconian Tiles from Bassai,"in N. Cooper, 1983, pp. 100-111, and eadem, "The ArchaicSanctuaryat Bassai,"AJA 84, 1980, p. 202.

23 K. Kourouniotis,

<<'AvaO-Ka

E'vKrTALcp?>,'E'ApX

1903 (pp. 151-188),

pp. 151-158;

idem (footnote

21 above), pp. 271-332; Kavvadias(footnote21 above), pp. 171-179; K. Rhomaios, <<'EKTOVf'apXaLoTpov FLyaAlaX , 'ApX'E4 1933, pp. 1-25. Rhomaiosdates the roof to 625-600 B.C. Also, some akrotevao3v (D rion-diskand antefix fragmentshave been publishedin GFR, figs. 56, 59, and 60. 24 See footnote20 above. I must rely on the published photographs,drawings, and descriptionsof the disk and antefix fragmentsfor my reconstructions. 25 Yialouris, "Problems," pl. 43:c. 26J. Boardman,"ArtemisOrthia and Chronology,"BSA 58, 1963 (pp. 1-7), pp. 2-4. 27 Rhomaios (footnote23 above), p. 2.

-

-

1-1.#

.0

-.

-

.

, .

j

42,

IT~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_

*.."t'.~C OW

~~~~~~~~cr

~ ~

~

*-51 1.-b~~~~~~~~~~~~~l *~~~~~

~~-.. ~~~~~~~~~~~~~~~~~~~~

~

~

P

~ ~

U~*(

i

t, _

sLy(

O~~~~~~~~~~~~~~~~~~~~~~t O

~

_

_

_

W

IL.

vo

'o~~~~ t:

'

_

W

W

WVVW

W

W

W

cZ'

.**.4

"-

It~~~~~~~~~~~~~~~~-?-

NO

W

s/

b.,

~~~~~~~~~~~~~~~~~~~~ ?~~~~~~~b%~~0 ~~~ ,*.~~~~~~~~~~L -'

NANCY K. COOPER

86

0A5

OA4

--

0.23 -

0.21

Q2

2

0.035 <

025 ---

BASSAI

HALIEIS

0.5m

0

L FIG.

1

_-_~~==

17. Bassai and Halicis, fragmentsof Laconiantiles

Pan-tile width can be measured from three joining fragments from the upper, overlapped edge of one "B"tile, which also providesa profileof the curvature(Fig. 17). The pan is 0.435 m. wide. The overlappingedge of one covertile is preservedon which the width can be measuredas 0.21 m. The length of the pan and covertiles is not known. Color, both red and black, was found on both the "A"and "B"types of tiles. Some covertiles had red color, othershad black, and the same was true for the pan tiles. Red is less commonthan black,but no reconstructionof color patternis possible. Some fragmentsof the ridge cover of type "A"have been identified.One preservesthe cut-out section that fit over the covertile and has the same type of bevelededge as the ridge cover of Temple II at Halieis. The color is well preserved:solid black with no pattern

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

87

visible. A reconstructionof the roof at the ridge is made possible by these fragments, although some details, such as how the ridge-coversectionsjoined, are lacking (Figs. 18 and 23). Antefixesof an unusual bi-level design were recovered(Figs. 19 and 20:c, d). Antefixes of both types of clay, "A"and "B",were found. There is some slight variationin detail:"A" antefixesshow more fine detail and have fewer filling ornaments(Fig. 20:c). The "B"antefixes have a recessedborder 0.02 m. wide, outlining the entire shape of the antefix, while "A"antefixesdo not. It is possiblethat "A"antefixeswere slightly smallerthan the "B"type. Reconstructeddimensions of the "B" antefix are 0.31 m. wide and 0.245 m. high. Kourouniotesdepictsa wavy line paintedbeneaththe sphinxes in one example and says that the top portion of the antefix, the palmette, was dark red and the back black; the sphinxes' bodies were painted red.28The faces of the antefixeshave flat bottoms,which indicatesthat they sat on the level surfaceof a geison-tile course. The sphinx motif of the Bassai antefixesis similar to that on an architecturalfragment found at Delphi.29 There is also a similar Laconian antefix from Capua, which has two heraldicsphinxes with one head.30The Corinthiantiles from the ArtemisTemple at Corfu are illustrated together with an antefix fragment, which bears a relief design of a seated sphinx in a heraldicpose, its body facing the centerline of the antefix and one paw raised.31 The size of this antefix is close to that of Bassai, but it is of the pentagonalCorinthiantype, not Laconian.All similar antefixeshave been datedto the late 7th-early 6th centuryB.C. The Bassai antefix differsin designfromthose with similar decorativemotifs,becauseit has a palmette perched above the lower half. The palmette is borrowedfrom Corinthian antefixes. Two Corinthian antefixes from Thasos have figures decoratingthe pentagonal face, with a palmetteabove (Fig. 20:a, b).32The treatmentof the figures is like that of those from Bassai: black or dark figures on light ground. Some of the empty space is filled by painted dots in a circle. The Thasos antefixes are from a Corinthian-typeroof, not Laconian. The Bassai palmette seems to be less developedthan that of the Thasos examples but more developedthan the small palmette which decorateda series of Corinthian-type antefixes found at Corinth, Nemea, Kalapodi, and other sites, which have a palmette perched on the peak of a pentagonal face.33This series of antefixes does not have figural decoration(Fig. 21). If comparedto other Laconian antefixes, the Bassai antefixes look strange. Laconian antefixeswere normallyof the semicirculartype, with a well-markedcenterand decoration Kourouniotes(footnote21 above), p. 276 and p. 277, fig. 4. Le Roy, 1967, p. 90, A.185, pl. 31. The fragment is not Laconian and is not even clearly an antefix. Le Roy dates it to the end of the 7th centuryB.C. without excludingthe possibilitythat the type was perpetuated in the first part of the 6th century.There is also a sima fragmentfrom Delphi with heraldicsphinxes. 30 H. Koch, Dachterrakottenaus Campanien,Berlin 1912, pl. 12:3. 31 KorkyraI, p. 135, fig. 107. Two brokenantefixes are in the museum at Corfu, nos. 447 and 449, labeled "fromArtemis temple 6th C.";GFR, pl. 21, fig. 68. There is a variantwith lions replacingthe sphinxes. 32 P. Lemerle, "Chroniquedes fouilles (1934)," BCH 59, 1935 (pp. 234-309), pp. 293, 296-297, and pI. XVIII. 33 This series of antefixes was arranged for the author's paper, "Early Development of the Corinthian Antefix"(84th General Meeting of the ArchaeologicalInstituteof America,Philadelphia, 1982, abstractAJA 87, 1983, p. 230) and "The Developmentof the CorinthianAntefix"in N. Cooper, 1983, pp. 52-73. 28 29

I

FIG.

18. Bassai, reconstructionof eaves, covertile, and ridge cover

89

0.5m

0

FIG.

a

19. Bassai, reconstructionof eaves

b

/'

c

d

FIG. 20. a, b: Thasos, two Corinthianantefixes.c, d: Bassai, "A"and "B"antefixes. Scale 1:5

90

NANCY K. COOPER

CORINTH

~~ ~ATHENS

DELPHI

NEMEA

NEMEA I \ I'

TIRYNS

ANTEFIXES 0.5m

0 F

a

FIG.21. Early palmetteantefixes

ANTEFIXES UNATTACHED

ATTACHED PHASE

PHASE

I

IV

IV

A

zI

Fn

m ~~0.5

FIG. 22. Development of the antefix

0

FIG.23. Bassai, ArchaicTemple of Apollo, reconstruction

5.0,~~&O

ARCHAIC ARCHITECTURAL TERRACOTTAS FROM HALIEIS AND BASSAI

93

of concentriczones or radiatingpatterns.The antefix of the Heraion at Olympia has raised molded decoration,a further developmentof earlier designs incised and painted on a flat surface. The only element of the decorationon the Bassai antefix that even remotely resemblesthese patternsis the placementof one of the wreaths or spheresmarkingthe center. The palmette at the top completely changes the semicircularshape. The Bassai antefix looks like a Laconian antefix disguised as Corinthian. It must come near the end of the periodwhen the Laconian-styleroof was used on templesand betweenphases III and IV on the left side of the chart of antefix developmentin Figure 22. The original surfacesof the "A"tiles are well preserved;the "B"tiles appear worn by comparison.The excellentpreservationof the surfaceand the colorsof the "A"tiles suggests that they were not on the roof long enough to undergosevereweathering. The presenceof pieces representingtwo nearly identical but distinct revetmentscan be explained by the assumptionthe roof was repairedby replacingonly damaged"B"tiles with "A"tiles. This would also accountfor the weatheringof the "B"tiles and the fact that tiles of this type are found at the site in greater numbers than those of type "A".Both types show the result of burning. Van Buren and Rhomaios claim that roof "A"preceded"B".Both seem to have based their conclusiononly upon the style of disk and antefix fragments,without considering the ratio of type "A"to "B"tiles at the site.34The "B"roof was the original roof of the Archaictemple, the "A"roof a replacement,most likely of only part of the roof. The foundationsof the temple and surroundinglandscapeindicate a simple structure with mud-brick walls on a roughly worked stone socle (Fig. 23). Three small (0.30 m. diameter) stone columns were found, none taller than about 0.60 m. The columns have flanges projectingfrom them, as if they were added to the temple interior when terracotta revetmentreplaceda lighter roof.35 As has been remarkedabove,the profilesof pan and covertiles from Bassai are remarkably similar to those of the Laconian roof at Halieis. The dimensionsof the tiles are also similar (Fig. 17). Since the building sizes and the tiles are so similar, the missing element at Bassai, the ridge cover, probably would have been similar to those at Halieis also. The Halieis akroteriondisk was about 1.0 m. in diameter,as are the disksfrom Bassai (Fig. 23). The architecturalterracottasof the Archaictemples at Halieis and Bassai are of special interest because they can be assigned to the foundationsof specific buildings. In size and proportion,all three templesare comparable,and all have the same north-southorientation. What little is known aboutthe superstructureof the threebuildingsis also in agreement. NANCY K. COOPER CORNELLUNIVERSITY

Departmentof Classics Goldwin Smith Hall 120 Ithaca, NY 14853 34

GFR, pp. 18-19. Rhomaios (footnote23 above), p. 2.

35

Seefootnote16 above.

PLATE 9

s~~~~~~~~~~~~~W Z

P.

.~~~~~~~~~~~~~~~~~~~~~~~~

4

i

a. Halieis, Roof II: three akroterion-diskfragments.Scale ca. 1:2

V)

E

b. Haijeis, Temple of Apollo roof: antefix

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE NOTES DE TYPOLOGIE ET D'HISTOIRE (PLATES 10, 11) A la memoired'A. Mallwitz.

U MATERIEL ARCHAIQUE mis au jour 'aArgos, a l'Heraion d'Argoset a Epidaure,1j'ai choisi de presenterici les principauxtypes de simas et d'antefixes,ceux qui contribuentle plus efficacement'a une meilleure connaissancede la typologie des terres cuites architecturales,de leur productionet de l'histoire des sites. Si pres de Corinthe et d'Isthmia,il faut s'interrogersur l'absencequasi totale de pieces du vIIesiecle. En revanche, celles du vIe siecle attestent l'originalite des productionslocales, probablementliee a une importanteactivite edilitaire de la cite d'Argos. Quant aux terres cuites d'Epidaure,elles D

I Cette premiere et rapide presentationd'un materiel inedit qui merite une publication beaucoupplus detaillee m'offre l'agreable occasion de remerciertous ceux a qui je dois d'avoir pu l'etudier. L'importante collectionmise au jour par W. Vollgraff s'est beaucoupaccrue depuis que l'Ecole Frantaise d'Archeologiea recommencea fouiller a Argos en 1952. Il serait trop long de citer tous ceux qui ont bien voulu me cederleurs droitsde publication,et je ne nommeraiici que P. Aupert,J.-F. Bommelaer,P. Courbin, D. Feissel, P. Marchetti, M. Pierart,J.-P. Sodini et J.-P. Thalmann dont les decouvertesont contribuea une meilleureconnaissance des toitures archaiques. Je me suis aperue en 1970 que le fonds ancien du Musee d'Argoscomportaitaussi un lot de terrescuites de l'Heraion, probablementreunies lors des fouilles du General Gordon,de A. R. Rangabeet a l'occasionde divers nettoyages.Avec la bienveillanteautorisationdu Prof. J. R. McCredie, alors Directeur de l'American Schoolof ClassicalStudies,du Prof.J. L. Caskeyet du Prof. P. Amandry,alors Directeurde l'Ecole Frantaise d'Archeologie,j'ai entreprisde reunirce materieldisperse.Les Directeursdu Musee National, MM. Callipolitis et Yalouriset Mme0. Alexandri-Tsachou,ainsi que MmesKrystalli-Votsiet Rhomiopouloum'ontreserve le meilleur accueil. Avec beaucoupde devouement,MmeI. Trianti a identifie le materiel dans les reservesau cours des annees 1970-1973. MmeSR. Colonia et V. Katsouli m'ont tres efficacementaidee dans les dernieres phases de l'etude. Par ailleurs, M. Ch. K. Williams, II, Directeurdes fouilles de Corinthe,m'a toujoursreque avec la plus grande cordialite.Je lui dois, ainsi qu'a R. Mason, l'identification,dans le Musee de Corinthe, d'un autre lot de terres cuites de l'Heraion qui fut transporteau Musee d'Argosen 1973 afin que l'etude du materiel regroupefuitplus aisee. Le Prof. R. Martin avait entrepris en 1940-1942 l'etude des terres cuites architecturalesd'Epidaure. Avec la generosite que lui connaissenttous ses eleves, il m'a propose de reprendrece dossier, avec l'accord bienveillant de la Societe Archeologiqued'Athenes. Lorsqu'en 1974 le Prof. V. Lambrinoudakisa repris la fouille du Sanctuaired'Apollon Maleatas, cet accordest devenu un appui efficaceet genereux puisqu'il me permet de presenterici une piece archaiquede ce sanctuaire. Ce travail n'aurait pas ete mene a bien sans les autorisationset l'aide active et precieusedes Ephores des Antiquites d'Argolide, MmesDeilaki, Karaghiorga,Krystalli-Votsi, Dimacopoulou et Pachyghianni, et des EpimeletesA. Archontidou,C. Barakari,Ch. Kritzas, H. Palaiologou-Kourachaniet Ch. Pitteros. Les dessinssont l'oeuvredes architectesK. Kolokotsaset N. Papayannakos,la plupartdes photoscelle de Ph. Collet. Que tous ceux que j'ai nommesou non, ceux qui m'ont ouvertleurs collectionset montreleur materielde fouille, ceux dontj'ai sollicite l'avis par lettre ou lors de conversationsveuillentbien trouverici l'expressionde ma profondegratitude.

96

MARIE-FRANQOISE BILLOT

refletentassez bien ce que nous connaissonsjusqu'a presentde I'architecturedes deux sanctuaires 'aI'epoquearchaique. I. QUEL MATERIEL POUR LE Vlle SIECLE?2 1.1. Le materiel actuellementdisponiblene semble comporteraucune piece que l'on puisse surement dater du VIIe siecle. Ce peut etre dufaux circonstancesdes fouilles: 'a 'Heraion comme ai Argos, les archeologuesdu siecle dernieret du debut de celui-ci n'ont manifestement conserveque les terres cuites decorees.A supposer que les sites aient livre les restes de toits
AH I ApollonPythe'en

= Ch. Waldstein et al., The Argive Heraeum I, Boston/New York 1902 = W. Vollgraff, Le sanctuaire d'Apollon Pythe'en(Etudes pe'loponne'siennesI), Paris 1956 Broneer,Isthmia I = 0. Broneer,Isthmia, I, Temple of Poseidon,Princeton 1971 = J. J. Coulton, The Architectureof the GreekStoa, Londres 1976 Coulton, Stoa Friuharchaische Baukunst = A. E. Kalpaxis, FriuharchaischeBaukunst in Griechenland und Kleinasien, Athenes 1976 Lauter H. Lauter, <>= A. Mallwitz, <>, AA (JdI 96) 1981, p. 599-642 = P. Amandry,<>, Hesperia 21, 1952, p. 222-274, pl. 61-72 = E. 0stby, <>, 0stby, <> OpAth 16, 1986, p. 75-102 = Th. Wiegand, Die archaischePorosarchitekturder Akropoliszu Athen, CasPorosarchitektur sel/Leipzig 1904 = G. Sotiriadis,<<'AvacKa4al Sotiriadis eIv0epp?>>, 'E4'ApX 1900, col. 161-212 = K. Wallenstein, KorinthischePlastik des 7. und 6. Jahrhundertsvor Christus, Wallenstein Bonn 1971 3En dernierlieu P. Amandry,?Surles concoursargiens?,Etudes argiennes (BCH-Suppl. VI), Paris 1980 (p. 211-253), p. 236-238. Heiden (1987, p. 22) proposed'expliquerl'absencede toit protocorinthiena l'Heraion par les mauvaisesrelationsd'Argoset de Corintheau temps de Phidon. Encorefaudrait-il que les dates du regne de Phidon soient assurees: il paralt preferablede les placer dans le premier quart ou tiers du vIe siecle. Et puis, quel que soit l'etat des relationsentre cites, il fallait que les batimentsfussent couverts...orles toits <<protocorinthiens>> ne sont pas tous realises en argile corinthienne, ni meme necessairementpar des artisanscorinthiens. 4 Les fouilles recentes ont beaucoup enrichi notre connaissancede l'agora, de,ses abordset des itineraires de circulation au vIe siecle. Elles n'ont pas modifie les bilans dresses en 1979 par P. Aupert et E. Protonotariou-Deilaki, l'un dans <> ASAtene60, 1982, p. 2132, I'autredans <<'AWro 7o 'ApyogToV80V Kal 70V at. r. X.>>,ibid., p. 33-48. -

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

97

encore atteint les principales installations de cette epoque. Mais il apparait aussi que les amenagementssuccessifsde la ville ont profondementbouleverseles etats precedents. Enfin,jusqu'au vIe siecle, les batimentsde l'Heraion ont du se limiter 'a la terrasse<> du milieu du vIe siecle mis au jour par P. Amandryet J. L. Caskey(<>, p. 225-226, note 19);Tomlinsonlui-meme,qui proposeque la peristasis ait ete ajoute a un premier temple (op. cit., p. 234-236); Wright (op. cit., p. 191) sans referencea Bergquist.La zone non dallee offre dans la directionEst-Ouest une longueurdisponiblede 25 m environ.Une peristasisne paraltpossibleni sur cettelongueur,ni a cetteepoque (ci-apres,p. 98-100). Il faut envisagerplutot un oikos un peu allonge. Quant a la maquette votive en terre cuite, elle ne peut etre invoquee ni comme reproduction,ni meme commepreuve directede l'existenced'un temple aux environsde 700 av. J.-C. car les dedicacesde ce genre sont frequentesdans toutes sortes de sanctuaires,ceux d'Hera notamment: I. Trianti, <>, AM 99,1984, p. 113-119, pl. 19; H. van de L6cht,?Ein archaischesArchitekturmodell aus demOrthia-Heiligtumin Sparta>>, ibid.,p. 145-153. Tout au plus peut-ondire que s'il y avaitun temple vers700 lorsquela maquette- la seule qui soit de fabricationargienne(P. Courbin[ci-dessusnote 5], p. 248) a t fa~onn&e, elle peut en avoir 'te une image, pas nesessairementfidele. 7 C. Weickert, Typen der archaischenArchitektur,Augsburg 1929, p. 42-43. K. Schefold,MusHelv 3, 1946, p. 88-89. W. B. Dinsmoor, The Architectureof Ancient Greece,3e ed., Londres (...) 1950, p. xi, 53, note 1, p. 55-56. FriuharchaischeBaukunst, p. 44. G. Gruben, Die Tempel der Griechen, 2e ed., Munich 1976, p. 105: deuxieme moitie du Vlle siecle. I. Beyer, AA (JdI 92) 1977, p. 53, fig. 6 et 7. Mallwitz, ?Kritisches>>, p. 634-635; idem, ?Osservazionisull' architetturanella Grecia dei secoli VIII e VII a. C.>>, ASAtene59, 1981 (p. 81-96), p. 92, etc. Admise par E. 0stby (?Tegea>>) sans discussionprealable,cette date haute, en tout cas nettement anterieure au temple de Hera a Olympie, me paralt influencer par trop son raisonnementsur l'architectureet la date du temple archaiquede Tegee. Il tient aussi un trop large comptede la restitutionhypothetiquede l'Heraion par A. Kalpaxis (Friuharchaische Baukunst,p. 42-46, fig. 24-29).

98

MARIE-FRANQOISE BILLOT

succedede peu a la mise en place du dallageet que l'on n'envisageaitpas de predecesseur.Par ailleurs, d'apresVitruve (IV.I.3), le temple d'Hera 'aArgos etait le premieroCul'on eut fait usage de l'ordredorique.Mais ce n'etait pas necessairementun temple periptere.... Templespe'ripteresdu VIle siecle De fait, en quelques annees de recherche, le nombre des temples peripteres du vIIe siecle a nettementdecru. II est maintenantsu^rque les premierstemples de Corinthe,d'Isthmia et de Nemee (vers 560) ne l'etaient pas.8 II est desormaiscertain que le megaronB de Thermos etait dejadetruitet qu'on en avait nivele les ruines lorsqu'ona disposedes bases de pierre alentour, mais 'a un niveau plus eleve et sur un trace absidal d'axe different9: en d'autrestermes, le temple B n'a jamais existe. L'actuel temple C est hellenistique, archaisant, du moins sa peristasis.10Contrairement'a ce que l'on avait d'abordcru, le premier temple en pierre d'Apollon a Eretrie n'avait pas non plus de peristasis.11Tel qu'il se presente aujourd'huien Grece continentale,le paysage architecturaldu vIIesiecle offriraittout au plus trois temples peripteres,celui d'AthenaAlea 'aTegee12et deux temples d'Hera, l'un construit'aOlympie vers 600 av. J.-C.,13l'autre 'aArgos.14 Observationsdes fouilleurs sur les couchesde terre accumuleessur la terrasseet sur le materielqu'elles contenaient,implantationdu stylobate,blocs laisses bruts dans leur partie inferieure, facture des tambours ou bases travailles au tour, tenons de bardage, trous et canaux de bardage en U,15anathyrosesetroites et faces de joints concaves,J. C. Wright a 8 Robinson, TH, p. 228; Robinson,NFGH, p. 247. Mallwitz, <>, p. 635-637; idem, <>, AA (JdI 97) 1982, p. 149-157. p. 621-624. B. Wesenberg,?ThermosB 1>>, zu Thermos B>>,AA (JdI 95) 1980 10FriuharchaischeBaukunst, p. 47-48. B. Schmaltz, <> p. 633-634. l Mallwitz, ?Kritisches>>, 12 0stby, ?Tegea>>. Mais il n'est pas absolumentassure que le temple archaiquede Tegee etait periptere. 13 A. Mallwitz, Olympia und seine Bauten, Munich 1972, p. 138. H.-V. Herrmann, Olympia, Heiligtum und Wettkampfstatte,Munich 1972, p. 92-93. 14 Le temple de Hera a Samos reputune peristasis (Heraion 2) lors d'un remaniementau tournantdu vlle (ci-dessus note 7), au vie siecle: Mallwitz, ?Kritisches?,p. 624-633, fig. 23-27; Mallwitz, <> p. 86-91, fig. 3-5. Certains mythes faisaient venir d'Argos l'effigie de culte samienne (Pausanias, VII.4.4; Athenee, xv.672 a). Est-ce un hasard si au nombredes premierstemples peripteres,trois reviennentpresque simultanement a Hera? Peut-etre faudrait-il ajouter a la liste ceux que I. Beyer replace sur la fondation II - l'architectureH - qu'il date Dorpfeld, le ?PeripterosI? qu'il date vers 650 av. J.-C. et le <>, Athenatempels alten des Reliefgiebel de 625-600: <
TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

99

reuni plusieurs indices qui invitent a dater le temple periptered'un siecle environ apr's la terrasse.16D'autres traits d'architectureont paru anciens,la crepisreduiteau stylobateet la possibilite que les colonnes aient ete de bois. Le dallage de la terrassene suffit pas 'aexpliquer que le soubassementdu temple se soit limite au stylobate,puisqu' aussi bien celui-ci est etabli pour partie sur la zone de terre laissee par le temple precedent.Mais une crepis a un degre n'est pas necessairementnon plus un trait de caracteretres ancien: les stylobatesdes temples archaiques Nord et Sud de Kalapodi ne sont fondes, pour partie, que sur une euthynteria;17or ils datent des annees 570-560. Quant aux colonnes de l'Heraion d'Argos, elles pouvaientetre en bois sur base de pierre, ou en pierre, ou les unes en bois et les autres en pierre. Or celles des temples Nord et Sud de Kalapodipresentaientla meme alternance de materiaux, et les chapiteaux en pierre y sont de dimensionsvariables.18L'idee s'impose progressivementque les premierstemples peripterespourraientbien n'etre pas anterieurs au premier quart du VIe siecle.19

Chapiteaux Depuis l'etude de P. Amandry,20ni les nouvellesrecherchesde P. de La Coste-Messeliere fondees sur les proportionset la facture des chapiteaux, ni l'approchestylistique de B. Wesenberg - qui souligne que les plus anciens chapiteaux ne peuvent etre anterieurs au dernier quart du vIIe siecle -, ni les classificationsmathematiquesde J. J. Coulton n'ont permis de preciserla date des chapiteaux archaiquesde l'Heraion,21faute, il est vrai, premiere moitie du vIe siecle (pour une nouvelle datation du chapiteau de Marmaria vers le milieu du Vle siecle, voir E.-L. Schwandner,NFGH, p. 116-120; Schwandner,1985, p. 117). Quelques canaux de bardage en U existent dejaau premiertemple d'Isthmia,mais les raresanathyrosessont tres larges: Broneer,Isthmia I, blocs nosAR 41, fig. 18 et AR 79, fig. 47. Sur ces questions,voir en dernierlieu R. F. Rhodes,<), Heraion d'Argos-templede Tegee-temple d'Olympie, paralt tres aleatoire. L'emprise des trois edifices au sol etait certainementtres diff6rente. Trop de variables interviennenta cette epoque d'experimentationpour que le rapportdu diametreinferieura l'entraxepermetted'etablirune chronologieassuree.A l'Heraion d'Argos,il s'etablita 78-80/350, soit 1/4,37 pour le temple, a 70/292 soit 1/4,13 pour la colonnade interieure de la stoa Nord (Coulton, Stoa, p. 29 et 37), a 60/235 soit 1/3,90? pour la colonnadeexterieure selon Kalpaxis (Friuharchaische Baukunst, p. 84). L'entraxe de la colonnadedes temples est sensiblementle meme a Argos et a Olympie, pour un stylobatelarge de 1,34 m a Olympie, et 1,05 m a l'Heraion d'Argos. 19 Cf. par ex. Felsch, 1987, p. 24. 20 ?Observations>>, p. 229-235. 21 P. de La Coste-Messeliere, <>, completerpar P. Aupert, ?Un edificedoriquearchaiquea l'emplacementdu stade>>, Etudes delphiques(BCHSuppl. IV), Athenes/Paris 1977 (p. 229-245), p. 234-240, oiu l'on trouvera les principaux rapports de proportionsdes chapiteaux D et E de l'Heraion. B. Wesenberg, Kapitelle und Basen (BonnJbb,Beih. 32),

100

MARIE-FRANQOISE BILLOT

d'exemplairesprecisementdates par un contextearcheologique.Devant l'extremedifficulte que presente le classement des chapiteaux de cette epoque, on invoque generalementles diversitesregionales,ou plus exactementla necessited'etudesregionales.22Toutefois, aucun particularismene se manifesteencoretres nettementdans le Nord-Est du Peloponneseoiule materiel disponiblereste de toute facon epars et prive de contextearcheologique.Pourtant, confirmantles indicationsde B. Wesenberget de J. J. Coulton en faveur d'une chronologie plus basse, les recentespublicationsde chapiteauxde Corinthe(VIesiecle), d'Egine (ca 575) et de Kalapodi (deuxiemequart du VIC siecle),23ainsi que la revisionde ceux de Tirynthe et du premiertemple de Marmaria- qui aboutita les daterrespectivementau debutet vers le milieu du vIe siecle24- invitent 'arevenir sur les dates hautes initialementproposeespour ceux de l'Heraion. E.-L. Schwandnera signale que le gorgerinlisse des chapiteaux C, D et E et la partie superieure, egalement lisse, de la colonne de l'edifice Nord-Est25pouvaient avoir recu une couronnede feuilles en bronze.26Cette technique precedecertainementl'integration des feuilles 'ala sculpturedes chapiteaux, mais elle a pu durer au-dela : le chapiteau D date de 570 environ.27 De maniere generale, au fur et a mesure que les publicationsparaissent,il devientplus clair que des formes differentes existaient concurremment,'a des epoques plus recentes qu'on ne l'aurait cru naguere. J. J. Coulton place les cinq chapiteaux les plus anciens de l'Heraion (B, M, C, H et N) a la fin du vIIe ou au debut du vIe siecle.28II n'est pas impossible que l'on puisse encore rajeunirtel ou tel d'entre eux (cf. ci-dessus note 22). A titre indicatif, pour le chapiteau C (proche de H), le rapport du diametre inferieur 'ala longueurde l'abaque,egal a 0,49, se retrouvesur les chapiteauxde Marmariaet de Kalapodi. Le rapport de la hauteur de I'echine'acelle de l'abaque, egal a 0,60, compte parmi les plus faibles et l'apparenteaux chapiteaux corfiotesdu premier tiers du VIe siecle et 'aceux de la basilique de Paestum. Enfin, le rapport de la hauteur de l'echine 'a la longueur de l'abaque,egal a 0,103, l'apparenteaux memes chapiteauxde Corfou et a ceux de Kalapodi. Grace aux rapportschiffres reunis par J. J. Coulton, on peut operer 'apeu pres les memes rapprochementspour le chapiteau B, analogue a M. Dusseldorf 1971, p. 49-62, fig. 92-110. J. J. Coulton, ?Doric Capitals: A ProportionalAnalysis?,BSA 74, 1979, p. 81-183; les chapiteaux de l'Heraion d'Argossont inscrits au groupe 1, lui-meme integre dans l'ensemble des groupes 1-3, dont l'heterogeneitereste irreductiblea tout classement: p. 84-85, 92, 97-98, 108, fig. 10-17. 22 Par ex. Ch. K. Williams, II, ?Doric Architectureand Early Capitals in Corinth>>, AM 99, 1984 (p. 6775), p. 67-68. Schwandner,1985, p. 113; mais le groupe du Nord-Est du Peloponnesese limite pour l'instant aux chapiteaux de Tirynthe et d'Egine. Peut-etre pourrait-onlui adjoindrele chapiteau E de l'Heraion, plus fig. 4), ainsi que les nos 4 et 7 de Corinthe: Williams, op. cit., p. 72-74, fig. 1. recent (?Observations>>, 23 Corinthe: Williams (ci-dessus note 22). Egine: Schwandner, 1985, p. 29-32 et 113-117, pl. 7 et 8. Kalapodi: Felsch and Schuler, 1980, p. 73-77, fig. 51-53. Les chapiteaux B, M, C, H et N de l'Heraion (AH I, fig. 61; ?Observations?,p. 229-230, fig. 3 et 4) n'en sont pas tres diff6rents. 24E.-L. Schwandner,NFGH, p. 116-120, fig. 14-17; Schwandner, 1985, p. 115-117, fig. 73; pour une chronologiegeneralementplus basse, ibid., p. 128-129 et Wesenberg (ci-dessusnote 21), p. 49-50. 25 AH I, fig. 51; ?Observations?,fig. 3-6. 26 AH I, p. 113-115 et notes 155 et 157; cf. Wesenberg (ci-dessusnote 21), p. 52, nos1 et 2, fig. 92 et 93. 27 ?Observations>>, p. 234-235, fig. 5. 28 Coulton, Stoa, p. 28.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

101

Edifices Les etudes parues dans les quinze dernieresannees tendent aussi a abaisserla date des edifices, sans toujours en expliciter les raisons sous-jacentes.29Le portique Nord (AH I, pl. V, no II), que P. Amandrydatait du VIIe siecle, serait, selon H. Lauter, de la fin du VIIe ou du debutdu vIe siecle.30J. J. Coulton parvienta une conclusionsimilaire31en se fondant surtout sur l'appartenancepresumeedu chapiteau C a cet edifice, mais aussi sur le plan et les techniquesde construction.De meme A. Kalpaxis.32On notera cependantque les bases des colonnes interieures et certains blocs du stylobate presentent sur un cote des entailles allongees dont on a reconnu, notammenta Olympie, a Tegee et 'aKalapodi, qu'elles servaient a dresser les colonnes en bois. J. J. Coulton avait deja formule l'hypothese qu'en facade, une premiere colonnadede bois aurait ete remplaceepar des colonnes en pierre.33 Mais on peut aussi envisager que la colonnadeinterieure ait ete de bois puis de pierre, la colonnadeexterieurede bois et de pierre, sans pour autant retenirune date tres haute (cf. cidessus p. 99, a propos du temple periptere). Dans ces conditions,il ne serait plus etonnant que les chapiteaux C et D aient appartenusimultanementau portique. Date par P. Amandrydu VIIesiecle ou de la premieremoitie du VIC, le portiqueNordEst (III) n'aurait connu, selon R. Mason, qu'une seule phase de constructionau cours du ve siecle. Apres L. Tilton et P. Amandry,H. Lauter,J. J. Coultonet G. Kuhn s'accordenta en reconnaitredeux, la secondefaisant, du reste, l'objetd'evaluationsdifferentes.34G. Kuhn a rtabli a l'angle Sud-Est le chapiteau d'ante originel - que L. T. Shoe date de 57056035- et, de ce fait, restitue a l'edificeune facade ouverterythmee par des piliers ou des colonnes.36 L'edifice Nord-Ouest (VIII), mal connu, n'a jamais beaucoup retenu l'attention. P. Amandryle jugeait archaYqueet proposaitde lui attribuerle chapiteauD. S. G. Miller a admiscettedate, mais sur son plan du sanctuaire,il a fait figurerl'edificeparmiles batiments 29 Lauter, p. 175-177. S. G. Miller, <
102

MARIE-FRANQOISE BILLOT

classiques.En revanche,il apparaitsur le plan que dresseH. Lauterdu sanctuairea la fin de l'epoquearchaique.37 Apres etre descendujusqu'a la fin du ve siecle, l'edifice Ouest (hestiatorionVII) regagne maintenantla place que P. Amandrylui avait assignee dans le derniertiers du vIe.38 Enfin, H. Lauter parait supposerque l'edifice Est, hypostyle (IV) a connu un premier etat archaiquesans prostoon.Mais les vestiges ne semblentrien indiquer de tel. En definitive,le temple peripterepourraitetre date, avec la stoa Nord, dans le premier tiers du VIe siecle. Le portique Nord-Est et l'edifice Nord-Ouest dateraientd'une seconde phase de construction,aux environs du milieu du siecle, l'edifice Ouest (VII) n'ayant ete edifie qu'en troisiemelieu. I1n'est pas exclu que l'edificeNord-Ouest soit contemporainde l'edificeOuest. 1.2. Une seule terre cuite architecturale pourrait a la rigueur remonter au VIIe siecle, la piece d'angle Musee National EA 293, qui provient de l'Heraion d'Argos (Fig. 1, P1. 10:a, b

et c).39 Argile beige-verdatrea degraissant rouge, de petit calibre et nombreux. Engobe ivoire tres soigneusement poli. Lit de pose plan. Face A: Long. = 13,4; Haut. = 14,8; Haut. or. de la tuile plate = 6,4 cm. Face B: Long. = 9,5; Haut. = 13,4; Haut. or. de la tuile plate = 6,4 cm. L'arete formee par les faces A et B est partiellement epaufree. Chaque face est ornee d'une incision a/a' sensiblementparallele a l'arete, d'une incision b/b' parallele au lit de pose et tracee dans le prolongement du bord superieur de la tuile plate, enfin d'une incision c/c' qui ferme le triangle. Au-dessus du triangle, deux cercles concentriques incises ornent l'acrotereproprement dit. L'arete formee par

les faces A et B n'est pas rigoureusementrectiligne: a hauteur des incisionsb/b', elle s'incline vers l'interieur de la piece. Les incisions a et a' suivent cette nouvelle direction.Sur la face A, les incisions a et b forment un angle de 91-92?; sur la face B, les incisions a' et b' forment un angle de 900, approximativement.L'arete d'angle forme avec le bord inf& rieur de la tuile un angle de 840 40' sur la face A et de 870 sur la face B. Les bordsinferieursdes faces A et B formentun angle de 870. A la face superieurede la tuile, deux pans inclines se rejoignent selon une aree qui forme un angle de 2 a 30 avec le plan bissecteur des faces A et B. Toutes les incisions sont remplies de vernis rouges.

Si cettepiece est une tuile d'aretierplaceea l'angleinferieurdu toit, il faut admettred'une part que l'angle de I'edificene mesuraitpeut-etrepas exactement900, d'autrepart que sur la face A, I'egoutdu toit etait 1egerementinclinevers l'anglede 1'edifice(frontonsurbaisse?). Cet acrotere discret ne ressemble pas aux massifs tetragonaux restitues aux angles du toit <protocorinthien>> de Corinthe40et connus sous une forme un peu differente a 234-235. Miller (ci-dessusnote 29), p. 9, note 7, p. 10, fig. 11. Lauter, fig. 4. <
37 ?Observations?,p. 38

M. S. Goldstein, The Setting of the Ritual Meal in Greek Sanctuaries, 600-300 B.C., diss. University of California, Berkeley 1978, p. 233-245. Ch. Borker (Festbankett undgriechische Architektur [Xenia, Konstanzer

althistorischerVortrdgeund Forschungen,Heft 4], 1983, p. 16 et note 54) se prononcepour la fin du vIe siecle, apres Coulton,Stoa, p. 103-105 et 217. H. Lauter fait figurerl'hestiatoriondans le plan du sanctuairea la fin de l'archaisme(fig. 4). 39 Au lit de pose, au crayon, sur deux lignes, l'indication<<West BLDG, [-] W corner (...) South Stoa? se restitue<<West Building, Southwest (ou Northwest) Cornerof the South Stoa>>. Peut-etre s'agit-il de la fouille stratigraphiqueeffectueepres de l'angle Nord-Ouest de la Stoa Sud, au voisinagede l'edificeOuest, fouille oCu il fut possible de reconnaitreles dechetsde taille du portique: AH I, p. 98-99. 40 Robinson,TH, p. 233, fig. 9; Robinson,AM, p. 61, fig. 1.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

.177

1'11'?<

,,

,~~~~~~~~~

B

FIG.

103

/_

A

1. Heraion d'Argos.Tuile d'aretiera acrotere.Athenes, Musee National EA 293

Thermos41ou-le motif complet des antefixes decore l'un au moins des cotes de I'acrotere. EA 293 se presente plutot comme une simple tuile d'aretierdotee d'un petit ajout decoratif, mais pour le reste semblablea celle que S. G. Miller restitue pour le temple de Zeus a Nemee (ca 560),42 et dont le sanctuaire d'Artemis Limnatis a Kombothekraa peut-ere livre un autre example.43 41 Sotiriadis, col. 195, 198-199, fig. 5; AntDenk II, 1902-1908, p. 8, fig. 11. Koch, p. 72, fig. 30. GFR, p. 70-71 et 177, n? 29, fig. 144. 42 at Nemea, Hesperia 49,1980 (ci-dessusnote 8), p. 186, fig. 3. Pour la date,voir S. G. Miller, <<Excavations les 1977?,Hesperia 47,1978 (p. 58-88), p. 63; 49,1980, p. 187; 50,1981 (ci-dessusnote 8), p. 52. Si lesdelements plus recentsde la ceramiquemelee aux dechetsde taille datent de 560, la constructionne peut etre envisagee avantle secondquartdu siecle. Neanmoins, R. F. Rhodesen parle commede l'?EarlySixth-centuryBuildingat Nemea>> (ci-dessusnote 8), p. 477. Notons que EA 293 est plane a la face inferieure,alorsque la tuile de Nemee formeun angle obtus. Il est possible,mais non assure,que EA 293 provenaitd'un >. 43 U. Sinn, <>, 71), p. 71, n? 135, pl. 15:7.

104

MARIE-FRANQOISE BILLOT

Les triangles dessines sous les cercles incises suggerent une demi-antefixe: ?A 293 evoque ainsi un acrotere d'angle d'Olympie, massif tetragonal aussi, mais dont les faces exterieuresaffectentla forme d'une demi-antefixea bord superieurconcave.44A cet egard, 1A 293 apporteune certainevraisemblancea la restitutiondu toit d'Isthmia.5 Quant au decor de cercles concentriquesincises, on le retrouvesur l'acrotered'un tout petit edifice ou parapet d'autel naguere publie par G. Bakalakis46et sur un assez grand parapet d'autel d'Egine que G. Welter date du haut archaisme.47Dans la premieremoitie du VIe siecle, les marbriersfont un usage frequent et des incisions garnies de rouge et des cerclesincises pour orner les simas en marbred'Athenes48et de Delphes.49 On ne peut pas ne pas etablir un rapprochemententre cette piece et les acrotereslateraux en spirale de Prinias, Larissa de l'Hermos, Lesbos, Milet, Athenes et Selinonte.50 Mais ces pieces sont elles-memesd'ampleurtres variable.Faut-il considererque les cercles sont anterieurs aux spirales? Ou bien en sont-ils un succedane commode sur un toit de petites dimensions? Les proportionsde l'acrotereet l'usage du vernis rouge rappellent egalement tout un ensemble d'antefixesd'Athenes,Delphes, Corinthe, Nemee, Nauplie et Argos que nous allons etudier en 11.1et 11.2.3et 5. En definitive, cet acrotereatypique pourrait dater de la fin du VIIe ou de la premiere moitie du VIC siecle.

II. POUR UNE TYPOLOGIE DES ANTEFIXES DU VIIe AU VIe SIECLE Les plus anciennes antefixes d'Argos et d'Epidaure nous donneront l'occasion de revoir aux antequelques questionsde typologieet d'histoire,des premierstoits <<protocorinthiens>> fixes pentagonalesdu VIe siecle.51 F. Graber,dans Olympia II, p. 169, pl. C:14 a et b. Broneer,Isthmia I, p. 49, fig. 64. 46 G. Bakalakis,?(FwvLaKaKpCOT7jpLo 14,1955-1956, p. 15, fig. 3. a7rorT7MapcveLa>>,'EAXXqVLK4 47 G. Welter, ?(Aeginetica XI-XII?, AA (JdI 33) 1938, col. 23-26, fig. 15-17. 48 a) Athenes, Acropole,tresse au rampant,chevronsa l'egout:Porosarchitektur, p. 178-179, pl. IX:4 et 5; AA (JdI 78) 1963, col. 811, fig. 13 et W. H. Schuchhardt,?(ArchaischeBauten auf der Akropolisvon Athen>>, 14. - b) Athenes, Acropole,?(KleineRosettensima>>: Porosarchitektur,p. 178, pl. IX:2a et 2b; Schuchhardt, loc. cit., col. 802-809, fig. 1-9. - c) ?(GrossePalmetten-LotosSima>>:Porosarchitektur,p. 44-47, fig. 62-64, pl. IX: la, b et c; W. H. Schuchhardt,?(DieSima des Alten Athenatempelsder Akropolis>>, AM 60-61, 19351936, p. 1-111, pl. 1-21; AA (JdI 78) 1963, col. 816-822, fig. 18. Traditionnellementdatee de 570-560, associee a l'architectureH dont Schuchhardtvoulait abaisser la date, cette sima remonteraitau contraireau dernier quart du vIIe siecle selon I. Beyer (?(PeripterosII>>):AA (JdI 92) 1977, p. 69-70. Mais a propos de cette chronologietrop haute, voir ci-dessus note 14. 49 M.-F. Billot, ?(Notesur un sima en marbrede Delphes>>, Etudes delphiques,p. 161-177. 50 M. Y. Goldberg, Types and Distributionof Archaic Greek Acroteria,diss. Bryn Mawr College, 1977 (University Microfilms International,Ann Arbor 1986), p. 295-318. Les acroteresd'Athenes,qui prolongent la grande sima a palmetteset fleurs de lotus (ci-dessusnote 48) sont geographiquementles plus proches.Mais l'echelle et le principe de la compositionsont tout differents. 51 Dans cette optique,je ne donneraiici ni cataloguecomplet,ni descriptionsexhaustives,sauf de quelques pieces isolees. D'autres types seront simplementillustres par les planches et les figures. 44 45

105

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE 11.1. DOCUMENTS POUR UNE TYPOLOGIE

11.1.1.Argos,antefixe C. 19249 (Fig. 2, P1. 1O:eet f). Trouvee en 1955 lors de la fouille du champ Granias, dans le sondage S. 63, carreW 14: P. Courbin, BCH 80, 1956, p. 366, fig. 1:B; BCH 81, 1957, p. 665-673, fig. 1-3. Cette piece a ete trouvee sur le sol F date de la fin de I'epoquearchaique.52 Palmette superieure mediane d'une antefixe dont la partie principale a disparu. Larg. = 6,8; Haut. = 7,5; Ep. = 10 cm. Argile pulverulente

beige-rose a fin degraissantgris; pas d'engobe. Audessus de deux spirales symetriquesdivergentes,une palmette a trois feuilles. Les canaux des spirales et les feuilles sont concaves, garnis de vernis rouge. L'oeil des spirales presente une petite depression centrale bordee d'un mince bourrelet. Le decor ne paralt pas avoir ete moule, mais plutot imprime a I'aided'un poincon.

0_

cm

FIG. 2. Argos. Antefixe C. 19249

D'abord connu par les exemplaires d'Atheneset de Delphes,53puis d'Argolide54et de Corinthe,55ce type est maintenanttres bien representea Nemee par la nombreuseserie des antefixes du premiertemple de Zeus, construitvers 560.56 52Je remercieP. Courbin de ces precieux renseignements. 5 Athenes: TdA II, p. 26-27, fig. 35, 36; Vlassopoulou,nOs5, 6. Delphes: Le Roy, 1967, serie 26, p. 64-65.

Musee de Nauplie, inv. 17260, sans provenanceconnue: Hiibner, 1975, p. 119-120, fig. 1:a et b, pl. 64:4. Williams, IrT?AX?,p. 348-349, pl. 155, Corinthe FA 547, FA 404, FA 24, Nemee F. 17. Roebuck(P1. 5). 56 Une douzained'exemplairesou fragmentsdecoresen creux, trouvesen 1979: Miller, Hesperia 49, 1980 (ci-dessus note 8), p. 185 et 190, pl. 38:e et pl. 36:b a droite, pl. 39:b a gauche et pl. 40:e. Deux exemplaires trouves en 1980: idem, Hesperia 50, 1981 (ci-dessusnote 8), p. 52, pl. 15:f. Je remercievivement M. S. G. Miller de m'avoirpermis par deux fois d'examinerle materielde Nemee. La date haute proposeepar Ch. K. Williams, II dans IrT?X)?, p. 349, pour les antefixesd'Athenes,Delphes et Corinthene semblepas pouvoiretre retenue, puisque le temple de Nem&eparait devoir 'tre date dans le second quart du vIj siele, ce qui, fnalement,resout l'apparentediscordancestylistiqueet chronologiqueentre les antefixeset les tuiles faftieresa palmettes (Hesperia 50, 1981, p. 52-54, fig. 4, pl. 15:e;Stella G. Miller, Hesperia 53, 1984, p. 176, pl. 35:c: AT 244). Ces dernieres ressemblentnotammenta celles du second temple d'Apollon a Corinthe que H. S. Robinson propose de dater vers 560: Robinson, TH, p. 210, 217-218, 236, pl. 53:c; Robinson, NFGH, the Date of p. 239-240, 244 et 250-251. Mais pour une date basse, vers 540 av. J.-C. : S. S. Weinberg,<
106

MARIE-FRANQOISE BILLOT

Un moulage et un frottisde C. 19249 m'ont permis de verifierque cette partie centrale, du moins, etait identique a celles des series athenienneet delphiqueet de CorintheFA 24.57 Par ailleurs, C. 19249 et Corinthe FA 24 sont fabriqueesexactementdans la meme argile corinthienne,alors que la terre est toute differentea Athenes (rouge-brun)et a Delphes. Ce constat et la dispersion des exemplaires laissent penser qu'ils ont ete fabriques a l'aide de trois poinSonsde la meme generation,tires de la meme patrice: peu importaitque certaines antefixes fussent lateralementmoins hautes, la variation n'intervenantqu'au bas de chacune. De surcrolt,la seule piece a decoren relief, Nemee F 17, paralt avoir ete aussi estampee,58mais a l'aide d'un poinSon en creux de la meme generation que les antefixes de Delphes, Athenes, Corinthe F 24 et Argos : sensiblementde meme largeur que toutes les antefixesdu groupe (17,5 ? cm), elle est nettementplus haute, mais le motif est plus petit, et les baguetteslaterales en relief - qui correspondenta des sillons sur les antefixes decorees en creux - ne descendentpas jusqu'au bas de la piece dont la hauteur est accrue par un listel inferieur. A l'exception possible de Nemee F. 17, 'elaborationde toutes les antefixes du groupe connaltquatre etapes: 1) moulage du couvre-joint;2) adjonctionmanuelle de paquets d'argile au sommetet aux extremiteslateralesdes versants;3) impressiondu decor;4) modelage des parois de l'antefixe et finitions. Si l'on ordonnece groupe d'antefixesdu decorle plus simple au decorle plus charge,on definit trois groupes: - Athenes, Delphes, Corinthe,59Argos, Nemee F. 17: le champ de l'antefixe reste vide; Nemee premier temple de Zeus: deux arceaux symetriques, paralleles aux arceaux superieurs,garnissentle champ de l'antefixedivise en deux par une rainure mediane; Nauplie, inv. 17260: un sillon supplementaire,horizontal,decorele bas de l'antefixe.60 Les couvre-jointsne sont jamais attenants a une tuile plate, mais leur section varie, tantot triangulaire,tantot pentagonale (Fig. 3).61 Ainsi, a l'exceptiondes appendicesdecores, les antefixes du sanctuairecorinthiende Demeter et Core (CorintheFA 547 [P1.5]) et the Temple of Apollo at Corinth?,Hesperia 8,1939, p. 191-199; J. Wiseman, <<Excavations in Corinth, The Gymnasium Area, 1967-1968?, Hesperia 38, 1969 (p. 64-106), p. 94-95; cf. Williams, DrT?X)?, p. 348, note 13; AM 99, 1984 (ci-dessusnote 22), p. 69. 57 M. G. Dontas, alors Ephore des Antiquites de l'Acropole,Mme. Roebuck et M. Ch. K. Williams, II, Directeur des fouilles de Corinthe, m'ont permis de faire cette verificationen 1970 et 1971. Pour Corinthe FA 24, voir aussi CorinthIV, i, p. 50. 58 Williams, DrT?Ax?,pl. 155. 59 Trois series diffrentes a Corinthe: Williams, DrT?X)?,p. 348-349, pl. 155. Winter (ci-dessus p. 24), fig. 9:a, b. 60 Une antefixe de Nemee, AT 80, inedite, ressemblea celle de Nauplie (Winter [ci-dessusp. 24], fig. 9:d). Nemee, premiertemple de Zeus: Winter (ci-dessusp. 24), fig. 9:c. 61 Au cours du developpementqui suit, j'utiliserai trois expressions conventionnellespour designer les couvre-jointsd'apres la forme de leur section (Fig. 3). Celle du couvre-joint<<protocorinthien> est arquee, convexe a la face superieure (Fig. 3:1), concaveen sous-face. Celle du couvre-joint<> se compose de deux versants,parfois inflechis (Fig. 3:2 et 3); il est vrai que l'epaisseurde l'argile donne a la section une forme proprementhexagonale;mais je reserveraice terme a la descriptiondes antefixes dont les deux cotes inferieurs et/ou les deux cotes superieurs peuvent etre incurves ou rectilignes. Le contour exterieur de la section du couvre-joint?pentagonal>s'inscritdans un pentagone (Fig. 3:4); en sous-face,elle est concave.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

1

34

~~~~2

0

FIG.

107

10

50cm

3. Divers types de couvre-joints,sectionstransversales

celles du temple de Zeus a Nemee sont la stricteprojectiondu couvre-jointtriangulaire.Les series d'Athenes et de Delphes, Corinthe FA 404 et FA 24 (P1. 5) ont un couvre-joint sensiblementpentagonal et les ant'efixespresententa la base une decoupe triangulaire, de meme Nemee F 17. Les repartitionspar decorset par formesde couvre-jointne se recoupent donc pas et sont l'indice de fabricationssimultanees. De toute maniere, quelle que soit la forme du couvre-joint,seules les spirales laterales, les spirales et la palmette centrales en depassentla section vers le haut, tandis que les arceaux s'inscriventtoujours,pour la plus grandepartie de leur trace, dans les limites de cette section. La fabricationintensivede ce type d'antefixesne paraftpas avoir dure plus d'une generation, de 580 au plus tot a 550 environ. E. Buschorassociaitles antefixes de l'Acropoleaux tuiles d'egoutde la serie 1,62 dont la fapaden'est pas decoree. Des tuiles archaiques tout a fait semblables (par ex. C. 27701, P1. 10:d) pourvues d'un rebordlateral de section triangulaire,ont ete trouvees a Argos sur l'agora, dans la region des Thermes A et dans la fouille du champ Granias. Parfaitement planes au lit de pose, elles sont seulementmunies d'un leger ressaut a la face superieure,en tete. Bien que depourvuesd'un appendicemedian,elles ressemblentaussi a celles du premier temple d'Aphaiaa Egine.63Elles pourraientdonc dater de la premieremoitie du vIe siecle. Leur argile uniformementjaune-verdatreindique peut-etreune fabricationcorinthienne. En tout cas, C. 19249 est certainementimporteede Corinthe. 11.1.2.Epidaure, ME a7r353 et ME arr381 (Fig. 4, P1. 11:a). GFR, p. 36 et 145, n? 5 et p. 130, n? 9: deux notices egalementfautives.

Deux antefixespentagonalesidentiques,qui epousent la sectionpentagonaledu couvre-joint.Larg. or.

TdA II, p. 6, fig. 1 et 4. Vlassopoulou,n? 7. Schwandner,1985, p. 72-77, fig. 47 et 48, pl. 25 et 26. D'autres elementsd'architecturearchaiqueont ete trouvesdans la region du theatre d'Argos: le chapiteauen sofa, publie par G. Roux (L'architecturede l'Argolide aux IVe et IIIe siecles av. J.-C. [BEFAR 199], Paris 1961, p. 383-386, fig. 104) et le chapiteau dorique remployedans les Thermes A: P. Aupert, BCH 78,1974, p. 773, fig. 1. E.-L. Schwandnervient de publier un chapiteau en sofa de Tirynthe, qu'il associe au celebrechapiteaudorique archafquedont il abaisse la date au debut du vIe siecle : <>,AA (JdI 103) 1988 (p. 269-284), p. 276-283, fig. 8-12 (cf. ci-dessus p. 100 et note 24). 62 63

0 5.0

10.5 0)

V.14.2

|t

aS~~~~~~~~~~~~~

4.Eiar.AteieM FIG.~~~~~~~~~1.

81

opee

'arsM

i ?i

FIGA

1Ee

5

0

1

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE restituee = 21; Haut. or. = 9 + cm. Argile brun-rou-

geatre, pas d'engobe,vernis noir et rouge fragiles et ecailles. Au-dessus d'un triangle isocele noir, nu, limite par un petit cadre saillant, deux tiges vegetales rouges, symetriques,serrees par un anneau noir se deploient lateralement en S et s'enroulent en spirale

109

aux extremites de I'antfefixe.De part et d'autre de I'anneauet au centre des spirales, quatre yeux convexes peints en noir. Dans F'coinson des spirales, une feuille noire. Au centre, une palmette a coeur triangulaire noir et a trois feuilles, une rouge entre deux noires.

11.1.3. Antefixes argiennes hexagonales, decorees d'une fleur de lotus renversee, de tiges vege"taleset d'une palmette (Fig. 5, P1. 11 :b-d).

20--~~~~~~r.

20

10

0

FIG. 5. Argos, agora. Antefixe 76/1835.1 et tuile d'egout

81/7011.1. Axonometriede la bordured'un toit

E. D. Van Buren en avait signale un certainnombre,mais avec des confusionssur leurs provenanceset sur leur repartition entre le Musee National d'Athenes, le musee de Corinthe et celui d'Argos: GFR, p. 131-132, nos 14-19, fig. 6 et 10. Cf. Apollon Pythe'en, p. 19, fig. 14; Htibner, 1975, pl. 64:5 et 6 et pl. 68:6. Argos et l'Heraion d'Argos en ont livre dix series, moulees en relief plus ou moins saillant, ou estampillees. Le couvre-jointest toujourspentagonal. L'antefixe le depasse en hauteur de toute sa partie superieure, bords incurves compris. Les cotes inferieurs, convexes, chevauchaientles rebordslateraux des tuiles d'egout.

110

MARIE-FRANQOISE BILLOT

Le decor ne varie pas: une fleur de lotus renversee,reduite a deux petales largement surgissent deux tiges vegetales ecartes et retenus par un anneau plus ou moins etroit d'oCu symetriques. L'ensemble est surmonte par une palmette a petit coeur et cinq feuilles qui lateralesde celle-ci restentvides. occupe l'angle superieurde l'antefixe. Les <
W. Vollgraff, BCH 31, 1907 (p. 139-184), p. 155-156, fig. 4. Apollon Pytheen, p. 18-19, fig. 15. 66 Sotiriadis,col. 197-198, fig. 5 et pl. 10:4 et 5. 67 Le Roy, 1967, series 33 et 34, p. 87-88, pl. 30 et 123. 68 A. Vlachopoulou-Oikonomou,HyqLuO'VfE KaL KOpv4aLESfKEpaLouL /uE bLaKoo,uL)-7- a7ro mOvHfEtpo, Jannina 1986, p. 30-37, fig. 6. Corinthe,fragmentinedit FA 555. Isthmia, fragmentinedit IT 20. 69 Tous les types sont representesdans ATK, pl. 53, 56 et 57. La plus ancienne serie paralt d'importation. 70 Voir surtout E. Walter-Karydi,Samos,VI, i, SamischeGefassedes 6. Jahrhundertsv. Chr., Bonn 1973. 71 Le Roy, 1967, p. 33-34, nOs1-10. 72Le Roy, 1967, n? 10, p. 33, pl. 5et 118. 73 Huibner,1975, p. 118-119, fig. 1, pl. 64:2 et 3, Beil. 8:2. Ni la figure 1, ni I'aquarellene donnent une reproductionsatisfaisante. 65

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

111

A2, le relief est seulementun peu plus plat qu'en Al, et peint en rouge sur fond reservealors qu'il est reservesur fond clair en A2. Les series Al et A2 se repartissententre l'edifice e de l'Aspis, le sanctuaired'Apollon Pytheen, l'agora et l'Heraion d'Argos, la serie D entre l'agora et l'Heraion. Jusqu'a present, la serie C n'est attesteequ'au sanctuaired'ApollonPytheen, les series B, E, F, G, H, I a l'Heraion. Mais plusieurs fragmentsde provenanceinconnue, appartenantaux series B, H et I peuventvenir soit de l'Heraion, soit d'Argosmeme. De ces remarques,on peut tirer quatre conclusions: la permanencedu type paralt definirun style local. La repartitiondes memes series entre plusieurs sites indique une remarquableactivite edilitairevers le milieu du VIe siecle. Argos y prendsans doute une large part: le conflitavec Sicyoneest resorbe,et malgre ses difficulteslatentesavec Sparte,Argos cherchea s'affirmer par le developpementde sa propre ville et par une participationactive a l'organisationdes Jeux Nemeens. Elle contribue,peut-etre plus que les autres cites d'Argolide,au developpement de l'Heraion. Huit series, A2, B, D-I sont attestees a l'Heraion. Un meme batiment pouvait en accueillir plusieurs, des l'instant que les differencesde facture n'etaient pas trop accusees. Certaines series ont pu servir a des refections,telle G qui semble imiter E et F. Mais il est remarquableque cet ensemble, fabrique a partir de 560 environ, confirmela chronologie basse que j'ai retenue pour la plupart des edificesarchaiquesdu sanctuaire,ceux du momns qui nous sont conserves. Cette production s'insere directement dans la typologie des antefixes des VIIC et VIC sie-

cles, et jette meme un eclairageretrospectifsur son histoire. 11.2. TYPOLOGIE

ET HISTOIRE DES ANTEFIXES AUX VIIC ET VIC SIECLES

Cette typologie se fonde sur les procedesde fabricationet les recherchesd'assemblages qui permettentde passer de l'extremitepure et simple d'un couvre-jointordinaire a l'ement a la fois fonctionnel et decoratif qu'est l'antefixe. Quelques reperes chronologiques semblentacquis, encorequ'ils ne soient pas absolumentindiscutables.Pour autant, les combinaisonsde procedes,ou de resultatsobtenus,ne dessinentpas une evolutionlineaire ni une sequencechronologiquedefinitive. 11.2.1 Du couvre-joint <<protocorinthien>> al'antefixe hexagonale a) Corinthe,premiertemple d'Apollon,ca 680 av. J._C.74 M. C. Roebuck, <<Excavations at Corinth, 1954?, Hesperia 24, 1955 (p. 147-157), p. 156; Robinson, TH, p. 231; Robinson,NFGH, p. 248, fig. 8; Robinson,AM, p. 59-60, fig. 2, pl. 14:4 et 15:1.

Sur 10 a 15 cm vers le bord du toit, le couvre-jointprend une forme triangulaire. En sous-face, son adherence contre le rebord lateral longitudinal des tuiles plates n'est pas parfaite, mais acceptable.Ce rebordexiste sur toute la longueur des tuiles d'egout,jusqu'a 74 De la bibliographierelative au temple qui a precedecelui du VIesiele, je ne donne ici que les titres et les passages qui decriventles tuiles d'egout. Pour la date, voir Robinson, AM, p. 57 et note 5. Je lui suis tres reconnaissanted'avoirbien voulu me presenterlui-meme les tuiles de Corinthe en 1981.

MARIE-FRANQOISE BILLOT

112

leur face anterieure,non seulementdans la realite sur le cote libre de chaque tuile a couvrejoint,75mais virtuellementsous le couvre-jointattenant, puisqu'il reapparalten facade de l'edifice,la face anterieurede la tuile plate debordantde 8 mm environsur celle du couvrejoint. Nous verronsque ce dispositif sera decisif pour la typologie des antefixesdu VIC siecle (ci-apres 11.2.6). Vue de l'exterieure,la forme triangulairedu couvre-jointne paralt dictee que par un souci purementdecoratifd'harmoniegenerale76:les triangles se superposentaux triangles, ou plus exactement s'emboltent.L'extremite anterieure du couvre-jointd'egout fait donc l'objet d'un traitement specifique qui conjugue necessite fonctionnelleet recherche ornementale: sur la facade de l'edifice,l'antefixeest nee.77Celles de Corinthesont hexagonales a cotes rectilignes. b) Isthmia, premiertemple de Poseidon,premieremoitie du VIIe siecle av. J.-C. Broneer, Isthmia I, p. 49, fig. 60, 61 et 64, et catalogue p. 51-52; idem, <,NFGH (p. 39-62), p. 43, fig. 3.

La presence d'un element decoratiftriangulaire - dit parfois et abusivement?fausse antefixe>- au milieu des tuiles d'egoutest le seul argumentque l'on avancepour affirmer que le temple d'Isthmiaest posterieura celui de Corinthe.0. Broneer,qui estimait d'abord que les deux toits avaient ete fabriquesdans le meme atelier corinthien,a peu pres a Meme epoque,78vint a penser que celui de Corintheetait anterieura celui d'Isthmiad'a peine une decennie.79H. S. Robinson a insiste sur l'anterioritedu temple de Corinthe qu'il evalue a un quart de siecle, ce qui revient a dater celui d'Isthmia vers le milieu du vIIe siecle.80 Recemment, Ch. K. Williams s'est aussi prononce en faveur de cet ordre chronologique, mais avec un ecart reduit: le temple de Corinthen'est que legerementplus ancien, et le toit d'Isthmia,qui copie celui de Corinthe,doit etre date de la meme generation.8' Le processusprincipalde la formationde l'antefixeest le meme qu'a Corinthe.Mais en le couvre-jointprend une forme triangulaire,les deux versants outre, sur les 15 a 20 cm oCu sont legerementincurves: les cotes superieurset inferieursde l'antefixesont respectivement concaveset convexeset l'adherencedu couvre-jointau rebordlateral libre de la tuile voisine est parfaite. La fabricationdu moule n'est pas plus complexe,mais demandeun peu plus de tile>> 7?5<<Pan/cover ou <>, pan-covertile>>. les Les faces superieuresdes versantssont tantot reellementplanes, tantot tres legerementconcaves,d'oCu informationsapparementcontradictoiresdonneespar M. C. Roebuck,loc. cit. (ci-dessusp. 111) et par Robinson, TH, p. 231. En fait, cette faible concavite occasionnelleest due au demoulage et a la cuisson; elle est negligeable,de sorte qu'elle n'apparaltpas ni dans le texte, ni sur les photos et les dessinspublies dans Robinson, AM. 77 De toutes les formules conclusivesde H. S. Robinson a ce sujet, la plus juste, a mon sens, se trouve certainementdans NFGH, p. 248:
79 80

Idem,NFGH, p. 43. Robinson,TH, p. 216, 227 note 1, 231-234; Robinson,NFGH, p. 247; Robinson,AM, p. 57. p. 346-347.

81 Williams, ITnAX,

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

113

soin.82En termes de recherchedecorativecomme du point de vue technique, la solution d'Isthmia,oCules courbesdu couvre-jointet de l'antefixe reproduisentcelles des tuiles d'egout, est a la fois plus efficaceet plus spontanee, plus naturelle que la stylisation geometrique des couvre-jointsde Corinthe dont l'adherenceest de surcroftmoins bonne. A cet egard, le toit d'Isthmiapeut paraitreplus ancien. L'anterioritedes tuiles plates de Corinthe se trouveraitainsi contreditepar celle des couvre-jointsd'Isthmia! En fait, commele montre la formule intermediairede l'antefixe delphique A.176 (ci-apres), la sequence chronologique des deux toits paralt devoir etre demontreepar d'autresdonneesarcheologiquesque les toits eux-memes. En tout cas, les deux variantesextremes d'antefixeshexagonales- a cotes principaux rectilignes ou courbes - sont sensiblementcontemporaines.En theorie, elles peuvent des lors connaitredes evolutions separees. Du moins les premierstoits connus - dont 0. Wikander dit a juste titre qu'ils ne sont tres probablementpas des prototypes83- offrent-ils, pour le developpementdes antefixes, deux tetes de serie etroitementapparentees. c) Delphes, antefixe A. 176 (<>) Le Roy, 1967, p. 28, pl. 5.

Partie anterieure (Long. = 14 cm) d'un couvre-jointprotocorinthiencomme l'indique sa face superieure, convexe vers l'arriere. Vers l'avant, la face superieure des versants est recreusee, donnant a l'antefixe des cotes superieurs concaves, comme a Isthmia. La face inferieuredes versantsest plane, donnanta l'antefixedes cotes inferieursrectilignes,comme a Corinthe. Il s'agit donc d'une formule intermediaire,oCules premiersprocedesdecoratifs coexistent avec la plus grande simplicite technique, sans-preciserun ordre chronologique. A.176 pourrait etre associee aux series 1 a 4 de tuiles protocorinthiennes(Le Roy, 1967, p. 21-28). d) Olympie F. Graber, dans Olympia II, p. 169, pl. LXXXXVII (XCVII):7 a, b et c.84

Le couvre-jointprotocorinthienest independant- sans tuile plate attenante- mais similaire a ceux de Corinthe et d'Isthmia. Pour la formationde l'antefixe, la comparaison s'etablit directementavec Isthmia: vers l'avant, les versants du couvre-jointse creusent. Mais le sommet de l'antefixe depasse legerementl'arete du couvre-joint.L'antefixe n'est 82 Voir l'article fondamentald'E. Gebhard et W. Rostoker,<>, JFA 8, 1981, p. 211-227. 83 0. Wikander, 1988, p. 204-205. Considerantla complexite des toits protocorinthiens,il conclut "It is difficultto accept them as the first step of a development,even though no preliminarystages are known. Yet, the possibility should not be ruled out completely:the historyof technologypresentsquite a numberof inventions developing not from the simple towards the complicated,as might be suspected, but the other way around." 84 En 1981, A. Mallwitz avait bien voulu me montrercette piece et celles qui doiventlui etre associees: deux couvre-jointsd'aretieregalementsans tuile plate attenante,et une tuile plate d'aretier.Un ensemblesimilaire, de dimensionsun peu plus fortes, comportedeux couvre-jointsd'aretier,plusieurs couvre-jointscourantdont un reproduitdans Olympia II, pl. XCVII:8 a, b et c, et deux couvre-jointsd'egout dont l'antefixe depasse la section, donc comparablea celle qui nous sert ici d'exemple. [Voir aussi J. Heiden, ci-dessus p. 41-46.]

MARIE-FRANQOISE BILLOT

114

donc plus seulement engendreepar le couvre-joint,elle n'en est plus la seule section anterieure. Meme si techniquementla piece est moulee d'un seul tenant, l'antefixecommencea se presentercommeun element autonome,ajoute. 11.2.2.Du couvre-jointprotocorinthienau couvre-jointtriangulaire Les modificationsopereesen facadede l'edificea l'extremiteinferieuredes couvre-joints d'egout semblentavoir tres vite affectetoute la longueur de ces pieces : leur fabrications'en trouvaitsimplifiee. Les premierscouvre-jointstriangulairesqui nous soient connus se trouvent en Grece occidentale,ce qui n'exclut evidemmentpas qu'il aient ete utilises tres tot ailleurs, notammenta Corinthe. a) Thermos: couvre-jointsfermes par des antefixesa tetes dedaliquesde deux types Sotiriadis,col. 192, fig. 4 au milieu, pl. 10:1;AntDenk II, pl. 53 A:2. Koch, p. 55-56. H. Koch, <>, fig. 124 et 125. Rhomaios, p. 103-104, fig. 68. Wallenstein, II A 7 et 8, p. 29-32 et 100-101 (avec bibliographie): la datationhabituelle, en 630/620 av. J.-C., y est confirmee.Mertens-Horn, 1978, p. 44-46, 52-53 et 55-56, fig. 19, 20.

A l'arrierede ces deux series d'antefixes,les couvre-jointstriangulairespresententdes versantsconcaves. b) Corfou, Mon Repos: couvre-jointsa tete feminine dedalique G. Dontas, AEXT 18, 1963, B' (1965), p. 167-168, pl. 198:a, /, 8, E et 207:3 et b; AEXT 19, 1964, B' 3 (1967), p. 320, pl. 357:y et 358:a; AEAT 23, 1968, B' (1969), p. 306-307, pl. 246:y, b. Wallenstein, III A 5, p. 38-39 et 107 (610-600 av. J.-C.). Dontas, NFGH, p. 125-126, fig. 5, 6. Mertens-Horn, 1978, p. 56-59 passim, fig. 23 (ca 600 ou debut du v1esiecle).

Le couvre-jointtriangulairepresenteegalementdes versantsconcaves.85 c) Calydon, couvre-jointsa tete feminine dedalique Dyggve, <<BuntesDach>>,p. 143-144, fig. 156, pl. XVI:M et N. Wallenstein, III A 4, p. 39-40 et 107 (bibliographie) : ca 610 av. J.-C. Dontas, NFGH, p. 127-128. Mertens-Horn, 1978, p. 54, fig. 21 (610600 av. J.-C.).

Les versants des couvre-jointss'inflechissentvers la base, tout commeceux des couvrejoints courantsde ce toit, qui sont, par chance, conserves: Dyggve, p. 143-144, fig. 157:0, pl. XVI:0 et P et pl. XVII. Dach> d) Calydon,<
Meme remarque. Dach>> e) Corfou,<
Peut-etre est-ce l'exemple qui symbolise le mieux le passage du couvre-jointprotocorinthien au couvre-jointtriangulaire. Couvre-jointet tuile plate sont attenants. Leurs angles diagonalementopposes sont abattusde manierea permettrele meme assemblageque sur les toits protocorinthiensde Corinthe, d'Isthmia, de Perachoraet de Delphes. La face superieuredes versantsest concave,leur face inferieureplane (fig. 106, 107). 85La tentativede M. Mertens-Horn d'attribuerles antefixesde Corfouet de Thermos a des ateliersargiens n'est pas tout a fait convaincante,du seul point de vue de la plastique et de l'esthetique. Au surplus, il ne semble pas que l'on ait jamais trouve d'antefixesa visages feminins en Argolide,pas plus qu'a Corinthe.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

115

Les antefixes decorees de figures feminines portant le polos et flanquees tantot de sphinx, tantot de pantheres(fig. 108-112) sont dateesvers 580 av. J.-C.: Wallenstein, IV A 33 et 34, p. 52-53 et 123; Mertens-Horn, 1978, p. 59, fig. 27. Il me paraft indispensable de souligner ici la persistance d'une formule ancienne de couverture,dont les transformationsmorphologiquesserontdecisivespour 1'avenementdes antefixespentagonales. Le <de Corfou etablit aussi la jonction chronologiqueavec la toiture du premiertemple d'Aphaia a Egine, dont tous les couvre-joints,independantsdes tuiles plates, sont triangulaires(ci-apres II.2.3).86 Avant de quitter la Grece de l'Ouest, notons que toutes ces antefixes developpentleur hauteur vers le bas du couvre-jointpuisqu'elles vont jusqu'a passer devant les tuiles d'egout.87A l'exception des antefixes du <<BuntesDach? de Calydon, oCule polos des figures femininesdepasseun peu l'arete du couvre-joint,toutes inscriventleur decorplastique dans un cadrepentagonal dont la hauteur s'etablit a 15-17 cm pour une largeur de 20-22 cm Dach? de Corfou),ou bien atteintet depasseles Dach? de Calydon,<
Le sommetde l'antefixe n'est pas plus haut que l'arete superieuredu couvre-joint.Les versantssont recreusesvers l'antefixe,lui donnantainsi des cotes superieursconcaves.Mais leur concaviteest surtout accusee par deux fortes cornes laterales rapportees.Ailleurs, et notammenten bas, l'antefixecoinciderigoureusementavecla sectiondu couvre-jointtriangulaire. De memeque Delphes A. 176 etait au memeniveautechniqueet esthetiqueque les antefixes d'Isthmia et de Corinthe, Nemee AT 103 se trouve au meme niveau que l'antefixe Olympia II, pl. XCVII:7 (comptenon tenu, evidemment,de la forme du couvre-joint):sur l'antefixed'Olympie,c'estla pointe medianequi se developpe;sur Nemee AT 103, se sont les corneslaterales. [C. Pfaff (ci-apresp. 149-156) presenteici meme deux series de l'Heraion d'Argos.Du point de vue de l'evolutiondes antefixes, la premiereserie, peinte en rouge, pr&cedeNemee AT 103; la secondeserie est au meme niveau techniqueque AT 103.] Rappelons l'existencede couvre-jointstriangulairesa Olympie: OlympiaII, pl. XCVII:11 a et b et 16 a, b et c. Mais ils ne sont, pour l'instant, pas dates. 87 Schema d'assemblagedans Dyggve, p. 229, fig. 228. 86

116

MARIE-FRANQOISE BILLOT

b) Kombothekra,Temple d'ArtemisLimnatis, premieremoitie du VIe siele Sinn (ci-dessusnote 43), n? 133, p. 71, pl. 15:5 et 16:1 et 2.

La face superieuredes versantsdu couvre-jointn'est pas recreusee,leur face inferieure est a peine inflechie. La pointe medianeet les corneslateralessont, commestoujours,modelees dans des masses d'argile fichees sur le couvre-joint. [Presenteespar G. Hubner lors de ce meme congres(ci-apresp. 167-174), les antefixes de Kalapodi sont construitessuivant les memes principes.] c) Egine, premiertemple d'Aphaia,ca 575 av. J.-C. Schwandner,1985, p. 76-77, fig. 46-48, pl. 26 et 27:1.

Les versantsdes couvre-jointscourantssont plans ou a peine concaves;ceux des couvrejoints d'egout sont legerement recreuses mais surtout releves vers l'avant, du sorte que la pointe sommitale et les deux cornes laterales de l'antefixe connaissentun importantdeveloppement;mais les cotes superieursconcavesde l'antefixe restent en grande partie engendres par la formedu couvre-joint.Par ailleurs, la surfaceinferieuredes versants,et par suite les cotes inferieurs de l'antefixe s'inflechissentde maniere a epouser le rebord lateral des tuiles, commetoujours. d) Halieis et Mases88 Les couvre-jointset les antefixes trouves sur ces deux sites89sont, dans leur principe, similaires a ceux d'Egine, comme l'avait deja note E.-L. Schwandner.90Leur surface est seulement un peu plus haute, annoncant les proportions des antefixes d'Argos, et leurs pointes, commetoujoursmodeleesdans des masses d'argilerapportee,sont plus accusees. e) Premiertemple de Zeus a Nemee, ca 560 av. J.-C. Cf. ci-dessus11.1.1.S. G. Miller,Hesperia49, 1980(ci-dessusnote8), pl. 38:e,39:bet 40:e;Hesperia50, 1981(ci-dessusnote8), pl. 15:f.[Seenoteat footnote8 regardingdate-Editor.]

Trois petites masses d'argile sont manuellementfichees au sommet et aux deux extremites inferieuresdu couvre-jointtriangulaire;le decory est ensuite imprime en creux. Deja notee par E.-L. Schwandner,91leur similitude structurelleavec les antefixes d'Egine, Mases et Halieis est frappante. Nous avions signale que, sur toutes les antefixes du groupe <>, seules les spirales laterales et centrales ainsi que la palmette mediane depassaient le couvre-joint.En d'autres termes, les 88Je remercieW. W. Rudolph qui a bien voulu me permettred'etudierce materielaux Musees de Nauplie et de Porto Cheli en 1978 et 1981, et N. Cooperavec quij'ai pu debattrede ces questionsde typologieen 1981. [Voir N. Cooper, ci-dessus p. 65-93.] 89J. Dengate, "The Archaic Doric Temple at Mases," Abstractsof Papers 76th General Meeting of the ArchaeologicalInstitute of America, 1974, Section II A, p. 22; cf. AJA 79, 1975, p. 149. N. K. Cooper, "Two Roof Tile Systems from Halieis" Abstractsof Papers 79th GeneralMeeting of the ArchaeologicalInstitute of America, 1977, Section III B, p. 22; cf. AJA 82, 1978, p. 250. 90 Schwandner,1985, p. 127 et notes 236 et 239. 91 Schwandner, 1985, p. 127-128. E.-L. Schwandner,AA 1988 (ci-dessus note 63), p. 280, note 10. [Les deux series d'antefixesde Kalapodi presenteespar G. Huibnerlors de ce colloque (voir ci-apres p. 167-174) relevent exactement du meme type: les antefixes hexagonales coYncidentavec la section du couvre-jointa l'exception de l'appendice superieur et des <
TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

117

antefixesdu temple de Zeus a Nemee donnentla versiondecor&ede celles d'Egine, Mases et Halieis. Dans l'histoiredes terres cuites architecturales,trop souventlivree a de subjectives appreciations stylistiques, pareille convergence,doublee d'une remarquable cofncidence chronologique(ca 575 a Egine, ca 560 a Nemee) est trop rare pour ne pas valoir d'etre soulignee. Elle est d'autant plus frappante que le decor affecte justement les parties par lesquelles les antefixesque nous avons enumereesse sont progressivementdeveloppees,cornes lateraleset pointe mediane,et qu'il s'y cantonne. f) Athenes, Acropole,serie II. Ca 560 av. J.-C.92 TdA II, p. 6-7, fig. 6 et p. 27-29, fig. 37, 38.

On notera les deux corneslaterales. La palmette se developpehors du cadrehexagonal sous l'influence des premieres antefixes et acroteresa palmette decoupee a jour (ci-apres p. 117-118). Commenceea Isthmia, la lente evolutiondes antefixeshexagonalesa cotes inferieurset superieursconcavess'achevepratiquementa Nemee. Meme si ce type d'antefixecontinuea etre prise, comme le prouvent,entre autres, nos antefixes d'Argos (ci-apres, II.2.5 et 6), il n'est plus susceptible d'un developpementautonome l'interet architecturalde la forme reste inconteste,mais les nouvelles series, celles d'Argosnotamment,empruntentun decor d'abord experimente sur les antefixes pentagonales. De meme, la serie II de l'Acropole releve d'autresinfluences. De Corintheet Isthmia a Nemee, le processusse serait deroule en un siecle. C'est bien long pour bien peu. N'est-ce pas trop long? Du moins est-on fonde a se demandersi la date des premierstemples de Corinthe et d'Isthmia,de Perachoraet de Delphes, ne devraitpas etre abaissee. Surgitaussi la questionde l'originedu motifdecoratif.Les arceauxs'adaptenttropbien a la formepreacquisedes antefixespour qu'il faille chercherautrechoseque l'inspirationnaturelle de l'artisanqui a creela premiereserie du groupe<>. les antefixeshexagonaleset pentagonalesbassesa un seul registrede tiges vegetales(ci-apres II.2.5 et 6) qui, nous le verrons,ne peuventetre anterieuresa 580 av. J.-C.93 Enfin, quel role a pu jouer ce groupe importantdans l'eventuel<<passage>> des antefixes hexagonaleset pentagonales,peu ou prou limitees a la sectiondu couvre-joint,aux antefixes a palmette libre? Les premiers exemples de ces compositionsa palmette decoupee a jour apparaissentdans la decennie 580-570: acroteredu premier temple d'Aphaia a Egine et fausse antefixe du toit IV-V de l'Acropole,94antefixeset palmettesfaitieresdu <> de Calydon,95puis une serie eginete naguere attribueeau premiertemple d'Aphaia 92 Les gros arceaux bombes, ourles d'un lisere, evoquent pour le traitementdu relief, les series Al et A2 d'Argos(ci-dessusp. 110-111). L'anneau large se retrouvesur de nombreusesantefixesde cette epoque. Vlassopoulou, no 8. 93 Liste dans Le Roy, 1967, p. 33-34, nOs1-7 et 9. Elle doit etre completee par la liste des antefixes a S couches: ci-apres, p. 122-123. 94 Schwandner,1985, p. 84-85 et 127-128, fig. 53, 54 et pl. 30. TdA I, p. 4 et 12, fig.1 et 11. Sur le groupement des simas IV et V de l'Acropole,ci-apres p. 133. 9 Dyggve, p. 164-169 et 230-236, pl. XXI, fig. 167-169.

118

MARIE-FRANQOISE BILLOT

mais toujoursdatee du deuxieme quart du VIC siecle,96les antefixesdes toits 14 et 16 et de la serie 21 de Delphes, a placer entre 560 et 550.97 L'ecart est trop grand de nos antefixes hexagonales, chichementdecoreesde maigres spirales et de petites feuilles imprimees en creux, a ces massives et riches compositionsdu second quart du VIe siecle. Celles-ci, bien que reproduisantdes motifs decoratifsdes longtemps connus par la peinture sur vases, sont certainementdes creations originales plutot que le resultat d'une <<evolution>> dont les etapes, desormais,s'evanouissent.Et puis, l'ecart chronologiqueest pratiquementnul. II.2.4. Naissance du couvre-joint pentagonal

Le couvre-jointtriangulaireparaitd'unetelle commodite,et si simple a fabriquer,qu'on peut s'etonnerde sa disparitiona partirdu secondquart du vIe siecle, et de la concurrencedu couvre-jointpentagonal.A dire vrai, tout le poids du couvre-jointtriangulairerepose sur les ar'tes inferieures,et il risque toujoursde se casserle long de l'aretesommitale.98 Par ailleurs, s'il a pour origine l'amenagementde la partie anterieuredes couvre-joints protocorinthiens(ci-dessus II.2.1), et si les couvre-jointsprotocorinthiensont ete fabriques assez longtempsdans le courantdu VIIe siecle et peut-etre memejusqu'au debut du vIe, il y avait place pour une forme hybride. C'est a nouveau en Grece occidentale que nous en trouvonsles premiersexemples connus. a) Corfou, premiertoit de Mon Repos, dernierquart du VIIe siecle d'apresles antefixes a gorgoneionet a tetes feminines99 Un couvre-jointcourant,pentagonal,a grand rayon de courbureinterieure,repose sur les aretes inf6rieures.100 A l'egout, le couvre-jointsembleprendreune formetriangulaire.101 Dach>>. Ca 580-570 av. J.-C. b) Calydon,<
Meme commentaire. II est inutile de multiplierles exemples.Jusqu'a l'aube de l'epoqueclassique,les couvrejoints <> sont ainsi conpus.Le rayonde courbureest toujourstres grand.Par suite, l'epaisseurlaterale des couvre-jointsdiminue au point de ne leur assurer que rarementune veritableassise plane. Et si tel pouvait etre le cas, on cherchea l'eviteren taillant les parois interieuresen biseau de manierea ce que le couvre-jointne reposeque sur des aretes. L'inventionde cette formehybridemultiplie les difficultesd'ajustement: au bas du toit, les rebordslateraux des tuiles plates opposent a la courbe interieure du couvre-jointdeux courbesinverses. Pour occulterce vide, il n'y a que deux solutions: 96 E.-L. Schwandner, <>,AA (JdI 86) 1971, p. 534, fig. 2; idem, NFGH, p. 112; Schwandner,1985, p. 128, note 244. 97 Le Roy, 1967, toit 14, p. 50, pl. 10; toit 16, p. 53, pl. 16 et 119; pour la date, p. 57-62; serie 21, p. 56 et 62, pl.21. 98 Cf. Dyggve, fig. 157; Schwandner,1985, fig. 46 et 49, pl. 27:2, 5 et 6. 99 Wallenstein, III A 2-5, p. 37-38 et 106-107 (avec bibliographie),pl. 6:1 et 2. Dontas, NFGH, p. 123124, fig. 2-4. Mertens-Horn, 1978, p. 36-40. 100 KorkyraI, p. 149, fig. 125. 101Ibid., p. 149-152, fig. 130.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

119

utiliser des antefixeshexagonalesa cotes inferieursconvexesqui epouserontle profil des tuiles plates; fermerl'espacepar une antefixe pentagonale. II.2.5. Couvre-jointpentagonal et antefixe hexagonale a) Olympie Olympia II, pl. LXXXXVII (XCVII):10 a, b et c.

Cette piece associe la plus grandeefficacitea la plus grandesobriete.Vers le bas, l'antefixe obtureparfaitementl'espacelibre. Pour ce qui est de l'esthetique,on note la persistance d'une forme hexagonale, reguliere et equilibree, mais un peu plus haute qu'aux origines. L'antefixe est desormaisplaquee contrele couvre-jointet ses c6tes superieursse situent un peu plus haut que les versants.102 b) Argos Ensembledes 10 series d'antefixeshexagonalesa cotes inferieurset superieursincurves et a decorde palmette sur fleur de lotus (ci-dessusII.1.3). Elles ne different pas fondamentalementde la precedente.Mais le developpementde leur partie superieure tres au-dessus du couvre-jointetait indispensable: il offre l'espace necessaireau motif decoratifque l'on avait choisi d'emprunter.103 c) Dans le groupe <> (ci-dessus 11.1.1), la serie I de l'Acropole,les exemplairesde Delphes (serie 26), certainsde Corinthe (FA 404 et FA 24) et Nauplie 17260 presententun couvre-jointpentagonal. Du point de vue typologique, ce sont les pieces les plus evoluees du groupe mais, a cette epoque, ce ne sont pas necessairementles plus anciennes. d) Certaines antefixes a palmette libre utilisent les qualites fonctionnellesde l'antefixe hexagonale a cotes inferieursconvexesjusqu'a preserveret mettre en valeur son image: Delphes, toit 12, antefixe A.23, ca 570 av. J.-C. Le Roy, 1967, p. 46, pl. 7.

Les cotes inferieurs convexes adherent parfaitementaux rebords lateraux de la tuile plate (R.3, Le Roy, 1967, p. 47, pl. 7). Mais de surcroit, de petites nervures saillantes completent, en relief, sur le fond d'A.23, le dessin de l'antefixe hexagonale, comme s'il importait qu'une nouvelle compositiondecorativen'occultepas la partie proprementfonctionnelle de la piece. troisiemequart du VIe siecle Halae, antefixes du toit <<megarien>>, H. Goldman,<, Hesperia 9, 1940 (p. 381-514), p. 440, no 4, fig. 102.

Seuls subsistentles cotes inferieursconvexes,toujourspour l'efficaciteet l'etancheitedu systeme. Mais dans le meme temps, des antefixes identiques, mais attenantes aux tuiles plates, et donc a base rectiligne,sont employeessur un autre cote du temple ou sur un edifice rigoureusementcontemporain: ibid., p. 440, no 3, fig. 103. [Voir F. Dakoronia (ci-apres p. 175-180) sur un portique de Gardinitsa-Oponteprotegepar une toiture identique.] 102

Cf. aussi deux antefixes inedites de Nemee, AT 37 et AT 78. De meme Nauplie 17264: Hubner, 1975, p. 120-121, fig. 3:a, pl. 64:6; Nemee AT 65: Hesperia 48, 1979, p. 87 et 89, pl. 30:c;et, dans une moindremesure, Acropoleserie IV: TdA II, p. 32-35, fig. 43; Vlassopoulou, no 25. 103

120

MARIE-FRANQOISE BILLOT

11.2.6. Couvre-joint pentagonal et antefixe pentagonale

Il ne faut pas ici se meprendre. En theorie, il s'agit bien de resoudre un probleme nouveau: occulterun volume pentagonalpour des raisons d'etancheiteautant que d'esthetique. Mais en pratique,la solutionest dejaamorceepar les pieces qui combinenttuile plate et couvre-joint,depuis les toits protocorinthiensde Corinthe, d'Isthmia et de Corfou (cidessus 11.2.1 et II.2.2.e): le problemey est, pour ainsi dire, a moitie resolu du cote oCule couvre-jointest attenanta la tuile plate. Reste a le resoudrede l'autre cote: ou bien le rebordlongitudinal libre reste intactjusqu'a la face anterieurede la tuile, et donc visible sous l'antefixe: mais il faut que celle-ci soit hexagonale. Autrement dit, on retrouvela situation decrite ci-dessus en 11.2.5 et c'est la formule que retient en definitive l'artisan des series III d'antfefixeset de tuiles de l'Acropole(ci-apres).104 Ou bien la partie anterieuredu rebordlongitudinallibre est abattue et l'antefixe pentagonale passe par devant,prenant appui sur la face superieure,desormaisplane, de la tuile. II n'est donc pas etonnant que plusieurs series d'antefixes pentagonales, dont celles d'Epidaure (ci-dessus 1.1.2, Fig. 4 et P1. 11:a), toutes limitees a la section du couvre-joint, presententun triangle inferieur plein, mais prive de tout decor: il evoque les deux rebords lateraux des tuiles platesjointives,rebordsqui, dans la realite, ne sont qu'exceptionellement decores.105Nous ne citerons ici que les exemplaires suffisammentconservespour donner une forme complete: a) Eleusis Koch, p. 79, fig. 36; Le Roy, 1967, p. 34, n? 4.

b) Corinthe FA 101 et FA 101 a (P1. 5) CorinthIV, i, p. 11 et 57-58; Le Roy, 1967, p. 34, n? 5; Williams, lrT7A7), p. 347-348, note 13, pl. 154.

Piece semblablea la precedenteet, pour le decor,a la serie III de l'Acropole. c) Tirynthe Hubner, 1975, p. 118-119, pl. 64:2 et 3, Beil. 8:2; Le Roy, 1967, p. 34, no 5.

Ce n'estjustementpas un hasardsi ces trois antefixessont attenantesa une tuile decoree d'une tresse dont, a l'autre extremite,le rebordlateral libre a ete necessairementabattu. Au-dessus du triangle nu, la surface hexagonaledecoree,soigneusementdelimitee par un cadre en relief, representeen fait la veritable antefixe. Objectera-t-onque les contours inferieursde ces images d'antefixessont rectilignes,alors qu'ils sontnormalementconvexes? II s'agit, commesur les antefixesdu toit 12 de Delphes, d'une stylisationqui, ici, privilegiele parallelisme des cotes superieurs et inferieurs. Mais d'emblee, les antefixes pentagonales apparaissent comme un succedane, un substitut des antefixes hexagonales.

tout a fait La meilleure preuveen est donneepar la serie III de l'Acropoled'Athenes106: 104

TdA II, p. 6-7, fig. 6 et p. 29-32, fig. 39-42; Vlassopoulou,nos 1-3. Seul exemple connu, la serie II de l'Acropole:une demi-palmette,en principe,orne la face anterieuredes rebordslateraux. En fait, le travail est tres neglige: TdA II, fig. 6 et 38, pl. I; Vlassopoulou,no 8. Ces rebords lateraux sont le plus souventpeints en noir ou en rouge au-dessus de la tresse. 106 TdA II, p. 29-32, fig. 39-42. Un exemplaire beaucoup mieux conserve que K230 et K231 justifie la restitution proposee fig. 40. L'hexagone decore est bien percu comme la partie principale et significative, puisqu'il marque la limite des ajustementsoperes apres cuissonsur K230; Vlassopoulou,n?s1-3. Par ailleurs, 105

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

121

similaire a celles d'Eleusis et de Corinthe pour ce qui est du decor, elle conservela forme hexagonaledans sa purete fonctionnelle,lui associel'imaged'une antefixehexagonaledecoree, et presente au-dessous de cette image decoree,une zone angulaire nue qui rappelle le triangle nu des antefixes de Corinthe et d'Eleusis. Comme si l'artisan avait hesite entre la mode, nouvelle mais nostalgique, des antefixes de Corinthe et d'Eleusis qui inscriventun hexagone dans un pentagone,et la vieille forme hexagonale si pratique qu'il l'a finalement retenue,et d'autantplus volontiersque le couvre-jointn'etait pas attenanta la tuile plate.107 Tout le probleme est de savoir quand on a commencea fabriquer des antefixes pentagonales, et donc, conjointement,a supprimer la partie anterieure du rebord lateral des tuiles plates. Les premierestuiles a tresse de l'Acropoled'Athenesconserventce rebord;108 E. Buschorles associe aux simas en cavet a feuilles doriques.Sur les tuiles a tresse plus recentes, la partie anterieuredu rebordest abattue.109A Delphes, le rebordest conservesur toute la longueur des tuiles du toit 12110- celui dont les antefixes a palmette libre conserventle contourd'une antefixe pentagonale(ci-dessusp. 119) - mais il est abattu sur les tuiles du toit 14111et du toit 16.112Le toit 12 de Delphes est sensiblementcontemporaindu de Calydon oiules antefixes similaires sont au contraireattenantesa la Dach>> <
122

MARIE-FRANQOISE BILLOT

d'Isthmia et de Corinthe, dont Delphes A.176 et les plus anciens exemplaires d'Olympie (ci-dessus II.2.1, c et d) peuvent etrejustement contemporains.Ainsi, les deux systemesde toiture pouvaient coexister a meme epoque. De toute maniere, la fabricationde couvrejoints independantsest infinimentplus simple, et apparentecelle des toits protocorinthiensa celle des toits laconiens.Rappelonsici qu'en amont, la sectiondes tuiles d'egoutde Corinthe et d'Isthmiaprend une forme incurvee.[Voir maintenantaussi, a ce sujet, E.-L. Schwandner, ci-apres p. 291-300.] 3) Couvre-jointpentagonal attenant, donc rebord longitudinal oppose abattu en fapade: toits 14 et 16 de Delphes; <> de Calydon;antefixespentagonalesa triangle nu d'Eleusis, Corintheet Tirynthe (ci-dessusp. 120). A moins que ces dernieresn'anticipentde beaucoupsur les toits 14 et 16 de Delphes ce qui obligerait a etirer demesurementla chronologiedes simas a feuilles doriques d'Athenes et de Delphes -, il y a de fortes presomptionspour que la troisieme solution technique n'intervienneguere avant 580 av. J.-C. On observepar ailleurs qu'aucunedes authentiquesantefixeshexagonalesn'est decoree avant celles du premiertemple de Zeus a Nemee (ca 560), celles du sanctuairede Demeter et Core (Corinthe FA 547 [P1. 5]: decor exclusivementpeint) et la serie III de l'Acropole, contemporainedes antefixes pentagonales similaires d'Eleusis et de Corinthe (FA 101 et FA 101 a [P1.5]). Tout invite donc a croire que les antefixes pentagonales qui integrent l'image d'une antefixe hexagonale n'ont pas ete fabriquees avant 580. On a cree, specifiquementpour elles, le motif des deux tiges vegetales symetriquesqui prennentappui sur les rampantsdu triangle nu et se deploient dans les deux moities de l'hexagone, laissant place a une petite palmette sommitalea un coeur et trois feuilles. Ce scheman'est absolumentpas conu pour garnir une surfacepentagonale. II.2.7. De'cor des antefixes pentagonales

En revanche,des lors que l'antefixe pentagonaleest desencombreede l'image-souvenir d'une antefixe hexagonale,elle est susceptibled'accueillirun decormieux adaptea la forme et a l'etenduede sa surface.Ce peut etre une creationqui lui est precisementdestinee,ou un emprunt. a) Creation specifiqueet parfaitementadequateque la compositionde deux tiges en S horizontaux, a spirales centralesdivergentesserreespar un large anneau;une petite feuille dirigee vers le bas garnit les ekoinSonsdes spirales exterieures, une palmette a coeur et trois feuilles ceux des spirales intferieures;une troisieme palmette similaire couronnele tout cependantqu'une petite feuille renverseese glisse encoresous l'anneau,dansl'axe de l'antefixe. De cette composition,on connalt au moins cinq versions dont les trois premieres,contemporaines,garnissentdes antefixesjustement attenantesa des tuiles a tresse: -

Trezene

Ph. E. Legrand, <, BCH 29, 1905 (p. 269-315), p. 273, fig. 2; Le Roy, 1967, p. 35, n? 16.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

123

Elle est associeea la sima trouveeau meme endroitet generalementdatee aux environs de550av.J.-C.114 -

Corinthe, FA 543

Robinson,TH, p. 236 et 237, pl. 53:a;elle est attribueeau secondtemple archaiqued'Apollon;mais voir Williams, 2rnAiW, p. 348, note 13. Si l'attributionse confirme,l'antefixe devraitetre datee soit vers 560 selon la date proposeepar H. S. Robinson (TH, p. 215, 216, 217, note 36 et 218), soit vers 540 si l'on retient la date basse proposeepar S. S. Weinberg, Hesperia 8, 1939 (ci-dessusnote 56), p. 191-199.

[Une autre antefixe de format beaucoup plus petit, et au decor plus leger, a ete presenteepar Mrs. Roebucklors de ce colloque.] -

Polychronon

D. Feissel et M. Seve, <,BCH 103, 1979 (p. 229-320), p. 231, fig. 1 etp. 261. Egine GFR, p. 4 et 130-131, n? 11, fig. 5 en bas a droite. -

Eleusis

Antfefixeinedite, presqu'identiquea celle d'Egine.15 Spirales tres enroulees, petites palmettes a trois feuilles a peine plus souples: il est difficiled'admettreun large ecart chronologiqueentre ce groupe et les autres antefixespentagonales, notammentla serie III de l'Acropole,116Corinthe FA 101 et 101 a'17et l'exemplaire de Tirynthe.118 b) Emprunt,mais adaptationoriginale,que la ?miseen page>> du motif de palmettesur fleur de lotus dans un cadrepentagonal,a Delphes19 puis a Thermos,120mais aussi a Ambracie, Corintheet Isthmia (ci-dessusp. 110 et note 68). Ch. Le Roy a date les series delphiquesde 570-560 av. J.-C. et signale combienl'interpretationdu motif etait prochede celle que nous offrent les vases corinthiens.Nous en avons tire plus haut les consequencespour la date des antfefixesargiennes (ci-dessus p. 110 et 119). Mais les antefixes de Delphes et de Thermos ont aussi un lien avec l'antefixepentagonalede Tirynthe: si l'on compareles proportionset la facture des palmettes, on est conduit a dater cette piece de 570-560, comme du reste les antefixes du <> de Calydon, celles du toit 12 et de la serie 21 de Delphes, 114

Le Roy, 1967, p. 62. Lowenkopf-Wasserspeier, p. 36: vers 540 av. J.-C. Travlost et a Mme Th. Karaghiorgad'avoirpu etudier le materiel d'Eleusis. 116 Ci-dessus p. 120-121 et note 106. 117 Ci-dessus p. 120 (II.2.6:b). 118 Ci-dessus p. 120 (II.2.6:c). 119 Le Roy, 1967, series 33 et 34, p. 87-89, pl. 30. 120 Sotiriadis,col. 195, pl. 10:4. Par le relief ourle d'un lisere, et par les deux pointes saillantes'ala naissance du lotus, la facture des antefixes de Thermos ressemblea celle des series Al et A2 d'Argos. Cette premiere serie de Thermos est certainementposterieure'acelles de Delphes. Une secondeserie reproduiteibid., pl. 10:5 pourraitpasser pour une fabricationlocale maladroite,ou pour une copie tardive. Mais ces lotus aux formes tres pointues et tres maigres se retrouventdans le decorimprimeau poinconsur le bordd'un bassin decouvert a Halae; de petites protomes lui sont aussi appliquees, dont le caracterearchaique ne fait pas de doute H. Goldman,Hesperia 9, 1940, p. 408, n? 11, fig. 39. Pour d'autresantefixes a palmette sur lotus renversea Ambracie,Corintheet Isthmia, voir ci-dessus note 68. 115 Je dois a J.

124

MARIE-FRANQOISE BILLOT

des antefixesd'Egine,121 les antefixes FA 16 (P1. 6) et FA 20 autrefoisattribueesau temple d'Apollonde Corintheetant certainementplus recentes.122 c) Empruntdu meme ordre,et adaptationegalementoriginaleque la <<mise en page>>, dans le cadrepentagonal,du motif de palmettesur deux registresde longues tiges serreespar un anneau, motif dejaconnu,justement,par ces antefixesa palmettelibre de Calydon,du toit 14 et de la serie 21 de Delphes. De tous les representantsde la version pentagonale,123celui qui provientdu sanctuairedu heros Ptoios a Kastraki(Le Roy, 1967, p. 35, no 15, pl. 118) serait le plus ancien;or il ne peut etre anterieura 570 av. J.-C.124:une fois de plus, nous avons a faire a des creationssensiblementcontemporainesqui ne peuventrien nous apprendrede certain sur une eventuelle<evolution>> des antefixesdans la premieremoitie du vie siecle. Inscrit dans un espace restreint, le motif subit de serieuses modificationsd'equilibre. Repoussee vers le haut par les deux registres de tiges vegetales disposes au-dessus et audessous de l'anneau, la palmette se reduit a trois (Ptoion) ou cinq petites feuilles (Corinthe, Perachora,serie 24 de Delphes); les spirales gagnent les angles inferieursde l'antefixe. Avec ses tiges larges, sa compositionserree, ses palmettesa trois feuilles dans les ecoincons, l'antefixe du Ptoion est traitee dans le meme esprit que les antefixes pentagonalesa S horizontauxdu groupe<> et dateraitdoncde 550 environ. Les exemplaires de Corinthe et de Perachora,plus legers, plus aeres, sont-ils plus recents?Le sommetde l'antefixede Perachoradepassea peine celui du couvre-joint,125 celui des antefixes de Corinthe le depasse beaucoup: il y a-t-il vraiment, entre toutes ces pieces, de grands ecarts chronologiques?A partir du milieu du vie siecle, toutes les interferencessont possibles. d) C'est ainsi que le decor des antfefixesd'Epidaure et de Tirynthe ne represente pas un motif en cours de developpement,mais un motif transfere et mutile par la necessite de garder nu le triangle inferieur de l'antefixe. Dans le cas precis des antefixes d'Epidaure,il est clair que le transfert s'est fait par l'intermediairedu groupe <>: les yeux disposesde part et d'autrede l'anneau et au centredes spirales, les feuilles dans les ecoin~onsle prouvent. On note aussi que la base des triangles est plus courte que celle des antefixes, qui, par ailleurs, ne sontjustement pas fixees a une tuile plate : l'image originelle de l'antefixe hexagonale s'effaceen meme temps que son role architecturals'evanouit. Il paraftpossiblede dater ME ar 353 (P1. 11:a)et 381 vers 540 av. J.-C. On retrouve le meme relief saillant et sec d'une part sur les antefixes d'Egine et d'Eleusis a S horizontaux (ci-dessus 11.2.7),d'autrepart sur les series argiennesles plus -recentes(B, H, I). 121

Schwandner,AA 1971 (ci-dessusnote 96), p. 534, fig. 22. Sur l'ensemblede la question, Williams, 1r7AiM,p. 348, note 13 et pl. 156. 123 Liste des antefixespentagonalesa palmettesur longues tiges vegetalesserreespar un anneau et disposees en deux registres,trouveesau Ptoion, a Corinthe et a Perachora: Le Roy, 1967, p. 35, nos 11-15; cf. p. 37 et pl. 118. Ajoutermaintenanta cette liste une autre antefixe de Corinthe: Wiseman (ci-dessusnote 56), p. 99, pl. 31:e. 124 Cf. maintenantJ. Ducat, Les kouroidu Ptoion (BEFAR 219), Paris 1971, p. 442: <<(....) l'occupationdu site de Kastrakine semble pas remonterau-dela de 580 au plus tot>>. L'antefixe Delphes A.3 (toit 9: Le Roy, 1967, pl. 6) doit tres probablementetre completeesur le meme modele. 125 H. Payne, Perachora,I, The Sanctuariesof Hera Akraiaand Limenia, Oxford 1940, pl. B:2, p. 113-115, fig. 18 et pl. 127:E. 122

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

125

Longtemps consideree comme la plus ancienne des antiefixespentagonales, FA 204 (P1. 5) trouvee au Quartier des Potiers de Corinthe fait plutot figure de creation marginale.126Quant a la piece delphique A.9,127elle est probablementla plus tardive de toutes, posterieurea nos antefixes d'Epidaure. Au chapitredes interferencesentre differentstypes, on notera enfin que certainesantefixes pentagonalesfinissent par avoir des cotes superieursconcaves,ainsi Corinthe FA 518 (ci-dessusnote 123), et l'une des deux series de Thermos.128 CONCLUSIONS

1) Les couvre-jointstriangulaireset pentagonaux apparaissentpeu de temps apres les couvre-jointsprotocorinthiens: les premiersexemplairesconnus remontentau derniertiers du vlle siecle. 2) Longtemps ignorees, les antefixes hexagonalesa cotes inferieurs et superieurs courbes, non decorees,jouent un role important: durantpresqueun siecle, a partir du temple d'Isthmia, ce sont pratiquementles seules antefixesqui soient utilisees. Leur histoire s'acheverait dans le second quart du VIe siecle avec le groupe <> si certaines antefixes pentagonalesn'en reproduisaientl'image et si les ateliers argiens ne leur redonnaientun regain de succesen leur adaptantun nouveaudecor. 3) Bien que du point de vue techniqueleurs originesremontentaux toits protocorinthiensde Corinthe et d'Isthmia, les antefixespentagonalesn'apparaissentpas, dans le Nord-Est du Peloponnese, avant 580 au plus t6t.129Elles sont alors les succedanesdes antefixes hexagonales dont elles conserventl'image, inscritedans leur cadre,jusque vers 540. Elles comptent parmi les premieresantefixes a decorvegetal. On ne sera pas etonne de constaterqu'il faut aussi abaisserla chronologiedes simas a feuilles doriquesauxquelles elles sont souvent associees (ci-apres IV. 1 et note 161). 4) Les antefixes pentagonales accueillent plusieurs compositionsdecorativesdifferentes, l'une creee pour cette forme, les autres emprunteeset adaptees. Ces compositionsexercent les unes sur les autres des influences qui permettentde preciserla chronologierelative des differentsgroupes. Mais ces effets reciproquesn'interviennentpas sur l'apparitiondes antefixes a palmette decoupeea jour, qui releventde creationsparalleles a celles des antefixes hexagonaleset pentagonales. 5) De manieregenerale,les dates auxquelles apparaissentles differentstypes d'antefixesdu vIe siecle sont, le plus souvent, plus recentes qu'on ne l'a cru jusqu'ici. Entre 580 et 550 surgissentpresque tous les schemas de composition.Certains durerontjusqu'aux ve et Ive siecles. A partir de 560-550, l'interet commencea se deplacervers le decordes simas. 6) Principauxgroupes, re'capitulation,chronologie a) Antefixes hexagonalesnon decorees: du VII siecle a 570 environ. Le Roy, 1967, p. 33-34, no 1 et p. 36-37. Cf. Williams, -7 A7j, p. 346-347, pl. 155. Roy, 1967, p. 34, no 10, serie 7, pl. 5:1 (attention,sur la planche 5, les legendes ont intervertiA.9 et A.5)etpl. 118. 128 Sotiriadis,pl. 10:5. 129 Contre la date haute admise depuis H. Payne, Necrocorinthia,Oxford 1931, p. 252: debut du vlI siecle. Le Roy, 1967, p. 36-37: vers 620. De meme Williams, -7 Ai, p. 347-348 et Heiden, 1987, p. 37-38. 126

127Le

MARIE-FRANQOISE BILLOT

126

b) Antefixes hexagonales a arceaux, spirales et palmette (groupe <>) c) Antefixes hexagonalesa palmette sur lotus renverse,argiennes: de 560-550 au dernier tiers du VIe siecle.

d) Antefixes pentagonalesa image d'antefixehexagonale (Eleusis-Corinthe-TiryntheEpidaure) : de 580 au plus tot a 540 environ. d') Antefixes hexagonales de la serie III de l'Acropole,oCus'inscrit l'image ?traditionnelle? de l'antefixehexagonale: contemporainesdes premiersexemplairesdu groupe d. e) Antefixes pentagonales a S horizontaux, spirales centrales divergentessurmontees d'une palmette,130feuilles et palmettesdans les ekoinvons(groupe<>) f) Antefixes pentagonalesa palmette sur lotus renverse(Delphes, Thermos, Ambracie, Corinthe, Isthmia) : de 570 a 550-540 environ. g) Antefixes pentagonalesa palmette sur longues tiges vegetales serreespar un anneau : de 560 a 540et disposees sur deux registres (groupe <>) 530 environ. h) Antefixes a palmette libre sur longues tiges vegetales serrees par un anneau et disposees sur deux registres(meme motif que precedemment,mais avec la palmettedecou'ajour): de 570 environau troisiemequart du VIC siecle. i) Antefixes a palmettelibre sur lotus renverse: a partir de 570 au plus tot. 7) Quelques correspondances stylistiques

Premieresantefixes des groupesb et d: palmettesa trois feuilles. Antefixe de Tirynthe (groupe d) et premieres antfefixesdes groupes h et i: larges palmettesa cinq ou sept feuilles courteset raides souventordonneessur un arceau. Premieresantefixes du groupe e (Trezene, Corinthe, Polychronon)et premiereantefixe du groupe g (Ptoion) : palmettes-etfeuilles dans les ecoin~ons. Premieresantefixes des groupes i et f: motif de palmette sur lotus renverse. Dernieres antfefixesdu groupe d (Epidaure), du groupe e (Egine, Eleusis) et du groupe c (argiennes G, H et I) : tiges vegetales etroites, compositionplus lache, relief saillant et sec. -

8) Quelques influences possibles

Des premieresantfefixesdu groupef sur les premieresantefixesdu groupec: le motif de palmette sur lotus renverses'inscritd'aborddans un cadrepentagonal,puis dans un cadrehexagonal. Du groupe h sur le groupe g : le motif de palmette sur deux registresde tiges vegetales serreespar un anneau, d'abordutilise pour des antefixesa palmettelibre, s'inscrit ensuite dans un cadrepentagonal.

-

130 La compositionequivalente,mais a palmettelibre, est connue par une antefixe de marbrede I'Acropole: Porosarchitektur,p. 48, n? 4, fig. 69. Un fragmentd'antefixed'Argosde la fin du Vle siecle ou du debut du ve siecle interpretecette compositionen relief. Par la suite, les S seront disposesen oblique, par exemple sur les antefixes du toit 50 de Delphes: Le Roy, 1967, p. 122-123, pl. 44.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

127

Du groupe h sur l'antefixe de Tirynthe (groupe e): le decor du groupe h s'inscrit dans un cadrehexagonal qui le mutile. - Du groupe g sur l'antefixed'Epidaure(groupee) : meme remarque.

-

III. RELIEFS DE BRONZE ET DE TERRE CUITE Entre la peinture des vases et le decordes antefixes et des simas, la communautedes repertoires n'est plus a demontrer.Mais il faut souventconstaterd'importantsdecallagesd'interpretation: repertoirecommunne signifie pas rendus ou effets similaires. On sait aussi que les compositionsdecorativesles plus souventreproduitessur les antefixes decorentaussi, a toute petite echelle, la ceramiquecorinthiennea reliefs estampilles: vasques, bassins, perirrhanteria,pithoi en offrent de nombreuxexemples de part et d'autre des anses, sur les 1evreset les rebords,et en zones sur les panses. Ce sont alors des emprunts directsaux terres cuites architecturales,et le rendu des motifs est identique, preuve que les memes ateliers assuraientune productiontres diversifieeet faisaient appel, pour la fabrication des moules et des poinvons,a une meme corporation.131 Mais faute de documentsbien publies, on ne s'est pas encore aper~ud'une etroite parente entre certaines antefixes et les ornementsterminaux des <<Schildbander>>. Deux antefixes argiennes,differentes,en portenttemoignage. -

Antefixe C. 11482 (P1. 11:e)

BCH 92, 1968, p. 1020, fig. 31. Trouvee par J.-P. Sodini sur le radier d'un egout de l'agora dans un contextedes ve et v1esiecles ap. J.-C. Piece cassee et epaufree de tous cotes. Larg. = 19; Long. = 10,5; Haut. = 15,5 cm. Argile beige-grisatre, roseeen profondeur,a nombreuxdegraissantde petit calibre,gris ou rougefonce.Engobesur la face anterieure.

Au-dessus de deux spirales divergentes,un arceau supporteles neuf feuilles d'une palmette. Le relief du canal des spirales et de l'arceau est particulierementbombe, celui des feuilles est plus plat. Spirales, arceau et feuilles sont ourles d'un lisere arrondi. Sous l'arceau, le coeur de la palmette est creux, pyramidant.Le canal des spirales et l'arceau sont reserves,leur lisere rouge. Les feuilles sont rouges, leur lisere reserve. -

Antefixe 77/507.1 (Fig. 6)

Trouvee a l'agorapar P. Marchetti, lors d'un nettoyagede surfacepres du Monument Rond. Piece cassee a l'arriere, et mutilee par un gros eclat a la face anterieure. Larg. = 18,5; Long. = 8,5; Haut. = 18 cm. Argile rose, engobe beige-rosesur toutes les faces.

Decor similaire mais plus elance. Le coeur de la palmetteest plan. Cette piece pourrait etre une copie de la precedente. Du point de vue stylistique au moins, les antefixes d'Argos, notamment C. 11482, sont directementcomparablesa une piece de Lousoi.132Du point de vue technique, elles 131 Sur ce sujet, deux articles fondamentaux: S. S. Weinberg, ?Corinthian Relief Ware, Pre-Hellenistic Period?,Hesperia 23, 1954, p. 109-137, pl. 25-33; M. lozzo, ?CorinthianBasins on High Stands?,Hesperia 56, 1987, p. 355-418, pl. 63-82, oCul'on trouvera, au fil des notices, d'abondantescomparaisons,ainsi que toute la bibliographieou presque. 132 GFR, p. 47 et 145, no 3, fig. 17 'agauche. Cette piece decoupeea jour ne parailtpas non plus presenterde couvre-jointnormal.

128

MARIE-FRAN?OISE BILLOT

0

19.0

8.5 coupe

FIG.

a

6. Argos, agora. Antefixe 77/507.1

appartiennenta des toitures identiques a celle du temple d'AthenaSoteirad'Asea en Arcadie.133Elles se retrecissenten effet vers l'arriere,en forme de pavilion. Le couvre-jointn'est pas conserve, mais il ne fait aucun doute que ces antefixes prenaient place sur une sima d'egout, exactementcomme a Asea. La toiture de ce temple archaique- dont il n'est pratiquementpas conserved'autresvestiges- est generalementdateeentre 570 et 540. Ses antefixes, deposees au musee de Tegee, presentent une palmette a sept feuilles assez raides et tassees. Leur relief y est aussi cerned'un lisere. Un large anneau serreles tiges vegetales;sur les antefixesd'Argos,il n'est plus conserve;il n'existaitpeut-etrememepas sur 77/507. 1. Ces quatre series d'antefixessimilaires,associeesau mqrnedispositifarchitectural,proviennent certainementdu meme centre de production,sinon du meme atelier. Il travaillait au contactetroit des fabricantsde brassardsde boucliers,dont, au premierregard,nos antefixes rappellent l'ornementterminal. On pense notammentaux ornements,aux palmettes que E. Kunze range sous le type A, qui couvrela premieremoitie du VIe siecle.134D'apres sa chronologie, les antefixes d'Asea dateraient effectivement de 550 environ.135C. 11482 et l'antefixe de Lousoi, un peu plus anciennes, seraient a placer dans le second quart du 133K. Rhomaios, <<'Iepov'AO-vasa comEpas KaL HoTEa8Lwvos Kara ri-v 'ApKabKrjv 'Aoe'av>>, 'ApX'E4 1957 (p. 114-163), p. 119-124, fig. 8-14. 134 E. Kunze, OlForsch,II, ArchaischeSchildbdnder,Berlin 1950, p. 204-206, Beil. 10, pl. 1-3, 52, 61 et 74. 135 Ibid., n? 17, inv. B 1975, p. 15, 204 et 243, pl. 31, 32 et 74.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

129

siecle.36 Bien que differente des antefixes d'Asea, 77/507.1 en serait contemporaine,ou meme un plus plus recente.137 La tradition ecrite attribue la fabricationdes boucliers a Argos.138Malheureusement, les fouilles n'y ont mis au jour ni fonderie, ni meme vestiges assures d'une activite metallurgique, de sorte que Cl. Rolley garde la prudencede situer les manufacturesde boucliers et d'armesen general - dans le Nord-Est du Peloponnese.139 Toutefois, et bien que les relationsentre Argos et Lousoi soient moins directesqu'entreArgos et Tegee, la decouverte de deux antefixes du meme type sur l'agora d'Argos autorise a considererl'ensemble, au moins provisoirement,comme l'oeuvrede tuiliers argiens. De surcroit,les antefixes de Tegee surmontaientune sima a profil rectiligne,tres frequenten Argolide. IV. SIMAS A PROFIL DROIT Sept simas a profil droit sont actuellementconnuesen Argolide.Certainesd'entreelles sont couronneesd'un demi-rond, mais cette moulure, implantee en retrait du profil principal, n'en modifiepas l'allure generale.Certainessimas sont orneesde feuilles doriques,d'autres d'une frise de palmettes,ou de palmetteset de fleurs de lotus. IV. 1. SIMAS A PROFIL DROIT COURONNE D UN DEMI-ROND ET DECOREES DE FEUILLES DORIQUES

Deux simas presque semblablesproviennentl'une de l'edificee construitsur le flanc Est de 1'Aspis,I'autredu sanctuaired'Apollon Pytheen. Malgre de menues differencesd'execution, elles appartiennenta deux toits contemporains140 qui se composentcommesuit: a) Aspis Trois fragmentsde sima dont le mieux conserveest C. 9890 (Fig. 7). W. Vollgraff (ci-dessusnote 64), p. 155-156, fig. 4. GFR, p. 10 et 78, n? 22. Apollon Pythe'en,p. 24. Argile beige, rosee au centre,a degraissantrouge. Engobede la meme argile. Haut. or. = 18,4-17,9 cm. Sur fond entierementrouge, dont une zone inferieurehaute de 3 cm, le contouret la nervuremedianedes feuilles sont reserves.Sur le demi-rond,des zones d'egale longueur, alternativementrouges et reservees,surmontentles feuilles. Travail soigne;l'arrondides feuilles est parfois prepareau compas.

Une tuile d'egout C. 26728 a face anterieurenon decoree(Fig. 8). Antefixes de la serie A2 (ci-dessusp. 110-1 11). b) Sanctuaired'ApollonPytheen -Sima C. 26701 (P1. 11:f) Meme argile, meme decor.Haut. or. = 18,5 cm. Haut. or. de la zone inferieurerouge = 3,5+ cm. Sur le demi-rond,les zones rouges et reserveescorrespondentrespectivementaux intersticeset a la partie mediane des feuilles. Vernis rouge beaucoupplus dilue. 136

Ibid., nOs8 et 9, p. 10-1 et 242, pl. 12,17 et 74. Voir aussi les nOs5,14 et 55, de type B, pl. 75, p. 207-208. Ibid., n? 43, inv. B 1650, p. 29, 205 et 243, pl. 52 et 74. Voir aussi le n? 40, de type B, p. 26-27 et 243, pl. 49 et 75. 138 Ibid., p. 215-230. 139 Cl. Rolley, Les bronzesgrecs, Fribourg 1983, p. 143. Depuis, on a toutefois decouvertun brassardde boucliersigne de l'Argien Aristodamos: GettyMusJ13, 1985, p. 166-167, n? 12; SEG 35, 1985, 266 bis. 140 La fouille de l'agora a livre une tuile d'egout et des antefixes d'un troisiemetoit identique. 137

0

C4~~~~~~~~~~~~~~~~~~

48.0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

FIG. 7. Argos, Aspis, 'Edificee. Sima C. 9890

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

e

8.8

131

-34.0

<

FIG.8. Argos, Aspis, Edifice e. Tuile d'egoutC. 26728

des series Al et A2 (ci-dessus p. 110-111) : il y avait donc alternancede -Antefixes motifs clairs sur fond rouge et rouges sur fond clair. Le rapprochements'etablitimmediatement14' avec a) la sima de Mycenes a profil similaire Hubner, 1975, p. 121-122, fig. 3:b, pl. 65:1,2, Beil. 9:1. Plus petite (Haut. or. = 11,5 cm), elle est decoreeau bas d'un motif de damier a deux

rangs de carreaux,suivi d'un filet reserve. b) Trois simas a simple profil droit -

Delphes, S. 3 + S. 118 (Toit 6)

Le Roy, 1967, p. 31-32, pl. 5, 98 et 118. 141 Une sima inedite a profil droit surplombe d'un demi-rond est conserveeau Musee de Corinthe. Une tresse decore la partie inferieure,tandis qu'une bande rouge sur fond reservela parcourta mi-hauteur [voir Roebuck,ci-dessus p. 62].

132

MARIE-FRANQOISE BILLOT

Elle aussi plus petite que celle d'Argos(Haut. or. = 10,8 cm), elle porte des feuilles sans nervuremediane au-dessus d'une zone non decore. -

Kalapodi, sima Z. 10

R. Felsch et H. Kienast, <<EinHeiligtum in Phokis>,AAA 8, 1975 (p. 1-24), p. 21, fig. 26 et 27. [Voir Hubner, ci-apres p. 173, fig. 2.]

Au-dessus de la zone inferieure non decoree, serie de feuilles a nervure mediane; un triangle reservea cotes concavess'inscritentre les arrondisde deux feuilles successives.142 Kalapodi,sima de la fapadeoccidentaledu temple archaiqueNord R. Felsch et G. Huibnerdans Felsch and Schuler, 1980, p. 78 et 113, fig. 101. [Voir Hubner, ci-apres p. 169, fig. 1.]

Les feuilles, a double contourcoloreet reserve,ne presententpas de nervure,et descendentjusqu'au bas de la sima. La piece Z. 11, plus haute, a profil concave,et plus richement decoree,aurait trouve place sur la fapadeorientale.143 Quelle que soit l'exacte repartitiondes simas de Kalapodi,toutes datent de l'epoque ou les deux temples de l'archafsmetardif ont ete construits,plus precisement,pour le temple Sud, des annees 570-560.144Les antefixes qui leur sont associeessont de formehexagonale, a cornes,celles du temple Nord etant surmonteesd'une palmette en relief.145 Bien qu'une certaine permanencedes types ne soit pas a ecarter, il paraft souhaitable d'abaisserla date de la sima de Delphes a la premieremoitie, peut-etre meme au deuxieme quart du VIe siecle. Et bien que les simas de Kalapodisoient de hauteurs variees et presentent un soffite decore,le groupe Delphes-Kalapodi a l'unite et les caracteresd'une production regionale. Les simas de Mycenes et d'Argos paraissent plus proches de la production corinthienne. et sur le cote Le damierde la sima de Mycenes se retrouvesur la sima I de l'Acropole146 de la sima angulaire du toit 13 de Delphes,147toutes deux de proportionssi prochesqu'elles peuvent etre attribuees au meme atelier.148Ch. Le Roy date le toit 13 de Delphes vers 560 av. J.-C. 142 Dans le rapportpublie en 1980, la sima Z. 10 ne semblait attribuablea aucun edifice connu: Felsch and Schuler, 1980, p. 78; G. Hubner dans Felsch and Schuler, 1980, p. 113. Selon le rapportpublie en 1987, elle appartiendraitau temple archaiqueNord: Felsch, 1987, p. 21 et note 43, ce qui paralt contredirel'attribution de deux autres simas, dont Z. 11, a ce meme edifice. D'ocula necessite de se reporterici meme a Huibner,ciapres p. 168. 143 Felsch et Kienast, AAA 8, 1975 (ci-dessus), p. 21, fig. 28 et 29. Felsch and Schuler, 1980, p. 78. 144 Felsch, 1987, p. 19-24. 145 Felsch et Hubner dans Felsch and Schuler, 1980, respectivementp. 78 et 113, fig. 102. Felsch, 1987, p. 21 et 24. Winter, ci-dessus p. 25, fig. 10. 146 TdA I, p. 6-8, fig. 1 et 6-8. Vlassopoulou,n? 14. 147 Le Roy, 1967, p. 48-49, pl. 8 et 9 et p. 57-62. 148 Le rapportIII = e/d utilise par L. T. Shoe (GM, p. 130-13 1, pl. LXII) se revele tres aleatoiredans son principeet d'applicationdifficile.En outre, il ne se combinejamais avec aucun autre criterede classement.En revanche, les rapports I = b/a, II = c/a, et le rapport c/b utilise par Le Roy (1967, p. 57-59) confirment d'autres rapprochements.Si l'on accepte la restitution de Buschor, on obtient pour la serie I de l'Acropole: I = 1/4,02; II = 1/4,83; c/b = 1/1,18; pour le toit 13 de Delphes, I = 1/4,7+; II = 1/5+; c/b = 1/1,05.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

133

La monochromieen rouge caracterisela sima du premiertemple d'Aphaiaa Egine, vers 575 av. J.-C.,149les simas I a V de l'Acropolequi s'ordonnentdans le quart de siecle suivant,150celle du Quartier des Potiers de Corinthe datee de 560-550151 et la serie 19 de Delphes situee vers 540.152 De ces exemples, on retiendraqu'a deux reprises, sur des pieces d'angle, des cavets a peine marques coexistent avec des cavets tres profonds;ainsi, sur la sima V de l'Acropole, que Buschor hesitait a associer a la sima IV153:pourtant, munie qu'elle est d'une fausse antefixe en guise d'acrotered'angle, elle peut tres bien avoirtrouveplace a l'egout d'un toit, sur le retourd'un angle inferieurgauche, a l'instardu no244/247 d'Egine.154Sur ces cavets peu marques,la partie superieuredes feuilles doriquesn'etait pas masqueepar le surplomb du bandeausuperieur. Le demi-rondpeut etre emprunteaux premieressimas de type megarien.155Quelques simas a feuilles doriqueset a profil en cavet ou en bec-de-corbinsont aussi couronneespar une moulure arrondieou par un listel: celles des toits 12 et 13, S. 157 et S. 202 (serie 23) a Delphes,156une sima de Corinthe157et, traditionellement,les simas corfiotes,158dont celle du toit 27 de Delphes.159 Compte tenu de toutes ces donnees, les simas de Mycenes et d'Argosont du etre fabriquees vers 550 av. J.-C. L'etude independantedes antefixes hexagonales a palmette sur lotus renversenous avait conduiteaux memes dates. Ces simas ne constituentpas un debut, une formeoriginelle de sima, mais une variante assez largementdiffusee. Elles attirentl'attentionsur la coexistencede plusieurs formes de profil - cavets a peine marques, cavetsprofonds,profils droits,becs-de-corbin,cyma recta a Corfou.160Ainsi s'explique qu'il soit si difficiled'ordonnerdans l'espace de quatre a cinq decennies, tout au plus,161la quarantaine de simas a feuilles doriques que l'on connaft 149 150

Schwandner,1985, p. 80-85, fig. 51-54, pl. 28, 29, 34 et 35.

TdAI, p. 3-12, fig. 1-11.

A. N. Stillwell, Corinth, XV, ii, The Potters' Quarter, The Terracottas,Princeton 1952, p. 280-281, nOS 59 et 60, pl. 58. Cf. Le Roy, 1967, p. 62. 152 Le Roy, 1967, p. 54-59 et 62, pl. 15 et 100. 153 TdA I, p. 10-12. Sima V: Vlassopoulou,no 13. Sima IV: Vlassopoulou,nOs11, 12. 154 Schwandner,1985, p. 80-81, fig. 52, pl. 28:3 et 4. 155 Celle de Calydonest peut-etre la plus ancienne: Dyggve, p. 169-190, pl. XXII et XXIII. Il ne faut bien entenduplus la dater a l'aide du sphinx d'acrotereAthenes, Musee National, 17870 qui ne lui appartientpas. D'apres la plastique des tetes de lion, M. Mertens-Horn propose 575 av. J.-C.: Lowenkopf-Wasserspeier, p. 30-33. 156 Le Roy, 1967, pl. 7, 8, 17 et 100; premieremoitie du vie siecle (p. 62). 157 GM, pl. LXII:8. 158 Sima>>:ibid., p. 125-130, fig. 94-98; Temple d'Artemis : KorkyraI, p. 97-124, fig. 73-93. <
134

MARIE-FRANQOISE BILLOT

actuellement.I1ne s'etablitjamais de correlationunivoque et exclusive entre les profils, les elements du decor, sa composition, sa monochromieou sa bichromie, sa facture plus ou moins soignee. Le decor de la sima de Mycenes est plus riche que celui des simas d'Argos, son execution incontestablementplus negligee: mais quelle est donc, des trois, la sima la plus ancienne? DU PROFIL DROIT IV.2. ORIGINES

Notons d'abordque le cavetaffectetout au plus le tiers superieurde la sima dont le reste du profil est justement droit. Des lors que des cavetsplats et tres peu profonds(AcropoleV, sima d'angle d'Egine) offrent au jour toute la surface des feuilles, le pas est aise a franchir au profil tout a fait rectiligne. Les fouilles de l'Agora d'Athenes ont livre les restes d'une toiture en marbre insulaire composeed'une sima de rampant en cavet tres ouvert et d'une sima d'egout a profil droit: les marbriersn'avaientpas juge 1'ensembletrop heterogene.162 Pour ce qui est des simas a feuilles doriques, le profil droit ne peut etre finalement que l'aboutissementd'un schematismetoujoursplus sobre. Par ailleurs, les simas en marbred'Athenesoffrentune bonne serie de profils droitsdes la premieremoitie du vIe siecle.163Directementaccessibleau marbrier,ce profil est aussi le plus economedu materiau;frequent pour les simas d'egout, il rend plus aisee la sculpture des emissairesd'eau et se prete facilementa la realisationd'un decorincise ou de faible relief 12 et 13 de Delphes (Le Roy, 1967, pl. 98) sont similaires;le rapportc/b oscille entre 1/1,05 et 1/1,3. De la meme maniere,AcropoleVI (Vlassopoulou,n?24) est comparableaux series 16, 18 et 19 de Delphes (Le Roy, 1967, p. 100): c/b = 1/0,75-0,87; on notera, au-dessus de la tresse, les bandes obliques sur l'une (TdA I, fig. 8), les chevronssur une autre (serie 18, Le Roy, 1967, pl. 14). La sima d'Egine, naguerejugee tres archaique, date en fait de 575 environ;le profil en cavety varie d'une piece a l'autre,tantot anguleux et sec, tantot recreuse et plat, et a peine marque a l'angle; la tresse tient en hauteur une proportionnon negligeable: Schwandner,1985, p. 80-85, fig. 51 et 52, pl. 34:4 et 5 et pl. 35. Il faut admettreici qu'une certainevariete dans le dessin du creux des cavets - variete souvent irreductible aux rapportschiffres - releve tout simplementde la libre initiative de l'artisan,et n'autorisepas necessairementa trop disperserdes simas que leur facturerapproche.Des lors, il est probableque les simas <(Le Roy, lue. Mais apres tout, rien ne prouvede manieredecisiveque le cavetderivede.la <>, quement reuni la retombeedes feuilles et le couronnementen un unique bandeau surmontantle cavet (pour voir KorkyraI, p. 132-133, n? 5, fig. 104:d). Meme prudence, mesures diffeune formule de <
TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

135

que la peinture a l'encaustiqueviendrarehausser.II n'est pas impossiblequ'il soit passe du marbrea la terre cuite dans le secondquart du vIe siecle. Par la suite s'installe chez les marbriersdes fles et d'Athenesune traditionde la sima a profil droit, susceptible de recevoir le couronnementd'un petit cavet, ainsi qu'un decor vegetal peint ou sculpte. N'ont de decor que peint deux seies de l'Acropole,164la sima d'egout du temple des Alcmeonidesa Delphes,165et celles du temple A de Paros, en cavet, a bandeau superieur assez developpe mais tres hautes (28,7 cm) et planes sur les deuxtiers de leur hauteur.166 Quant aux simas a decorssculptes, elles sont l'oeuvrede marbriers insulaires ou micrasiatiques aux tresors delphiques de Siphnos167et de Marseille,168ou encore a Thasos169 et dans le Peloponnese a Amyclae.170Elles n'exercent pratiquement aucune influence sur la productionen terre cuite de Grece continentale.Par contre,la plupart des simas en terre cuite d'Asie Mineure, a profil droit le plus souvent couronned'un bandeau,et a decorvegetal en relief, sont certainementles succedanesmodestesde simas en marbre.171

IV.3. LES SIMASD'ARGOLIDEA PROFILDROITET DECORVEGETAL

Au nombrede quatrejusqu'a present, elles n'ont de communavec les simas en marbre sculpte que leurs frises de palmetteset de lotus relies par des arceaux,et font plutot figurede parentes pauvres des simas attiques a decor peint. Leur style, parfois empreint de maladresse,n'est marqued'aucunparticularismelocal. Sauf sur la sima de Mazi, la morphologie des palmettes releve de la meme koine que celles des simas attiques, ou encore des simas <<megariennes>>. Des lors, il devient assez aleatoire de preciser une date dans la seconde VIe moitie du siecle. Toutefois, la sima du sanctuaire d'Apollon Pytheen (ci-apres), au meme profil que la sima a feuilles doriques, provient peut-etre du meme atelier (cf. les antefixes), de meme qu'une sima tres soignee de l'Heraion. Sur toutes trois, on note l'utilisation du meme vernis rouge vineux. 164

a) Temple <<des Pisistratides>,sima d'egout,ca 520 av. J.-C.: Porosarchitektur,p. 124, pl. X:1 b. b) Sima a palmettes aux rampants dans deux compositionsdifferentes, a palmettes et lotus a l'egout, egalement dans deux compositions: Porosarchitektur,p. 182, pl. X:2 a et b et X:3; W. H. Schuchhardt, AA 1963 (ci-dessusnote 48), col. 812-815, fig. 15-17: debut du ve siecle. 165 F. Courby, FdD II, i, Le sanctuaired'Apollon,la terrassedu temple, Paris 1927, p. 104, fig. 82, pl. 13. 166 A. Ohnesorg, <<Einparisches Marmordach>,AA (JdI 93) 1978, p. 333-342. Autre sima de marbre en cavet a l'Heraion de Delos: A. Plassart, De'los,XI, Les sanctuaireset les cultes du Mont Cynthe, Paris 1928, p. 200-203, fig. 169 et 170. La petite sima en terre cuite de l'Heraion fait l'effet d'un succedanepeu couteux. 167 G. Daux, E. Hansen et M.-C. Hellman, FdD II, iii, Le tresor de Siphnos, Paris 1987, p. 216-222, pl. 101-103. Voir aussi, evidemment,le decordes larmiers,pl. 89-93. 168 G. Daux, FdD II, ii, Le Sanctuaired'Athe'na Pronaia, fasc. 1, Les deux tresors,Paris 1928, p. 68, fig. 69, pI. XXIII. 169 M. Launey, Le Sanctuaireet le culte d'Hracle's a Thasos(EtudesthasiennesI), Paris 1944, p. 108, fig. 6. Cf. J. des Courtils, , BCH 107, 1983, p. 133-148. 170 Corniche d'Amyclae: E. Fiechter, <,JdI 33, 1918 (p. 107-245), p. 154-158, nos 65-71, fig. 28-35, pl. 13 et 14. Sur l'ensemblede ces questions,voir E. Langlotz, Studienzur NordostgriechischenKunst, Mayence 1975, p. 45-48, pl. 7. 171 ATK, pl. 10, 46-47, 50.

136

MARIE-FRANQOISE BILLOT

a) Argos, Sanctuaired'ApollonPytheen, C. 23662 J. Deshayes, Lesfouilles de la Deiras (EtudespeloponnesiennesIV), Paris 1966, p. 229, pl. XXVI:8 et XXXII:8 a gauche.

Petit fragment d'angle. Comme sur les simas a feuilles doriques, le profil droit etait couronned'un demi-rond. Du cote du rampant, amorce, cassee, de ce qui pouvait etre un acrotereet ne semble pas avoir eu d'implantationsur l'autre cote: ce ne pouvait donc etre, comme a Egine et sur Acropole V, une fausse antefixe. Restes d'une palmette peinte en rouge et d'une tige vegetalequi remonteet se recourbevers l'angle : ce peut etre une maniere d'acheverle decor.Ce peut etre aussi l'indiceque le motif se composaitde palmettesalternativementdroiteset renversees:dans ce cas, la sima serait un peu posterieurea 530. Elle peut etre associee aux antefixeshexagonalesde la serie C (ci-dessusp. 110-1 11). b) et c) Heraion d'Argos Deux simas decoreesde palmetteset de fleurs de lotus relies par des arceaux. b) Une sima tres soignee, mais tres fragmentaire,peinte en rouge commecelle du sanctuaire d'Apollon,est associee d'une part a des tuiles d'egout decoreesde fleurs de lotus renverseeset de palmettesdroites,d'autrepart a une thtede lion. Les spiralestres larges et tres enroulees du decor de la sima et la plastique de la gargouille conduisenta une datationdans les annees 540-530. c) Une sima de facture negligee, peinte en rouge-brun (Musee d'Argos C. 23732 et C. 27510 [P1. 11:g]) GFR, p. 25-26 et 85, no 48, fig. 74. 172 Dernier quart du VIe siecle.

I1 est sur que l'une et l'autre sima doivent etre associees a certaines des series d'antefixes hexagonales B et D-I. d) Mazi (Oinoe), Musee de Nauplie, inv. 2899 Htibner, 1975, p. 128-129, fig. 7, pl. 68:1 et 71:2.

Au vu de la palmette, il s'agiraitaussi d'une sima plus recente,du dernierquart du vIe siecle. Commepour les antefixeshexagonalesa palmettessur lotus, l'impressionprevautque cet ensembleest une productionlocale.Mais evidemment,les termesde comparaisonfont defaut. En tout cas, les argilesne paraissentpas corinthiennes.Rappelonsaussi que la sima d'Asea,a decorcertesgeometrique,ajustementun profil droit.173N'est-ce qu'une coincidence? V. TOITS <<MEGARIENS>> ET DIVERS Pour offrir au lecteur un meilleur aperSude l'architectured'Argolide,nous enumeronsici rapidementplusieurs pieces ou ensembles qui ont toute leur importance,mais contribuent moins que d'autresa la connaissancedes particularismesde la region. 172 La piece a ete rapporteeen 1973 du Musee de Corintheau Musee d'ArgosoCu elle porte le n?d'inventaire C. 27510 (P1. 1l:g). E. D. Van Buren a reuni sous la meme notice, outre ce fragment,un bloc d'angle appartenant a la sima megariennede l'Heraion (ci-apres). 173A. K. Rhomaios (ci-dessusnote 133), p. 120-122, fig. 8 et 9.

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

137

V. 1. Ce sont d'abordtrois toits de type <<megarien>>: 1) Heraion d'Argos: troisiemequart du VIe siecle Plusieurs elements d'une sima de rampant, dont la piece d'angle: AH I, pl. 23, G. GER, p. 7 et 87, no 56, fig. 9 et 75. GM, pl. XVIII:4. <>, p. 248, note 59.174 Cf. ici, Musee d'ArgosC. 27517 (P1. 11:h). Une antefixe a palmette libre sur fleur de lotus renversee.Elle est attenante a une tuile a tresse. Fragmentsd'un acrotered'angle et d'une palmettefaitiere. Cet ensembleprovientcertainementd'une refectiondu toit du temple periptere. 2) Epidaure: dernierquart du VIe siecle Un fragmentde sima: GM, pl. XVIII:6. Une antefixe a palmette libre sur fleur de lotus renversee.Elle etait attenante a la tuile d'egoutqui a disparu. 3) Argos, agora Un fragment de sima decoreed'un meandresur le bandeau vertical, et d'une sinusoide sur l'ovolo. Ce motif est assez bien date par les loutrophoresatheniennesdu debut du ve siecle. Une tuile d'egoutdecoreedu meme meandre. Un fragmentde palmette faitiere. V.2. Par ailleurs, lors d'un nettoyage du Sanctuaired'Asclepiosa Epidaure, on a mis au jour une antefixe a spirales sur lotus renverse',d'un type apparemmentnouveau mais pour lequel les elements d'une reconstitutioncompletefont encoredefaut.175 V.3. Au nombre,enfin, des rares vestigesd'un edificequi aurait precedele temple classique d'Apollon Maleatas,176il faut compterla partie gauche (pour le spectateur)d'une criniere de gargouille leonine recomposeede deux fragmentstrouves en 1977 et 1978, 0 15/57 et P 16/72 (P1. 10:g): Gargouille decolleede la sima et cassee de tous cotes Argile corinthienneuniformementjaune-verdatre;engobe de la meme argile. Larg. = 11+; Haut. = 13; Long. = 10 cm.

A la cassure inferieure gauche, on observe qu'il n'y avait, comme il est frequent, que deux rangs de meches autour de la gueule et du menton,mais trois rangs autour des tempes et du front; epaisses, plates, bien individualiseeset flammees, elles retroussentleur pointe 174 On trouve mention dans GFR (p. 89, n? 69, fig. 9) d'une secondesima dont L. T. Shoe pretendrapporter le profil en GM, pl. XVIII:14. Mais ce fragmentn'est autre que C. 27517 (P1. 11:h) naguere au Musee de Corinthe avec le bloc d'angle C. 27509 (GFR, fig. 9). Son profil, identique a celui des autres fragments(GM, pl. XVIII:4), ne correspondpas a GM, pl. XVIII:14. 175 A. Archontidou-Argyri,AeXr 32, 1977, B' 1 (1984), p. 49. Je remercieMadame Archontidoude m'avoir permis d'etudiercette piece exceptionnelle. 176 V. K. Lambrinoudakis, <<'Avao-Kac/7j a-r0 Leporov 'A7roXXcvo MaXfarao, rlpaKrLa 1977 (p. 187194), p. 189-190, pl. 118 et 119; HpaKTLa 1978 (p. 111-121), p. 117.

138

MARIE-FRANQOISE BILLOT

vers le haut. Au sommetdu front, une mechetriangulairemarque,au premierrang, l'axe de la criniere. Le seul rapprochementpossible - mais il est etroit - se fait avec les lions du temple des Alcmeonidesa Delphes, gargouilles des angles177et fauves terrassantun taureau178et un cerf179au frontonEst. On peut egalementsongeraux lions atteles au char de Cybele a la frise Nord du Tresor de Siphnos,180au moins pour la dispositiongeneraledes meches;mais elles sont plus espaceeset leur sectionrondereleve d'une esthetiqueplus ancienne. CONCLUSION A partir du vile siecle, l'histoire du systeme ?corinthien>de couverturese joue simultanement dans plusieurs regions de Grece, par approximations reciproques, adaptations successives, avec utnecontinuite - et une lenteur - remarquables. Corinthe, Isthmia, Olympie, Thermos et Corfou, Kombothekra,Egine, Halieis, Mases, Delphes et Kalapodi, Nemee et l'Argolide: rien de ce qui se fait ici ou la n'est ignoreailleurs. Monopole et <>de Corinthe? Ou seulement circulation des idees, des modeles et des modes? Nous n'avonspas (encore)trouve, dans la grandecite de l'Isthme, les elements qui nous auraient permis de tenter la meme typologie ni de retracerla meme histoire. En definitive,les modes et les initiativeslocalesjouent un role majeur. Une forme s'impose presque partout pour son efficacite,l'antefixe hexagonale, surtout a cotes inferieursconvexes.Le succesqu'elle rencontresous des formesplus ou moins deoratives'aEgine, dans les cites cotiereset au centrede l'Argolide,ou encorea Nemee peut etre considerecomme l'un des premiers indices d'un style regional d'architecture.En tout cas, c'est en misant sur ce succes qu'Argosdeveloppesa propreproduction,oCu les simas 'aprofil droit tiennent par ailleurs une large place. De son cote, Corinthe s'essaie, a partir de 580-570, a des formules plus variees, plus stylisees aussi, plus decoratives,ce qui fait le succes de ses exportations,ou de son style. Mais elle ne jouit d'aucuneexclusivite, et jusqu'au milieu du siecle au moins, son rayonnement ne depasse guere les regions limitrophesdes deux golfes. Encorefaut-il nuancer: si le toit <<megarien>> de l'Heraion d'Argosest suirementimportede Corinthe,celui d'Epidaurea du etre fabrique sur place par un atelier itinerant,tandis que celui d'Argosest une authentique productionlocale. Hormis la tete de lion du sanctuaired'ApollonMaleatas et une antefixe du sanctuaire d'Asclepios,aucune serie archaYqued'Epidauren'a de provenanceconnue. Du moins peuton affirmerque rien n'est anterieurau milieu du vIe siecle, ce qui correspond'ace que nous savons a peu pres des debuts de l'architecturedans les deux sanctuaires. "I

F. Courby (ci-dessusnote 165), p. 103, fig. 2. Lowenkopf-Wasserspeier,p. 45-46, pl. 7:b. F. Courby (ci-dessusnote 165), p. 104, fig. 83, pl. XII. Ch. Picard et P. de La Coste-Messeliere,Sculptures grecques de Delphes, Paris 1927, p. 19, pl. XXIII:2 et 2 a. P. de La Coste-Messeliere et G. Mire, Delphes, Paris 1943, pl. 148, p. 326-327. Lowenkopf-Wasserspeier,p. 45-46, pl. 7:c. 179 Picard et La Coste-Messeliere,op. cit., pl. XXII; La Coste-Messeliereet Mire, op. cit., pl. 146. 180 Fr. Willemsen, OlForsch, IV, Die Lowenkopfwasserspeier vom Dach des Zeustempels, Berlin 1959, pl. 1: excellente photo de detail. Lowenkopf-Wasserspeier,p. 48. 178

TERRES CUITES ARCHITECTURALES D'ARGOS ET D'EPIDAURE

139

La collection de l'Heraion d'Argos - deux simas, huit series d'antefixes,une toiture megarienne- s'echelonneentre 560-550 et 520 environ. Trois edifices au moins ont ete construitsou refaits. Dans la ville d'Argos,les terres cuites les plus anciennesremontent'a570 environ. On constatela meme activitequ'a l'Heraion entre 560-550 et 520, notammentdans les centres vitaux de la cite, agora et sanctuairesoiuplusieurs chantierss'ouvrenten memetemps. Prosperite et grandes entreprisessuscitent une productionlocale dont l'originaliteest bien plus vive qu'elle ne sera au ve siecle. On note un certainralentissementde l'activiteedilitairevers la fin du vie siecle, et une remarquablerepriseaux environsde 470. Argosest alors sortiedes difficultesconsecutivesa la defaitede Sepeia. Une autre phase de son histoirecommence. MARIE-FRANQOISE BILLOT

CNRS - IRAA 38, rue Lacepede 75005 Paris, France

PLATE 10

b. Face A

c. Face superieure

a. Angle exterieur a-c. Heraion d'Argos,tuile d'aretiera acrotere.Athenes, M.N. IA 293

d. Argos, tuile d'egout C. 27701, face anterieure

f. Face

e. Cot'egauche e, f. Argos, ant'efixeC. 19249

g.Eiare

_~~ M,!-FAyOS

acuar

'polnMlets

et

elo

~

1557P167 BILOT

TERE

CUTE

ARHTCUAE_'RGSE

'E

AR

PLATE 11

a. Epidaure, antefixe ME asr 353

w

c. Face

b. Heraion d'Argos, antefixe. Athenes, M.N. 18413 (serie F)

w

d. Cote droit c, d. Argos, agora, antefixe 76/1835.1 (serie D)

e. Argos, agora, antefixe C. 11482

f. Argos, Sanctuaired'Apollon Pytheen, sima C. 26701

_or~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

g. Heraion d'Argos, simas C. 23732 et C. 27510

h. Heraion d'Argos, sima C. 27517

W q IL uo -rIaodttsvdDi 'ISaoitvyL 1f3Xt3 aoL StL aioftLn 10o'X'L 'o#fa) 'DLtrLAOoo?)1 O 73gdXy X s(o 37r 7o soduL saotlrl ao.L t/l ofrvarn ava3 3r/a0oo7q3X.0o aV OoD /wi vqvanorl W/Z SOiQa o)iiyoaao o-i sod naod>vOLD so/ifa 0- l7 'C/Lso4/ a sodumsao.LvoyL otoy aaoX) aoWLS7DL.^DqOv2 sI7ant aoiL iLV 3JJ :1'.V73) orlvdiodov oimdiL o0j W ?rl?aofyyf 10 'odamD2x1 atinOLL7Iva (.Su 3D loD .oa4) a tcit [ n aWi )naoaao0rl I o nd lt *o0?-LorI oI7LttLrlDoo)D 9dI) ivw14na(04dorl rIao IL)ySfvdg O7OLO ao[LDioV noL o rlaaQL o.Lvadg7L? O -vo)rl aQOLStDILtoavi[ L2 'sI Os ao naavn n7ova dorl amt or>r7dXc7yaj..o ox?dn(-)Do .L.ODrIonv9D1D .t aou 'odtiJLoa orload O oL tti otItd) adw tL)? v0rio r7i?7C 713X? Warlo ?OIL? tLD)9,Qnao t 'n(a)IQ.LodL n(am)Ia nq a n vrl 73n73 svrl naoDav D -itgoDLO QOIL'ItDDLtDdvXa? -Ad ro t ILtdo Df H ILtx.dw *D 'S ySj3)io.Lao-(o3y70 5)1LsOsf dorI S?)17C anoXdDva rvanrlnorlbrl r OZL0 7t73.YVItytotV Ma3 '1VlVrlnb0 vY.La4q) 17) S3aL)OvDtLv s vd 9Ii,A H LtDtrlOO)1V2 '*oA13d7os3 vinrlvr/vdmL o.LDoyorl7-xo 'Vrlv^ 737M/ on o74 'StDtoty.O)D7zg sltzt loltvav nali i7,CtdaoIrlQltstk .Lo s.L ILDo3 ILaoyyv -3 'vwv.DD.oodLtltanrlao I.?l n3L 7[3D)7y D)i Iyt k 'D73Xo0 t aIn7oLLo anIi ILymvIt 7 ) I ta3)o o) sao.L amSt3r ltorldv4 -v+o1S>3 0L 7v rDyyD 'ao7dt.L vdof)xgq OILo) 7I, t).IOILv 'SItOQao s$Y71diDsL oILDoldtL.I o.L 010g77 aQO odoX o,7d3c.L)3 no73.L L/LoD H 7IaLvLDodIL aozL tLa.a.LDvi tLtLfnaaD Dmrl7nl slt.tar

slt^.y3rlI

D. s(t i oa.rl7.4. nm o.Lo aaoXdiLQa aOmLam2X73o.1 ami t oay\vn ap7rl/mam PoI rmd] I sao.L ao4Cdao7irlfgao.i saoXoLo oiLrD sao.L aaot3dtiLaS3 S(^r.Do I7vi 'ILL^3 QOLos?2/dQom^3y S?1iLDr7llDt1ooSaa1o s71oD Dl3 S naoX.4 on3trlrl7d.XaD ona 'oyod oanrlrl7dv3YtaDonvn' 'I1inX.t ntLtD )700oDoy(b nWao)m oric nc)doVi7g 1 ma3 nmnidLndt 3 Lmg27 c'nc&odDoL 31.73' 7.oD lD 34t7n aaoi)dUttaQ3 tlj.L.do()vi not 07C m MiVi nl am n.' Irla 'madAd n[nornoi.iXddv orlOnri.rlr/lvdodiL nt[LOD0aaD Uttorldvo3

avna3 hanodXvi2 naaoy3ioIL si)3.vdDX sis,[anorIdv 10 7-Lo D.M^.L.7ILi2 aOILO'so/Ad, >vO7g2)Y aoaa(vm ao7I v aQOL OLIDaorlod2odi QaOL so(r[ sod3i o sm Lo 'sIXo0idIL stit v7idttiLtvd3.L)3a[ 3.0 71v Dyyo '[amXoIL3amdod?72 'sao/CdVy aot svdoCv s 7vXdv siLL vidt^ ^ Do.L sCmdQaI 34Xrl V7aodX v7va.Ly .L, vZj It 13DL0yOXDd0) 'amSvdDvX nmminorldvrnami ty3rf 73IX3 saoCdV aoi.t Qao7dH ao.t iDv Sao/dV QO.t 'aodav-Q1L sIL n(m7dtdLi n[c)7m Xdr a t Ly37rlD([7yUiL V)l7not^L3.LXdv DuL D(ott3oy.rl Dn iDxwroo72)g3X. vXIL It2t 3X73 al 7.1.7iXdv QOn D,1diQa? n( nOt*7o.Lv.7DXdvD amt &It3yn 'IammytIL m a(iI -D7.ov0d?3aa aoIL ';olig Lrl nv 'ItyoX: tt/t7oyo7oXdV F7 sdN sLI )sio5ueij-iuWjN IV I 74 QaoI ZHJL VI3IVVLI VL & ItLyyIoj aIO pLLD 'oLVNAg3d3 :HNIJOVOIVXdV

AOd,AVVIU 3HL IVYIJAOJdV AOlJ VNIVHLI VIINOI)L3IXdV

VNIVXdV VJ:L lI35VdVX 33NINOINdV

1 1----= |

----

-5 5

1

~

L___ 1 _ __ _ ____

i

rl r--*///-w/^'

14.

~~~~~~ _ _ _ ____________

_ __

_______

____

___

/~~~~~~~~~~~~~~~~~~~~~~~~~~ I~~~~~~~~~~~~~~~~~~~~~~\ /~~~~~~~~~~~~~~~~~~~~~~~~~~'

I

/

I

L - - - . J- - -

ELK. 1.

- -

-

_ - - -

-

-

- -C.-

- -

-

E7rbavpos (ME a7r 381 KaLava7rapao-rao-r e TroME a7r 353)

-

-

A14

I

1

CI

\

ocl

0

I I

mtl

" -

1

18.7

-5. 15.2

kOdACVVIIAI3HI IVN LkOJdVk OI 3I3SVdVX I3NINOAIdY

144

K. G. KOLOKOTSAS

ELK.3. 'Apyo (C. 19109)

ofO0a?uotl eival rTTO

/lEOO

TOv

EVOELWv rov rerpayovo

eyyeypa/LElvoL viovs,

Kal

KaVaJ3ov,

T?ETeTpaycova

TO avT(epo TETIrT

TO E

-vvoXo

TWrv

1 X 1. To avw /LEpos T7)s TaLVLas 8pUT-KeraL

a Ts rLa

EXLKO?ELOVS' Ka/LArvA7X)s (oa7TTETaL

TPLCV fvXAWV T7)S KOpvLq7s) eyypaeTra

TWV

euE eva

2 x 2.

Ao-aAXcws p7ropovic

va 7rapaT7rprjr-ovjLE? KaL rrfv KiLV?(r-7 7T0V XapaKr71piLe?

Latpya

T77s XEtporeXvLaS. ITO 8evTEpo

KO/.4LaTt (ELK. 2)

AXE'TrOVy/E MLacrL-a

a7ro rT?v Aor-rdba, 07Tov TO a7rXo Or opOoy4vto /LAE avaXoyLa vi/ovs 7rpoS 7rXAaTros 1,5/1. YEW.ftPLETKOcroX,ya eyypacTraL ITO TpLTOKOL/art, TO CLKpOKepap.o(ELK. 3: C. 19109), avaXoyla vfrovs 7rpos 7TraTOs? eLvat 3/4, KaL Lc 7rl/v oXeal(Tarr Tov KaTraAXrX0ov Kava/3ov 8La7rLrrTcovov?E, OTL 73 bLaKOorOTCS /8AcE7Tov/fle a7ro Tn/V Oe(T^ Tcov ofOaaX?i()v TOV av0r4'iov /XJ)Oc- TOY aKoXov0r)-e, KaL T?7Sratvtas. Ta E7ro/jEva bVOKOpluaTia EivaL /.La abOKLILT.0rE /LOoLECvyEVa aKpOKEpaula. avTo (ELK. 4) /LE Tr/v (XE?8lao(?/ rov KaTraXX7Xov Kava'fov, pXE7Trov/JE Tro aKpOKEpal/o ont / btaKo'rO-7ffr?rKaLTO E6TEPLKO crXrlLa aKOXovOeL rov Kava8o, Tro avOel4o Trr KOpV/)riS

APMONIKEI XAPA,EI

TOY APFOYE KAI THE EHIAAYPOY

ELK.4. TdA II, or.27,

rv -

I \\

7

ELK.

36

I

A_

/I

i-Ma/ 'I 2/3

I_

N.I

1/1

.._-l/

I I

ELK.5. Corinth IV, 1, 7riL. II

I

145

j~~~~~~~~~ ~~~I .I

I I

14.0

.I I l

-------- - ----.- - ----1 '--L------------------..1 -~~~~~~~~~ ~~~~~18.&~~~~~~~~~~~~~~~~2

ELK.6. E'rlbavpov (ME aw 351)

APMONIKEEXAPASEIE TOY APrOYE KAI THE E1IIAAYPOY

147

0.S cr)

C4

(.

0 CL

~~~~ I

~

~

~

-

aaa9JQ'su9aqv 08 901-"IO aaJ1Snolop!a 9 S3N3HIVUa

3SIv5NVHIJ[

ao03

SVSIOMO1ON[ *0) *N

aoAod3 soan 'anoy,j Dz daoi7rlhQ atl. I3r nif?)7D07LoILorI7DttdX D 'o07 otlov. aa f n Vt" 'atndIt 57L van 'tDfotrloait afiL naDo adL 'totdd4rl anti 31 's73DvoLvdVdiL s3oDaonaSf7ry t. (7 a X.n yy aO O.LV I.L4a7/t 'so?vrlQ0 ao1 tt D.oDaoi r -DodIL S?, 730oa0iDw 3v D'.DL3 tt iv?i tfd dorlvLrI tJL 't lDttovydLor t va 73LudiL 'orl0ad oNid3umox? 'Lo rIlao.LX2odv?idD no?oLoo

.rlaonamLjDtLmvQ no-oo()4

7^

tltav

tDjtoLm7d.L

at.Do

'a7di

3r,D2Q7 aomL s3Q)mo

va s7idX 'tg0oaaLD aftL v.i VX7(VdV 7C 30t rofl ?3DvQ.ooma.oLz i -vdvvw ?7 tn a Md Dm Lid3 ftti v iVXdv?a scoanrloXaqn4 scwrio od4aDrf) .oftidaoi7?y vL iva7 7L0 crdHona's7v1pdvX sinaorldv s? DtLQnaao Vt shaX t7yy'DV 0 730 ioTIorLdX 0no '?ar7df vyyv 'sanrloidoo"i,odL 7XoSIo .D XD) s)7.Ltri -37rI3c, afl 3I aao,S7vmL lo 0 st.otDrlowQV st. t t0.aa at.L -07id7v7i s)17id?.rkc aaonioLor70lttdX sa50 7daoitrlq2 t>a vL.LT rlaodolLrl'namSvdvX amn7iaorldvamnoamnLanttdLtdltLvdrvDlaL z?iA uop 3?raoiL ni 7 skt.i aman sDLt.otrlow iqS ao.L aDrl t k X.o ftiid.t3drlwm 3 Lt tii o ai?ynitrlft D 3,r stotstrl/.ooviq7 stt.L ftoc.iLraoD i" ant.ktrHaoa"LD.oi.Dvi? rliaoSvL3 aoa (L i*^gi) 7irlrlo, o0v.aLQy3 a or7o-fD3 O.LD7Dv o o n wLoD aoL odtftco oLn 0o SLiL 0s1X10ft ti.Lrl/TDdIL ia?)3 stD.otrp.Doiq ao8fvnvx aoL. o. no.i oirlha,a DTLD7LoDooI aod4d)viq72a a,QyOIL?1ntr/4xdu. MIa7f *saoI/fa i70aoyod orlgad -VDL"D s)7diLr'3ot

su S1

0)07DLOd.7dLLOL aiLOm soyvav o i7vi 'td,rl VD.7 OaLD 1?3X3 VL -rI ftDI odt3rlIq 07Wr/ g c.LmIO 3oDsaoiMa aoJ Itodivi7omla atkt OUIoi7iLi,avodmLaomlodfvar) avan 7iaoyow tLt.Dtrl.DoV t OLv 'i.rlrlovi du akti O.LD o.0x orlvd-3odxD 0ot -102 g^ :9 'v7g) (ISC 'odav%mLgj v>vi'saoifna Qaoi /Z 0-L7Da)I 07irlaV o.?qi)vdbid^L noxl aoyniav aot sod.3r7liq H S).?y43 S7.L Mi saorlyrf)o sao.L mvl ThLsodZLi7), vio7dv altiL monav? t sodiL i7atnuv no.?rl o -tdrI V0 7 DdoDD3L D SQO/aQaOJ.L 3)1itD7Ioyoa ttod,wivd oJua atLi. OILvD QaosL 7iw '?ltLoDl73X. aoQL olaflfav ao.Lv7)rlft vyyoL ft ao.f tOftrloovwD7i 3o 73lftoyioyovv 1?3X aorlvd,odv (S V7g) 7LDr0rlOV0oITI3L 0.L3 M oaoavD acoacvdi n nc.Lnao a vian taL nra H i3na7 *stQi3o07iy3 )D yD) 5 x s Ix ? x ? 'I x I vnoCvd.id.L oL7it" 'vjovrlftX I 1i 3473 3.173 i7o sftomJL3rnaOan.LiD71LViQ ?riaon7idi7 vna rIaodolLrlond DL. xDw7.Lnadv r?LD 0 7 Z x/ t sd ao.i t/C

Z x Z 0/oaotzd.L232 I

std 3.0 Ii 7/23(^1od4U3 I()dC/)stdvstL

5fty3

svsioyoriox,

i I

tt Ij X ^ 0a(0dC.?23. Oa~iVd?.

0 N8

3-0 ?23<.?4d im4vb

3

8bHl

THREE-PEAKED ANTEFIXES FROM THE ARGIVE HERAION (PLATE 12)

AMONGTHE EARLIESTATTESTEDSYSTEMSof roofrevetmentis theprimitive horned roof (Hornerdach), a roofing system characterizedby unattached, pitched covertiles, flat pan tiles, and undecorated,three-peakedantefixes.This type of roofis thus far attested at Nemea, Halieis, Mases (near Halieis), Kombothekra(near Olympia), Delphi, and Aigina (Fig. 1).1 In this paper I shall presentevidencefor at least one and perhaps two differentroofs of this type at the Argive Heraion. The evidenceconsistsof four fragmentary three-peakedantefixesand the rear portionof one covertile. Three of the antefixes(1-3) and the single covertile fragment(4) probablybelongto the sameroof (Figs. 2, 3, P1.12:a,b): they all share similar dimensions,the same tan, gritty fabric,and a reddishbrown to black wash paintedover the exposed surfaces.The formof these three antefixesis very similar to that of the well-known antefixes of the Early Archaictemple of Aphaia on Aigina, which likewise have a brownto blackwash.2The fourthantefix fromthe ArgiveHeraion (5, Fig. 4, P1.12:c, d) is somewhatnarrowerthan the otherthree. It is unpainted(at least as it is now preserved), and its fabric is rather more orangish in color. Even more significant is the fact that its undersideis not angular like that of the otherthree antefixesbut curvedlike the undersideof later Corinthiancovertiles. A parallel for this apparentlyadvancedfeatureis providedby an unattributedthree-peakedantefix at Nemea.3 All the fragments from the Argive Heraion seem to have been unearthed during the excavationsundertakenby Charles Waldstein at the end of the 19th century,but none was ever published.4Unfortunately,there is no recordof the exact circumstancesof their discovery, henceany attemptto attributethem to a specificbuildingor buildingson the site must depend upon our dating of the antefixesand the chronologyof the buildingsof the Heraion. At present, the available evidence for dating these antefixes is meager. With the exceptionof the Temple of Apollo at Halieis, none of the buildings associatedwith a primitive hornedroof has been scientificallyexcavatedso as to providea firm stratigraphicdate. Even I

Nemea: N. Cooper, 1983, p. 64; S. G. Miller, "Excavationsat Nemea, 1979," Hesperia 49, 1980 (pp. 178-205), p. 185, pl. 39:b (antefixes of more than one roof). Halieis: N. Cooper, 1983, pp. 33-47 (fragmentssurviveof all elementsof the roof except the raking sima). Mases:J. Dengate, "The Archaic Doric Temple at Mases,"Abstractsof Papers 76th GeneralMeeting of the ArchaeologicalInstituteof America, 1974, Section IIA, p. 22 (two antefixes). Kombothekra:U. Sinn, "Das Heiligtum der Artemis Limnatis bei Kombothekra,"AM 96, 1981 (pp. 25-71), p. 50, pls. 15:5, 16:1, 2 (one antefix). Delphi: Le Roy, 1967, pp. 28-31, pl. 5 (one antefix, one cornertile with false antefix on one side and protomeon other, one pan tile). Aigina: Schwandner,1985, pp. 72-85, 126-128, pls. 25-27 (fragmentsof all elements of the roof). 2 Schwandner,1985, p. 76, pl. 26. 3Cat. no. AT 78, from Well K14-4, on display in the museum of ancient Nemea; for the well, see S. G. Miller, "Excavationsat Nemea, 1978,"Hesperia 48, 1979 (pp. 73-103), pp. 77-81. 4 For the results of the excavations,see C. Waldstein, The Argive Heraeum I and II, Boston/New York 1902 and 1905.

150

CHRISTOPHER A. PFAFF

DELPHI

RAvx .

KOMBOI<

ARGIVE HERAION

rD

XR'A

0

FIG. 1.

50 km

Map showing distributionof primitivehornedroofs

at Halieis the date is not secure owing to the uncertaintiesinherent in the excavationof a submerged site.5 Faced with the dearth of excavation data, we are left to rely upon the evidenceof the architecturalstyle of the associatedbuildings. In fact, the only building well enough preserved to provide a firm basis for stylistic assessment is the early temple of Aphaia on Aigina. The date of ca. 580-570 B.C. proposedby Schwandnerseems reasonable, allowing as it does for a sensible developmentof general proportionsand of specificdetails of the Doric order within the first half of the 6th century B.C.6 With the degree of caution 5 Initially the excavator,M. Jameson, dated the Temple of Apollo, which is the building associatedwith the hornedroof, to ca. 675 B.C.: PECS, s.v. Halieis, p. 375. Accordingto Schwandner(1985, p. 127, note 239) this date has now been lowered to the early 6th century.This later date, however,seems to result not so much from a reappraisalof the stratigraphyas from a re-evaluationof the probabledate of the roof revetment. 6 Schwandner, 1985, pp. 128-129. The attempt by D. Williams ("Aegina,Aphaia-Tempel IV, The Inscription Commemoratingthe Constructionof the First LimestoneTemple and Other Features of the Sixth Century Temenos," (AA [dI 97] 1982 [pp. 55-68], pp. 61-64) to lower the date to the middle of the 6th centuryB.C. is, in my opinion, unconvincing.

THREE-PEAKED ANTEFIXES FROM THE ARGIVE HERAION

J

f

.046

....

A ...

~

~

FG

2. ArieHrin

.20m neie

-

151

CHRISTOPHER A. PFAFF

152

BACK

TOP

SIDE

4030

.0099 _

036

0

.05

.20m.

FIG. 3. Argive Heraion: covertile 4

that must always accompanyinvestigationsinto the art and architectureof the Early Archaic period, we may, I think, accept that the primitive horned roof was used at Aigina in about the 570's. With that point fixed, let us examine the moreindirectevidencefor determiningthe lifespan of this type of roof before and after the Aphaia temple. At present the only evidence pertinentto establishingthe dateat which the primitivehornedrooforiginatedis the factthat on Aigina and in the Argolidand Hermionid,where this type of roof seems most at home, no earlier or more primitive type of roofing system is attested.Therefore we should probably posit its originat the verybeginningof monumentalarchitecturein the area,probablybackin the last quarterof the 7th century.As for determiningthe lower limit of the chronologyof the

THREE-PEAKED ANTEFIXES FROM THE ARGIVE HERAION

153

BACK 027

TOP

SIDE

BOTTOM

A

to

-i-044

-X

-+.021.016

.093

1 .058---1

FRONT .053 0

.20m

05

t 0

.020.186

FIG.

4. Argive Heraion: antefix 5

primitivehornedroof, recentlyexcavatedfindsfrom Kalapodi(Phokis) providean apparent breakthrough.7The roof associatedwith the southern,peripteraltemple of this site, which is dated stratigraphicallyby Felsch to 570-560 B.C., clearly follows upon the tradition of the earlierhornedroof, the only significantdifferencebeing that the centralpeak of the antefix is enlargedand the face of the antefix is decoratedwith moldedlinear patterns.On the basis of this evidencewe might hypothesizethat as earlyas the 560's B.C. the plain peakedantefixesof the hornedroof were being supersededby decoratedones. With a general date for plain peaked antefixes from the last quarterof the 7th century to the beginning of the second quarter of the 6th, let us considerwhich buildings might be associated with the antefixes from the Argive Heraion. The most obvious building is of coursethe Archaicperipteraltemple of Hera, which stoodin its commandingpositionat the top of the site until its destructionby fire in 423 B.C.8 Although only part of the stylobateof I

Felsch, 1987, pp. 22-24, 80, figs. 77, 78. E. L. Tilton in Waldstein (footnote4 above), I, pp. 110-111, pl. 8; P. Amandry, "Observationssur les monuments de 1'Heraiond'Argos,"Hesperia 21, 1952 (pp. 222-274), pp. 223-226, fig. 1; A. E. Kalpaxis, FriiharchaischeBaukunstin Griechenlandund Kleinasien,Athens 1976, pp. 42-47. 8

154

CHRISTOPHER A. PFAFF

this temple survives,the evidence it provides-most significantlythe relationshipbetween the lower diameter of the pteron columns and the interaxial spacing-would seem to indicate that this temple representsa stage of developmentprior to the Heraion of Olympia, which is dated to about 600 B.C.9 Since there is no evidenceto indicatethat the architecture of Olympia was particularlyavant-gardeor that that of the Argolid was particularlyretardataire,it would seem reasonableto date the early temple at the Argive Heraion to the last quarterof the 7th century.The thesis that the three-peakedantefixesbelong to the Archaic temple is perhaps strengthenedby the fact that at the two sites where we know which buildingswere associatedwith hornedroofs-at Halieis and the Aphaia sanctuary'?-these buildings were temples. At Kombothekraand Mases the buildingswere also probablytemples, although we cannotbe absolutelycertain."I If we allow for the possibility that subsidiary buildings of the early Archaic period might also carry a primitivehornedroof, a secondbuilding at the Argive Heraion may also I For the date of the Heraion at Olympia, see A. Mallwitz, Olympia und seine Bauten, Munich 1972, p. 138. The following list illustrates to what extent the early temple at the Argive Heraion falls outside the norm of the 6th centuryin terms of the ratio of lower column diameterto interaxial spacing. Argive Heraion, Old Temple . . . . . . . . . . . . . 4.375 Kalapodi, South peripteraltemple . . . . . . . . . . ca. 3.660 Aigina, early Aphaia temple. . . . . . . . . . . . . . 3.217 Olympia, Heraion (side, using column NI1) . . . . . . . 3.176 Corinth, Apollo temple (side) . . . . . . . . . . . . . 2.287 Even in the Temple of Apollo at Thermon, which is usually datedback into the 7th century,the ratio is about 3.857; there is, however, the uncertaintyhere of whether the pteron as rebuilt in Hellenistic times precisely reflects the proportions of the original; see G. Soteriades, <>, 'E4'ApX 1900 (pp. 161-212), p. 174 and Kalpaxis, op. cit., pp. 47-49. Anotherargumentfor dating the temple at the Argive Heraion earlier than the Temple of Hera at Olympia is the fact that whereas at Olympia the stylobate is a fully visible step course resting upon a cut-stone euthynteria,at the Argive Heraion the stylobate serves the combinedfunctionof euthynteriaand stylobateand was originally buriedto half its height. 10At Halieis, the remains of the roof were found right on the site of the long rectangularbuilding that is identifiedas the Temple of Apollo: M. Jameson, "Excavationsat Halieis, Final Report,"AeXtr 27, 1972, B' 1 (1976), pp. 233-236. At the Aphaia sanctuarya large quantity of fragmentsof the hornedroof were found in the fill that containedthe debris of the superstructureof the temple. As Schwandner(1985, p. 72, note 70) argues, the quantity of tiles is so great that statisticallythey must be associatedwith that building. The fact that the visible length of the two best-preservedpan and covertiles of the horned roof from Aigina (nos. 207 and 224) does not correspond to the visible length of the segment of the raking sima of the temple (no. 236/237) might be taken as evidencethat the hornedroofis incompatiblewith the sima and thereforedoes not belong to the temple. A careful review of the survivingelements of the apex of the sima and horned roof shows, however,that tiles 207 and 224 are in fact compatiblewith the sima, so long as they are restoredto the first series of tiles below the ridge pan and cover tiles. For scholarsbotheredby the discrepancybetween the fabric of the sima and the other elements of the hornedroof or by the greatersophisticationof the sima, there remainsstill the possibilityof interpretingthe sima as a later additionto the roof. This possibility,however,in no way alters the fact that the hornedroof belongs to the original design of the temple. I The antefix from Kombothekraprobablyrepresentsthe original roofof the Temple of Artemis Limnatis, for the simple groundplan and mud-brickwalls of this temple seem sufficientlyearly for the antefix, which is the earliest extant roofing element from the site: see Sinn (footnote 1 above), pp. 47-52. The antefixes from Mases were found together with fragmentsof early Doric capitals on or near a long, narrow terrace which Dengate believes supporteda temple: Dengate, loc. cit. (footnote1 above).

THREE-PEAKED ANTEFIXES FROM THE ARGIVE HERAION

155

be proposed as a candidateto receive our antefixes. This is the so-called North Stoa, an impressivetwo-aisled porticothat runs along the north side of the later, Classical temple.12 Its early date has been questionedon occasion,13but a re-examinationof its remainsin situ suggeststhat it was probablythe next significantbuildingto be constructedafterthe temple. Although the associationof architecturalelementswith specificbuildings at the Heraion is difficult, I am convincedthat the earliest Doric capitals now lying in the stoa do in fact belong to it.14These capitals,if accepted,would date the stoa within the first half of the 6th century.15 In light of the probablechronologyof the plain peakedantefixesand the probabledates for the early temple and stoa, an associationof these buildings with the antefixes is quite possible. Given the fact that we have both a painted and an unpaintedseries of antefixes,it is tempting to suggest that the temple receivedthe former and the stoa the latter. With so few fragments,however, and with such an incomplete knowledge of the chronologyand developmentof the primitivehornedroof, we must, at least for the present,remain cautious in makingsuch associations.It is, after all, possiblethat the unpaintedantefix, with its more developed underside, is much later than the other examples and that it is a replacement piece made for the repair of an earlier hornedroof. In that case we should have to consider whether the other, painted antefixes belonged originally to only one building, either the temple or the stoa, or perhapsto both. In conclusion,it should be noted that despite the uncertaintiesthat remain with regard to the exact chronologyand attributionof the three-peakedantefixesfrom the Argive Heraion, the mere fact that we can now documenttheir existence in the sanctuary should improve considerablyour ability to assess the later, more developedroofingsystemsof the site, for it is now clear that the earliest antefixespreviouslypublishedneed not be associatedwith the earliest buildings. In addition, the evidence of a primitive horned roof or roofs at the Argive Heraion should help us better understandthe distributionof this type of roofing system in Greece and perhaps even focus upon a centerof production. 12 Tilton (footnote8 above), p. 112, pl. 12; Amandry (footnote8 above), pp. 226-235; J. J. Coulton, The Architectural Development of the Greek Stoa, Oxford 1976, pp. 27-29, 215. 13 B. Bergquist (The Archaic Greek Temenos, Lund 1967, pp. 21-22) dates the building to the late 5th

centuryB.C., making it contemporarywith the Classical Temple of Hera. 14 These capitals (two fragmentspreservingpart of the echinus and abacus)correspondto capitals B and M illustrated by Tilton (footnote8 above), p. 113, fig. 51. Whether these capitals are identicalwith those illustrated by Tilton (but now more poorly preserved)or whether they are others of the same series is not absolutely certain.As I hope to demonstrateelsewhere,Tilton's capital C, which now lies in the North Stoa and which is generallytaken to be one of the earliestattestedDoric capitals,does not belongto the original phase of the stoa;in fact, it is not Archaicat all. 15 The date of these capitals is no easier to determinethan the date of the peakedantefixes.The fact that the echinus of these capitals is appreciablybroaderand flatterthan the echinus of the capitalsof the early Aphaia temple might be taken as evidencethat they are earlier. If, however,the depth of the curvebelow the annulets is taken as a more decisivecriterion,as Schwandnerargues (1985, pp. 113-117), the Heraion capitals should be placed after those of the Aphaia temple. It is hard to imagine, however,that these capitals could date after ca. 560 B.C.

156

CHRISTOPHER A. PFAFF

CATALOGUE16 1. Three-peakedantefix Fig. 2:A, P1. 12:a, b Pres. H. 0.091; pres. L. 0.110; pres. W. 0.121; restoredW. 0.208; Th. of covertile 0.041 m. Fabric:pinkish buff clay (5YR 8/4) with many large reddish brown inclusions. Surfaces unslipped. Dark red (1OR 4/4) to brownish black (2.5YR 4/2) wash on front, top, and side. Preserved:left side of antefix and small portion of covertile. Peaks of antefix largely brokenaway. Pitched cover tile terminatingat lower end in plain, three-peaked antefix. Angular underside of antefix follows line of underside of cover tile. Front face slopes backward0.003 m. from bottomto top. Peaks of antefixcurvegraduallyintotop of covertile behind; their original heights cannotbe determinedprecisely. 2. Three-peakedantefix Fig. 2:B, P1. 12:a, b Pres. H. 0.082; pres. L. 0.110; pres. W. 0.115; restoredW. 0.206; Th. of covertile 0.034 m. Fabric: Pinkish buff clay (between 5YR 8/4 and 5YR 7/6) with many large reddishbrown inclusions. Surfaces unslipped. Dark red (between lOR 5/4 and lOR 4/6) to brown (2.5YR 5/4) wash on front, top, and side. Preserved:left side of antefix with small portion of covertile. Peaks of antefix largely brokenaway. Form similar to preceding. 3. Three-peakedantefix Fig. 2:C, P1. 12:a, b Pres. H. 0.071; pres. L. 0.103; pres. W. 0.103; Th. of covertile 0.048 m. Fabric: same as 2. Dark red (1OR 5/4 to IOR 4/6) to black wash. Preserved:most of the left side of antefix with small portion of cover tile. Left peak largely broken away. Form similar to 1 and 2.

4.

Cover tile Fig. 3 Pres. L. 0.158; pres. W. 0.099; Th. 0.0390.049 m. Fabric:buff clay (7.5YR 8/4) with many large reddish brown to dark gray inclusions. Surfaces unslipped. Brown (5YR 4/4) to blackwash on top and side. Preserved:small portionof left, rear part of tile. Pitched covertile with flat underside.Area of upper surfaceoriginallyoverlappedby next covertile in the series is 0.090 m. wide (front to back) and recessed 0.007-0.010 m. Back edge ratherrough.

5. Three-peakedantefix Fig. 4, P1. 12:c, d Pres. H. 0.081; pres. L. 0.213; pres. W. 0.147; restoredW. 0.186; Th. of covertile 0.027 m. Fabric:rathersoft clay with many large reddish brown inclusions, fired buff to yellowish buff (7.5YR 8/6 to 1OYR8/6) at surface,orangishtan (5YR 7/6) below. No slip or wash on surfaces. Front face of antefix perhaps self-slipped. Preserved:left side of antefix with left, forward portion of cover tile. Left peak of antefix slightly worn; centralpeak brokenaway. Pitched covertile with roundedundersidefrontedby three-peaked antefix with angular underside. The transitionfrom roundedundersideof covertile to angular undersideof antefix is effectedby a lip of clay projectingdownwardfrom bottomof covertile. Face of antefix slopes backwardca. 0.005 m. from bottom to top. Preserved lateral peak rather low and rounded;it merges gradually with top of cover tile ca. 0.055 m. from front of antefix. Central peak (now missing) originally merged with central ridge of covertile ca. 0.048 m. from front face.

CHRISTOPHER A. PFAFF AMERICAN

SCHOOL

OF CLASSICAL

STUDIES

54 Souidias Street GR-106

76 Athens, Greece

16 The antefixes cataloguedhere (1-3, 5) are now storedin the museum of Ancient Corinth;the covertile is still in the tile pile at the Argive Heraion. The colors of the fabrics and washes of the pieces are assignedthe notationsof the Munsell Soil ColorCharts,Baltimore 1975.

PLATE 12

a. 1 2 3, top view

b. 1 2 3, front view

d. 5, front view c. 5, top view

THE ASINE SIMA (PLATE

13)

HE SANCTUARY OF APOLLO PYTHAEUS at Asine in the Argolid is archaeologically documentedas existing alreadyin the third quarterof the 8th centuryB.C.1 In the early period the architecturalfocal point within the temenos (Fig. 1) on the Barbouna hill was a simple, apsidal structureon a narrow stone socle. This building, B, had walls of pise or mud brick, as indicatedby burnt lumps of clay retrievedduring excavation,and a thatched roof. Associatedpottery dates the destructionof the building to around 720 B.C. Building A was erectedslightly further east. Its early history is somewhat obscure,as pottery found in the trenchesby the walls only gives a terminus post quem for its erectionat the end of the 8th century. On the other hand, nothing contradictsits being an immediatesuccessor of Building B. If we suppose that A was constructedat the end of the 8th century, it must have been rebuilt a numberof times. In its later phase the socle, 0.8 m. wide (1.0 m. at the north end), carriedmud-brickwalls as shown by fragmentsof burnt clay.2 The excavators in the 1920's note in the diaries "large quantities of tile fragments" inside the structureand, underneaththis layer, "blackmatter",the charredremains of the woodworkof the roof. Further, by the outer wall of the northernshort end a deposit of tiles of differentkinds, both roof tiles and sima fragments,is reported.Fragmentsof the sima lay by the door at the south end. Of the apparently vast amounts of terracottapieces found during excavation, few now remain. Illustrations in the notebooks enable us to identify where some of the extant fragmentswere actually found (Fig. 1:T). Pieces of both Laconian and Corinthiantiles can still be found on the Barbounaterrace,but the questionis how they should be related to the excavatedstructureA. Roofs were not necessarilylaid consistently in one system or the other, or the Asine temple may have had an earlier, less pretentious, T

remains on the top terrace of the Barbouna hill were investigatedduring two brief seasons in the 1920's, in 1924 and 1926. In the publicationby 0. Fr6din and A. W. Persson (Asine, Results of the Swedish Excavations 1922-1930, Stockholm1938, pp. 147-151) however,the materialwas not given quite the attention it deserves. A complete study of the sanctuaryand the related fortificationson the Barbounahill is forthcomingas [Wells] Asine III, i in the Skrifterutgivna av SvenskaInstituteti Athen, series in 4?. In 1986 an initial presentation of the materialwas made at the A' Evv(8plo Apyo;LKc.V 7rov8Zv, held at Argos from May 30 to June ? EVVE8plOV AE\o7ToVV?pYLaKWV 1. By mistake, however, the paper has appeared in flpaKTlKa P ALE0voV'g ?7rrov82v, KaXaua,ra 1985, pp. 349-353. 2 A remark in a recent article by 0. Wikander (1988, p. 207) calls for a cautionary comment. I do not disagreewith him in that tiled roofs certainlypresupposesturdywalls to carrytheir weight, but wide soclesdo not necessarily imply tile roofs. Early Greek buildings often present wide socles but certainly did not have these roofs. The instances are many and can most convenientlybe studied in H. Drerup, GriechischeBaukunst in geometrischerZeit (ArchHomII, 0). On the other hand, if Building A was constructedin the late 6th centuryB.C. we could argue that the tile roof, then in existence, called for a wide socle. K. Fagerstr6m,in his dissertationGreekIron Age Architecture.DevelopmentsthroughChanging Times (SIMA 81), Goteborg 1988, p. 28, reaches the unfortunate conclusion that Building A was a stone-wall structure.The excavatedremainsbelie such a contention.It should also be noted, as is evidentfromthis paper, that the roof was gabled and not, as Fagerstr6msuggests, flat. 1 The

158

BERIT WELLS

(2; .

,;D

--

3.

-'~

M

1'~~~~~~~~~~~~~~~~~~~~~~~~~~~4q ;

. Adnotshet

Plan ofthetopterraceoftheBarbouna

FIG.

iD

e\

e

ee

X.t\\Xx:0e^e __

1'\

(+0~~~~~

\

IV+

,,>~~~~~~~~~~~

\

t

\

1S'=--ishEcvtion

192

N1~

90 Stokhl

O

/o

2,0rf

1938 fIg13

FIG. 1. Plan of the top terraceof the Barbounahill. A denotesthe 8th-centurytemple, the replacementof the

apsidal structureB, with which a large vase deposit, D, is connected.T representsthe findspotsof the sima pieces discussed.C is a terrace wall. After 0. Frodin and A. W. Persson, Asine, Results of the Swedish Excavations 1922-1930, Stockholm1938, fig. 130

THE ASINE SIMA

I'~

FIG.

~

~

~

~

~

159

'

2. As 5382 in the Asine Collectionat Uppsala. Drawing A. Grenberger

Laconiantile roof and a later, more elaborate,Corinthianone. On the other hand, the Laconian tiles may once have been associatedwith an ancillarybuilding of humble characteron the terrace.3In the Nauplion Museum storeroomsare kept a large piece of sima, 32.0 cm. long (P1. 13:a) and several pieces of tiles from the later roof, all found in 1924 by the north wall of the building. Amongthe materialfound in the 1926 excavationand now in the Asine collection at Uppsala are two pieces of sima, both catalogued as As 5382, found lying in the doorway at the south end of the building.4The larger of these, 16.2 cm. long (Fig. 2, P1. 13:b,c), is of crucial importancefor the reconstructionof the roof of the temple.5 3This was suggestedto me by Dr. Nancy A. Winter and couldvery well have been the case:a large portion of the terraceremains to be investigated,as is obviousfrom Figure 1. 4The Asine Collectionat Uppsala was createdin the 1930's from the materialdonatedto Swedish universities by the Greek government;see, e.g., Frodin and Persson (footnote 1 above), p. 12. 5 The identificationof the structure,based on Pausanias, II.36.5, as the temple of Apollo Pythaeus was made during the excavation (Fr6din and Persson [footnote1 above],p. 149). I see no reason to question the attribution.My arguments and the debate concerningthe issue are set forth in my forthcomingmonograph (footnote 1 above).

160

BERIT WELLS

The fabricof the sima pieces shows some variation,the Uppsala fragmentsbeing white (2.5YR 8/2) and the Nauplion one pale yellow (5Y 8/4), both with brick-red core and large, dark, brownish red grit and a white slip. A large tile fragmentin Nauplion differs in that its fabric has a greenish tinge and black grit. The absence of visible mica is of significance. On the fascia and the ovolo are opposing friezes of lotus and palmette, and the astragalis stylized into squares,surfacecolor alternatingwith blacksand reds, the colorsof the friezes. The Nauplion sima preservesthe completeheight of 18.0 cm. The two most important pieces of the sima here presented show features that reveal their positionson the gables. As we have noted,the Nauplion sima comesfromthe northern pediment, the flange indicating its placement on the slope at the spectator'sright when facing the building. Greaterinterest,however,attendsAs 5382, found lying by the doorway at the south end and thus fallen from the south pediment. At the back of this sima piece (P1. 13:b, c) are the remnantsof the attachedhorizontalbase into which the cornerakroterion was sunk and nailed into place. The nail hole pierces the base and continues slightly diagonallyinto the ovolo. As 5382 thereforealso reveals the slope of the roof as 250. However high the temple once was, its squatnessseems assuredfromthe inclinationof the roof in the mannerof the many small chapels scatteredover the Greek landscapetoday.6 Simas closelyresemblingthe Asine one have been foundin severalplaces, for instanceat Olympia, Athens, and Delphi.7 The ovolos and astragals of the Olympia simas are more roundedthan the Asine ones, while the ovolo of roof 45 from the tufa temple in the Marmaria at Delphi is a trifle flatter. The closest parallel to the Asine sima, however, is Nauplion Museum inv. no. 17282, which G. Hubner thought came from Tiryns.8The Tiryns piece is slightly less tall, 16.5 cm. high. Huibner'sdate of ca. 500 B.C. shouldalso be valid for the latest roof decorationof the Apollo temple on the Barbounahill. The Asine sima is not of local manufacture.Although clay beds may vary considerably, even within a limited area, Asine potteryis micaceouswhereasthe Apollo temple sima lacks mica. The fabric both of the sima and of the tile fragmentsdoes, however, conformvisibly with Argive productsof similar date, and it thereforeseems logical to seek the origin of the Asine sima in an Argive workshop.9In the early 5th centuryB.C. human activitywas again increasingin the Asine area,10which had been under Argive controlsince the end of the 8th century.11But we may be fairly certain that the cult of Pythian Apollo was still highly 6I

wish to acknowledgemy gratitudeto Dr. Nancy A. Winter for drawingmy attentionto the significance of the piece in the Uppsala Collection. I OlympiaII, pp. 194-195; Tafelband, 1896, pls. 118:4 and 119:2 and 4. A similar sima was also found on the Athenian Akropolis:TdA I, p. 16, fig. 16. Le Roy, 1967, pp. 101-102 and 219-221, pls. 38, 102, 109. 8 Huibner,1975, pp. 123-125 with fig. 4:b, pl. 66:2 and colorpl. 8:1. Note that the piece is storedat Tiryns. 9 This became obvious to me during the first excursion after the conference as we visited the Argos Museum. 10Tombs datable to Late Archaic and Early Classical times were excavatedat Asine in the early 1970's: B. Rafn, Asine, II, Results of the Excavations East of the Acropolis 1970-1974, vi, The Post-Geometric Periods, 1, The Gravesof the Early Fifth CenturyB.C. (Skrifterutgivna av SvenskaInstitutet i Athen, 40, 24:6:1), Stockholm1979. l Asine was laid waste by Argos around 720 B.C., but even though the area no longer prosperedor was densely populated, it certainlywas not deserted,as archaeologicaldiscoveriesin the 1920's led the excavators to believe:Frodin and Persson (footnote I above), p. 437.

THE ASINE SIMA

161

popular in the first half of the 5th century,or the two fragmentssurvivingof Bacchylides' paean to this god at Asine would make no sense.12 After that, rapid declinemay have set in. Except for a referencein Thucydides (v.53) to a conflictin the year 419 between Argos and Epildaurosover Epidaurianfailure to sacrificeto Apollo Pythaeus, our sourcesfor the cult, archaeologicalas well as literary,becomesilent for the remainderof the Classical period. BERIT WELLS G6TEBORGUNIVERSITY

Departmentof Classics Institute of Ancient Culture and Civilization Vastra Hamngatan 3 S-411 17 G6teborg, Sweden

I2

W. S. Barrett,"Bacchylides,Asine, and Apollo Pythaieus,' Hermes 82, 1954, pp. 421-444.

PLATE 13

a. Fragment of sima, Nauplion Museum

b. As 5382, Uppsala, Asine Collection. (Photograph0. Lindman)

bae (h

j,su

.~~0 .1~

~

~~

~EI

WELS THEA::E.IM

Lidmn

EIN NEUER ANTEFIXTYP AUS DEM HEILIGTUM DER ARTEMIS IN LOUSOI (PLATE

14)

N IHREM SORGFALTIGEN BERICHT, erschienenin den OsterreichischenJahresheften,machen die beiden ersten Ausgraberin Lousoi, W. Reichel und A. Wilhelm, die Ergebnisseihrer drei Kampagnenin den Jahren 1898 und7l899 bekannt.IDiese drei Ausgrabungendes damals gerade gegruindetenOsterreichischenArchaologischenInstitutes in Athen waren die einzigen systematischenUntersuchungenim Gebiet neben einer langen Tradition von Raubgrabungenseitens der Bev6lkerung,und die beiden Osterreicherbeklagten schon damals den durchwuihltenZustand des Erdreiches. Ihre Untersuchungen beschranktensich auf den Tempelbereich, der auf einer Hiigelkuppe liegt, sowie weitere Bauten des Heiligtums am Fusse des Hiigels. Ihre Funde befinden sich im Athener Nationalmuseum.Gleichzeitig gelangten zahlreiche Kleinfundeaus Lousoi uiberden Handel in verschiedeneMuseen im Ausland, die unsere Kenntnis vom Kult der Artemis Hemera

I

A

bereichern.2

Reichel und Wilhelm datiertendas Tempelfundamentmit dem eigenartigenGrundriss in das "dritte,frihestens das vierte Jahrhundert"und setzten aufgrundvon Einzelfunden die Existenz eines Vorgangerbauesaus dem sechstenJahrhundertvoraus.3Weiters bilden sie auf drei Seiten4 14 Typen von Dachterrakotten-Antefixen und Simen-ab, die sich zeitlich in zwei Gruppen teilen lassen: a) diejenigenaus der Zeit vor dem hellenistischenTempel und b) solche, die gleichzeitigund auch spater sind. Seit der Wiederaufnahmeder Arbeitendurchdes OsterreichischeArchaologischeInstitut Athen im Jahre 19815 trugen wir die Ruine des spaten Kirchleins,in dem viele antike Bauteile verbaut waren, bis auf einen ungefahr einen Meter hqhen Mauersockel ab und konnten dabei 200 bearbeitete Architekturstiickebergen; das empelfundamentwurde wieder gereinigt und die Arbeitenim Inneren sowie in der Umgebung des Tempels weitergefuihrt.Obwohl die Befundegestortsind, durchRaubgrabungenund durcheinen Friedhof aus der Tuirkenzeit,konnten doch fur einige der Dachziegeltypengenauere Hinweise gewonnen werden, sodass sich in einigen Fallen ein neuer oder genauerer Zeitansatz vorschlagenlasst. Dazu tragt auch der Fortschrittin der ErforschungantikerDachterrakotten seit damals wesentlich bei. 1

W. Reichel und A. Wilhelm, "Das Heiligtum der Artemis zu Lusoi,"OJh 4,1901, S. 1-88. Ich danke Dr. Nancy Winter fur die Moglichkeit, den Neufund aus Lousoi im Rahmen der Konferenz uber ArchaischeDachterrakottenvorstellenzu konnen. 2 Vgl. U. Sinn, "Ein Fundkomplexaus dem Artemis-Heiligtumvon Lusoi im BadischenLandesmuseum," Jahrbuchder StaatlichenKunstsammlungenin Baden-Wiurttemberg 17, 1980, S. 25-40. ' Reichel und Wilhelm, a.a.O., S. 32. 4 Ebd., S. 61-63. ' Zuletzt V. Mitsopoulos-Leonund F. Glaser, "Lousoi1985/86," OJh 57, 1986, Grabungen,S. 17-22.

164

VERONIKA MITSOPOULOS-LEON

FIG.

1. Lousoi, Heiligtum der Artemis, Antefix Tka 116/1987, Skizze

Heute interessiertvor allem, dass den bereits bekanntenTypen aus Lousoi ein Neufund hinzugefugt werden kann. Es handelt sich um ein Antefix vom "lakonischen"Typ (P1. 14), 0.185 m hoch, 0.15 m breit erhalten,mit einem Absatz an der Unterkanteund dem Ansatz des Ziegels an der Ruckseite.6Bis jetzt liegen drei Bruchstuicke(von mindestens zwei Exemplaren)vor, die sich etwa zu folgendemSchemaerganzen lassen (der Mittelteil fehlt, weshalb die Anzahl der Palmettenblattersowie die Mittelkonstruktionnoch unklar sind), s. Skizze (Fig. 1). Der Dekor besteht aus zwei schragliegenden,Rucken an Rucken gesetzten Doppelvoluten. Ein doppeltes Band zwischen den beiden oberen Voluten verbindetdie (9- oder 11blattrige) Palmette an ihrem unteren Ende. Die Hohe des Mittelblattes ist nicht gesichert. Die untere Volute greift weiter aus; sie ist doppelt und setzt sich in einem Feld zur Mitte hin fort. Auf dem Ziegelansatz ist schwarze Farbe erhalten, schwarz auch auf dem Reliefgrund;auf den Voluten rote Farbe. Das Ornament mit den aneinandergesetztenDoppelvoluten ist offensichtlichselten.7 G. Hubner publiziert ein in mancher Hinsicht verwandtes Beispiel vom "korinthischen" Typ aus Tiryns und datiert es in das erste Viertel des fuinftenJahrhunderts.8Sie verweist 6 Tka 116/1987, aus Flache 0 6 Mitte, braune Erde, Steine, knapp uber dem gewachsenen Boden; H. 0.185 m; r. Halfte erhalten;Ansatz des Ziegels; Absatz an Unterkante;Ton beige, viele bunte Einsprenkelungen;schwarze Farbe auf dem Ziegel; auf der Vorderseiteschwarze und weinrote Farbe. 7 ATK, S. 59; Hubner, 1975, S. 126-127, Abb. 5:a, b; A. Vlachopoulou-Oikonomou,HyuLAOVCES Kal Kopv4 alEs/Kepa,uoL Mue8LaK6o-/L?)o-? a7ro r?lv H7reLpo,Ioannina 1986, S. 50, Anm. 25. 8

Hiibner, 1975, S. 126-127.

EIN NEUER ANTEFIXTYP AUS DEM HEILIGTUM DER ARTEMIS IN LOUSOI

165

auf eine "Ableitung"aus Kerkyra,vielleicht vom Kardakitempelin Mon Repos.9Um 525 ist das sehr kraftvolleVolutenantefixaus Chios datiert,10bei dem die Palmettenblatterorganisch in den Zwickeln zwischen den Voluten sitzen, welche gross und schwungvoll das Feld ftillen. Als "archaisierend" wird ein "lakonisches"Beispiel aus Elateia in Phokis11 angesehen, dessen beide Voluten im oberen Abschnitt durch das Band zusammengehalten sind. Es erinnert an ein Beispiel von der Athener Akropolis.12Ein "korinthischer"Antefixtyp aus Kassope in Epirus13 mit gestrecktemDekor und steileren Voluten, wird in die zweite Halfte des fiinften oder das frtihevierte Jh. datiert. Dem Umriss und der allgemeinen Anordnungnach, steht das Lousoi-Exemplar dem Typus von der Akropolis und aus Elateia/Phokis am nachsten;in einzelnen Details, wie der Lage der Voluten und dem Band, welches die Palmette am unteren Ende zusammenhalt, erinnertes an das Beispiel aus Tiryns. Das unterste Blatt der Palmette folgt elastisch dem Schwung der Volute wie auf dem Beispiel fortgeschrittenemDatums aus Kassope. Das ungegliederte Feld unter und zwischen den Doppelvoluten k6nnte als missverstanden angesehenwerden, besondersim Vergleich mit dem Tiryns-Dekor. Die Tatsache, dass die Palmette gegentiber der Volute klein gehalten ist (an sich ein fruihesKennzeichent4)wirkt wie ein Bemuhen um Altertiimlichkeit. Schliesslichemnweiteres Detail: der Dekor des Lousoi-Antefixesfillt nicht den gesamten Hintergrundder Scheibeaus, wie bei einem "archaisierenden" Antefix-allerdings mit anderemDekor-aus Delphi.15 Wichtig ist der Fundplatz der Exemplare: Vor der NO-Ecke des jungeren Tempels stiessenwir auf eine etwa zwei Meter tiefe Grube, die in den lehmigen gewachsenenBoden vertieftist. Dem Bau des hellenistischenTempels, den wir jetzt durch Funde aus der Fundamentgrube,besonderseine Munze und eine Lampe, um die Mitte des dritten Jhs. ansetzen m6chten,16gingen offensichtlichgrbssere Planierungsarbeitenvoraus. Falls es auf der Hiugelkuppeeinen Vorgangerbaugegeben hatte, fiel er diesen Arbeiten zum Opfer. Vielleicht diente auch das lehmige Material der Grube der Produktionvon Lehmziegeln fur die Wande des neuen Tempels. Anschliessendwurde die grosse Offnung neben dem Tempel zugeschuittetmit Steinen, Erde und Votivgaben;diese Votive stammen,soweit wir bis jetzt iiberblicken,aus der Fruihzeitdes Heiligtums: aus spatgeometrischer,archaischer, und klassischerZeit. Es sind teilweise bereits bekannteTypen; dazu kommeneinige Neufunde, z.B. Nadeln, Fibeln, Tonfiguirchen,weiters auch einzelne Gefasse. In diesem Rahmen findet das Antefix seinen Platz. 9 GFR,

S. 158, Nr. 67, Abb. 67.

10ATK, S. 35-36, Abb. 8.

11Koch, S. 74-75, Abb. 32. 12 TdA II, S. 37, Abb. 49. 13Vlachopoulou-Oikonomou,a.a.O., S. 48, Reihe Klk, Taf. 3:b, Zeichn. 8. 14 Ebd., S. 17. 15 Vgl. Le Roy, 1967, S. 90, Taf. 31, Serie 36 alle. 16 V. Mitsopoulos-Leon und F. Glaser, "Lousoi 1987" (im Druck).

166

VERONIKA MITSOPOULOS-LEON

Die Grube ist noch nicht vollkommenausgegraben,es bestehtsomit die Hoffnung, dass ein weiteres Fragment zutage kommt, welches zur Erganzung des Dekors, vor allem des Mittelteils, beitragt. VERONIKA MITSOPOULOS-LEON OSTERREICHISCHES ARCHAOLOGISCHES INSTITUT

26 Alexandras Boulevard GR-106 83 Athens, Greece

PLATE 14

__~~~~~~~~~~~m

r.~~~~~~~~~I

-..,

Lousoi, Antefix Tka 116/1987 aus dem Heiligtum der Artemis

V. MITSOPOULOS-LEON:EIN NEUER ANTEFIXTYPAUS DEM HEILIGTUMDER ARTEMISIN Lousoi

DIE DACHTERRAKOTTEN DER ARCHAISCHEN TEMPEL VON KALAPODI (PHOKIS) (PLATES

15, 16)

DIEDACHTERRAKOTTEN,die im Folgendenin der Formeiner summarischen Ubersicht vorgefuhrtwerden, stammen vom sog. N- und S-Tempel des Heiligtums beim Dorf Kalapodi in der Phokis.1Die Existenz von Altertumernin dem hugeligen Gelande 6stlich des Dorfes Kalapodi,linker Hand der modernenAsphaltstrassenach Atalanti, war in der Vergangenheit bereits von Reisenden bezeugt worden.2 Seit 1973 wurde der Platz unter der Leitung von R. C. S. Felsch systematischuntersucht.Geschichteund Vorgeschichteder Grabungsind seitdemder Wissenschaftkontinuierlichzuganglichgemacht.3 Die Identifizierungdes Heiligtums ist inzwischen hinlanglichgesichert.4 Die Autorin hatte die Bearbeitung der Dachterrakotten 1976 ubernommen.5Auf Grund des bis dahin zu Tage gekommenenDachterrakottenmaterialswar zunachst lediglich eine katalogartigeVorlage geplant gewesen. Die Funde der Folgezeit erwiesenjedoch, neben den beiden archaischenDachern, drei TondacherklassischerZeit (P1. 15:a), Ziegel lakonischenTyps unterschiedlicherZeitstufen und eine Reihe nicht uninteressanterVaria aus verschiedenenEpochen.6Dass in dem derart reichen Bestand die archaischenDachterrakottendes sog. korinthischenSystems quantitativ uberwiegen, beruht auf dem Umstand, dass die zugehorige Architektureiner Brandkatastrophezum Opfer gefallen war.7 Ein Grossteil der entsprechendenZiegel fand sich entwederin Sturzlage8oder in einer Art Planierungspackung,mit der, im Zug der Aufraumarbeitendanach, die nahere Tempelumgebung unterfuttertwurde.9Auch fur den Bau des provisorischenKultraumeswaren I

Fur vielfaltige praktische Hilfe und Diskussionsbereitschaftbin ich R. C. S. Felsch bei der Durchfuihrungder Bestandsaufnahmeverpflichtet.Da die Arbeitenmeist nicht wahrend der laufenden Kampagne hatten erfolgen k6nnen, gebuihrtzusatzlich dem damaligen Wachter in Kalapodi, N. Papageorgiou,besonderer Dank. Daruberhinaus danke ich dem Restaurator Chr. Vaporakis und dem DAI-Photographen A. Tzimas. Auf den Letzterengehen die Photovorlagender Tafeln zuruck. Abkiirzungsverzeichnis: Felsch und Kienast = R. C. S. Felsch und H. J. Kienast, "Ein Heiligtum in Phokis,"AAA 8,1975, S. 1-24 = G. Huibner,"Die Dachterrakotten,"AA (JdI 95) 1980, S. 112-115 Hiubner,1980 2 Felsch und Kienast, S. 2-4, Anm. 2-4; Felsch and Schuler, 1980, S. 111-112. Felsch, 1987, S. 1-2, Anm. 1-5. 4 R. C. S. Felsch und P. Siewert, "Inschriftenaus dem Heiligtum von Hyampolis bei Kalapodi,"AA (JdI 102) 1987 (S. 681-687), S. 686-687. 5 Felsch and Schuler, S. 38, Anm. 6 R. C. S. Felsch, "BoiotischeZiegelwerkstattenarchaischerZeit," AM 94, 1979 (S. 1-40), S. 26-29; Felsch and Schuler, S. 77-78; Huibner,1980, S. 112-115. 7 Felsch and Schuler, S. 84. 8 Felsch and Schuler, S. 77. 9 Felsch and Schuler, S. 67, Abb. 41.

GERHILD HUBNER

168

Dachziegel verwendetworden.10Fur die Erbauungder beiden archaischenTempel, denen jeweils Vorgangerbautenmit Schilfdachvorausgegangenwaren,11 ist durchdie Zerstorung der letzterendas 2. Viertel des 6. Jhs. v. Chr. als terminus post quem gesichert.12 Die fiir die Dachziegel selbst aufschlussreichstenFundkomplexestellten die Grabungsareale 6stlichund siidlichbeiderTempel, d. h. also auch in einem Bereich,der, zwischen Nund S-Tempel gelegen, eigentlich Dachversturzbeider Gebaudeenthalt.13Und tatsachlich lIsst sich das Ziegelmaterialmehr oder weniger eindeutigbeiden Tempeln zuteilen, wobei die dem N-Tempel zugewiesenen Dachelementeein weniger vollstandigesBild ergebenals die nahezu komplett rekonstruierbareEindeckungdes S-Tempels (P1. 15:b). Auffallig ist dabei, dass technischeund ornamentaleQuerverbindungendie Produktionbeider Dacher vordergruindigzeitlich eng miteinanderverknupfen.Davon hebt sich, bei gleicher Zurichtung des Tons, die formale Andersartigkeitder N-Tempelsima ab (Fig. 1). Sie ist,-ohne dass sich direkte Vergleiche im allgemeinen Denkmalerbestanderhalten haben,14-wohl einerjungeren Phase zuzurechnen.U. U. gehortesie zu einem Modernisierungsprogramm des N-Tempelgiebels, das bei der Zerstorungdes Heiligtums vielleichtnoch garnichtabgeschlossenwar; scheint es doch bis zu einem gewissen Grad verwunderlich,dass von diesem Tempel keinerlei Firstfragmente oder Traufeckstucke, geschweige denn Uberreste des Akroters, feststellbarwaren, auch wenn man dafur den Zufall resp. die durch die Ausgrabung bedingte selektive Unvollstandigkeit der Funde verantwortlichmachen wollte. Dass der obereAufbau des N-Tempels in konstruktiverHinsicht dem des S-Tempels nicht entsprochenhaben kann, zeigt zumindestauch das Fehlen all solcherBaugliederaus Lehm oder Ton, wie sie mit der Decke in der Peristasis des S-Tempels hatten in Verbindung gebrachtwerden muissen.15 Die normalenundekoriertenZiegel beider Dacher stimmenjedoch in ihrenjeweiligen Dimensionen, in der Konsistenzdes Tons, seiner Engobierungund Magerung uberein. Es handelt sich um flach aufliegende, auf der Oberseite zu den Seitenkantenhin leicht aufgebogene Platten, deren dem Dachrand zugewandte Schmalseite zu einer vorkragenden Zunge zugerichtet ist, unter die der jeweils daruntergelegtenachste Ziegel hineingeschoben worden war (P1. 16:b). Zu diesen sog. Stroteres gehorten separat gearbeitete Reiter (P1. 16:a). An allen Beispielen lassen sich sowohl die Spuren des Herstellungsverfahrens selbst als auch die nachtraglicheAbarbeitungvor dem eigentlichenVersetzen ablesen. Da praktisch also jeder Ziegel individuell am Bau eine Art "final touch" erhielt, existieren uberall Massunterschiede.Die Verjungungnach hinten, die etwa 1-3 Zentimeterbetragt, 10

Felsch and Schuler, S. 85, Abb. 45, 68, 70. 11Felsch, 1987, S. 14-15. 12 Felsch, 1987, S. 17-19. 13

Vgl. Felsch,1987,Abb.3.

Es lasst sich nur an Simen erinnern,die in der mehr oderweniger willkuirlichenKompilationvon Einzelteilen eine offenbar ahnliche Freiheit der Erfindungfuirsich in Anspruchnehmen, vgl. die Sima Nr. 17283; 17286 (Huibner,1975 [S. 117-136], S. 121-122, Abb. 3:b, Taf. 65:1-2) oder Nr. G 6 von Dach 27 (Le Roy, 1967, S. 66-67, Taf. 19:7, 100). Die Datierung in die 80iger Jahre des 6. Jhs.: ebd., S. 69. 15 G. Hubner, "Dach und Decke,"AA (Jdl 102) 1987, S. 76-82; vgl. Felsch, 1987, S. 24, Anm. 50. 14

DIE DACHTERRAKOTTEN DER ARCHAISCHEN TEMPEL VON KALAPODI

0.0

FIG.

0.05

169

m

1. Profil der vermutlichostlichen Giebelsimamit zugeh6rigemGeisonziegelvom archaischen N-Tempel, Kalapodi

ist jedoch sowohl bei den Flach- als auch bei den Deckziegeln vor dem Brand durch den Formrahmenfestgelegtworden. In abgerundetenmodernen Zahlen ausgedrtickt,liesse sich fur die Stroteres ein allgemeines Durchschnittsmassvon 0.58 m Breite x 0.77 m Tiefe nennen, d. h. ein Proportionsverhaltnisvon 3:4. Dem entsprachein Kalyptervon 0.19 m Breite x 0.77 m Tiefe, also ein Proportionsverhaltnisvon 1:4. Fur das antike Fussmass mit Daktyleneinteilung, das dem Systemim Prinzip zugrundelag,hat sich moglicherweisesogardie skizzenhafteAnlage auf einem Firstziegel mit Ritzlienen erhalten. Ein bezeichnendesMass bildete offenbardie Halfte der Kalypterbreite,begegnet doch die Einheit von 8.5-12 Zentimeter bei der Gestaltung des Traufrandeshaufiger. Die hier genannten Zahlen durfen im Moment nur als Hilfswerte betrachtetwerden, um dem Leser eine genauere Vorstellung vom Objekt selbst zu vermitteln. Sowohl der

170

GERHILD HUBNER

Brandschwundals auch die zusatzliche Einpassung am Dach machen es vorlaufig noch schwer, das antike Grundmasseindeutigerzu postulieren.Dies bleibt der Endpublikation vorbehalten. Vom antiken Farbeindruckhat sich der heutige Zustand,-ein ins Braun spielender Beigeton mit creme-gelblicherOberflache,-wohl wenig unterschieden,da ein Schub das Gros der Ziegelmassenzur Erde hatte sttirzenlassen, langst bevor der holzerne Dachstuhl in Flammen stand. Jene Ziegel, die tatsachlich intensiver mit dem Feuer in Beruhrung gekommen waren, zeigen Verfarbungen ins Grau-Grune, andersartigeBruchstellen und Oberflachenverbiegungen.Starker unter Feuereinwirkunggelitten hat die Ziegelreihe an den Traufrandern, die in einem schmutzigen Grau resp. Dunkelrot erscheint. Die unten behandeltenGeisonziegel (P1. 16:d)wirken danebenverhaltnismassigunbeeintrachtigtund jedenfalls heller. Grossomodo kann man anhand der Ausdehnungdes N-Tempels16eine Quantitat von ca. 1 000 Flachziegeln annehmen, mit ungefahr ebensovielenKalypteren. Beim kleineren S-Tempel17kommenungefahretwa 650 Stuickpaarein Frage. Beide Tempel besassenGeisonziegel (P1. 16:d). Es handelt sich beim S-Tempel ausschliesslich um ebenfalls leicht trapezformige Platten von 3 Zentimeter Dicke und einer Ausdehnung von 0.74-0.77 m Breite x 0.46-0.476 m Tiefe. Die gr6ssere Langseite verdicktsich zu einer vorgezogenen "Nase"von 6.6-7.8 ZentimeterHohe. Deren Unterseitetragteinen rot angelegtendorischen Blattstab von ca. 12 Zentimeter H6he, der 8 Elemente auf einem Farbstreifendstehend zeigt. Die durch Ritzlinien vor der Bemalung markiertekursorischeEinteilung wurde nie eingehalten. Die Stirnseite der Platte ist einheitlich rot eingefarbt. Die Geisonziegel des N-Tempel-Giebels zeigen eine modifizierte,bislang unbekannteForm (vgl. Fig. 1). Uber den Geisonziegelnsassen die eigentlichenTraufziegel (Hegemones). Sie existieren in zwei Ausfuhrungen. Das grossere Format von 0.58 m Breite x 0.77 m Tiefe geh6rte zum N-Tempel, das kleinere mit 0.58 m Breite x 0.64 m Tiefe zum S-Tempel (vgl. P1. 15:c). Da sie ebenfalls-bis zu 2 Zentimeter-Verj"ungungennach hinten aufweisen, variieren die genannten Werte jeweils von Fall zu Fall. Die rote Frontleiste greift bei den Traufziegeln des S-Tempels leicht auf die Unterseite uber. Verhaltnismassighaufig treten sonst auf der Unterseite aber auch Abarbeitungenauf. Die formale Gestaltung des Umrisses hebt sozusagen die Monotonie des Farbbandesan der Traufe auf: er alterniert in einem wellenartigen Hoch und Tief, das durchplastischeMittelspitzen zustandekommt,18 die zwischen die hochgezogenen "Horner"der Seitenkanten gesetzt wurden. Die so erreichtenHohen schwankenzwischen 5.6-6.6 Zentimeteran der Seite und ?8.5 Zentimeter in der Mitte. Auch andernortsin Griechenlandbegegnen ahnliche ornamentale Grundmuster des Dachranddekors.19 Die Fugen der aneinanderstossendenTraufziegel deckt das plastisch gebildeteAntefix ab, das gleichzeitig den vorderstenKalypter an der Front schliesst (P1. 15:c). Die Breite Felsch, 1987, S. 21. Felsch, 1987, S. 22. 18 Felsch und Kienast, S. 21 mit Abb. 25; BCH 99, 1975, S. 638, Abb. 101; BCH 104, 1980, S. 627. 19 Zuletzt Schwandner,1985, S. 77, Abb. 48; vgl. ebd., S. 127; N. Cooper, 1983, S. 62-65, 70-71. 16 17

DIE DACHTERRAKOTTEN DER ARCHAISCHEN TEMPEL VON KALAPODI

171

betragtetwas mehr als 0.19 m, die Lange entsprichtdemjeweiligen Hegemon. Das Schema der artikuliertenMittel- und Seitenbetonungen,das den Traufziegel ausgezeichnethatte, wird hier durchdie Umrissftihrungdes Ornamenteswiederholt. Das mehr oder weniger in flachem Relief gegebene Motiv erhalt seine Wirkung durch den Gegensatz von Rot und gelblicherTongrundierung.Vom Aussehen her lassen sich zwei Versionen scheiden. Das niedrigeFormatvon etwa 12 ZentimeterHohe stellt dabei lediglichdie dezimierteVariante des hoheren Formates von 0.194 m Hohe dar, das dem N-Tempel zuzurechnenist (S. 21 oben, Fig. 7),20 wahrend das erstere zum S-Tempel gehorte. Der beiden OrnamentenzugrundeliegendePrototyp,der auch den eigentlichenFormmodelnals Vorlage gedienthaben muss, wurde zur Herstellung des S-Antefixesuber der Knospe gekappt. So verschwanddie Palmette, wahrend der tibrige Aufbau unverandertubernommenwerden konnte. Sie entfaltet sich uber zwei Aussenvoluten,die ein Mittelsteg halt. Die hangendegespreizteLotosbltite darunter offnet sich uiberder dreieckigenAussparung der Antefixplatte, um damit uber den beiden zusammenstossendenSpitzen der Traufziegel aufliegen zu konnen. Im Bestand des tibrigen gebranntenArchitekturschmuckesGriechenlandsfehlen fur das Ornament des Antefixes bislang Parallelen. Die typologischeAuflosung in a) formal tektonische Elemente (Umriss, Proportionen,etc.) und b) formal gestalterischeElemente (hangenderLotos, nach aussen eingedrehteVoluten, deren Strangeals Rahmenleistenfungieren) ergibt eine mehrschichtigeAbleitung. Cum grano salis liegt der Umrissfuihrungein Schemazugrunde,das fur die fruhen Ausbildungenvon Antefixen in Anspruchgenommen wird,21aber wohl auch noch im 2. Viertel des 6. Jhs. v. Chr. durchausin Gebrauchwar.22 Jedoch wird dieses einfacheGrundschemavon dekorativenKomponententiberlagert,die in ein Dreieck-Rahmensystemeingespanntsind, das so sonst im Mutterland nicht vorkommt. Die ausgedtinnte,hangende Blute, die von einem den Konturdes Antefixes nachzeichnenden Steg begleitetwird, begegnetzum ersten Mal bei archaischenAntefixlosungenin Didyma.23Umgekehrt erinnert die gedrungene,niedrige Form des S-Tempel-Antefixes,-ohne im Detail vergleichbarzu sein,-an eine spezifische Modifikation des Festlandes.24Indirekt suggeriertdas Palmettenantefixdes N-Tempels einen noch alteren Eindruck,k6nnte hier doch die Uberbetonungdes PalmettenfachersErinnerungenan den sog. Kamirosstil der Malerei wachrufen.25Das Gesagte verdeutlichtjedenfalls die Ambivalenz moglicher Beurteilungskriterien,die erst dann zu konkretenErgebnissenftihren, wenn sie sich untereinander kontrollieren oder relativieren bzw. wenn zusatzliche Indizien hinzukommen. Jedenfalls passt die Ornamentikder Kalapodiantefixenicht in die so schluissigvon N. K. CooperaufgestelltePhasenabfolge.26Geradebeim N-Tempelantefix aber weisen die 20

Hiibner, 1980, S. 113, Abb. 102: in der Abbildungslegendefalschlichals Firstpalmettebezeichnet. N. Cooper, 1983, S. 55-56, 62-65. Schwandner,1985, S. 126-129: nicht nach 570 v. Chr. 23 ATK, S. 108-1 10, Taf. 56:4. 24 Zur festlandischenVersion, vgl. das argivische Antefix, Huibner, 1975, S. 121, Anm. 19, Taf. 68:6; N. Cooper, 1983, S. 59-60; Schwandner,1985, S. 128 mit Anm. 247. 25 Vgl. W. Schiering, WerkstattenorientalisierenderKeramikauf Rhodos,Berlin 1957, Beil. 6. 26 N. Cooper, 1983, S. 61, 68, Taf. 20: Rahmung, Gleichwertigkeitder ornamentalenTeile kommenoffenbar in der Kategorie"unattached"nicht vor, bestimmenaberwesentlichden Habitus des Kalapodi-Antefixes. 21 22

172

GERHILD HUBNER

Ausgewogenheitdes Aufbaus (d. h. Palmettenhoheund untere Bildhalfteentsprechensich) und die feinlinige, geschlosseneVerflechtungdes Ornamentesselbst bereits auf Tendenzen hin, die in der VasenmalereidurchAmasisvertretensind,27auch wenn die direktenmorphologischen Entsprechungeneher gering sind. Postuliert man mit dem Didyma-Antefix eine gewisse Sicherheitder Datierung, so sollten die Kalapodi-Antefixejedenfallsjuingersein. Die ostliche Giebelsima des S-Tempels (Fig. 2; P1. 16:c) sass, ebenso wie die gerade beschriebenenTraufstucke,auf den oben genannten Geisonziegelnauf. Geritzte Nummerierung auf der Unterseite legte die Reihenfolge im Einzelnen fest, wobei offenbar die Zahlung vom unteren Dachrand nach oben erfolgte und mit einer geringfugigenBreitendifferenz pro Simenglied im oberen Bereich korrespondierte.Anscheinend sollte die Stimmigkeit untereinander,aber auch der Ubergang zur dahinterliegendenStroterreihe hergestelltwerden. Bei einer Tiefe von ?0.58 m betrug die jeweilige Simenfrontdann eine, z. T. auch rechnerischerschliessbareBreite von 0.668-0.754 m. Dazu kamjeweils ein Falz von 4.5 Zentimeter an der der Traufe zugewandten Schmalseite. Die H6he des Simenwulstes belauft sich an der Front auf ca. 10 Zentimeter.Ihn gliedertvorn ein ausgesparter, 10teiliger dorischerBlattstab,der auf dem doppelt so tiefen Vorsprungder Unterseite wiederholtwird. Eigens zugeschnitteneEckstuckeverbandenSima und Traufe. Die fur den First gefundeneLosung (vgl. P1. 15:b), die auf gesichertenEinzelteilen beruht, ist innerhalbdes Dachterrakottenbestandesohne direktvergleichbareParallelen. Die jeweils in der Dachmitte zusammenstossendenund auseinanderklaffendenStrotere uiberdeckte ein leicht gew6lbter Deckziegel von 0.58 m Lange in horizontaler Richtung. Die normalerweisehochgezogenenSeitenstegeder Stroteredarunterwaren zu diesem Behufe extra abgearbeitetworden. Vertikal dartiber, in der Verlangerung der Kalyptersequenz, wurde ein gekruimmterDeckziegel gestuilpt,der eine Streckevon etwa 0.25 m tiberspannt.28 Wieweit dasselbeSystem auf den N-Tempel zutraf, ist nicht mehr uberprufbar. Auf der Spitze des S-Tempelgiebels befand sich am First, frei modelliert,eine Sphinx, von der mehrere Fragmente erhalten sind.29Sie erweckt unter den Dachterrakottenvon Kalapodi zweifellos den alterttimlichstenEindruck. Man konnte sich fast tatsachlich fragen, ob hier nicht uiberhaupteine bereits fur den fruharchaischenVorgangerbaukonzipierte (und dann wieder verwendete)Figur bewahrt blieb. Das stilistischjuingsteElement innerhalb des bisher vorgetragenenKomplexes bildet die elaborierteAusgestaltungdes N-Tempelantefixes. Die gedrungeneSima des S-Tempels (Fig. 2) fur wesentlich alter zu halten, besteht wegen der Geschlossenheitdes Gesamtbefundes kein hinreichenderGrund.30Dagegen m6chte man die hohe, sehr viel schlanker 27 Vgl. z. B. die Henkelpalmetten, D. von Bothmer, The Amasis Painter, London/Malibu/New-York 1985, S. 43, Abb. 32; S. 114, Abb. 19; S. 137, Abb. 25. 28 Nach freundlicherAuskunftvon A. Ohnesorgbefinditsich ein ahnlicherFirstkalypterin den Magazinen des Heraions von Samos. 29 Vgl. Hubner, 1980, S. 114, Abb. 103; BCH 99, 1975, S. 638, Abb. 102 = AAA 1975, S. 21 mit Abb. 30; M.-F. Billot, "Recherchessur le sphinx du Louvre CA 637," BCH 101, 1977 (S. 383-42 1), S. 387. 30BCH 99, 1975, S. 638, Abb. 100; Felsch und Kienast, S. 21, Abb. 26, 27; vgl. die (altere) Sima Nr. S 3 von Dach 6 in Delphi: Le Roy, 1967, S. 31, Taf. 5:1, 98, 118.

173 E 0

0.58

0

1o

m

9 m

0-0

l

0-

52 NK

7 3 m

4

4

tH___~~~~057

FIG 2. Un_anih

FIG.

2.

Unteransicht

un PfIldrsthnGbeimdearashnSTmpl,Kapi

und Profi der desarchaischn-S-Tempes,Kalapo

Giebls'ma

4i

174

GERHILD HUBNER

proportionierteSima von geringererTiefe (?0.48 m), die im westlichen Bereich des Tempelgelandeszu Tage kam, als juingeransehen.3IDie Fundumstandemachen eine Zuweisung sowohl an den N- wie an den S-Tempel m6glich. Ob ein bereits bekanntgemachtes Simenfragment,das eine modifizierteWiederholungdes Ornamentesund eine profilierte Umrissgestaltungzeigt,32ein Ersatzstuickder schlankenSima darstellt,oder gar eine selbstandige (noch juingere)Ausbildung verk6pert, die zu einem bislang unbekannten Bau geh6rt haben k6nnte,bleibe dahingestellt.Jedenfalls Iasstsich die schlankeSima chronologisch mit jenen eingangs erwahntenFragmentenzusammensehen,die ftir den O-Giebel des N-Tempels in Anspruchgenommenwurden. In typologischerHinsicht haben beide nichts miteinanderzu tun. Eine intensive Er6rterungder daran anschliessendenFragestellungen geh6rtjedoch mit Fug und Recht in die Endpublikation.In jedem Fall ergibt das Gesagte fur beide Tempeldacher einem grundsatzlichenzeitlichen Rahmen, der die Jahre 545 v. Chr. bis 520 v. Chr. umfasst. GERHILDHUJBNER ARCHAOLOGISCHES INSTITUT DER UNIVERSITAT,

Am Hofgarten 21 D-53 Bonn 1

Federal Republic of Germany 3' Vgl. Hibner, 1980, S. 113, Abb. 101. Felsch und Kienast, S. 21, Abb. 28, 29.

32

BONN

PLATE 15

a. Traufrand des 1. klassischenTempels. (DAI Athen, Neg. 86/510)

b. Dach des archaischenS-Tempels. (DAI Athen, Neg. 86/477)

c. Traufrand des archaischenS-Tempels. (DATAthen, Neg. 86/528)

-A

a. ArchaischerKalypter. (DAI Athen, Neg. 86/513)

'

~~~~~~~~~~~

b. ArchaischerStroter. (DAI Athen, Neg. 86

1

c. Ostliche Giebelsima des archaischenS-Tempels. (DAI Athen, Neg. 86/482)

d. Geisonziegel vom archaischenS-Tem Neg. 86/490)

GERHILD HUBNER: DIE DACHTERRAKOTTEN DER ARCHAISCHEN TEMPEL VON KALAPODI (PH

APXAIKEE KEPAMIJEZ AHO THN ANATOAIKH AOKPIJzA (PLATES 17-19) T

A ETH 1978 KAI 1979 OT7OVS P0pELOVs

Tov 7p07To7robfE

7 KOKKLAo#ov KaT76pa6KL

V0d8payosgj ]FapIiLVLToa T7)S KOLVT7T77-as KvvrapLorcrov 0T7)V AvaTOALK I AoKp&'a KaTaT?JvapxaLa To7roypaOLa,lo o07TLO0 (v,u4ova /ue 7roAA\\ovs'ue\eATTes 7rapovaTLa'EL TLg 7TLOaVO'T2TfES y&a va TavTL-TTEL /uE T7v) aKpO7ToX7oTOv apxaLov 07rovra,2 7TEp
KT?)pLO T?)9 apXaLK?,S

Ta avao-Ka#LKa 8EbeOULEa p/asa

VTO)X19.3

obI7yovv 0E TavTLO-7)TOVKT7)pLOV avTov

cso -TToas,7TOV

aUV7Kf CE Ka7TOLOLEpo.4

H (rroa

(TTLv avoLKTES'

Xto. EL'XE 0-7O

- 8L'PPT7)

tX v 7zXEVpE T?-ri Ka

aLST WMaTa

LE 7TW&p)VOVsb(p&KOVS KLOVUSKcaL FWCp&VO C7TLTTrv-

To yeyovos av-o

TLrSbrVo rTfEVES 7rAEvpEST7).

f7TL/3E/3aLWETa a7ro Tyl/ua

Tv/.L7ravov 7TWpLVOV KaL av64fuLa KOpv#patw KEpauLLIi 7rov /3pEO7)Kav KaTa T7rv avao-Ka#7,. Io-COS!ILE'OET KaL EOCTfpLK7' lVAe?LV7 KLoVoO-TroLXLa, 7paK-

e

T7LK?)/3e3aL0,LEfV?) 0Tov 6o aL. 7r.X.5 S7apayuara aAAovEatuLaa dT1IL?),7TTV)XWeoV a7ro 7rIr7rLvaayac uara 71aKpwrj?pLa, 7rLoavov T,/.u)/.aToq 7T7 XLV?)9 /E7-07r?)s K.a. ,uas uLAOVV y&a Eva 7rAoV Lo apXLTEKTOVLKo bLaKKo7r,OT7jq crToaT. AAAa To TOTOLXECLO J.E TO 07To10 Oacras a7ra-xroAX?io- ELvaL Ta 7rAoV'-La KaraXoL7a T7-i r

ty77s

av7-779

7-77g

a-7oag.

17TSL TO/t's; 7uV avao-KaT77?Kav

aIiLaTapaKTOorTpcw/a

xcp.aTa

H

orToa

oTav

OLKObOU ?O'qKf

arV7KOVV CEoVO IiLa#opeTLKOVs

/3p'O7)Kf

KaTaOrTpo#?7S cTTEya-T7KE

0Ef /LKpO a7To /.f

00os9KaTW

a7To

Ta fETXLaVELaKa

KEpaULbes! (P1. 17:a). KEpa/L'bfS

KOpLVOLaKOV

TV7TOV KaL aVTES

TV7TOVs.

O ayckAalf

KaL KaAv7rnj7KEpa1u'ibEsf7)TavTov crvvnOovs Tv7rov, T7-pwnrpas E7r&rELbs9 pas~ TpLyWrKos~,XwPLa (P1. 17:b).

'W. A. Oldfather,RE, s.v. Lokris. Alzinger

= W. Alzinger, et al., "Aigeira-Heperesiaund die Siedlung Phelloe in Achaia,"Klio 67, 1985, Oae.

389-451.

Goldman = H. Goldman, "The Akropolisof Halae," Hesperia 9, 1940, 0-E. 381-514. Payne = H. Payne, Necrocorinthia,Oxford 1931. 2 'E.a., C. Blegen, "The Site of Opous," AJA 30, 1926 (pp. 401-404), a-EA. 401 K.E., A. Philippson, Griechische LandschaftenI, 2, 0-EX. 355, ErnstMeyer,RE, s.v. Opous K.a. 3 (. AaKopdvta, AEXi- 33, 1978, XpoVLKa, 0EX. 139 K.E., 34, 1979, XpoVLKa,-EX.187. 4 Elval yvwoo-ro OTl ia IALaa7rorTls Op-le rovs

o-i-tK4

OlKO8bOl.L7ara

r7flSapXatKflS' c7roX?1saVflKOVVKara

KVpLOAOyO0E Lepa KaL

71ravyLa i-ly -ireyao- a4Lepw/.Aairwv U. Coulton, The Architectural Development of the GreekStoa, Oxford 1976, OEX.9). 1OES 'E.a., OEX.23, 189, ELK. 11,12, 83A.KaLava7rpaporai-ao i-D i-roaisicov AOqvaiov oalr-ovs AEA aA. 187, ELK. 7.

F. DAKORONIA

176

O07TOLOVOL/uEET7T ESt OL 7)ye/OVES T7)S OpLCOvTLas7rAevpasgazV7KOVV(TTOVTV7TO TTOV 8)A. o KaAv7TT7pas /e a7rob&'IOvv KaTa KVpLO Aoyo Tov apxaLO opo <>A ELvaL EVLaLOsFuETOV o-TpWoT7pa, 0 07T0sL0 apXLCELa7ro TO KtVTpO 7TEpL7TOVTOV 7ravTa 7rpos! Ta bee4a (P1. 18:a, b).7

To avO,uo

avOqu'ov

KaL EKTELveTaL

H roXoAGcr7-LK?7TEpLypa.h7 TPV Kepa/LbLIOVVOIuLC, oTL hEv Oa ElXE vrorva E#'ooov 21 ELILK v. Ea 7TEPPLOpTTCW 7TapOVo-Laorr7TOVs yLveTaL EVW7TLOV I.OVOV -T7)V E7TLOi)/.avO1) ELTE7TOV La#opo7TOLOVVTL9sKEpadLbes!avTEs! a7roaAAES O,LOLES,ELTE OpLOulEVWV (OTOL)XELWV, 7TOV fE 4EL'vaLEVTEAWsopaTa.

oAoKAMp?). 'ET0-L a7r a7ro TLs KEpa aLiESg hEy 3pE`O27KfE

Kaula

)(ap

va EXOVILE /OVOV ,LpLKfS

a-OaAELS!

haOTaOELS!

oA.

T7V O-VyKOAA7X)o1 E7TLTV-

r\aTos!

Tov 7)yepova OTpW-

0,546 ,. 0V VOALKO'7AaTos! ual .e Tov av0E,.L7o KaAv7rT7)pa0,666 ,u., To vifos! Tov T rVTpW T7pa 7rov ELvaL 0,07 ,., To vifos! Tov avOquLov0,275 u. To uf'yLo-ro ,UETowov Tov OTL TO OTVVOALKO' EVpEOE,V1c?jKOS avOcEpiToTvKaAv7rTT7pa ELva 0,54 u. aAAa EKa'CETaL /.47KOS TL oTL OL a7ro ayEAaLES. 7)TaV Tov 77Tav 0,58 ,. ELvaL )yE/uOVErs KOV7TvTpcv /3e,3aLO 7r7pa

OAf9

OL KEpauLLhiEs ELvaLKaTao-KEVaor7.EvEs a7TO GTKA?7pO KOKKLVW7TO 7T?Ao KaAaa'r7J7uE-

VO, 7TOV 7TpELL)XE 7rpoIG.lLU? pe a4upo. H avo oparT7 E7TLAaVELa Tovs 7)T-av aAEL/4LEV7) /.E Ta 8LaKoG-/rjTLKa KLTpLVOKaOapo 7rqA;O,7raXVWOTOV O7TOLO ELXav CWypaULTTEL TTOL)XELa. 17O

KaITo (epoS

TOV9 #aL'VovTaL aKo,A7) Ta a7ToTv7T

AELWV TOVS. bLaIuop#c0s To /Lt' w7TO TWIV ?7yqAOVW0V /laTa /lavpo,

4ELVaL bLaKOO,L?/.4E'VO

(OTpT7)pwV

KOKKLVO, /.47rE' KaL 0

7XAouos

aTa T7)9 Irpag

aXEhLa'rT?1KE ap(LKa

/7)T71 KaLLxapaKa, O7TWS/Laq /3E/3aLWvOVv Ta

-XErTLKa XapayIaTa8

JAE 7TAOXYLO

KaL

JE

Tov epyaTpLa

Xpc-

T07o /uaAaKo 7T7)A;o fE lLa7TOV ELVaL EvhLaKpLTa.

Ta bhLaKOOW,TqrLKa(-TOLXELa Tov aVOqLloV, 7rETaAa, E`\LKES, beO,uos, opOaA,uoL, arTa/uE o-KOvpaXPW8aOos!KaL TOVLCOvTaL yova, 7rpo,3aAAovTaL avayAv4a 0cEaVOLKToXpWIuo paTa, KOKKLVO8aOtV KaL uavpO. 1To G-?flqLEOEK#Va-EOS! TOV av0e/LLOv awo Tov 014uo#To O-TpWT7)pa br3yLOVppyELTaL AE7rTos ava,83a,uo's 7rov Orn/LaTobo7eLTaL MEKOKKLV7) TaLVLa. -TOLOL KepauLLES av vT7)sq T7)Sq O-ToaS pAas!8L'VOVV I.LEpLKa EVhLa#epovTa KaTaOKEvacrOTLKa

cTO opwTO 0-Ta, O a-cETpLa 0-Talia. KaTao-KEvaO-T7?Kav KaL a-E -VI4VTW0V COV avT7)vT7)/v KEpaL'L ayEAaLCV (Ul)VV7?lov PaO-7),evo 0 a-T7)v TOV TpLyWVOV X EPa Ta EKavav 83aOovAalKc? 7r7)Aof 7auaAaKOS, s)TaV /.4E aKOp?j T7oPpo4ecxov Tyr7fua Tov avOduLov LE ,uara (P1. 17:c, d). le bEV'TEpOOTaIbLOE7rAaaa-av )Ga.

OL avOEucwToL KaAVTT7)pES

ELXVaT?jv

,Op+?1

6

2., Rhomaios,aTEA.12 K.a. GFR OEA.104, ap. 33, Payne, OEA.250, v7roo-,qu. ,ua a7roEvaKaXov7lt, ELvaLovvbr1X.7Aao,uAvovs, KaX v7r-rpa cvLaLovg, ipaL 7r-pcor-7pa M. C. Roebuck, "Excavationat rov 7. aL. Payne, aTEA. 253, 7r.X. a7o' ?)lOoAE'ves o-nfv KopLvOo 28bq (,3A. Corinth: 1954," Hesperia 24, 1955, OEA. 156 K.E., 7rLV. 162:f, S. S. Weinberg, "Excavationsat Corinth E 7roAA ,.t7 i-D EXXabos, ElTe cs ayeXales dlTE Cs 1938-1939," AJA 43, 1939, OEA.595). 'EXovvf3pEOda a'p ?)VyfuoEvs MEavO4uLa,66cos 7r.X.o-frlv HlepaXopa,a- oria IoOIALa(Roebuck,f'.a.Oscar Broneer,"Excavations at Isthmia, 1954," Hesperia 24, 1955, OEA. 110 K.f), i-rqvAOq'va(GFR; TdA aTEA.27 K.E. ELK. 6, 37, 38, a (Rhomaios,OEA. 18, 7dlv. 1), i-rqvKE'pKvpa (KorkyraI, OEX.9 K.E.,TEA.135 K.E., ELK. 106,107), a-rqVALrtALELK. 7; Dyggve, OEA. 191 K.E., 7TLV.XXIV:A), o-rovs AEX4ovvS (Le Roy, 1967, OEA. 24 K.E.), orTlD YELTOVLKE( AA's (Goldman, aTEA.441, ELK. 102). 8 Dyggve, TEA. 191, K.E., 7TLV.XXIV:A. 9 Er-v KaAv8ova 7rapo1oLaRhomaios,OEA.18, ELK.7. 7 KepaAubles, JE

APXAIKEE KEPAMIAEZ AHO THN ANATOAIKH AOKPIAA

177

TOV KaAv7rT7)pa aVTLo-TOLXa 4eOyKW.LaTa To 07T0L0 KOAA7)0-av 0-To KVpLO 0-c/la 1T-0 TpLTO a-TaIbLO ETLaeaVa KaL KOAA7)r-av To avayAvco o7)r osr?s Tov avOLdov.

(P1. 18:c).

Eva aAAo -TOLXECLO ELvaL OTL OL7jyE/lOVfE KaAv7rTT7,pes a7rO KaTW 7rapovrLa ovv lLa KOLAo'T7Ta 0TTO 7TULfr pyfpos! (P1. 18:d). AvTo 1asq OXEiboVopOoy.4v&a 71 o7rOLa o4iys KAELVEL opOoy.vo V7TOIELKVVfL OTLfLXEyLVELy&a va a-TpL)XOOV'V 0e Ka7rOLOY7TPO4eEXOa0L0oLXa /Le/.oOVW.LEVO

G-TOLaXELO

T7) 130 pW0-?1s

T719 0- TEyrls

yCyOVO9s

7TOV TOvAaXrL-TOV

y&a

T7)V 7rEpL-

7TTW0-7)1avT77s1 T7sq 0rToas epPeXTaL -cE aVTLOEO-7).E I

I

I

TWV Kepa/.LIV

0-Ta

I

Tr7)

TL9sOVVqOfLgsava7rapao-Tcra-CLs TOV 10 I~~~~~~~~~~~~ -itEy7SX.

GTfptwu79 evAa To G-TOLXEloavro .yastE7TEt /pee va V7roOfE0OV/.LE OTLTO /Ji)KOS TOW 7)yE/.oV@WV 7TpE7Tf va

Tp07TOV

7Trav 0,58 F. H a-TrepE`W07)0/.L9

TWV )yE/.lOVW0VKepa/.LIbcv

KapLa'a. OL ayEAaLoLKaAV7TrT?pEsELXav a7ro

eaac0aA

L

aU KaL/E cIEpELa

Eva e6yKW/ja opOoydwVo7rAaTOVS 0,04 FL. KaL/.L?KOVS0,09 U. 0-TO FL7TpO0-TLVO oAl00CLXe va e/7TObLe' T) ltpOSu TOV9, 7TOV0-KO7TO Ta KMTm(P1. 18:e). T&oLo 4eoyKW/La vTapXEL 0-E avOq.J07TOVS 7)yE/uVfEs KaAv7TT)pES 7Tpo0 Kaiw

aT7}v KaAVX1vva KaL 0T7)V ALyELpa1 aAAa loovov 0r7TV ALyELpa /3pL`-KeTaL KaL 0rTovq aye'12 Aal`ovs KaAv7TT7pes. H bLaKo00-/.?)0-?) TrV KEpa/LLIWVTOV Kv7rapL0--L'ov ELvaLa7roAVTW0 LooLa/ue avTV TWv W ELvaL Alyo /.LKpOTepa AAKepaFlL'VzWTWV}AA v LE Fuv7 bLa4opac OTLTa avO4La TW

(v4os! 0,21 UL.KaL7rAaTos!0,193 FL.EVaVTL0,275 U.. KaL0,225 ,L. T}V 3LKV /LaS!).l3 A7IoXAvTa -TLS AAE'TO (LoL ELvaLKa Ta avOEf',aaTV KOpv4aLCwV,7TOV07r09 E'rL KaL0TTo Kv7rapL0-0L elvaaravI 14 ULE aAAa hEy} Ypa( -a TOs 1V S b EL,vaL aXa G-TL9 v' O1/IEL9 9 (TEE7rL7rEhi)E7TL#a6ELa avayAv4a

cyaL--'KaL

Ta LIa xp4aTa

LE Ta o7Tol

aL ClwypacrLu

/tva

KaL /lavpo EKELva, KOKKLVO

0re aVOLKTO

KaL G-T9sb1VO7TEpL7T-EL9s OLEALKES! a7roAvT7) aVTLG-TOLXLa b13?A. ELvaL uatpoL, ,3aOo0, KaL TcE ra 7reraAa TrEtu-Epa KOKKLva KaL TpLa /Lavpa, o bea-L0,9sKaL )1 0rTayova KOKKLVES (P1. 19:a, b). To 7rpoypa/L/La bLaKo00/crlJ.?)T?j)s 0t-TrEy?s ELXE7ToAV /.EAET?)OEL,co0rTe va TaLpLa'oVVTa av0'UELLa rTcVKOpvcaLC)V LE ra avO4LLa T-1CVr/yEfL0Vo0V KaL hEy 7TL-TEVCO 07TLOLKOpv0aLES avOE'eLa u7ropovv va XapaKrTPL0-Trovv4027ves'("cheap") ocsy pLOvTraL ra 7rapacAA7Aa T7-r KopL'VOov.15

XFpls avcyAv4a 'DaL'veTaL

e

XaPaaKI71-

aAAov O,TL7Tarv?1 0-VV)OLu0I-CFV-77TspaKTLK?70Tovs

apXaCKovs! Xpovovs! ra KaLEKELva TCOVKOpvcaL(Ov Ccoypa4 Wm-a.6

av0E'LLa Trov 7YyeFLoo va ELvaL aayAyvIa OL KEpaFlLhESr TOVKv7rapL0-0-Lov 7rapoV0-LaOVV KaLaLGr07)TLKes` EKAE7rVV0-ELS! 8rA. 06OS! KaL0-rTrv KaAvhi4va ErTl KaLEI EXOVFLET?)V Aav0avov0a KLV?)0-?) 7TOVELXE7raparT?)pr?)EL0

PoLaaoL17 KaL KALVOVVEAa0pa 7Tpos Ta apLrTepa. 10 IH.X.Martin, 0-E. 97, ELK.46, 104, ELK.52, K.a.Av. Opa'vbos-,Ta AO2'vaL 1955-1956, I, OEX. 93, ELK. 45 KaL94, ELK.46.

VALKaabo,ts-

" Dyggve, 0-E. 193, 7vL'.XXIV:E Alzinger, OEX. 432, ELK.29. Alzinger, OEX. 433. 13 Goldman, a-EA.440, ap. 3 KaL4, ELK. 102, 103. 14 Goldman, a-EA.440, ap. 5, ELK. 104. 15 CorinthIV, i, OEX.10. 16 IH.X. -7TivKaXvbcs'vaRhomaios, OEX.27, ELK. 11. Alzinger, OEX. 433. 17 Rhomaios, OEX.101 K.E. 12

rcovapxalov EXA71vwV,

178

F. DAKORONIA T0O7rXal'o-Lo T)sv KopLvOLaKr

OL KEpaLLbfES7 TOV Kv7rappLLov a,/L'aorTa eVTaO-oVTaL

va ELvaLairo K'pLVOO. To a4rapaLT7)TOl TOVS' T7)V a7roXvT7) .e aVTes T7O yELTOVLKWV LO3pKOr7TO;v\7rLOavo .e 3ao-) 0,.OL0oT7-Ta AAr'V zKaL fE/a3a717 7rapaT7)p)cEfLvS a7ro aXAxEs 7rapolOLEs18 Va Ex(OVV KaTacTKvfaTTEL (-T7V AOKpL'ba, q oroLa ouwcs yYVpLCOVfEKaLrTqXo KaKaAx?)xo pTOpEL va Tpoo4rEpEL KaLKEpa-

'rapaVLo-rva. 7v' .

H JpOEXEVO? T

,uELKa epyacrXpLa

ELXEa-T7)VapXaLOT71ra. HaflvTS! ELvaL,3E,3aLO OTLKaLOLbvo 7rpotpXOVTaL

awro TO L'oLOepyaoTrpqLO. ,Nov

H GoldmanXpoVoAOYEL TV TLS9OIOLEsAT IiVEL ELvaL roA xa,Xai IcaLjEa8fEL`vaL

o

o-T-a TEA) Tov

6ov aL. i.X. H ?)pfpoprvLa

UOVoV 8K7LK a7rop?I. uoV

Kepaulbes Trcwv AXAv KaL Tov Kv7rapnlLov a7roreXOvV ,ula 7rapaAAay7 EVOSTVITOV 0 OJTOLOs 0-TOV yEVLKaE4rTL7TTEL0-TO 'JTpWTOJllTO TOV 6ov aL.r.X. KaL eVpEWsybLabeboIeOVov, OJTOLoarV7KOVV Trow AOqvc4v,20 T77s KaAXv8@vas,21 T7o lELy/uara o'rJs' T71s!AKpOFrOArXS AEA#WV,22 TOV fTdov,23 Tvjs KopiLOov,24 a rTvs OXvpridas25 Awo oAa ra lpoavaVaEpOe'vTa 7rapabelylAaTa /Ji7ropov/uE va bLaKpL'VOVIAEbo viroTtV ov: (1) avra 'ov Karw a7ro TOV bEO7fLoeXovv tIEva aveTTpaJ4evo avO4Eo ouws avTa T77s' AKpoIrO;7)s, TnS KaXvc'wvas', T7OVAfX#WV, T7)s!KopiVOov, Tov HT7.0ov, KaLTr7sOAvprLdas!'idv.CXIX:1 KaL (2) avrrarov OL

0-T7 OfEO?1 T7s) wraXpe'Tag EXOVV/Ula

7SATV owcoT

po/,43o0o-X7l) cr7ayova

TOV Kv7rapLO--

OxvuLaas! 'idv. CXIX:3. 'rov VylKpLVOVTasg aT-Est Tow AA4V KaL Tov Kv7rapLO-o-LOV ue avTes! T7-s KaAvbovas!, 580 'TEpl 6OV XPOvOAOYOvVTaL TO uJ.X.,26 /E avTEs T2s AKpo0AkS, TO ap7w TOV a&. KaLT7)s OTLOV,

7T.X.,7-T1OV

AEA4)WV Jrov TO7TOfETOvvTaL/LEcTa$t? 600-560

7TpOA)TOVJMUTOVTOV

KaL.X.28 Tov 7cL TV)s KopLvOov 6ov aL. ir.X.,29 l3AE'TOV/uE OTLOL 7Tp(WTST ELvaL ITLOpabLVEs vKa Ka

AOTIEpEs,Ta 77EtTaAa eLvaL rTLOKLV7)/LEvaKVpL0s9Ta aKpaLa KaL EXOVV$E#VYEL 7TEpLOOTEpO TWV aKpwT7-7pLoV TOV 7ov aL. a'ro T)V TPLrWVLK7) ILOPvp4fll Kv7rapurO-Lov KaLTwvAAdw 7TpE7TEL va ELvaLVE.')TEpEs. CEILO-7)S/ue ?VyKpLVOVTaS TLS!

avTes' TOV O70-avpov

'r.X. 'ETO-Lao4oaAs

Tco

Meyacpwv

avTES9 TOV

TT7V71OAvpLa30

oL

c crav terminus yLa xpovoXoAy?If KaL /,L3L,3ALOypa4bLa t-ptc,oT'V) a'roI/fl arV27KOVV 0-To TeXos Tov 6ov aL. 'i.X.31 uJTopov/ue va

07roles! XP?10,u07roLovzTat OL OJTOLES!KaTa yEVLK7)v

18 Goldman, 0-cE.453, Corinth VI, i, aTEX. 40, Rhomaios,0-cE. 12. 19 TdA II, orCA. 35 K.E. [Editor'snote: Goldman dates the roof ca. 490 B.C.].

20TdA II, aCA. 35, ELK. 46,

7LV.

II.

Dyggve, 0-cE. 236, Rhomaios,0-cE. 301. Le Roy, 1967, 0-EX. 46, A23, 7dv. 7:5. 23 Le Roy, 1967, 0-EX. 61, 7dv. 17. 24 CorinthIV, i, 0-EX. 12, ELK. 12, Payne, 0-EX. 160, ELK. 109, B. 25 Olympia II, orX. 189, 195, 7L'V.CXIX:1 KaL CXIX:3. OLKepaALbes, OTl arV)KOVV avres 7rLOavaXoyEL'raL 21 22

rE Eva Orqoavp0rcov BvCavrL-v 26

i1 r-s

EirtLb4avov.

'E.a. v7roo-,qu. 21 KaLPayne, OE-. 255, 258, 260. 27 TdA II, E.a. 28 Le Roy, 1967, E.a. 29 GFR, OEX. 103, ap. 20, ELK. 71. H Van Buren Tro (OEX.260) 4uWs'T0E0copel'VfYiepOiepO. 30 Payne, TEX261. 31 Olympia II, OEX.50 K.E.

pOvoAoyVE O7lw apxE's, T-ov

6ov 7r.X. 0 Payne

APXAIKEE KEPAMIzEZ AHO THN ANATOAIKH AOKPIA

179

&OO-OV/.E lda ao4aA?I )(poVOoy71?1 TWV KepaFldILW Tov Kv7rapO-dLov KaL aVTLOrToL)(a Trv AAXd 7repL ra u4E'o-aTov 6ov aL. 9r.X.32 'avrT?v T?)V ppoooyl asg obX?yoV KaL 7roAAa a7ro ra v7roAoL7rao-TrOLXLa Trs) avao-KaITs, o7rws ra oo-rpaKa KaL a XaAKLva. 'O'Trw avaO'pO27KE, 7) 0-TOa oTaV KTLO-T?7KE ELXEuTEy7 KOpLVOLaKOV TV7TOv. MeTa$f TOV (TVyKEVTpWOV7TWV KEpaLbOiV 0/UO@S9 KVpTOVS AaKCOVLKOVS! KaAv7TT7)pes.

Eva 7TOAV .LLKpO7TooooTo

arV7KEL CE Ko/4LaTLa a7ro arVKOVV oE .EcTayEEo-TEp7) f7TLOKEV7)33 T7ST

Hpo0avco

y&a Tr7l O7rOLa/.7TOpOV/E KaL va efpOV.E 7TOTEEyLVE .E 7)V8/3 0OELa btfo o4pay&o-OrTEy?)S a0aTW34 /l3OLrLKOv epyao-T7)pLOV (Pl. 19:c), 7TOV XP oAo yovvTaL 7TEpLTO 480 7.X.35 KaL O.LOLaTovs EXovV /pEOcl oTo KaAa7rofIL KaL Trv AL&y&va.36

To cpcfT7-ua yLacTL&E (P,OpLUF0o7rOLqo-aV KopLVOLaKov TV7TOV KEpqbLesg yLa T7)qV E7TLOfKEV7)/.7TOpELva beXOeLula avarro7r dTL EaV 8fEX0o0L or)(av ELOaXOEL araoSTv7VKoSpLVOo a /OEVToVV ov va 7rpo 7000 a7ro KOVTa Kepa/.L3es! ocro 7) YEL7OVLK7 v.l7ropov%-av h?qA. BoLWTLfa yLarL va KaTacvyovv oTr7v KopLVOo.To yeyoVo s4 c AAfov /as! ilEXlVEL KaL KaTL afAAo,OTL Or^rV

BoTLa

rT7V e7roTX? OLKOfO/.l?)lo?) T77s o-Troaq beV v7T7)PXVa /FLyafA\s! E//3EAfLas! epyaoT7)pLa apXLTEKTOLKWzV OTL o TTLq ap(Es! Tov 5ov To Kepa/LELKWOV,KaL TEA f OVo7T(AO 7) 7) 7rp(oTo7ropLa T7ls KopLvOov 0rT7)V OXET-LK 1Lo0LXaVLK7) 7rapay)y7 KaL

/br/egacrderlere.y(Ily egay^yELXE / elXe 7)h7) 4eaoOEVUTEL.

4eaywynf

fHavcTos a7ro T7)V OCEO17TOV /pEO7)Kav Ta AaKAVLKa KEpa/l8La caLveTaL OTL7)1f7TLUTKfV7

'~~~~

f f f a4opovc-Ee 0cETr/.L7)a Tr)s OTrEy79 OXLa/.cL-a oparo. f 37 H c-roa AOL7ToV 7)Tav ?TEXP7)()l IyEXPLro 480 7r.X. H KaTa-T7poqf7 rTs! 7rpo07AOe a7ro ELvaL yvorTog evas rEL07,Lo. H OfEO71 ELvaL KOVTacTro yVWa-ro p7)yLa T7-r AraAarrr/s. o.X.38 OTV o7olo a7ToILIETaL KaL P KaTaorTpE7rrTLK09O-cLO-.L0rG-TT7V7TEPLOX7)cTTa 425

KaTrao-rpoc?7 Tov vaov cTroa

rov

Kv7rapLwrrlLov,

AAsx39 (a TrTLs 9Ea7ropovOE aV KaL GTE7Trp(7)

4ETacT7)

va ELvaLavTrO, 7TOVKaTEOtppe*/E KaLrTv TOV VALKOV T7-S7 aVaGTKaO?cs, EKTOT rTWv

bev caLveTaL va v7rapXovv evp7/.araTa rov reAovs Trov5ov aL. 7r.X. 'ErTL apXLTrEKroVLKWv, Troevp7yla EXEL/ua LbLaLrTEp7) o-7/.Lao-La-KaL yLa rovs TcELo7.OAOyovs, 3EIO/.EfVOV OrTLEvaL bvvaTo fva KaraypacEraL E&8 KaL afAos0 evasg OELO/Los0, evwpLTEpa, 7rov bev aVa4e peraL OTLS7T71yET.

32 METv

LI, o-EA. 35

3 HapaI3'yj.ara o-irqv AtyEdpa 3

Ta

yqo-?/ ovupcwvovv Ot Van Buren, GFR, o-EA.103, ap. 23, Buschor, TdA

xpovo 7TpOTELVOMtEV7 K.E.

/3E/3atwjAEva

,BA. Alzinger,

o-.payto-jtara

E7rLO-KIEVO)V OTEY71S

avacEpovTaL

apKETa

oT7/v

0cos' 7r.X.

/3/XLtoypa4)La

o-EX. 438.

7rapLo-rTovv

1.a

pOCETTra rov 7rEpL,aAAErat a7or E7rLypaOj.

Ot E7rypaEs-

TAHO-

Z\ABYXEIAN [EMI AA]LOONKat [T] AIIO Z0[ABY] XEIAN [EMI AAIOON]. 35 R. Felsch, "BoiotischeZiegelwerkstattenarchaischerZeit,"AM 94, 1979 (OEX.1-40), 0-ex. 20 K.E. 36 'E.a., orEA.14, ELK.4, 5, 7, 8, 10 K.a. 37 0

TvvIbvaqo-

7ravToS

KOpLtvLaKOV E7T7L7rE3V 0T7pT07/pOV

JE KVpTOVS' 7//tLKVXLVIpLKOvS'

(XaKwVLKOvS)

ELvatayvwo-rTos oT7qv apxaLor7ra Kat KaAElTat OTLKEXLKOS IE7TELI7 ?rav OT-7V ELKEALa bLabeboIAeoVosg OpXa'vbos, E.a., o-rA. 102, ELK. 56, E, Z, H, 0. W. B. Dinsmoor, The Architecture of Ancient Greece3, New York 1973, o-EA. 44, ELK. 16. AAAaKaL o7-rv M. Aota elvat avv7/OLO-EC'V7/ 7/ 0TEy7J f ETL7EbOVS O-Tp(OT7pfsE Kat 7,MLKVALVIPLKOVsg KaAv7r7-7/pES: /3A. Martin, o-cE. 70, ELK.26. 38 ?OVKVbL'b7Sg, III.89. AZLo'bpOT, XII.59. zIppa/3wv, i.6o. 39Goldman, oEA. 454. 8Ev

F. DAKORONIA

180 TE?EuovTas'

Oa 7OE?a va ava#EpW KaLva o-as b

ev

Ea

TEPLEpyO, 0TwS0 v.CwO,

apXLTEKTOVLKO Kopl,aTL), 'rOV 3pEO7qKE 0T7-0 4ErOovT7?)sa7ro0r0Taor-s pyeTaev TOV -7Tv?Xo,arT7 7raEvWt KaLLTOV VOTLOVTOLXOVT7?) O-ToaS. HpoKELTaL yta uLa 7T7 XLV? 7rAaKa 0T7)V 0V TLa, 0rV/uVTOS

yL'avro

uE avTrJv, evas KvALVibpOsbLapu.0,24 j. (P1. 19:d). H ep/.7vELa xov oKEfrrqKa , y&a /aoT?) ?) e7rev83V07, av 717Tav To aVTLKEL/AEVOELvaL OTL7TpOKELTaL v*/XoTepo,

~~~~~~,

fl iALvEs KOAXVES $V\LV7s KOXcAWvas. &Ey yvwp'Cw va eXovv /3peOel aAAoovKaL,uaX&-ra ,le TO o7rolo XfLTOVpYLKO XapaKT7-pa. T7)-v ALyELpa eXEL /pEOELl 8I3pLKO KLOVOKpavO 1T?r1XLVO, OEWpELTaL OTL 27Tav yta

va 0-T7JpLCEL KaTOLo ava.Oqta.4

o-av T)7v oToa Tov Kv'napL-dLov K7T?pLa To00o o-7rnIaVT-LKa a7To apXLTEKTOVLK7) a7loT71r?) OTL aVt)KOVV 0E f7TL4av7) Lepa TrS apXa>LK7) E7TOX?S KaL ?q afia TOV EVp-7

eLvaL YVWTO

eva aKojla E7TLXCLP7.Lpa ME TOv apXaLO TOV Kv7rapL-o-L'OV 7rCptOX71S 7) o7rola Oa 7)Tav eVvo-qTO O7rOvvTa, T7qvo-7)/LaVTLKWTCp7) ITOX?)T7-Sq avaTOiVK,7)' AoKpaas',

I,.LXPL0-71/iepa puaTos,p.ova&LKOV ytac

Tv7V TavtTW7) TOV X4oov

va baOE'TEL TETOLa

OTL 7IpOfOETEL O-T7p 7TEpLOX7), ELvaL

Kao-TpaKL

KaL TjS'

KT-7pLa.

F. DAKORONIA IA' EPHOREIA PROISTORIKONKAI KLASSIKONARCHAIOTETON

Lamia 35100, Greece

40

Alzinger, ek. 449, ELK. 41:a.

a.

IrTpwxa

TOAa KaTaoTpO#7

7

KEpaM

T 04T7V

M

b. AyEAaL'C KopLvOLaKEs' KEpa/lb8e

d. 'O*/7rqvpa'X719 c. M7rpoorLvil' 0'Inq c, d. BaOovA?o,Aara,Uf ra

F.

DAKORONIA:

akKTvAca OT7v

aK/Kt7 rov TpgyoVov

7y)E,ovP

APXAIKE? KEPAMLAEI AHIOTHN ANATOAIKH AOKPIAA

PLATE 18

a. HyE,uOVES KEpaiMLtE

7TAEVpLKc'g

mb. Hye,uovav

d. Karo 047n7lyequova

c. EfoyKuaTa

e.

OrT7)V Kam) fE7rLtavELaavOq.ulov qlyeqova

Ka5T o0fl ayeActov KaXv7TTrrpa

PLATE 19

b. flFILOC I a, b. AV@UO' KOVal IS8a

a. E,A7rposg

Skk

3

c. 4ISpaylo-para 0EcAaKovuJCOVs. KaAv7rTr7pfS

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS (PLATES

20-22)

DURINGTHE PASTEIGHTYYEARSOF EXCAVATIONSin theprecinctof the Heraion of Samos,tiles belongingto severalArchaicroofshave been found.1 Although the recordsof some tiles discoveredbeforeWorld War II have been lost, most of the tiles (or rather, fragmentsof tiles) have been extractedfrom well-stratifiedcontexts for which full documentationexists. This is especially true in the case of finds made since Kyrieleis and Kienasttook chargeof the excavationsin 1976.2 The correctdating even of the stratifiedtile fragmentspresenteda slight problemat first, becausethe contextstend to be mixed or offer too wide a range. A few informative fragments could, however, be providedwith a safe terminusante quem;on these fragmentsrests the dating of completeroofs. I decided,therefore, first to classify the tiles typologically. I hope to show that at least some 40% of the resultingroof types may, with the help of the few "safe"examples,be put into chronological orderand datedwith a satisfactorydegreeof accuracy. The tile fragmentsfrom the Heraion of Samos may be divided into a Laconian and a Corinthianclass. LACONIAN TILES

The Laconian tiles are comparativelythin, averaging 2-2.5 cm., with a minimum thickness of 1.2 cm. We may distinguish five groups: there are five types of pan tiles3 of differentappearanceas well as five types of covertiles. Pan Tiles

Pan tiles which have simple, straightedges, evenly cut off all around,I call "plain"pan tiles (Fig. 1). These are slightly tapering and so support each other just as do Laconian marbletiles, for example, those of the Oikos of the Naxians on Delos.4 I

This is a preliminaryreport on my catalogueand reconstructionof Samian roof terracottas,which was started in 1978. I would like to thank K. Godeckennot only for her interest and subsequentanalyses of the tiles but also for help with the English version of this short report. Thanks are due also to Irene Ring, who helped with the drawings. Frequently cited works are abbreviatedas follows: Kyrieleis, 1981 = H. Kyrieleis, Fu2hrerdurchdas Heraion von Samos,Athens 1981 Kyrieleis et al., 1985 = H. Kyrieleis et al., "Ausgrabungenim Heraion von Samos 1980/81," AA (Jdl 100), 1985, pp. 365-450 2 See Kyrieleis, 1981, esp. p. 57. 3Greek: stroteres,pace M.-Chr. Hellmann ("Aproposd'un lexique des termesd'architecturegrecque,"in Compteset inventairesdans la cite grecque. Actes du Colloqueinternationald'epigraphietenu a Neuchatel ... , Neuchatel 1988, pp. 239-261, esp. 256), I preferto follow the definitionof strotergiven by Orlandosand Travlos. [The termsstroter,kalypter,and hegemonhave been changedto pan tile, covertile, and eaves tile, respectively,to conformwith the usage in the other English texts-Editor.] 4 P. Courbin,De'los,XXXIII, L' Oikosdes Naxiens, Paris 1980, pp. 81-85 with pl. 59; A. Ohnesorg,"Ein

AENNE OHNESORG

182

FIG.

1. Laconianpan tiles ("plain"),Type I

FIG.

2. Laconian pan tiles, Type 2

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

FIG.

183

3. Laconian pan tiles, Type 3

Other pan tiles show raised rims along the non-concavesides in combinationwith the two plain, cut-off edges on the concave(Type 2; Fig. 2). A third type has similar raised rims along the non-concavesides and a-similarlyraised rim along the rear concaveedge; there is sometimesin addition a rim bent downwardsin front, which I would like to call an "apron"(Type 3; Fig. 3). Another criterion for classificationis whether glaze (engobe) was applied to both top and bottomof the Laconian pan tiles or only to the top. The plain pan tiles of Type 1, like the cover tiles, show glaze only on the top. Types 2 and 3 can be divided into Type 2a, rimmedpan tiles with glaze on top, and 2b, rimmedpan tiles with glaze on top and bottom; Types 3a and 3b are correspondinglydistinguished. Thus there are five distinctgroups of pan-tile types in the Laconianclass. A word on the colors of the glaze, black and red. My own observationsagree with the notes on these tiles made by Walter:5I have found only indicationsfor intentionalblack or naxisches Marmordachauf der AthenerAkropolis,"in Bericht iuberdie 34. TagungfiurAusgrabungswissenschaft und Bauforschung vom 7.-11. Mai 1986 in Venedig (Koldewey-Gesellschaft), Karlsruhe 1988, pp. 11-16, esp. p. 14, fig. 3. 5 ProfessorH. Walter and architectAngelika Kubanek-Clementesorted the Samian tiles, most of which were found duringtheir excavationsin the 1960's;these fragmentswere made availableto me. A final publication of the material is planned.

AENNE OHNESORG

184

FIG.4. "Plain"Laconian pan tiles, Type 1 and Laconian covertiles, Types Ia (top) and lb (bottom)

0

10

20cn

LA

FIG.

5. Laconian covertile, Type lb (Inv. no. 533)

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

185

FIG.6. Laconianpan tiles, Type 2 and covertiles, Type 3

FIG.7. Laconianfront covertile, Type 1

for red, but not for the two together.Some fragmentsbearing both colors are obviouslythe result of varying conditionsin the kilns. CoverTiles As far as I can say now, the treatmentof the cover tiles seems to bear out my classification of Laconian tiles. For example, the "plain"cover tiles must be combinedwith the "plain"pan tiles and vice versa and show the same tapering form (Fig. 4). Laconian cover

AENNE OHNESORG

186

FIG. 8. Eaves of the Laconianroof with pan tiles of Type 1 and covertiles of Type 2

tiles divide into two distinct groups which may then be further subdivided.The first group comprisesboth a "plain"cover-tiletype and a similar type with a notch at the back, hereafter called Types la and lb (Figs. 4 and 5); both types have averagediametersof about 15 to 20 cm. and show a full semicirclein crosssection. The second group, with diametersabout or larger than 30 cm. (which yields a much smaller segment in cross section), divides into a large, "plain"cover-tile type, henceforth Type 2 (P1. 20:a), and covertiles with two parallel rims that are even but bent slightly inward and with plain concavesections;these can be dividedby diameterinto Types 3a and 3b (Fig. 6). It may be of interestthat about 60%of the cover-tilefragmentsshow an averagediameter larger than 20 cm. The coloringof all five types of covertiles-either black or red, and presentonly on the top-seems again simply to reflecttaste or the choiceof color for individualroofs. Other tiles of the Laconianclass have been found with the exceptionof eaves tiles, but I think it quite probablethat they were identicalwith ordinarypan tiles (cf. for example the roof of the Heraion of Olympia6). Examples of at least one type of antefix can be identified(Fig. 7). They must be seen as belongingto a roof with the smaller Laconiancovertiles, i.e., Type 1 (althoughI am not yet sure whether to la or lb). 6

Olympia II, pp. 27-36, pl. 98.

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

rA

LA 0

I

FIG.

Im

I,

I 11

1

9. Fragmentof disk akroterion(?; Inv. no. 950), suggestionfor the reconstruction

FIG.

10. Laconianpan tiles of Type 2a and covertiles of Type lb

187

188

AENNE OHNESORG

The flat, undecoratedfront of these antefixes was coveredby black glaze and extended about 3 cm. below its tile, thus hiding the joint of the pan tiles and probablylocking neatly into the geison;until now this featurewas known only in marbleroofson Delos and Naxos.7 For roof groupswith the secondtype of covertiles we have no conclusiveevidenceso far, but I would like to suggest that they were built very much like mediaeval and later roofs, with "monksand nuns",8probablyin an arrangementlike that shown in Figure 8. Some fragmentsof ridge covertiles are known, with a diameterof about 50 cm.9One of them (P1.20:b) belongs to a roof of the smaller cover-tileTypes la or lb. Fragmentsof disk akroteriaare also known and were publishedby Buschor.10A third, as yet unpublished,fragmentcame from the excavationin 1981 (Fig. 9, P1.20:c). This short survey of the currentevidencefor Laconian roofs on Samos has left out the question of dimensions because the measurementshave largely to be reconstructedfrom fragments.Pan tiles seem to have been about 0.50 m. wide and up to 1 m. long. A few words on indicationsof date. About 25 pan tile fragmentsof Type 2a, uniformly blackish brown on top, are dated by three well-stratifiedfragmentsearlier than 540/520, even as early as 590 B.C.These pan tiles I would like to combinewith the narrowercover tiles of Type lb mentionedabove (Fig. 10). Another fragmentdates a Laconian roof with "plain"Type 2 covertiles to the turn of the 7th to the 6th centuryB.C.This roof seems to have used pan tiles of Type 2b with glaze on top and bottom;the combinationcan be restoredon the basis of excavationresults. CORINTHIAN TILES

The same criteriamay be used to define specifictypes of Corinthiantiles, but the variations within groupsare on a much larger scale;the classificationsare thereforenot quite satisfactory,and some degreeof overlappingtakes place. Extra help, however, comes from the evidenceof the many antefixes. It should be noted that Corinthiantiles show paint only on the top. Five or six different types of antefixes can be distinguishedso far; they must be assigned to differentroofs and may, in some cases, be combinedwith fragmentsof pan and covertiles. The following list of roof-tileassembliesis not necessarilyin chronologicalorder. 1. The well-known "Rhoikosantefixes"(P1.20:d),togetherwith the equally well-known pan and covertiles (P1.20:e) of the roofof the "Rhoikostemple"publishedby Buschor,11 represent one of the earliest and most completegroups. They are tiles of exceptionalquality in 7For the Oikos of the Naxians on Delos, see footnote 4 above. For the temple of Demeter at Sangri on Naxos, see G. Gruben and M. Korres,preliminaryreportsin HpaKWrLa1976 and subsequentyears. 8 Cf. W. Dorpfeld et al., "Qber die Verwendung von Terrakotten am Geison und Dache griechischer Bauwerke,"4!. Programmzum Winckelmannsfeste,Berlin 1881, p. 16. See also Schwandner,p. 292 below and P1.49:a. 9 The marbleridge covertiles of the Oikos of the Naxians on Delos have the same diameterand show similar semicircularopenings to accommodatethe adjoiningordinarycovertiles of the slopes. 10 Buschor, 1957, pp. 2-3, figs. 1, 2. " E. Buschor, "Heraion von Samos: Frihe Bauten,"AM 55, 1930 (pp. 1-99), pp. 87-90, figs. 40-43, Beil. XXIII; Buschor, 1957, p. 3, Beil. 1.

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

189

form as well as in material, quite in keeping with all other known details of the "Rhoikos temple". 2. Importanttoo are the "SouthBuildingantefixes"publishedby Ziegenaus12(P1.21:a), to which I would add some fragmentsof covertiles, pan tiles, and possibly of eaves tiles to completethe roof;these fragments,like the antefixes,were paintedred on top. 3. A group of palmette antefixes in rather high relief (P1. 21:c, d) in a bright red color alternatingwith white and perhaps five fragmentsof pan and,covertiles in a similar bright red are part of anotherroof. 4. The existenceof a fourthroofis indicatedby a set of Late Archaicgorgoneionantefixes (P1.21 :b),but no additionaltiles have yet been identifiedwhich definitelybelongwith them. This set of antefixesis interestingbecausethreeothergorgoneionantefixesare known so far, almostidenticaland attachedto the sametype of covertile (P1.21:e,f). As thesetiles shouldbe datedto the Classicalperiodon stratigraphicand stylisticgrounds,and especiallyas they copy the paint as well as the modeledforms,I would like to suggestthat they representrepairsto the originalArchaicroof. Perhapsthese antefixesbelongto the "NorthBuilding".13 5. Last but not least, there is a group of plain, undecoratedantefixeswith a sometimes unevenly applied blackish paint (Fig. 11, P1. 22:a), to which may be added some cover-, pan- and eaves-tile fragments of suitable dimensions and similarly applied non-lustrous paint. They constitutea fifth roof. The five groups of roofs defined by these antefixes may be dated on various grounds. Groups 1 through 3 representthe earliest types, with the "Rhoikosroof"(No. 1) between 570 and 560 B.C.,14 the high-relief palmetteantefix type (No. 3) between 540 and 530 B.C.,15 and the South Building roof (No. 2), on the evidenceof the antefixes, in the last quarter of the 6th centuryB.C.16 The original gorgoneionantefixes (No. 4) are datedstylisticallyto the last years of the 6th centuryB.C., while the Classicalreplacementsmust have been manufactured early in the 4th centuryB.C.17 The correctdating of the fifth group (No. 5) presentsa slight problem, as the antefixes were found in the Classical debris along the processional way, which containssherdsdating from the 7th to the late 4th centuryB.C. But if the combination which I have suggestedaboveof the antefixesof this group and the variousfragments of similar size and surfacetreatmentis accepted,this roof can be datedby some well-stratified pan- and cover-tilefragmentsto the turn of the 7th to the 6th century. These are, so far, the five surest roof assembliesof Corinthiantiles from the precinctof the Heraion of Samos. 0. Ziegenaus, "Der Siidbau,"AM 72, 1957 (pp. 65-76), pp. 72-73, fig. 1, Beil. 92, 93. Kyrieleis et al., 1985, pp. 432-433, fig. 75. 13 H. Kienast and A. Furtwangler,Samos, III, Der Nordostbauim samischenHeraion (in press). 14 Buschor 1957, p. 4. Kyrieleis, 1981, pp. 73-78. 15 E. Diehl, "Fragmenteaus Samos,"AA 79, 1964 (cols. 493-612), cols. 502-503, figs. 1, 2. 16 Ziegenaus (footnote 12 above), p. 72. Kyrieleis, 1981, pp. 91-94, fig. 69. 17 Kyrieleis et al., 1985, pp. 432-433, fig. 76; E. Buschor,AEXr 17, 1961/1962, p. 280, pl. 342; J. Floren, Studien zur Typologiedes Gorgoneion,Munster 1977, pp. 62-64, 171-173, pl. 15:3. 12

190

FIG. 1 1.

AENNE OHNESORG

Corinthianantefix and eaves tile of group 5

FIG.

12. Corinthiancovertiles and pan tiles of group 8 (top) and group 7 (bottom)

There are four other groups which thus far have no typical antefixes of their own, but definitelyrepresentseparateroofs: I would like to distinguishas a sixth group the "PO-stampedtiles" (P1.22:b).18Twelve pan-tile fragmentswith this stamp have been found so far and I am hopeful of ascribing cover-tilefragmentsas well. The meaning of the Po will be furtherdiscussedbelow. A seventh and an eighth group are characterizedby their "white"tiles: both pan and covertiles are coveredwith a creamywhite paint (Fig. 12). If they are imitationsof marble tiles, Samian architectswere ahead of their time, for the earliest examples of Corinthian tiles in marble known so far are from Naxos, and these are dated to the years around 550 B.C.19 It should be noted that only the later-lookinggroup of Samian "whitetiles", the eighth group, shows a form comparableto the marble tiles of Naxos with the same type of pronouncedrims on the pan tiles. A last group of tiles, a ninth type, always painted black, may be seen in Plate 22:c. To add some absolutedates for the Corinthiantypes 6 to 9: I can suggestonly a general date of the 6th centuryB.C. for roof 6, the PO-stampedtiles, and for roof 9, the black tiles. A 18 G. Schmidt, "Heraion von Samos: Eine Brychon-Weihung und ihre Fundlage," AM 87, 1972 (pp. 165-185), pp. 167-168, pl. 71. In the inventoryof the inscriptionsbegun by G. Dunst which I was permitted to see, two rO-fragments bear the numbers 12 and 13 (or J 39 and 40); Dunst completesPO(AEQI) and dates them generally to the 6th centuryB.c. [See R. C. S. Felsch, pp. 301-323 below.] 19 G. Gruben and W. Lambrinoudakis,"Das neuentdeckteHeiligtum von Iria auf Naxos," AA (JdI 102) 1987 (pp. 569-621), p. 600.

ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

'

~~~~0 10

\

0

1^1al i,

I ~~~~I I

191

20cm

1m I,

I

I

FIG.13. Corinthianpan tile (?) with opaion (?; Inv. no. 1334)

date earlier than 550 B.C. is indicatedby the better-stratifiedexamples of the "white tiles" (roofs7 and 8). So far only the so-calledRhoikosroof and that of the South Building have been securely assigned to a known monument.The other roof types of Laconian and Corinthiantiles I have mentionedstill await future discoveries;for example, fragmentsof Archaic simas are still lacking.Thus, no certainreconstructionof the front or back view of the temples can be drawn, just as no satisfactoryevidence for Corinthian akroteria yet exists.20Corinthian gables cannotbe properlyreconstructed. I would like to draw attentionto anotherrather curious and so far unique Corinthian pan tile which, in my opinion, could be an opaion tile. The zigzag rim (Fig. 13) might then be intendedto providea betterhold for a lid or a similar feature.21 A last remarkconcerningthe "Rhoikos"and PO-stampedtiles: Karin Godeckenis analyzing the clay of architecturalterracottasand pottery from Samos, especially the Archaic finds from the Heraion. I can reportone of her preliminaryresults, that the "Rhoikos"and the "PO-stampedtiles"are made of a distinctlySamianclay, one found in the vicinityof the sanctuary.These pieces are comparablein quality and clay technologyto the excellent Archaic pottery of Samian manufacture,even though that was made from clay of a different 20 Buschor, 1957, pp. 3-4, Beil. 2. The two fragments,which Buschor suggests are sphinxes, are in my opinion not conclusiveevidence. 21 Two other "special"Corinthian tiles, with secondary engravings, were presented in the preliminary report:Kyrieleis et al., 1985, pp. 432-433, figs. 77, 78.

192

AENNE OHNESORG

source. We thereforeagree that the r0 stamp does not signify "Polykrates",which would have been tempting as the tiles could have been manufacturedduring his reign, but should ratherbe read '7rorAXeso". AENNE OHNESORG INSTITUT FOR BAUFORSCHUNG UND BAUGESCHICHTE DER TECHNISCHEN UNIVERSITAT

Arcisstrasse

21

D-8000 Muinchen 2 Federal Republic of Germany

MONCHEN

a. Laconian cover and pan tiles (inv. nos. 1189 and 1190)

C.Fragmentof disk akroterion?(inv. no. 950). (DAT Athen)

b. Laconian ridge cover tile (inv. no. 3)

e. "Rhoikos"pan and cover tiles. (DAI Athen, Neg. 88/615)

-

rrr

-

w

n

a

d. "Rhoikos"antefix (inv. no. 861; found 1983). (DAT Athen, Neg. 83/1089)

PLATE 21

Em

mm

-

mm-

....

a. "SouthBuilding"antefix (inv. no. 860; found 1980). (DAI Athen, Neg. 80/893)

c. Front

b. Fragmentsof Late Archaic gorgoneionantefixes (inv. nos. 864-867).

d. Back

c, d. Palmette antefix, Berlin Antikenmuseum(inv. no. 494x). (Photograph,AntikenmuseumBerlin, SMPK)

e. Front

f. Back

e, f. Classical gorgoneionantefix, Samos Museum (inv. no. 448). (DAI Athen, Neg. Samos 5371, 5372)

a. Antefixes of group 5

c. Pan-tile fragmentsof type 9

b. Pan tile with ro stamp. (PhotographDAI Athen) AENNE OHNESORG: ARCHAIC ROOF TILES FROM THE HERAION ON SAMOS

CHIAN RELIEF POTTERY AND ITS RELATIONSHIP TO CHIAN AND EAST GREEK ARCHITECTURAL TERRACOTTAS (PLATES23, 24)

rT

HE RELATIONSHIP betweenthe relief decorationof temple entablaturesand relief potteryhas been discussedmany times in the past.1Scholarshave suggestedsome possible ways that architecturalreliefs might have affectedthe arrangementof the decoration on relief pottery (i.e., compositionin metopes or in a continuousband2),but for the most part they have focusedon the analogies existing between the moldingsof temple simas and the formationof the lip of some Cretan relief pithoi, especially those from Arkades. Features such as the strongly flaring lip of these pithoi, the decorationof the rim with molded tori or with other patternswhich was currentas early as the first half of the 7th century,and particularlythe ridgedprotrusionshanging from the lip, frequentlyin the form of the head of a lion, bull, panther,or horse, have led many scholarsto comparethe design of these lips with the similar arrangementof the temple sima and to correlatethe protrusionsand the heads with the water spouts of the sima.3 l Some of the problemsset forwardhere were outlined briefly in Simantoni-Bournias,1987. In the paper which follows, comparisonswith Chian material are limited for the most part to parallels from the realm of East Greek art, since influenceis more likely to come froma culturallyrelatedarea. I am gratefulto Dr. M. E. Caskey,who read a draftof this paper and polishedmy English. I also wish to expressmy thanksto the Ephor of Chios-Mytilene, Mrs. A. Archontidou,for facilitatingmy studyof the materialin the Museum of Chios and kindly giving me permissionto present the perirrhanterionstand from the Attaliotis'plot. Works frequentlycited are abbreviatedas follows: = J. K. Anderson,M. S. F. Hood, and J. Boardman,"Excavationson the Kofina Andersonet al., 1954 Ridge, Chios,"BSA 49, 1954, pp. 123-182 = M. E. Caskey,Relief Pithoi. A Surveyof Some8th and 7th CenturyGroupsfrom Caskey, 1972 Mainland Greece,Creteand the Aegean, diss. Bryn Mawr College, 1972 = W. Hornbostel,"KretischeReliefamphoren,"in DadalischeKunst auf Kretaim Hornbostel, 1970 7. Jh. v. Chr., Museum fur Kunst und Gewerbe, Hamburg 1970, pp. 56-93 = S. Miller, "ArchaicRelief Wares from the Nemea Area," in 'IILua f`771 iLSr. Miller, 1987 Mvxwva^v,Athens 1987, pp. 266-284 Simantoni-Bournias,1984 = E. Simantoni-Bournias, NaetaKot 'Ava'yXv4oLIItOoL,diss. University of Athens, 1984 Simantoni-Bournias,1987 = E. Simantoni-Bournias,"Ceramique'areliefs de Chios,"in TableRonde Internationale: Les ate'liersde potiers dans le monde grec aux epoques geometrique,archaiqueet classique,Ecole FranSaised'Athenes,2-3 octobre, 1987, to be published as a supplementto BCH 2 Hornbostel, 1970, p. 65. 3 P. Demargne, "Un pithos archaiquecretois au Musee du Louvre,"RA 1972 (pp. 35-46), p. 40 and note 1. For different views on the subject, cf. Hornbostel, 1970, p. 60; L. H. Anderson, Relief Pithoi from the Archaic Period of Greek Art, diss. University of Colorado, 1975, pp. 47-48, where she gathers most of the examples with pendentprotrusionson the lip.

194

EVA SIMANTONI-BOURNIAS

Archaic Chian relief ware has survivedin very poor conditionand has, therefore,little to contributeto this discussion. Nevertheless, Chian relief pottery permits another kind of comparisonwith terracottarevetments:its decorativepatterns especially, but often its figurative subjects,are duplicated almost exactly on architecturalrelief terracottasfrom the island of Chios and from other East Greek centers. In a few instances it is clear that the Chian potter and the Chian maker of architecturalrevetmentwere using similar clay4and the same motifs, made in matricesobviouslycarvedby the same hand, a fact which perhaps implies that they both workedin the same establishment. The manufacture of a variety of artifacts with relief decorationin a workshop specializing in the productionof relief potteryhas alreadybeen proposedfor certain areas, for example Crete, where votive plaques and relief pithoi of the 7th century bear the same stamps.5Miriam Ervin Caskeywas the first to call attentionto the productionof clay akroteria in Corinthian and Laconian relief-potteryworkshops.6Her remarkswere based on the use of continuousmoldedtori in the decorationof both akroteriaand pithoi. Although a relationshipbetweenthe two cannotbe denied,no featuresuggeststhe use either of the same mold or of molds carvedby the same person in the productionof Laconian or Corinthian relief vases and clay architecturalreliefs. The influence of one category of works of art upon another can be demonstratedin Greek art of the 6th centuryB.C. As examples we can mentiona series of roulette-decorated pithoi with relief bands representingprocessionsof chariots and warriors, or chariots and dogs hunting hares shown beneath the bellies of gallopinghorses, or riders accompaniedby their dogs. Reliefs with similar subjects,made in molds, decoratedterracottaarchitectural revetments;these comemainly from Asia Minor but are also known from Thasos and Palaikastro in Crete.7 Such scenes even played a special role on Etruscan terracotta simas.8 There is an undeniable affinity between the horse's heads on some of the Thasian Late Archaic relief-pithos sherds9and on relief and other clay simas from Thasos and Klazomenai or from Larisa on the Hermos.10To judge from photographsof these monuments, 4 No clay analysis has been undertakenuntil now. The similarityin texture and color between the clay of Chian relief potteryand that of Chian architecturalterracottashas been establishedby careful examinationof the material in the Museum. For the clay analysis of Chian paintedand plain pottery,see P. Dupont, "Classification et provenancedes ceramiquesorientalesd'Istros,"Dacia 27, 1983 (pp. 19-43), pp. 24, 30, 41; R. E. Jones et al., Greek and CypriotPottery. A Review of Scientific Studies, Athens 1985, pp. 662-663. For the fabricof Archaic Chian relief pottery,cf. Simantoni-Bournias,1987. 5Hornbostel, 1970, p. 65 and note 24. Demargne (footnote3 above), pp. 42-43. This was not the case in Corinth, however, where it seems that potters and makers of votive relief plaques worked in separate establishments;see Caskey, 1972, p. 33 and note 77. 6 Caskey, 1972, p. 36. Cf. Miller, 1987, pp. 275-276. fK KaL KoptvOtaKrs ayyELoypatas- -7-qv av4yXv4nq KE7 E. Simantoni-Bournias, <"E7drlpaoq7s paE.LLK27?, in lpaK-lKa F' AcOvo's- Tvvcbpiov flXo7rovvr,oLaKwv i.rovb 'v, KaXapa'ra 1985 II, Athens 1987/1988 (pp. 175-189), pp. 177-181, pls. Kr', I, 2, KA', 3, 5. 8 See, e.g., A. Andren, "Osservazionisulle terrecotte architettonicheEtrusco-Italiche,"OpRom VIII:1 (LectionesBoethianaeI), Lund 1974, pp. 1-17, pls. I-XXX. 9 E.g., Ecole FranSaised'Athenes,nos de negat. 30170, 30174, 30175, and 30177. 10ATK, pp. 54-6 1, pls. 22-24, A.Akerstr6m,"Ein Tonrelief aus Klasomenae,"in XapL-r7jpLov Eds'A. K. 'OpXa68ov IV, Athens 1967/1968, pp. 366-368, pl. 104; G. M. A. Richter, Archaic GreekArt against its HistoricalBackground,New York 1949, fig. 161.

CHIAN RELIEF POTTERY AND EAST GREEK ARCHITECTURAL TERRACOTTAS 195

FIG.

1. Chios, Kofinas,decorationof sima fragment.Scale 3:4

however,even here we may not concludethat the makerof the pithos stampsand the maker of the sima matrices were one and the same person. The same is true also for some decorativepatterns,notablyfor the simple or multiple guilloches,11 a few maeandervariations,12 and certain combinationsof lotus and palmette chains13which are found on relief vases as often as on terracottaarchitecturalreliefs. Archaicrelief potteryand clay revetmentsfrom Chios may, in my opinion,help support the thesis outlinedabove,for which there has been limited evidencebut no real proof:that in the case of this island at least, both classes were producedin the same potteryworkshop,or, if the architecturalreliefs needed to be made near the building they were destined to decorate, by the same potters,using the same clay and tools. II For simple and multiple guilloches on relief pottery and their distribution, see Simantoni-Bournias, 1984, pp. 125-130. The pattern is very popular for relief terracottarevetments:to mention only examples from East Greek centers,see ATK, pls. 16, 53:1-3, 56-57, 63:1, 67:4-5. P. Schneider,"Auseinem archaischen Bezirk an der Heiligen Strassevon Milet nach Didyma: Das Tonziegeldachdes Ostbaues,"Bericht uiberdie 34. Tagungfiur Ausgrabungswissenschaftund Bauforschungvom 7.-11. Mai 1986 in Venedig (KoldeweyGesellschaft),Bonn 1988, p. 16. 12 L. Marangou, "Reliefkeramikaus Amorgos,"in Praestant Interna. FestschriftfiurU. Hausmann, Tubingen 1982 (pp. 190-198), p. 197, notes 44-46, figs. 3-5, pl. 41:2. Excellent examples of the pattern on architecturalreliefs are two almost intact Chian eaves tiles from Rizari: N. M. Kontoleon,?'AvaoKa#aUEv XL(,, flpaKrTK I 1952 (1955; pp. 520-530), p. 528. There are many moreunpublishedfragmentsof eavestiles with the same pattern, less well preserved,in the storeroomsof the Chios Museum. They are chance finds, broughtto the Museum by Mr. A. Stephanou. I3 Marangou,op. cit., p. 197, notes 50 and 51, fig. 6. For the distributionof this patternon relief potteryand on architecturalterracottarevetments, see Simantoni-Bournias,1984, pp. 133-135. Miller, 1987, p. 275, note 35.

196

EVA SIMANTONI-BOURNIAS

FIG.2. Chios, Kofinas,decorationof eaves tile. Scale 1:2

One can see immediatelythat the bow-volutes and palmettes on a Chian pithos sherd with an underwaterscene14 (a chancefind dating from 540 to 530 B.C.) and a sima fragment from Kofinas, in the town of Chios (Fig. 1, P1.23:a),15are almost identical.The sima fragment may be slightly later, for its palmettehas nine insteadof seven leaves and as a result its incurving volutes are less compressed.No other difference,however, can be detected.The relief bulges in the same way on both, and in both cases the forming of the matrix and the quality of the impressionbetraythe same attentivecare. Given the similar quality of the fabric the logical conclusionis that the molds were producedby one person and that the vase and architecturalrelief were made in one establishment. To the same workshop can also be attributeda fragmentof an eaves tile, presumably made from the same clay and contemporarywith the sima fragment(Fig. 2, P1.23:c). It was found along with the sima in the Kofinas excavationsof the British School at Athens and was dated by context to the Late Archaic period by Anderson.16The way the volutes coil around their hemisphericalcenters, the form of the darts emerging from between the volutes, and the plastic impression in general are alike on the pithos sherd and the two architecturalreliefs. Although very little survives of the seven-leavedpalmettes on the eaves tile, what is left is enough to confirm the elegance and charming grace we have already observedon the other two pieces. Datable to the last decade of the century is a relief perirrhanterionstand17 bearing a later versionof the decorationon the eavestile (unfortunatelybadly worn), which I consider 14 Best photographin N. M. Kontoleon, 'Friihgriechische Reliefkunst,"'ApX'E4 1969 (1970; pp. 215236), p. 217, note 6, pl. 58:b. Cf. also W. Fuchs, "The Chian Element in Chian Art,"in Chios.A Conference at the Homereion in Chios, 1984, J. Boardman and C. E. Vaphopoulou-Richardson,edd., Oxford 1986 (pp. 275-293), p. 276, fig. 1. II Andersonet al., 1954, p. 144, no. 106, pl. 8:a. The height of the bow-voluteswith palmettesis 0.05 m. on the pithos and 0.056 m. on the sima. Dimensions of the sima fragment:H. 0.09, W. 0.06, Th. 0.025-0.035 m. 16 Andersonet al., 1954, p. 144, no. 107, pl. 8:a. Dimensions of the eaves-tile fragment:H. 0.057, W. 0.09, Th. 0.035-0.04 m. (the back is broken). 17 Unpublished. It was found during the excavationsof Attaliotis'plot in the Scaramangaarea, Chios town, togetherwith severalmore relief-warepieces:A. Tsaravopoulos,<'H apXaia7ro1ALri7sXlov>>, Hopo 4, 1986 (pp. 124-144), p. 132, plan 2, site 17. Cf. the pattern on a terracottaperirrhanterionstand from Abdera: D. Lazarides, AEAT 20, 1965, B' 3 (1968), p. 437, pl. 552:b and Ch. Koukouli-Chrysanthaki,
CHIAN RELIEF POTTERY AND EAST GREEK ARCHITECTURAL TERRACOTTAS 197

FIG. 3. Chios, Scaramangaarea, decorationof stand fragment.Scale 3:4

to be the last, but not the least, productof the workshopwilthwhich we are dealing (Fig. 3, P1. 23:b). In spite of the chronologilcaldifferencewhich is apparent in the wide spacing of the decoratilveelements and in the more slender nine-leaved palmettes and the elaborate lotus flowers of the stand, the arrangementof the pattern 'Itself,as well as the half-round renderilngof the relief, betraysthe same orilginas the eaves tile. The only possible parallels for this decoratilonin relief pottery are to be found among some of the Laconianrelief amphorasof the last decadeof the 7th and the beginning of the 6th century B.C. (e.g., the "Hunters",amphora'18). The main resemblancelies in the halfround relief of the tendrils and the volute-and-palmettechailnsdecorat'ingthe Laconian vase,19despite the fact that a chronologicaldifferenceof more than half a century must be T7jA7g a70 Ta "A 7pa>, in Kpvov, Tt7+TtKO TO/.O!g r. MraKaAK77, 1972, p. 78, note 44, for its dating in the late 6th-early 5th century B.C. I use the word perirrhanterionthroughout this paper for convenience, although the pieces so termedmight as well be louteria:with the exceptionof the fragmentunder discussion, they are all chance finds. The fragmentin question could plausibly be called a louterionstand, given the fact that the large building in which it was found seems to be a private house. Excavation of the site is not yet finished. For the terms perirrhanterionand louterion,see recentlyMiller, 1987, p. 272, note 18. 18 I follow the chronologyproposedby Dr. Miriam Caskey for the Laconianpithoi: Caskey, 1972, p. 40. 19C. Christou, <
198

EVA SIMANTONI-BOURNIAS

taken into consideration.The Laconianinfluenceon Chian potteryhas recentlybeen pointed out by Dr. A. A. Lemos.20Strongstylisticrelationshipsand similar handlingof the relief are also to be found in some Chian marble moldingsof the last quarter of the 6th century from Phanai, Emporion,and Olvia.21 The two instancesmentionedaboveare not the only Chian examples of relief ware and terracottaarchitecturalreliefs for which we proposeproductionin one workshop.A pithos sherd with a tongue pattern and rosette, also found in the Kofinas excavations(P1.23:d),22 and a sima fragmentwith a rosette (P1. 24:a), of unknown provenance,23constitutea third case. Anderson,when publishingthe Kofinasmaterial,thoughtthat the pithos sherdwas in fact a piece of an architecturalrevetment,but not being entirelycertain,he addeda question markto his identification.A closerexaminationof the fragmentprovesit to be slightly curved, however,andthus it is moreprobablethat it camefromthe neckof a largepithos.The quality of fabricis the same in bothfragments,the rosetteson bothpiecesare inscribedin a relief disk, and they each have twelve concavepetals arounda hemisphericalkernel.The only difference between the rosettes lies in the thin band circumscribingthe petals of that on the pithos, which does not exist on the sima fragment.The concaveform of the petals of the Chian rosettes, not found on any relief vases, clearly reproducesthe rosetteswhich decoratemarble sculptureon grave reliefs (e.g., the Samian stelai24)or on marble architecturalmembers.25 This affinityto the rosetteson the Samianstelai of Polycratean(and early post-Polycratean) times helps date the Chian clay examples within the decade530-520 (the pithos sherd was datedby Andersonon groundsof contextto the "LateArchaicperiod"). The most strikinganalogies,however,betweenChian relief ware and clay architectural reliefs are to be found in the large tongues or eggs-and-dartsdecoratingthem. The tongues on the pithos fragmentsmentionedabove are not the only examples of the pattern on Archaic Chian relief pottery.There are many more.26From the end of the 6th centuryB.C. on, they tend to becomemorepointed,and they acquirea dart,thus becomingthe canonicaleggand-dart(P1.24:b, c).27Together with the bead-and-reelthey are the usual decorationof a large series of Chian perirrhanterionbasins28(P1. 24:d) and stands, most of which, unfortunately, are chancefinds, datablefor the most part to the first half of the 5th centuryB.C. 20A. A. Lemos, Chian Figure-decoratedPottery of the Archaic Period, diss. Oxford University, 1984, 0a XLaK' ayyEta?, in IlpaKTtKa F' EvvIbpiov (footnote 7 above), II (pp. 69-79), pp. 75-79, pl. E:7-9. 21 J. Boardman,'Chian and Early Ionic Architecture," AntJ 39, 1959 (pp. 170-218), p. 189, A, B, p. 190, L, pp. 191-192, pls. 29:a, 31:a, b. 22 Andersonet al., 1954, p. 144, no. 105, pl. 8:a. 23 Simantoni-Bournias,1987. Diameter of the pithos rosette:0.055 m.; of the sima rosette:0.065 m. 24 E. Buschor, "AltsamischeGrabstelen,"AM 58, 1933 (pp. 22-46), pp. 28, 32, 34, pls. X:2, XII:1, 2, XIV:1. B. Freyer-Schauenburg,Samos XI, Bonn 1974, p. 176, no. 89, p. 178, no. 93, p. 180, no. 96, p. 181, no. 98, pls. 72-74. 25 E.g., Buschor, 1957, pp. 17-18, pls. 16:2, 17. 26 For the differentformsof tonguepatternon relief potteryand their distribution,see Simantoni-Bournias, 1984, pp.123-124. 27 One of the earliest Chian examples is the sherd published in Anderson et al., 1954, p. 144, no. 102, fig. 14, dated by the excavatorto the Late Archaicperiod. Cf. Simantoni-Bournias,1987, note 52. 28 Simantoni-Bournias,1987, fig. 6. Recently on perirrhanteria:M. Iozzo, "CorinthianBasins on High Stands,"Hesperia 56, 1987, pp. 355-415; Miller, 1987, pp. 270-278. pp. 227-246; eadem, <
CHIAN RELIEF POTTERY AND EAST GREEK ARCHITECTURAL TERRACOTTAS 199

Eggs-and-dartsor tongues on Chian architecturalrevetmentsof the Archaicperiod are thus far unknown. The earliest example appears on a fragmentof an eaves tile from the Kofinas excavations,29dated by Anderson to the end of the 5th century B.C. But even if comparisonwith Chian material is not possible, the relationshipof pottery relief to architectural relief becomesmore than apparentwhen we draw comparisonswith architectural terracottas from Asia Minor. The large tongues on the two pithos fragments discussed above find exact parallels in the eggs on friezes VIII and 4 from Larisa on the Hermos and on some of the simas from Sardis,30parallelswhich extendthroughoutthe secondhalf of the 6th century. The egg-and-dart pattern on many Chian perirrhanterionbasins from the early 5th centuryB.C. matchesthat on the later Larisa friezes (Akerstr6m'sfifth group31). As in the case of the rosettes,there are strikingsimilaritiesbetween the large tongues or eggs-and-dartson Chian relief potteryand the moldingsof the great marble Ionian temples of the second half of the 6th century,which have obviouslyservedas models for these clay replicas.32Very instructive in this case is - comparisonof the decorationon the lip of a Chian perirrhanterionbasin of unknownprovenance(P1.24:e) with some marblemoldings from Managros and Emporioon Chios.33The bead-and-reeldesign which frequentlyborders the egg-and-darton Chian relief ware is missing fromthe productionof any other relief pottery workshop. Its presence on Chian relief vases stresses their dependenceon architecturaldecoration. We have mentionedabove the relationshipin style and iconographybetween Thasian relief pithoi and terracottarevetmentsfrom Asia Minor; the same provesto be true in the case of Chios. Bow-voluteswith palmettesvery much like those on the Chian sherdwith the swimming Triton decorateone of the best-conceivedLarisan simas.34The sima must be ten to fifteen years younger than the pithos sherd because the half-round relief and the compressed, seven-leavedpalmettesof the latter have been supplantedon the sima by a flatter relief and eight-leaved palmettes with pointed edges which splay over part of the large, circumscribingvolutes. The strong decorativesense and the greater elegance of the sima contrastwith the stylisticallymore archaic,but not ungraceful,renderingof the relief on the Chian sherd. Given that the pithos is dated about ten years after the middle of the 6th century (compare the upper body and thorax of the Triton on the sherdwith the kourosin Munich from the Anavyssos-Ptoion12 group35or with some East Greek gems of the "SlimSatyr Group", 29Andersonetal., 1954, p. 144, no. 111, pl.,8:b. 30ATK, p. 51, pls. 19:1, 20:3 (from Larisa), p. 76, no. 17, p. 77, nos. 18 and 21, p. 78, nos. 32-34, pl. 48:1-3, fig. 24:1, 2 (from Sardis). A. Ramage, Lydian Houses and ArchitecturalTerracottas(Sardis Monographs, 5), Cambridge,Mass. 1978, p. 24, nos. 32-39, figs. 67-77. 31 ATK, p. 63, pl. 34:2. 32 Some of the closest parallels are to be found among moldings on marble architecturalmembers from Samos, e.g., Buschor, 1957, pls. 10, 11, 16:1. This observationpertains only to the tongues or eggs-and-darts on Chian relief pottery;the same motifs on Cretan and Laconian relief vases are much earlier. 33 Boardman(footnote21 above), p. 180, nos. 25-27, pl. 26:c-d; J. Boardman,Excavationsin Chios 19521955. GreekEmporio (BSA Suppl. 6), London 1967, pp. 69-72, pl. 16, figs. 36, 37. 34 ATK, pp. 48, 58, pls.31:2,32:4. 3 G. M. A. Richter, Kouroi,2nd ed., London 1960, figs. 391, 392. For the Chian sherd and its dating, see Simantoni-Bournias,1987.

200

EVA SIMANTONI-BOURNIAS

dated by Boardmanin the third quarter of the 6th century36),we tend to place the Larisa sima at the end of the third and the beginningof the fourth quarterof the 6th centuryB.C. Multi-leaved, concave rosettes and large tongues such as those on the second Chian pithos sherd mentioned above (P1. 23:d) are also to be found on another clay sima from Larisa.37Although the rosette type is not exactly the same and on the sima the Chian tongues have been turned into eggs-and-darts,the decorativeconceptof combiningthe two elementsis the same. We have dated the pithos sherd within the decade 530-520, relating its rosette to the similar examples on Samian stelai, as did Akerstromwhen he datedthe sima from Larisa to about the same period.38Supportfor this date is offeredby Anderson,who proposesfor the pithos sherd a date in the "LateArchaicperiod"on the basis of its context. The lotus-and-palmettechain on the Chian eaves tile finds its closestparallel on a sima from Mytilene (now in the Istanbul Museum) dated by Akerstrom in the years around 525.39 Although on the perirrhanterionstand this pattern is obviously later, it still bears strong similarities to the two architecturalfragments discussed above. The motif is very popular and has a long life in East Greek art, on painted pottery40and marble vases41as well as on marbleor terracottaarchitecturalreliefs. The decorativepatternson Chian relief ware are not the only ones to find parallels on the architecturalrevetmentsof Asia Minor. At least one subjectin the iconographicrepertory of Chian relief pottery may likewise be compared:the dispositionof the lion on a now lost Chian lekane fragmentfrom Rizari42recalls the lions on clay relief simas from Akalan on the Pontus, datableto the third quarterof the 6th century.43 The scantinessof the Chian material precludesdrawing conclusionsabout the priority of Chios or of the Ionian and Aeolian centers in the use of the patterns and subjectsdiscussedabove.Only for the bow-volutesand palmettescan we be certainthat their use on the Chian pithos sherd precedesany appearanceof the motif on East Greek clay or stone architecturalreliefs. EVA SIMANTONI-BoURNIAS UNIVERSITY

OF ATHENS

Department of Archaeologyand Art History Panepistemioupolis-Zographou GR-1 57 84 Athens, Greece

36J.Boardman,ArchaicGreekGems, London 1968, p. 59, nos. 93, 94, 110, 111, pls. 6, 8. 37ATK, pp. 49, 58, fig. 18, pl. 30:4. 38ATK,p. 58. 39ATK, p. 24, p.1 10. 40 Mainly on Caeretanhydriai (J. M. Hemelrijk,CaeretanHydriae [Kerameus5], Mainz 1984, pp. 96-99 and 169-170) but also on Fikellura pottery:E. Walter-Karydi,SamosVI, i, Bonn 1973, p. 23, no. 335, fig. 27 (the rim pattern);G. Schaus, "Two Fikellura Vase Painters,"BSA 81, 1986 (pp. 251-295), p. 271, no. 64, pl. 16:e. Cf. Walter-Karydi,p. 39, note 129, no. 504, fig. 37. The potteryexamplesare to be datedafter 525 B.C. 41 E.g., Buschor, 1957, pp. 8-10, pls. 6:2, 7:1, 2. 42 N. M. Kontoleon,<<'Avao-Ka4albvXl>, FlpaKTLKa1953 (1956; pp. 267-274), p. 268, fig. 1; SimantoniBournias, 1987, fig. 3. 43 ATK, pp. 122, 127, pls. 61:1-2, 62.

PLATE 23

_

a. Chios, Kofinas, fragmentof sima

c. Chios, Kofinas, fragmentof eaves tile

.

b. Chios, Scaramangaarea, fragmentof basin stand

d. Chios, Kofinas, sherd of relief pithos

PLATE 24

a~~~~~~~~~~~~~.

Chios, unknown provenance,fragmentof

basin

_~

-~~

__i12~~~~~~~~~~.

1-, I

I

i

i

L

-I

=

d. Chios, unknown provenance,fragmentof basin SIMANTON-ORIS CIAN RELIE EVA~~~~~~~~~~~~~~~~~~~~~~~~~~

e POTER

ho,ukonpoeac,rmo

ai

AND ARCHITECTURAL L-TERRACTTAS1

L'ATELIER DE POTIER ARCHAIQUE DE PHARI (THASOS) LA PRODUCTION DE TUILES (PLATES

25, 26)

E 1985 A 1988, l'Ecole Franaise d'Athenes et l'Ephorie de Kavala ont fouille en collaborationun atelier de potier a Phari,1un site au Sud-Ouest de l'fle de Thasos, a deux kilometresa l'Est du village modernede Skala Maries. Situe sur un plateau a l'extremite d'une large baie, l'emplacementrepondait a toutes les conditions indispensablesau fonctionnementd'un tel atelier : presence d'argile, de bois, d'eau douce, et d'un debouche maritime (et sans doute terrestre)pour le transport. Les fouilles ont revele, outre les principalesinstallationsliees a la production- fours, appentis de travail, et bassins de decantation(Fig. 1) -, une quinzaine de formesde vases attestant de la diversite de la productionde l'atelier.2Dans les nombreux depots de ceramique, plusieurs fragments de tuiles avaient ete decouverts,mais ce n'est que lors de la deuxieme campagne de fouille qu'un sondage, ouvert dans le secteur Nord-Est du site oCu affleurait l'assise superieure d'un grand mur, confirmal'existence d'une importanteproduction de ce type de materiel. Sous ce mur et s'etendant de chaque cote de celui-ci, on degagea une large fosse remplie de tuiles couranteset de couvre-joints.Ces tuiles n'etaient pas tombees a la suite d'un effondrementmais avaient ete stockeessur place dans l'attente d'une utilisation future. Elles etaient posees a la verticale, en plusieurs rangees : dans la moitie Est, quatre rangs de cinq a huit couvre-joints,l'extremitelarge vers le haut, etaient disposes devant deux rangs de trois a cinq tuiles courantes,la plus grande largeur vers le bas.3Dans la partie Ouest du sondage,deux autres groupesde couvre-jointsetaientrepartis de maniereidentique (P1.25:a). D

' La fouille etait dirigee par Catherine Peristeri (Ephorie de Kavala), Francine Blonde, et l'auteur (Ecole Francaise d'Athenes).Le plan a ete releve par T. Kozelj, les dessins effectuespar V. Anagnostopoulos,et les photos P1. 25:c et 26:e, f par P. Collet. Les rapportspreliminairesdoivent paraitredans les AAA. Un article resumantles principalesdecouvertessera publie dans les Actesde la table rondesur les ateliersde potiers dans le mondegrec aux epoquesge'ome'trique, archafqueet classique(supplementdu BCH a paraitre). 2 Cette ceramiqueimite tantot des formes cycladiques: coupes subgeometriquesa decor lineaire (groupes de cerclesconcentriquesalternantavecdes filets verticaux),coupesbasses avecdecorde rang de points entre les anses, tantot des formesattiques: craterea colonnettes,oenochoes,olpes, lekanes, etc. 3 La disposition des tuiles courantes et des couvre-jointsindique clairement que ces tuiles avaient ete stockees.En effet, si elles etaient tombeesd'un toit, on aurait vraisemblablementdecouverttuiles couranteset couvre-jointsmeles les uns aux autres et fort probablementbrises,mais surtout,les tuiles auraientete trouvees dans le sens inverse, la partie large des couvre-jointset l'extremite etroite des tuiles courantesvers le bas, correspondantau sens dans lequel on les disposaitsur la toiture (cf. Fig. 4).

202

JACQUES Y. PERREAULT

r

A

D7I

T

fati

eFT

CT

E

+

d

+

T

+

f

G y

+

H~~~~~.',1-s \+

3

[..-_e

7~~~~~

1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ::H

H~~~~~~~FG DECITO5

DSTIE

ulspoeatd LeH truelsdu

.Pangnrldie(97

omshbtels

et os patenettue utp aoin4Nu vn atiecornee

faltiere~ ~ ~~~~~~ t eae.L +u tul-erv''

ecurjit

asacn

ul

arcainlcl ~- e ulsd hr efi

4Sur les caracteristiquesdes tuiles laconiennes, cf. 0. Wikander, 1988; Orlandos, p. 81-92; Martin,

p. 65-87.

L'ATELIER DE POTIER ARCHAIQUE DE PHARI: LA PRODUCTION DE TUILES

203

aucun doute: elle a ete demontreenon seulementpar des analyses de 1'argile,5mais de plus on a decouvert, un peu partout sur le site, de nombreux fragments de tuiles surcuites (P1.25:b). Toutefois, l'argileemployeedifferede celle utilisee pour d'autrescategoriesd'objets faconnesdans l'atelier. La compositiondes argiles de Phari varie beaucoupd'un endroit a l'autre du site et les artisansont su parfaitementtirer parti de cette diversite.Alors qu'on employait une argile jaunatre deja bien epuree a l'etat naturel pour la fabricationde la ceramiquefine, on preferait,pour les tuiles, une argile beaucoupplus grossiere,habituellement de couleurrougeatre,mais qui assurait une plus grandesolidite a ces objetslourds. TUILES COURANTES

1 (Fig. 2, P1. 25:d). NOinv. 87.892.02 - tuile fragmentaire, manque une partie de 1'extremitesuperieure, legerementdeformeepar la cuisson. Long. 91,5 cm; plus grande largeur conservee en haut 37 cm; larg. en bas 35,8 cm; ep. 0,2 cm. Argile grise (surcuite), vernis gris/noir sur la face externe concave presque entierementdisparu. Face interne lisse, non vernissee.Extremitesuperieurecomportantun epais rebordvernisse (0,5 cm), extremite inferieure creusee d'un mince sillon horizontal.

2 (Fig. 3, P1. 25:c). NOinv. 86.981.08 - tuile fragmentaire,manque 1'extremiteinferieure.Long. conservee 83,5 cm; larg. en haut 40,5 cm; larg. en bas, a une dizaine de centimetres avant 1'extremite36,5 cm; ep. 0,2 cm. Argile rougeatre micacee avec plusieurs particulesblanchatres,vernis rougeattresur la moitie superieurede la face externe concave,disparu dans la moitie inferieure.Extremitesuperieurecomportantun epais rebordvernis (0,5 cm). Face interne non vernie presentant une serie de stries, horizontales dans le tiers superieur,obliques et en forme de zigzag ailleurs.

Les travaux de restaurationetant a leur debut, il ne nous a ete possible de restituer qu'une seule exemplaire d'une tuile courante complete sur sa longueur (91,5 cm). En revanche, les nombreux fragmentsdes extremites superieureset inferieures fournissentune moyenne de ces dimensions: la plus grande largeur, dans la partie superieure,varle de 39 cm a 41 cm, alors que la plus petite largeur varie de 35 cm a 37 cm. A l'extremite la plus large, les tuiles possedenttoutes un large rebordde 0,4-0,6 cm de haut. Ce ressaut, qui est recouvertpar la tuile superieure, empeche toute infiltrationd'eau sous l'effet du vent.6A l'extremiteinferieure,les tuiles possedentpresquetoutes un mince sillon situe a mi-hauteur du bord sur toute la largeur. Cette marque n'est peut-etre due qu'a une particularitedu moule - on retrouveegalementun sillon sur l'extremitela plus etroitedes couvre-joints , mais elle pourraitaussi influer sur l'ecoulementdes eaux pluviales. En effet, un bord completementverticalne peut empecherl'eau de remonter1egerementle long de la face interne, par capillarite, avant de tomber sur la tuile suivante. L'addition de ce sillon assure que l'eau, en fin de course sur la face externe de la tuile, tombe immediatementsur la tuile suivante (Fig. 4).7 I

Les analyses ont ete faites par M. Maurice Picon, directeurdu Laboratoirede ceramologiede Lyon. Y. Grandjean,Recherchessur l'habitatthasien a l'e'poquegrecque (Etudes thasiennesXII), Paris 1988, p. 391, note 1. I Lors des essais effectues sur certainestuiles courantesne possedantpas ce sillon, on a pu constaterque l'eau remontait le long de la paroi interne sur une plus grande longueur (15 cm en moyenne) que celle du recouvrement(qui devaitetre.autourde 10 cm), alors que sur les tuiles munies de ce sillon, l'eau ne remontait 6

37cm

4

+-8.2cm-

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

I

.

.

.

n

1,

FIG.

2. Tuile courante 1, n? inv. 87.892.02 (ech. 1:10)

~~

11 ~~~~~~~~~~~~~~

FIG.3. Tuile courante2, no inv. 86.981.0

L'ATELIER DE POTIER ARCHAIQUE DE PHARI: LA PRODUCTION DE TUILES

0

1

2

3

4

205

5cm

FIG. 4. Reconstitutionde la position des tuiles sur le toit

Quant a la face interne de la tuile, aucun n'exemplairene presented'incisionhorizontale dans la partie inferieurequi permettraitd'emboitercelle-ci dans le rebordde la partie superieurede la tuile suivante.8Dans la majoritedes cas la face interne est lisse mais quelques exemplairescommele no 2 possedentdes stries en relief, qui devaientassurera la tuile une plus grandestabilite sur la charpente. Enfin, toutes les tuiles courantes sont vernies sur la face externe, mais ce vernis, de mauvaise qualite, a presque entierementdisparu sur la plupart des exemplaireset sa couleur varie beaucoupd'une tuile a l'autre. que 3 a 5 cm. II ne faut pas oublier que meme si, theoriquement,le recouvremententre deux tuiles courantes devrait etre etanche, en pratique il y a toujours un jeu de quelques centimetresqui devait permettre une infiltrationde l'eau, notamments'il y avait du vent. 8 Comme sur le type L3 de 0. Wikander, 1988, p. 210 et fig. 3.

206

JACQUES Y. PERREAULT

COUVRE-JOINTS 3 (Fig. 5, P1.26:a, b). NOinv. 86.981.07 - Presque complet, manque quelques fragments de la partie superieure.Argile rougeatre,presencede mica et de degraissants;vernis brun-rouge sur la face externe. Long. 91 cm; Larg. en haut 21,5 cm; en bas (approximatif) 27,5 cm; haut. de 1'extremitesuperieure 10,5 cm; haut. de 1'extremiteinferieure 14 cm; ep. laterale 1,5 cm. Sur 1'extremitesuperieure,presence d'un large sillon au centre. 4 (Fig. 6, P1.26:c, d). NOinv. 87.986.01 - Completa 1'exception de deux fragments. Argile beige-rougeatre, presence de mica et de degraissants;vernis brun sur la face externe,disparua plusieursendroits. Long. 90,5 cm; larg. en haut 24 cm;en bas (approxi-

matif) 27,5 cm;ep. laterale 1,5 cm;haut. de 1'extremite superieure9,3 cm; haut. de 1'extremiteinferieure 13,8 cm. Sur 1'extremitesuperieure, presence d'un large sillon au centre.Tuile deformeepar la cuisson, legerementcourbedans le sens de la longueur. 5 (P1. 26:e, f). No inv. 86.981.01 - Profil complet, manque plusieurs fragments.Argile rougeatre,presence de mica et de degraissants;vernis brun-rougeatre sur la face externe. Long. 91 cm;larg. en haut 23 cm; en bas 27 cm; ep. laterale 1,5 cm; haut. de l'extremite superieure 9,4 cm; haut. de l'extremite inferieure 14 cm. Sur la tranche superieure, presence d'un large sillon au centre.

La forme des couvre-jointsne se distingue guere du modele laconien habituel. Ils sont presque semi-circulaireset trapezoidaux, plus etroits a 1'extremitesuperieure, puis s'ouvrant progressivementpour atteindre leur largeur maximum a 1'extremiteinferieure. La longueur varie de 90,5 cm a 92,5 cm, la largeur dans la partie superieurede 21,5 cm a 24 cm, alors que la largeur dans la partie inferieure, autour de 27 cm, semble constante. De meme, la hauteur totale est sensiblementla meme d'un exemplaire a l'autre, autour de 10 cm a 1'extremitesuperieureet de 14 cm en partie basse. L'extremiteinferieurepossede un epais rebord (0,3-0,4 cm) qui donne une courbe montante a la face externe de la tuile, permettant ainsi a l'eau qui s'ecoule le long du couvre-jointd'etre dirigee vers les tuiles courantesdisposees de chaque cote. L'extremitesuperieureest toujourscreusee d'un large sillon et tous les couvre-jointssont vernis sur la face externe. DATATION

Nous ne possedons,pour le moment,que des indicationslimitees pour la datationde ces tuiles. Le comblementde la fosse ne contenait, mis a part plusieurs fragmentsd'un grand pithos, qu'un seul fragment de vase, le rebord d'une grande lekane imitant une forme attique du secondquart du ve siecle (P1. 26:g).9Quant aux differentsniveaux correspondant au mur qui fut construitau-dessus de cette fosse, une premiereappreciationdu materieln'a pas fourni d'indicationschronologiquesprobantes.Toutefois, l'ensemblede la ceramique decouvertsur le site indique que la grande periode de productionse situe entre le dernier quart du vIe siecle et le premier quart du siecle suivant, et que l'atelier paralt avoir cesse toute productionpeu apres cette periode. La date indiquee par ce fragmentde lekane s'insere donc bien dans cette chronologieet on peut ainsi, du moins provisoirement,dater le comblementde la fosse de la premieremoitie du ve siecle. 9 NOinv. 86.981.02. Cf. B. A. Sparkeset L. Talcott, The AthenidnAgora,XII, Black and Plain Potteryof the 6th, 5th and 4th CenturiesB.C., Princeton 1970, p. 213.

L'ATELIER DE POTIER ARCHAIQUE DE PHARI: LA PRODUCTION DE TUILES

-

207

21.5cn A-- ---10o.

1

'8m-

A

E

| _ 24cncn

~~~~~~~~~

I~~~ ~ -Q2--

~~ ~ ~

1

f-It 13.8 C

-

-14c-

_FIG.

FIG. 5. Couvre-joint3, no inv. 86.981.07 (ech. 1:10)

6.

Couvre-joint

4,

no

mv.

87.986.O1(&

110.7

FIG.6. Couvre-joint4, n? inv. 87.986.01 (e'ch.1:10)

JACQUES Y. PERREAULT

208

REMARQUES

1 - Meme si les tuiles de Phari s'apparententau type des tuiles laconiennes, quelques particularites,tels les sillons creuses dans l'extremite de la plus petite largeur des tuiles couranteset des couvre-joints,paraissentetre d'originelocale. D'autre part, ces tuiles ayant ete fabriqueessur place et donc a partir d'un meme modelede moule, on n'est pas etonne de constaterque les couvre-jointspossedenttous, a quelques centimetrespres, les memes dimensions,ce qui est egalementle cas des tuiles courantes,et que les deux types ont la meme longueur, soit autour de 92 cm. Ailleurs dans l'ile, seules les tuiles mises au jour dans les fouilles de la Porte du Silene ont fait l'objet d'une etude approfondie.10Si on les compare aux notres, on constateque les dimensionsdes tuiles courantesdes deux sites correspondent a peu pres, mais l'on releve des differencesavec les couvre-joints,la longueur des exemplaires de la Porte du Silene variant de 86 cm a 95 cm, ce qui n'est pas le cas a Phari. 2 - La fabricationde tuiles et de vases dans un meme atelier va a l'encontrede ce qui est generalementadmis. On considereen effet que les ateliers etaient specialisesdans l'une ou 11 L'atelier de Phari demontreque cette regle ne s'appliquaitpas l'autre de ces productions. partout,du moins a la fin de l'epoque archaique,et que dans un atelier qui fabriquaitavant tout de la ceramiqueon pouvait aussi, et manifestementavec une grandemaltrise,fabriquer des tuiles. La presence de tuiles a Phari remet donc en question tout le probleme de la specialisation des ateliers de potiers durant cette periode et on peut se demander s'il ne les fortes faudrait pas etablir une distinctionplus specifique entre les ateliers urbains, ou% agglomerationsfavorisaientla specialisation,et les ateliersregionaux,desservantune population restreinteet donc obliges de diversifierleur productionpour survivre. 3 - On considereegalement qu'il existait dans l'antiquite un type de four pour la ceramique, conpusur un plan circulaire,et un autre, de formerectangulaire,pour la cuissondes tuiles.12Or, les deux fours mis au jour a Phari appartiennentau premier type, piriforme avec une chambrede chauffe circulaire. Le diametrede la sole du plus grand de ces deux fours, ca 2,80 m, qui suffisaitamplementpour accueillirnon seulementde la ceramiquefine mais aussi de grands pithoi, pouvaient sans aucun doute servir a la cuisson d'une grande quantite de tuiles, notammentsi celles-ci etaient disposeesverticalement.13D'autre part, les 10

Y. Grandjean(ci-dessusnote 6), p. 387-392. I Orlandos,p. 92; Martin, p. 78. 12

Martin,p. 78.

Ch. K. Williams, II nous a indique que des experiencesfaites sur la cuisson des tuiles laconiennesont montre que celles-ci, placees a plat dans le four, se brisaientet que la meilleure cuisson etait obtenue lorsque les tuiles etaient placees a la verticale. Un four circulaireetait donc plus propice a la cuisson des tuiles laconiennes, de formecourbe,alors que les fours rectangulairesdevaientmieux s'adapteraux tuiles corinthiennes, plutot carreeset plates. D'autre part, il existe au moins un autre exemple de l'utilisationd'un four circulaire pour la cuisson des tuiles: sur le site de Nemee, trois fours de tuiliers dont un circulaireont ete mis au jour; cf. S. G. Miller, "Excavationsat Nemea, 1975,"Hesperia 45, 1976, p. 186-189; idem, "Excavationsat Nemea, 1977,"Hesperia 47, 1978, p. 80-81. 13

L'ATELIER DE POTIER ARCHAIQUE DE PHARI: LA PRODUCTION DE TUILES

209

fours rectangulairesdestines a la cuisson des tuiles paraissentetre une innovationde la fin du ve siecle,14et par consequenton a donc du^s'accomoderde fours circulairesdurant les periodesanterieures. JACQUES Y. PERREAULT

Directeur INSTITUT

CANADIEN

D'ARCHEOLOGIE

A ATHENES

2B Gennadiou Street GR-1 15 21 Athens, Greece 14 A notre connaissance,les plus anciens fours de tuiliers de forme rectangulairesont ceux degages a Corinthe et a Pella; cf. pour Corinthe: Ancient Corinth,A Guide to the Excavations,6e ed., 1954, p. 85; J. A. 0. Larsen, <
PLATE 25

~~~~~~~~~~~~m - -

b. Fragment de tuile surcuite no 85.250.38

a. Couvre-jointsen place dans la partie Ouest _4; du site. Vue vers le Sud

* Tr-ii c. Tuile couranten? 86.981.08 (2)

d. Tuile courante n? 87.892.02 (1)

PLATE 26

a, b. Couvre-jointno 86.981.07 (3)

c, d. Couvre-jointno 87.986.01 (4)

Frgetderbr

>.~~~~~g

*G novrjon c,~~~~~~~~~arcto

e eaed

.028.14

e, f. Couvre-jointno 86.981.01 (5) . ERAULT:LAEIRD iJCQE S *

locle

8698

OIRACHIU

EPA(HSS

NEW INFORMATION FROM THE DISCOVERY OF AN ARCHAIC TILED ROOF IN IONIA (PLATE 27)

THE COURSE OF THE WORK of the German ArchaeologicalInstitute in Didyma, Turkey, an Archaic temenos was excavatedin 1985 and 1986 on the Sacred Way between Miletos and Didyma. It dates from the secondhalf of the 6th centuryB.C.(Fig. 1).1 IN

D

F

E

Z

G

IL

H

/

10

M

!

L

/

N

Z1

13

1

14

18 wSS'

I _ _ J__

__,__

_

15

_

_. _ _

_ _ _

f

_f

/

19

1188

1/P

t')<16

.

18

16/ oX_,X-j

0----

t 20

lO /

O

10

+

< | ,

'

[

/t

,

z/A

*O /

40

5Om

E

F

TEMENOS/ ~~~~~~~~EIN SSE 1 ANDERHEILIGEN STRA NACH VONMILET DIDYMA/

~~~~~~~~~~Rek
I D

g :/

/r

/'

/;;

j/

/

,

t t

/

G

||H

Z

|

IK

L/

M

N

O

Ql

P P Schneider

/ 1CCordessis

l98

FIG.1. Temenos on the SacredWay to Didyma. Reconstructionof the groundplan (P. Schneiderand C. Cortessis) 1 The excavationsat Didyma are directedby K. Tuchelt, for whose permissionto discussthis material I am

grateful. I would also like to thank W. D. E. Coulson and N. Winter for inviting me to the conferenceon Archaic Greek architecturalterracottas.A preliminarypaper on the excavationsin the temenos and on the archaeologicalfinds will be published in AA 1989.

364 ifv6

t

e

t

164

+o

32

I L-6.101

",

4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

i231LiX ~~~~

5~~~~~~ -tt Zi231

5 0 1 1 0 1 K|

10 i

|

20 i

i

30 i

i

40 i|

50 cm ^19671 M

~

~

-I

5 FH

+ t-

0 +

+

tO -

If----

20 t--tf

-

- 30

,

i

40

FIG. 2. Didyma, Sacred Way, temenos, "Ostbau", tile roof: left, cover tile; right, pan tile (P. Schneider and M

17, }.h--

---14?--

a(- -T

t-

---81t

1'

-----

---17.1

-w3

4 t

t4.0

ST

-47_

s+ ...... p I >

t

-

__

--

___-

uot-

IF---e -+F L

\.~~~~~~~~I

~I Wyteeo

203yaSce

Zi.

tlro:lf,nei;rh,ev

tie(.cnieadM

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~5 15

50

50

5

10

--

0

10

20

3

m

FIG. 3. Didyma, SacredWay, temenos, "Ostbau",tile roof:left, antefix;right, eaves tile (P. Schneiderand M

214

PETER SCHNEIDER

An artificiallyfilled terracelies closeto the SacredWay, on a slope behinda huge retaining wall. It was probablya cult site or "theaterof worship".Its northwestside is borderedby a 23-meter-long,semicircularbase which carriesenthronedmarblefigures.A secondgroupof six winged female sphinxes stands on the terrace wall as if on a high pedestal;this wall defines the limit of the site beside the Sacred Way. On the summit above and behind the semicircularbase emergesa spur-likeblockof rockon the centralaxis of the temenos. In the northeast section of this terrace were discoveredthe ruins of an oikos (Fig. 1, "Ostbau").By measuringthe preservedfoundations,we were able to determinethe ground plan of this rectangularbuilding, which is about 12.15 m. wide and 6.72 m. deep. Its wide, double-dooredentranceopened onto the cult site. About 8000 fragments of tiles were recoveredfrom inside and around the building. Some tiles were still intact, and others could be restored.The large number of fragments makes possible not only observationson the construction,coloring,and manufactureof the tiles, but also an almost completereconstructionof the roof. The tiles are of a fired clay low in mica, which is reddish brown at the breaks. An intense brown-to-red or brown-to-blackpaint was applied to their exposed surfaces. All 8000 fragmentswere examinedand classifiedinto six differenttypes of tiles which comprise the structuralelementsof the roof (PI. 27:a). Judging by the fact that no ridge tiles of a hip roof were found and on the basis of two matchingfragmentsof the raking sima, we come to the conclusionthat the roof was double-sloped.2 The roof is a mixed type of flat "Corinthian"pan tiles and convex"Laconian"covertiles (Fig. 2). The eaves tiles bear on their front side a painted relief guilloche, the antefixes a gorgoneionwith four beardedsnakes (Fig. 3). At a point aboutmidway in their length, their pitchedoutline becomesconvex,a changewhich accentuatestheir plasticity;the same is true of the ridge pan and covertiles at the peak of the roof (Fig. 4). Traces of spatula-like tools can be observedon the undersideof almost all the tile fragments. The process of pressing the clay into a mold producedan uneven lower surface in contrastto the smooth upper sides. For the productionof the plaques of the antefixes only one mold was necessary,as was confirmedby comparingall 33 examples of Gorgonsfound; they appear identical in every detail. The few variations are due to insufficiently filled matricesor to soiled material. The bottomedge of the guilloche is often incomplete,sometimes showing a diagonal cut which indicatesthat the matrix was not accuratelyfilled.3 All the measurementsof the tiles follow a strict metrologicalsystem, which is based on metrical units of inches and yards (24 Archaic Ionic inches = 1 Archaic Ionic yard = 2 Both fragments (Inv. No.

Zi 393a, b) are badly disintegrated.They are lighter than any of the other tile fragments because they contain a different percentage of fired-clay temper. The raking sima can only be reconstructedon analogy with comparablesimas: Le Roy, 1967, pl. 98 (Roofs 9, 10, 11, and 12). 3On analogy with the work of W. Rostoker and E. Gebhard, "The Reproductionof Rooftiles for the ArchaicTemple of Poseidonat Isthmia, Greece,"JFA 8, 1981, pp. 211-227. Carelessnessduring production, even of the ornamented antefixes, seems contradictoryto the precise metrological design of the roof. The deviationfrom the intended"idealform"of the tiles can be explainedby the characteristicsof the material and by the fact that the series of tiles was craftedby hand.

Q__.___-4et

-,

s

X1 26

,

-t@_ _'_ _ _ _ _ _ _6 -

__

_ _

(

,,

041't1253 "4 -

\

22

4 __

-

_

6121

7(4512414

I i

_ _ _

i ..<9

X 94!

X

l

X

SI,2~~~~~~~~~~~~~~~~~I

~ ~ ~~ ~ ~ ~ ~

0,3)~

__ ,(5

0

10

I

+_+-1t-

0-,2

II,._

f^O

30

3

7_

A0

I tFH fi

t

_,L;Z

5t30m

cm30'A

5

t

-

2(23

s

K

t

-

--

-

-j

. 30

-~~

; 6(3

~ ~ ~ ~ ~ ~ ~ ~~

zi 199

o

0

10

20

0

40

50 crn

0

0

10

20

30

40

FIG. 4. Didyma, SacredWay, temenos, "Ostbau",tile roof: left, ridge covertile; right, ridge tile (P. Schneiderand

216

PETER SCHNEIDER

Stirnziegel

Traufzieget

8D Q

|16

4D

D 24 6D

SD

1,Fup 1I&Etle 1/2 Fup

10

0

10

20

8Dl

i

30

16D

24 Daktylen

1Fup

lElie

40

50

1

60cm

FIG.5. Didyma, SacredWay, temenos, "Ostbau",tile roof:metrologicalsystem of the eaves, detail (P. Schneider)

52.5 cm.).4The moduleof this metricalsystem is the number24.5 The measurementfor the horizontalaxis of the tile surface,52.5 cm., correspondsto the length of an Ionic yard;the 24 elements of the braided ornament correspondto the 24 inches (Fig. 5). Furthermore, 16 braidedunits of the eaves tiles, which are two-thirdsof the guilloche,are 16 inches equaling one foot;one third equals the 8 units of the antefix (8 in. = 1/2foot). The lateral height of the antefixes is 4 inches, and the total height is 6 inches, which is three-quartersof the width. The flat eaves tiles are 25 inches long, one inch over the yard grid, while on each side they are an inch less (24

-

2 in. = 22 in.).

The total length of the lateral eaves of the building can be reconstructedthrough the measurementsof the foundationsto which the small size of the overhanginggeison tiles, 0.18 cm., should be addedon each side: 12.15 m. (length of the building) + 2 x 0.18 m. (the overhangof the roof) = 12.51 m. Althoughthis figuredependson an assumption,it is sufficiently accurateto permit calculationof the numberof rows of tiles. Dividing the length by the measurement of the axis shows that the roof had 24 rows of tiles (24 x 0.525 m. = 12.60 m.).

The dimensionsof the elementsare identicalto their number.6 The ornamentedparts of the antefixesand of the eaves tiles are paintedred or black. Of the 33 Gorgon antefixesfound, 10 are black and 19 red;4 fragmentslack any trace of color. The ratio of red to black tiles is exactly 1:2. As this statisticcoversmore than 50%of all the ' Cf. H. Frankel, Dichtung und Philosophie desfriuhen Griechentums2, Munich 1962, p. 314.

is the product of 1 x 2 x 3 x 4. Accordingto Pythagoras'theory, such series of numbers with an inherent mathematicalstructureinclude their AXoyos: ". . . ursprunglichoffenbardas Zahlenbundel, das in einem Ding steckt und das es moglich macht, seine Gestalt zu fassen und daher zu reproduzieren" (K. von Fritz, RE XXIV, col. 199, s.v. Pythagoras). 6 For the three terms "Mass","Gewicht",and "Zahl"and the contributionof rEKroveS to this "Theoriebildung", see F. Preisshofen, "Zur Theoriebildung in Bauplanung und Bautheorie,"Diskussion zur archaeolog.Bauforschung,Berlin 4 (n.d.), pp. 26-30. I remain indebted to F. Preisshofen for many exciting 5 Twenty-four

discussions.

Gebaudemitte 12 Ziegetachsen

It E

IE i

sch

FIG.

ar

1Elte

~ ~

rot

~

|D

t

-_E

t

IE

1E

IE

U

16

6

|

8

rot

__

sctvar-

rot

rot

16

roat

t

ET-

It

-

.112F _____

80

160 I

24Doktylenz lF8 r 1 1 F 16

t

1

rot

t

IE

IE

6

16

schy4rz

rot

6. Didyma, SacredWay, temenos,"Ostbau",tile roof:metrologicalsystemand color schemeof the eaves (P. Sch

218

PETER SCHNEIDER

antefixes (the total number would have originally been 46), we can deduce that the color arrangementwas a sequenceof black/red/red (Fig. 6). Even the preliminaryresults made it apparentthat the lengths of some completelypreservedcovertiles were not uniform. As more tiles were restoredfrom fragments,it became clearthat, amongeighteen mendedcovertiles, the lengthsvary from 36.0 to 49.7 cm., i.e., by more than 13 cm., and that the lengths of the flangeson their back ends differ by 3.6 to 13.5 cm. (Fig. 7). In comparing these individual tiles, it becomes clear that the length of the exposed surfacesdiminishesand the length of the flangesincreases.7This variationmust not be interpretedas resulting from carelessnessduring productionor adjustmentsduring the laying of the tiles. The intentional gradation of the length of exposed tile surfaces and flanges can be easily proven by tool marks. The surface of the flanges was increasedby a sharp-edgedtool following an inch-basedsystem of measurements(P1. 27:b).8The dimensions of these adjustments9can be preciselymeasured:they are 2.2 cm., 4.4 cm., or 6.6 cm., which are multiples of the inch. There are seven differentsizes of covertiles (see Table). Only one example from among the eighteen mendedtiles cannotbe classifiedwith certainty. TABLE:

Didyma, temenoson the SacredWay, "Ostbau".Comparisonof sizes of covertiles of roof

Inv. No. Zi 237, 225, 427, 262, (423) Zi (423), 403, 559, 238 Zi 233, 426 Zi 464, 234, 235, 236 Zi 231 Zi 558, 232 Zi 512

Total length 25 in. 24 in. 23 in. 23 in. 23 in. 23 in. 23 in.

Exposed surface length 23 in. 22 in. 21 in. 20 in. 19 in. 18 in. 17 in.

Flange length 2 in. 2 in. 2 in. 3 in. 4 in. 5 in. (6 in.)

The numberof the rows of tiles fromthe eaves to the ridge of the roof can be determined from the dimensionsof the foundations,which are 6.72 m. wide. If we assume that the roof overhung the walls on either side between 20 and 30 cm. in order to provide protection against rain for the mud brick, then the width of the roof should be 7.2 m. The tiles at the peak of the roof were made to fold over the ridge. If we measure the angle between their inclined sides and the horizontal,340, which correspondsto a gradientof 2:3, we can determine the inclinationof the slopes with sufficientaccuracy(Fig. 4). The frontsof these ridge tiles stood at an angle of 100 to the verticalaxes so that they would meet the covertiles at a right angle. 7Traces of workingleft on some examples show that the surfacesof the flangeswere deliberatelyenlarged, probablyafter firing.That would mean that the covertiles were manufacturedthe same size and adjustedlater accordingto the length required. 8 I would like to thank H. Kienast for his adviceon the tool marks on the fragments. 9 Zi 244, 401, 414, 433, 512, 518, 532, 711.

Stirnziegel

Befund Deckziegel

16D- --

220

250

240

230

23D

Rekonstruktion

+i~~

-240

-

24D

8D

Decklange + Fazlange

0

16D

Gesamtlange

24

E

23W2D 25D

7~~~~~~~~~~~~~~~~~~~~~~~S 22D2D D 20D+3D 212D+2D 24D

230

230

23D

2303

167D 19D+4D 17D+6D 230

18D+5D

-7-_

_R

39D

19D+3D 22D

__D

L_

+

7. Didyma, SacredWay, temenos, "Ostbau",tile roof:gradation ~~~FIG. to their lengths;cross-sectionof the roof (P. Schneiderand C. C

220

PETER SCHNEIDER

All this informationpermits us to concludethat the cover tiles were set 100 from the horizontal.The inclinationof the roof shouldbe 40 steeperthan that of everyindividualtile, which would mean an angle of 140 and a gradientof 1:4. If we arrangethe covertiles accordingto their lengths, combiningthem with the eaves and ridge tiles at the 140 angle, we can reconstructthe entire roof. Its width is now 7.30 m. (Fig. 7),10which correspondsperfectlyto our calculatedminimum width of 7.20 m. If another row of tiles whose exposed surfacemeasuresbetween 36 and 50 cm. were added,the roof would have an overlap of 66 to 100 cm.; in terms of construction,this length is very problematic. Leaving out a row would mean that the roof would be only 6.40 m. wide, certainly too small to fit the oikos. The exact inclination of the roof is irrelevant to the definitionof the numberof the rows. The graduated arrangementof the cover tiles accordingto length carries over to the exposed surfaceof the pan tiles, becausethese interconnectwith the upper ends of the cover tiles. The pan tiles themselves do not show any differentationin size: they are 22 inches wide and 25 inches long. Therefore the degree of overlappingon individual rows of tiles must be differentlyreconstructed. In a preliminarycross-sectionof the roof incorporatingthe seven differentsizes of tiles, we arrangedthem accordingto their lengths and in such a way that the smallest tiles were assignedto the upper rows and the largest to the lower. The result was surprisinglylogical. Although in the first three rows the visible surface of the antefixes and cover tiles ranges from 22 inches to 25 inches, when combined with the individual measurementsof their coveredparts, the visible surfaceof the eaves and pan tiles remainsthe same:22 inches each time. In the fourth and following rows the covertiles, as they graduallybecomeshorterand shorter,influencethe dimensionsof the pan tiles. This structurallyand metrologicallyvery complicatedarrangementends with the shortestcovertile that has the longest flange of six inches. Some latitude is obviouslya necessity,since the final length of the roof is only determined during the process of laying the tiles on the rafters. The sequence in length of the covertiles from the eaves to the ridge is 19 inches, 23, 22, 21, 20, 19, 18, 17; that of the pan tiles, 3 x 22, 2 x 21, then 20, 19, 18, and, last, 17 inches. The visual impression of the gradually decreasinghorizontal stripes which articulate the sloping roof, which an observernotices only after it has been drawn to his attention, makes the distanceto the peak of the roof seem longer, and is accentuatedby the plastically projectingline of the ridge tiles (Fig. 8). To achievethis effect must have been the intention of this arrangement.The spectator sees the surfaces and outlines of the roof from a disadvantageousangle in comparisonto the rest of the building, which lies more or less within a suitable angle of view.1"This distortion,therefore,which is due to "true"dimensions,is 10 The sum of the individual lengths of the tiles from the eaves to the ridge is 169 inches = 3.70 m. This dimension must be reducedby the inclination factor of the tiles and multiplied by two. This product is the width of the roof projectedon the ground plan: 3.70 m. x coslO? x 2 = 3.64 m. x 2 = 7.28 m. l' Cf. Vitruvius (111.5.13). By this fact he explains the forward inclination of all architecturalelements above the capitals.

~~~~~~~~~~~

FIG.

tile roof:reconstruction (P. SchneiderandC. Cor 8. Didyma,SacredWay,temenos,"Ostbau",

222

PETER SCHNEIDER

correctedby using tiles of different length,12which gradually become shorter and shorter fromthe eaves to the ridge of the roof. In such a cross-section,no straightlines are produced, but instead a slightly oscillatingcurve results, like an interconnectingline of corresponding points (Fig. 7). This fine arrangementof tile lengths can be interpretedas an arithmetical definitionof curves.t3It is comparableto a staircasein which the height of each step remains the same while at the same time the width of each tread decreasesupwards. Still, the first three rows of pan tiles form a straight line 3 x 22 inches, and as the cover lengths start to decrease(21, 20, 19, 18, . . . in.) it becomesa flat curve. The resulting curvatureof the cross-sectionwas a matterof only a few centimetersand was probablynot a noticeablefeatureof the roof. The conceptionand metrologyof the roof, like its construction,plasticity, and colorfulness,testify to an extraordinaryunity of form. The hithertounknown differentiationof the plastic surfacesof an Archaic Ionian roof goes beyond the functional requirementsof a roof. Although its form is based on a strictly rational metricalsystem, the roof approachesthe quality of a free, organicstructure.It is this law of orderthat characterizesevery single elementof an ArchaicIonian building. PETER SCHNEIDER Dragonerstrasse9 D-7500 Karlsruhe21 Federal Republic of Germany 12 Cf. H. Busing, "OptischeKorrekturenund Propylaen-Fronten," JdI 99, 1984 (pp. 27-73), p. 31: "Den Korrekturen... fallt lediglich die Aufgabe zu, die idealiter konzipierte Form wirklich wahrnehmbar zu machen, so das sich dem Auge die unverfalschteeurythmiadarbietet." 13 Cf. A. Petronotis,Zum Problemder Bauzeichnungenbei den Griechen,Athens 1972, p. 6, note 27, with further bibliography. B. Gunn, "An Architect's Diagram of the Third Dynasty," Annales du Service des Antiquite'sde l'Egypte 26, 1926, p. 197, pl. 26.

PLATE 27 4

-

-

.

i

'

a. Preliminary reconstructionwith the tile fragmentsin 1985. (Photograph W. Schiele)

b. Traces of working on the flanges of the cover tiles. (PhotographW. Schiele) Didyma, Temenos on the Sacred Way, Tile

Roofof the "Ostbau"

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA (PLATES

28-32)

ER UBERLIEFERUNG nach im 7. Jh. v. Chr. von Milet aus als alteste griechische Siedlung im Schwarzmeergebietgegrtindet,1kamen in Histria bei entsprechendtief reichendenGrabungenauch DachterrakottenarchaischerZeit ans Licht. Auf einige Proben hatte bereits Xkerstr6m2hingewiesen;wenig spater konnte Theodorescu3weiteres, bis zum Jahre 1960 gefundenes Material bekanntmachen.Fast alle diese Fundstucke entstammender Tempelzone von Histria (Fig. 1), wo bis vor etwa zehn Jahren systematische Ausgrabungendurchgeftihrtworden sind.4 Ohne dass dabei das stadtische Kultzentrum vollstandig freigelegt wurde, konnten hier zwei dem Zeus Polieus und der Aphrodite geweihte Tempel, dazugehorigeAltare, moglicherweiseein Propylonund eine Vielzahl von Basen fur verschiedeneVotive aufgedecktwerden.Alle diese Monumentezeigen zusammen mit dem iibrigen Fundmaterial,dass dieser Bereichvom 6. bis 1. Jh. v. Chr. als Heiligtum genutzt, ein Teil seiner Bauten dabei immerwieder erneuertwordenist. Speziell die innerhalb und ausserhalbdes Aphroditetempelsangelegten,bis auf den gewachsenenBodenherabgefuihrtenSondagenhaben weitgehendeKlarheituber die stratigraphischenVerhaltnisse und die damit korrespondierendeBaugeschichtedieses Tempels und der anderen Bauten, aber auch einen erheblichenZuwachs an Dachterrakottenerbracht.Dabei ist festzustellen, dass die der archaischen Epoche angehorendenStuickeheute nicht nur in ihrer Menge, sondernauch nach der Vielfalt vertretenerTypen den Hauptteil ausmachen.5 Hinsichtlich der Fundsituationist auf eine Besonderheitaufmerksamzu machen: Bei der Tiefgrabungim Pronaosdes zu Beginn der zweiten Halfte des 6. Jh. v. Chr. errichteten Aphroditetempelsstiess man auf eine Fuillschicht,die ausserordentlichviel archaischesMaterial, darunter auch zahlreiche Fragmente von Dachterrakottenenthielt (P1. 28:a). Als spatestensin den Jahren um 500 v. Chr. nicht nur dieser spatarchaischeTempel, sondern I Abkiirzungsverzeichnis: = Histria. MonografiearheologicaI, II, Bucureqti1954, 1966 Histria Pippidi, 1962 = D. M. Pippidi, "Gli scavi nella zona sacradi Histria,"Dacia, N.F. 6, 1962, S. 139-1 56 Theodorescu, 1970 = D. Theodorescu,"Notes histriennes,"RA 1970, 1, S. 29-48 Zur Griindungsgeschichteund archaischenPeriodevon Histria vgl. zuletzt P. Alexandrescu,"Histriain epoca arhaica,"Pontica 18, 1985, S. 41-53; ders., "Histria in archaischerZeit,"Xenia. KonstanzerAlthistorische Vortrageund Forschungen,W. Schuller, Hrsg. (im Druck). 2ATK, S. 3-4. 3 Theodorescu, 1970, S. 29-43. 4 Vgl. K. Zimmermann,"Ausgrabungen in der Tempelzonevon Histria,"Ethnographisch-Archaiologische Zeitschrift22, 1981, S. 453-467. 5Vollstandige Bearbeitung der Dachterrakottenvon Histria durch K. Zimmermann, Dachterrakotten griechischerZeit aus Histria. Untersuchungenzur Typologie,Datierung und VerbreitungkeramischerBauelemente im Schwarzmeergebiet,Dissertation B (Habilitation) Berlin, Humboldt-Universitat,1983 (ungedruckt). Vgl. thesenartige Zusammenfassung unter dem gleichen Titel, Ethnographisch-Archdologische Zeitschrift27, 1986, S. 229-239; ferner ders., "Zu den DachterrakottengriechischerZeit aus Histria," in Xenia (Anm. 1 oben).

224

A

j~~~

I

Us(l 97 D5

I/o

--

I/oc(1973)

|

1 976

I/0( 1973-75

1/

FIG.

i

j4 >dl

)-~

-6(1 974/5)-t

3

1. Plan der griechischenTempelzone von Histria (SektorT) mit schematischerAngabe der 1973-1976 im Bereich des Aphroditetempelsausgefuhrten Schnitte (A = Zeustempel, C = vermutliches Propylon, I = Aphroditetempelauf PodiumJ)

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA

225

auch die uibrigenBauten der Tempelzone tiefgreifendzerst6rt wurden, legte man offensichtlichvor dem Wiederaufbauals Ausgleich diese sogenanntepostarchaischeFiillung an. Sie findet sich uiberallin der Tempelzone und hat wohl das bei der Zerst6rungangefallene Material insgesamt aufgenommen. Es kommt hinzu, dass die im Pronaos erschienenen Dachterrakottensich dort auch nicht in Sturzlagebefundenhaben konnen. Denn drei hier in einer Reihe und aufrecht stehend angetroffeneFlachziegel unterstreichen,dass hinter ihrer Anordnung eine Absicht steht. Vielleicht handelt es sich um eine Abgrenzung;die kompakte Masse davor befindlicher Fragmente von Dachterrakottenverschiedener,von mehr als einem Gebaude stammenderTypen lasst zusammen mit den anderen Fundkomplexen an ein beabsichtigtesNiederlegendieserwenn nun auch zerstorten,so doch ehemals im Heiligtum verwendetenTeile, an eine Art "kultischenBegrabnisses"denken. Die in jener FiillschichtgefundenenDachterrakottenstammeninsgesamtaus der zweiten Halfte des 6. Jh. v. Chr. Das ist die Zeitspanne, in der die histrianischeTempelzone ihre erste monumentaleAusgestaltungerfahrenhat. Ein Teil dieses Materials kann durch typologische Beobachtungenund bei Stuickenmit Schmuckelementen(Bemalung, Reliefleisten) auch durch stilistische Erwagungen zeitlich sogar noch genauer eingegrenztwerden. Von dieser Gruppe archaischerDachterrakottenist eine zweite abzuheben, die noch vor Errichtungdes Aphroditetempels,also vor 550 v. Chr. entstandensein muss, weil sich einzelne Beispiele in den Fundamentenjenes Tempels wiederverwendetbzw. als bereits "verworfenes"Material in den der Errichtung dieses Baues gleichzeitigen oder vorangehenden Schichtengefunden haben. Sie gehoren allgemein in die erste Halfte des 6. Jh. v. Chr. Denn die fruihestenin Histria gefundenen geschlossenenKomplexe an Gefasskeramik6stammenohnehin erst aus den Jahren um 600 v. Chr.; und selbst fur die Mutterstadt Milet sind bisher keine alteren Ziegeldacherbekannt.7Diese friihe Gruppevon Dachterrakotten belegt Dacheindeckungenund demzufolge die Existenz von Gebauden, von denen sich sonst-aus welchen Grunden auch immer-bis heute keine weiteren baulichen Reste haben nachweisenlassen. Die demnach in zwei zeitliche Gruppen zu scheidendenarchaischenDachterrakotten von Histria sollen hierjedochals Typen bzw. Einzelstuickevorgestelltwerden. Ihr uberwiegend schlechterErhaltungszustanderschwertnicht selten deren Rekonstruktionund exakte Positionsbestimmungam Dach; eine Zusammenstellungganzer Dacher und deren Zuweisung an bestimmte Gebaude bleibt vorerst stark hypothetischbelastet, weshalb derartige bleiben. Fragen im hier beabsichtigtenUberblickbesser unberuicksichtigt Vom FundortHistria waren bisher nur einige unscheinbareFragmentevon Flach- und Deckziegelnbekannt,8aus denen der Dachbelageines Gebaudesja im wesentlichenbesteht. 6 FundkomplexeHis. 1981 T. (6) + (7): enthalten u. a. Fragmentezweier Teller mit hohem Fuss (fruitstand), Nordionischesdes Middle Wild Goat style, Mittelkorinthischesund ostgriechischeSchalen der Lambrino-Gattung. 7 Der gegen 620 v. Chr. errichtetealteste Athenatempelscheint noch mit einem Flachdach ausgestattet gewesen zu sein, wie A. Mallwitz und W. Schiering ('Der alte Athena-Tempel von Milet," IstMitt 18, 1968 [1970; S. 87-160], S. 125) darausschliessen,dass im Umfeld dieses Tempels aus Poroskeine Dachziegelfunde gemacht wurden. Aus ahnlichen Grunden nimmt auch Xkerstr6m(ATK, S. 205) fur den um 600 v. Chr. zerstortenTempel von Altsmyrnaein flaches Lehmdachan. 8 Vgl. D. M. Pippidi, in Histria I, S. 260, Abb. 106; Theodorescu, 1970, S. 31, Abb. 3:a, b.

226

KONRAD ZIMMERMANN

FI

F.I

FY

FYJ

FIG.2. Typen in Histria gefundenerarchaischerFlachziegel

Hier bieten die neueren Funde im Gebiet des Aphroditetempelsund unbearbeitetealtere Bestandemit dem summarischenFundvermerkTempelzone den Ansatz zu einer typologischen Systematisierung. Unter den histrianischenFlachziegelnist die konkavgeschwungeneForm mit muldenformigemQuerschnittnicht belegt. Vielmehr kommtausschliesslichdie korinthischeForm, namlich die flache, rechteckigePlatte mit angehobenenSeitenrandernvor. Nur zwei solcher Flachziegel sind nahezu komplett erhalten (P1. 28:b).9 Als Bestandteile der sogenannten postarchaischenFullung stammen sie den Fundumstandennach aus der zweiten Halfte des 6. Jh. v. Chr. Bei annaherndgleichen Abmessungen-namlich als abgerundete Grundmasseeinmal 62 x 51 cm, zum anderen 67 x 55 cm, aber fast identischerHohe des Seitenrandesvon 6.3-6.6 cm-vertreten sie zwei leicht variierendeTypen. Als typenbildendes Kriteriumhat die Zurichtungim Bereichder oberenSchmalseitezu gelten (Fig. 2). Danach sind solche mit glatter Oberflache (Typ F.I; P1. 28:b oben), solche mit rillenformiger Einsenkung (Typ F.IV; P1. 28:b unten) und solche mit aufgesetztem Wulst (Typen F.V-VI) zu unterscheiden;letzterer kann direkt an die obere Schmalseiteherangeschoben (Typ F.V), aber auch leicht abgerucktsein (Typ F.VI). Die unteren Schmalseiten zeigen regelmassigeine herabhangende,dabei entweder sichelformiggeschwungene oder mehr gerade gefuhrte Stirn, sind aber aus konstruktivenGrunden weniger differenziert, wie die fast einheitlich angelegte und als Widerlager dienende Aussparung an der Unterseite beweist. Zwischen den kompletten Vertretern beider Typen sowie den ihnen jeweils anzuschliessendenFragmentengibt es neben den geringfuigigenAbweichungenin der ausseren Form und den Massen auch gewisse Unterschiedeim Ton und in der Farbgebung. Der "a la brosse" gestrichene Farbuberzug spielt beim erstgenanntenTyp von verbreitetemWeinrot uber Violett- und Brauntonebis zu gelegentlichemSchwarz, kennt aber auch die elfenbeinfarbeneEngobe, die fur den zweiten vorgestelltenTyp verbindlich zu sein scheint. Alle diese bestehendenUnterschiedeschliessenaus, dass beide Typen zum gleichen Dach gehorthaben, was angesichtsder in der Mitte des 6. Jh. v. Chr. einsetzenden regen Bautatigkeitin der Tempelzone von Histria auch nicht notwendigerscheint. 9 Oben Typ F.I, unten Typ F.IV; beide Fundinventar:His.1976 T (18).

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA

D.T

D.7

227

LXf

FIG.3. Typen in Histria gefundenerarchaischerDeckziegel

Gleiches gilt ftir die histrianischenDeckziegel, wo ebenfalls mehrere, allerdings noch deutlicherunterschiedeneGrundtypenvorkommen(Fig. 3). Zuerst sei die lakonischeForm mit halbrundem Querschnitt, halbmondformigerStirn unten und meist abgebrochener oberer Muffe genannt (Typ D.I). Davon sind zwei Exemplare fast komplett (PI. 29:a [b, c])10und viele Fragmente(P1.28:c)1Ierhalten,die unter mehrerenGesichtspunkteneine in sich geschlosseneGruppe bilden, deshalb sicherlichauch zu einem Dach gehort haben und nach den Fundzusammenhangenwiederum in die zweite Halfte des 6. Jh. v. Chr. datiert werden konnen. Ihre unbedecktenOberflachensind in den Farben Rot, Hell- bis Dunkelbraun und Schwarz "a la brosse"gestrichen.Sie erreichenbei Abzug der unter den nachstfolgenden Kalypter verschwindendenMuffe eine "Nutzlange"von ca. 78 cm, wegen ihrer konischenForm eine zwischen oben ca. 22 und unten ca. 24-26 cm variierendeBreite sowie eine in gleicher Richtung von ca. 10 auf 12 cm zunehmende Scheitelhohe. Bruchstucke eines lakonischenDeckziegeltypserheblichgr6ssererDimensionen (Typ D.II; P1.29:a) mit nur einer bewahrten Schmalseite in einfachem Ziegelprofil konnten wegen der sonstigen Ubereinstimmungenmit dem vorgenanntenTyp als Firstkalypteredes gleichen Daches in Anspruch genommen werden.12Die eckige, d. h. korinthischeForm histrianischerDeckziegel mit annahernd senkrechtenSeiten und einem Satteldach ahnlichen Schragflachen daruiberweist an der oberen SchmalseiteZiegelprofil und seitliche Aussparungen,an der Gegenseiteeine giebelf6rmigeStirn auf (Typ D.VII; P1.29:b).13Kein Stuckdieses Typs ist Von links nach rechts: a = Typ D.V, Fundinventar:His.1976 T (18); b = Typ D.I, Fundinventar: His.1975 T (12) + 1976 T (77); c = Typ D.I, Fundinventar:His.1976 T (18); d = Typ D.II, Fundinventar: His.1976 T (12), (15) + (18). 1 Fundinventar:His.1976 T (77). 12 Es mag mit dem fragmentarischenErhaltungszustandzusammenhangen,wenn halbrundeseitliche Einschnitte zur Aufnahme der die DachschragenhinablaufendenKalypterreihenfehlen, falls diese uiberhaupt vorhandenwaren. 13 Fundinventar:Histria, Alte Funde, 1927-1942. 10

228

KONRAD ZIMMERMANN

kompletterhalten.Dennoch sind die FragmentedurchZurichtung,einheitlicheTonzusammensetzung,elfenbeinfarbenebis gelblicheEngobeund vermutlichgleichen Fundplatz-es handelt sich meist um sogenannte Alte Funde-zu einer Gruppe zusammengeschlossen, deren allgemein archaische Entstehungszeit gesichert ist. Zu den beiden vorgestellten Grundtypen kommt eine diese beiden Moglichkeiten kombinierendeMischform hinzu, deren senkrechteSeitenflachendurch einen gewolbten Rucken als Dachflache verbunden werden (Typ D.V; P1. 29:a [a], c). Von dieser Kalypter-Mischformkonnte ein Exemplar wieder vervollstandigtwerden;es zeigt unten eine halbrundeStirn, oben das einfache Ziegelprofilund nahebei im Scheitelein Nagelloch.14Charakteristischin Ton und ockerfarbenem UiYberzug, misst der sich von unten nach oben leicht verjuingendeDeckziegel max. x 65 20 cm und gehort den Fundumstandennach wiederum in die zweite Halfte des 6. Jh. v. Chr. Bisher fur eine singulare Mischform gehalten, zeigt das kurzlich ver6ffentlichte, derselben Jahrhunderthalfte zugewiesene Dach15 von einem Gebaude an der Heiligen Strasse zwischen Milet und Didyma bei offenbar geringeren Abmessungenden-wie es scheint-vollkommen gleichen Typ. Obwohl Tonzusammensetzung und Farbuberzug nicht tibereinstimmen,bleibt die milesische Herkunft dieses auch in Histria vertretenen Typs nicht ausgeschlossen. An den eckigen Deckziegeltyp (Typ D.VII) kann als Einzelstuck das Fragment eines sattelf6rmigen Firstkalypters mit Loch an oberster Stelle zur Aufnahme eines Firstschmuckesangeschlossenwerden (P1. 29:d, e).16Dieses Stuck zeichnet sich dartiberhinaus noch durch eine Bustrophedon-Weihinschriftan Aphrodite aus: 'A4pobIirrjt a'VeO)KEV ex.... Der altertumliche Schriftcharakterwie die Fundzusammenhangeweisen auf die Zeit vor 550 v. Chr. hin. Demnach hat es in der Tempelzone von Histria bereits in der ersten Halfte des 6. Jh. v. Chr. ein der Aphroditegeweihtes Kultgebaudegegeben. Einige stratigraphischebenfallsvor 550 v. Chr. anzusetzendeFlachziegelfragmente(Typen F.Va, F.VIa) sprechendeutlich von einer Bautatigkeitin jener Zeit, obwohl sich-wie schon angedeutet-daftir bisher keinerlei andereSpuren haben nachweisenlassen. Zu dieser fruhen Gruppe histrianischer Dachterrakottensind schliesslich noch die Uberresteeiner Serievon Traufziegeln (Typ T.I; P1.30:a)17zu zahlen. Denn die drei davon erhaltenen, in Zurichtung, Ton, Farbgebung und Ornamentik einen geschlossenenTyp bildendenFragmentestammenubereinstimmendaus der Fundschicht,die dem in der Mitte des 6. Jhr. v. Chr. errichtetenAphroditetempelzeitlich vorausgeht.Entsprechendder Position dieserGattungvon Dachterrakottenam Traufrandnimmt der ruickwartigeFlachziegel 14 Sowohl zu dieser Misch- wie zur lakonischenKalypterformsind in Histria Deckziegelfragmenteaufgetaucht, die im Inneren einen Querriegel aufweisen (Typen D.III und D.VI). 15 P. Schneider, "Aus einem archaischen Bezirk an der Heiligen Strasse von Milet nach Didyma: Das Tonziegeldach des Ostbaues,"Bericht iuberdie 34. TagungfiurAusgrabungswissenschaft und Bauforschung vom 7.-i1. Mai 1986 in Venedig (Koldewey-Gesellschaft),Bonn 1988, S. 16-19, Abb. 3, 4; vgl. ders., "Zur Topographie der Heiligen Strassevon Milet nach Didyma,"AA (JdI 102), 1987 (S. 102-129), S. 106, 108, Abb. 5. [S. P. Schneider,S. 211-222 oben.] 16 Fundinventar:His.1974 T ([obiecte]58) + 1976 T (18). 17 Fundinventar:His.1974 T ([obiecte]56).

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA

229

die Dachschrage auf, wahrend der abgebogene Teil mit seiner kuinstlerischgestalteten "Ansichtsseite"senkrechtherabhangt;beide Flachen treffen also stumpfwinkligaufeinander, und die Front bildet unten eine Wassernaseaus. Bei ihrer vorrangigenAufgabe, herabstr6mendesWasser gefahrlos fur die holzernen Teile darunter abzuleiten, sind diese beiden Flachen mit einem hier elfenbeinfarbenenUberzug versehen. Die sichtbare,durchschnittlich 7.5 cm hohe Front wird von einem Weinrot und Braunschwarzausgeftihrten Flachenornamentgeschmuickt,fur das auch geritzte Vorzeichnungen zu erkennen sind: Die einfachste Form des Maanders lIsst Rechteckfelderentstehen, die im Wechsel von Schachbrettmusterunten und achtteiligem Stern oben ausgefuilltwerden.18Keiner dieser Traufziegel ist kompletterhalten, doch bewahrt ein Fragmenteine Ecke, uber der die Ziegeloberflacheleicht ansteigt. Die Frontbreitewird vermutlichaus einem Mehrfachen des aus jeweils einer Stern- und Schachbrettfuillungbestehenden Ornamentpaaresvon ca. 15.5 cm Breite zu berechnensein. Wegen einer gewissen ausseren Gleichartigkeitist ein weiterer Traufziegeltyp (Typ T.II; P1.30:b, c)19unmittelbaranzuschliessen.Seine drei fragmentarischerhaltenenStuicke wurden-soweit nachweisbar-auf dem Plateau von Histria ausgegraben.Hier lag in archaischer Zeit ein Teil der Zivilsiedlung, innerhalb derer ein Gebaude mit solchem Schmuck am Dachrand besonders hervorgestochenhaben mag. Untereinander in allen wichtigen Merkmalen uibereinstimmend,gestattet der beklagenswerteErhaltungszustand dieser drei Fragmente nur wenige gesicherteAngaben zum Typ. So unterscheideter sich vom vorangegangenenvor allem durch die breite, von einer Mittelkerbein zwei Stege aufgeteilte Unterseite der Front, die auf knapp 7 cm Hohe zu erganzen ist, und durch deren auf elfenbeinfarbenemGrund erscheinenden,heute violettbraunenDekor. Akkuratausgeftihrt, handelt es sich bei diesem um versetztangeordnete,jeweils an die Ober- und Unterkante geruckte Rechtecke; diese bestehen aus einem umlaufenden Streifen mit inneren Fortsatzen,wobei urspruinglichdiejenigennahe der Unterkantegeschlossen,die an der gegenuberliegendenKante dagegen nach oben offen erscheinen. In vielen anderen Punkten wie Teilen der Zurichtung, Anheben der Ziegeloberflacheim Bereich der Ecken, Verteilung und Farbe des Uberzuges ahnelt dieser Typ dem zuvor behandeltenauffallig. Eine aus mehrenenFragmentenkomplettzusammengesetzteTonplatte mit plastischer Ornamentleistean einer dadurchals Front charakterisiertenLangseite muss eine ahnliche Position wie die Traufziegel eingenommenhaben (Typ T.III; P1.30:d-f).20Sie war ihrerseits jedoch ganz bedeckt,denn nur die Kombinationaus Eierstaboben und Astragal unten '18 Zu einem weitgehend ahnlichen Ornamentpaarauf Dachterrakotta-Fragmenten aus Olbia vgl. A. S. Rusjaeva, "Archaiceskajaarchitekturnajaterrakottaiz O1'vii,"in Antic'nyedrevnostiSevernogoPricternomor'ja,V. A. Anochin, Hsrg., Kiev 1988 (S. 33-51), S. 40, Nr. 25-28, Abb. 7. 19 Fundinventar:His.1952 X-Z 43/44 = Generalinventar:V5760. V. Canaracheund S. Dimitriu, Histria I, S. 204. Abb. 73 oben; M. Coja, "Activitateamestesugareascala Histria in sec. VI-I in.e.n.,"Studii ji cercetari de istorieveche 13, 1962 (S. 19-46), S. 23, Abb. 3:3; dies., "L'artisanata Histria du VIe au ler siele avant notre ere,"Dacia N.F. 6, 1962 (S. 115-138), S. 120, Abb. 3:3; S. Dimitriu, in Histria II, S. 482 = Taf. 50:D (ausser in Histria I stets um 1800 zu drehen!). 20 Fundinventar:His.1976 T (70).

230

KONRAD ZIMMERMANN

zeigt SpureneinstigerBemalung;sonst fehlt selbstan der OberseitejederUberzug. Stattdessen finden sich hier zwei original erhaltene, stark gerosteteEisenstifte und biegt die rickwartige Langseiteebenfalls,aberohne Schmuck,nachunten um. Platten dieserArt scheinen etwas von oben her, also horizontalabgedecktzu haben, verlangenihrerseitsaber ebenfalls nach einem daruberliegendenBauglied. Dienten solche Platten m6glicherweiseals Mauerbekr6nung,so stellen sie zwischen aufgehenderWand und Dachbelag ein vermutetesZwischenglieddar, dessen allein sichtbareFront einen Schmuckaufweist. Diese ca. 63 x 40 cm messendePlatte war bei den Grabungenim Pronaosdes Aphroditetempelsso wie die beiden komplettenFlachziegel aufrechtstehendin der sogenanntenpostarchaischenFullung angetroffen worden (P1. 28:a). Sie geh6rt deshalb aus stratigraphischenGrunden in die zweite Halfte des 6. Jh. v. Chr., wobei die stilistischeBeurteilungspeziell des Eierstabeseine Datierung an den AnfangdieserZeitspannenahelegt.Theodorescuhat bereitsdas Fragmenteiner weiteren solchen, ebenfalls in der Tempelzone gefundenen und in den Details fast massgleichen Platte ver6ffentlicht2tund als "moglichenTeil eines Geisons"angesprochen. Ein von ihm ausserdem vorgestelltesweiteres archaisches Dachziegelfragment(Typ T.IV),22 allerdingsnur mit Perlstaban der Front, kann hingegen kaum als Teil einer Sima angesprochenwerden, sondern mag-bei aller damit verbundenenProblematik-einem ahnlichenVerwendungszweckgedient haben wie die hier neu eingeftihrtAbdeckplatte. Das kleine Fragmenteiner schmalen,nur knapp 6 cm hohenSima (Typ S.IV;P1.31:c)23 bewahrt in der mehrgliedrigdekoriertenFront, ihrer Neigung nach vorn, dem rtickwartig geschwungenenUbergang zu dem als Auflager dienendenFlachziegel und dem auch diese wetterseitigen Partien bedeckenden,hier elfenbeinfarbenenUberzug alle Merkmale, die diese Gattung von Dachterrakottencharakterisieren.Die oberen Dreiviertel der Front werden von einem aus dunkelbraun gehaltenen Zickzackstreifenbestehenden Flachenornament, das auf den abgerundetenoberen Rand uibergreift,eingenommen;den unteren Abschlussbildet ein Rundstabmit Feldereinteilung.Die Fundsituationdatiert dieses Einzelsttickwieder in die zweite Halfte des 6. Jh. v. Chr. Ein solcherzeitlicherAnsatz erscheint auch unter dem Gesichtspunktgerechtfertigt,dass aus ostgriechischemBereichbisher keine Sima bekanntgewordenist, die sichervor der Jahrhundertmitteentstandenware. Als prachtigstes Sttick unter den archaischen Dachterrakottenvon Histria hat eine weitere, aus mehreren nicht aneinanderpassendenFragmenten und deshalb auch nicht vollstandig erschlosseneSima zu gelten (Typ S.1; Pl. 31:a, b).24Ihre Hohe kann auf etwa 32 cm berechnetwerden. Bereitsvon Theodorescupubliziert,25weist auch sie die fur Simen 21 D. Theodorescu, "Un fragment architecturalceramic, cu ove, la Histria," Studii ji cercetaride istorie veche 16, 1965, S. 163-168, Abb. 1; Theodorescu, 1970, S. 32-35, Abb. 4, 5. 22 Fundinventar:His.1952 X-Z 43/44. Canarache und Dimitriu, in Histria I, S. 205, Abb. 74; Coja, "Activitatea"(Anm. 19 oben), S. 23, Abb. 3:2 bzw., "L'artisanat"(Anm. 19 oben), S. 120, Abb. 3:2; Theodorescu, 1970, S. 34-35, Abb. 6, 7 (ausser in Histria I stets um 1800 zu drehen!). 23 Fundinventar:His.1975 T (13). 24 Fundinventar:His.1960 T. Pippidi, 1962, S. 153-154, Abb. 15; G. Bordenacheet al., "*antierularheologic Histria,"Materiale ji cercetariarheologice9, 1970 (S. 177-223), S. 180-181. 25 D. Theodorescu, "Remarquessur la compositionet la chronologiedu kymation ionique suscitees par

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA

231

typische Zurichtung auf. Ein neu zugeordnetesFragment26sichert jetzt sogar den ruckseitig gekrtimmtenUbergang zum Auflager. Die einheitlich elfenbeinfarbengrundierte Front wird durch mehrere, zusatzlich Weinrot und Braunschwarz verwendende Ornamentstreifengegliedert,von oben beginnend:plastischerPerlstab,auf eine Flache gemalter doppelter Maander mit eingestreutenSternen, plastischer Eierstab und unbekanntesunteres Abschlussglied.Die nahe Verwandtschaftin Zurichtung,allgemeinerAnordnungder aus Schmuckglieder,Teilen des Ornamentsund dessen Farbigkeitzur Kalabaktepe-Sima27 Milet fallt sofortauf; das histrianischeExemplar weist jedoch eine plastischreichereForm und auch gr6ssereDimensionenauf. Da BruchstuickeahnlicherSimen auch in Olbia aufgetaucht sind,28werden die im SchwarzmeergebietgefundendenExemplare dieses Typs mit hoher Wahrscheinlichkeitaus der Mutterstadtdorthinexportiertworden und ebenfalls als Giebelsimenanzusprechensein. Eine Datierung der histrianischenSima etwa an den Anfang des letzten Viertels des 6. Jh. v. Chr. wird sowohl durch die stratigraphischenFundverhaltnisseals auch durchdie Moglichkeitnahegelegt,den plastischenEierstab-bei aller Vorsicht-auch in eine stilistische Reihe einzuordnen;die milesische Kalabaktepe-Sima geh6rt bereitsmehr an das Ende dieses Jahrhunderts. Auf der Entwicklungsstufeder Kalabaktepe-Simastehen dagegen andere, dieser und der histrianischenGiebelsima nachtsverwandteBruchstuckeeiner Tonplatte, von der jedoch nur der untere Teil-zudem fragmentarisch-erhalten ist (Typ S.II; P1. 31:d).29 Sie bewahren Reste eines in den bekannten Farbtonenbemalten,wiederum plastischen Eierstabes. Dieser schmale Plattentyp verfuigtuber eine senkrecht geebnete Ruckseite ohne Uberzug und bildet am muldenf6rmigunterschnittenenunteren Rand eine deutlicheWassernase aus. Eine derart vertikal orientierte Zurichtung spricht daftir, dass dieser Typ schtitzendvor einer wohl holzernen Konstruktionangebrachtwar, demzufolge nicht als Sima, sondern unterhalb einer solchen als Verkleidungsplattegedient hat. Entsprechende Nagell6cherfehlen, lagen aber vielleichtim Bereichder uber dem Eierstabanzunehmenden weiteren farbigen Ornamentteile;deshalb konnen fur diesen Dachterrakottentypweder seine obere Gestalt noch Plattenhoheund -breite erschlossenwerden. Die vielfaltigen Bezuge dieses Typs zur histrianischenund Kalabaktepe-Simalegen auch hier den Gedanken an unmittelbareAbhangigkeit,wenn nicht gemeinsamemilesischeHerkunft nahe. Zumindest in der Ornamentikvon diesen m6glichen Importsttickeninspiriert, aber in Ton und sonstigerMachart dochdeutlichabzuheben,prasentiertsich ein weiterer,zwanzig quelques exemplairesdecouvertsa Histria,"Dacia N.F. 11, 1967 (S. 95-120), S. 117, Anm. 92; Theodorescu, 1970, S. 38-40, Abb. 10, 11. 26 Fundinventar:His.1952. 27 A. von Gerkan, Milet, I, viii, Kalabaktepe,Athenatempelund Umgebung, Berlin 1925, S. 20-21, 26, Abb. 11, Beilage 2; ATK, S. 103-104, 106-107, Taf. 54:1; G. Kleiner, Die Ruinen von Milet, Berlin 1968, S. 40, Abb. 22:a. 28 Zu diesen Parallelstuicken vgl. V. M. Skudnova,"IzneizdannykhMaterialovOlvii VI-V vv. n.z.,"Sovetskaja archeologija29-30,1959 (S. 248-262), S. 250-251, 255, Abb. 3, 8; Rusjaeva(Anm. 18 oben), S. 34-38, Nr. 3-19, Abb. 3-5. 29 Fundinventar: (von links nach rechts) His.1974 T ([obiecte] 61); 1975 T (12) + 1976 T (23); 1976 T (23).

232

KONRAD ZIMMERMANN

Fragmente umfassender und dabei ganz einheitlicher Dachterrakottentyp(Typ S.III; P1. 32:a, b).30Nur die ca. 13 cm hohe und dabei senkrechteFront ist grundiertund bietet unter Verwendung der uiblichen,sinnvoll variiertenFarbtonedrei verschiedeneSchmuckglieder: oben ein vorspringenderRundstab mit Zickzackstreifen(wie an KalabaktepeSima), in der Mitte ein auf die Flache gemalter doppelter Maander mit eingestreuten Sternen (wie beim histrianischenSima-Typ S.1; P1.31:b), unten eine etwas zuruckgesetzte Wolbung mit Feldereinteilung(wie an Kalabaktepe-Sima).Die gemeinsameFundlage im Bereich des Aphroditetempelsdatiert den Typ in die zweite Halfte des 6. Jh. v. Chr. Doch setzt er offensichtlich die Kenntnis der vermuteten milesischen Importstuckevoraus, als deren locale Nachahmung er sich dann vielleicht erklaren lasst. Problematischbleibt vor allem sein Verwendungszweck.Zwei Fragmente dieses Typs hat Theodorescu3lbereits publiziert und als Geisonplattenangesprochen.Ist die rechtwinkligabbiegendeOberflache mit einer tiefen Rinne und gelegentlichen Lochern versehen, so bildet die Unterseite nur einen Steg aus-jedenfalls ist an keinem Fragment mehr erhalten. Sollte auch dieser Typ irgendwie mit zu schutzenden Holzbalken im Sinne von Verkleidungsplattenverbunden gewesen sein? Nur noch kurz hinzuweisen ist auf die beidenbisher in Histria gefundenenarchaischen Antefix-Typen, da sie ebenfalls schon von Theodorescubekanntgemachtworden sind. Urspruinglichjeweils mit einem Kalypter verbunden, ist zu dem Vertreter des halbrunden lakonischen Typs mit plastischem, farbig gefasstem, aber stark erganztem Gorgoneion (Typ A.I;,P1. 32:c)32neuerdings das Fragment (P1. 32:d)33eines weiteren solchen Stirnziegels hinzugetreten.Bei annaherndgleicher Form besitzt es allerdings einen einheitlich dunkelrotenU0berzug;doch man kann sich am Dachranddurchauseinen Wechsel verschiedenfarbigerGorgoneion-Antefixevorstellen. Auch zu dem kompletten,den korinthischen Typ vertretendenAntefix mit hochgezogenerStirnplatte und flachem Relief, deren einst bemalte Palmette aus einer Doppelvolute entspringt (Typ A.II; P1. 32:e, f),34sind weitere Fragmentedieses Typs hinzugekommen,ohne dass sich dadurchNeues ergebenhatte. Die VertreterbeiderTypen stammeneinheitlichaus der Tempelzone und gehbrenin das beginnende dritte Viertel des 6. Jh. v. Chr., aus stilistischenGriindenpraziserwohl in die Jahre um 540 v. Chr. Deshalb liegt es nahe, sie mit den beiden bisher dort ausgegrabenen,ungefahr zur gleichen Zeit errichtetenarchaischenTempeln zu verbinden. Die typologische Vielfalt und die teils beachtliche kuinstlerischeQualitat der in Histria aufgetauchtenDachterrakottenarchaischerZeit sprecheneinerseitsfur die Beliebtheit 30 Fundinventar:His.1960 T. Pippidi, 1962, S. 153-154, Abb. 14 (um 1800 zu drehen!); Bordenache (Anm. 24 oben). 31 Theodorescu, 1970, S. 40, Abb. 12 (um 1800 zu drehen!). 32 Fundinventar:His.1914/15 + 1951 T. Pippidi, in Histria I, S. 258-259, Abb. 105; ders., 'Les fouilles d'Istros(1914-1957)," BCH 82, 1958 (S. 335-350), S. 343, Abb. 11; Pippidi, 1962, S. 148, Abb. 9;ATK, S.4, Abb. 1:1;Theodorescu,1970, S. 29-31, Abb. 1, 2. 33 Fundinventar:His.1975 T (12). 34 Fundinventar:His.1960 T. Pippidi, 1962, S. 153, Abb. 16; Theodorescu, 1970, S. 36-38, Abb. 8, 9.

ARCHAISCHE DACHTERRAKOTTEN AUS HISTRIA

233

dieserdekorativenGattung in jener Epoche,andererseitsvon einer auch sonst nachgewiesenen Blutezeit dieses Fundplatzes im 6. Jh. v. Chr. Dabei tritt auch an diesem Material die Bindung an das griechischeMutterland deutlich hervor, die sich fuirHistria natiirlich auf den ostgriechischenBereich,namentlichMilet bezieht. KONRAD ZIMMERMANN WILHELM-PIECK-UNIVERSITAT ROSTOCK

WB Altertumswissenschaften Universitatsplatz DDR-2500 Rostock German DemocraticRepublic

a. Fundplatz archaischerDachterrakottenim Pronaos des Aphroditetempelsvon Histria, Sondage I/y2 (1976)

c. Fragment eines Deckziegels lakonischerForm b. Zwei komplette Flachziegel untersehiedlicherTypen

PLATE 29

a. Vier teils komplette Deckziegel unterschiedlicher Typen (a-d)

c. KompletterDeckziegel in Mischform

e. Weihinschriftan Aphrodite

b. Fragment eines Deckziegels korinthischerForm

d. Fragment eines Firstkalyptersmit Weihinschriftan Aphrodite

PLATE 30

a. Fragment eines bemalten Traufziegels

b, c. Fragment eines bemalten Traufziegels

r

~~KNA

ZIMRANACASH

d-f- Komplettesog. Abdeckplatte

AHERKTE

U

ITI

PLATE 31

a, b. Fragmente einer

bemaltenEierstab-

-

Sima

Wi'

c. Fragment einer bemalten Sima

d. Fragmenteeiner bemalten Verkleidungsplatte

PLATE 32

a, b. Fragment eines bemalten Dachterrakotten-Typs

.

c. Erganzter Stirnziegel mit Gorgoneion

e, f. KompletterStirnziegel mit Palmette

d. Fragment eines Gorgoneion-Antefixes

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE IN DELPHI UND OLYMPIA (PLATES

A

33-38)

LFREDMALLWITZhat in mehrerenPublikationendie komplexenProblemeder

Schatzhauserin Olympia und der ihnen zuzuweisendenTondacher dargestelltund viele von ihnen gelost.I Von seinen Ergebnissenwerde ich im Folgenden auch im Zusammenhangmit den Fundumstandenund den Werkstattbeziehungenausgehen.1 Pausanias nennt bei seinen Gang tiber die Schatzhausterrasse (Pausanias, vI.19.I-I5)

zehn Hauser, zwolf Fundamente sind jedoch erhalten. Das kleine Fundament VIII trug indessenwohl einen Altar und scheidetsomit fur die Bennenungder Bautenaus. Pausanias 'Mallwitz, 1968; Mallwitz, 1972; Mallwitz, 1980. Abkurzungsverzeichnis: = Convegnodi Studi sulla Magna Grecia, Napoli CMGr Griffin = A. Griffin,Sikyon, Oxford 1982 Mallwitz, 1968 = A. Mallwitz, "Ein Scheibenakroteraus Olympia,"AM 83, 1968, S. 124-146 Mallwitz, 1972 = A. Mallwitz, Olympiaund seine Bauten, Munchen 1972, S. 163-179 Mallwitz, 1980 = A. Mallwitz, "Die Dacher aus gebranntemTon," in: Die Funde aus Olympia.Ergebnisse der Ausgrabungstdtigkeit,A. Mallwitz und H. V. Herrmann, Hrsg., Athen 1980, S. 141-151 2 Vgl. auch K. Herrmann, "Beobachtungenzur Schatzhausarchitektur Olympias,"NFGH, S. 321-359; H. V. Herrmann, Olympia,Heiligtum und Wettkampfstdtte, Munchen 1972, S. 240, Anm. 390. Von dieser ursprunglichauch von Mallwitz vertretenen(Olympiamodell),wenn auch spater wieder in Frage gestellten (Mallwitz, 1972, S. 173) Annahme gehe ich aus folgenden Grundenaus: Der von Dorpfeld und Mallwitz erwahnte grobe, spate Mortel an den Stufen des Baus VIII bezeugt, dass dieser in romischer Zeit neu verputzt und hergerichtetworden ist. Er duirftezu Pausanias' Zeit also noch aufrecht gestanden haben. Handelte es sich bei dem Bau um ein Schatzhaus,dann also wohl nicht um eines der "unbekannten Schatzhauser,"sondernum eines der von Pausaniasgennanten,und zwar um das einer Stadt,die in r6mischer Zeit noch die wirtschaftlicheMoglichkeit und die kulturelle Motivationhatte, ihr Schatzhausin Olympia zu renovieren.Gegen die Annahme von zwei anderen unbekanntenSchatzhausernsind jedoch bei dem in Anmerkung 4 gennanten Verstandnis des Pausaniastextes schwerwiegendeEinwande anzufuhren. Folgende Moglichkeitenkommenin Betracht: a. Die zwei unbekanntenHauser lagen zwischen denen von Kyrene und Selinunt und zwischen denen von Metapont und Megara (Fig. 1:c). b. Die zwei unbekanntenHauser lagen beide zwischen denen von Kyrene und Selinunt (Fig. 1:d). c. Die zwei unbekanntenHauser lagen beide zwischen denen von Metapont und Megara (Fig. 1:e). Zu a: Das Haus der Selinuntinerfiele dann auf FundamentVIII, dessen Bau in spatererZeit restauriertworden ist. Nach der fruhen ZerstorungSelinunts durch die Karthagerist es unwahrscheinlich,dass diese Stadt ihr Haus so spat erneuern liess. Das Haus von Megara fiele auf das Fundament IX, das nach seiner eingeengten Lage zwischen den alteren FundamentenVIII und X zu denjuingerengeho6enmuss. Nach der von Pausanias erwahnten, mit dem Gruindungsmythosder Stadt in Verbindung stehenden Endymionstatueim Haus der Metapontiner(s. unten S. 248) gehortediesesjedoch sicher zu den altesten. Zu b: Eines der unbekannten Hauser fiele auf Fundament VIII, was aus den oben genannten Gruinden unwahrscheinlichist.

III a.

IV

li Sikyon

I Sikyon

c.LIL X

I

Sikyon

Syrakus

Epidamnos

II

IIIl I

II

I I

I

I:l

IALA

Epidamanos

Sybaris

Sybaris

Sybaris

Altar

I v I vI IE

Kyrene

I<

I

Byzanz

xr -

Selinunt

Kyrene

I

Byzanz

i

-

I,

Epidamnos

Syrakus

V---

Byzanz

Syrakus

il

L

V

II I

I

I

Altar

?

I Kyrene

T

vllr III

l

I]

I

Selinunt

?

I

Metapont

xl Selinunt

I

M

-

K1 Metapont

-I I I I I I I I I I ~~~~~~~~~~~~~~~~~~~~~~~~~Vll III1

I d.

L

L Sikyon

Syrakus

Epidamnos

XL Byzanz

mL Sybaris

I

II

Kyrene

?

VI

VI I

Kyrene

Selinunt

?

I

Selinunt

M

IX

e. Sikyon

Syrakus

FIG. 1.

Epidamnos

Byzanz

Sybaris

Metapont

?

SchematischeDarstellung der Schatzhausreihe.a, b: Nach dem Pausaniastextm6gliche Lo unbekanntenSchatzhauses.c-e: Nach dem Pausaniastextm6gliche Lokalisierungzweier u Schatzhauser

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

237

hat also nur ein Haus nichterwahnt,und zwar wohl deshalbweil es zu seinerZeit nichtmehr zu sehen war. Bei seiner Beschreibungverbindeter immerzwei Hauser mit einer Konjunktion wie "gleichanschliessend","das nachste","an der Seite", oder mit einer andern ihre Lage bezeichnendenBemerkung.Nur zweimal tut er das nicht:zwischen den Hausern der Kyrenaerund der Selinuntinerund zwischendenender Megarerund der Metapontiner.An einer dieserStellen,durfenwir annehmen,lag das von ihm nichtgenannteHaus.4 Legen wir die Lucke zwischen die Hauser von Metapont und Megara auf Fundament X (Fig. 1:a), verschiebtsich die Reihe so weit nach links, dass das Haus von Selinuntjenseits der Altars VIII auf Fundament VII fallt, was wohl nicht moglich ist, da Pausanias vom Haus der Metapontinerja ausdriicklichsagt, es schliesse gleich an das von Selinunt an. Wahlen wir jedochdie Stelle zwischenden Hausern von Kyreneund Selinunt (Fig. 1:b),bleibenalle vom Text gebotenenVoraussetzungengewahrtund wir stehenvor dem FundamentVII, von dem nur noch eine ganz unvollstandigeQuaderschichtvorhandenist, weshalb auch Mallwitz dazu neigte, in ihm das des verlorenenHauses zu sehen.5 Zu c: Das Haus von Metapont fiele auf das Fundament VIII des in spaterer Zeit restauriertenBaus. Da Metaponts Blute die Pyrrhuszeit nicht ulberdauerthat, war dieser Bau kaum das Schatzhaus dieser Stadt. Auch die Placierung des Hauses der Selinuntiner auf dem am schlechtestenerhaltenen Fundament VII ist unwahrscheinlich,da es zu Pausanias' Zeit noch stand und ein Standbilddes Dionysos beherbergte. 4 Die enge Nachbarschaftzwischen dem von Pausanias als erstem genannten Haus, dem von Sikyon, zu dem zweiten, dem von Syrakus,geht aus dem Wort f4E$ivEf(das nachste)hervor.Die Nachbarschaftzwischen dem zweiten Haus und dem von Epidamnosist dadurchgesichert,dass Pausanias dieses "das dritte"nennt, und zwar in einem Atemzug mit dem anschliessendenvierten, das offenbarden Byzantinerngeh6rte. ("Dass nach dem Namen der Epidamnierder der Byzantiner ... ausgefallen ist, geht aus S. 505,5 unwiederleglich hervor;im Folgenden spricht er allerdings nur von dem Schatzhausder Epidamnierallein, doch standen in der Lucke vielleicht auch Angaben uber das Schatzhaus der Byzantiner."H. Hitzig, Das Pausanias Beschreibungvon Griechenland,Leipzig 1904, S. 632.) Das fuinfteHaus, das der Sybariten, wird durch das Wort fX4LuEvog(das nachste) an das von Byzanz angeschlossen,und das der Kyrenaeran das der Sybariten durch das Wort 7rpog(nahe bei). Erst bei der Erwahnung des Hauses der Selinuntinergibt Pausanias keine topographischeAngabe. Erst hier, zwischen ihm und dem der Kyrenaerkonntesomit ein unbekanntesSchatzhaus gestanden haben. Das Haus der Metapontiner wird dann wieder durch das Wort lrpoOEX7S' (daran anschliessend)direkt an das von Selinunt angeschlossen.Nun geht Pausanias wieder ohne Ortsbezeichnung zum Haus der Megarer uiber,somit k6nnte auch zwischen ihm und dem der Metapontinerein unbekanntes Haus angenommenwerdern. Dass das als zweitletztes genannte Haus der Megarer auch wirklich das zweitletzte der Schatzhausreihe ist, ist dadurch gesichert, dass die fast ganz erhaltene Frontarchitektureiner Schatzhausesmit der BauinschriftMEyap'wv und der von Pausanias genannten Gigantomachieim Giebel genau auf das entsprechendeFundament passt. Vom Haus der Geloer sagt Pausanias, es sei das letzte. Vgl. K. Herrmann, "Die Giebelrekonstruktiondes Schatzhausesvon Megara,"AM 89, 1974, S. 75-83. Zur Frage der Zuverlassigkeitder Angaben des Pausanias s. Ch. Habicht, Pausaniasund seine BeschreibungGriechenlands, Munchen 1985, S. 40-63, 93-100, 150-152; 0. Regenbogen,in: RE, Suppl. VIII (1956), s.v. Pausanias, bes. Sp. 1070-1076, 1090-1091. 5Mallwitz erwog (in Mallwitz, 1972, S. 173) die Placierungder Schatzhausesder Kyrenaerauf Fundament VII nur deshalb,weil ihm das FundamentVI fur das seit F. Studniczkanach einem Relieffragmentmit bewegter weiblicher Figur in die erste Halfte des 6. Jahrhunderts datierte Kyrenaerhauszu jung zu sein schien. Nachdem F. Chamouxund N. Bookidisklar ausgesprochenhaben,dass in der Figur wohl eine Herrin der Tiere, sicher aber nicht speziell die Nymphe Kyrene zu sehen ist, womit jede Beziehung zum Kyrenaerhaus wegfallt, besteht auch kein Hinderungsgrundmehr, dieses auf dem jungeren Fundament VI zu lokalisieren. (F. Chamoux, Cyrenesous la monarchiedes Battiades [BEFAR 177], Paris 1953, S. 378-385; N. Bookidis, A Study of the Use and GeographicalDistribution of ArchitecturalSculpture in the Archaic Period, diss. Bryn Mawr College, 1967, S. 85, 87).

238

MADELEINE MERTENS-HORN AND LUISA VIOLA

Die hier interessierendenHauser, die einst westgriechische Dacher getragen haben konnten,sind nun: II Syrakus,III Epidamnos,V Sybaris,VII unbekanntesSchatzhaus,IX Selinunt, X Metapont, XII Gela. Das Dach das Schatzhauses der Geloer (P1. 33:a):6 Ch. Wikander hat zuletzt die "par-

tial abnormality"des Dekors dieses Daches innerhalb des Gesamtbildes der sizilischen Dacher und die geringen Beruhrungspunktemit den Dachern in Gela hervorgehoben.Die auf Suisserottzuruckgehende Datierung um 560 wird allgemein noch als Fixpunkt der Chronologieder sizilischen Dacher verwendet,dochhat V. Kastnerletzthin auch eine Entstehungszeitkurz nach der Mitte des 6. Jahrhundertsvorgeschlagen.SuisserottsDatierung griindete, ausser auf der nun fast hundertjahrigenChronologie Dbrpfelds, auf dem Vergleich der umschriebenen Palmetten auf dem Simasockel der Suidseitedes Hauses mit Ornamentender Vrouliaschalen.Die Problematikder Datierung der Dachterrakottenanhand der Vasenmalereiist indes seither oft genug betont worden. Das Datum "um 560" ist also noch alles andere als gesichert. Von mehrenenAutoren ist bereits ein Dach von einem unbekanntenBau in Kaulonia7 mit dem das Geloerhausesverglichen worden, da die Ornamente seiner Sima denjenigen der Suidseitensimades letzteren typologisch sehr nahe kommen (P1. 33:b). Die Reihe der umschriebenenPalmettenist hier als fortlaufendesAnthemionverstanden,in das die Wasserspeierunregelmassigeingesetzt sind. Die zugehorigeGeisonverkleidung(P1.33:c) weist ein versetzes Flechtband auf, wie es gelegentlich in Unteritalien,8 nicht aber in Sizilien vorkommt.An einer weiteren Sima eines kleinen Daches von einem unbekanntenBau in Paestum (P1. 33:d-f) finden sich nun die paarweise zwischen den Wasserspeiernangeordneten umschriebenenPalmetten des Geloerhausesfast in der gleichen Form. Die zeitliche Nahe aussert sich vor allem in der Form und der Lage der Blattchen und im Verlauf der Bogen, die sich gleichsametwas um jene zu dehnen scheinen. Das sch6ne Kehlenornament des Geloerhausesaus langen Efeublattchenzwischen unten zugespitzten,hohen Blatternist hierjedochzu einer Reihe aus stehendenund hangendenTropfen mit breitenweissen Randern vereinfacht,und der fortlaufendeMaander auf der Stirnleiste ist zu einer Folge aus Maanderkreuzen und Lotosbluitenauf rechteckigen,einst roten Feldern geworden. Ein Geisonfragment(P1. 34:a) dieses Daches zeigt dasselbeversetzteFlechtbandwie das Dach in Kaulonia in eng verwandterAusfuihrung,auf dem doppelt geschwungenenoberen Abschlussprofiljedoch einen einfachendorischenBlattstab. 6 Zuletzt: C. Wikander, 1986, S. 19, 50, Nr. 80, mit der alteren Literatur;V. Kastner, "Studienzu den Dachern sikeliotischerund italiotischerTempel," WissenschaftlicheZeitschriftder Humboldt-Universitatzu Berlin 25, 1976, S. 520-522; H. K. Siisserott,"GrossgriechischeDachterrakotten.Das Schatzhausvon Gela," OlForschI, Berlin 1941, S. 83-110. Auffallig ist die Verschiedenheitder Zahlbuchstabenauf dem Dach des Geloerhausesvom Bezeichnungssystemder archaischenDachern in Gela! Vgl. OlympiaV, S. 792, Nr. 943; L. BernaboBrea, "L'Athenaiondi Gela e le sue terrecottearchitettoniche,"ASAtene27-29,1949-1951, S. 46, Abb. 35, S. 56, Abb. 43. 7Wikander, 1986, S. 19, Anm. 47, mit der alteren Literatur;A. De Franciscis, "MonasteraceMarina (Caulonia),"NSc 1957 (S. 184-190), S. 187-188, Abb. 9, 10. 8 Lokri: A. De Franciscis, I1 santuario di Marasa in Locri Epizefiri, I, I1 tempio arcaico, Napoli 1980, S. 95-96, Abb. 66, 67. Sybaris:A. De Franciscis, "Contributiall'archeologiadi Sibari,"RendNap 36, 1961 (S. 63-84), S. 67, Taf. XI, Abb. 2.

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

239

Dieses Dachlein in Paestum ist nun sehr wichtig, weil es sich anhand des Vergleichs mit der grossen Baldachinsimader sog. Basilika in Paestum ziemlich genau datierenlasst. Diese zeigt wieder in der Kehle die stehendenund hangendenTropfen mit breitenweissen Randern und auf der Stirnleiste die Maanderkreuzeund Lotossterneauf roten Feldern (P1. 34:c). Die wesentlich grosseren Palmetten des Anthemions auf dem Simasockel (P1. 34:d) wirken durch die schwarze Fullung der Blatterreicher,in Form und Proportion sind sie jedochvon den umschriebenenPalmettendes kleinen Daches nicht zu trennen. Die Geisonverkleidung(P1.34:e) weist wieder das doppeltgeschwungeneobereAbschlussprofil mit dem Blattstab auf, doch das Flechtband hat die kanonischesizilische Form. Die vor allem aus den plastischen Lowenkopfenzu erschliessendeEntstehungszeitdes Daches der Basilika um 530/209 darf somit auch fur das kleine Dach mit den umschriebenenPalmetten gelten. Dass nun das Dach des Geloerhausesetwas alter sein muss, geht aus der hoheren Qualitat seiner Bemalung hervor,sowie aus dem typologischalteren kontinuierlichenMaanderauf der Stirnleisteund aus dem Kehlenornament,das noch den ursprtinglichenBlattstab verrat. Allzuweit mochte ich es jedoch nicht abruicken,was uns in die Jahre 540/30, jedenfalls in das dritteViertel des 6. Jahrhundertsfuihrt. Das eben erwahnte Kehlenornamentmit den Efeublattchenzwischen hohen, unten zugespitzten Blattern findet sich auch an einer Sima in Reggio Calabria,10und zwar uber demselben,aus gegenstandigenhellen Zungen gebildetenKreuzornamentwie an den Giebelsimen des Geloerhauses.Auch das Wellenbandmit den die unteren Schleifen fullenden Blattchen ist dort zu Hause.11Die hangende umschriebenePalmette begegnet, ausser an einem weiteren, qualitativ minderwertigenDach aus Kaulonia,12an einer grossen Sima in Paestum (P1. 34:b), die nach dem geraden Blattstab in der Kehle wohl alter ist und dem Schatzhausdachvorausging.Diese Beispiele zeigen, dass in Unteritalien zu dieser Zeit eine wichtige Terrakotta-Werkstatttatig war, die diese Schmuckformenverbreitete,und die wohl auch den ehrenvollenAuftragubernommenhat, das Dach des Geloerhausesin Olympia auszufuhren. Das Dach des Hauses der Epidamner (P1. 34:f),13 Fundament III: Diesem Schatzhaus

hat Mallwitz ein Dach zugewiesen, das in seinen Schmuckelementenenge Abhangigkeit vom ersten Dach des Artemistempelsin Korfu zeigt.14Von Korfu war die Kolonisierung von Epidamnosausgegangen,Beziehungenzwischen den Stadtensind auch spaiterbezeugt. Diese Zuweisung findet eine glanzende Bestatigung durch Funde in Cavallino, einem kleinen stark hellenisierten Zentrum in Messapien nahe bei Lecce an der adriatischen Kuiste,und zwar uibersMeer ziemlich genau gegenuiberder Stadt Epidamnos-Dyrrachion, 9 M. Mertens-Horn, Die Lowenkopf-Wasserspeierdesgriechischen Westensim 6. und 5. Jahrhundertsv. Chr., RM-EH 28, Mainz 1988, S. 131-137. 10 N. Putorti, "Terrecottearchitettonichedi Reggio Calabria,"in: L'Italia Antichissima I, 1929, S. 33, Taf. II:3. " E. D. Van Buren, Archaic Fictile Revetments in Sicily and Magna Graecia, London 1923, S. 104, Taf. IX:37. 12 P. Orsi, "Caulonia,"MonAnt 29, 1923 (S. 410-490), S. 426, Taf. II:b. 13 Mallwitz, 1972, S. 170-171, Abb. 130. 14KorkyraI, S. 97-106. Le Roy, 1967, S. 70.

240

MADELEINE MERTENS-HORN AND LUISA VIOLA

dem heutigen Durazzo. Die Sima,"5die direkt auf dem steinernen Geison auflag, zeigt in der Kehle einen schwarz-rotenBlattstabmit plastisch hervorgehobenenRandern, und auf dem Sockel eine grosse Spiralwelle mit Zwickelpalmetten.Es sind dieselben Motive, die, gemalt, auf der Sima in Olympia, und wiederum in Relief am Dach des Artemistempels erscheinen. Zur Datierung ist von Francescod'Andriadas 3. Viertel des 6. Jahrhunderts vorgeschlagenworden, und zwar vor 530, in welcher Zeit der Artemistempelein neues marmornesDach erhielt, weshalb das alte nicht mehr als Vorbild gedient haben kann. Das Haus von Syrakust6 ist erst im 5. Jahrhundert,wohl nach der Schlachtvon Himera erbaut worden. Es k6nnte, wie alle wichtigen Bauten Siziliens des fortgeschrittenen5. Jahrhunderts,ein Dach aus Marmor oder Stein gehabt haben."7 Das Haus der Selinuntiner'8 geh6rte nach Aussage des Fundamentes IX, vor allem nach dessen zwischen die alteren FundamenteVIII und X eingeengten Lage ebenfalls zu den jtingeren. Unter den noch nicht zugewiesenen Tondachern Olympias k6nnen nur die nahe beim Buleuteriongefundenen,aber nicht zu ihm geh6rendenFragmenteeiner hohen, mit R6hrenwasserspeiernversehenenTraufsima mit recht ungewohnlichemProfil bedingt als sizilisch angesprochenwerden.'9 Das einfache Flechtband auf der Sockelzone ist an spatarchaischenSimen aus dem sizilischen Naxos20geliufig, und das aus gegenstandigen hellen Zungen gebildete dunkle Kreuzmusterin der Mitte k6nnte von der Sima des Geloerhauses kopiert worden sein. Das lesbisches Kymation, das paradoxerweise auf eine Hohlkehle gemalt ist, findet seine engsten Parallelen in klazomenischenSarkophagender Albertinum-Gruppe,die in die ersten Jahrzehnte des 5. Jahrhundertsdatiert sind.21Auch die Bemalungstechnik,bei der die Farbennur nebeneinanderaufgesetztsind, sowie die stark aufgefachertenPalmetten der seltsamenAntefixe sind Anzeichen eher spater Entstehung. Bei einer Datierung um die Jahrhundertwendeoder an den Anfang des 5. Jahrhunderts wtirdedas Dach also gut zum Schatzhausder Selinuntinerpassen. Nach derTonbeschaffenheit, der im Detail ungewohnlichen Profilform,und dem gemalten Kymation m6chte ich indessenannehmen,dass es in Olympia selbstausgeftihrtwordenist. Nun bleiben von den archaischenFundamentennoch die zwei grossenund relativalten V und X, die, einst an beherrschenderStelle, die Hauser der wichtigstenAchaerstadteSybaris und Metapont trugen.22Auch das unbekannteSchatzhausVII dtirftejedoch ein sehr altes Dach gehabt haben. Zudem wissen wir, dass dem fruihklassischenHaus der Sikyonier ein altererBau vorausgegangenwar, der am Ende des 6. oderzu Beginn des 5. Jahrhunderts 15 F. d'Andria, "Cavallino (Lecce). Ceramiche ed elementi architettoniciarcaici,"MEFRA 89, 1977, 2 (S. 525-562), S. 552-553, Abb. 17:a, b. 16 Mallwitz, 1972, S. 169; ders. "Architektureines Schatzhauses,"OlBer VII, 1956-1958, S. 29-55. 17 Vgl. Mertens-Horn (Anm. 9 oben), S. 89-114, 165-172. 18 19

Mallwitz,1972,S. 173.

Olympia II, S. 194, Taf. 118:5a, b; vgl. J. Schilbach, "Die Silen-Manaden-Gruppeaus Olympia. Ein neuer Rekonstruktionsversuch," AM 97, 1982 (S. 25-34), S. 34, Anm. 45; Mallwitz, 1972, S. 238. 20 C. Ciurcina, "Nuovi rivestimentida Naxos e da altri centridella Sicilia orientale,"CronCatania16,1977 (S. 66-81), S. 68, 76 Nr. 11, Taf. V:2, 3. 21 R. M. Cook, ClazomenianSarcophagi(KerameusIII), Mainz 1981, S. 31-32, Abb. 20, Taf.-39:3. 22 Mallwitz, 1972, S. 170-172.

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

241

abgetragenwordenist, und dessenDach ebenfallsunter den altestenzu suchensein dtirfte.3 Tatsachliche stehen uns noch vier Dacher zur Wahl, die ich kurz die zwei Dacher mit Elementen des dorischenGebalks (Pls. 3:d, 4:a-c),24das Rosettendach(P1.4:e),25und das Dach mit der Blattstabsima(P1.4:f)26nennenwerde. Wichtig sind die Fundumstandedieser letzten beiden Dacher, die zuletzt von Mallwitz beschriebenworden sind.27Die Fragmente der Blattstabsimasind alle im Nordwall des Stadions ausgegrabenworden, was darauf schliessen lasst, dass das Dach am Ende des 6. oder um die Jahrhundertwendeabgetragenworden ist. Die Fragmentedes Rosettendaches fanden sich an verschiedenenStellen in den Stadionwallen,weshalb sie offenbarnicht alle gleichzeitig abgetragen,sondern im 5. Jahrhundert nach und nach, Stuck fir Stuck vom Dach gefallen sind und dann weggeschafftwurden. Allein daraus ware zu schliessen, dass das Dach mit der Blattstabsimadem um 500 herum planmassig abgetragenenVorgangerbaudes Sikyonierhausesgeh6rte und das zu mehrerenZeitpunktenim 5. Jahrhundert,Stuckum StuckheruntergefalleneRosettendach dem frtih zur Ruine gewordenenund zu Pausanias Zeiten nicht mehr sichtbarenHaus des FundamentesVII. Die beiden Dacher mit Elementendes dorischenGebalksmtisstendann den alten Hausern von Sybarisund Metapont zugewiesen werden, die zur Zeit des Pausanias noch standen. Wie wir sehen werden, konnen neuere Funde aus Unteritalien diese erste Arbeitshypothesenur bestatigen. Das Rosettendachdes FundamentesVII: Eine kleine Simaecke(P1.35:a) mit rosettengeschmtickterglatter Stirn uiberdem Ansatz einer sehr flachenHohlkehle hat sich in Metapont gefunden.28Sie ist kleinerals die Sima in Olympia, die Rosettensind dementsprechend vereinfacht,die Blattrandersind nicht eigens plastisch abgesetzt.Eine dem Dach in Olympia vergleichbareGeisonverkleidungmit grossen, schweren Rundstabenneben einem einfachen Flechtbandist unter dem reichen Material in Metapontjedochnicht vorhanden.Da die Sima etwas jtinger sein dtirfte als die in Olympia, hatte sie wohl ein konventionelles Geison mit dtinnen Rundstaben. Das Fragment ist seltsam schief und verzogen, es hat sicher keinen wichtigen Bau geschmtickt. 23 C. Weickert, Typen archaischerArchitekturin Griechenlandund Kleinasien, Augsburg 1929, S. 88, Nr. 1 (Datierung wie Heraion). Herrmann (Anm. 2 oben), S. 99, 100; Griffin, S. 101-106. 24 Vgl. Mallwitz, 1980, S. 148-149, Taf. CXVI:101; Mallwitz, 1972, S. 178, Abb. 140; Le Roy, 1967, S. 80-84, Taf. 27-28; OlympiaII, S. 200-202, Taf. CXVI:5, 6. Die Bezeichnung"Dachermit Elementendes dorischenGebalks"scheint mir zutreffenderals der Begriff "H6rnerdacher",da die sog. Hornerantefixeim Ursprungsland Unteritalien auch mit anderen Verkleidungsplattenvorkommen.Vielleicht ware fur diese Antefixform der Begriff "Blultenantefixe" besser, da er ihrer wahrscheinlichenEntstehung gerechterwird. Der Begriff "Hornerantefix"liesse sich dann auf die Antefixe mit den hornartigausgeformtenspitzen Ecken (Typus Aegina I) beschranken. 25 K. Suisserott,"Ein altsizilisches Dach," OlForsch,I, Berlin 1941, S. 128-138; Mallwitz, 1972, S. 171172, Abb. 132-133; Mallwitz, 1980, S. 148, Taf. CXVI: 100:2;ATK, S. 260-265. 26 Olympia II, S. 202, Abb. 26; Mallwitz, 1972, S. 173, Abb. 134; Mallwitz, 1980, S. 147, Taf. CXVI: 100:2. 27 Mallwitz, 1980, S. 147. 28 Publikationin Vorbereitung.

242

MADELEINE MERTENS-HORN AND LUISA VIOLA

Mehrere Fragmenteeines dem Rosettendachin Olympia aufs engste verwandtengrossen Dachs sind in den Ausgrabungender Stadt Siris, dem heutigen Policoro, im Umkreis des archaischen Tempels zum Vorschein gekommen (Pls. 37, 38).29 Die achtblitterige Palmette ist zwar stark verrieben, doch lassen sich noch die plastisch abgesetzten Blattrandermit dem Ring um die Mitte erkennen.Auch ein kleines Fragmentder Sima mit dem aufgemalten Blattornamentist vorhanden.Die Fragmente der zugehorigenVerkleidungsplatten-weisen Paare von Rundstabenauf, die mit rechts- und linksliufigen Bindern bemalt waren und ein einfaches Flechtbandbegleiteten, das wie am Dach in Olympia nicht hoher war als die beiden Rundstabezusammen.Mit jenem hat es jedochvor allem auch die nur im Profil sichtbaren leichten Uberstaindeder Rundstabe uiberdie Plattenrandergemeinsam, die oben den Ubergang zur Sima und unten den zum Geison kaschierten.Mehrere im benachbartenDemeterheiligtumgefundene Fragmenteeiner Geisonverkleidung30 zeigen etwas duinnereRundstabe,aber dasselbe Flechtmusterwie in Olympia, an dem die Bander in drei gleich breite Bahnen aufgeteilt sind, von denen einmal die eine innere, einmal die zwei ausserenschwarz sind. Dieser Verkleidungstypuswar also in Siris heimisch. Innerhalbdes Gesamtbildesder DachterrakottenSiziliens und Unteritaliens schliessen sich diese Fragmente und das Rosettendachin Olympia so nahe zusammen, dass zu erwagen ist, ob das so fruihzur Ruine gewordeneHaus des FundamentesVII, das nach den Folgerungen aus den Fundumstandenmit dem Rosettendachgeschmticktwar, nicht eben das Schatzhausder ionischenStadt Siris war. Die Geschichtedieser Stadt sprichtzweifellos daftir.31 Die sehr alte kolophonischeGruindungist schon in der ersten Halfte des 6. Jahrhundertsvon den machtigenbenachbartenAchaerstaidtenMetapont, Sybaris, und Kroton zerst6rtworden, nachdemdiese, wie PompeiusTrogus wusste (Just., xx.2.3-9), schon fruih beschlossenhatten, alle nicht achaischenGriechen aus dem Lande zu werfen. Der Ruf der LandschaftSiris war bereits im mittleren 7. Jahrhundertbis nach Thasos gelangt, wo Archilochossie als besonderssch6n pries (apud Athenaios, xII.523d). Fur die Bedeutungder Stadt im frtihen6. Jahrhundertsprichtdie Nachricht,dass der Sirite Damasos als einer der Bewerber um Agariste, die Tochter des Kleisthenes von Sikyon galt, deren Hochzeit bei einer der Olympiaden der 70er Jahre des 6. Jahrhunderts in Olympia ausgerufen wurde (Herodot,VI.I26-I27).32 Nach der Zerst6rungder Stadt gelangte die Siritis zunaichstunter metapontinischeHerrschaft. Doch ist auch daran zu erinnern, dass Herodot in den vierziger Jahren des 5. Jahrhundertsden Themistokles sagen lasst, dass er, wurden seine Ratschlage nicht befolgt, in die Siritis auswandernwollte, die zu besiedelndas Orakel geboten habe (Herodot, vIII.62).33In den Fragmentender Tragodie Die gefesselteMelanippe des 29 D. Adamesteanu,"Siris.Il problematopografico," in 20. CMGr, 1980, S. 87, Taf. V oben, VI unten, und VII oben. Vom gleichen Fundplatz wie die FriesfragmenteTaf. III:a, b! S. L. Viola, unten S. 249-250. 30 B. Neutsch, "Documentiartisticidel santuariodi Demetra,"in 20. CMGr, 1980, S. 155, Taf. XIII, XIV. 31 M. Lombardo, in Siris-Polieion, Incontro Studi, Policoro 1984 (1986), S. 57-62; F. Mele, in ebd., S. 99-100; G. L. Huxley, in 20. CMGr, 1980, S. 27-43; J. Berard,La colonisationgrecque de l'Italie Me'ridionale et de la Sicile dans l'antiquite',Paris 1975, S. 187-199; M. Guarducci, "Siris,"RendLinc 33, 1978, S. 273-288. 32 Griffin, S. 43-44. 33 R. Crahay, La litte'rature oraculairechez Herodote,Paris 1956, S. 144-145.

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

243

Euripidesist davondie Rede, dass, entgegenden historischenTatsachen, ein Metapontiner einst Siris gegrtindet habe,34was fur den Theaterbesucherin Athen eine ausreichende Rechtfertigungdarstellte,frei tiber das Gebiet zu verhandeln.Die alte jonische Stadt war offenbarschonim 5. Jahrhunderteiner Art der damnatiomemoriaeverfallen,weshalb auch ihr Schatzhausnicht mehr gepflegtund damit dem Verfall preisgegebenwurde. Nach der oben aus dem GrabungsbefunderschlossenHypothesewairedas Dach mit der Blattstabsimadem alteren Schatzhausder Sikyonierzuzuweisen, wahrend die Ddcher mit Elementen des dorischenGebdlksden Hausern der Sybariten und der Metapontiner gehorten. Nach Beobachtungenvon Alfred Mallwitz bestehenalle&drei Dacher aus dem gleichen lokalen Ton und sind wohl in Olympia in derselbenWerkstattgeschaffenworden,35 ebenso wie eine Blattstabsimain Elis36und das Dach 30 mit Geison und Mutuli in Delphi.37Ch. Le Roy hat der Gruppe dann noch das kleine Dach 31 in Delphi hinzugeftigt.38 In diesem Falle muissenwir einerseits davon ausgehen, dass diese Werkstatt Traditionen ihrer Heimat vertrataber auch anderen,neuen Einfliissenoffenstand,und anderseits,dass sie, da sie Dacher fur verschiedeneBauten zu schaffenhatte, nicht nur eine Dachtypologie vertretenkonnte, die ausschliesslichin einer grossgriechischenStadt vertretenwar-dass sie sozusagenuiberregionalenCharakterhatte. Die Idee, auf VerkleidungsplattenElemente des dorischenGebalks darzustellen,war sicherin der Gegend der AchaerstadtKrotonbesondersverbreitet,was Dieter Mertens veranlasst hat, in dem verlorenenSchatzhausdas von Krotonzu sehen. Hier haben wir die bekannte, ganz ungewohnliche, und schwer datierbareVerkleidung des alteren Tempels in Ciro-Crimisa(P1.35:b), welche eine genaue Parallele in Kaulonia hatte.39An der Geisonverkleidungfinden sich zwei Taeniae mit versetzten Regulae und Guttae und unter den Pentagonalantefixenwiederum regulaartige Leistchen mit Guttae. Das neugefundene, wohl juingereDach aus ContradaS. Anna bei Kroton40ist ganz andersaufgebaut (P1.35:c, d). Die Geisonverkleidungzeigt einen Triglyph, oder besser einen Tetraglyph, und wiederum Regulae mit Guttae und flechtbandgeschmtickteViae, sowie, am "Hornerantefix" tiber der Hohlkehlensima, die Guttae des Daches aus Ciro. Ein recht hypertrophesGebilde! Genau die Kombinationvon Elementen der Dacher in Olympia finden wir in Kroton jedoch nicht.41Der Reiz der Dacher in Olympia liegt ja gerade in der ganz bewussten H. Van Looy, Zes verlorenTragediesvan Euripides, Brussel 1964, S. 247-248, 327. Mallwitz, 1968, S. 145, Anm. 73. 36 N. Yalouris, <<'Avao-Ka4 'ApXaL'og "HAXLo0>>, flpaKTLKa1972 (S. 139-142), S. 142, Taf. 122:b. Ein weiteres Fragmentfand sich bei den Ausgrabungenin Elis von 0. Walter, vgl. Yalourisa. a. 0. Vgl. auch die Terrakottatriglypheaus Elis: R. Demangel, "AnecdotaDorica,"BCH 71-72, 1947-1948, S. 366, Abb. 6, 7. 37 Le Roy, 1967, S. 80-84, Taf. 25, 26, 101, 122. 38 Le Roy, 1967, S. 84-86, Taf. 28, 101. 39 Ciro: P. Orsi, "Templum Apollonis Alaei," AttiMGrecia 1932 (S. 1-184), S. 61-67, Taf. VI, VII; D. Mertens, in "Crotone,"23. CMGr, 1983, S. 213-22 1, Abb. 5, 6, Taf. XXVII:a, b. Kaulonia (heute in Kroton): Mertens a. a. 0. Taf. XXVIII:c; Van Buren (Anm. 11 oben), S. 150, Nr. 2, Abb. 69. 40 Mertens (Anm. 39 oben), Abb. 5:a, Taf. XXVIII. 41 Krotonentsandtezwar hervorragendeAthleten zu den Spielen, doch das Verhaltnisder Stadtzum olympischen Heiligtum war zeitweise alles andere als gut, s. Athenaios, XII.522C: "Spaterjedoch versuchtendie Krotoniaten,wie Timaeus erzahlt, die olympischen Spiele auszuschalten,indem sie zum selben Zeitpunkt 34

35

244

MADELEINE MERTENS-HORN AND LUISA VIOLA

Gegenuiberstellungder mit bunten Ornamentenbelebten grossen glatten Friesflachenund der stark plastisch hervorgehobenenElemente des dorischenGebalks:der Mutuli, Triglyphen, Regulae, und Guttae. Es ist die Spiegelung einer Architektur,die sowohl die Elemente des dorischen Gebalks als auch den Schmuckfriesals Bestandteileeines Bauwerks kannte und vielleicht gleichzeitig an verschiedenenStellen verwandte.2 Im ebenfallsachaischenMetapont haben sich vier leidernur sehr kleine Fragmentevon solchen Verkleidungsplattengefunden.43 Das wohl altere Paar (P1. 35:e, f) stammt von einer Verkleidung, die, aihnlichwie diejenigeaus Ciro-Crimisa,unter der Stirnfliche eine glatte Taenia mit Regulae und Guttae aufwies. Ein weiteres, aus gutem Ton sorgfaltigausgeformtes Stuck bewahrt den Rest einer Regula mit zwei Guttae unter einem Rundstab (P1.35:g). Es k6nntezu einer gleichenHaingeplattemit Doppelrundstabund Regulae unter dem Ornamentfriesgehort haben, wie sie in Olympia erhalten ist.44 Wie an dieser betraigt der Abstand zwischen der Oberkante des unteren Rundstabes und der Unterkante der Tropfen 7 cm, die Platte hatte also wohl die gleichen Masse. Das zweite, aus anderenTon sehr unprazise ausgeformteFragment mit Resten eines dicken weissen Uberzugs konnte von einer Ersatzplatte fur dasselbe Dach oder von einem weiteren solchen stammen (P1. 35:h). Sind die Fragmente in Metapont auch verschwindendklein, so bezeugen sie doch, dass es auch hier mindestens zwei verschiedeneDachverkleidungenmit Elementen des dorischenGebalks gab. Ein kleines Dach aus dem Stadtheiligtumin Paestum (P1. 36:a) bietet die naichsteParallele zu den sogenanntenHornerantefixender Dacher in Olympia. Hier sind die Antefixe unter den mittleren Spitzen mit hangendendreiblattrigenPalmettchenbemalt. Die Traufsima45ist im Profil und der Bemalung einer sehr alten Sima in Metapont eng verwandt. Die Giebelsima tragt ein Anthemion, das in derselben Technik eingestempeltist wie die Ornamenteder Schmuckfrieseder Dacher mit Elementendes dorischenGebalks in Olympia, und dessen Lotosbltiten und Palmetten in der Form denjenigen auf der Geisonverkleidung des Daches mit der Blattstabsima(P1. 4:f) sehr nahe kommen. Das Dachlein in Paestum ist nun besonderswichtig, weil die Zahlbuchstabenauf der Ruckseiteder Platten nach einstimmigemUrteil der Epigraphikerausnahmlos achaiischund um 580/70 zu datieren sind.46P. Moreno hat in dem kleinen Bau das Schatzhausvon Sybarisvermutet,da diese achaischeStadt die Gruindungvon Poseidoniaeinst veranlassthatte. Die eben genannte Blattstabsimain Olympia mit den ungewohnten, sehr weit uberhangenden Blattspitzen findet ihre nachste Parallele nun gerade wiederum in einer Sima Spiele mit sehr reichen Silberpreiseneinrichteten."Diese Konkurrenzhaltungware kaum vorstellbar,wenn Krotonin Olympia durch ein Schatzhausvertretengewesen ware! in 21 CMGr, 1981, S. 118-121. 42 Vgl. D. Mertens, "Urbanisticae architettura," 43 Publikationin Vorbereitung. Olympia II, Taf. CXVI:5; Mallwitz, 1980, Taf. 100:5. 45 N. Daniele, Paestum, Hypothese und Wirklichkeit, Paestum 1979, Abb. 55; vgl. Mallwitz, 1968, S. 142-144. RendLinc 18, 1963, S. 210-229; M. Guarducci, 46 p. Moreno, "Numerazionedi elementi architettonici," Epigrafiagreca I, Rome 1967, S. 113; L. Lazzarini, muindlich.

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

245

aus einem kleinen Heiligtum in Francavilla bei Sybaris (P1. 36:b),47deren Sockelzone jedoch mit einem Schachbrettmusterbemalt war. Einige Fragmentevon plastischenFiguren, von Menschen und Pferden (P1. 36:c, d), die als Akroteredieses Daches geschmuickt haben duirften,sind nach den grossen,schwerhangendenPerllockenwohl ebenfallsnoch im ersten Drittel des 6. Jahrhundertsentstanden.8 Aus all den eben genanntenBeispielenbestatigtsich die Annahme, dass die Werkstatt, die diese drei Dacher in Olympia geschaffen hat, aus Unteritalien kam, und dass ihr Wirkungskreis urspruinglichdie AchaerstadteSybaris, Metapont, Paestum, und Kroton umfasste. Man darf sie also sicher als achaisch, oder von Griechenlandaus, als kolonialachaischbezeichnen. A. Mallwitz datierte die Dacher mit Elementen des dorischen Gebalks jedoch viel spater, ins dritte Viertel des 6. Jahrhunderts,und Le Roy erwog bei der Besprechungdes Daches 30 in Delphi ebenfalls eine juingereEntstehungszeit,fur die vor allem historische Uberlegungen ausschlaggebendwaren. Fur den durch die Form der Zahlbuchstabenauf dem Dach in Paestum nahegelegtenfruiherenAnsatz sprechenjedoch auch folgendeGrunde: Die Spiralwellen der einen Gebalkverkleidungkennen wir vom ersten Dach des Artemistempelsin Korfu, und ich sehe keinen formalen Grund, diese als wesentlich alter zu betrachten. Das sehr einfache Flechtmotiv der Dacher mit Geison und Mutuli, dessen Bander in je drei genau gleich breite Bahnen aufgeteilt und auf den ausseren schwarz bemalt sind, findet seine nachste Parallele-ausser an den schon genannten Geisonverkleidungenaus dem Demeterheiligtumin Siris-im Simaornamentdes altestenTempels in Lokri mit den grossen Wandverkleidungenaus Terrakotta, der noch in 7. Jahrhundert entstandensein dtirfte.49Die Anthemienaus Lotosbltitenund Palmettenhaben einen wichtigen Vorgangerin einer Traufsima in Thermos (P1. 36:e), die von H. Koch dem Bau mit den kleinen Metopen zugeschriebenworden ist.50Sie zeugt von sehr friiher Verwendung dieses Motivs auch an Dachern Nordwestgriechenlands,von wo aus dieses dann auch nach Unteritalien gekommensein konnte. Das schon erwahnte Dach 30 mit Geison und Mutuli in Delphi, das ein genaues Pendant in Olympia hat, dtirfte, wie Le Roy schon erwogen hat, ebenfalls einem Schatzhaus einerwestgriechischenStadtgehorthaben.Von einemvon SybarisgestiftetenWeihgeschenk aus Kriegsbeutespricht Strabo (Ix.38) nur mit vagen Worten in Zusammenhangmit anderen Weihungen des Kroisos, des Gyges und der Spineten in delphischenSchatzhausern. Doch Metapont hat zur Zeit seiner ersten landwirtschaftlichenBlute den bertihmtengoldenen the'ros,wohl ein goldenesAhrenbundel,dorthingeweiht. Strabon(VI.I.I 5) nennt den the'rosim Zusammenhangmit der prakolonialenpylischen Besiedlung, was sicher falsch ist, da es zu deren Zeit in Delphi noch kein Apollonheiligtumgab. Dennoch dtirfen wir 47 D. Mertens, "Die Bauten auf der Motta," in AttiMGrecia,N.F. 21-23, 1980-1982 (1983; S. 143-171), S. 164-166 (Dach B), Abb. 65, Taf. XCV:b, d. 48 Ebd., Taf. XCV:e, f. 49 De Franciscis (Anm. 8 oben), S. 68, Abb. 46, 47. 50G. Soteriades,<<'AvaO-KabaL E'v0ep?>>, 'Eq 'ApX 1900 (S. 161-212), S. 210, Abb. 9; GFR, S. 115, Nr. 107; Koch, S. 68, Nr. 1.

246

MADELEINE MERTENS-HORN AND LUISA VIOLA

annehmen,dass die Weihung in der Fruhzeit der Stadt erfolgt ist. Das wertvolle Geschenk konntesehr wohl Anlass fur die Errichtungeines Schatzhausesgewesen sein,51dessen Dach dann eben von der achaischenWerkstattin Olympia geliefertwurde. Vielleicht konnenwir sogar so weit gehen zu sagen, dass das dem Dach 30 in Delphi gleiche Dach mit Geison und Mutuli in Olympia52ebenfalls das des Hauses der Metapontinerwar, wahrend das Dach mit den Triglyphen53das Haus der Sybaritenschmuickte.54 So bleibt uns jetzt noch das Dach mit der Blattstabsima,das nach den eingangs gemachten UCTberlegungen zu den Fundumstandendas altere Schatzhaus der Sikyonier gedeckt haben duirfte.Auf den ersten Blick wird das unmoglicherscheinen,man wird fur das Haus der so nahe bei KorinthgelegenenStadteine fruihekorinthischeBlattstabsimaerwarten. Doch hat sich von einer solchen eben nicht die geringste Spur gefunden. Wir wissen jedoch, dass die Politik des Kleisthenesin Sikyon den Kypselidenin Korinthausgesprochen feindlich war,55und wir wissen, dass Sikyon auch die Heimat des beruihmtenTonbildners Butadeswar, der als ersterTondachermit menschlichenKopfen schmtickte.56Diese Dachgattung, die Dacher mit Gesichterfriesen,wurde dann besondersin Nordwestgriechenland, in Thermos, Kalydon, und Korfu heimisch, doch dtirfenwir annehmen, dass die Heimatstadt des Butades sie ebenfalls kannte, m6gen die wenigen bisher aufgefundenenDachterrakotten aus Sikyon der zweiten Halfte des 6. Jahrhunderts auch alle korinthischsein.57 Betrachtenwir noch einmal das Dach in Olympia: Mit Anthemien geschmuickteVerkleidungsplattenund eine Blattstabsima,uiberdie kleine, an den Deckziegeln befestigteKopfchen hervorblicken.58Ich habe oben schon auf die fruihe,wohl noch in 7. Jahrhundert entstandene Sima mit Lotosbluten aus Thermos hingewiesen. Auch Schmuckplattenmit 51 Die Wichtigkeit der Weihung wird durch die Gerstenahrebezeugt, die im 6. Jahrhundertdas meistbezeugte MuinzbildMetaponts ist. G. Roux ("Tresors,Temples, Tholos,"in Templeet sanctuaire,Seminairede recherche 1981-83, hrsg. G. Roux, Lyon 1984, S. 157) zeigt, dass wertvolle Weihgeschenke,die geschuitzt werden mussten, der Anlass zum Bau der Schatzhauserwaren. 52 Le Roy, 1967, Taf. 27, 28. 53 Mallwitz, 1980, Taf. 101:4. 54 Eine schwere Bronzetafeldes fruhen 6. Jahrhundertsaus Francavillabei Sybaris mit der Weihung des Athleten Kleombrotosan die Gottin Athena zum Dank fur einen Sieg bei den olympischenSpielen bezeugt die Wichtigkeit,die das olympischeHeiligtum fur die Sybaritenschonzu dieser Zeit hatte, vgl. Guarducci(Anm. 46 oben), S. 110-111. Zum Dach 31 in Delphi: Nach Athenaios (xII.52oA)schicktendie Sybariteneine Gesandtschaftnach Delphi, um nach der Lebensdauerihrer Stadt zu fragen. Da an ihr auch der Weise Amyris aus Siris, Vater des Damasos, des Freiers der Agariste, teilnahm, duirftediese Gesandtschaftnicht lange vor oder nach die Hochzeit der Agariste zu datieren sein. Die reichen Sybariten sind sicher nicht mit leeren Handen nach Delphi gekommen.Vielleicht ist fur ein von ihnen geweihtes (das von Strabo,Ix.38 erwahnte?) Geschenkein kleiner Thesauros errichtetworden, der dann mit dem Dachlein 31 aus der achaischenWerkstatt geschmiicktwurde. 55 Nicolaus Damascus, FGrHist Nr. 90, F 61. Vgl. Griffin, S. 49. 56 Mertens-Horn, S. 30-65; N. Winter, "Archaic Architectural Terracottas decorated with Human Heads,"RM 85, 1978, S. 27-31; M. Torelli, "Terrecottearchitettonichedi Graviscae una nota a Plinio NH XXXV,151-52," in Studi in onoredi F. Magi, Perugia 1979, S. 310-314; Schwandner,1985, S. 126-127. 7 Griffin, S. 99; GFR, S. 60; A. K. Orlandos, <<'Avao-Ka4' ILKVW^VO! rov 1937?, HIpaKrLKa' 1937 (S. 94-96), S. 95-96, Abb. 3, 4. 58Vgl. Olympia II, S. 202, Abb. 26; Heiden, PI. 4:d.

ARCHAISCHE TONDACHER WESTGRIECHISCHER TYPOLOGIE

l'

~~~~~~

247

''s

.~~~~~~~~~.

FIG.2. Blattstabsimamit Kopfantefixen.a: Thermos. b: Korkyra

vielgliederigenFlechtbandernsind dort gefundenund von Koch dem Dach mit den altesten Kopfantefixenzugewiesenworden.59Vor allem aber gab es auch hier in Thermos sowie am Tempel in Mon Repos-Heiligtum in Korkyra die hinter einer Blattstabsimabefestigten und uber diese hinwegblickendenKopfe (Fig. 2).60 Das strukturellso Ungewdhnliche,ja Befremdlichedes Daches in Olympia findet in diesen nordwestgriechischenDachern seine Erklarung. Das schmale, sich nach unten verjiungendeAntefix-Gesicht (P1. 4:d) mit den mandelformigenAugen und dem in flachem Bogen die Stirn begrenzendenHaarwulst ist gewiss nicht das einer Gorgo, sondernwohl das eines menschlichenKopfes spatdadalischer Tradition. Die Einzelformender Blattstabsimasind sichervon denen der nordwestgriechischen Dacher verschieden,es sind die der acha'ischenWerkstatt.Nordwestgriechisch,oder eben urspriinglichsikyonisch,ist indessen die Idee, menschlicheKopfe tiber einem Blattstab, oder uberhauptam Dachrand anzubringen.Somit konnte dieses Dach sehr wohl das des ersten Schatzhausesder Sikyoniergewesen sein.61 59GFR, S. 83, Nr. 42, Taf. 38, Abb. 141; Koch, S. 70, Nr. 2. 60KorkyraI, S. 153-154, Abb. 136, 137. 61 Zum Datierungsproblemder von Pausanias genannten ratselhaften Bronze-Thalamoi im Sikyonierhaus, s. Griffin, S. 101-106. Ein "Schrein"mit Formen, die Pausaniasals "jonisch"bezeichnenkonnte,ist im

248

MADELEINE MERTENS-HORN AND LUISA VIOLA

Eine gleiche Blattstabsimaverzierte zur selben Zeit auch einen Bau auf der archaischen Agora von Elis.62N. Yalouris hat schon darauf hingewiesen, dass der sich in diesem Dach zeigende enge Kontaktmit dem olympischenHeiligtum erst nach der Wiedergewinnung der Herrschaftder Eleer uiberOlympia moglichwar. Fur die 50. Olympiade,um 580, haben die Eleer bereits wieder die beiden Hellanodikenbestellt. Dieses Datum bietet somit einen willkommenenterminus ante quem fur das Dach des Sikyonierhauses,das damitbald nach 580 entstandensein duirfte. Daruber hinaus wirft die Sima in Elis aber auch neues Licht auf weitere historische Zusammenhange.Nach Pausanias war die wichtigste Weihung im Schatzhausder Metapontiner eine chryselephantineStatue des Endymion. J. Berard hat dies mit einem alten Kult fir diesen Heros in der Stadt Metapont erklart,deren Griindungzuweilen ja auf den Sohn des Endymion,den Eleer Epeios zurtickgefuihrtwurde.63Es handelt sich um denselben Endymion, der als K6nig von Elis als Mitbegriinderder olympischenSpiele galt und dessen Grab im Heiligtum verehrtwurde (Pausanias,VI.20.9). In diesem gemeinsamenInteresse der Eleer und Metapontiner fur den alten Kult dtirfte auch die sonst schwer verstandlicheTatsache begrtindetsein, dass der unbekannteBau in Elis mit einem Dach aus einer achaischenWerkstattgeschmuicktwar. MADELEINE MERTENS-HORN DEUTSCHES ARCHAOLOGISCHES INSTITUT

Via Sardegna,79 I-00187 Rome, Italy

mittleren 7. Jahrhundert in der Tat sehr schwer vorstellbar,wahrend gegen den Schrein mit den dorischen Formen (Elementedes dorischenGebalks?)nichts einzuwendenist. A. Griffinversuchtdas Problemzu l6sen, indem sie im dorischenThalamos das ursprunglicheWeihgeschenkdes Myron sieht, im jonischen dagegen das Weihgeschenkdes Kleisthenes,das dieser anlasslich seines Wagensieges (wohl 576 v. Chr.) und in Erinnerung an den Wagensieg des Myron gestiftet hatte. Bei dieser Gelegenheit konnte das Schatzhausrenoviert und mit einem neuen, zeitgemassenDach gedecktworden sein. 62 S. Anm. 36 oben. Vgl. N. Yalouris in 20. CMGr, 1980, S. 12-13. 63J. Berard, La Magna Grecia (italienische Ausgabe des in Anm. 31 oben genannten Werkes), Torino 1963, S. 333. Epeios wurde als mythischerGruindervon Metapont in einem Heiligtum in Lagaria bei Metapont verehrt, wo er sein Werkzeug, mit dem er das troianische Pferd geschaffenhatte, der Athena geweiht haben soll. Zur Vermischungdes elischen und des phokischenEpeios s. Berarda. 0. S. 330-333.

PLATE 33

_

_

~~~~~~~~~~~I

a. Olympia, Schatzhausder Geloer

b. Kaulonia d-. Paestum M.r--EN-HONADLVOAACASH

c. Kaulonia . ODCE

INDLH

NLMI

PLATE 34

a. Paestum b. Paestum

c. Paestum, sog. Basilika

d. Paestum, sog. Basilika

e. Paestum, sog. Basilika

f. Olympia, Schatzhausder Epidamner

PLATE 35

~~~~U I

a. Metapont

b. Ciro-Crimisa,Apollontempel

d. S. Anna bei Kroton

c. S. Anna bei Kroton M.MRESHR

N

.Vo4

RHISH

ODCE

NDLH

N

e. Metapont

.

ap

LMI

PLATE 36

a. Paestum

b. Francavillabei Sybaris, Acropoli sulla Motta

X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

c. Francavillabei Sybaris, Acropoli sulla Motta

' e Thermos

e. Thermos

'j

d. Francavillabei Sybaris, Acropoli sulla Motta

..-

I

PLATE 37

-----------~'?-

K

j

;

&

IP

g ; f~~~~~~~~~~~~~~~~~~~~~~~i

nhI.-X

>\>

j

~~~~P

1 Tetto fittile arcaico da Siris (Policoro)

7\ _;____

/t

;;-.rV

PLATE 38

4

-4

_ ------------1

~~~~

-m

m

4

2 Tetto fittile arcaico da Siris (Policoro)

ANHANG: UN TETTO FITTILE ARCAICO DA SIRIS

249

ANHANG: UN TETTO FITTILE ARCAICO DA SIRIS64 (PLATES

37, 38)

Nel Museo Nazionale della Siritide di Policoro, la cittadina sorta oltre trent'anni fa sulla costaionica della Basilicata,nella zona che un tempo appartennea Siris65e successivamente ad Heraclea, sono conservatiquattroframmentiarchitettoniciprovenientidall'areadel cosiddetto"tempioarcaico"(zona B). Essi sono: 1. Frammentodi cassettafrontonale Pls. 37, 38 Inv. 38834. Alt. 0.189, Largh. 0.141, Alt. totale doppio tondino 0.108, Spess. piastra frontale 0.028 m. 20. CMGr, 1980, p. 87, tav. V:1. Argilla rossiccia, scarsamentedepurata, con granuli biancastri (pietrisco) di varia grandezza (2 mm di diametroin media). Strato d'argilla superficiale beige-rosato, piuttosto sottile. II frammento appare scheggiato soprattutto sul secondo tondino dall'alto, oltre che fratturato nella parte inferioredella piastra frontale e sul lato destro. II lato sinistro,invece,conservaben evidenteil taglio verticaleoriginario. Superficieerosa in piCu punti, ma ancoraben leggibile nella sua decorazione. Profilo (dall'alto): due tondini dall'andamentoirregolare,il primo dei quali si presentaunghiato,all'incontrocon la tegola superiored'appoggiodella lastra; parte della piastra frontale originaria,formantecon la residuategola d'appoggioun angolo di circa900. Decorazionedipinta (dall'alto):sui due tondini,fasce trasversalinere sul fondo chiaro, correntiverso sinistra e affiancatesu entrambii lati da liste ugualmente nere e parallele;nella sovrapposizionedei tondini ne le fascene le relativeliste si corrispondonodall'altoal basso, bensi le sole liste chiare, interposte e risparmiate;l'effettocomplessivoe cosi quello di un'insolita scacchiera,sottolineatainoltre dalle righe nere orizzontali che delimitano in alto e in basso il secondo tondino. Sulla piastra frontale, motivo a treccia,

molto probabilmentesemplice," dalle volute alternativamentenere e-rosso-brune,queste ultime profilate di nero; una linea ugualmente nera racchiude superiormentee inferiormentele volute, mentre un cerchiello concentricoavvolge l'occhio della treccia, riempito da un dischetto nero. L'andamento complessivodel motivoe da sinistraa destra.67 La lastra doveva essere lavorata in un unico pezzo (non compare,infatti, alcun segno di sutura tra la tegola verticalee quella d'appoggio)e successivamente rifinitaa seccocon una lama (si notino,a questo proposito, i due solchi orizzontalialla base della unghiatura sul retro del primo tondino, oltre ad una certa irregolaritanella lavorazionedel retrodella lastra, in particolareall'altezzadella curvatra piastra frontale e tegola d'appoggio). 2.

Frammentodi cassettalaterale P1.38 Inv. 47265. Alt. 0.122, Largh. 0.255, Alt. totale doppio tondino 0.101, Spess. piastra frontale 0.0275 m. 20. CMGr, 1980, p. 87, tav. VI:2. Argilla giallastra tendente al grigio, con qualche piccola area giallo-rosata, ricchissima di grani bianco-grigi (pietrisco)di notevoli dimensioni (fino a 7 mm di lunghezza). Denso strato d'argilla superficiale giallastra, scarsamente depurata. II frammento,fratturatoquasi all'incontrodella piastra frontale con il doppio tondino e nella tegola superiore d'appoggio,appare ampiamentescheggiato sul primo tondino dall'alto, sbrecciato sul

64 Desidero ringraziarevivamentela Dott.ssa Madeleine Mertens-Horn per i suggerimentie la fertilissima discussione,oltre che per l'opportunitaoffertamidi presentarequi alcuni dei frammentiarchitettonici,ancora in corso di studio, oggetto della mia prossimadissertazionedi laurea. 65 Sul problemadella reale ubicazionedi Siris, v. Adamesteanu(sopra, nota 29). 66 II confronto,alquanto persuasivo, con il rivestimentodel geison ad Olimpia (v. sopra, Heiden, p. 44, P1. 4:a: con identico motivo decorativodei tondini, struttura ugualmenterobusta degli stessi, conformazione assai simile della treccia) induce a considerarecome semplice la guilloche del frammentodi Policoro, nonostante alcune differenzenella disposizionedei colori all'internodelle volute. 67 Nella ricostruzionedi profilo e decorazionesi e preferitonon avanzare alcuna ipotesi a proposito della terminazioneinferioredella cassetta.

250

MADELEINE MERTENS-HORN AND LUISA VIOLA

secondoed eroso in piCupunti della sua superficie decorata. Esso conserva inoltre parzialmente intatto il taglio lateraledestro. Profilo (dall'alto):due tondini dall'andamentoirregolare, pressappococome nel frammentoInv. 38834; stessa unghiaturadel primotondino,anche se leggermente diverse sono qui le proporzioni;diversoe invece l'angolo che la piastra frontale, in questo caso assai parzialmente conservata, forma con la tegola d'appoggio, notevolmente massiccia al suo attacco (circa970). Decorazione dipinta (dall'alto): sui due tondini, l'identicadecorazionea fascetrasversalie liste parallele del frammentoInv. 38834 (1); anche il senso di percorrenzadelle fasce,da destraa sinistra,e lo stesso. La zona decoratadella residuapiastrafrontalemostra soltantola riga nerache anchenel suddettoframmento delimita inferiormenteil piCu basso dei tondini. La faccia superiore della tegola d'appoggioreca inoltre una larga pennellatadi colorenero in senso verticale, affiancatada un punto a brevedistanza,pressappoco al centrodell'attualelunghezzadella tegola. Anche questa lastra dovevaessere lavoratain un unico pezzo e rifinita successivamentea secco con una lama: ben visibili sono qui i segni di un rimodellamento del retro,all'altezza della larga curvatra doppio tondinoe tegola d'appoggio,come al passaggiodi una "pialla"dalla tegola ai tondini e viceversa.68 3.

Frammentodi sima (frontonale?) Pls. 37, 38 Inv. 49881. Alt. 0.173, Largh. 0.158. Spess. mass. conservato0.037 m. Inedito. Argilla rosatama tendenteal giallastroe al grigio, ricca di granuli bianco-grigi (pietrisco) di varia grandezza (fino a 4 mm di diametro).Stratod'argilla superficialegiallastra, piuttosto sottile e depurata. Interamentefratturata,la lastra presenta pressoccheintegra buona parte del taglio laterale sn.; i colori sono mal conservati,mentre la superficie decoratae erosa in piCupunti, nonche leggermente scheggiata. Profilo (dall'alto):il frammentocomprendela parte inferioredel cavettoe l'inizio del listello di base,senza

elementi di separazione rilevati (tondino) tra una zona e l'altradella sima;il cavettosi preannunciaalto e poco profondo,mentrelo spessoredella lastra si assottiglialeggermenteall'altezzadella curvatra lastra verticaleintera e tegola inferiored'appoggio. Decorazione dipinta (dall'alto): sul cavetto, foglie dorichenere, doppiamenteorlate di bianco e di nero; gli interstizi tra le foglie sono riempiti di colore rosso-bruno. Sul listello di base, separato dal cavetto per mezzo di una sottile fascia nera, l'accennodi un triangolougualmentenero. II disegno e tracciatocon un andamentoirregolaree piuttostotrascurato(si notino, ad es., la grossolanita dell'orlonero delle foglie o l'imprecisionecon la quale sono calcolatele distanzetra un elemento e l'altro della decorazione;o la fatturaaffrettatadel triangolo sul listello di base, sovrappostoalla fascia terminale del cavetto). 4.

Frammentodi rosetta P1. 38 Inv. 47266. Largh. mass. conservata0.069, Diametro nucleo centrale 0.019, Spess. mass. conservato 0.024 m. 20. CMGr, 1980, p. 87, tav. VII:1 Argilla giallastratendenteal grigio, con lievissime sfumature giallo-rosate, ricca di granuli biancogrigi (pietrisco)di varia grandezza (2 mm di diametro in media). Strato d'argilla superficialeocra chiaro, piuttosto sottile. Interamentefratturato,il frammentoconservacospicuetraccedi colorenero nei solchi dei petali e alla base degli stessi, oltre che nei cerchiellicircondantiil nucleo della rosetta. Profilo: maggiormenteconcavo al centro, in prossimita del nucleo, esso si rialza ai bordi, mantenendo inoltreuna zona leggermentecavaper ciascunpetalo. Decorazione: sulla faccia anteriore sono parzialmente visibili 6 degli 8 petali originari, dai contorni impressi e riempiti di nero, intorno ad un nucleo centralenon dipinto,a rilievo leggermenteappiattito e circondatoda due cerchielliconcentriciugualmente neri ed impressi. La facciaposteriore,pressoccheverticale,e rivestitadi uno stratod'argillapiCu finee fluida,dellostessocolore dell'argillain frattura,levigatoe lievementeporoso. LUISA VIOLA

Via FrancescoBaracca 1Oa I-85013 Genzano di Lucania, PZ, Italy 68

Per una simile lavorazionedel retro delle sime, cfr. Le Roy, 1967, p. 203.

PLATE 37

-----------~'?-

K

j

;

&

IP

g ; f~~~~~~~~~~~~~~~~~~~~~~~i

nhI.-X

>\>

j

~~~~P

1 Tetto fittile arcaico da Siris (Policoro)

7\ _;____

/t

;;-.rV

PLATE 38

4

-4

_ ------------1

~~~~

-m

m

4

2 Tetto fittile arcaico da Siris (Policoro)

SCHEIBENFORMIGE AKROTERE IN GRIECHENLAND UND ITALIEN (PLATES D

39-42)

IE BISHERIGEN UNTERSUCHUNGEN zum scheibenformigen Firstschmuck

archaischergriechischerTempel haben gezeigt, dass dieser wahrscheinlichin Lakonien entstandenist und dortauch besondersverbreitetwar.1 Die sich in einzelnen Fallen zu imposanterGrosseentwickelndenFirstscheibennahmenoffenbarden kreisformigenQuerschnitt der lakonischenFirstziegel auf, deren giebelseitigenEnden sie zu verschliessenund zu schmiickenhatten. Viele Details der Genesis und der Verbreitung dieser lakonischen Scheibenakroteresind jedoch noch ungeklart. So sind zunachst die Entwicklungsphasen dieser Akroterformdarzustellenbevor die Frage erortertwird, inwieweit ahnliche schildformige Firstakroterean den Tempeln der Griechen in Suiditalienund an etruskischen Bauten von lakonischenVorbildernabhangigsein konnen. Eine Betrachtungder lakonischenAkroteregeht am zweckmassigstenvonjener Gruppe der Firstscheiben aus, die stilistisch eng an das langere Zeit bekannte Akroter des Heraions von Olympia angeschlossenwerden mtissen. Es handelt sich besondersum die nachfolgendaufgefuihrtenFundstuicke: LISTE

B (AKROTERE

DER HERAION-GRUPPE)

1. Artemis-OrthiaNr. 10 (zusammenmit Lakonisch-III-Keramik,580-550 v. u. Z., gefunden):2

inneresRandstuick mit 3 Schuppenreihen, Zahnband,undkonvexemProfil. 2. Artemis-Orthia Nr. 5: allseitig gebrochenesRandsttickmit Treppenmuster(?)und 4 Schuppenreihen.3 3. Aigina, Apollontempel: allseitig gebrochenes Randstuck mit 9 Schuppenreihen und Treppenmuster.4

4. Amyklai (Fig. 1, P1.39:a):fast vollstandigesAkrotermit geraderBasis;Rand mit Zackenkranz, Blattkranzkehle,Rundstab, 5 Schuppenreihen,doppeltem Flechband und konvexem Profil sowie Innenscheibemit Blattkranzkehle,dreifachemRundstab,abgesetztemZahnband,abgesetztem Flechtband,Blattstabkehle,dreifachemRundstabund zentralemWulstring.5 1 Neben den alteren Arbeiten von 0. Benndorf,H. Koch, K. A. Rhomaios, und K. Volkert sind v.a. die nachfolgendabgekiirztzitierten Beitragezu nennen: = W. S. George and A. M. Woodward in R. M. Dawkins, The Sanctuaryof Artemis Orthia AO (JHS, Suppl. 5), London 1929, S. 117-144 = M. Y. Goldberg,"ArchaicGreek Akroteria,"AJA 86, 1982, S. 193-217 Goldberg Lauter-Bufe = H. Lauter-Bufe, "Entstehungund Entwicklung des kombiniertenlakonischen Akroters," AM 89, 1974, S. 205-230 = A. Mallwitz, "Ein Scheibenakroteraus Olympia,"AM 83, 1968, S. 124-146 Mallwitz Yalouris = N. Yalouris, "Das Akroterdes Heraions in Olympia,"AM 87, 1972, S. 85-98 2 AO, S. 118, Abb. 87, 136, Nr. 10. 3 AO, S. 118, 135, Nr. 5, Taf. 24; Goldberg,Nr. D 32. 4A. Furtwangler, Aegina, Munchen 1906, S. 485-486, Abb. 403; W. Wurster, Alt-Agina 1.1, Mainz 1974, S. 79; Goldberg,Nr. D 2-3. 5 Mallwitz, S. 133, Taf. 48; Goldberg,Nr. D 4.

VOLKER KASTNER

252

A.1.1 - 4 I

~~I

L

C.6

B.10

FIG. 1.

B.4

B.6

C.1

B.21

Masstab 10 cm. Oben: Fragmente von Scheibenakroterenaus dem Artemis Orthia Heiligtum in Sparta (A.1.1-4) Unten: Heraion in Olympia (B.10), Amyklai (B.4; nach Fotos), Bassai A (B.6), Mantinea (C.1), Olympia (B.21), Tegea (C.6)

SCHEIBENFORMIGEAKROTEREIN GRIECHENLANDUND ITALIEN

253

5. Sparta, Akropolis:Randfragmentmit Zackenkranz,Blattkehle, Rundstab, 5 Schuppenreihen und doppeltem Flechtband sowie Fragmente der Innenscheibe mit abgesetztem Zahnband, Band mit Zickzackmuster,doppelterRundstab,Streifen .. . abgesetztesZahnband,abgesetztes Treppenmuster,Blattkranzkehle,doppelterRundstabund zentralemWulstring mit Offnung.6 6. Bassai, AkroterA (Fig. 1): Randfragmentemit Zackenkranz,Blattkehle,Rundstab, 10 Schuppenreihen,doppeltemFlechtbandund konvexemProfil sowie Innenscheibemit Blattkehleund Rundstab,dreifachemRundstab,Blattstab,laufendemHund, Zickzack-Treppenmuster,Blattstab, abgesetzterStreifen,dreifachemRundstabund zentralerOffnung.7 7. Artemis-Orthia Nr. 2 (zusammen mit Lakonisch-III-Keramikgefunden):8Randsttick mit Zackenkranz,Blattkehle,Rundstab, 5 Schuppenreihenund Flechtband. 8. Artemis-OrthiaNr. 6 und 7: allseitig gebrocheneRandstuickemit 6 Schuppenreihensowie mit 3 Schuppenreihenund Flechtband.9 9. Thasos, Herakleion (vom gegen 540-525 v. u. Z. datierbarenPolygonalbau?):mehrereRandstuickemit Zackenkranz,Blattkehle,Rundstab,4 Schuppenreihen,und Streifen.1I 10. Olympia, Heraion (Durchmesserca. 2.40 m, um 590 v. u. Z.; Fig. 1, P1.40:a, b):1IRand mit Zackenkranz, Blattkehle, Rundstab, Schuppenreihen (innerer Abschluss unbekannt) sowie Innenscheibe mit Blattkehle, dreifachem Rundstab, Kranz mit reliefierten Blattzungen, abgesetztes Volutenhakenband, abgesetztes Treppenmuster, dreifachem Rundstab (innerer Abschlussunbekannt). 11. Artemis-OrthiaNr. 11 (unter der Sandschichtvon 570-560 v. u. Z. gefunden zusammen mit Lakonisch-II- und III-Keramik):2 zwei Fragmenteeiner Innenscheibemit Kranz reliefierter Blattzungenund Volutenhakenband. 12. Artemis-OrthiaNr. 12: Fragmentder Innenscheibemit Kranz reliefierterBlattzungen.13 13. Artemis-OrthiaNr. 13 (zusammen mit Lakonisch-III-Keramikgefunden):14 Fragment einer Innenscheibemit reliefiertenBlattzungen. 14. Artemis-Orthia Nr. 30 (zusammen mit Lakonisch-III-Keramikgefunden):15Randfragment mit Zackenkranz,Blattkehle,Rundstabund Schuppen. 15. Artemis-OrthiaNr. 18: Randfragmentmit Zackenkranz,Wulstprofil(?),abgesetzterLeiste.16 16. Kynouria:zwei Randstuckevon unterschiedlichgrossen Scheibenmit Zackenkranz,schmaler Leiste, Blattkehle,Rundstab,und noch zwei Schuppenreihen.17 6 A. M. Woodward,"Excavationsat Sparta 3. The Acropolis,"BSA 28, 1926-1927 (S. 37-48), S. 40-42, Abb. 2; Lauter-Bufe, S. 214, Anm. 39. 7 K. A. Rhomaios,?<'EK 'rov'apXaLo'repovvaov r719(FLyaXelasvo,'ApX'E4 1933 (S. 1-25), S. 1-7, Abb. 1, 2, Taf. 1, 2; N. Yalouris in Acta of the XI InternationalCongressof ClassicalArchaeology(London 1978), London 1979, S. 94, Taf. 43:c;Goldberg,Nr. D 8. 8 AO, S. 118,135, Nr. 2, Taf. 23; Goldberg,Nr. D 29. 9 AO, S. 118, 135 f., Nr. 6,7, Taf. 24; Goldberg,Nr. D 33, 34. 10 Etudes thasiennesI, Paris 1944, S. 46 if., Nr. 18-20, Abb. 18. Yalouris;A. Mallwitz in Die Funde aus Olympia,Athen 1980, S. 145 zu Taf. 98:1; Goldberg,Nr. D 15. 12AO,S. 118, 136, Nr. 11, Abb. 87; zur Datierung der Sandschicht,J. Boardman,'Artemis Orthia and Chronology,"BSA 58, 1963, S. 1-7; zur Ornamentik,H. E. Searls und W. B. Dinsmoor, "The Date of the Olympia Heraion,"AJA 49,1945 (S. 62-80), S. 70-72; Goldberg,Nr. D 24. 13AO,S. 118, 136, Nr. 12, Taf. 24; Goldberg,Nr. D 25. 14AO,S. 118,136, Nr. 13; Goldberg,Nr. D 38. 15 AO, S. 131, 139, Nr. 30 (als Giebelsima gedeutet),Taf. 26. 16AO,S. 120, Abb. 90,137, Nr. 18; Goldberg,Nr. D 27. 17

K. A. Rhomaios,<
Abb. 4-6; Goldberg,Nr. D 11, 12.

flpaKrtKa'

1953 (1956; S. 251-253), S. 253,

254

VOLKER KASTNER

17. Artemis-OrthiaNr. 16: Randstuickmit Zackenkranz,Rundstab,und abgesetzterLeiste.18 18. Menelaion: Randstuickmit Zackenkranz,breitem und schmalen Streifen in einer Ebene; zuruckgestufterStreifenmit Zinnenmaanderund Ansatz eines stark zuruickweichenden Profils.19 19. Bassai, AkroterB: Randstuckmit abgerundetenZacken und schmalerLeiste sowie Fragmente der Innenscheibemit Profilansatzdes Randes, abgesetzterLeiste, dreifachemRundstab,einem Kranz reliefierter Stabe alternierend mit aufgemalten spitz endenden Blattern, dreifachem Rundstab und Wulstring mit Offnung.20

20. Sparta (ca. 120 m nordlich des Artemis-Orthia-Sanktuarsgefunden):Fragment einer Innenscheibe (Antefix ?) mit gemaltem Zungenmuster, Hakenmaanderband,einem abgesetztem Band mit Zickzackmuster,und Wulstprofileines zentralen Ringes(?).21 21. Olympia, Akroter eines Schatzhauses (Durchmesser 1.12 m, etwa 3. Viertel 6. Jh. v. u. Z.; Fig. 1):22 stark vorgew6lbter schmaler Rand mit Blattkranz (Zwickelspitzen), aufgemalter nach innen gerichteterBogenreihemit Knospenund Lotosbluiten,dazwischenScheiben,Innenscheibe mit Blattkehle, dreifachem Rundstab, alternierend gegenstandigen LotospalmettenGeschlinge, dreifachemRundstab und konkaverZentralscheibemit ausserem Blattkranzund Offnung. Anzuschliessen sind weitere nicht genauer bestimmbare Fragmente aus Sparta (ArtemisOrthia Nr. 1, 3, 4, 8, und 19, sowie Neufunde aus dem Stadtgebiet),23Delphi (AC.1, 2),24 Amyklai,25Tegea (Artemis-Knakeatis-Sanktuar),26 Lepreon,27Kaliani,28Olympia (Staatliche Museen zu Berlin, AntikensammlungInv. 1548x; P1.39:b),29und von Korfu.30 Die hier aufgefuihrten Akrotere und Fragmente weisen eine grundsatzlich gleichartige Struktur der Scheibe auf, die vergleichbar ist mit dem Profil archaischer Teller. Auf eine unterschiedlich breiten vorgew6lbten Rand folgt eine flache Innenscheibe mit zentraler S. 120, Abb. 90,137, Nr. 16; Goldberg,Nr. D 26. 19AO, S. 118, Abb. 89; Goldberg,Nr. D 30. 20 Rhomaios (Anm. 7 oben), S. 7-12, Abb. 3, Taf. 3; Goldberg,Nr. D 7. 21 AO, S. 124 f., Abb. 96. 22 Mallwitz. 23 AO, S. 118 if., 135-137, Nr. 1, 3, 4, 8,19, Taf. 22, 23, 24 sowie Abb. 90; Fundnotizenin den AR 198586, S. 50 und AR 1976-77, S. 35 f., Abb. 23, 24 ("AltesMenelaion"). 24 Le Roy, 1967, S. 89, Nr. AC.1 und 2, Taf. 30; Goldberg,Nr. D 9,10. 25 W. von Massow, "Vom Amyklaion (Einzelfunde),"AM 52, 1927 (S. 34-64), S. 44, Abb. 43, Beilage 10.4; GoldbergNr. D 5. 26 K. A. Rhomaios,<>, lrEPLOX71v Lauter-Bufe, S. 214, Nr. 1. 28 E. Meyer, Neuepeloponnesische Wanderungen,Bern 1957, S. 14 ff., Abb. 1, Taf. 1,2; Mallwitz, S. 133, Anm. 21. 29 Unpubliziert: doppelter Rundstab (einer erhalten), abgesetzteseinfaches Flechtband,vertieftes Innenfeld, nach dem errechnetengrossten erhaltenen Durchmesser (ca. 59 cm) und der groberenTechnik sicher kein Heraion-Antefix. Das Fragmentin Olympia II, Text S. 192, Abb. 5 (hier P1.39:c) vielleicht zugeh6rig. Wie mir K. Herrmann und J. Heiden freundlicherweisemitteilten,wurden in Olympia Fragmentevon mindestenseinem weiteren Scheibenakroter,das dem Kaliani-Fragmentahnelt, gefunden. 30KorkyraI, S. 141, Abb. 113; Goldberg,Nr. D 45. C. K. Williams, II, berichtetmir von einem weiteren Fund eines lakonischenAkrotersin Pheneos. 18AO,

SCHEIBENFORMIGE AKROTERE IN GRIECHENLAND UND ITALIEN

255

Offnung, die wahrscheinlichzur Befestigungeiner Metallapplikation(Rosette wie bei den Geraki-Akroteren?)31 diente. Beide Zonen werden von plastischenProfilen eingefasstoder gegliedert und besitzen zusatzlich gemalte Ornamentstreifen.32Die Randzone fasst ein Zackenkranzmit Blattkehle und Rundstab ein. Darauf folgt ein mehrreihigesSchuppenmuster, das durch ein aufgemaltes Ornamentband(Zahnreihe, Treppenmuster, Flechtbander, einfacher Streifen) vom unverziertbleibendeneinbiegendenProfil getrennt wird. Die eigentliche Scheibe innerhalb einer uiberleitendenBlattkehle wird aussen und innen von dreifachenRundstabengesaumt. Das verbleibendeRingfeld zeigt bei den einzelnen fur diesen Bereich aussagekraftigenFragmenten unterschiedlicheOrnamentfolgen,die wohl im Sinn einer fortschreitendenEntwicklungvon kleinteilig-addiditivenzu grossflachigeren ubergreifendenMustern zu deuten sind. Die in diesem Sinne alterttimlichsteStrukturzeigt eine nahezu vollstandigeScheibemit geraderBasis aus Amyklai (B.4; Fig. 1, P1.39:a). Von aussen nach innen erscheinthier eine reliefartigabgestufteFolge mit einem Zahnbandund einem einfachemFlechtbandzwischen Blattkranzprofilen.Das AkroterA aus Bassai (B.6) hingegen zeigt auf einer kaum mehr reliefiertenRingscheibezwischen Blattkranzeneinen laufenden Hund und ein Zickzack-Treppenmuster.Ein Treppenmuster mit Zahnband und Zickzackornamentbesass auch ein Akrotervon der spartanischenAkropolis(B.5). Auf dem Ringfeld des Akrotersdes olympischenHeraions (B.10; Fig. 1, P1.40:a, b) dominieren aber schon grosse Reliefzungen, die ein abgesetzter Volutenhakenstreifenvom inneren Treppenmusterbandtrennt. Das gleiche Ornamentsystembegegnet auf Fragmenten aus dem Artemis-Orthia-Sanktuar (B.11), die teilweise unter der Sandschicht gefunden wurden. Eine spatere Entwicklungsstufeduirftedann das Akroter B aus Bassai (B.19) verk6rpern,bei dem die Ornamentrapportereduziert, die Blattzungen und der Zackenkranz verkuimmerterscheinen. Anfuigenkann man vielleicht ein ausserhalb des OrthiaBezirkesgefundenesFragment (B.20), das die Ausgraberaber als Antefix gedeutet haben. Die Ornamentfolge-gemalter Zungenkranz,Maanderband,Zickzackband,und profilierte Zentralscheibe-erinnert jedoch deutlich an die Musterung der aufgeftihrtenAkrotere. Den Ubergang zu der von Mallwitz33 bereits charakterisiertenjtungerenspatarchaischen Akrotergenerationbildet schliesslicheine Scheibe,die zwar einem grossgriechischen Schatzhauszugeschriebenwurde aber dennoch strukturelleher mit den lakonischenGiebelaufsatzenzu verbindenist (s. B.21). Es besitzt zwar die traditionelleplastische Struktur der Scheibender Heraion-Gruppeaus der erstenJahrhunderthalfte,die aufgemaltenfloralen Ornamentrapporte-Bogenfries und Anthemion-zeugen jedoch von einem neuen Dekorationsprinzip,das so erst im 3. Viertel des 6. Jh. v. u. Z. denkbarist. In die zweite Jahrhunderthalftegeh6rt dann eine Gruppe kleinerer Ton- und erster Marmorakrotere,die eine starkeVereinfachungdes Heraiontypusbezeugen:

Lauter-Bufe, S. 216 if., Nr. 14-16, Taf. 85. Zur Ornamentik im Vergleich mit lakonischer Vasenmalerei siehe Searls und Dinsmoor (Anm. 12 oben), S. 70 if. 33 Mallwitz, S. 133, Anm. 21. 31 32

256

VOLKER KASTNER

LISTE C (SPATARCHAISCHEUND JUNGERE AKROTERE AUS TON UND MARMOR)

1. Mantinea (Fig. 1): mehrere Fragmente mit eingesetztem Zackenkranz, Blattkehle, Hakenmaanderzwischen doppelten Stabchen und zuruickbiegendemProfil, Innenscheibemit Blattkehle-sonst undekoriert(?).34 2. Petrobuni (Durchmesser ohne Zacken ca. 30.52 cm): mehrere Fragmente eines konvexen Ringes mit ausseremZackenkranzund Rundstabsowie inneremreliefiertemBlattstab.3s 3. Lusoi (Durchmesserca. 37 cm): konvexes Ringfragmentmit schmalem Zackenkranz,Reliefblattstabaussen und Reliefblattstabund doppeltenStabcheninnen (riickseitigerInnenrandgebrochen).36

4. Lusoi: weiteres Randfragmentmit Zackenkranz,Rundstabchenund reliefierterLotosbliite(?) auf dem konvexemRingband.37 5. Asea: Randfragmentemit Zackenkranzund Rundstabchen.38 6. Tegea (Artemis-Knakeatis-Sanktuar;Fig. 1): Randfragmentmit Kyma-recta-Profiluiberdoppeltem Rundstab (Marmor).9 7. Geraki (in der Stadt eingemauert;P1.41 :a): Marmorakrotermit unteremgiebelformigemEinschnitt-Randprofil mit Blattkehle und Rundstab, konvexer Aussenring, Innenscheibe mit einfassendemProfil (Blattkehleund Rundstab ?) und zentralerRosette.40 8. Geraki (Museum Sparta,erhaltenerDurchmesser38 cm; P1.41 :b, c): walzenf6rmigerAkroterblock aus Marmor, Aufbau analog (C.7)-Aussenprofil jedoch ganz abgeschlagen.4 9. Geraki (eingemauert in der Metropolis): Marmorakroter(gerade Basis ?) mit stark bestossenem Rand, Aussenring und Innenscheibeals flaches Kyma-recta-Profil,getrennt durch ein doppeltesStabchenprofilsowie in der Mitte eine flachreliefierteRosette.42 10. Kurno (Durchmesser 78 cm): oben gebrochenesMarmorakrotermit konkav eingeschnittener Basis und abgebrochenenseitlichenAnsatzen,ScheibedurchblattkehlenartigesProfil vom konvexen Ring abgesetzt,Innenfeldmit sternartigerReliefrosette.43 11. Kurno: stark beschadigtesMarmorakrotermit gerader Basis, schmaler abgesetzterRand mit vertiefterInnenscheibe,darauf in flachemRelief ein radartigesRosettenornament.4 Wahrend sich die Scheibe aus Mantinea (C.1; Fig. 1) noch eng an die altere Akrotergruppe anschliessen lasst, bestehen die Fragmente aus Petrobuni (C.2) und Lusoi (C.3) nur noch aus einem von Blattprofilen eingefasstem Ring, wobei noch unklar bleibt, ob die Koch, S. 88 f., Abb. 42; GER, S. 180, Nr. 7, Abb. 15,109; Goldberg,Nr. D 14. F. Hiller von Gaertringenund H. Lattermann,ArkadischeForschungen,Berlin 1911, S. 35 f., Abb. 9 und 11; Goldberg,Nr. D 20. 36 GFR, S. 180, Nr. 6, Abb. 17; W. Reichel and A. Wilhelm, "Das Heiligthum der Artemis zu Lusoi," OJh 4,1901 (S. 1-89), S. 61-62, Abb. 128; Goldberg,Nr. D 13. 37 GER, S. 182, Nr. 17, Abb. 17. 38 K. A. Rhomaios, Kara rv 'ApKatKx)v 'Ao-E'avo, lwretpas KaltH0flo82w0vos 'ApX'E4 <
SCHEIBENFORMIGEAKROTEREIN GRIECHENLANDUND ITALIEN

257

mittlereOffnung mit einem Einsatz erganztwerden muss.45 Im spateren6. Jh. v. u. Z. treten schliesslichauch steinerneScheibenakrotere(Geraki;C.7, 8, und 9) auf, die sich mit einigen Nachzuiglernbis in hellenistischeZeit verfolgenlassen (Kurno;C.10 und 11).46 Sie sind insofern fur die Beurteilung der fragmentarischeruberliefertenhocharchaischenTypen von Bedeutung, da sie die bei diesen nicht mit Sicherheit rekonstruierbarengiebelf6rmigenEinschnitteneben einer tangentialabgeschnittenenBasis uberliefern.Ausserdem zeigen sie bei einer grundsatzlichahnlichen Einteilung in konvexeRandzoneund vertieftes flaches Mittelfeld zudem einen rosettenf6rmigenMittelknauf (Geraki-hier C.7 und C.8; P1.41). Dieses zentrale Motiv kann als steinerneUmsetzungjenes Metallzieratesgedeutet werden, der vermutlich in die Zentral6ffnungder Tonscheiben eingesetzt wurde und der heute ausnahmslosverloren ist. Die in Marmor ausgefuihrtenlakonischenAkroterelassen dartiberhinausnoch eine weitere wichtige Entwicklungstendenzerkennen: LISTE D (LAKONISCHEAKROTERE MIT FIGURLICHEN DARSTELLUNGEN)

mit konvexem 1. Tegea (Artemis-Knakeatis-Sanktuar): Randfragmenteines Marmorakroters RandprofilundKopfeinerlaufendenGorgone.47 32 cm)mit konvexemRand,eingefasstvon 2. Sparta(Museum):Marmorscheibe (Durchmesser vertieftemMittelfeld,darauf einerausserenLeisteundmiteinemdurchein Profilabgesetztem die ReliefbusteeinergehorntenGorgone.8 3. Kyrene (sekundar als Brunnenmiindungverbautes Volutenakroter, Inv. 14.017): Marmorakroter, bestehend aus einer Gorgoneionscheibemit schmaler Randleiste zwischen Volutenbandern.49

Offenbar wurde die Akroterscheibe-zunachst ausserhalb Lakoniens-und nach einem Fragmentaus Tegea zu urteilen bereits im frtihen 6. Jh. v. u. Z. auch mit figtirlichen Darstellungengeschmtickt.Aus Sparta selbst ist bishernur ein Beispiel eines solchenAkroters mit einem gehorntenGorgoneionaus dem 3. Jahrhundertviertelbekannt.50Deutlich ist aber auch hier die Struktur des traditionellen lakonischen Typs mit breitem konvexem Rand erkennbar.Am Gorgoneion des Apollontempelsvon Kyrene ist diese Randzone zu einem schmalen die Maske einfassendenRandprofilgeschrumpft.Dieses um 500 v. u. Z. datierbareAkroterkann zudem als Vorform der kombiniertenlakonischenGiebelaufsatze angesehen werden, deren Entwicklung seit der spaten Klassik H. Lauter-Bufe bereits naiheruntersuchthat. 45Vgl. das Fragmentaus dem Menelaion, hier B.18. 46Auch in Gestalt der kombiniertenlakonischenAkrotere-dazu Lauter-Bufe. 47 Rhomaios (Anm. 26 oben), S. 15, Abb. 11, 18-19, Abb. 16; Lauter-Bufe, S. 214, Nr. 5, 218; MertensHorn, 1978, S. 44, Anm. 90 (Datierung); Goldberg,Nr. D 55. 48 Lauter-Bufe, S. 215, Nr. 11, Abb. 2. 49 E. Paribeni, Catalogodelle sculturedi Cirene, Rom 1959, S. 19 f., Nr. 22, Taf. 30. Vgl. auch ein Antefix aus Kalaureia (S. Wide und L. Kjellberg, 'Ausgrabungen auf Kalaureia,"AM 20, 1895 [S. 267-326], S. 284-285, Abb. 15) aus dem 3. Jh. v. u. Z. 50Vgl. auch die Verbindung von Gorgoneion und traditionellerlakonischerAntefixformbei den Stirnziegeln des Bouleuterions in Olympia: A. Mallwitz, Olympia und seine Bauten, Muinchen 1972, S. 238, Abb. 191 (2. Halfte 6. Jh. v. u. Z.); Heiden, P1.3:c.

258

VOLKER KASTNER

Ausserhalb Lakoniens wurde das mit dem fruihzeitlichenMaskenbrauchtumverbundene apotropaischeGorgonenhauptin Nordwestgriechenlandjedoch schon sehr fruihmit der Akroterscheibekombiniert.51 Davon zeugen Akroterfragmentein Thermos52und Kalydon.53Wahrend die Fragmenteder noch in das dritte und letzte Viertel des 7. Jh. v. u. Z. datierten Reliefgorgoneia54aus Thermos bisher aufgrund ihrer mangelhaftenPublikation noch kein naheresUrteil erlauben,zeigen die Fragmenteaus Kalydoneine den lakonischen AkroterenvergleichbaretellerartigeStruktur.55Der etwas schragund kaum merklichvorgew6lbte Rand ist hier allerdings nicht besondersakzentuiert, sondernwird von dem das gesamte Scheibenrunderfassendengemalten Gorgoneionbedeckt.Sehr wahrscheinlichist das Akroterhier auch-wie vielleicht ebenso bei den Thermos-Firstziegeln-als vollrunde in das Giebelfeld herabhangendeScheibezu ergainzen.Danach ware diese um 575 v. u. Z. datierbareScheibeein Pendantzu den uiberdie griechischenKolonienStiditaliensund Siziliens bis nach Etrurien verbreitetenSchildakroteren.Der Unterschiedzwischen den aufgesetzten lakonischenScheibenund dem vollrundenam GiebelfirsthangendenSchild scheint mir jedoch zunachst nur graduellerNatur zu sein, auch weil die konstruktiveVerbindung mit dem gebogenenFirstziegelbei beidenAkrotertypenin technischgleicherWeise erfolgte. Bevor wir uns jedoch diesem Problemkreiszuwenden, m6chte ich noch einmal auf die Ubernahme lakonischer Akrotere des "Heraion-Typus"im ostgriechischenGebiet hinweisen. Innerhalb der verschiedenenStudien zur lakonischen Kunst ist bereits mehrfach auf die engen Verbindungenzur ostgriechischenKunst hingewiesenworden. Interessantist hierbei nun, dass auf dem Gebiet der BaukeramikLakonien auch als gebenderTeil dieser Wechselbeziehungerscheint.Eine direkteUbernahme des lakonischenAkrotertypskonnte durch Funde vom Apollontempel auf Aigina und aus dem thasischen Herakleion nachgewiesen werden. Auf Thasos wurden die lakonischenScheibenfragmentemit einer gegen 540-525 v. u. Z. datierten Erneuerung des "Polygonalbaues"verbunden. Zum gleichen Dach wuirdendann ebenso die strukturellund im Ornament von lakonischenVorbildern abhangigenGorgoneionantefixegeh6ren (Winter, P1. 2:d).56Die Fragmenteaus Phokaia, Larisa am Hermos, und dem samischen Heraion sind dagegen offenbar lokale Nachahmungen des "Heraiontypus".Das Randfragmentaus Larisa am Hermos57zeigt eine malerische Umsetzung des Zacken-Blattkehlenprofiles,wobei der zwischen Blattkranz und SchuppenfeldeingefuigteZinnenmaanderin ahnlicherWeise bereits auf einem Bruchsttick auch Mertens-Horn, 1978, S. 65 (VerbindungKultmaske-Kopfantefix). Koch, S. 56, Nr. 2 (Inv. 21) und 71, Nr. 1-zu letzteremauch Mertens-Horn, 1978, S. 42, Abb. 10 und J. Floren, Studien zur Typologie des Gorgoneion,Mtinster 1977, S. 134 ff. (b), Taf. 12:2 (Datierung nach 480-470 v. u. Z.?). 53 Fig. 2; Dyggve, Abb. 158-160 und S. 146-149, 235-236; Mertens-Horn, 1978, S. 54, Anm. 134; Floren (Anm. 52 oben), S. 22 f., Taf. 2.4 (um 575 v. u. Z.); Goldberg,Nr. D 54. 54 Mertens-Horn, 1978, S. 40 ff. 5 D. h., eine Unterteilung in flache Innenscheibe und vorgewolbten Rand bei einem rekonstruierten Durchmesservon 84.5 cm. 56 Gorgoneiamit halbrunderEinfassungaus Rundstab, Blattkehle,und Zackenkranz:Etudes thasiennesI, Paris 1944, S. 44 ff., Nr. 12-17, Taf. 10, Abb. 16 und 17; Winter, oben, S. 31, Anm. 28. 57ATK, S. 48 (Gruppe 2), 63 (festlandischerEinfluss, um 550-530), Taf. 20:1, 2, 21:3; Goldberg, Nr. D 46,47. 51 Dazu 52

SCHEIBENFORMIGEAKROTEREIN GRIECHENLANDUND ITALIEN

259

aus dem Menelaion (B.18) vorkommt.Am Akroter aus Samos58erinnern die reliefierten Tropfenblatteran die entsprechendenlakonischen Blattzungenkranze.Eine eigenstandigere Abwandlungdes lakonischenSchemasverk6rpernvielleicht Fragmenteaus Neandria mit einem gemalten lesbischenKymation(?)unter dem Zackenkranz59und aus dem pontischen Nymphaion mit einem Lotosbltitenornament(?)umgeben von einem reliefierten Schuppenfeld,60die wohl erst gegen Ende des 6. Jh. v. u. Z. zu datieren sind. Die alteren ostgriechischenScheibenakroteregeben ausserdemAnhaltspunktefur eine zeitliche Eingrenzungder Lebensdauerder Giebelbekr6nungendes "Heraiontypus".Ausgehendvonder DatierungderthasischenFragmenteund dervonA.Akerstr6mvorgeschlagenenZeitstellung des larisaischenDaches mit Scheibenakroter(550-530 v. u. Z.) darf man annehmen,dass die Produktionder lakonischenScheibendes "Heraiontypus"im drittenViertel des 6. Jh. v. u. Z. auslief. Wahrendso das Ende der Entwicklungdes "Heraiontypus"einigermassenfixierbarist, ergebensich umgekehrtfur die Ermittlungder Vorstufenund der Anfange des lakonischen Giebelaufsatzes allgemein aufgrund des Mangels an interpretierbarenFunden weitaus grossere Probleme. Es sind besondersvier Fragmenteaus dem Artemis-Orthia-Sanktuar, die als Vorlaufer der schon sehr komplex strukturiertenScheibender Heraion-Gruppe in Frage kommen: Fig. 1) 1. Artemis-Orthia Nr. 15 (gefundenmit geometrischer undLakonisch-I-Keramik, 650-620 v. u.

LISTE A.1 (FRUHE LAKONISCHE SCHEIBENAKROTERE;

Z., im "Early Temple"): schwarzgefirnistesstark gebogenes Randstuckmit konkavemStirnprofil.61

Nr. 17:Randfragment einerkonkavenScheibemitfalkenschnabelartigem 2. Artemis-Orthia Abschlussprofilund ruckseitigemWulst (Ziegelansatz?),schwarzgefirnist.62 3. Artemis-OrthiaNr. 14: Randfragmentmit abgesetzterSchachbrettleisteund konkaverScheibe,

bemaltmit Blattzungen(Maltechnikerinnertan Lakonisch-II-Keramik, 620-580 v. u.Z.).61 4. Artemis-OrthiaNr. 9 (gefundenmit Lakonisch-III-Keramik): mit drei glattes Randstuick undvorgew6lbtem Schuppenreihen Profil." Das Randprofil A.1.1 geh6rte wahrscheinlich zu einer braunschwarz gefirnisten Scheibe und stellt das alteste architektonischeTerrakottafragmentdes Artemis-OrthiaBezirkesdar. Nach der mit ihm zusammengefundenenKeramikkann es spatestensum 620 v. u. Z. datiertwerden. Ein ahnlichesProfil in elegantererAusfuihrungverk6rpertauch das zweitgenannteBruchstuick(A.1.2), das juingersein duirfteaber leider nicht exakt datierbar 58 ATK, S. 99, Abb. 31:2; Goldberg,Nr.

D 51, 52. Dazu weitere Funde von Randprofilen,die A. Ohnesorg demnachstpublizierenwird. 59 ATK, S. 10; Goldberg,Nr. D 48. 60 M. M. Chudjak,Iz istorii Nimfeja, Leningrad 1962, S. 18 f., Taf. 7:1 (oben);ATK, S. 34 erwahnt noch zwei Fragmentevon ScheibenlakonischenTyps aus Phokaia (Goldberg,Nr. D 49, 50). 61 AO, S. 120, Abb. 90,137, Nr. 15; Goldberg,Nr. D 21. 62 AO, S. 120, Abb. 90,137, Nr. 17; Goldberg,Nr. D 22. 63 AO,S. 118, Abb. 88,136-137, Nr. 14; Goldberg,Nr. D 23. 64 AO, S. 118,136, Nr. 9, Taf. 24; Goldberg,Nr. D 35. Ein Neufund aus Olympia zeigt eine sehr ahnliche Gestaltung (Hinweis von K. Herrmann).

260

VOLKER KASTNER

ist. Gegentiberdiesen frtihen schwarzgefirnistenScheibenstellen die AkrotereA.1.3 und 4 ein fortgeschritteneresStadium der Entwicklungdar. Sie besitzen ein komplizierteresAbschlussprofilund zeigen einen ornamentalenDekor in der bekanntenschwarzbuntenlakonischen Technik. Das AkroterA.1.4 besitzt bereitsdas regularelakonischeAbschlussprofil, weist aber noch keinen Zackenkranzund keine Blattkehle auf. Es ist bis zum glatt abschliessendenRand mit einem Schuppenmusterbemalt. Die Ausgrabererwogen hier eine Verbindungmit dem FragmentNr. 10 (hier B.1). In diesemFall lIge hier bereitsder Prototyp der Heraiongruppevor. Zu den fruihenScheibengehort sicher auch das Randfragment A.1.3, das ebenfalls noch keinen Zackenkranzaufweist aber mit einer Schachbrettleiste abschliesst.Ausserdem war die mit einem Zungenmusterbemalte Scheibe konkav vorgewolbt. An diese frtihen lakonischen Exemplare lassen sich nun eine ganze Reihe westgriechischerSchildakroterenach der Ornamentikund dem konkavenProfil anschliessen: A.2 (GROSSGRIECHISCHE,SIZILISCHE,UND ETRUSKISCHEAKROTERE) 1. Lokri,Marasa-Sanktuar undzentralemBuckel (Fig.2): konkaveScheibemitSchuppenmuster

LISTE

(Durchmesser54 cm).65 2. Lokri, Marasa-Sanktuar:Scheibe mit Randstreifen,Zahnmuster,und einfachem Flechtband

(Durchmesser ca. 60 cm).66 3. Gela, Akropolis (Fig. 2): konkave Scheibe mit ausgeschnittenemBlattkranz (Durchmesser ca. 70 cm).67 4. Gela, Akropolis:konkaveScheibemit profiliertemRand und aufgemaltemkonturiertemBlattkranz (Durchmesserca. 78 cm).68 5. Himera, Insula III in Zone VI (Inv. 71.1152, 72.1077): Fragmente einer leicht konkaven Scheibemit gekerbtemBlattkranz,Rundstab,und glatter Innenscheibe.69 6. Cumae, Akropolis (Heidelberg, UniversitatssammlungInv. 25/54 und 25/55): Randfragmente von Scheiben mit innerer Wirbelrosette und breitem, leicht konkaven konusartigen Rand, der mit einem konturiertenBlattstabbemaltist und mit einer schmalenLeiste (Treppenmuster) abschliesst.70 7. Cumae (Nationalmuseum Neapel, Inv. 85553; Fig. 2): Scheibe mit innerer Wirbelrosetteund breitem aufbiegendemRand, der einen konturiertenBlattkranz mit Zwickelspitzen aufweist (Durchmesser38 cm).71 8. Rom, Regia (Phase III, gegen 540 v. u. Z. zerstort):Scheibenfragmentmit gemaltemBlattstab (Durchmesser54 cm).72 A. De Franciscis, Il santuariodi Marasa in Locri Epizefiri. 1. IEtempio arcaico, Napoli 1979, S. 68, Nr. 3, Taf. B, Abb. 48-50. 66 Ebd., S. 68-69, Abb. 51-53. 67 L. Bernabo Brea, "L'Athenaiondi Gela e le sue terrecotte architettoniche,"ASAtene, N.S. 11-13, 1949-1951 (S. 7-102), S. 66, 68-69, Abb. 64-65; weitere ahnliche Fragmenteebd., S. 66, 69-70, Abb. 66 und ein Bruchstuck aus Akragas (Museo Nazionale Agrigento Inv. 9500) in Cron Catania 4, 1965, S. 77, Taf. 28:2b. 68 Ebd., S. 66, 69-70, Abb. 66. 69 Himera II, 1, Palermo 1976, S. 332 f., Nr. 1, Taf. 50:4. 70 H. Koch, Dachterrakottenaus Campanien,Berlin 1912, S. 18 (2. Art), Abb. 22. 71 Ebd., S. 17-18 (1. Art), Taf. 2:1; L. Scatozza, "Le terrecottearchitettonichecumane di eta arcaica," Klearchos13, 1971 (S. 45-111), S. 108, Abb. 63, 64. 72 F. E. Brown, "La Protostoriadella Regia,"RendPontAcc47, 1974-1975 (1976; S. 15-36), S. 28. 65

A.2.13>

Kalydon

A.2.1

A A.2.3

A.2.7

A.2.11

A.2.14

FIG.2. Masstab 10 cm. Kalydon (nach E. Dyggve), Cumae (A.2.7), Capua (A.2.11), Locri (A.2.1 Fotos), Gela (A.2.3), Monte San Mauro (A.2.13), Gela (A.2.14), Monte San Mauro (A.2.12)

262

VOLKER KASTNER

9. Acquarossa,Zone F (Inv. 69-118): kreisrundesAntefix(?) mit drei konzentrischenRillen und konkaverMittelscheibe (Durchmesser 17.5 cm).73 10. Pompeji (Porta Ercolanense, Basilika): Randfragmenteeines Schildakroters(?)-Randleiste bemalt mit konzentrischenHalbkreisenund Ruickseitemit umlaufendemFlechtband.74 11. Capua (aus einem Kindergrab, Louvre D 159/Campana 5206; Fig. 2): Scheibe mit aufgemaltem Gorgoneion, konvex vorbiegenderRand mit konturiertemBlattstab und Zwickelspitzen sowie abgesetzterRandleistemit Zinnenmaander(Durchmesser59 cm).75 12. Monte San Mauro (Huigel1-2; Fig. 2): konkaveScheibemit am Rand vorbiegendemBlattkranz und schmalerzuruckgesetzterSchachbrettleiste(Durchmesserca. 55.5 cm).76 13. Monte San Mauro (Grab unterhalb der Hugel 1-2, Museum Syrakus Inv. 77844; Fig. 2): Firstziegel mit parabelformigenSchild (geradeabgeschnitteneBasis), bestehendaus einer konkaven Scheibemit Mittelbuckel (Blattzungenund Rosette konturiert)und abgesetzterSchachbrettrandleiste,wobei die einzelnen Teile durchdunne Rundstabegetrenntwerden (Basisbreite 68 cm, Hohe 48 cm).77 14. Gela, Akropolis (Fig. 2): Vier Fragmente einer in mehrere Ornamentzoneneingeteilten konkaven Scheibe-Randleiste mit Rosettenkranz,eingefasst von Rundstibchen, konkav vorbiegendes Ringfeld mit Wirbelmuster, weitere konkav vorbiegendeMittelscheibe mit ausserem Kranz aus konturiertentropfenf6rmigenBlattern und innerem Rosettenring um einen zentralen flach vorgew6lbtenBuckel (Durchmesser73 cm).78 15. Rhegion (Museum Reggio di Calabria):E. D. Van Buren79beschreibtmehrereFragmentevon konkavenScheibenmit abgesetzterRandleiste,bemaltmit Zungenmustern,reliefiertemSchuppenmuster,und anderenOrnamenten. Die Schildakrotere A.2.1 und 2 wurden von De Franciscis dem altesten Bau des Marasa-Sanktuars zugewiesen. Zu dessen Dach gehorten neben maandergeschmuickten Schenkelplatten und einer einfachen Flechtband-Simaleiste gebogene Flachziegel und halbkreisf6rmige Antefixe mit Wirbelrosetten.80 Die Struktur des Daches wurde also offensichtlich zu einem wesentlichen Teil von Elementen lakonischer Dacher gepragt. Auch die in Lokri vollrund gestaltete Akroterscheibe lIsst sich mit den vorgenannten Fragmenten aus dem Artemis-Orthia-Bezirk vergleichen (A.1.1, 2, 4). Ebenso entspricht die Dekoration mit einem etwas provinziell anmutendem Schuppenmuster oder mit Zahnreihe und Flechtband 03O. Wikander, "Etruscan Roofing-Tiles from Acquarossa," OpRom 8, 1972 (S. 17-28), S. 21-22, Abb. 9:e. 74 A. Maiuri, "Studie ricerchesulla fortificazionedi Pompei,"MonAnt 33, 1929 (S. 113-276), Sp. 230231, Nr. 1, Abb. 29; idem, "Pompei,"NSc 1951 (S. 225-260), S. 256, Abb. 21. 7 Koch (Anm. 70 oben), S. 7-8, Abb. 13, 73, Taf. 20. 76 p. Orsi, "Di una anonime citta siculo-grecaa Monte S. Mauro presso Caltagirone,"MonAnt 20, 1910 (S. 729-850), Sp. 786-789, Abb. 48. 77 P. Pelagatti, "L'attivita della Soprintendenzaalle Antichit'adella Sicilia Orientale," Kokalos 22-23, 1976-1977 (S. 519-550), S. 532, Taf. 81:1, 2; eadem, in II tempio greco in Sicilia. Architetturae culti (CronCatania16) Palermo 1977, S. 55 if., Abb. 13. 78 BernaboBrea (Anm. 67 oben), S. 66 if., Abb. 62:f.;Pelagatti, Tempio (Anm. 77 oben), S. 60-61, Abb. 14. 79 E. D. Van Buren, Archaic Fictile Revetments in Sicily and Magna Grecia, London 1923, S. 151 f., Nr. 3, 4. 80 De Franciscis (Anm. 65 oben), S. 67 if.

SCHEIBENFORMIGE AKROTERE IN GRIECHENLAND UND ITALIEN

263

dem ublichen lakonischenOrnamentrepertoire.Der mit einer Rosette bemalte Mittelbukkel kehrtwieder an Ziegeln aus Monte San Mauro (A.2.13) und Gela (A.2.14) oder an dem Tempelmodell von Sabucina.81Vergleichbare Scheibendarstellungenlassen sich dariiber hinaus auch im etruskischen Gebiet beobachten (z. B. Basis, Staatliche Museen Berlin, AntikensammlungInv. Sk 1222, um 520 v. u. Z.; P1.42:a).82Vielleicht war auch hier das lakonischeVorbild des zentralen Metalleinsatzeswirksam.83Die fur die lokrischenSchildakroterevorgeschlageneDatierung (nach Mitte 7. Jh. v. u. Z. von De Franciscisoder um 600 fur den Oikos als altesten nachweisbarenKultbau von Gullini)84scheint zudem chronologisch die postulierte Abhangigkeitvon friihen lakonischenAkroterenzu bekraftigen. Weitere einfache Scheiben mit Blattkranzenwurden in Gela (A.2.3-5) und Himera gefunden.85Interessantersind jedoch die komplexerstrukturiertenAkroterevom Monte San Mauro und der geloischenAkropolis(A.2.12-14). Ein Fragment(A.2.12) erinnertmit dem leicht vorbiegendenBlattkranz und der zuruckgesetztenSchachbrettleistean lakonische Randprofile (besonders A.1.3). Der zweite fast komplette Firstziegel vom Monte San Mauro (A.2.13) wurde sekundarfur ein Grab aus dem 3. Viertel des 6. Jh. v. u. Z. verwendet. Mittelbuckel, Blattkranz,und Randleiste mit Schachbrettmusterlassen wiederum wie die gerade abgeschnitteneBasis (vgl. B.4) an lakonischeVorbilder denken. Die kompliziertere Zonengliederungder geloischen Scheibe A.2.14 dtirfte dagegen direkt auf die Giebelaufsatze der Heraiongruppe zuruckzufuhren sein. Das Wirbelrosettenmotiv,das auch fur mehrere kampanischeFirstziegel belegbar ist (A.2.6, 7), kennen wir bisher allerdingsnur als StandardmusterlakonischerAntefixe. Die kampanischen Firstziegel mit konisch vorbiegendemAussenring und schmaler Randleistebilden anderseitseine relativ einheitlicheGruppe, die man nach der relativ einfachen Struktur und der geometrisch bestimmten Ornamentik sicher noch in die erste Halfte des 6. Jh. v. u. Z. datieren kann. Dafur sprache auch der Fund eines ahnlichen Ziegeltyps im Material der Phase III der r6mischenRegia (A.2.8). Der capuanischeFirstziegel mit gemaltem Gorgoneion (A.2.11) wird nach den stilistischen Merkmalen der Maske aber schon in das 3. Viertel des 6. Jh. v. u. Z. geh6ren. Er leitet uber zu den zahlreichen seit spatarchaischerZeit in Kampanien und StiditaliennachweisbarenFirstschilden mit Reliefschmuck(meist Gorgoneion-oder Acheloos-Masken).86Auch in Etrurien werden diese Schilde weitaus haufiger gewesen sein als die bisherige Fundevidenz 81 G. Castellana, 'II tempietto votivo fittile di Sabucina e la sua decorazione figurate,"RdA 7, 1983, S. 5-11, Fig. 1-3. 82 Die Welt der Etrusker,Kat. Berlin 1988, S. 213 f., Nr. B 10.4. 83Vgl. auch die quadratischenEinsatz6ifnungender steinernenSchildakrotereder Tomba a casa con portico di Pian di Mola in Tuscania (Architetturaetruscanel Viterbese.Ricerchesvedesia San Giovenalee Acquarossa 1956-1986, Viterbo 1986, S. 141, Nr. 1, 2). Diese Akrotere(Durchmesser47.5 cm und 44 cm) erinnern auch durch ihre Gliederungin Wulstring und konkaveMittelscheibean lakonischeScheiben. di LocriEpizefirii, Taranto 84 De Franciscis(Anm. 65 oben), S. 67 if.; G. Gullini, La culturaarchitettonica 1980, S. 23. 85 Das Blattkranzmotiverscheint auch auf lakonischenAntefixen (AO, S. 138, Nr. 24-26, Taf. 25; 139, Nr. 27, Taf. 25). 86 Zahlreiche Beispiele aus Capua, Fratte bei Salerno,und Paestum.

264

VOLKER KASTNER

vermutenlIsst. Dafur sprachendie haufigen Darstellungendieses Giebelschmuckesin der Bildkunst.87

In Italien setzte sich offensichtlichseit dem 3. Viertel des 6. Jh. v. u. Z. die mit dem Maskenelement verbundeneAkroterscheibedurch. Hier-wo bereits sehr frtih der Kopf oder die Maske als architektonischeDekorationvielleichtnach nordwestgriechischemVorbild aufgenommenwurde-ergab sich anders als im griechischenMutterland daraus eine weiterwirkendeSynthese.Fur die gleichartigeVerwendungvon Maske und ornamentalem Schild im Giebel italischer Bauten bietet eine etruskischeAschenkisteim Pergamonmusem (StaatlicheMuseen zu Berlin, AntikensammlungInv. Sk 1293; P1.42:b, c) aus dem 4. Jh. v. u. Z. ein bemerkenswertesBeispiel.88Wahrenddie eine Giebelseitedes dachf6rmigenDekkels drei angehefteteKopfe zeigt, schmticktdie andereeine Trias ornamentalerScheiben. Der hier vorgelegte Uberblick uber die Entwicklung des scheibenf6rmigenGiebelschmuckesbedarf sicher im Detail noch weiterer Prazisierungen.Das Grundanliegenwar jedoch die GegenuiberstellungvergleichbarerErscheinungenin verschiedenenRegionender griechischen Kultur, wobei sich Beziehungen zwischen Entwicklungen im griechischen Mutterland und den Randzonen im Osten und Westen herauskristallisierten.Dabei ergaben sich m. E. auch Ruickschltisseauf die bisher nur wenig bekannten Entwicklungsphasen des Scheibenakrotersin Lakonien selbst. Dies gilt besonders fur dessen frtiheste Form, die offenbargeradein den westlichen Kolonienund vermitteltuber Kampanienauch in Etrurienzur HerausbildungahnlicherAkrotertypen(Schildakrotere)geftihrthatte. VOLKERKASTNER STAATLICHEMUSEEN zu BERLIN

Antikensammlung Bodestrasse1-3 DDR-1020 Berlin German DemocraticRepublic

87 Ein annaherndvollstandigesSchildakroter,das vermutlicheinen Bronzeeinsatzals Mittelscheibebesass, wurde im Sanktuarvon Punta a Vipera gefunden (Mitte 4. Jh. v. u. Z.); S. Stopponi, Studi Magi, Perugia 1979, S. 251 if., Taf. 2:2. 88 Welt der Etrusker(Anm. 82 oben), S. 314, Nr. D 5.1.

PLATE 39

A_i'A a. Scheibenakroteraus Amyklai (B.4). (DAI Athen)

b. Fragment eines Scheibenakroters, Berlin Antikensammiung, Inv. 1548x VOLKER KASTNERSCHEIBENFLRMIGE AKROTERE IN GRIECHENLAND UND

ITALIEN~~~~~~~~~~~~~~~~

54~~~~~~~~~~~~~~~~~

~~~~~~c. Fragment eines Scheibenakrotersaus

i

.

*

A

Olympia

(Olympia

IIS.

Abb.

5)

iLL~

~~

~

~

~~~~~~~~~~~~

a. Vorderseite

a, b. Scheibenakroterdes Heraions in Olympia (B.1O)(DAI Athen)

b. Ruickseite

VOLKER KASTNER: SCHEIBENFORMIGE AKROTERE IN GRIECHENLAND UND ITALIEN

PLATE 41

?

*~~~~~~~~ VOKRKSNR

a Marmorakroterin Geraki (C.7). (DAI Athen)

~~~N

.........

CEBNFMIEAOTEINGECNLDUDIAIN

b. Marmorakroteraus Geraki (C.8). Athen) Mamoaroe c_ (DAI auerak C8

.10L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~T

a. EtruskischesRelief mit Gebaudegiebel,Berlin Antikensammlung, Inv. Sk 1222 (urn 520 v. u. Z.)

b. Giebelseite mit Kopfen

c. Giebelseite mit Scheiben b, c. EtruskischeUrne, Berlin Antikensammlung,I (PhotographieStaatliche Museen zu Berlin, A VOLKER KASTNER: SCHEIBENFORMIGE AKROTERE IN GRIECHENLAND UND ITALIEN

AN EAST GREEK MASTER COROPLAST AT LATE ARCHAIC MORGANTINA (PLATES43-46)

N ORIGINALLYDEFINING THE TOPIC for this paper,the materialto which I might claim some small level of expertise seemed, in and of itself, too far removedfrom the majorfoci of this conferenceto be of other than peripheralinterest.I thereforeintended to limit my observationsto instances in which Sicilian architecturalterracottas,especially those of the communities in the mesogeia, are related to revetmentsfound on Mainland Greece. Upon reflection,such an approachseemedunfruitful,for these relationships,while formativefor Sicilian architecturalterracottas,are few and would be generallywell known. My researchon the Archaic architecturalterracottasof Morgantina, however, has led to some exciting discoverieswhich draw extensively for comparandaon East Greek revetments) certainly a focus of this conference.' Indeed, through both their iconographyand their arrangement,the revetmentsof one of the many Late Archaic naiskoiat Morgantina seem to reveal in some detail specifics of the life of an otherwise anonymousmaster coroplast. His training was in East Greece, and he probablyarrivedin Sicily at Zankle in the mid-490's among the Milesian, Samian, and other Ionian refugees whom both Herodotos (VI.I-24) and Thucydides (VI.4.5-6) tell us were invitedby the Zankleansto join in founding a new colony at Kale Akte on the north coast. Morgantina lies in centraleasternSicily at the head of the Plain of Catania and astride the pass through the mountains to the Plain of Gela in the south (Fig. 1). In its Archaic phase the intramuralcity was situated on the conical hill that servedas an akropolisand is still known to the locals as La Cittadella(Fig. 2). The building in question,the "Farmhouse Hill naiskos",lies on the western side of the summitof this akropolis(P1.43:a). Excavations in 1967 and 1968 under the supervisionof ProfessorMalcolm Bell revealeda long, narrow building of a type well known at other sites in the interiorof Sicily. Although the downhill corners and west wall of the building have suffered somewhat from erosion, the original exterior dimensionscan be establishedas ca. 34.70 x 6.90 m. The heavy rubble socle, I Thanks are due to the following individualsand institutions:Shari Taylor Kenfield,Curatorof Research Photographs in the Department of Art and Archaeology at Princeton University, for invaluable research assistance;ProfessorMalcolm Bell of the University of Virginia and Directorof the Morgantina Excavations for helpful discussionsas always; Stephen Falatko for his painstakingand insightful reconstructions;Nancy A. Winter of the American School of Classical Studies at Athens for arrangingthis conference;David Connelly, photographerfor the Departmentof Art and Archaeologyat PrincetonUniversity, for the productionof the slides for the talk and the illustrations in this article; the Rutgers University Council on Research and SponsoredProgramsfor making possible my attendanceat this conference;the AmericanSchool of Classical Studies at Athens for its generoushospitality. In addition:o the abbreviationslisted on pp. 5-6 above,the following is used: Tempiogreco = II tempiogreco in Sicilia: Architetturae culti (CronCatania16, 1977), Palermo 1985

JOHN F. KENFIELD

266

L;~~~~~~y1'ac~~~~~~~~~m ~ ~ ~ )~ anPanrrlie(?al(r i

tanqramndaic h nNta Wilybaeum

~

j4nr4f(l~salltygi) (Caltaniseta)

~

SICILY k

.ya

heh1e)umesasa na ?4organtna

,are hownin C)m

N.~N

tia

*

C7nt4riape

d i4

Akrjuini s

(icata) fca*ttniifet)

;Worgantina

ANClENT SlClLY x~~~~~~~~~~~~~~~~' r"ceg a ikE 'Nanseswhichareon

e 0

FIG.

jo 10

!modern

mrna

(gA

in 20

3

oyXk

1. Map of Sicily (M. I. Finley, Ancient Sicily: To the Arab Conquest,London 1968, p. 2)

FIG. 2. Plan of Archaic Morgantina (Cittadella)

AN EAST GREEK MASTER COROPLAST AT LATE ARCHAIC MORGANTINA

267

ca. 0.75 m. wide, supportedmud-brickwalls reinforcedwith timber and coveredwith hy-

draulicplaster.2 Althoughthis naiskoshad been strippedof its contentsprior to the collapseof its roof in the early 4th century,that it was probablyreligious in functionis suggestedby its splendid roof revetments (P1. 43:b). The eaves of this otherwise simple building are truly ornate (P1.43:c). The eaves tiles are each paintedwith a swastikaand box maeanderon the vertical face of the drip;the boxes encloserosettesin reserveagainstthe blackbackground,the tip of each petal containinga black dot (P1.43:e). The horizontalface on the bottomof the drip is paintedwith a leaf pattern in red, black, and reserve(P1.43:d).3 Centeredabovethe drip is a revetmentwhich I call an eaves-tilepalmette (P1.43:e). Its decorationcombinespainting with relief of the greatestsubtlety.These eaves-tile palmettes are providedwith a tripartite, broadly spreading base behind the vertical face and were attachedwith a slip to the eaves tiles below (P1. 43:f). At some time, probably soon after their installation,some of these palmetteswere damagedslightly and becamedetachedfrom their eaves tiles; they were reattached with lead clamps (P1. 44:a). To the best of my knowledge,the eaves-tile palmetterevetmentis thus far unique to Morgantina (Fig. 3), but since eleven variants of this revetmenthave been found at Archaic and Classical Morgantina (Fig. 4), it is hardly unique to the FarmhouseHill naiskos.4 2

H. L. Allen, "Excavationsat Morgantina (SerraOrlando), 1967-1969. PreliminaryReportX," AJA 74, 1970 (pp. 359-383), pp. 375-378. For naiskoi at other Sicilian sites, see D. Adamesteanu, "'AvaKTopao sacelli?"ArchCl7,1955, pp. 179-186; D. Pancucci,"Precisazionisul sacellodi Monte Bubbonia,"in Tempio greco, pp. 119-124, with completebibliography.See, too, P. Pelagatti, "Sacellie nuovi materialiarchitettonici a Naxos, Monte San Mauro e Camarina,"in Tempiogreco, pp. 43-65; G. Fiorentini, "Sacellisull'Acropolidi Gela e a Monte Adranonenella valle del Belice,"in Tempiogreco, pp. 105-114. For a model of a building of this kind, see P. Orlandini, "Sabucina,"ArchCl 15, 1963 (pp. 86-96), p. 88, pls. 27-28; idem, "L'espansione di Gela nella Sicilia meridionale,"Kokalos 8, 1962, pp. 103-106; E. Sj6qvist,Sicily and the Greeks, Ann Arbor 1973, p. 72, figs. 43-46; M. S. Migliore, Sabucina,Caltanisetta 1981, p. 93, fig. 58; G. Castellana, "II tempiettovotivofittile di Sabucinae la sua decorazionefigurata,"RdA 7, 1983, pp. 5-1 1; P. Griffo and L. von Matt, Gela, Greenwich, Conn. 1968, pi. 47; R. R. Holloway, "Early Greek ArchitecturalDecoration as FunctionalArt,"AJA 92, 1988 (pp. 177-183), pp. 177-179, fig. 1. Eaves tiles that are almost identical and were certainlypainted by the same hand were found by Paolo Orsi on the northernside of the agora at Megara Hyblaia. These eaves tiles were associatedwith two types of gorgoneionantefixes,as is the case with the Morgantinaroof underconsideration;see below, pp. 268-269 and 270. Both the Megarian eaves tiles and antefixes are now in the Museo Nazionale Archeologicoin Syracuse, unpublished and uncatalogued. It might also be pointed out, however, that none of the revetmentsof the Morgantinaroof which I will argue are particularlyEast Greekin inspirationwere found at Megara Hyblaia. 4 The eaves-tile palmette shown in Figure 3 was found with other revetmentsbeneath the floor of the FarmhouseHill naiskos, indicatingthe presenceof an earlier decoratedbuilding on the summit of the Cittadella. Since no walls of another building appeared in the excavationsconductedon the western side of the summit, it is assumed that this earlier building lies on the unexcavatedeastern side. The eaves-tile palmette shown in Figure 4, along with many other ornate revetments,decoratedthe eaves of a large, four-roomed building that borderedthe southernside of the Archaicagora on the Upper Plateau of the eastern slope of the Cittadella (Fig. 2). The finds associatedwith this building suggest, among other functions,its use for symposia, and it may well have servedas the prytaneionof ArchaicMorgantina. A revetmentcalled an antefix by its excavatorand generally similar in the appearanceof its decorated frontal surface to the eaves-tile palmettes from Archaic Morgantina was discoveredat Knossos in a well

JOHN F. KENFIELD

268

,,~

FIG.3. Morgantina,eaves-tile palmettereconstructedfrom fragments found below floor of FarmhouseHill naiskos

4.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4 WA

FIG. 4. Morgantina,eaves-tilepalmettewhich decoratedfour-roomed

building boundingsouthernside of Archaicagora

Four sets of gorgoneionantefixeswere found in the ruins of this naiskos. In the first set, the wavy hair of the Gorgon is held in place by a stephane and surroundedby a coronaof eight snakes (P1. 44:b). Variety within this "Coronaof Snakes"type is achieved through painteddetail such as the patternon the stephane(P1.44:c). Associatedwith the firsttype on the basis of identical fabric is a second set of gorgoneionantefixes, the "FlankingSnakes" type in which the hair is arrangedin a triple row of spit curls surmountedby a flange at the point where the end of the cover tile joined the back of the antefix (P1. 44:d). The vertical depositdated securelyto 500-480 B.C.: see J. N. Coldstream,"Knossos1951-1961: Orientalizingand Archaic Pottery,"BSA 68, 1973 (pp. 33-63), p. 60, no. 117, pl. 25. I would like to thank Dr. Nancy A. Winter for bringingthis comparisonto my attention.

AN EAST GREEK MASTER COROPLAST AT LATE ARCHAIC MORGANTINA

269

face of this flange is paintedwith a patternof alternatinglarge and small leaves in black and reserve (P1. 44:e). It seems certain that these two sets of gorgoneion antefixes alternated along the eaves;when combinedwith the revetmentspreviouslydiscussed,they yield the reconstructedelevationof the eaves seen in Plate 43:c. The other revetment seen in this reconstructionis an element which might, in other circumstances,be called the geison. Although this example is typical of Sikeliote geisa, it is an especially fine example (P1. 44:f), guidelines for the guilloche having been incised with the aid of a compassin the slip before being fired and painted. I might add that the date of the constructionof this building suggestedby the style of the antefixes is considerablylater than the cutoff date proposed by Charlotte Wikander for geisa decoratedwith a single guilloche.5 There are two reasons, I suspect, for this discrepancy.The first is visual and is to be found in the types of revetmentsused above the geison or, in this case, pseudo-geison;the more lofty sima invites elaboratedecorationwithout being visually overwhelmed,while the arrangementof lower eaves tile and antefix invites the opposite. The second reason is that this revetmentseems not to have servedas a true cassetta coveringthe geison beam;it is not reconstructedas such in the axonometricview (P1.43:b). Evidencefor the eaves of another, similar naiskos at Archaic Morgantina suggests that they overhung the walls by about 2.0 m., almost one-third the width of the building, confirmationof Vitruvius' adviceto the builders of Italic temples for the span of the eaves overhang, and an apt formula for any building with mud-brickwalls regardlessof location.6With such a broad overhang,there would be no need for a protectiverevetmenton the geison beam. Moreover, the shadows thrown by eaves would obscureany decorationon the geison beam, and, even if it could be seen, its distance from the eaves would break its visual relationshipto those revetments.It seems more likely, then, that this revetmentwas used on naiskoi with mud-brickwalls and broadlyspreadingeaves in order to protectthe ends of the rafters,in a position in which it would be readily seen and visually relatedto the other revetmentsof the eaves.7 A single corner fragment (P1. 45:a) providesthe informationthat this revetmentfunctioned like a rakinggeison in the tympana,turningthe cornersand rising at an angle of 160. The pitch of the roof has been reconstructedaccordingly(P1.43:b). I

C. Wikander, 1986, pp. 28-30. Even the single example of a "Geloan"lateral sima at Archaic Morgantina was accompaniedby a geison revetmentpainted with lotus buds in the crotchesof a running horizontal zigzag, a motif highly divergentfrom Wikander'sparadigm. 6 The evidence comes from a naiskos in Area V. See the preliminaryreports on the excavationsat Morgantina:Sj6qvist,AJA 66, 1962, pp. 142-143; Stillwell, AJA 67, 1963, pp. 170-171; Sj6qvist,AJA 68, 1964, pp. 146-147; idem, Sicily and the Greeks,Ann Arbor 1973, p. 46; Vitruvius,de arch.IV.7.5. For a clarification of this passage, see A. Andren, Architectural Terracottasfrom Etrusco-Italic Temples, Lund 1940, pp. LXII-LXVI. 7 For similar examples of this pseudo-geison,perhaps executed by the same workshop, see C. Ciurcina, "Nuovi rivestimenti fittili da Naxos e da altri centri della Sicilia Orientale,"in Tempio greco, pp. 66-67, pl. III, fig. 1 and pl. IV, fig. 3. Unpublishedexamples,perhapsby the same workshopas well, have been found in the continuing excavationsin the courtyardof the cloister of the Benedictinemonasteryof San Nicolo in Catania. These excavationsare under the auspices of the Istituto di Archeologiaof the University of Catania and the directionof ProfessorGiovanni Rizza and Dr. Massimo Frasca.

270

JOHN F. KENFIELD

With the exceptionof the two other sets of antefixesfound associatedwith this building, the remainingrevetmentsare not particularlygermaneto our discussion.The ridgepoletiles, however,are as spectacularand as carefullyexecutedas the otherrevetments(P1.45:b). Perhaps in orderto guidethe eye, partsof the backsof the imbriceswere paintedblack,creatinga pleasingvisual rhythmbetweenthe eaves and the ridgepole(P1.43:b). Except for the corner fragment from the rafter end, almost nothing is known of the revetmentsthat definedthe tympana. Fragmentswere found, however, of large gorgoneia, the most typical of Archaic Western Greek pedimental decorations,as were fragmentsof lion akroteria.8 The gorgoneionantefixesof the first phase of the roof (P1.43:c), with their combination of naturalistic features and Archaic monster format, seem typical of the Severe Style and probablydate to the early years of the 5th century;any time in the late 490's suits my argument.9The date of constructionof the naiskos should be the same. The excellent preservationof the painted colors on the fragments indicates that they suffered damage, probably contemporarywith that suffered by the eaves-tile palmettes, within a few years of installation. The damage to the building was repaired, apparently immediately,with two other sets of antefixes (P1.45:c) which are groupedtogetherbecause of their identical fabric. For the sake of argument, let us say that this damage and repair occurredin the 480's. The first set (P1.45:e) consistsof gorgoneionantefixeswhich are the second generation of the Corona of Snakes type that decoratedthe roof in its first phase. While this antefix is obviouslylike its predecessorin most ways, it suffers defectstypical of second-generationterracottas.It is, for example, 16%smaller than the original from which its matrix was formed,resulting in some loss of the finer plastic details of the original, and the painted decoration,while still effective,is cruder.The only majordesign change is that the corona of snakes from which the backgroundhad been partially cut away in the originals of the first phase (P1.43:c) is now incorporatedinto a borderthat surroundsthe entire face. This change was no doubtintendedto inhibit the breakagethat must have plaguedthe visually more effectivebut delicate extremitiesof the first generation.The thicknessof the fabricof the secondgenerationwas increased,too, surely for the same reason.10 The secondset of antefixesfor the secondphase of the roof (P1.45:f) sharesboth composition and thicknessof fabricwith the secondgenerationCoronaof Snakesgorgoneionantefixes, with which they must have alternatedalong the eaves.11Unlike the gorgoneiawith which they shared the eaves, these lion's head antefixes are for the most part unrelatedto 8 For pedimentalgorgoneia,see N.

Bookidis,A Study of the Use and GeographicalDistributionof ArchitecturalSculpturein the ArchaicPeriod (Greece,East Greeceand Magna Graecia), diss. Bryn Mawr College, 1967, pp. 429-437. Fragmentsof a similarlion akroterionwere foundat Megara Hyblaia: see E. Douglas Van Buren, ArchaicFictile Revetmentsin Sicily and Magna Graecia,London 1923, pp. 125-126, no. 3, pl. XII, fig. 51. 9 C. Laviosa, "Le antefisse fittili di Taranto," ArchCl 6, 1954 (pp. 217-250), pp. 227-235; J. Floren, Studienzur Typologiedes Gorgoneion,Miinster Westfalen 1977, pp. 116-127. 10For an explanation of "generations"in coroplasticproduction,see R. V. Nicholls, "Type, Group, and Series:A Reconsiderationof Some CoroplasticFundamentals,"BSA 47, 1952, pp. 217-226. 11Allen (footnote2 above), pp. 375-378, pl. 95, fig. 26.

AN EAST GREEK MASTER COROPLAST AT LATE ARCHAIC MORGANTINA

271

their predecessors,the Flanking Snakes gorgoneia, of the first roof (P1. 43:c). They do, however,hold flankingsnakes in theirjaws and in the paws that dangleon either side of the face. In fact, the entire arrangementof feline forms suggeststhat the image is meant to be a lion skin rather than a lion's head. Variety was achieved in these lion-skin antefixes by alteringthe combinationof colorson such details as the mane and the part of the pelt at the top of the antefix (P1.45:d).12 Lion-headedantefixes are rare in the Greek world, being unknown, to my knowledge, on Mainland Greece, and appearing sporadicallyin East Greece and West Greece and in Etruscan contexts. Besides these examples from Morgantina, other Western Greek lion's head antefixes have been found at Croton, Vibo Valentia (Hipponion), Naxos, Adrano (Aetna-Inessa), and Messina.13With the exception of the example from Messina, other Italiote and Sikeliote lion-headedantefixes are renderedin higher relief than the examples from Morgantina. Moreover, there is, as far as I know, no other example from Western Greece which is suspectedof having alternatedwith gorgoneionantefixes on the eaves of a building. The Etruscan examples share the low relief of the Morgantina antefixes but are otherwise not similar and, as far as I know, are not known to have alternatedwith gorgoneion antefixes. The combinationof lion-headedantefixesin low relief alternatingwith gorgoneiaalong the eaves of a building does, in contrast,occur with some regularity in East Greece. They are found as alternatingappliques in combinationwith lion-headedwater spouts on a sima from Larisa. It might also be pointedout that althoughthe positionof the snakesis different on the gorgoneionappliques from Larisa, the formsof the face are the closestparallels so far to those of the gorgoneionantefixesof the first phase of the Morgantinaroof. From Temnos, too, come fragmentsof a sima which appearsto have been decoratedwith alternatinglion's head and gorgoneionappliques. In this case, it should be pointed out that the lion's heads are dappledwith leopardlikespots, and the snaky locksof the gorgoneiaare enclosedwithin a flange in relief. Equally cogentparallels can be found amongthe revetmentsfrom Miletos 12 A lively discussionwith ProfessorEvelyn Harrison and others arose after the deliveryof this paper as to whether the Morgantina antefix representsa lion skin or a lion thoughtof as an active, living being. Professor Harrison pointedout the similarityof forms, especiallythe apparentlydanglingpaws, with the lion protomes that appear on the Early Orientalizing shields from Crete that are locatedon that island and at the Panhellenic sanctuaries.See, for example, R. Hampe and E. Simon, The Birth of GreekArt, London 1981, p. 113, fig. 168. Two of these shields were in the Museum at Delphi when the Conferencevisited that site to see its architecturalterracottasat the kind invitationof ProfessorChristian Le Roy. 13 For the best surveys of this material, see K. M. Phillips, Jr., "Terrecottearchitettonichecon protomi di leopardoda Poggio Civitate (Murlo, Siena),"BdA 18, 1983, pp. 1-24; P. Muiller,Lowen und Mischwesenin der archaischengriechischenKunst: Eine Untersuchungeniuberihre Bedeutung,diss. University of Zurich, 1978; review by M. Mertens-Horn, ArchCl 31, 1979, pp. 423-429. See also M. Mertens-Horn, Die Lowenkopf-Wasserspeierdes griechischen Westensim 6. und 5. Jahrhundertsv. Chr. (RM-EH 28), Mainz 1988. For the example from Messina, see P. Pelagatti, "Sacellie nuovi materialiarchitettonicia Naxos, Monte San Mauro e Camarina,"in Tempiogreco, p. 49, note 26. For Naxos, ibid., fig. 5. Of course,both lions and Gorgons often appear in the decorationof Archaictemples on Mainland Greecebut never,as far as I know, in the form of alternatingantefixes. They also appear on the obversesand reversesof Archaic coinage;see, for example, Archaic issues from Athens and Eretria:C. M. Kraay, Greek Coins, London 1966, pl. 115, no. 349; A. Baldwin Brett, Catalogueof GreekCoins:Museum of Fine Arts, Boston, Boston 1955, no. 1023a, pl. 54.

272

JOHN F. KENFIELD

(P1.46:a, b) and Didyma, where lion's head and gorgoneionantefixesof the same type were found together, a circumstancesuggesting that they decoratedthe same building, alternating along the eaves.14 Althoughthe image of the true lion's head at Miletos and Didyma is not the same as the lion skin at Morgantina, the use of low relief is similar. Frontal images of a lion's face similar to that on the antefixes from Morgantinado occur,however,on the Archaiccoinage of both Miletos and Samos.At Samos,moreover,as on the Morgantinaantefixes,the head is shown as removed with the skin. Even more significant is the fact that the same image appears on the obverseof coins issued by the Samians, Milesians, and other lonians after their arrival in Zankle. Indeed, the lion's face continues to appear on the coinage of Zankle/Messina (P1. 46:c) even after the arrival of new refugees from Messene and the change of the city's name. These issues of the 480's provideby far the closestparallel for the form of the lion's face as it appears on the Morgantina antefix.15It is for this reason that I date the secondphase of the roof to the 480's B.C. The use of images on the obverseof a coin for comparisonmay raise some objections,but in spite of the greatly differentdimensions, the influenceof coin types on antefix types seems reasonablesince both sculptors,each in his own medium, were faced with similar problems of filling a circular or semicircularfield with a bold image in low relief.16 In any event, the presencein Messina of the master coroplastresponsiblefor the Morgantina roof is certain, since the only other known fragment of a lion-skin antefix in the same series was found in the excavationsof block 224 of that city precedingthe construction of the Hotel Reale.17 It might be difficultto understandwhy, in the secondphase of the roof, a decisionwas made to alter so radicallythe secondset of antefixes (P1.45:c). The answer, I believe, is to be found in the identity of the god to whom the building belonged. Evidenceis providedby a graffito inscribed on a sherd found beneath the floor. Given its context, the sherd must antedatethe constructionof the naiskos, but it can be assumedthat the sanctuarybelonged 14 Lion-headed antefixes have also been found at Neandria but are not associatedwith gorgoneionantefixes. See ATK, p. 9, pl. 3 for Neandria; pp. 37-41a, pls. 13, 14 for Temnos; pp. 60-61, pls. 30 and 32 for Larisa; pp. 103 and 107, fig. 32, pl. 53 for Miletos; pp. 109 and 113, pl. 57 for Didyma. For Larisa see also L. Kjellberg,Larisa am Hermos, II, Die architektonischenTerrakotten,Stockholm1940, pp. 91-95, figs. 27, 28, pls. 44,45; see also G. Andreassi,"Simefittili tarantine,"RM 79,1972, pp. 188-189, pl. 96. Dr. P. Schneider addressesthe Didyma examples in his paperon the temenoslocatedat the midwaypoint on the SacredWay between Miletos and Didyma (pp. 211-222 above). The Didyma antefixes, though of the same series as the Miletos examples, were producedin differentmatricesand show differencesin the treatmentof forms;most notably,the snakesat the top of the head face in a differentdirection. 15 C. M. Kraay, GreekCoins, New York 1966, p. 355, no. 588, pl. 177, for Miletos; p.357, no. 611, pl. 182, for Samos, and no. 613 for the Samians at Zankle; G. Vallet, Rhegion et Zancle, Paris 1958, pp. 338-340, pl. XIX, figs. 1, 2. For the coinage of the Samians, Milesians, and Messenians at Messina under Anaxilas of Rhegion, see ibid., pp. 335-354, pl. XIX, figs. 3-5. 16 Such influence is implied if not actually stated by Phillips ([footnote 13 above] p. 18) with regard to fig. 54. 17 Pelagatti (footnote2 above), p. 49, note 26.

AN EAST GREEK MASTER COROPLAST AT LATE ARCHAIC MORGANTINA

273

to the same deity throughoutits history.18The first letter in the graffitois theta in the form of a rectangleon end with a horizontalbar. This letter is followed by E. In both the Chalcidian and Dorian colonies of the West during the 6th century, these forms would signify the spiritusasper followed by E or H. Taken in conjunctionwith the lion-skin antefixes,the inscription might be restored as 'HpaKXE'L,suggesting that the very pinnacle of Archaic Morgantinawas occupiedby the deity most closely associatedwith the divine right of Hellenic colonization.19 Let us return to the gorgoneionantefixes of the Corona of Snakes type (P1.43:c). The example from Miletos (P1. 46:b) is also closely connectedwith both phases of the Morgantina roof. With the exceptionof a single fragmentin Syracusesaid to be from that city,20 gorgoneionantefixes in this series are unknown in Sicily outside Morgantina.They are, in contrast,well representedin Magna Graecia, in the extensive series from Taras and elsewhere.21The examples from Magna Graecia, however, are differentin that the Gorgon's entire face is often surroundedby snakes, thus greatly increasing their numbers. On the antefixes from Miletos, Didyma, and Morgantina large snakes flank the Gorgon'scheeks, two on each side at Miletos and Didyma, a single snake at Morgantina. In both cases, the top of the Gorgon'shead carries six coiled snakes which are of equal size but smaller than those that flank the face; on the Milesian antefix the crowning snakes face out, away from the center,while at Morgantinathe two centersnakeson each side face each other acrossthe top of the head. To the best of my knowledge, the Corona of Snakes gorgoneion on the Milesian antefix is the closest parallel to the antefixesof that type at Archaic Morgantina. Once again, the roof of this temple of Herakles contains a type that is specifically East Greek and can be associatedmost closely with Miletos.22 Since both phases of the naiskos roof show such specific connectionswith East Greek types of antefixes and, indeed, combinationsof those types, and since the second phase appears to follow the first in quick succession, I believe that the revetmentsof both phases are productsof the same workshop. These precise East Greek referencesindicate that the 18 The sherd and its graffitoshould probablybe associatedwith the hithertounexcavatedbuilding of which some architecturalterracottaswere found beneaththe floor of the naiskos. See footnote4 above. 19 Allen (footnote 2 above), p. 377, nos. 32 and 33; L. H. Jeffrey, The Local Scripts of Archaic Greece, Oxford 1963, p. 79; M. Guarducci,Epigrafia greca I, Rome 1967, pls. 1, 11. For the use of Herakles as a symbolof the divine right of Greek colonial expansion, see Sjoqvist(footnote2 above), pp. 13-14; T. J. Dunbabin, The WesternGreeks,Oxford 1948, pp. 330-331. 20 Inv. no. 51003. The forms of this antefix have more in commonwith South Italian antefixes of this type than with the East Greek or Morgantinaexamples. 21Van Buren (footnote 8 above), pp. 137-144, figs. 59-61; Laviosa (footnote 9 above), pp. 229-243, pl. LXX, figs. 2, 3, pl. LXXI, fig. 1; M. La Cava, Topografiae storiadi Metaponto,Naples 1891, pl. IV:3; A. Campi, Magna Grecia 1924-1925, pp. 6,14-15, figs. 3,4; M. Galli, Magna Grecia1926-1927, pp. 66-71; D. Adamesteanu, "Problemesde la zone archeologiquede Metaponte,"RA 1967 (pp. 3-38), pp. 33-34, fig. 41; Andreassi(footnote 14 above), p. 185, pl. 94. 22 Floren (footnote9 above), pp. 62-73, pl. 5. The snakes at the top of the head on the gorgoneionantefixes from Didyma face each other across the top of the head although they are of the same series as those from Miletos. See P. Schneider,pp. 211-222 above.

274

JOHN F. KENFIELD

masterin chargeof this workshophad directexperienceof the relevantEast Greek roofs.As this group includes those roofs with a sima from Larisa and Temnos, one should not argue too strongly for the master's being Milesian. Since his presence is indicated in Zankle/Messina in the early 5th century,however,I suggestthat he was an Ionian refugeefrom the Persians who, traveling to Miletos in preparationfor the general exodus to Zankle, studied the decoratedroofs of the buildings in that great and soon-to-be destroyedcity, as might be expectedof a man of his profession.Once in Sicily, this master set his genius immediatelyto work, adapting East Greek decorativeschemesto typical Sikelioterevetments, and, in the second phase of the Morgantina roof, when a decision was made to have the revetmentsrefer iconographicallyto Herakles, he chose a form symbolic both of his East Greek heritage and of his new mothercity Zankle/Messina. JOHN F. KENFIELD RUTGERSUNIVERSITY

Departmentof Art History Voorhees Hall Hamilton Street New Brunswick, NJ 08903

PLATE 43

a_E;~~~~~~~~'v

a. Aerial photographof foundations

b. Axonometricreconstructionwith second-phaseantefixes

Morgantina: Farmhouse Hill Naiskos

~ ~

EKE

~

~

~

~

c. Reconstructionwith first-phase antefixes e.Fotve

~

~ ~

.Ve

M

rmabv

d. Bottom of drip of eaves-tile fragment

~ ~ fragment..............from..roof e f. Eaves tile and eaves-tile palmette

_a. J eH

F.

KEFED

NES

RE

ATRCRPATA

. AEACACMRATN

PLATE 44

a. Central part of tripartitebase of eaves-tile palmette with lead clamp

b. First phase: fragment of Corona of Snakes type of gorgoneion antefix

c. First phase: fragmentof Corona of Snakes type of gorgoneionantefix d. First phase: Flanking Snakes type of gorgoneion antefix Morgantina:

e. Fis phs:Fakn nkstpeo gogninatfxsoigpitdtnu patr

.Gio

Farmhouse

Hill Naiskos

rget

nfaneaoehi

JOH FKEFEDANESGRE

MATRCRPATALAEACACMRA

IA

PLATE 45

a. Corner geison fragment

b. Ridge-pole fragments Morgantina: Farmhouse Hill Naiskos

.~~~~~.

c. Second phase: reconstructionof eaves

JOH F.X3 KNEL:

ANES_RE

d. Second phase: fragmentsof lion antefix showing different color combinations

ATRCRPATA

AEACACMRATN

PLATE 46

a. Miletos: lion antefix (ATK, pl. 53, fig. 1)

_

5

_

;

_

b. Miletos: gorgoneion antefix (ATK, pl. 53, fig. 2)

c. Obverse and reverse of coin issued at Messina in 480's during the reign of Anaxilas of Rhegion (G. Vallet, Rhe'gionet Zancle, Paris 1958, pl. XIX, fig. 5)

THE ARTEMISION SIMA AND ITS POSSIBLE ANTECEDENTS (PLATES

47, 48)

ASEARLY AS 1944 Herrmann Siisserott,in his article on early Sicilian roofs, had emphasized the essential differencebetween two terracottamembers,the sima and the geison revetment:the one belongs to the roof proper, the other to the building under the roof.1He also noted that while this functionaldifferencewas fully appreciatedin Sicily, as indeedit was, appreciationof it had disappearedfrom Greeceby the early 6th century.Only thus could the Corfiotesima of the Artemisiontemple and the similar sima from Delphi be explained. Now, after more than forty years, there may be reasons to look at this feature again.2

It is quite true that the Artemision sima (P1. 47) and the Corfiote sima of Delphi (P1. 48) are unique in Greece.3A terracottacover for the geison, a wall-plate, or a comparable beam is a consistentfeature on Sicilian and South Italian roofs from the early 6th century;it is a major feature in the repertoireof architecturalterracottasin Central Italy and Asia Minor, but it does not occur in Greece.4In the terracottarevetmentof the Artemision, not only is such a memberused but it is also fashionedin one piece with the sima so that a technically new member is created, a combinationpiece with one raised and one overhangingpart. In this respect, Siisserott'sanalysis was perfectlyaccurate:this combination piece eliminates the distinction between the roof itself and the beam that carries it. What I would like to question is Stisserott'sassumption that this blurring of the tectonic parts of the building is, properlyspeaking,a featureof Greek roofs, as opposedto the Greek or Greek-influencedroofs locatedon Italian soil. ' H. K. Susserott,"Herkunftund Formgeschichtedes sizilischen Traufsimendaches,"OlForschI, Berlin 1944 (pp. 110-125), p. 115. Works frequentlycited are abbreviatedas follows: De Franciscis, 1979 = A. De Franciscis,II santuariodi Marasaiin LocriEpizefiri, I, II tempioarcaico(Monumenti antichi della Magna GreciaIII), Naples1979 Himera I = Himera, I, Campagne di scavo 1963-1965, A. Adriani, N. Bonacasa, C. A. di Stefano, E. Joly, M. T. Manni Peraino, G. Schmiedt, and A. Tusa Cotroni, edd., Rome 1970 Scichilone = G. Scichilone,"Tre rivestimentifittili selinuntini e alcuni problemidella produzione siceliota arcaica,"ASAtene39-40, 1961-1962, pp. 173-217 C. Wikander, 1988 = C. Wikander, Acquarossa,I, ii, The Painted ArchitecturalTerracottas.Typological and DecorativeAnalysis (Skrifterutgivna av SvenskaInstituteti Rom, 40, 38:1, 2), Stockholm1988 2 I have touchedupon this theme before, in C. Wikander, 1986, pp. 26-27. 3 Corfu: KorkyraI, pp. 100-124, figs. 73-87, 91, and 93; Delphi: Le Roy, 1967, pp. 65-68 ("Toit 27"), pls. 19, 20, and 100. 4 Unless, of course, we refer to a place like Olympia, with treasuryterracottaswhich are not of mainland manufacture.

276

CHARLOTTE WIKANDER

It is perfectly true that the Sicilian roof systems in the 6th century generally retain a sharp functional differencebetween the hanging geison revetmentand the standing sima, whether lateral or raking. They are distinct, separatepieces, each with its differentformal characteristics.This distinctionis also clearly made in the painted decoration,where, as far as we know, the geison revetment plaques are restrictedfrom the very beginning to one decorativefeatureonly, the single or doubleguilloche.5For the simas a much wider scopeof decorationis permitted. While the cavetto always has some kind of tongue pattern or derivativethereof, there rapidly developsa great profusionof variation, and the lower fascia shows many differentpatterns:checkers,lozenges, rosettes,etc.6 It should not be forgotten,however,that to a great extent the edge of such a roof would present a unified aspect to the viewer looking up from below. The technical and functional separationsare, so to speak, internal. Owing to the close proximityof the members,the eye is still led to see a unified edge to the roof, extending upwards in one sweep. This effect is enhancedby the fact that the undersideof Sicilian simas, unlike, for example, that of Etruscan ones, does not normallyprojectbeyondthe edge of the roof.7The same tendencyis well illustrated by that particularlySicilian phenomenon,the insertion of a horizontal sima in the gable.8 In this position, the sima serves no practicalpurpose whatsoever,and it is distinctly not, as it shouldbe, a memberof the roof cover.What seems to be essentialhere is the unity of the sima-geison revetment,which prevents a separation of the two elements and producesa filling for a tympanumfield which would otherwise be at least partly empty. In Etruria, in the late 7th and early 6th centuries,the distinctionbetween the roof cover and what is placed below the edge of the roof is much stricter.Here we seem to have a true separationof the two parts, illustrated,for example, by the fact that at Acquarossathere is a general tendency to use revetmentplaques, both painted and in relief but without simas. This is a fundamentallydifferent attitude from that seen in Greece, where the sima is the dominant feature. When relief revetment plaques are used with simas, the separation is made additionally clear by the strigil course of the revetmentplaque, which separates it 5C. Wikander, 1986, pp. 28-29; concerningthe use of architecturalguilloches in general, see C. Wikander, 1988, pp. 100-107. 6 See C. Wikander, 1986, pp. 13-21. 7 On Etruscansimas it is normal to find a decoratedborderon the undersideof the horizontalplaque. This borderis usually from 7 to 15 cm. wide. 8 This practiceis securelydocumentedat Gela on the Athenaion:L. BernaboBrea, "L'Athenaiondi Gela e le sue terrecottearchitettoniche,"ASAtene27-29, 1949-1951 (pp. 7-102), pp. 22-32, 36-38, 56-59, figs. 14, 15, 18-21, 26, 47-49; C. Wikander, 1986, pp. 32-35, nos. 6, 9, figs. 1, 2; at Selinus, on Temple Y and Temple C: E. Gabrici, "Per la storia dell'architetturadorica in Sicilia,"MonAnt 35, 1933 (cols. 137-262), cols. 200203, 193-198; C. Wikander, 1986, pp. 41-42, nos. 47, 48, figs. 3, 10, and 12, and also in a smaller fragment: see Scichilone,pp. 204-214, figs. 41-57; C. Wikander, 1986, p. 43, no. 50, fig. 10; at Syracuse,fromthe Athenaion: P. Orsi, "Gli scavi intorno all'Athenaion di Siracusa negli anni 1912-1917," MonAnt 25, 1918 (cols. 353-754), cols. 637-642, 660-666, figs. 223-225, 234-236, pls. XVIII, XXII; C. Wikander, 1986, pp. 44, 46, nos. 54, 56, figs. 4, 13; from the Apollonion:Orsi, op. cit., col. 663, fig. 237; C. Wikander, 1986, p. 47, no. 65, fig. 13. They are also present on the two Sicilian treasuriesin Greece, at Delphi: Le Roy, 1967, pp. 70-73, pls. 21-23 and 101; C. Wikander, 1986, p. 50, no. 87, fig. 13; H. Schleif and H. K. Siisserott,"Das Schatzhausvon Gela," OlForschI, Berlin 1944 (pp. 83-110), pp. 96, 100, pls. 40, 47, 48, 51; C. Wikander, 1986, pp. 50-51, fig. 6.

THE ARTEMISION SIMA AND ITS POSSIBLE ANTECEDENTS

277

emphaticallyfromthe sima.9The simas themselvestend to sit with their horizontalplaques projectingbeyond the edge of the roof. In a similar way the decorationgenerally enhances the effect of separation. The question now is how we are to view the phenomenonof the two-part member,the Artemisionsima. Is it a completelylocal variant,tried once and producedagain, in a slightly less ambitious manner, for a Corfiote treasury at Delphi?10To begin with, there is no suggestion here that the Artemision decorationis a part of the repertoireof the so-called Northwest Greek circle, as are many other known terracottaremains from the island. Neither at Thermon nor at Kalydon, the best-known exponents of this style, is there any generaltendencyto coverthe geison beam,11 much less to combinethe two functions.There is also no connection at all with the Corinthian praxis of the early decades of the 6th century:the developing cavetto sima combinedwith palmette antefixes on the long sides representsan entirely differenttradition. But if we are not to see the Artemision sima as a completelyisolated specimen, what about the West? There are two immediatepointers toward that direction:first, of course, the fact that a revetmentis providedfor the geison at all, and secondly, the fact that the lateral sima pieces carry tubular waterspouts,the normal practice in Sicily. In fact, some finds of recent decadessuggest that in the colonial environmentthere exists the same tendencyto combinethe two functions. For a long time, the position of the geison revetmentas such in the developmentof the Sicilian systems has been somewhat obscure.12The two single occurrencesof very simple simas with a cavetto alone which were long generally agreed to be the earliest representatives of terracottaproductionon Sicily, the simas of Grammicheleand Syracuse,have no known geison revetmentsto accompanythem.13This fact is, of course,very likely to result from chance, and these two single examples, only three fragments altogether, are hard pressedto representan entire 7th-centurytraditionin the island. Nevertheless,most scholars who have treated the question at all tend to concludethat in fact there were no early geison revetments.They are consideredto be a feature of the period shortly after 600 B.C., A course of concaveor convex strigils, low and squat in the early 6th centurybut gradually growing in height, is a standardfeature of Etruscan relief-decoratedrevetmentplaques, with very few exceptions. For these, see C. Wikander, 1988, pp. 19-21, figs. 2, 3. One may comparethis practicewith the upper terminating rolls on Sicilian geison revetments,which create a transition rather than a separation, since the roll is then echoed on the sima itself, between the cavetto and the lower fascia, and is even sometimes also present as terminationof the sima at the bottomedge. 10 As suggestedby Le Roy (1967, p. 69); there is no sure indicationof the findspotof the fragmentsfrom Delphi. II With one possible exception, however:the fragments from Thermon, possibly from "Dach A", which were briefly mentionedby Koch and Van Buren:Koch, pp. 69-71; GFR, p. 83, no. 42, fig. 141. These terracottas, made up of two plaques at right angles to each other, may be consideredto be either simas with a completelystraightprofile or revetmentplaques. They are decoratedwith rather primitivedouble guilloches; see C. Wikander, 1988, pp. 103-104. 12 W. Darsow, Sizilische Dachterrakotten,Berlin 1938, pp. 45, 63; C. Wikander, 1986, p. 27, note 100. 13 C. Wikander, 1986, p. 36, no. 18, fig. 8 and p. 47, no. 61, fig. 13, with references. I

278

CHARLOTTE WIKANDER

FIG.

1. Sima from Himera (after Himera I, pl. XIII)

when the typically Sicilian, so-called Geloan, sima was finding its form, and they are thought to belong completelywith this sima.14 The isolation of the Grammicheleand Syracusefragmentsas the sole representativesof an early productionin the island has, however, now been ended by the discoveryof several fragmentsof an early lateral sima at Himera (Fig. 1). This piece decoratedand protecteda small naiskos, "TempioA", the constructionof which is placed by the excavatorslate in the third quarterof the 7th century,around 630.15 This exceedingly unassuming sima, no more than 15 cm. high, is a true combination member:a low edge with a flat roll on top, fromwhich protrudesa tubular spout and, below this, a low, overhangingpart. It is thus again a combinationpiece coveringboth the function of a standing membercollectingrain water and that of a hanging one protectinga wooden beam and keeping dripping water away from walls of a friable material. No painted decoration is preserved,but some traces of red, white, and black paint remain, enough to show that the piece was originally decorated. The precisepositionof this sima in the developmentalsequenceof Sicilian terracottasis a matterof debate.At presentit is unique, but does this mean that it should be seen as lying completelyoutside the developmentof terracottaroofingsystems?16Its unusual nature may be due merely to the fact that the building to which it belongedwas small and unassuming. Although the sima itself is very low, however, it is by no means unsophisticated:the combination of two functions is in itself an ingenious feature, and there are easier ways of 14 One of the earliest known examples of this association is probably the "RivestimentoA" of Selinus: C. Wikander, 1986, pp. 42-43, no. 48, fig. 10, with references.For the early date, see Scichilone,pp. 188-189. 15 Himera I, pp. 84-87, pl. XIII:3 (N. Bonacasa);C. Wikander, 1986, pp. 36-37, no. 19, fig. 7. 16 My attention has since been drawn to two other simas from Himera: E. Epifanio, "Nuovi rivestimenti fittili di Himera," in II tempio greco in Sicilia. Architetturae culti (CronCatania16, 1977 [Palermo 1985; pp. 165-173]), pp. 171-173, pls. LI:2-3, LII:1-2. I would like to thank D.ssa Epifanio for this reference.

THE ARTEMISION SIMA AND ITS POSSIBLE ANTECEDENTS

279

divertingrain water than by fabricatinga tubularspout. The remainsof paint show that the piece also provideda way to adornthe building that carriedit. The very fact that this was a sacred building which was rebuilt in the 6th century and furnishedwith highly elaborate terracottadecorationshould also be a warning against assuming that the revetmentis unusual becauseof the small size of the building ratherthan becauseit is an early feature.17 The dates of the Sicilian sequence are admittedlysomethingof a problem. If this sima was producedaround 630 B.C., we are left with a period of ca. 30 to 40 years with no finds other than the two simas mentioned previously.18Yet, the Himera finds are unusual for Sicily in that they were discoveredin situ and excavatedunder modernconditions;if we, for once in this area, are given an archaeologicaldate based on votive deposits,it deservesto be taken seriously. The votive deposits contain a concentrationof material from the Transitional, Early, and Middle Corinthianperiods,and this, of course,is a long span of time. In addition,the generaldifficultiesoften alludedto in the datingof roofsby findsin and around a building still obtain.19 Even if one were to doubtthe high date, maintainingthat the unusual appearanceof the piece is due to the type of building ratherthan to its age, the chronologicalgap remainsthe same:there is nothingfrom Sicily, except for the two fragmentsof sima, for the whole period from ca. 640 B.C. onwards, the very time when we see intense activity in the spread of the terracottaroof not only over Greek territorybut also in the fringe areas of the Greek world, mainly Central Italy. Since this diffusion also seems heavily directedtowards the West, it seems inconceivablethat Sicily could stand completelyapart from this development.Yet, in the present state of our knowledge,this seems to be preciselythe case: intense activity after 600 B.C., but not before. To meet this case, some possible explanationsmay be offered.The first, of course, is that there are indeed terracottasfrom this period, but we simply have not foundthem yet. One may comparethe situationin Etruria25 years ago:nothingwas known of any material before 600, but two new excavationsites, Acquarossaand Poggio Civitate, changedthat picture completely,and once this had happened,new materialkept appearing. The Himera fragmentsmay well be consideredthe first pointersin such a direction. Another way of attackingthis problem is to accept the fact that we do indeed have no particularlyflourishingterracottaindustryin the island before 600 and try to explain it. In that case, I believe we have one importantfactorto take into consideration:the fact that in the Greek world, the terracottaroof does not seem to be used early on private houses, as it does in Central Italy. It is, for example, a curious fact that the extensive excavations at Megara Hyblaea have not, to my knowledge,yieldedany early terracottas.It is possiblethat at least part of the explanation for this lack lies in the early date of many of the colonial foundationson the island, such as Megara (728 B.C.). M.-F. Billot, rev. of C. Wikander, 1986, RA 1988, 1, pp. 149-150. For the currentlyacceptedchronologyof Sicilian terracottas(which was, however,workedout beforethe discoveryof the finds at Himera), cf. Scichilone,pp. 185-193; see also C. Wikander, 1986, pp. 10-12. 19 Votive deposit:cf. Himera I, pp. 83 and 86; for the particulardifficultiesin datingarchitecturalterracottas, see 0. Wikander, "Opaiakeramis. Skylight-tilesin the AncientWorld,"OpRom 14, 1983 (pp. 81-99), p. 94; 17 18

E. Rystedt, Acquarossa, IV, Early Etruscan Akroteria from Acquarossa and Poggio Civitate (Murlo) (Skrifter utgivna av Svenska Institutet i Rom, 40, 38:4) Stockholm 1983, pp. 149-150; C. Wikander, 1988, p.119.

280

CHARLOTTE WIKANDER

Colonial cities which were acquiringan urbanizedcharacterin the period between the late 8th century and ca. 650 could not use tiled roofs, since at that time these had not yet been invented.When they were, in their early years they were probablyrestrictedto highly prestigious buildings and to a limited geographical area, the Corinthia. It seems an inescapableconclusionthat the establishmentof coloniesin the West promptedurbanization at a very early stage but that this developmentdid not include tile roofs. Once you had laid out your city with private houses, public buildings, and sanctuarieswith some other type of roof cover, your immediate needs were satisfied. You also certainly had buildings whose walls were not strong enough to carry the extremely heavy weight of a tiled roof without extensive rebuilding. Thus we find the appearanceof terracottaroofs primarily in the period that inauguratesambitiousbuilding programsof stone temples, that is, definitelyafter 600. What may be significant, as far as Himera is concerned,is its late foundation date, 648 B.C. Since it was foundedpreciselyin the periodwhen the initial diffusionof the terracotta roof took place, its sanctuarywas roofedin this materialfrom the beginning.We may comparethe situationin Etruria;here the processof urbanizationseemsto be a phenomenon beginning for the most part shortly before the middle of the century in the larger Etruscan metropoleis(althoughit must be admittedthat we know very little of the urbanenvironment of the large centersin this period) but reaching its height in the secondhalf of the century, when terracottaroofs, with tiles and, in at least some cases, decorativeterracottaswould unquestionably have been part of the repertory of urban construction.20I would like to suggest that the apparent tardiness of Sicily in the field of architecturalterracottasis one result of its comparativelyearly urbanization.21 Returning to the sima from Himera, I would still maintain a date not later than ca. 600 B.C., mainly because of its originality. Within the early group of low revetmentsfrom Selinus, Leontini, and Megara, the distinct profiles of the sima and geison revetmentsare alreadypresent, and the pieces separated.22If we want to push the Himera sima more than only slightly down into the 6th century,we are also forcedto producean explanationof why Himera stoodapart from the developmentin the island that otherwise seemsboth rapid and remarkablycoherent. Turning to mainlandItaly, the soundingsmadein the 1970'sin the temenosof the Ionian temple at Marasa in Locri Epizephyriirevealeda sequenceof buildingactivityprecedingthe Ionian temple, namely a small naiskosof oikos-typeand the subsequentrebuildingof this oikos and its extension by means of a peripteros,presumablywith wooden columns.23The R. Drews, "The Coming of the City to Central Italy,"'AmericanJournal of Ancient History 6, 1981, pp. 133-165; M. Torelli in Case e palazzi d'Etruria (Catalogue of the exhibition at Siena, 20/5-20/10, 1985), S. Stopponi,ed., Milan 1985, pp. 21-32; G. A. Mansuelli, "L'organizzazionedel territorioe la citta," in Civilta degli Etruschi (Catalogue of the exhibition at Florence, 16/5-20/10, 1985), M. Cristofani, ed., Milan 1985, pp. 111-116; C. Wikander, 1988, pp. 132-136. 21 Cf. A. Di Vita, "L'urbanisticapiCu antica delle colonie di Magna Grecia e di Sicilia: problemi e riflessioni,"ASAtene 59, n.s. 43, 1981 (1983), pp. 63-78. 22 This group is presentedby Scichilone(pp. 185-190). 23 See De Franciscis, 1979; G. Gullini, La cultura architettonicaa Locri Epizefiri, Taranto 1980; idem, "Origini dell'architetturagreca in occidente,"ASAtene 59, n.s. 43, 1981 (1983; pp. 97-125), pp. 102-103, fig. 2. 20

THE ARTEMISION SIMA AND ITS POSSIBLE ANTECEDENTS

281

architectural remains included several groups of architectural terracottas, which their publisher, De Franciscis, divided into differentphases and in some cases attributedto the successivebuildingphases of the naiskos. The first roof of this building already presents an extremely interestingpicture. It is a distinctly Laconian roof, with painted decorationonly. The preservedparts include fragments of the tiles and cover tiles; one well-preserveddisk akroterion,54 cm. in diameter, with a large, scalelikepattern, and smaller fragmentsof anotherakroterion;a semicircular antefix with the canonical crescent pattern;a raking sima; and the one local deviation:a revetmentcoveringfor a wooden beam.24The profiles of the sima and revetmentplaques are the simplest imaginable:for the sima, merely a straight raised edge, ca. 20 cm. high, without any cavettocurve or molding whatsoever.The revetmentplaques are equally simple, again 20 cm. high but with a length of almost 60 cm., with two parts at right angles to each other, one vertical, presumablyoverhanginga beam, and one horizontal. The revetment plaques are consideredto have run along all sides of the building. We have here then what is, in its fundamentals,a fairly orthodox Laconian roof, but one which has one functionalpart that does not belong:the "cassetta",the revetmentplaque. It seems an inescapableconclusionthat this additionto the roof was promptedby local preference, reflecting a striking flexibility in the use of mainland models. The painted decoration shows a similar mixture of orthodoxyand flexibility. While the decorationof the antefix is perfectlyin accordwith mainland Laconianpractice,the motifs used for the sima and geison revetmentdeservenotice:the sima carriesonly a single guilloche, the cassettaa maeander.Thus the decorationconformscompletelyneither to an early Greek sima (one would expect a Doric tongue) nor to what became the established decorativepraxis for geison revetmentsin the West, the guilloche. A very high chronologywas proposed for this roof by its excavator, shortly after the middle of the 7th century.25This high date has subsequentlybeen modifiedby other scholars, and the roof is now generallyplaced in the last quarterof the century.26 The secondphase of this structureconsistsprimarilyof the additionof an adyton,and it is uncertainwhether any majorchangeswere made to the terracottasof the roof. The excavator suggeststhe possibilitythat the raking sima was renewed.27 What is of primary interest for our discussionis the third phase of the temple, a major rebuilding in which a peristasis was added. This addition entailed a major change in the roof. The terracottamembersof the new roof, as groupedby the excavator,show five slightly differentversionsof simas varying mainly in the decoration;four of them, however, consist of lateral simas and geison revetmentsin one piece (Fig. 2).28 The painted decoration consists of two kinds of Doric tongues on the cavetto (one straight, the other slightly tearshaped) and doubleguillocheson the overhangingpart. As seen in Figure 2, these guilloches 24

De Franciscis, 1979, pp. 66-71, pls. A and B, figs. 38, 44-57. De Franciscis, 1979, p. 71. 26 Gullini (footnote 23 above), p. 23; C. Sabbione, in II Museo di Reggio Calabria, E. Lattanzi, ed., Rome/Reggio Calabria 1987, p. 70. 27 De Franciscis, 1979, p. 76. 28 De Franciscis, 1979, pp. 94-98. 25

282

CHARLOTTE WIKANDER

i

RKo

H~~~~NI \__111

\

@03

FIG.

2. Schematicdrawing of sima from Locri, Marasa, third phase (drawing by 0. Wikander)

lack the central palmette which ties the two strandstogether and which is typical of architecturaldouble guilloches.There are no raking simas;to explain their absence,the hypothesis has been offeredthat the roof may have been hipped. The pieces are ca. 38 cm. high and 55 cm. long. The profile is of great simplicity,a trait further enhanced by the provision of simple, crescent-shapedopenings for draining rain water. The closest parallels for the simple cavettoof the sima are the two Sicilian examples mentionedabove29and an Etruscanpaintedrakingsima fromAcquarossadatedat the latest ca. 575 B.C. I would suggestthat the painteddecorationpresupposesthe low sima edges with Doric tongues known from Thermon, Kalydon, and Delphi.30 But does the whole piece, with its integrationof sima and revetmentplaque, presupposethe sima of the Artemisionof Corfu? The answer, of course, hinges very much on chronology.Professor De Franciscis, who published the terracottas,dated them somewhere between the later years of the 7th century and the early part of the 6th. There is, however, some controversyhere, since Professor Gullini, who has studied the other, non-terracottaarchitecturalremains, wants to place the addition of the peristasis around the middle of the 6th century.31Frankly, if the attributionof these terracottasto that particularrebuilt roof is correct,I would hesitatevery much to see such types in the extremely rapid developmentof South Italy placed as late as the middle of the century. If we keep to the date proposedby the excavatorin the initial publication,and from a formal point of view this seems very reasonable, that date of around 600 B.C. or shortly 29

See footnote 13 above.

136; Kalydon:Dyggve, pl. XVII; Delphi: Le Roy, 1967, pp. 31-32, pl. 5. G. Gullini, "Architetturaa Locri,"in Locri Epizefirii (AttiMGrecia 16), Naples 1977 (pp. 409-440), p. 421. 30 Thermon:KorkyraI, fig. 31

THE ARTEMISION SIMA AND ITS POSSIBLE ANTECEDENTS

283

thereafter would make these combinationsimas slightly older than, or roughly contemporarywith, the Artemisionsima. Here lies the heart of my hypothesis:given the interestin the West in an overhangingmemberand the lack of such interestin Greeceitself, and given the example of the Himera sima, would it not be possible that the contactsbetween Corfu and the West were not one-way only? This is not to deny completelythe originality of the Corfiotecraftsmen;simply their wanting to cover the geison is proof enough of that in the Greek environment. Finally, there are also some instances of the survival in Western architectureof this system of joining the sima and the geison revetment.These include two more fragments from Himera, dated to the middle of the 6th century.32Another, later, example is a sima from Naxos with a very low, overhangingpart.33Finally, an unpublishedsima of the same type is reportedfrom Ischia,34demonstratingyet again the continuinginterestin the West in this particularkind of sima. CHARLOTTE WIKANDER LUNDSUNIVERSITET

KlassiskaInstitutionenoch Antikmuseet Solvegatan2 S-227 62 Lund, Sweden 32 Epifanio

(footnote 16 above), pp. 171-173, pls. LI:2-3, LII:1-2. Influence from Asia Minor in the formationof this type is proposedby Epifanio (p. 172) on the basis of a sima from Sardisof the same construction: ATK, pp. 78, 80-81, fig. 24:1, pl. 51:1. From the same area also comes the frieze from Diuver,ATK, pp. 218-222, figs. 70, 75. These share the same idea, the combinationof functions,but have a very rudimentary sima part, a simple raised edge. For these two examples from Asia Minor I would, at least until further evidenceappears,preferto proposean independentdevelopment,unconnectedto either the Italian or the Corfiote examplesof the same trait. It must be rememberedin discussingIonian influenceson terracottasthat sites such as Sardis and Duver are not Ionian Greek cities but inland sites which in many cases developtheir own types under Greek influence;cf. ATK, pp. 230-239. 33 C. Wikander, 1986, p. 40, no. 43, fig. 11, with references. 34 L. A. Scatozza, "Le terrecottearchitettonichecumane di eta arcaica,"Klearchos49-52, 1971 (pp. 45111), pp. 48, note 12 and 59, note 54.

PLATE 47

a.Crfu,,Atmsin

latrlsiaKokr

,fg75

,,

9a ___ 1_~~~~~~~~IE

__

|

:

zz

_r

L

EX

f,54777=

~~~~~~N

7-w;

EXb-z:1 .

x s E | 11

/~~~~~~~~vt

I fi1 A5 b. ,Co,,r11u. AreiionraE simIKrkr

PLATE 48

1_

1111

~~~~~~~~~~~~~~~~~~~~~~~~~~~1

a. Delphi: sima (Le Roy, 1967, pl. 20)

b. Delphi: sima (after Le Roy, 1967, pl. 100)

ARCHAIC ROOF TILES THE FIRST GENERATIONS

Tn HIS PAPER furtherdiscussesissues which I touchedupon brieflyin a studyin the last volume of the Opuscula Atheniensia: Why did the tiled roof appear at all, why did it spread as fast as it did, and how was this diffusion achieved?1

The basis for my argumentis a comparisonbetween documentedfinds of roof tiles from the period 700-650 and those from 650-600 B.C.From the first half of the 7th century,we know of a small group of tiled roofs from a very restrictedarea (markedwith triangles on the map, Fig. 1). The centralmonumentsare the first temple of Apollo at Corinth, datedto approximately680 B.C., and the slightly later first temple of Poseidon at Isthmia. The almost identical shape of the tiles from these buildings recurs at two additional sites: Perachoraand Delphi. The latter cannotbe associatedwith particularbuildingsand thus cannot be dated by external evidence.Their close similarity, however, to the specimensfrom Corinth and Isthmia gives us reason to believe that they are roughly contemporary.2 In the second half of the 7th century,the situation changes completely.Tiled roofs are now in evidenceat a large numberof sites over most parts of the Hellenized world. The list presentedhere does not aim at completeness.Publicationsare often very vague on dates, and some items may prove wrong; many more could certainly be added. Still, even in this preliminary state, the list should be more than enough to prove my point. Besides the four places mentionedabove, architecturalterracottas-plain tiles or decorativeones-are known from the following sites down to ca. 600 B.C. (markedwith circlesin Figure 1):3 Sukas. Tiles from a Greek sanctuary.Second half of 7th century. P. J. Riis, Sukas I, Copenhagen 1970, pp. 52 and 58-59; G. Ploug, Sukas II, Copenhagen 1973, pp. 93-94. Sardis.Tiles from privatehouses. End of 7th century.A. Ramage,Lydian Houses and Architectural Terracottas(ArchaeologicalExploration of Sardis, Monograph 5), Cambridge, Mass. 1978, pp. 9 and 38-41.

Ephesos?Antefix. End of 7th century. Mertens-Horn, 1978, p. 61. Samos.Tiles from various buildings. End of 7th century.Ohnesorg,pp. 181-192 above. Athens. Antefixes from the Akropolis. Late 7th century. TdA II, pp. 26-27; Williams, TEr?A7, p. 349. 1 0.

Wikander, 1988, pp. 203-216. The following special abbreviationwill also be used: Winter, 1978 = N. A. Winter, "ArchaicArchitecturalTerracottasDecoratedwith Human Heads," RM 85, 1978, pp. 27-58 2 Robinson, AM, with bibliographyfor all four sites in note 1 on p. 55. I accept here the dates given by Robinson, even though they are not completely uncontroversial.[As pointed out by R. C. S. Felsch (footnote 40, pp. 313-314 below), the date of 680 B.C., providedby the pottery found in the working-chiplayers, providesonly a terminuspost quem for the date of the temple and its roof-Editor.] 3 Since this list is not intended as a proper catalogue, instead of full bibliographiesI give only one or two referencesfor each site.

FIG.

1. Distributionof tiled roofs:700-650

B.C.

(triangles)and 650-600

B.C.

(circles)

ARCHAIC ROOF TILES: THE FIRST GENERATIONS

287

Eleusis. Tiles from the Solonian Telesterion. Ca. 610 B.C.? K. Kourouniotes, ?'AvaarKa, 'EA7 erltvo0 Kara TO 1933?, AXAr 14, 1931-1932, HapapTrpa (pp. 1-30), p. 4; G. E. Mylonas, Eleusis and the Eleusinian Mysteries, Princeton, N.J., 1961, pp. 66-67, 70, and 76. Argive Heraion. Tiles and antefixes from the North Stoa or the Old Temple of Hera, or both. Ca. 600 B.C.? J. J. Coulton, The Architectural Development of the Greek Stoa, Oxford 1976, pp. 29 and 215; Pfaff, pp. 149-156 above. Halieis. Tiles from the Temple of Apollo. Ca. 600 B.C.? M. H. Jameson, "Excavations at Porto Cheli. Excavations at Halieis, Final Report," AcEXA27, 1972, XpovLKa(pp. 233-236), p. 234; N. Cooper, pp. 65-93 above. Epidauros Limera. Antefix from the Temple of Apollo. Ca. 600 B.C.? Koch, p. 95; GFR, p. 38. Sparta. Tiles, antefixes, and akroteria from the Temple of Artemis Orthia and the Old Menelaion. Ca. 600 B.C. W. S. George and A. M. Woodward, "The ArchitecturalTerracottas,"in The Sanctuary of Artemis Orthia at Sparta (JHS Suppl. 5), R. M. Dawkins, ed., London 1929, pp. 117-144; H. W. Catling, "Excavations at the Menelaion, Sparta, 1973-1976," AR 1976-1977 (pp. 24-42), p. 36. Bassai. Antefixes and disk akroteriafrom the early temple of Apollo. Ca. 600 B.C.? K. A. Rhomaios, <<'EKTOv apXaLorTepov vaov rijs FtLyaXAas?>, 'ApX'E< 1923, pp. 1-25; N. Yalouris, <'AvaAAA 6, 1973, vaov Tov 'E7rLKovpLov O-KacaKL ELz TOV Ev Baocraa^sQOLyaAelas 'AroAAXWYos>?, 39-55. pp. Olympia. Tiles from two unidentified buildings. Second half of 7th century. Heiden, pp. 41-46 above. Thermon. Roof I. Ca. 640 B.C.Koch, pp. 51-74; Mertens-Horn, 1978, esp. pp. 62-63. Kalydon. Tiles, antefixes, geison revetments, akroteria, and metopes from temple buildings. Ca. 630/610 B.C. Dyggve, pp. 135-164; Mertens-Horn, 1978, pp. 53-54. Corfu. Tiles, antefixes, and simas from the Mon Repos temple. End of 7th century. Korkyra I, pp. 143-158; Dontas, NFGH, pp. 121-133, esp. p. 126. Apollonia. Antefix. End of 7th century? Winter, 1978, p. 30, note 4; Mertens-Horn, 1978, p. 60. Taranto. Antefix. Ca. 620 B.C. Winter, 1978, p. 31, with note 6. Metaponto. Tiles from Buildings Al and C. End of 7th century? D. Adamesteanu, "Santuari Metapontini," in NFGH (pp. 151-166), pp. 162-166. Locri Epizefyrii. Tiles, antefixes, simas, akroteria, and revetment plaques from the Marasa sanctuary. Second half of 7th century. A. De Franciscis, II santuario di Marasa in Locri Epizefiri, I, II tempio arcaico (Monumenti antichi della Magna Grecia 3), Naples 1979, pp. 67-71; G. Gullini, La cultura architettonica a Locri Epizefiri, Taranto 1980, p. 23.

Syracuse.Decorativeterracottasfrom the temples of Athena and Apollo. Ca. 600 B.C.? G. Gullini, "Origini dell'architettura greca in Occidente," ASAtene 59, n.s. 43, 1981 (1983; pp. 97-126), p. 116; C. Wikander, 1986, pp. 46-48, nos. 59, 61, and 66. Grammichele. Raking sima. Late 7th century. C. Wikander, 1986, p. 36, no. 18. Himera. Lateral sima from Temple A. Ca. 630 B.c.? Himera, I, Campagne di scavo 1963-1965, A. Adriani et al., edd., Rome 1970, pp. 84-87; C. Wikander, 1986, pp. 36-37, no. 19. Satricum. Tiles from the first sanctuary of Mater Matuta and Building A. Second half of 7th century. G. Colonna, "I templi del Lazio fino al V secolo compreso," Archeologia Laziale 6, 1984

288

ORJAN WIKANDER

(pp. 396-411), pp. 400-401; Case e palazzi d'Etruria (Catalogue of the exhibition at Siena), Milan 1985, pp. 181-183. Ficana. Tiles and antefixes from private buildings. Second half of 7th century. C. Pavolini and A. Rathje, in Ficana. Una pietra miliare sulla strada per Roma, Rome 1981, pp. 75-87.

Gabii. Tiles from private buildings. End of 7th century. M. Guaitoli, "Gabii," PP 36, 1981 (pp. 152-173), p. 159; idem, "Le citta Latine fino al 338 a.C.-Urbanistica," Laziale 6, 1984, p. 378.

Archeologia

San Giovenale.Tiles from privatebuildings.650/625 B.C.O. Wikander,"ArchitecturalTerracottas from San Giovenale,"OpRom 13, 1981, pp. 69-89; I. Pohl, "SanGiovenaleda villaggio protovillanoviano a citta etrusca," PP 35, 1980, pp. 131-142. Tarquinia. Tiles. Late 7th century. M. Bonghi Jovini, in Gli etruschi di Tarquinia (Catalogue of

the exhibition at Milan), Modena 1986, p. 105.

Acquarossa. Tiles, antefixes, simas, akroteria, and revetment plaques from private buildings. Ca. 625 B.C. E. Rystedt, C. Wikander, and 0. Wikander, in Architettura etrusca nel Viterbese (Catalogue of the exhibition at Viterbo), Rome 1986, pp. 60-79, 99-108, and 130-134; C. Wikander, Acquarossa, I, ii, The Painted Architectural Terracottas, Typological and Decorative Analysis (Skrifter utgivna av Svenska Institutet i Rom, 40, 38:1:2), Stockholm 1988, pp.119-126.

Poggio Civitate (Murlo). Tiles, simas, and akroteria from the Orientalizing Building and the Southeast Building. 650/625 B.C. K. M. Phillips, Jr. et al., in Case e palazzi d'Etruria (Catalogue of the exhibition at Siena), Milan 1985, pp. 64-74; E. Nielsen, "Some Preliminary Thoughts on New and Old Terracottas," OpRom 16, 1987, pp. 91-119.

The difference in numbers between tiled roofs of the early and late 7th century is obvious, but there is also a typological difference. In the first half of the century, the tiles are rather complicated in construction; considering their shape, I would prefer to describe them as "proto-Laconian" combination tiles. After the middle of the century, we are dealing instead with tiles of more or less classical Corinthian and Laconian types. The changes are considerable, but seldom great enough to obscure their obvious origin in and development

from tiles of the "Isthmiatype". Real problems arise when one tries to explain the backgroundof this type of tile. Its sophisticatedconstructionhas caused some scholarsto look upon the type as the climax of a slow developmentfrom Mycenaean forerunners.4I was once inclined to do the same, but the complete lack of evidence for 8th-century predecessorsis not likely to be fortuitous.5 However advanced the Isthmia type may appear, it is now my belief that these tiles represent a first stage of a development towards simplification rather than the result of a gradual refinement. If so, they are presumably the invention of one inspired Corinthian potter, using his skill in a completely new field. An analysis of the technical details of the tiles in Martin, p. 68, note 2; ATK, pp. 195 and 261, note 28; Le Roy, 1967, p. 26. Drerup, GriechischeBaukunstin geometrischerZeit (ArcheologiaHomerica II, 0), Gottingen 1969, pp. 116-121; A. Mallwitz, "Kritischeszur ArchitekturGriechenlandsim 8. und 7. Jahrhundert,"AA (JdI 96) 1981, pp. 599-642; K. Fagerstrom,GreekIron Age Architecture.Developmentsthrough Changing Times (SIMA 81), Goteborg 1988, pp. 101-103. 4

5 See, for instance, H.

ARCHAIC ROOF TILES: THE FIRST GENERATIONS

289

comparisonwith those of contemporaryCorinthianpotterymay possibly reveal some common features. The basic structureof the Isthmiatype is a combinationof two elements,pan and cover, which can be seen most easily in profile. This combinationseems odd if separatepan tiles and cover tiles were not already in use. There may, however, be a different explanation. Besides thatch and tiles, we should reckonwith a third possible way of coveringa sloping roof:wooden shingles. Such shingleswere used in western Mediterraneancountrieslater in antiquity,6and they are to be found in variousparts of the world well into our own century. They do not normally leave any archaeologicaltrace, and so may well have been used in Late GeometricGreece, too. If that be the case, the shape of the Isthmia type of roof tiles is easily explained as a local potter'stransferenceof the basic structureof overlappingwooden shingles to terracottaand additionof the necessaryalterationsand refinements. The complicatedshape of the Isthmia tiles, what Mrs. Roebuck has called the "ingenious system of oblique cuttings"(p. 49 above), made them efficient,but it also made their productionextremelydifficult and time-consuming.A modernestimate suggests that seven workmen may have been occupied for two years with this single roof.7 It goes without saying that tiles of this type were no real alternativefor buildings other than monumental ones, temples in particular. The questionthen is: Why did builderschange over to tiled roofs at all? Thatched roofs are as goodor even betterin all respectsbut one:they presenta greaterfire hazard.8This fact alone makesthe early use of rooftiles on the templesof Corinthand Isthmiaunderstandable. Not only were templesparticularlycostlybuildings,they were also exposedmorethan others to fire becauseof the large altars in their immediatevicinity.9Rapid urbanizationincreased the risk of fire for privatearchitecturealso, and it is no surprisethat the secondhalf of the 7th centurywitnesseda change-overto tiled roofsin a large numberof towns. Urbanization and fire hazard may have been motivationenough for the change, but it couldnot have been achievedwithout a drasticreductionin the cost of producingtiles.10The contradictionbetween an urgent need for safer roof coveringon the one hand and unacceptable productioncostson the othermeant a tremendouschallengeto local potters.Concurrent with the rapid spreadof tiled roofsover the Hellenized world, the late 7th centurywitnessed dynamicand unhamperedexperimentationin an effortto simplifytile productionand make it moreefficient.Some solutionswere obviousand were arrivedat independentlyin different places;for instance,the similaritiesbetween Mycenaeanpan tiles from Berbatiand Archaic ones from inland Asia Minor (Pazarli) and Etruria11cannot reasonablybe explained by 6

For references,see 0. Wikander, 1988, p. 206, note 28.

7 W. Rostoker and E. Gebhard, "The Reproductionof Rooftiles for the Archaic Temple of Poseidon at

Isthmia, Greece,"JFA 8, 1981, pp. 211-227, particularlypp. 224-225. 8 See 0. Wikander, 1988, pp. 206-207. 9 I owe this suggestionto ProfessorB. Bergquist,Stockholm. 10Roof tiles neverthelessremained rather expensive throughout antiquity: 0. Wikander, 1988, pp. 206207, with notes 33-35. 11A. Akerstrom, "Zur Frage der mykenischen Dacheindeckung,"OpArch 2, 1941, pp. 164-173; ATK, Abb. 49. Cf. also the "Laconian"tiles found at Mycenaean Gla (I. Threpsiadis, <<'AvaoKafa't 'ApviJs (FXa) iaos?, IlpaKrLKa 1961 [pp. 28-40], p. 38, pl. 14b), and 0. Wikander, 1988, p. 205, with note 19. T7)s Kw7r

290

ORJAN WIKANDER

directcontacts.Othersolutionswere peculiarto particularareas;someof these were destined to spreadand survivefor a longerperiod,while somewere morelocal and short-lived. The ultimateresults of this experimentationvary considerably.Generallyspeaking,the Greek solutions have more in common with the Isthmia type than do the more radical Central Italic ones. This fact may go some way to explain why the final breakthroughof tiled roofs had to wait for so much longer in Greece than in Latium and Etruria:while in the 6th century in Greece, even monumental buildings like large stoas were mostly not tiled,12almost everyprivatehouse in Central Italy was coveredwith roof tiles. We must also considerthe fact that the solutionsarrivedat in ArchaicCentral Italy were in many respects similar to, or even identicalwith, those that finally prevailedin the Roman Empire. The only comparableachievementin the Greek world was the simplified Laconian system, which was finally to prevail in eastern Mediterranean countries. We thus have reason to question the view of the pre-eminenceof the Corinthiansystem;in my opinion, it proceeds from the wrong premises. While Corinthian tiles preservedat least parts of the "ingenioussystem of oblique cuttings",it was the extremelysimplifiedItalic tiles that made the completebreakthroughof tiled roofs possible;consequently,it is these which represent the true technicaladvancement,if tiled roofs should be consideredan advancementat all.1'3 Finally, I would like to commentupon the geographicaldistributionof 7th-centuryroof tiles. As alreadymentioned,my distributionmap can certainlybe criticizedin many details, but I would be surprised if the general picture were to change. Besides the eastern Greek world, which is representedby only four sites, there are three main areas of diffusion: Mainland Greece,with fourteensites;Northwest Greece,South Italy, and Sicily, with eight sites; and Central Italy, with seven. These three groups representthe main areas of developmentof Early Archaicarchitecturalterracottas.Ionia, which was later to becomea fourth such area, had barely begun its productionby 600 B.C.This fact strikinglyconfirmsa suspicion that has been growing for the last few decadesamong studentsof Etruscanterracottas: the long-acceptednotion of Ionian terracottasas an early inspirationfor (or even the origin of) Central Italic ones has no foundationat all.14 The local variationsamong 7th-centuryroof tiles were numerous;we havejust begun to grasp their true extent. But all this experimentationshows the local efforts to exploit a foreign idea. It has nothing to do with the ultimateorigin of the tiled roof as such, which, for the time being, may still with good reasonbe ascribedto Corinth. ORJAN WIKANDER LUNDSUNIVERSITET Klassiska Institutionen och Antikmuseet Solvegatan 2

S-227 62 Lund, Sweden 12

J. J. Coulton, The ArchitecturalDevelopmentof the GreekStoa, Oxford 1976, p. 37. The firing of roof tiles must have caused grave damageto Mediterraneanforests in antiquity. For other disadvantagesof tiled roofs, see 0. Wikander, 1988, pp. 206-207. 14 See 0. Wikander, "The Archaic Etruscan Sima," in Murlo and the Etruscans, R. De Puma and J. P. Small, edd., in press. 13

UBERLEGUNGEN ZUR TECHNISCHEN STRUKTUR UND FORMENTWICKLUNG ARCHAISCHER DACHTERRAKOTTEN (PLATES

IINES

49, 50)

Tondacherbleibt DER UNGEKLARTENPROBLEMEfruihergriechischer

die Frage nach der Einfuihrungund Erfindung dieser Dachdeckung. Als fruihestes bekanntesTondach, das stratigraphischdatiert werden kann, gelten die Ziegel des alteren Apollontempelsvon Korinth, die nach 680 v. Chr. entstanden sein mtissen,1und es gibt Forscher, die annehmen, dies sei das griechischeTempeldach, fur das die Tonziegeldekkung erstmals in Griechenlandangewandt worden sei. In der Folge dieses Daches hatten sich die bekanntenDachtypen, wie "lakonisches","korinthisches",und "sizilisches"Dach, erst herausgebildet.2Nun ist aber gerade das Dach des alteren Apollontempelsschon dermassen kompliziertund diffizil ausgekliigelt,dass mir jedes Verstandnisfur die Annahme, es k6nntesich hier um einen Prototyphandeln,fehlt. Die Kompliziertheitder Ziegel von Korinthsetzt Erfahrungvoraus.Und die Beobachtung, dass die protokorinthischenNormalziegelnichts anderesdarstellen,als zusammengesetzte lakonischeFlach- und Deckziegel lIsst mich nicht daranzweifeln, dass das lakonische Dach der Erfindungdes protokorinthischenvorausgeht,auch wenn wir noch keine lakonischenZiegel des 8. Jhs. belegenk6nnen.3 Da es keine unumstrittenenDatierungen fur Tondacher des beginnenden 7. Jhs. v. Chr. gibt, meine ich, dass es sich rechtfertigenlasst, zunachst eine formgeschichtlicheEntwicklungsreiheaufzustellen, die sich streng an formalen Kriterienorientiertund logische Fortbildungenaufzeigt. Die daraus ableitbarerelative Chronologiekann nur ein vorlaufiges Hilfsmittel bleiben, bis ein dichteres Gertist stratigraphischgesichert datierbarer Dachterrakotteneine guiltigeAussageerlaubt.Auch muss einschrankendvorweggenommen sein, dass sich mein Versuch nur auf die fur mich bessertiberschaubarenordostpeloponnesichen und argosaronischeBauhuttenbezieht, auch wenn die kulturelleund technologische Ausbreitung architektonischerFortschrittezu Beginn des 6. Jhs. geradezu rasant die gesamte griechischeSiedlungswelttiberzieht. Wenn wir davon ausgehen, dass die geneigten Dacher der geometrischenZeit neben Schilf oder Reet mit Schindeln oder Steinplatten gedeckt waren, so sollten die altesten I

Da alle hier aufgefuihrtenDacher in anderenVortragenschonbesprochenwurden, sollen sich die Zitate auf der Verfasser der Beitrage dieser Konferenzbeschranken.Zu Korinth, hier Roebuck (S. 47-63 oben); zum minimalen Zeitansatz "nach680", Heiden, 1987, S. 20. 2 0. Wikander, 1988, S. 205, und in seinem Beitrag zu dieser Konferenz(S. 285-290 oben). I ImmerhinkonstatiertR. Felsch (1979, S. 25) "dassschon im fruhen siebtenJahrhundertdie lakonische Ordnung voll entwickelt war" und bootische Ziegelbrennereieneinen ausgedehnten Export lakonischer Dacher betrieben(vgl. auch seinen Beitrag zu dieser Konferenz,S. 301-323 oben).

ERNST-LUDWIG SCHWANDNER

292

Tonziegel nachahmendeigentlichderenflache Oberflachebewahren.Dabei ware aberverkannt, dass solch ein Flachziegel keinerlei Vorteile gegentiber der alteren Deckungsart, sondern nur eine aufwendige Herstellungsart mit sich brachte. Die Erfindung des Tonziegels sollte aber einen Fortschritt bringen, und diesen sehe ich darin, dass durch die W6lbung der Tonplatte der Wasserabflusskanalisiertwerden konnte, und dass damit das leidige Problem der Fugendichtunggenial gel6sst wurde. Die einfache, in einer Richtung gewolbte Platte konnte hierbei samtliche Funktionen der Dachziegel vom First bis zur Traufe erftillen,und ergab ein extrem variablesBauglied, das kaum Problemeaufwarf. Dass es diese gew6lbte Tonplatte gewesen sein muss, deren Erfindungdie Entwicklung der anspruchsvollengriechischenTondacherausl6ste, glaube ich daran zu erkennen,dass zum einen in Zeiten gr6sserer Not und Anspruchslosigkeitdie Dachdeckungwieder mit dem gew6lbten Einheitsziegel auskam (vgl. byzantinische Dacher, deren System von "Monchund Nonne" auch unserejungeren Wohnhausdacheruiberdas Mittelalter bis ans Ende des 19. Jhs. bestimmensollte [P1.49:a]). Und zum anderenverbluiffteuns bei unterschiedlichenAusgrabungenvon einigen Wohnhausernder klassischenund hellenistischen Zeit (so z. B. in Kassopeund in Orraon/Ammotopos)4der Befund,dass auf diesen Hausern offensichtlich nur die grossen flachen lakonischen Ziegel verlegt waren (Pls. 49:b, 50). Eigene Deckziegel von abweichendemFormat fanden sich hier nicht im Dachversturz.Es gab also noch in klassischerund hellenistischerZeit Dacher, bei denen die grossen lakonische Flachziegel im Systemvon "Monchund Nonne"verlegt waren. Der Vorteil extremer Anpassungsfahigkeitan die Abmessungen unterschiedlichster Dacher durch Uberschieben in der Richtung der Dachneigung und beliebiges Auseinanderruickender Flachziegelbahnenbis fast zur Breite des als DeckziegelverwendetenFlachziegels in der Trauflinie des Gebaudes ist ganz evident (Fig. 1). Dem Nachteil grosser, offener Lehmnester der Bettung am Traufrand konnte am Holzbau mit einem einfachen zugehauenen Deckbrett weitgehend abgeholfenwerden. Diese Lehmnesterwaren gleichwohl ein Nachteil, der nicht zu tibersehenwar. Und es scheint mir nur logisch, dass es eine nachste Entwicklungsstufesein musste, die Offnungen unter und zwischen den Ziegeln zu

FIG.

I

1. Skizze zur lakonischenDachdeckungmit Einheitsziegeln

Letzte Vorberichtezu Kassope in W. Hoepfner und E.-L. Schwandner,Haus und Stadt im klassischen Griechenland.Wohnenin der klassischenPolis I, Munchen 1985, S. 75-140; Orraon,ebd., S. 109, Anm. 195.

STRUKTUR UND FORMENTWICKLUNG ARCHAISCHER DACHTERRAKOTTEN

293

FIG.2. Skizze zum ausgebildetenlakonischenDach mit unterschiedlichenFlach- und Deckziegeln

verkleinern,was durch die Erftihrungeines kleinformatigerenspeziellen Deckziegels moglich war (Fig. 2).5 Hierbei ging zwar ein guter Teil der Verlegungsfreiheitin der Richtung der Trauflinie verloren, doch blieb die Korrekturmoglichkeitbeim Ziegelversatznoch erstaunlichhoch, zumal, da der gr6ssereliegende Ziegel im Problemfallefast beliebig seitlich gekuirztwerden konnte,ohne seine Funktion zu beeintrachtigen. Herstellungstechnischgab es fur diesen Einheitsziegelkaum Probleme.Wie der Ziegel geformtwurde, glaube ich noch in den spaterenuns erhaltenenMonumentenzum Ziegelstandardzu erkennen.Von den drei bekanntenBeispielen stellt die Marmorplattevon der Agora von Assos eindeutig einen Standard-Messtischdar, in dem einerseits Normmasse eingelassen waren und andererseitsnormierte Beispiele fur lakonische und korinthische

151

,3

AK

FIG.3-.S

FIG. 3. Skizze der "Patrize"von Messene (nach A. K. Orlandos)und einer ausgestrichenenMatrize

I Dass dieser Uibergangtatsaichlichstattgefundenhat, sehe ich bestatigtin archaischenlakonischenDeckziegeln aus Istria (vgl. Beitrag K. Zimmermann [S. 223-233 oben]) und Samos (vgl. Beitrag A. Ohnesorg [S. 181-192 oben]), die im Zuschnittweitgehendden Flachziegelnentsprechenund nur geringfiigigschmaler sind. Der Firnis auf der konvexenOberseiteweist sie aber eindeutigals Deckziegel aus. Dacher der Vorstufe (wie spaiterin Kassope und Orraon) mussten natiirlichbeidseitiggefirnisstsein.

ERNST-LUDWIG SCHWANDNER

294

FIG. 4. Skizze zum lakonischenDach mit geschlossenerStirn der Kalypteream Traufrand

Dachziegel abgebildetwaren.6Entsprechendder Ausfuihrungkonntensie nur dazu dienen, fertige Ziegel auf ihre normierteForm und Abmessungenhin zu uberprufenund allenfalls Masse zu nehmen, um eine Matrize zu formen. Ganz anders sehen zwei Beispiele von Athen und Messene aus.7 Hier sind nur lakonischeZiegeltypen dargestellt,aber nicht fast freiplastisch, wie in Assos, sondern flachanliegendin einer muldenformigenAushohlung des Marmorblocks.In diesenVertiefungenist das Massnehmenausgesprochenunpraktisch und auch der Vergleich mit einem fertigen Tonziegel nicht eben besonders gtinstig. Ich mochte daher annehmen, dass es sich in Messene und Athen um staatlich genormtePatrizen handelte, aus denen sich ohne grossenAufwand beliebig viele Matrizen streichenliessen (Fig. 3). Da die Patrize immerdie selbe blieb, erubrigtesichjedes weitere Massnehmen, da das Schwindenvor und wahrend des Brandessicherlicheinkalkuliertwar, und das einfache lakonischeDach ja grosse Toleranzen zuliess. Eine einschneidendeAnderung musste diese Ziegelkonstruktiondurchmachen,als an anspruchsvollenGebauden, zunachst wohl nur an Sakralbauten,die durchlaufendeglatte Trauflinie mit Geison aus Ton oder Stein eingefuhrtwurde. Die nur in satterLehmbettung stabil verlegbarenlakonischenZiegeltypen mussten konstruktionsbedingtmit diesem glatten Dachrand in Konflikt geraten und die doch immernochsichtbarenLehmnester unter den Ziegeln warfen auch ein asthetischesProblemauf. Eine moglicheLosung bot die Erfindung eines speziellen Deckziegels fur die Traufe, bei dem die Stirnseitedurcheine Schurze geschlossenwurde (Fig. 4). Diese Schurze konnte mit glatter Unterseite einen festen Halt auf der Geisonoberseiteherstellen, im Geison eingelassenwerden oder lang herabgezogen uber das Geison uberhangen,wobei sich durch Gewicht und Reibungswiderstandebenfalls eine gute Fixierung ergab. In beiden Fallen war damit auch das immer noch erforderliche Lehmnest verdeckt. Nun mochte man diese Erfindung gern dem Sikyonier Boutades zuschreiben,da diese Schurze sofort das Ornament,das Relief, ja die Maske als Schmuckan sich zog.8Doch muss diese Form nicht gleich mit dem Peloponnesierverbundenwerden, da H. Bacon,J. T. Clark, und R. Koldewey,Investigationsat Assos,Cambridge,Mass. 1902, S. 71. Athen: G. P. Stevens,"ATile Standardin the Agora of AncientAthens,"Hesperia 19, 1950, S. 174-188. Messene:A. K. Orlandos, "Epyov1960, S. 166; 1971, S. 168. 8 Vgl. Plinius, NH xxxv. 152. Eine andere Interpretationder "prostypa"und "ectypa"des Boutadesvertritt zuletzt M. Mertens-Horn in ihrem Beitrag zur Konferenz(S. 235-250 oben). 6 F. 7

STRUKTUR UND FORMENTWICKLUNG ARCHAISCHER DACHTERRAKOTTEN

_

._ . , __

_,

=

_

295

_-_

FIG. 5. Skizze zum lakonischenDach mit geschlossenerStirn der Stroteream Traufrand

in der Folge dieser Ziegeltyp eigentlich nur fur das westliche Griechenland,Unteritalien, und lonien bestimmendwurde. In der nordostlichenPeloponnes, genauer gesagt in Korinth, entschied man sich fur eine andere Losung der Umformung des lakonischenDaches. Hier wurde nicht der Deckziegel, sondernder liegende Ziegel dem Geisonrandangepasstindem man dem lakonischen Traufziegel an seiner Stirnseiteein horizontalesUnterlager anformte.Ein durchlaufendes Band der Ziegelstirn mit senkrechtenFugen war die Folge, ein asthetischwie konstruktiv sauberes Ergebnis mit nur einem ungunstigen Nachteil: die Korrekturim Ziegelversatz war nicht mehr durchden eine Fuge uberdeckendenDeckziegel moglich (Fig. 5). Das, was alle zuvor aufgefuhrten Dachdeckungsartennoch gemeinsam als Vorteil zeigten, konnte nur aufgegeben werden, indem nun die zu deckende Dachflache exakt in Ziegelbahnen vorberechnetwerden musste. Hierbei war insbesondereauch der Schwund des Ziegels von der Ausformungbis zum Brand einzukalkulieren,der immerhin zwischen 7 und 10%betragen konnteund somit grosse Erfahrungdes Ziegelherstellerserforderte,wenn seine Ziegel fugendichtund exakt die vorgegebeneDachflacheabdeckensollten.9 Mit der Einfuhrung des voll aufliegenden flachen Randziegels mit dichtem Fugenschlusswar also festgelegt,dass Fehler der Planung nicht mehr durchgeringereoderweitere Uberdeckung des Kalypters korrigiertwerden konnten, sondern nur noch durch geringe Abarbeitung an den Flachziegeln, den Stroteren. Damit war einerseits die Position des Dachziegels, des Kalypters,unverruckbargeworden,und es war nur eine logische Konsequenz, dass wir bei den altesten Dachern dieser Art den Kalyptermit dem Stroternun fest verbundenvorfinden,und andererseitsmusste sich an den seitlichenStossflachender Traufziegel die Anathyrosedurchsetzen,um an der einzig nochmoglichenStelle die Korrekturder Ziegelbreitenzu erleichtern. Auch wenn der Ziegelschnittdieser fruhen Dacher, fur die sich die treffendeBezeichnung "protokorinthischeDacher" eingeburgert hat, mit der Konstruktion lakonischer 9 Zum Ausmass des Schwindensvgl. Stevens,a. 0. (Anm. 7), S. 178.

296

ERNST-LUDWIG SCHWANDNER

Dacher eigentlich nichts mehr gemein hat, ist m.E. ihre Herleitung von lakonischenZiegeltypen evident. Der Normalziegel dieser Dacher besteht formal eindeutig aus der Verbindung eines normalen lakonischen Stroters mit einem lakonischen Kalypter. In der Auf- wie Untersicht und im Schnitt liegt der einzige Unterschiedin der festen Verbindung der ursprtinglichgetrenntenTeile. Und die Schwierigkeitfur die Erfinder,sich von einer vertrautenForm zu l6sen, zeigt am bestender Zuschnittder Traufziegel:10 wahrendauf der Unterseiteim Bereichdes Auflagersauf dem Geisonder Ziegel glatt abgestrichenwordenist, geht im ruckwartigenBereich der Ziegel wieder in die geschwungeneProfilierunglakonischerZiegel uber. Herstellungs- und versatztechnischist dies unsinnigund absolutunn6tig. Ich kannmir das Phanomennur so erklaren,dass auch die Traufziegelaus dergleichenlakonischen Form wie die Normalziegelgestrichen,und dass die Sonderheitender traufseitigen Stirnmit der Hand nachgeformtwordensind. Hierbei ist es naturlichauchm6glich,dassaus einer Patrize des Normalziegels eine ummodellierteMatrize gewonnen wurde. Allerdings bleibt es dann verwunderlich,warum in solch einer abgeandertenMatrize nicht auch die ruckwartigeZiegelunterseiteexakt geglattetwordenist. Die entscheidendeNeuerung bliebjedenfalls der geschlosseneTraufabschlussder Ziegelreihen, der dem neuen Monumentalitatsanspruchder griechischen Sakralarchitektur besser entsprach und einen organischen Ubergang vom breiten Geisonband zur Traufziegelreihe erst herstellt. Es kann deshalb nicht verwundern,dass sich sofort der Gestaltungswille dieserneuen Dachrandformannahm.Die nun uberder horizontalenGeisonlinie seitlich zu Spitzen aufgewolbten Strotere erforderteneine Anpassung der Kalypterstirn, da die runde, "lakonische"Kalypterformnicht mit dem geschlossenen Band der Stroterstirnflachenzu vereinbaren waren (Fig. 5). Die Angleichung erfolgte zunachst wiederum nur am Traufende der Kalyptere, wahrend die optisch nicht erfassbarenhinteren

p.

F

FI.6

kzezu

poooithshnDc"vo

yu

shi

10 Heiden (1987, S. 18-19 mit Anm. 45) m6chteallenfalls eine 'entfernte Ahnlichkeit"in der Kriimmung der Ziegel von Korinthmit lakonischenZiegeln gelten lassen. Dass hier sehr wohl eine "Mischform"vorliegt, hatt ihm deutlichwerden miissen, da die Unterseite des Hegemon eben nicht wie behauptet"anseiner Unterseite vollig plan gearbeitist."

STRUKTUR UND FORMENTWICKLUNG ARCHAISCHER DACHTERRAKOTTEN

297

.~~~~~~~~~~~~~~~~~~~~~~~~~~ .~~~~~~~~~~~~~~~~~~~~~~~~~

FIG.7. Skizze zum "Horerdach"

Ziegelteile (schon am Traufziegel) weiterhin ihren "lakonischen"Querschnittbeibehielten (Fig. 6). Auch hier gibt es keinen herstellungs- oder versatztechnischzwingenden Grund fuirdie kompliziertenFormverschleifungen,wenn man nicht das Bewahren einer gewohnten Form in Rechnung stellt. Und die kleinen Aufsaitzevon Pseudoantefixen,wie sie uns erstmalsam Dach des Poseidonstempelsvon Isthmiabegegnen,II halte ich dochweniger fuir Brechungsbarrierenfuirden Wasserstrahl, als fuirtastende Versuche, die neugewonnene Ansicht des traufseitigen Ziegelbandeszu rhythmisieren,d. h. die neue technische Form 12 ornamentalaufzuwerten.

Dies scheint mir deutlicher zu werden, wenn man die Fortentwicklungdieser Dachformen betrachtet (Fig. 7). Die "Hornerdaicher" stellen die direkte Nachfolge des isthmischen Daches dar.13Die vorsichtigenAnderungender Stirnziegelwerden nun konsequent auf den gesamten Ziegelzuschnittuibertragen,und "Lakonisches"findet sich nur noch im 0. Broneer,Isthmia I, Temple of Poseidon,Princeton 1971, S. 40-53. Heiden (1987, S. 20, Anm. 56) mochte im Gegensatz zum Ausgraberdiese Deutung nicht akzeptieren, wobei er aber einen logischen Grund fur seine technische Erklarung schuldig bleiben muss. Der dem isthmischen Ziegel entfernt verwandte Traufziegel des "buntenDachs" von Kalydon zeigt, dass man bei einer Wasserbarrierein der Ziegelmitte Gegenmassnahmentreffen musste, um zu verhindern,dass zuviel Wasser an die Kalyptere druckte (schrage Grate bei Dyggve, S. 141 und Taf. XVI:J, XVII:2). Die Grate auf dem Wasserspeierziegelvon Thermos sollten gerade dazu dienen, den Wasserstrahlin die Mitte des Ziegels zu leiten (AntikeDenkmdlerII, 1908, Taf. 53:A1). Was hatte die Wasserbarrierein Isthmia erreichensollen? 13Vgl. die Beitrage von M.-F. Billot (S. 95-139 oben), N. Cooper (S. 65-93 oben), und Ch. A. Pfaff (S. 149-156 oben). N. Winter (S. 13-32 oben) trennt die Hornerdacherstrikt als argivischeSonderformvon den korinthischenDachformen,da in Korinth die Kombinationsziegelbeibehaltenwerden und bisher keine Hornerziegelgefundenwurden. Das ist vermutlichzutreffend.Es bleibt aber untibersehbar,dass die H6rnerdacher unmittelbarauf das Dach von Isthmia zuruckzufuhrensind, und dabei mag es bedeutungsvollsein, dass dieses korinthischeHeiligtum am SaronischenGolf lag. Entwicklungsgeschichtlichandertsich bei dieser Annahme nichts, doch in der absoluten Chronologie sind naturlich regionales Beharrungsvermogenund Neuerungen schwererzu fassen. 12

298

ERNST-LUDWIG SCHWANDNER

~~~~~~~~~~~~~~~~~" /

FIG.

8. Skizze zur Herstellung alterarchaischerkorinthischerStrotere

Profil der Stroteroberseiten.Dafur wird das bescheideneSchmuckmotivder Antefix- und Pseudoantefix-Spitzenvon Isthmia nun zu einem kraftigakzentuiertemWellenbandan der Traufe. Dennoch sind die jetzt dreifachenSpitzen an den Antefixen und Pseudoantefixen nicht so bedeutungsvoll,dass sie eine eigene Matrize fur die Traufziegel erforderthatten. Sie wurden im lederhartenZustandvor dem Brandden Normalziegelnvon Hand frei angeformt. Ulberhauptist die Ziegelherstellungoffensichtlichbei diesen Daichernschon sehr rationell geplant. Am Dach des alteren Aphaiatempels14etwa waren mit Ausnahme der Sima und AkrotereZiegelformennur fur vier Typen hergestelltworden:je eine fur Strotere und Kalyptereund je eine fur Firstziegel und Firstkalyptere(die Sima mag vielleicht nicht zur urspruinglichenDachplanunggehort haben und sollte hier ausser Betrachtbleiben). Der Antefix unterscheidetsich vom Kalypter nur durch seine angeformtenHornerspitzen und seine ktirzere Lange. Er war also aus der Form des Normalkalyptersohne Schwierigkeitenherzustellen. Die Stroteregleichen in ihrer Oberseitenw6lbungso weitgehend lakonischen Ziegeln, dass ich mir vorstellen kann, dass ihre Matrize mit einem parallelseitigemlakonischenZiegel hergestelltsein konnte (Fig. 8). Jedenfalls zeigt die Unterschneidungder Ziegellippe an den Normalstroteren,dass sie mit einem Messer nach der Ausformungin der Matrize freihand herausgeschnittenworden ist. Wenn die Stirnpartie vor dem Herausschneidenleicht schraigausgestellt war, wie es bei den Hegemones, den Traufziegeln der Fall war, so konntenTrauf- und Normalziegelwiederum aus einer Form gewonnen werden. Dass bei den H6rnerdaicherndie Flach- und Deckziegelnicht mehr zusammenhangend aus einer Form gestrichenwurden, wie bei den protokorinthischenDachern, mochte man auf praktische Erwagungen zurtickfuihren:Die Matrize wairenwesentlich komplizierter und die Handhabe der grossen Kombinationsziegelist vom Gewicht und der Gr6sse her einfach schwieriger.Dennoch ist die Lage der Kalyptereso strengfixiert, dass man sie ohne 14

Schwandner,1985, S. 75-85, 126-128.

STRUKTUR UND FORMENTWICKLUNG ARCHAISCHER DACHTERRAKOTTEN

v to

FIG.

_

vC

299

c

9. Skizze zum korinthischenDach mit hexagonalenAntefixen

weiteres dem Stroterauch hatte anformenkonnen,wie es in Korinthnoch fur langere Zeit uiblichblieb. Das Hornerornamentder Traufziegel hat dem Schmuckbeduirfnisbald nicht mehr genugt und Bereicherungsversuchewie in Nemea, Kalapodi,und auf der AthenerAkropolis z. B. zeigen, das der Hornerziegelzur Ausbildungvon Schmuckfelderndenkbarungeeignet war.15 So fiihrt die Entwicklung von den Hornerantefixenzu Deckziegeln, die zum einen mehr der Stirnflacheboten, und zum anderensich mehr der Geradlinigkeitund kubischen Geschlossenheitder fruihenSteintempel anpassten (Fig. 9). Zwar wurde zunachst noch die gew6lbte Oberseitedes Flachziegelsbeibehalten,doch verliertdie Deckziegelkontur vollstandigihre geschwungenenKanten. Es ergibt sich dabei ein hexagonalerAntefix.16 Die dem Flachziegel folgende geschwungeneUnterseite geriet aber sofort in Konflikt mit dem geradlinig begrenzten Ornamentfeldder Antefixstirn. So war als logische Folge die Aufgabe der geschwungenenZiegeldeckeam Traufranderforderlich(Fig. 10). Die hierbei ganz der dorischenArchitekturentsprechendeparallele Bandzonedes traufseitigenZiegelabschlusseskonnte nun mit giebelartigenStirnfeldernder Deckziegel akzentuiertwerden, auf denen sich in der Folge jenes Antefixornamentaufbauen konnte, das als Lotus- und Palmettenkombinationseit der zweiten Halfte des 6. Jh. die dorischeArchitekturunserer Region fur Jahrhundertebeherrschensollte. Die Ausbildung der dorischen Trauf- und Normalziegel war hiermit soweit abgeschlossen, dass sich im Vergleich zur vorhergehendenEntwicklung keine gravierenden 15 Akropolis:TdA II, S. 27-29, Taf. 1, Stirnziegel I und II. Nemea: S. G. Miller, "Excavationsat Nemea 1979,"Hesperia 49, 1980 (S. 178-205), S. 185-187; ders., "Excavationsat Nemea, 1980,"Hesperia 50, 1981 (S. 45-67), S. 52. Kalapodi:Beitrag G. Hubner (S. 167-174 oben). 16Auf die unzutreffendeBezeichnung"antefixespentagonales"fur diese Ziegel (Le Roy, 1967, S. 32-37) habe ich schon fruiherhingewiesen (Schwandner, 1985, S. 128 mit Anm. 245). Da M.-F. Billot in ihrem Beitrag hier (S. 95-139 oben) nun auch deutlich zwischen hexagonalen und pentagonalenAntefixen unterscheidet und eine ganz ahnliche Entwicklungsliniesieht, sollen Belege und Argumentehier nicht nochmals aufgefuihrtwerden.

300

ERNST-LUDWIG SCHWANDNER

FIG.

10. Skizze zum korinthischenDach mit pentagonalenAntefixen

Anderungenmehr ergaben. Nur in technischenDetails wie den Ausbildungender Randaufbiegung des Ziegels gegen aufsteigendeFeuchtigkeitam oberen Ende des Flachziegels oder der seitlichenAufbiegungunter dem Kalypter,der Zungenausbildung,der Anathyrosenform in der Stossfuge der Hegemones etc. finden sich geringf6rmigeUnterschiede,die aber jeweils mehr ortliche Handwerkstraditionenbelegen als technische und formale Wandlungen,wie sie in standigerFolge im 7. und 6. Jh. v. Chr. Architektenund Ziegelproduzentenvon Erfindungzu Erfindungdrangten.Aber ob diese Entwicklung,wie sie hier zu skizzierenversuchtwurde, nun tatsachlichnur vom fruhen 7. Jh. v. Chr. bis zur Mitte des 6. Jh. v. Chr. ablief, oder ob sie fruher anzusetzen erst gemahlich anlaufend, dann in dichter Folge sprungartigsich ausbreitete,ist heute erst bestenfallszu ahnen. Solange zuwenig stratigraphischabgesicherteBefunde in solch einer Entwicklungsreiheeingeklinkt werden k6nnen, bleibt vieles spekulativ. Dennoch sollte nicht vergessen werden, dass die fast explosionsartigsich ausbreitendenNeuerungen der fruhen dorischenSteinarchitektur am Beginn des 6. Jh. v. Chr. ganz besondereAuswirkungenauf die Traufranderder Tonziegeldacherhaben musste, und dass hier eine besondereDichte von Wandlung und Fortschrittzu erwartenist. ERNST-LUDWIG SCHWANDNER DEUTSCHESARCHAOLOGISCHES INSTITUT

Podbielskiallee69-71 D-1000 Berlin 33 Federal Republic of Germany

PLATE 49

a. Aliverion, Euboa: Mittelalterliche Dachdeckung mit "lakonischen"Einheitsziegeln (Visantina= Monch und Nonne), 1976

f,'.

E

Kassope,Haus 1: lakonischerZiegel~~~~~~~~~~~~~~b. Deckziegel

PLATE 50

Kso Hu E.-L.

SCHWAN iNER: ST .UKTURUNDk4F' ARCHAIS'HER RMENTWIKLU G DA'HTERRAKON..A

.<st b?''''''w',' W; + *t5e

Kassope, Haus 5: Ziegellager in Raum f ohne schmalere Deckziegel

DACHTERRAKOTTEN ARCHAISCHER E.-L. SCHWANDNER:STRUKTUR UND FORMENTWICKLUNG

FURTHER STAMPED ROOF TILES FROM CENTRAL GREECE, ATTICA, AND THE PELOPONNESE (PLATES

51-56)

ALMOSTTEN YEARS AGO I assembleda class of ancientobjectswhich had hitherto been virtually unnoticed:stamp impressions on Archaic and Classical roof tiles of Laconiantype and on one Corinthiantile.1 A feature commonto all 60 fragmentsof tiles is that the impressionsare invariablylocatedon the lower face of the tiles which was eventually hidden, unlike the large group of tiles of Hellenistic and later date, which are always stamped on the upper face.2 The late stamp marks served, as has long been accepted,to facilitate checkingby the client or as protectionagainst theft of the finishedtiles. The early impressions,on the other hand, seem to have servedmerelyas an accountingaid for internal workshoprequirementsor perhapsalso as an advertisingmediumfor the workshop.3In my first study I was able to assemble 35 differenttypes of stamps, some of which had been describedin scatteredpublications,and some which were unknown. Since that time at least 34 furthertypes on more than 96 stampedtile fragmentshave come to my attention,so that the total volume of materialtoday amountsto over 147 tiles with at least 69 differentstamps. In view of the abundanceof material I can deal here only with the essential characteristics;for details the readermust refer to the Catalogue.The new stamps from the excavations at the Aphaia sanctuaryand near Kolonna on Aigina are only summarilytouched upon, since these finds will be publishedby the excavators.Similarly only the stamp representations from Nemea are considered,without discussion of the details of the roof. The new stamp impressionsare listed in the Catalogue in the order of their provenancefrom North to South;within the topographicalgroups they are arrangedaccordingto subject. At Kynos,the port of the LokrianOpous, a Laconiantile with a rosettestamp (E 1) was found by Fanouria Dakoronia. To judge from its description,this stamped tile, which I have not inspectedpersonally,should perhapsbe attributedto the workshopof Dabychos,4 the existence of which I established in my first report, or to its immediate sphere of influence. At Kyparissi, which is also situated in Opountian Lokris, Dakoronia found two Laconian tiles with the circular stamp of the Dabychos workshop;5this stamp had 1 Felsch, 1979, pp. 1-40. For permissionto study and publish here both new and older, so far unpublished material, I am greatly obliged to: M.-F. Billot, W. Coulson, F. Dakoronia, S. Miller, M. Ohly-Dumm, 0. Picard, E.-L. Schwandner,H. A. Thompson, and C. K. Williams, II. 2 G. Hubner, "Dachterrakottenaus dem Kerameikosvon Athen," AM 88, 1973 (pp. 67-143), p. 86; eadem, "AntefixadeorumAthenarum,"AM 91, 1976 (pp. 175-183), p. 180, note 29. 3 Felsch, 1979, pp. 18-19. 4 Felsch, 1979, pp. 14-17, fig. 10, pls. 5:6-7, 6:1-6. 5See F. Dakoronia, pp. 175-180 above.

302

RAINER C. S. FELSCH

a

_~~~~~~~~~~~~~~~~~~~~

b

FIG.

/L

c

1. a) Reconstructionof stamp Kalapodi B 2 and B 3: Stamp of OIONOS with representationof the mantic bird. b) Impression of Prikon's stamp, Kalapodi C 4. c) Stamp impression of EPMAIO*, KalapodiA 2

previously been attested only in the near-by Phokian sanctuaryof Artemis at Hyampolis, near the village of Kalapodi,and at Kolonnaon the island of Aigina. In the sanctuaryof Artemis near Kalapodithe number of 12 stampedtiles bearing ten different stamp marks has been increasedby a further nine Laconian tile fragmentswith three new stamp representations.Among the impressionsof the stampsalreadyknown, two appearancesof the Prikon stamp should be singled out (Kalapodi C 4 and C 5, Fig. I:b, P1.5 1), since this stamp had so far been attestedat Kalapodionly on a small fragmentwith the letter sequence-IA-.6Any objectionthat this fragmentis too small to warrantattribution 6 Felsch, 1979, pp.

14, fig. 10:9; 17; 27, C 2; pl. 6:6.

FURTHER STAMPED ROOF TILES

303

to the Prikon stamp, otherwise encounteredonly on Aigina, is invalidatedby these new examples. The identificationis especially significantbecausethe fragmentin question constitutesproofthat Dabychosand Prikonwere activetogetherin a workshopwhich supplied both Kalapodiand Aigina. Includedin the catalogueof the first report but illustratedhere for the first time is the papyrusor flower stamp KalapodiE 6 (P1. 51),7which on stylisticgroundsmight belong to the early tile stamps. New types of stamp impressions are preservedon six tile fragments from Kalapodi. Two, perhaps even three tiles show a bird with clearly set off head, eye, and hookedbeak within an elongatedpictorialfield (B 2-4, Figs. 1:a,4:a, P1.51). From the preservedletters the legend can easily be restored as Oicov6s,the bird of prey depicted, and obviously the visual translationof the tilemaker'sname. It shouldbe notedthat one tile is stampedtwice, a feature occasionallyencounteredon other tiles.8 Another impression of late origin shows within an irregular field a degeneratepalmette flanked by two equally degeneratevolutes (KalapodiE 7, P1. 51). Finally, two tile fragmentsbear a name stampwithin an irregularly roundedfield (A 2 and A3, Fig. 1:c, P1.51). One of the stampspreservesonly the first letter, while the other is intact. It bears the inscriptionEPMAIO written from left to right with letter forms which do not occur before the late 5th centuryB.C. A date for this stamp later than the 5th centuryB.C. also accordswith the findspot. One Laconian pan tile from Paralimni in Boiotia carries on its lower face the retrogrademonogramR within a large, rectangularfield (ParalimniA 1, P1.51). Althoughthe form of the tailed rho would point to the 5th century, the thin wash of black glaze would argue in favor of a later date. Many years ago a stamped tile was discoveredin the Amphiareionof Thebes;it was overlookedin my first report.9Within a circularfield this stamp shows an anchor, an emblem already known from Tanagra.10Evidently Laconian tiles with anchor marks are widespread, since such tiles have also been found in the Athenian Agora (Athens, Agora E8,E9,andE 10,Pl. 53). The tiles from the Athenian Agora constitute the largest unpublished group of Laconian tiles with stamp impressions. Among the anchor stamps, the complete specimen (Athens, Agora E 8, P1. 53) stems from a fill which should be dated in the early 5th century B.C. The context providesa concreteindicationfor dating the group of anchor stamps, which does not admit of stylistic classification;this indicationis supportedby the contextof the Theban find. Probablyall anchorstampsfromAthens, Tanagra, and Thebes originated from a single workshop.This assumptionis strengthenedby the fact that anothertile from the Agora bears a stamp impression in the shape of a scallop shell (E 7, Fig. 2, P1. 53) identical with one found at Thebes, and perhaps also with yet another from Kolonna on I

Felsch, 1979, p. 29, E 6. See Felsch, 1979, pp. 18-19; for emblems of the kind mentionedabove, see most recently F. Brommer, "RedendeZeichen,"AA (JdI 103) 1988, pp. 69-70. 9 A. D. Keramopoullos,<
304

RAINER C. S. FELSCH

FIG.2. Scallop-shellstamp impression,Athens, Agora E 7

Aigina.1IThe impressionsfrom the Athenian Agora and from Thebes representthe third stamp for which a supraregionaldistributionhas been proved.12 Of relativelyfrequentoccurrencein the Agora are simple round stamps dividedby two cross-barsinto four equal segments (E 12-14, P1. 54). Two of these tile fragments come from a deposit of ostraka, together with pottery dated not later than the middle of the 5th centuryB.C. (E 12, E 13). Further specimenswith the same or similar impressionsare listed in the excavationdiary from tin 702. One similar impressionfeaturesan additional,sunken bordercircle (E 14, P1.54). Finally, I would also include a tile fragmentwith a four-spoked wheel stamp (E 11, P1. 53), despite its coming from a Late Roman fill, among the early stamp marks on Laconian tiles. A comparablestamp impression is found on a tile in the Epigraphical Museum in Athens.13One tile fragmentfrom Gyphtokastroon the Boeotian border,the ancient Eleutherai,shows a similar motif (GyphtokastroE 1, P1. 54). Another tile from the Agora bears a circular impression which, as far as can be ascertained in its damaged state, depicts within a round field eight flowers in profile (E 6, P1. 53). They are set out in sunk relief and arrangedin such a way that four flowers with intersectingstems alternatewith four stemless flowers. The interplay of raised and sunken pictorialelementsmakes the exact representationdifficultto discern. Two Agora tile fragmentsare markedwith the same stamp (E 3 and E 4, P1.52), albeit in both instances only part of the complex palmette within an irregularlyroundedfield is "IThebes: Felsch, 1979, pp. 29-30, E 1, pl. 4:5; Kolonna:pp. 12, fig. 9:2, 37, E 6, pl. 4:6.

Felsch, 1979, p. 17: workshopsof Dabychos and of the "dancer'sstamp". A?eAr 25, 1970, EvAAoy7j'AqvIJ@v>>, Felsch, 1979, p. 39, C 1; D. Peppa-Delmouzou,
1'

FURTHER STAMPED ROOF TILES

305

preserved.For this stamp too the find contextprovidesa terminusante quem, viz., the mid5th century B.C. Yet another tile bears a stamp with a nine-petaled palmette composedof loops (E 5, P1. 52). This tile comes from a fill of the periodbefore480 B.C. Its linear style is somewhatreminiscentof the bird stamp from Kalapodi (B 2-4, Fig. 1:a, P1. 51). Of particularinterest is a larger fragmentof a covertile from the Agora (E 1 and E 2, P1. 52). Its lower face bears the impressionsof two differentstamps, both of excellent quality; except for the Dabychos-Prikontile from Kalapodi, this Agora tile is unique in being markedby two differentstamps.14The impressionon the left, of which only one-quarteris preserved,shows a palmette abovetwo volutes on a horizontalbase;on the right, the curved field is filled by a nine-petaledpalmetteabovetwo horizontaltendril-lyreswhich arejoined at the base and terminate in volutes. The lower angle between the lyres is filled by a bud shaped like a palmette heart. Stylistic considerationssuggest a date for this ornament around 510-500 B.C.15 This specimen is exceptional among stamped tiles because it confirmsthe suppositionstatedabovethat it was not simply by chancethat individualtiles were given two differentstamp marksin the workshops.16 Only two tiles from the Athenian Agora bear figural stamp impressions.One pan tile is stampedwith a bird, perhapsa partridge,within a roundfield (D 2, P1.52), while on another, a roundedfield with bulging rim containsa horse stridingright (D 1, P1.52). The highly expressive depiction of the horse, reminiscentin its slimness of the horse types of the 7th centuryB.C., neverthelessseems stylisticallycomparableto representationsof horses in the early 6th centuryB.C. on accountof its loweredhead, sharplyset off fromneck and mane.17 Stamped representationson Laconian tiles from Attica are not confined to Athens (Acropolis,Agora, Kerameikos).Besides the stamp from Eleutherai-Gyphtokastro(P1. 54) a Laconiantile with a stamp impressionwas recentlyreportedfrom Rhamnous.18 According to the excavator'sreport, the tile was found in the earliest strata of the sanctuary of Nemesis, directlyabovevirgin soil. The stamp shows within a rectangularfield with rounded cornersa lion striding right, head turned backwardsand tail raised. Between the legs a name runs right to-left, read by the excavatoras EZTON.In the published photographit appearsto me that there are two more letters,boustrophedon,at the beginningof the legend Felsch, 1979, pp. 14, fig. 10:9, 17, pl. 6:6. No exact parallel for this ornamentseems to be known in either vase painting or relief. The closest approximationsare palmette tendril-lyressuch as P. Jacobsthal, Ornamentegriechischer Vasen, Berlin 1927, pp. 122-124, pls. 72:d (Epiktetos) and, more developed,78:a (Douris). Comparableare some antefixes, although these feature only simple, rolled-in tendrils and no tendril-lyres:Olympia II, p. 194, pl. CXVIII:1. 16 Felsch, 1979, pp. 18-19. 17 While the slenderness of this horse is somewhat reminiscent of early Archaic equine representations (cf. J. Schilbach,"EineGruppe grosserprotoarchaischerPferdestatuettenaus Olympia,"AM 99, 1984 [pp. 515], pls. 1-4), neverthelessthe peculiarity of setting off the horse's head sharply from neck and mane and forming it as a distinct unit is in accordancewith the equine types of the 6th century (e.g., E. Simon and M. and A. Hirmer, Die griechischen Vasen, Munich 1976, pls. 28:a, 48 or the Attic horseheadamphorae, pl. 62). All the horses cited, however,have a thickerneck, an indicationperhapsthat the representationon the stamp might be earlier. 18 B. C. Petrakos,<<'Avao-Ka47j 1981 (1983; pp. 120-140), p. 121; idem, <<'AvaHpaKTrKca PauvoovTrosv>, 1982 (1984; pp. 127-162), pp. 135-136, pl. 91:c. HpaKTrKLa o-Ka4nlPauv.oivTros>, 14

15

306

RAINER C. S. FELSCH

a

b

c d FIG.3. a) Stamp impression with gorgoneion,Aigina, Aphaia Sanctuary D 17. b) Stamp impression with remains of letters, Aigina, Aphaia Sanctuary A 1. c) Lion stamp, Aigina, Aphaia Sanctuary D 8. d) Lion stamp, Aigina, Aphaia SanctuaryB 1

so that perhaps the craftsman'sname should be read as AAEZTON.So far as comparison between photographs and originals is admissible, the lion stamp from Rhamnous would seem to be identicalto a partially preservedstamp from the Sanctuaryof Aphaia on Aigina (Aigina, Aphaia B 1, Fig. 3:d). This is the fourth stamp,then, for which it has been possible to prove supraregionaldistribution. I do not intend to enter here into a detaileddiscussionof the stamp markson Laconian roof tiles found in the course of the new excavationsin the Sanctuaryof Aphaia on Aigina, since these will be published later. Twelve new tiles with stamp impressionshave come to my attention,although the total number is probablylarger. These stamps are of particular importancein that they all presumablycome from the terrace fill for the later Temple of Aphaia and thereforebelong to the 6th centuryB.C.19 The well-known stamp of a lion in a 19

For the dating of the terrace fill, see D. Williams, "Aegina,Aphaia-Tempel. XI. The Pottery from the

FURTHER STAMPED ROOF TILES

307

a

b

~~~~~~~~~~~c

FIG. 4. a) KalapodiB 3. b) Corinth D 1. c) Corinth E 1

round field is representedby three new impressions, of which only the best preservedis illustrated here (Aigina, Aphaia D 8, Fig. 3:c).20Three tiles likewise show the equally familiar running hare, and three others, the serpent stamp.21Previouslyunknown stamps includea rectangularimpressionof which is preservedthe upper right cornerwith the letter sequence ... Er'o . .. from a legend (A 1, Fig. 3:b); one rectangularstamp with a gorgoneion (D 17, Fig. 3:a); and one rectangularstamp with a lion striding right, head turned backwards,and a legend in the space between the legs (B 1, Fig. 3:d). In all probabilitythis impressionis identicalwith that on the tile from Rhamnousdiscussedabove. The excavations at Kolonna on Aigina since 1979 have also yielded new tile stamps. The director, Hans Walter, most generously put the older finds at my disposal, but my informationon the more recent finds is only second-hand.Accordingto the data available, the recently discoveredimpressions both repeat types I have already published, allowing correctionsto be made to my reconstructions,and represent new types. Thus one rectangular lion stamp is most probablya repetitionof the stamp known from the Aphaia sanctuary and from Rhamnous (cf. Fig. 3:d). It has also been claimed that for the first t'ime Corinthianridge tiles with rich stampeddecorationhave been found at Kolonna. From the Peloponneseonly one single Laconianrooftile with a stamp markhas cometo my attention,an old find from Corinth.The fragmentaryrectangularstamp (D 1, Fig. 4:b, P1. 55) shows a crouchingsphinx, facing right, of Late Archaicstyle. Presumablythe composition should be restoredas a pair of antitheticalsphinxes. At this time I know of no furtherLaconianrooftiles datingfromArchaictimes or the 5th century B.C. Systematicsearchesin museums and excavationstoreroomswilllprobablyadd more examples to the existing corpus, and furtherexcavations,as those at Kalapodiand on Aigina have shown, may furnish still others. In all.events,the amountof material presently availableis too meagerto permitconclusionsregardingeconomyor historyto be drawn. SecondLimestoneTemple and the Later History of the Sanctuary,"AA (JdI 102) 1987 (pp. 629-680), p. 629 for bibliography,pp. 669-674 for date around 500-490 B.C. 20 Cf. Felsch, 1979, p. 10, fig. 6:1, p. 37, Aigina, Aphaia D 1. 21 Cf. Felsch, 1979, p. 10, fig. 6:3, p. 38, Aigina, Aphaia D 3.

308

RAINER C. S. FELSCH

I have encounteredsimilar stamp impressionson the lower face of tiles of the so-called Corinthiantype only on the Corinthiancovertile in Munich depictinga more or less acrobatic symplegma and dating from the early 5th century B.C. A counterpartof this stamp from the same kind of roof was brought to my notice by Virginia Grace (ProvenanceUnknown, D 3, P1.56). This tile, which was given to her as a presentbefore 1939 and had ever since been regardedas a modernfake, was donatedto the collectionof the AmericanSchool of Classical Studies after publicationof the Munich impression.22 No other Archaic Corinthiantiles are yet known which bear stamp impressionson the underside. There are, however, Archaic Corinthian tiles with such impressions on the upper face; so far they are known only from the Samian Heraion, which yielded twelve specimensof the monogramro,23 and from Nemea.24 The roof of the Nemean temple, hipped but otherwise similar to the Archaic roofs of Kalapodi, should, accordingto the informationavailable from the excavators,be dated in the first half of the 6th centuryB.C.25 Accordingto the preliminaryreportsand my own investigations,there are 38 stampedtiles known from Nemea, the impressionsinvariablyappearing on the visible upper surface. Altogetherthere occur four different stamps. Three are completelylinear, featuringthe "S-ornament"formedby two impressedlines (Nemea, E 37, P1. 55), the "keyhole"(E 14, P1. 55), and the "teardrop"(E 29, P1. 55). The fourth stamp shows a thirteen-petaledrosette (E 1 and E 2, P1. 55). A feature commonto all these stamps is the absenceof a framed field. Only the rosettemight be regardedas evidenceof a possible relationshipto Laconianstamp impressions.Providedthat all the stampedtiles belong to the same roof, the distributionof the impressionson the roof of the Archaictemple reveals a certainsystem. The "S"-stampoccursexclusivelyon seven pan tiles and the "teardrop"only on three cover tiles, while the "keyhole"is found on sixteen pan tiles and three covertiles. The most impressivestamp,the rosette,occurson four pan tiles (eavestiles?) and on three antefixes, twice on each. The distributionof four differentstamps presumablyon one roof thus manifests a definite pattern. It may well be that four individualtilers within one workshop have dividedtheir tasks among themselves,each producingmainly one type of tile. Thus the "S"-tilerproducedonly pan tiles, the "teardrop"tiler only ordinarycover tiles, and the "keyhole"tiler chiefly pan tiles but also ordinarycovertiles, while the master tiler, who used the rosettestamp,was in chargeof producingthe morecomplicatedantefixes 22 I thank V. Grace for her kindness in acquainting me with the tile, which was then in her possession. W. Coulson most willingly grantedme permissionto publish it. 23 See A. Ohnesorg, pp. 181-192 above. The second type of stamp impression from the Samian Heraion (Felsch, 1979, p. 39, Samos D 1) shows the Archaic cult statue on the right, and on the left the forepartof a bull turned toward the top, a representationfamiliar from Samian coins, confirmingan early date for this stamp. The latter detail had remainedunexplaineduntil now. 24 For the roof of the Archaictemple at Nemea, see S. G. Miller, "Excavationsat Nemea, 1979,"Hesperia 49, 1980 (pp. 178-205), pp. 185-187, fig. 3; idem, "Excavationsat Nemea, 1980," Hesperia 90, 1981 (pp. 45-67), pp. 52-54. 25 S. Miller kindly offered me the opportunityto study the entire body of relevant pottery. This seems to contain nothing which would require a date in the secondquarter of the century. One find from one destruction phase of the temple still merits closer considerationas to its significance.

FURTHER STAMPED ROOF TILES

309

and perhapseavestiles. Such an interpretationis admittedlyboundto remainpure hypothesis as long as we do not know the essential componentsof the roof in their entirety. The particularsignificanceof the finds from Nemea, however, lies in the fact that they provide us with the oldest stratigraphicallydated stampedtiles. So far we cannot discern a continuous path leading from the oldest Corinthian tile stamped on the upper surface to the later canonicaltiles of the same class, since the previously mentioned Classical Corinthian tile bearing an erotic scene is stamped, after the fashion of Laconian tiles, on the underside.There are only sporadicspecimens of Corinthian tiles which might fill this gap. It is only with reservationsthat a Corinthiancovertile from Corinthcan be mentioned here (Corinth, E 1, Fig. 4:c, P1. 55). The upper surfacebears a stamp with a meticulously executed blossom ornament. It should perhaps be dated somewhere in the 5th or 4th century B.C. Nevertheless, a stamp from Tanagra bearing, in additionto a Hellenistic Damosios stamp, a mark in the form of a Boeotian shield (Tanagra E 3) should warn us to be cautiousas regardsdating.26 Only two furtherLate Classical stamp impressionson Corinthiantiles have so far come to my attention,both on covertiles fromArgos.Each tile carriesa rectangularfield in which stands a stag facing right (Argos, D 1 and D 2, P1. 56). The depictions are stylistically consistentwith the conventionsof Classical gems.27It is, however, impossibleto determine their precise date, which is probablyas late as the end of the 4th centuryB.C. With the exceptionof one specimenwhich I shall discussbelow, this is the sum of early roof tiles with stamp impressionswhich have come to my attentionsince the publicationof the initial catalogue. Decisive new insights are not providedby this new material. All the same, certain points can be made. A map of all the locationsof finds (Fig. 5) demonstrates that stamped roof tiles occur almost exclusively in eastern Central Greece, in Attica, on Aigina, and in the northeasternPeloponnese. Supraregionalworkshops have been identified throughtiles with identicalstamp markings(Fig. 6): the Dabychos-Prikonworkshop, stamps of which are found at Kalapodi, at Kyparissi,and on Aigina; the lion-stamp workshop, at Rhamnous and on Aigina; and the scallop-shell workshop, at Thebes and in the Athenian Agora. It is likely that the anchor stamps from the Agora, from Tanagra, and from Thebes also come from such a workshop.28The various workshopsemployedsimilar marking systems with a single or double stamp, or two differentimpressionson one tile.29 The significance of these marking systems has not yet been clarified. Furthermore, the names of the manufacturersare known to us in only a few instances; we may mention Dabychos and Prikon, then Brysas,30... eston, Agonippas,31and Oionos, as well as three Felsch, 1979, pl. 3:6. Cf., e.g.: D. Ohly, GriechischeGemmen, Insel-Verlag, n.d., pl. 12; P. Zazoff, Die antiken Gemmen (Handbuchder Archdologie),Munich 1983, p. 149, fig. 42:f. 28 Cf. also Felsch, 1979, pp. 17-19, "Blattstern-Werkstatt". 29 KalapodiC 2, stamps of Dabychos and of Prikon;Agora E 1, E 2, two differentpalmette stamps. 310Felsch, 1979, p. 12, fig. 8:1. 31 Felsch, 1979, p. 15, findspotunknown C 1. 26

27

Q

~~~~~~~~~~~~~~~

FIG.5. Sites where early stampedroof tiles have been found. Open dots:Corinthiantiles.

FIG. 6. Distribution of workshops.Full dots: Dabychos and Prikon. Open dots:scallop shell. Segmented

312

RAINER C. S. FELSCH

incompletelegends.32Simple name stamps with the name Ermaios are known from Kalapodi and others from Koroneia.33They are probablythe inauguratorsof the series of later name stamps and should not be datedbeforethe 4th centuryB.C., or perhapseven later. Regional differencesare apparentin the choiceof the stampmotifs,but it is not possible to draw detailedconclusions.The earlieststampsappearto be purely figuralor ornamental, without legends, while stamps which mentionthe manufacturer'sname emerge only in the late 6th centuryB.C. The chronologicalclassificationof the stamp impressionsimmediatelyraises the questions of which are the oldest tile stamps and, by implication,when did the Laconianroof in its canonicalform come into being? The clues furnishedby the tile stamps are so far rather scanty. On the evidence of the find contexts, the stamps from the sanctuaryof Aphaia on Aigina, several from the Athenian Agora, the stamp from Rhamnous, that from the Amphiareion of Thebes, and some stamps from Kalapodi must be dated within the 6th century B.C. On the strength of stylistic criteria some stamps should be dated even earlier, probablywithin the 7th century.34These assignmentspose certainproblems,since, as far as I am aware, not a single Laconian roof has been securely dated to such an early time, although some investigatorsare inclined to regard the roof of the Heraion at Olympia, the oldest that is datable by stratigraphiccontext, as the latest Laconian temple roof (as distinguished from house roofs). Considerationof all excavationresults suggest that this roof may well date after 600 B.C. and before ca. 570.35The question is thereforewhether the 32 Felsch, 1979, fig. 6:1-3; here Fig

1:a. HlpaKrLKa1975 (1977) (pp. 392-414), p. 401, 33T. Spyropoulos, (>, fig. 2:4.5, -MnlOBIAIand -nlO.I-; Felsch, 1979, p. 29, A 1. 34 Felsch, 1979, pp. 4-6, Findspot Unknown D 1; p. 7, Tanagra D 1, D 2; pp. 9-11, Aigina, Aphaia D 1, D 3;p. 12, Kalapodi E 1, E 6. 35 Althoughthe communisopinio places the Heraion ca. 600 B.C., this datingis not yet consideredabsolutely certain.This is surprising,given the fact that the archaeologicalcontextis unequivocal:beneaththe cella floor was found an Early Corinthianalabastronwhich can be dated around 600 (-590) B.C. and providesa definite terminuspost quem for the building of the temple: W. Dinsmoor and H. Searls, "The Date of the Olympia Heraeum,"AJA 49, 1945 (pp. 62-80), p. 68, with note 57; H. Riemann, "Die Bauphasendes Heraions von Olympia,"JdI 61/62, 1947/1948 (pp. 30-54), p. 50; A. Mallwitz, "Das Heraion von Olympia und seine Vorganger,"JdI 81, 1966 (pp. 310-313), p. 328, note 26 (recently discoveredpottery, still of the 7th cenMunich 1972, tury B.C.), p. 357, summarizedin H.-V. Herrmann,Olympia,Heiligtum und Wettkampfst&atte, pp. 93-97; W. B. Dinsmoor, The Architectureof Ancient Greece, 4th ed., New York 1975, p. 93, note 2 (althoughmaintainingthe erroneousassumptionof a previoustemple of the late 8th centuryB.c.). One would thereforehave to fix the inceptionof the building of the Heraion somewherearound or after 600 B.C. and its completion,allowing for adequate time for construction,around 570 at the latest. Further evidence for the dating of the Heraion, providingits interpretationis correct,is the limestonehead which U. Sinn, in the main Der sog. followingMallwitz, has interpretedas a sphinx fromthe pedimentalreliefof the Heraion: `"EKTVlroJv. Hera-Kopf aus Olympia,"AA (JdI 99) 1984, pp. 77-87. It is datedby Sinn ca. 590-580 B.C. but mightbe later: cf. U. Jantzen, EAA V, 1963, s.v. Olimpia (pp. 635-656), p. 643: "nonpiiuanticodel periodointornoal 570." This tallies with the assignmentof the Corfu pedimentaround 570 B.C., a dating supportedby B. S. Ridgway (The ArchaicStyle in GreekSculpture,Princeton 1977, pp. 191-195). The decade590-580 B.C. constitutesa terminusante quem for the inceptionof the Heraion, and so the durationof constructioncouldbe fixed between 600 and 570 at the maximum.A late date for the completionof the Heraion and consequentlyan even later one for the earliest stone replacementcapital can easily be reconciledwith the short time it took the Doric stone capitalsto evolvewithin the 6th century;this view is supportedby the datingof the temple at Kalapodi.For the

FURTHER STAMPED ROOF TILES

313

stamp representationson Laconiantiles, stylisticallyeven earlier, should indeedbe datedby their style or whether they belong instead to an archaizingartistictrend and should therefore be dated much later than their style would suggest. To formulate the question differently:are there any artifactsof such early times which are comparableto roof-tile stamps and which may perhaps have servedas modelsfor the earliesttile stamps? There exists a large body of pottery, characteristicof the 7th centuryB.C., which bears stamped decoration and features pictorial subjects generally related to those of the tile stamps. The relief pottery of Corinth is one example;36another, more important, is the stampedrelief potteryof Crete.37Particularlyrevealingand relevant'tQour argumentseem to be two stampeddepictionsof standingwarriorswhich on the basis of iconographicdetails have been convincingly dated by John Boardman to the time around or shortly after 700 B.c.38 As he points out, these stamp impressionsare exceptionallylarge, their pictorial field being over 10 cm. high. Even larger, with a height of 11.3 cm., is the tile stamp in the Kanellopoulos Museum, stylistically the earliest by far. It is related also to the Cretan stamps in its subjectmatter, a helmeted warrior-dancer.39It is undoubtedlyof a later date than the other representationsof warriors,but preciselybecauseof its similarityto them can hardly be dated later than 650 B.C. I thus place the earliest known stamped Laconian tile, the specimen in the KanellopoulosMuseum, at about the same time the first Protocorinthian roofs known to us, those of the temples at Corinth and at Isthmia, were built.40 distinctly later capitals at Kalapodi, the stratigraphiccontext has yielded a surprisingly low date, i.e., after ca. 570 and at the latest perhaps ca. 550 B.C.: Felsch, 1987, pp. 17-18, figs. 30-33, regardingthe destruction layer of the predecessorand the potterywhich dates the foundationdepositfor the Archaic SouthernTemple; pp. 22-24, concerningthe Archaic Southern Temple; detail of the capital: fig. 42. These datings virtually compel us to accept the late dating of the Temple of Apollo at Corinth in the third quarter of the 6th centuryB.C.: S. S. Weinberg,"Onthe Date of the Temple of Apollo at Corinth,"Hesperia 8,1939, pp. 191-199. C. K. Williams has demonstratedthe problematiccharacterof the issue by distinguishingbetween a high and a low chronology("DoricArchitectureand Early Capitals in Corinth,"AM 99, 1984, pp. 67-75) and, not yet acquaintedwith the dating contextsof Kalapodi,declaredhimself in favor of the high (p. 72, note 21). In my opinion, and providedthat the pottery chronologyis correct,the new contextsargue clearly in favor of a low chronologyand a rapid evolutionof the Doric stone capitaltowardsthe Late Archaicform. 36 S. S. Weinberg, "CorinthianRelief Ware: Pre-Hellenistic Period,"Hesperia 23, 1954, pp. 109-137, pls. 25-33; regardingthe earliest stamp impressionsfrom Corinth, precursorsof the later amphora stamps, from before 750 B.C., cf. C. A. Pfaff, "A Geometric Well at Corinth: Well 1981-6," Hesperia 57, 1988 (pp. 21-80), pp. 39-40, pls. 31-32. The only previouslyknown stamp on a vessel from the Geometricperiod comes from Pithekoussai:J. N. Coldstream,GeometricGreece,London 1977, p. 229, fig. 75:d. The stamp is identical to the impressionson a votive tablet from the Heraion of Samos: D. Ohly, "Friihe Tonfiguren aus dem Heraion von Samos II.," AM 66, 1941 (pp. 1-46), pl. 11:416;cf. J. Boardman,GreekGems and Finger Rings, London 1970, p. 112. 37 The earliest example so far is a pithos lid from Knossos,dated to the Protogeometricto Early Geometric period:H. W. Catling, "Archaeologyin Greece 1982-83," AR 1982-1983, p. 52, fig. 93. From the end of the 8th century B.C. there come some figural stamp impressions:J. N. Coldstream,GeometricGreece, London 1977, p. 276 with notes 15-17, fig. 87:f. 38J. Boardman,"ArchaicFinds at Knossos,"BSA 57, 1962 (pp. 28-34), pp. 31-32, fig. 3, pl. 4:a. 39 Felsch, 1979, p. 5, fig. 1. 40 Isthmia:0. Broneer,Isthmia, I, Temple of Poseidon,Princeton1971, pp. 40-55. The excavatorsdate the temple within the first half of the 7th centuryB.C., taking the perirrhanterionfound in the rubbleof the temple (op. cit., p. 3, pl. 7) as the terminusante quem. This assumptionis not altogethersecure,since the basin might

314

RAINER C. S. FELSCH

While it might be possible to rest content with this conclusion,one cannot help wondering whether there might not exist even older material. In this connectionit may well be worth examining a remarkablestamp impression from Eretria which Claude Berard has publishedand discussedin detail.41This impressionis on a clay plaque (EretriaD 1, Pl. 56) discoveredin the lowest stratumof the bothrosnear the Heroon. Its edge is roughlychipped into a roundedshape and the surfaceof the reverseis chippedaway; in its presentstate it is 5.1 cm. in diameterand 1.7 cm. thick. Within a rectangularfield a stallion trots to the left; under its hooves a fish faces right. Berard points out that the obverse side of the relief is completely flat and that consequently the possibilityof its being a fragmentof a pithos must be ruled out. He therefore regardsthe objectas a kind of disk. Berardrightly datesthe stratumof the find to the second quarterof the 7th centuryB.C., while placing the creationof the relief within the first quarter of that century. Stylisticallyhe relatesthe objectto the glyptic art of Melos and places it within the sphere of the early votive pinakes. Although Berardhas arguedthis attributionpersuasively,I cannotagree with it. In the first place, the stamp is considerablylarger than all known Melian and Argive seal stamps. Moreoverthe combinationof horse and fish, as Berardalso observed,is reminiscentof the representationson some Boeotianfibulae with decorativeplates. Furthermore,such "platefibulae"from Central Greece and Thessaly as a rule feature horses on one side of the plate and fishes on the other.42Finally, analogies may also be drawn from the contemporary three-dimensionalsculptureof Central Greece.A bronzehorse from Anavrain Lokris may be cited as undoubtedlyrelated to the relief picture, both in its conceptionof volume and proportionsand in the executionof severaldetails, such as the shape of the muzzle and legs (P1. 56).43 These factorsshow that the Melian origin of the stamp impressionfrom Eretria is by no means certainand that one would be equally if not morejustifiedin attributingit to a workshopin Central Greece. In the second place, the identificationproposed by Berard for the relief plaque as a discus or round votive tablet is not convincing.The ultimate function of the object,which has been roughly chippedinto a roundedshape, seems to have been as a lid for an amphora; hundredsof such tile fragmentsor sherds,unevenlychippedinto a circularshape, are found in most excavations,Kalapodi not excluded. In the case in point the maker of the lid has shown respectfor the existing relief picture,becausehe either enjoyedit or regardedit as an objectof curiosity.An exactly parallel situationis attestedby a tile fragmentfrom Kalapodi, a Laconiancovertile convertedinto a lid in such a way as to preservethe impressionon the well be earlier and may perhaps have been transferredto the pronaos only after the erection of the temple. Corinth:Robinson (AM) now dates the temple ca. 680 B.C. (p. 57, note 5) on the evidenceof the potteryin the layers of working chips, which, however, can provideonly a terminuspost quem for the temple and its subsequent roofing:0. Wikander, 1988, p. 205. 41 C. Berard, "Note sur la fouille au sud de l'Heroon,"AntK 12, 1969 (pp. 74-79), pp. 77-79, pl. 36:2. 42 Cf. K. Kilian, Fibeln in Thessalien von der mykenischen bis zur archaischen Zeit (Prdhistorische BronzefundeXIV, 2), Munich 1975, pp. 105-137; idem, "Trachtzubeh6rder Eisenzeit zwischen Agais und Adria,"PZ 50, 1975 (pp. 9-140), pls. 24:1, 25:3-4. 43Kilian, "Trachtzubehor"(footnote42 above), pl. 23:11.

FURTHER STAMPED ROOF TILES

315

lower surface,which is the circularstamp of Dabychos.4 The significanceattachedto these lids may in either case have been far greaterthan we are able to discerntoday, given the fact that the Eretria lid lay in a votive bothros, while the Kalapodi lid was discoveredimmediately behind the votive bench within the sanctuaryof the 5th century B.C.; these are circumstanceswhich are probablyno accident. Originally the clay plaque from Eretria must surely have served a different purpose. For the votivetablet assumedby Berard,however,this plaque, with an original thicknessof at least 2 cm., seems to be too bulky. The obverse,the surfaceof which would offer a reliable clue for determining whether it representeda votive pinax or a Laconian pan tile with glazed upper surface,is unfortunatelyobliterated.An examinationof the plaque at Eretria undertakentogether with P. Themelis revealed that both its clay and its manner of execution are in every way in keeping with the oldest Laconiantiles; consequentlythe stamp impressionmay well be that of an early Laconianpan tile. A clear-cut answer to this question is as yet not attainable. Nevertheless, on the evidence of the stamp impressionfrom Eretria, it does not appear impossiblethat the earliest Laconianroof tiles may be tracedback as far as the beginningof the 7th centuryB.C., a circumstancenot entirelywithout significancefor the elucidationof the emergenceof the Laconian tiled roof, and that there alreadyexisted specializedworkshopsat this early time. CATALOGUE OF STAMP IMPRESSIONS This cataloguecontinuesthe numberingbegun in Felsch, 1979, pp. 26-40. The stampedrepresentationsare listed by provenancefrom North to South and groupedby subject: A: Inscription B: Figural with legend C: Ornamentalwith legend D: Figural E: Ornamental The formatof the first cataloguehas been retainedexcept in the case of materialfromthe AthenianAgora;the entries for these stamps combineinformationfrom the architecturalinventoryof the Agora, a commentaryon the dating of the individual find contexts which H. A. Thompson kindly put at my disposal, and my own remarks.I have abstainedfrom giving descriptionsand measurementsof the tiles from Nemea in ordernot to anticipatethe publicationof the Archaicroof. KYNOS, SETTLEMENT

KALAPODI, SANCTUARY

E. ORNAMENTAL

A. INSCRIPTION

E 1. Laconiantile, fragment

A 2. Laconianpan tile, brokenon Fig. 1:c,P1.51 all sides L. 8.25, W. 9.33, Th. 2.1 cm. Clay gray to orange buff, fine inclusions, well fired. Upper surface, blackishred glaze. The unglazedundersidebearsan irregularelongated

On the unglazedunderside,roundstamp-impression rosette. Findspot: Kynos, rescue excavation directed by F. Dakoronia. Unpublished

44 Felsch, 1979, p. 27, C 3, pl. 6:5, completelyanalogousto the Prikon stamp Kalapodi C 4 (here P1. 51); furtherstampedtiles refashionedas lids: Felsch, 1979, p. 29, Kalapodi E 5, pl. 5:2; pp. 38-39, Samos A 1.

RAINER C. S. FELSCH

316

inscriptionstamp.Largelettersin high relief,inscription runningleft to right:EPMAIOL.H. 1.2-1.3 cm. Findspot: Kalapodi 1977, context no. 3739/1213, layer of Hellenistic dump. Inv. no. Z 486 A 3. Laconianpan tile, brokenon all sides P1. 51 L. 6.1, W. 10.9, Th. 2.06 cm. Clay reddishbeige, well fired. Upper surface,violet-brownglaze. On clay of underside,within a round field, the first letter of the stamp A 2 is preserved. Findspot: Kalapodi 31.8.79, I 22, western slope, without stratigraphiccontext. Inv. no. Z 258 B. FIGURAL

WITH LEGEND

B 2. Laconianpan tile, brokenon Fig. 1:a, P1. 51 all sides L. 10.2, W. 10.5, Th. 2.05 cm. Clay with reddish beige core. Outside yellowish beige, sand and clay inclusions, well fired. Upper face with purple to blackishglaze. On the unglazed underside,two impressionsof the same stamp, their respective front parts juxtaposed and slightly intersecting.Head of a bird with hooked beak and large eye, traced in narrow relief lines. Above the head the letters Ol . .. run right to left, to be restored by the letters on tile B 3 to read OION,the Greek name of the bird of prey (olcovOs) represented. Findspot: Kalapodi 1979, context no. 17053/1819, slope. Inv. no. Z 352 Figs. 1:a, 4:a, P1. 51 B 3. Laconianpan tile, fragmentof edge L. 11.7, W. 14.1, Th. 1.6-1.7 cm. Clay as B 2. Upper surfaceglazed, as B 2, more reddishat the edge. On the unglazed underside, impression of an elongated stampwithout distinctlyrecognizabletermination, ca. one-half preserved.Hind part of the body and tail of the bird shown in B 2. Probablythe same tile, althoughthe edges do not fit together.Abovethe bird, within the field, the letters ... ON ... Findspot: Kalapodi 1979, context no. 17033/9091, destruction layer of Classical Temple I, after 426 B.C.

Inv. no. Z 319

B 4. Laconianpan tile, brokenon all sides P1. 51 L. 25.2, W. ca. 15, Th. 1.9 cm. Clay with black inclusions.Upper surfaceglazed violet brown. Underside smoothed with spatula 5 cm. wide, remains of a stamp impressionpresumablyidenticalto stamps B 2 and B 3. Findspot:Kalapodi28.8.79, contextno. 12065/22, Late Classicaltile fill I. Inv. no. Z 353 C. ORNAMENTAL WITH LEGEND

C 4. Laconianpan tile, chipped into shape of lid; broken

Fig. 1:b, P1. 51

L. 7.82, W. 4.15, Th. 1.7 cm. Clay light gray, fine, fired extra well, identical to clay of C 3. Upper surface,thick yellow-brownglaze. On the clay of the underside,an abradedstamp impression of which ca. one-third is preserved.Round stamp with band of letters around a rosette with eight petals. Diameter not determinable.Clockwise ... .)0\iO. Identicalto inscription:TAn1P(IQONEIA stamp Kalapodi C 2, left stamp, C 5; Aigina, Kolonna C 5. Findspot:Kalapodi31.8.79, contextno. 11073/27, Late Classicaltile fill I. Inv. no. Z 259 C 5. Laconian covertile, brokenon all sides P1. 51 L. 10.3, W. 5.95, Th. 1.9 cm. Clay very fine, pale reddish to gray beige, fired ringing hard. Upper surface,shiny black wash. On the unglazed underside,one-half of a stamp impression,identicalto C 4. Diameter of stamp 6.5 cm. Clockwiseinscription: (TAnPI)QONEIA ... (1V0O) Findspot: Kalapodi 1979, context no. 15021/81, Late Antique rubble fill. Inv. no. Z 351 E. ORNAMENTAL

E 6. See Felsch, 1979, p. 29.

P1. 51

E 7. Laconianpan tile, brokenon all sides P1. 51 L. 15.5, W. 16.9, Th. 1.8-2.4 cm. Clay light reddish beige, somewhat porous, with inclusions of sand and small stones. Well fired. Upper surface, thin violet-brownwash.

FURTHER STAMPED ROOF TILES On the unglazed underside, stamp impression with irregularcontour.Stamp:H. 6.3, W. 5.7 cm. Within the pictorial field, stylized palmette flanked by two volutes on joint base. Findspot: Kalapodi 1979, context no. 15104/83, Late Antique or later pit. Inv. no. Z 318 PARALIMNI A. INSCRIPTION A 1. Laconianpan tile, fragment

P1. 51

L. 16.5, W. 16.4, Th. 2.0 cm. Clay reddishbrown, with some organic inclusions. Scatteredblackish stones. 7.54R 8/8. Upper surface, thin wash of black glaze. On the unglazed underside,rectangularstamp with large retrogradeR. Stamp:H. 6.5, W. 3.5 cm. Findspot:small tile fill beside the road about halfway along the southernshoreof the lake (E. Vanderpool, 1966). Collection of the American School of Classical Studies Inv. no. ASA.53

E 1. Laconian pan tile, brokenon all sides L. and W. unknown, Th. 1.1-1.6 cm.-Type of clay unknown. Upper surfaceglazed black brown or with red wash. On the unglazedunderside,roundstampimpression: anchorwithin circle. Diameterof stamp 6.5 cm. Findspot: Thebes, Amphiaraion, rubble layer with finds dated shortly after 500 B.C.at the latest. Publication: D. Keramopoullos, AcEAr 3, 1917, pp. 263-264, fig. 181. ERETRIA, HEROON FIGURAL

D 1. Laconian pan tile (? or votive pinax) chipped into shape of lid

GYPHTOKASTRO, CASTLE E. ORNAMENTAL

P1. 54

L. 15.5, W. 11, Th. 1.7 cm. Type of clay not mentioned. Upper surface,black glaze.

E. ORNAMENTAL

D.

On the clay of the underside (obverse)rectangular, presumably square, stamp impression. Within the narrow framed pictorial field a stallion trotting left over a fish swimming right. Different posture of front and hind legs. The similarity noted by Berard of the forelegs to the human front legs of Archaic centaurs is probably unintentional, as is shown by the shape of the rear hooves. Findspot:bothrosto the southwest of the Heroon, at the bottomof the votive deposit. Inv. no. Eretria FK 1437 Publication:C. Berard,"Notesur la fouille au sud de l'Heroon,"AntK 12, 1969 (pp. 74-79), pp. 7779, pl. 36:3, and p. 2. For the secondaryuse as a lid, cf. the stamped tiles Kalapodi C 3 (Felsch, 1979, p. 27, pl. 6:5) and C 4 (above),in which the stamped representation is respected in the same way, E 5 (Felsch, 1979, p. 29, pl. 5:2, where ills. 2 and 3 have been mistakenly interchanged), Samos A 1 (ibid., pp. 38-39).

E 1. Laconian pan tile, brokenon all sides

THEBES

317

On the rough, unglazed underside, a round stamp impression, of which ca. one-fourth is preserved: four-spoked wheel with unidentified filling ornament within sunken circular stamp field. Diameter of stamp ca. 6.5, of wheel ca. 5 cm. Findspot:Gyphtokastro,Western gate, 1961. Collection of the American School of Classical Studies, Athens. Inv. no. AST.19 I am indebted to Sarah Morris for drawing my attention to this item, as well as for its description and measurements. RHAMNOUS, SANCTUARY OF NEMESIS

P1. 56

Max. Diam. 5.1, Th. 1.7 cm. Clay brick red, coarse,little purified,with mica inclusions,somewhat brittle. Upper face (reverse) chipped over entire surface.

B. FIGURAL

WITH LEGEND

B 1. Laconian (pan?) tile, fragment L. 47.2 cm. Upper face black glazed. On the unglazed underside, rectangularstamp impressionwith roundedcorners.In the pictorial field,

318

RAINER C. S. FELSCH

lion striding right with tail raised and head turned back. Between the legs, inscription, right to left, reading (according to the excavator) EZTON; according to the published illustration it would seem insteadto read AAEETONor similar. Stamp:L. 7.3, H. 5.4, H. of letters 0.9 (E), 0.45 cm. (0). Findspot:Sanctuaryof Nemesis, terracefill of the Archaictemple. Inv. no. 500 Publication:B. Petrakos,<'Ava0-Ka4+ PaMuvovHlpaKTLKa 1982 (1984), pp. 135-136, pl. 91:c. TOs>>, Cf. Aigina B 1 and Fig. 3:d. ATHENS, AGORA D.

FIGURAL

D 1. Laconian pan tile P1. 52 Max. dim. 8.0, Th. 1.4 cm. Broken all around. Light brown clay. Dull red glaze on concaveside. On unglazed convex side, part of a circular stamp (est. Diam. 5 cm.) with a horse in relief: head and one foreleg preserved. From an early deposit in the valley to the west of the Areopagus(D 18) along with sherdsof early 5th centuryB.C., some Geometric.July 30, 1947. Inv. no. A 1284 P1. 52 D 2. Laconian pan tile, fragment Max. dim. 13.5, Th. 1.5 cm. Red glaze on the upper slightly concavesurface. On the underside, circular stamp with bird, probably a partridge.Diameter of stamp 7.0 cm. From road metal to northwest of the Areopagus (C 17); context of late 5th-early 4th century B.C. April 29, 1949. Relatively late tile. Inv. no. A 1486 E. ORNAMENTAL

E 1. Laconian covertile, fragment P1. 52 Pres. L. 28.4, pres. W. 18, H. of complete palmette 6 cm. One edge only preserved.Ash-gray clay coveredwith fine black glaze on outside. On the unglazed lower side, the impressionsof two different palmettes (E 1, E 2), one with single and one with double tendrils. The incomplete stamp partly covers the complete one. Complete stamp: H. 6.05, W. 6.5 cm.

From a disturbedcontext in area of Stoa of Zeus Eleutherios.August 1967. Inv. no. A 3622 E 2. See E 1: secondstamp

P1. 52

E 3. Laconiancovertile

P1. 52

Max. dim. 7.8 cm. Glazed outside. On the rough inside surface,part of a stamp with an elaboratepalmetteornament. From a pit in the valley between the Areopagus and the Hill of the Nymphs. A 18-19. Deposit A 1819:1. Cf. Hesperia 17, 1948, p. 193; 19, 1950, p. 337; 20, 1951, p. 253. Found togetherwith many ostrakaof the early 5th century,but the potteryruns down to the middle of the century.April 24, 1947. Same stamp as Agora E 4. Inv. no. A 1319 E 4. Laconianpan tile P1. 52 Max. dim. 8.0, Th. 1.3 cm. Dull red glaze on upper surface. Broken all around; only part of the stamp, a palmette like E 3, preservedon underside. Cf. A 1319 (E 3) from ostrakadeposit in section NN. From a well of the 5th-6th centuriesafter Christ at northwestcornerof Stoa of Attalos. Deposit P 7:4. April 20, 1951. Inv. no. A 2009 E 5. Laconian covertile

P1. 52

Max. dim. 5.8, Th. 1.7 cm. Dull black glaze on outer surface. Stampedon interior:palmette,only partlypreserved. From an early deposit to south of Tholos (G 14) along with early red-figuredand late black-figured sherds,probablyall pre-Persian. March 31, 1932. Inv. no. A 1113 E 6. Laconian covertile

P1. 53

Max. dim. 12.5 cm. Thin brown glaze on outside. Stampedon underside:eight-petaledrosette, from a defectivestamp (?). Great Drain, West Branch, clearing east side to southeast of Tholos (H 12). Context of early 5th century B.C., including ostraka of Hippokrates and Themistokles. Probably debris from destruction in 480 B.C. of Archaic predecessorof Tholos (cf. Hesperia, Suppl. IV, pp. 32-33). March 1, 1938. Inv. no. A 833

FURTHER STAMPED ROOF TILES E 7. Laconianpan tile

Fig. 2, P1. 53

Cornerfragmentof a curvedpan tile, in two joining pieces. Pres. L. 27.5, pres. W. 23.5 cm. Coarse,pinkish buff clay. Painted red on top surface and unglazed below. Stampedimpressionin shape of a scallop shell on the underside. From the destructiondebris of an early building (House D) to the south of the Tholos along with many roof tiles and pottery of the 5th-3rd centuries B.C. May 27, 1954. Same stamp as Thebes E 1. Inv. no. A 2485 E 8. Laconian pan tile

P1. 53

Fragment from pan tile, broken all around. Pres. L. 10, Th. 1.2 cm. Red glaze on one side. On unglazed side, circularstamp with a large relief anchor in the center; illegible smaller signs above cross arms. Diameter of stamp 4.8 cm. From an early deposit at the northwestfoot of the Areopagus along with many broken roof tiles and pottery of late 6th and early 5th centuries. July 6, 1949. Inv. no. A 1483 E 9. Laconian pan tile

P1. 53

Small piece, brokenall around,from a thin, nearly flat tile. Max. dim. 12, Th. ca. 1 cm. Gritty, yellowish buff clay. Top surface covered with thin black glaze wash. Unglazed underside:circularstamp with anchor. From an unfinished well in the valley to the northwest of the Areopagus in context of early 4th century B.C. Deposit D 17:9. Hesperia 20, 1951, pp. 179-180. May-June, 1949. Inv. no. A 1472 E 10. Laconiancovertile

P1. 53

Fragment. Pres. L. 10.2, pres. W. 11, Th. at edge 1.6 cm. Very coarse pale pink clay, glazed black on top surface. Circular stamp on undersidewith anchorin relief. From the gravelly fill of the West Branch of the Great Drain to south of Tholos (G 13); predominantly 2nd century B.C. but with some early material. May 2, 1934. Inv. no. A 434

319

E 11. Laconianpan tile P1. 53 Broken all around. Max. dim. 12 cm. Buff clay; black glaze on concaveside. On unglazed convex side, part of a circular stamp: wheel(?) in relief. Late Roman fill on northeastslope of Areopagus. June 10, 1938. Inv. no. A 1413 E 12. Laconiancovertile

P1. 54

Part of a convex cover tile with some of the edge preserved.Pres. L. 16.5 cm. Dull reddishglaze on top surfaceonly. On underside,a large circular stamp: sunken circle with crossbarsin high relief. Fragments of several others with similar stamps in Tin 702. Provenanceas for A 1319 (E 3). No wheel, but same stamp as E 13. Inv. no. A 1412 P1. 54 E 13. Laconianpan tile Brokenall around. Max. dim. 11.6 cm. Buff clay. The slightly concaveface is coveredwith red glaze wash. On the unglazed underside, part of a stamp: fourspokedwheel. Provenanceas for A 1319 (E 3). Cf. E 12 for motif. Inv. no. A 1538 E 14. Laconianpan tile Brokenall around. Max. dim. 9.6 cm.

P1. 54

An emblemof a wheel has been pressedinto the clay. Provenanceas for A 1319 (E 3). From an early depositto the west of the Areopagus(A 20) in a context of the early 5th centuryB.C. May 10, 1940. Stamp similar to E 12 and E 13 but with a sunken circularframe. Inv. no. A 2677 P1. 54 E 15. Laconianpan tile Fragment: 9 x 6, Th. 1.8 cm. Coarse red clay. Black glaze on the upper surface. Part of swastika-like stamp preserved on the unglazed underside. From the filling of a cisternon the northeastslope of the Areopagus (Deposit M 20:1). Context predominantlyHellenistic. February6-9, 1937.

320

RAINER C. S. FELSCH

Stamp of unknown type. Inv. no. A 978 AIGINA, SANCTUARY OF APHAIA

D 16. Laconiantile

A. INSCRIPTION A 1. Laconian tile with inscription stamp? Pres. L. 3.2, pres. H. 2.5 cm.

Fig. 3:b

Underside: rectangular stamp with lettering running right to left: Ero .... Possibly only the legend of a pictorial stamp. B. FIGURAL

Underside: same stamp, only back part of body preserved. D 17. Laconiantile L. 7.3, W. 4.8.

FIGURAL

D 8. Laconiantile Fig. 3:c Underside:circularstamp with striding lion. Stamp identicalto Aigina, Aphaia D 1.

Fig. 3:a

Underside: rectangular stamp, frontal gorgoneion. Only upper left-handcornerpreserved.Stamp:pres. H. 4.0, W. 2.9 cm.

WITH LEGEND

B 1. Laconian tile, brokenon all sides Fig. 3:d H. 5.9, L. 5.0 cm. Underside: rectangular stamp with lion striding right and head turned back. Between the legs, remains of an inscription:A(?) ... Stamp identical to Rhamnous B 1. D.

D 15. Laconiantile Underside: same stamp, only upper part of body preserved.

AIGINA, KOLONNA Accordingto the informationavailableto me, at least four more stamps have been found, among them apparently the first complete repetition of the "dancer stamp"Aigina, Kolonna D 1 and D 2, and Aigina, Vaghia D 1. In additionthere are a lion stamp, presumably a repetition of the stamp Rhamnous B 1 and Aigina, Aphaia B 1, and Corinthian tiles with hitherto unknown types of stamps, applied in a different fashion. These new finds will be publishedby the excavators. CORINTH

D 9. Laconiantile

D.

Underside:identical stamp, only back part of rump preserved.

D 1. Laconian pan tile, fragment Fig. 4:b, P1. 55 of corner. L. 8.55, W. 6.7, Th. of edge 2.8, of center 1.95 cm. Fine clay with some reddish clay particle inclusions, orange to reddish,firedto ringing hardness. Upper surfacered glaze, at edge somewhatblackish. Underside,beige slip of finely purified clay. On underside, rectangular stamp impression with roundedcorners:squatting sphinx; scene perhaps to be restoredas antitheticalpair of heraldic sphinxes. Pres. H. of stamp 3.8, pres. W. 4.2 cm. Findspot: 22.6.1962, Anaploga Well 1962:6, "well in grave 12",basket no. 35, lot 1231. Inv. no. FM 62 Publication of find complex: H. Robinson, "A Sanctuaryand Cemeteryin Western Corinth,"Hesperia 38, 1969, pp. 1-35.

D 10. Laconiantile Underside:identicalstamp, only paws preserved. D 11. Laconiantile Underside: rectangular stamp with coiled snake. Identicalto Aigina, Aphaia D 3. D 12. Laconiantile Underside:same stamp, fragmentary. D 13. Laconian tile Underside:same stamp, only cornerpreserved. D 14. Laconiantile Underside: rectangular stamp with hare running right. Stamp identicalto Aigina, Aphaia D 6.

FIGURAL

FURTHER STAMPED ROOF TILES E. ORNAMENTAL

E 1. Corinthiancovertile, Fig. 4:c, P1. 55 fragmentof corner L. 9.5, W. 10.55, Th. at vertex 3.3 cm. Yellow clay with black inclusionsand reddishbits of clay, core fired slightly reddish. Fine greenish yellow slip, polished. On upper surface square stamp impression: starshapedblossomwith four petals and five dots. Tile of 4th centuryB.C.? Findspot: 15.5.1934, South Basilica, Roman context. NEMEA, SANCTUARY OF ZEUS E. ORNAMENTAL

P1. 55 E 1. Corinthianantefix tile, fragment On either of the sloping upper facets, stamp impression of a thirteen-petaledrosette. Findspot:SectionJ 13. Inv. no. AT 91 Publication:Hesperia 49, 1980, p. 185, pls. 38:e, 39:b right. E 2. Corinthianantefix tile, fragment P1. 55 On upper face, as E 1, two rosetteimpressions,onehalf and two-thirds preservedrespectively.Identical to stamp E 1. Findspot:Section H 14. Inv. no. AT 82 Publication: Hesperia 49, 1980, pp. 185, 190, pl. 39:b left; 40:e. E 3. Corinthianantefix tile, fragment On upper face two stamps identicalto E 1 and E 2. Findspot:Section0 16. Inv. no. AT 34 Cited:Hesperia 49, 1980, p. 185, note 20. E 4. Corinthianpan tile; eaves tile?

321

E 6. Same Inv. no. AT 206 Cited:as E 4. E 7. Same Inv. no. AT 207 Cited:as E 4. E 8. Same Inv. no. AT 222 Cited:as E 4. E 9. Same(?) Findspot:SectionI/13,20-17/6,9, Early Christian context. Inv. no. AT 245 Cited:Hesperia 53, 1984, p. 176. E 10. Corinthiancovertile, fragment On upper face, "keyhole"stamp impression. Inv. no. AT 200 Cited:Hesperia 50, 1981, p. 52. E 11. Same Inv. no. AT 202 Cited:as E 9. E 12. Same Inv. no. AT 203 Cited:as E 9. E 13. Corinthianpan tile, fragment On upper face, "keyhole"stamp impression. Inv. no. AT 94 Published:Hesperia 49, 1980, p. 185, pl. 39:a. E 14. Corinthianpan tile, fragment On upper face, "keyhole"stamp impression. Findspot:SectionJ 13 Inv. no. AT 98 Cited:Hesperia 49, 1980, p. 185, note 21.

On upper face, identicalrosettestamp. Inv. no. AT 89 Cited:Hesperia 50, 1981, p. 52.

E 15. Same Inv. no. AT 99 Cited:as E 13.

E 5. Same Inv. no. AT 116 Cited:as E 4.

E 16. Same Inv. no. AT 100 Cited:as E 13.

P1. 55

322

RAINER C. S. FELSCH

E 17. Same Inv. no. AT 101 Cited:as E 13.

E 34. Same Inv. no. AT 224 Cited:as E 32.

E 18. Same Inv. no. AT 102 Cited:as E 13.

E 35. Same Inv. no. AT 225 Published:Hesperia 50, 1981, p. 52, pl. 15:d

E 19. Same Inv. no. AT 114 Cited:as E 13.

E 36. Same Inv. no. AT 226 Cited:as E 32.

E 20. Same Findspot:Section N 16 Inv. no. AT 27 Cited: as E 13.

E 37. Same Inv. no. AT 227 Cited:as E 32.

E 21. Same Findspot:Section0 16 Inv. no. AT 51 Cited: as E 13. E 22. Same Findspot:Section 0 15 Inv. no. AT 52 Cited: as E 13. E 23-E 28. Six further fragmentsof pan tiles. No data given Cited:Hesperia 50, 1981, p. 52. P1. 55 E 29. Corinthiancovertile, fragment On upper face, "teardrop"stamp on the verticaledge of the tile. Inv. no. AT 136 Cited: Hesperia 50, 1981, p. 52. E 30. Same Inv. no. AT 197 Cited: as E 29. E 31. Same Inv. no. AT 236 Published:Hesperia 50, 1981, p. 52, pl. 15:c. E 32. Corinthianpan tile, fragment On upper face, "S"-stamp Inv. no. AT 209 Cited:Hesperia 50, 1981, p. 52. E 33. Same Inv. no. AT 223 Cited: as E 32.

P1. 55

E 38. Same -Inv.no. AT 228 Cited:as E 32. ARGOS, TOWN D.

FIGURAL

D 1. Corinthianpan tile, fragmentof P1. 56 corner L. 25, W. 12, original H. of front 4.5 cm. Reddish clay with large red inclusions, top and underside with yellow-beige slip. On the slanting upper facet, rectangularstamp impression:within pictorial field a standingstag facing right. Stamp: L. 4.5, H. 4.7 cm. Tile of the late 4th centuryB.C.? Findspot:1953, sondagein the courtyardof Psirogiannis House: G. Roux, BCH 78, 1954, pp. 164167. Inv. no. C 9927 D 2. Corinthiancovertile, fragmentof P1. 56 corner L. 30, W. 9, original H. of front 6 cm. Reddish clay with large red inclusions. Top and underside with yellow-beige slip. On the slanting upper surface, rectangular stamp impressionof which only the lower half is preserved. Within the pictorial field, a standing stag facing right. Stamp identicalto D 1. Findspot:1955, excavationin the theater:G. Roux, BCH 80, 1956, pp.376-395. Inv. no. A 528

FURTHER STAMPED ROOF TILES PROVENANCE UNKNOWN D.

FIGURAL

D 3.Corinthiancovertile, fragmentof edge P1. 56 L. 15.1, H. 11.4, Th. 2.0 cm. Buff clay, hardly any inclusions,high mica content.Fired to ringing hardness.Upper face, thin beige slip. On clay of underside,oval stamp impression.Within pictorial field, an acrobatic symplegma group.

323

Stamp identicalto D 2. Representationblurredowing to vertical impressionof a straw and an oblique scratch.Stamp:H. 6.3, W. 4.6-4.7 cm. Findspot:unknown. Provenance:Acquiredbefore 1939 through Athenian antiques market. Dealer's mark: M/253 Dr. 500. Subsequentlyowned by V. Grace; now in collectionof AmericanSchoolof Classical Studies at Athens. Inv. no. ASA.72.

RAINER C. Silker Weiche 10 D-2057 Reinbek Federal Republic of Germany

S. FELSCH

PLATE 51

Kalapodi A 2 Kalapodi B 3

Kalapodi A 3

Kalapodi C 4

Kalapodi B 4

Kalapodi B 2

Kalapodi C 5

_

Kalapodi E 6

PLATE 52

R1

14

Agora D 2

Agora D 1

Agora El, E 2 Agora E 3

-

in

Agora E 5

PLATE 53

Agora E 6

Agora E 7

Agora E 8

Agora E 10

Agora E 9

Agora E li

PLATE 54

'S

Agora

I 12

13

Agora E 12

E 13

E 4Agora

Agora E 14

GyptkCtr E' I Agora E 15

Gptksr

PLATE 55

Nemea E I

Corinth D 1 Corinth E 1

-

A-

Nemea E 1, detail

Nemea E 2

4Nemea

2

E

PLATE 56

Argos D 1

Argos D 2

Eretria D 1

Provenanceunknown D 3

Lamia Museum: Geometricbronze horse from Anavra

DECORATED ARCHITECTURAL TERRACOTTAS FROM THE ATHENIAN ACROPOLIS

CATALOGUE OF EXHIBITION

CHRISTINA VLASSOPOULOU

DECORATED ARCHITECTURAL TERRACOTTAS FROM THE ATHENIAN ACROPOLIS

CATALOGUE OF EXHIBITION

AMERICAN SCHOOL OF CLASSICAL STUDIES IN ATHENS

Publication:ElectraAndreadi-

ChristinaVlassopoulou

Photographs:AristotelesAnagnostou

FOREWORD The exhibition of architecturalterracottasfrom the Athenian Acropolis coincides with the First InternationalConferenceon Archaic Greek ArchitecturalTerracottas,sponsoredby the American Schoolof Classical Studies (December2-5, 1988). The originalpublicationof this materialby Ernst Buschor,Die Tonddcherder Akropolis,Berlin 1929-1933, has never been supersededby a more recentpresentation.The exhibition,presentedby the First Ephorate of the Acropolis in the Weiler Building of the Center for Acropolis Studies, includesdecoratedelementsfromterracottaroofsof numerouslarge and small buildingswhich once stoodon the Acropolis,datingfromthe late 7th to the 5th centuryB.C. Severalof the earlierroofsbear witness to their destructionby fire at the hands of the Persiansin 480 B.C. The cleaningand conservation preparatoryto display undertakenby Nancy Lazarou, conservatorof ceramics,has revealed the bright colors which are still preservedon many of the pieces. The examples from the 5th century B.C., the acme of ancient Greek civilization,are made from the same fine clays and show the same lustrous black glaze as the Attic vases of the same period;they are painted by equally good hands and carrysimilar floral motifs. In sum, this display documentsthe early history of a traditionwhich continuedinto the NeoClassicalbuildings still remainingin Athens, Piraeus, and other cities in Greece. Evi TOULOUPA Ephor of the Acropolis

I would like to thank Dr. Nancy Winter for her help in the study and classificationof the Acropolis material and for her kindness in placing at my disposal an extract from Volume I of her forthcoming book, A Handbook of Greek ArchitecturalTerracottas.I would also like to thank the AmericanSchool of Classical Studies for publishing the English versionof the catalogue. C.V.

INTRODUCTION The earliest excavationson the Acropolis,which were carriedout by L. Ross, C. Hansen, and E. Schaubert(1835-1836) and by P. Kavvadiasand G. Kawerau (1837, 1885-1890), uncovered, among other splendid finds, numerous fragments of architecturalterracottas.These terracottas, which decoratedthe roofsof a numberof the buildingsin the great sanctuaryof Athena,were studied by E. Buschor in Die Tondacher der Akropolis, Berlin 1929-1933.

The Acropolisterracottasinclude antefixes,eavestiles, geison tiles, ridge tiles, rakingand lateral simas, and akroteria;so far, only decorativeelementsseem to have been preserved.They covera span from Early Archaicto Classical times and come from the construction,renovation,or simple repairs to the roofs of the buildings which adornedthe sanctuaryand servedits particularfunctions. The pieceswhich are now on displayhave been classifiedby material,techniqueof manufacture, and style into groups belongingto the following chronologicalperiodsin the building history of the citadel: I. The terracottasof the Early Archaic period (ca. 620-570/60 B.C.; 1-10), consisting only of antefixes and eaves tiles, belong to the roofs of three small buildings of which the foundationshave not been found. They were probably hipped and without pediments or simas, like the building shown in the Olive Tree Pediment of 570-550 B.C. Their style is strongly influenced by the Corinthianand possibly by other roofing systems. Brown clays, highly temp.eredand probablyimported, are their main technical feature. These roofs are contemporarywith the first monumental edificeson the Acropolis,the so-calledH-architecture,for which marbledecorationwas preferred. II. The reorganizationof the Panathenaic Festival in 566 B.C. by Peisistratos favored building activityin the sanctuary;new structuressuch as the poros Building C (ca. 550 B.C.) were added.Terracotta fragments 11-30 belong to several roofs with raking simas which carry painted decoration. Despite its dependenceon Corinthian types, this decorationdisplays a local characterwhich finds parallels in the polychromeporos architectureon the Acropolis. Clearly Attic is the style of the antefixesand akroteria,made of red Attic clays. Their high quality is comparableto that of contemporaryAttic pottery,ca. 550-530 B.C. To these years dates the fine terracottastatue of a goddess (30). Evidencefor the existence of an Attic roofing system comes also from other Athenian sites such as the Agora and the Academy.The buildings roofed by these tiles were perhaps "treasuries"to house votive offeringsor the "oikemata"mentionedby the inscriptions. III. In the years between the fall of tyranny and the destructionof the Acropolisby the Persians,the Old Propylonwas built (ca. 520-510 B.C.), and later the Old Temple of Athena Polias was renovated by the Peisistratids and adorned with marble decoration(ca. 500 B.C.). During the years 489 to 480 B.C., apsidal Building B and the terraceof the Older Parthenonwere added in the sanctuary. Contemporaryare five differentroofs with antefixes and floral akroteria(31-40). By the end of the century the Ionic style is predominantin some antefixeswith palmettes and gorgoneia.The gorgoneia are thought to belong to the roof of the Propylon.

In the layer of the Persian destruction,sima fragmentswith a new kind of profile (cymareversa) and light-on-darkdecorationhave been found;this sima (41-43), attributedto the Archaicsanctuary of Artemis Brauroneia,belongs to a series of simas which were fashionablethroughoutthe entire 5th century B.c. Tiles of similar types are known from other sites in Athens and Attica (the Agora and the Kerameikos;Rhamnousand Brauron)and the Athenian Stoa at Delphi as well. IV. Eleven terracottaroofs dating to the Early Classical period probablycome from the necessary repairsto buildingsburntby the Persians,such as the west roomof the Old Temple and Buildings D and E. From this material have been selected parts of four roofs which are characteristicof the evolutionof roof decoration.The new type of sima mentionedabove reaches its acme in 51-53 and 56, which are, respectively,the smallest and largest examples found on the Acropolis. The light-on-darkpainted decoration,enrichedwith lotuses, is executed in an archaizing style found also in contemporaryred-figuredpottery.To the end of the periodbelongtwo simas which are distinctivefor a more elaboratedand almost naturalisticdecorationharmoniouslyarrangedon their profiledface (62 and 63). Similar decorationis foundon the marblesima of the temple of Poseidonat Sounion. V. In addition to the entirely marble buildings of the Periklean program, there also stood on the Acropolis smaller edifices, such as the workshops of the Parthenon and the Erechtheion, the Chalkotheke,the House of the Arrephoroi,etc., which were probablycoveredwith terracottaroofs. The terracottasof a roof selectedfrom amongeight poorlypreservedones date to this period (64-67). The material consists of lateral simas painted with lyre-shaped volutes and palmettes which are flankedby beautiful lilies. The motifs are independentunits placed to either side of the lion's-head waterspouts.The designrecallsthe fine marbledecorationof the Erechtheionand the vase paintings of the so-calledRich Style of the last decadesof the 5th centuryB.C.

ABBREVIATIONS ATK = A. Akerstr6m,Die architektonischenTerrakottenKleinasiens, Lund 1966 Corinth IV, i = I. Thallon-Hill and L. S. King, Corinth, IV, i, Decorated Architectural Terracottas,Cambridge,Mass. 1929 = Corpus VasorumAntiquorum CVA GM = L. T. Shoe, Profiles of GreekMouldings, Cambridge,Mass. 1936 Huibner = G. Hubner, "Dachterrakottenim Magazin des Museums von Nauplia," in Tiryns. Forschungenund Berichte VIII, Mainz 1975, pp. 117-136 = C. Le Roy, Fouilles de Delphes, II, Les terres cuites architecturales,Paris 1967 Le Roy = E. Buschor, Die Tonddcherder Akropolis,Berlin 1929-1933 TdA Travlos = J. Travlos, Pictorial Dictionary of Ancient Athens, New York 1971 = T. Wiegand, Die archaische Porosarchitektur der Akropolis zu Athen, Wiegand Kassel/Leipzig 1904

Numbers prefixed by K refer to TdA."Colors of clay and slip are given the notations of the Munsell Soil Color Charts.

PLATE A

11

~~~~~~~~~19

I.

30

~~~~~~~~~~~~I

39

PLATE B

41

56

65

CATALOGUE

1. Acr. Mus. 10124 (K230). Fragment of an antefix of Corinthian type, decoratedwith a molded and painted three-leaf palmette and two volutes on a hexagonal plaque. Clay 1OYR6/4, slip lOYR 6/6, color lOR 4/4. H. 0.06, W. 0.16, L. 0.17 m. Date: 620-600 B.C. Bibliography: TdA II, pp. 29-32, figs. 39, 40, 42 (Type III). Cf. Le Roy, p. 34, nos. 2 and 9 (A.3, pl. 6); CorinthIV, i, fig. 1. 2. Acr. Mus. 10185 (K231). Fragment of an antefix of the same type as 1. Clay 5Y 7/3, slip 5Y 7/3, color 5YR 4/4. H. 0.07, W. 0.17, L. 0.115 m. Date: 620-600 B.C. Bibliography:TdA II, figs. 39, 41. 3. Acr. Mus. BHIep.327. Fragment of an antefix of the same type as 1. Clay lOYR 7/3-6/3, slip 10 YR 7/4, color lOR 4/3. H. 0.065, W. 0.14, L. 0.13 m. Date: 620-600 B.C. Bibliography:TdA II, pp. 29, 32. 4. Acr. Mus. 9570 (K125). Fragmentof an eaves tile decorated with a painted guilloche, of Corinthiantype. It is associatedwith antefix 3. Clay lOYR 6/4, slip 2.5Y 7/4, color IOR4/4. H. 0.05, W. 0.11, L. 0.08 m. Date: 620-600 B.C. Bibliography:TdA II, p. 7 (Type III). Cf. Le Roy, p. 38, serie 8, R.30, pl. 6.

5

6

5. Acr. Mus. 9667 (K241). Antefix with three molded peaks at the top. The plaque is decoratedwith a stamped design filled with red color, channels which follow the contourof the front and end in spirals on the lateral finials, and a palmette which fills the central finial. The shape and type of decorationmay be characteristicof the Argive system. Clay 2.5YR 6/4, slip 7.5YR 6/6, color lOR 4/6. H. 0.14, W. 0.14, L. 0.16 m. Date: 600-580 B.C. Bibliography:TdA II, pp. 26-27, figs. 35, 36 (Type I). The closest parallels are found in Delphi (Le Roy, pp. 64-65, A.41, pl. 19), Acrocorinth(C. K. Williams, II, "Demaratusand Early CorinthianRoofs,"in trTA?, Athens To,uosveds ,uv7),ujAvNucoAcaovKov-roAofv7os0, 1978 [pp. 345-350], pp. 348-349, pl. 155), Nemea (S. G. Miller, "Excavations at Nemea, 1979," Hesperia 49, 1980, p. 185, pls. 38:e and 38:a), and Mycenae (Hubner, pp. 119-120, fig. 26, pl. 64:4, no. 17260).

-~~~

6. Acr. Mus. 9668 (K246). Antefix of the same type as 5. It preservesthe bottom left edge of the cover tile which follows the line of the raised borderof the eaves tile. Clay 5YR 6/4-5/4, slip 5YR 5/4, color lOR 3/4 (red). H. 0.14, W. 0.19, L. 0.18 m. Date: 600-580 B.C. Bibliography:See 5 for the type.

7 7. Acr. Mus. 9659 (K121). Fragment of an eaves tile with part of an incised inscription on its undecorated front:AEH. It may belong with antefixes 5 and 6. Clay 7.5YR 6/4, slip lOYR 6/4, color lOR 3/2 (black). H. 0.04-0.06, W. 0.22, L. 0.165 m. Date: 600-580 B.C. Bibliography:TdA II, p. 6, figs. 1, 5 (Type I).

H. 0.27, L. 0.245, W. 0.28 m. Date: Ca. 570 B.C., on the basis of its resemblanceto an example from Kalapodi. Bibliography: TdA II, p. 7, figs. 6, 37, 38, pl. 1 (Type II); Travlos, p. 62, no. 72. Similar fragmentscome from the Athenian Agora (inv. nos. A815, A882, and A2580) and Brauron(inv. no. 53). 9. Acr. Mus. 9697 (K225). Fragment of an eaves tile of the same type as 8. Clay 7.5YR 6/4,2.5YR 5/4, slip 10YR 7/3,7.5YR 6/4, color 7.5R 4/2 (red). H. 0.05-0.06, W. 0.335 (restored0.67), L. 0.22 m. Bibliography:See 8. 10. Acr. Mus. 9704 (K258). Middle part of an eaves tile of the same type as 8. Clay 5YR 6/4, slip 5YR 5/4-5/6, color 1OR3/4 (red). Bibliography:see 8. 10 bis. Acr. Mus. 4938. Terracotta model of a hipped roof, possibly part of a votive model of a temple. One hipped end and part of one side are preserved.The roof consistsof flat pan tiles and sloping covertiles, the latter

8. Acr. Mus. 9708. Combination tile. The antefix consists of a molded five-leavedpalmette above two volutes, tied with a band, set off from the red backgroundof the plaque by narrow, red-lined bordersin relief. The front of the eaves tile is decoratedwith a painted guilloche and two palmetteswhich fill the area of the peaks. Clay 7.5YR 6/4, 2.5YR 4/4, and 5YR 6/3, slip 5YR 6/4, color 7.5R 4/2 (red).

8

lO bis

9

10

Date: Ca. 570 B.C. Bibliography:Wiegand, no. 19, p. 191, fig. 210.

10

bis

painted white. The edge of the roof carries antefixes topped by three finials and eaves tiles which peak beneath them; both elements are painted purple red. The closest parallels are found in antefixes 5 and 6 and combination tiles 8-10. L. 0.11, W. 0.095 m.

11. Acr. Mus. 118 (K13). Plate A. Part of a raking sima with a cavettoprofile, belonging to the right slope of the pediment. It has dowel holes at the top back for the attachmentof a cornerakroterion.The cavettois decorated with a guilloche and a tongue pattern,with recurvedtips of tongues on the top fascia. The decorationwas drawn by compassbeforefiring and painted in the dark-on-light style, and it imitates Corinthiantypes. Clay lOR 6/6, slip lOYR 6/4, color 5YR 4/4 (red). H. 0.18, W. 0.41 m. Date: 550 B.C., by comparisonwith Corinthian simas and eaves tiles. Bibliography: Wiegand, p. 184, fig. 193; TdA I, p. 10, fig. 10 (Type IV); GM, p. 131, pl. LXII:9; Travlos, p. 143, fig. 195. Cf. Le Roy, p. 43, S.5, pls. 6, 98. 12. Acr. Mus. 9467 (K12). Fragmentof a raking sima of the same type as 11, preservingthe tongue and guilloche patterns. Clay 1OYR7/4, slip lOYR 6/4, color 2.5YR 3/4. H. 0.102, W. 0.15, D. 0.105 m. Bibliography:See 11 (TdA I, p. 10, Type IV).

Bibliography: TdA I, pp. 6-8, figs. 2-8 (Type I); GM, p. 130, pl. LXII:1. Cf. Le Roy, pp. 48-49 and 86, note 1, LN.5, pl. 9. For the checkerboard,cf. Hubner, p. 121, pl. 65:1, Beil. 9:1, Abb. 3:b. 15. Acr. Mus. 9460 (K1). Fragment of a raking sima with a cavettoprofile, of the same type as 14. Clay 5Y 7/3, slip 5Y 6/4, color 1OR3/2 (red). H. 0.09, W. 0.08, D. 0.072 m. Date: 550-540 B.C. Bibliography:See 14. 16. Acr. Mus. 9571 (K124). Fragmentof an eaves tile of Corinthian type, decoratedwith a painted guilloche. It preservesa dowel hole. Clay 5Y 7/3, slip 5Y 7/3, color 5YR 3/1 (black). H. 0.043-0.055, W. 0.015, L. 0.07 m. Date: 550 B.C. Bibliography:TdA II, figs. 1, 8 (Type V). 17. Acr. Mus. 9461 (K16). Corner fragmentof a raking sima, preservinga small part of the waterspoutfrom the eaves side. On the face is a bichrome guilloche with creamborders,and on the soffit a red band. Clay 7.5YR 6/4, slip 2.5Y 7/4, colors 1OR 4/4 (red), 10YR 4/1 (black). H. 0.055, W. 0.055, D. 0.14 m. Date: 550-540 B.C. Bibliography:TdA I, pp. 16-17, fig. 17 (Type VII); Le Roy, p. 112 (datedca. 550-540).

13 13. Acr. Mus. 110 (K14). Right cornerof a raking sima with a floral akroterionattachedto the top. It may belong to the continuationof the sima onto the eaves. The profile is a flat cavettocarryinga tongue pattern and a fascia at the top with recurvedtips of tongues. The palmette has five moldedpetals, alternatelyred and black, over a thick double volute tied with a wide band. Clay 5Y 7/2, slip 2.5Y 8/4, colors 5YR 3/1 (black), 5YR 4/3 (red). H. 0.105 (sima), 0.172 m. (palmette). Date: 550 B.C. (by comparison with Corinthian simas). Bibliography:Wiegand, no. 189, p. 189, fig. 203; TdA I, p. 2, fig. 11 (Type V). 14. Acr. Mus. 93 (K2). Right corner of a raking sima with a deep cavetto profile and a recessed akroterionbase. The decoration consists of an upper band with tongues and a lower with a checkerboardpattern, painted free-handin the dark-on-lightstyle. Clay 2.5Y 5/4, 5Y 7/3, slip 2.5Y 7/4, color lOR 3/2 (red). H. 0.142, W. 0.125 (pediment),0.154 m. (eaves). Date: 550-540 B.C.

18. Acr. Mus. 9462 (K17). Corner of a raking sima preserving part of the akroterionbase and the upper part of a lion's head. The head has a handmademane of incised wavy locks painted black. Clay 7.5YR 6/4, slip 2.5Y 7/4, color 2.5Y N3/, 10YR 4/1 (black). H. 0.088, W. 0.196, D. 0.04-0.11 m. Date: 550-540 B.C. Bibliography:TdA I, pp. 17-18, fig. 18 (Type VIII) andII,p. 10;LeRoy,p. 112. 19. Acr. Mus. 10097 (K284). Plate A. Palmette antefix of Attic type with seven leaves in alternating red and black above a thick black volute which forms a peak at the center and a red-eyed spiral to each side. It may belong to the same roof as 17 and 18. Clay 7.5YR 6/4, 5YR 6/3, slip 2.5Y 7/4, colors 7.5R 4/4 (red), 5YR 3/1 (black). H. 0.17, W. 0.23, L. 0.15 m. Date: 550-540 B.C. Bibliography: TdA II, p. 40, fig. 52, pl. 4 (Type VIII); Travlos, p. 62, no. 77. 20a-b. Acr. Mus. 9713 (K295) and 9712 (K296). Two fragments of a molded antefix of Corinthian type con-

'I

sisting of a seven-leavedpalmette above two volutes, the right spiral of which is preserved. Clay 5Y 8/3-7/3, slip 1OYR6/3-2.5Y 7/4, colors 2.5Y N3/ (black), 7.5YR 4/4 (red). L. 0.12, W. 0.25 m. Date: 550-540 B.C. Bibliography:TdA II, p. 35, fig. 46, pl. 2 (Type V). Cf. Le Roy, nos. A.23 (pl. 7:5), A.27, and A.28 (pl. 16:2, 3); CorinthIV, i, p. 12, fig. 2. 21. Acr. Mus. 9463 (K3). Part of a raking sima with a cavetto profile, decorated with tongue pattern and guilloche. Clay 2.5Y 7/4, slip 2.5Y 7/4, color lOR 3/2. H. 0.12, W. 0.08, D. 0.065 m.

r:'4

Date: 550-540 B.C. Bibliography:TdA II, p. 8, fig. 1 (Type II). 22. Acr. Mus. 9464 (KS). Fragment of a raking sima with a cavetto profile, decoratedwith tongue pattern on the cavettoand recurvedtips of tongues on the top fascia. Clay 2.SY 7/3, slip 2.5Y 7/3, color 2.5Y 3/2. H. 0.19, W. 0.0132, D. 0.04-0.08 m. Date: 550-540 B.C.

26

Bibliography: TdA I, p. 8, fig. 1 (Types II, III); GM, p. 131, pl. LXIII: 11. Similar to Roof 10 at Delphi (550-540 B.C.): Le Roy, p. 45, nos. S.12, S.196 (pls. 7, 98), S. 15, pl. 12, and LN.5 (see 14), pl. 9. 23. Acr. Mus. 9465 (K9). Fragmentof a raking sima;the painted double guilloche has a central row of five-leaved palmettespointing right. Clay 2.5Y 7/4, slip 2.5Y 7/4, color IOR3/2. H. 0.068, W. 0.09, D. 0.096 m.

Date:550-540B.C. Bibliography: TdA I, p. 8, fig. 9 (Types II, III); GM, p. 131, pl. LXII:11. Cf. Le Roy, p. 45, nos. S.12 (pls. 7, 98) andS.196, S.15 (pl. 12), and LN.5, pl. 9. 24. Acr. Mus. 9468 (Kl 5). Right cornerof a rakingsima with a cavetto profile, preserving part of a recessed square base for the akroterion,and remains of the mane of a lion's head along the side. It is paintedwith tongues, a herringbonedesign, and a guilloche (probablydouble). Clay: 2.5YR 5/4, slip 2.5Y 7/2-6/2, colors 7.5R N3/ (black), 2.5Y 7/4 (yellow). H. 0.155, W. 0.15, D. 0.07-0.085 m. Date: 540 B.C. Bibliography:TdA I, pp. 12-15, figs. 12, 13 (Type VI). For the herringbonemotif, cf. Le Roy, nos. S.234, pl. 16 and S.6, pl. 14. 25. Acr. Mus. 9678 (K232). Fragment of an antefix of Attic type, with a molded and painted floral design on a pentagonal plaque. A five-leaved palmette rises above two intertwinedhorizontalvolutes ending in spirals and tied with bands at the sides. Between the spirals are small palmettes.

27

Clay 10YR 5/3 core, 5YR 5/4 surface, slip 10YR 7/4, color 1OR4/6. H. 0.117, W. 0.195, L. 0.16 m. Date: 540 B.C. Bibliography:TdA II, pp. 32-35, figs. 43-45 (Type IV); Travlos, p. 62, no. 73. 26. Acr. Mus. 9672 (K236). Fragment of an antefix of the same type as 25. Clay 10YR 5/3, 5YR 5/4, slip 10YR 7/4, color 1OR4/6. H. 0.158, W. 0.11, L. 0.08 m. Date: 540 B.C. Bibliography:See 25. 27. Acr. Mus. 9676 (K239). Fragment of an antefix of the same type as 25. On an inner corner where the plaque joins the covertile is a nail hole for attachmentto the eaves tile. Clay 7.5YR 6/4, slip 10YR 7/3, color 1OR4/6. H. 0.12, W. 0.11, L. 0.145 m. Date: 540 B.C. Bibliography:See 25. 28. Acr. Mus. 9684 (K276). Fragment of an antefix of Attic type. Remains of a palmette with seven leaves, alternatingred and black, over a thick black volute. Clay 5YR 6/4, 2.5Y 6/4, slip 2.5Y 7/4, colors 10YR 3/1 (black), 1OR3/3 (red). H. 0.105, W. 0.175, D. 0.105 m. Date: 540-530 B.C. Bibliography:TdA II, pp. 36-37, figs. 47, 48 (Type VI); Travlos, p. 62, no. 75. Similar examples from the Academy:Travlos, p. 46, fig. 54.

28

29 29. Acr. Mus. 9578 (K128). Fragment of an eaves tile decoratedwith a bichrome guilloche on a black background.Red bands paintedon the upper and bottomsurfaces. It probablybelongs to antefix 28. Clay 5YR 6/4, slip 2.5Y 7/4, colors lOR 4/3 (red), lOYR 3/1 (black). H. 0.13, L. 0.15, D. 0.037-0.056 m. Date: 540-530 B.C. Bibliography:TdA II, p. 10, figs. 2 and 11 (Type IX). Similar examples are found in the Athenian Agora and at Brauron(inv. no. 89).

A..

30~~~~.

30. Acr. Mus. 30. Plate A. Terracottastatue of a seated goddess.Preservedare the body below the knees, the intact left foot, and a thin, square plinth. The body is hollow, obviouslymade by hand of successivelayers of clay. The goddesswears a long, light red chiton with a black paryphe decorated with a simple red maeander and a foldedhem. The dark red himation falls in modeled,zigzag folds and is decorated with a black border which curves over the ankle. The foot is solid and has skilfully formed anatomicaldetails. The piece could be an akroterion or a votive statue. Clay 2.5YR 4/4, slip 2.5Y 7/4, colors 7.5YR N3/ (black), lOR 4/2 (red). H. 0.27; plinth 0.30 x 0.30 x 0.01-0.02 m. Date: 530-525 B.C. Bibliography:W. Deonna, Les statuesde terre-cuite en Grece, Paris 1906, pp. 48-51; N. Bookidis, "Attic Terracotta Sculpture, Acropolis 30," Hesperia Suppl. XX, Studies in Athenian Architecture,Sculpture, and Topography,Princeton 1982, pp. 1-8, pl. 1.

31a-b. Acr. Mus. 4923 a, b. Two fragmentsof a horse's head preservingthe front part of the face and the right eye. The bridleis renderedas red bandsheld in place by a round buckle. The flat white eyeball is outlined by a black eyelid; the pupil is black. The head is hollow and handmade and may belong to an akroterionor a votive statue. Clay lOYR 7/3, slip 2.5Y 7/4, colors 2.5Y 3/N3 (black), lOR 3/1 (red). H. 0.17, L. 0.187 m. Date: Secondhalf of 6th centuryB.C. Bibliography: Deonna (30 above), p. 34, no. 10; E. D. Van Buren, Greek Fictile Revetments in the ArchaicPeriod, London 1926, p. 178, no. 4. /

S

. ;r..:,y.'-

32. Acr. Mus. 10102 (K283). Antefix with a sevenleavedpalmette (broken)abovea thick blackvolute and a pendent five-leavedpalmette. The type is similar to 19, but as the shape of the palmette is differentand the proportions of the cover tile are lighter as well, it may be a later replacement. Clay 7.5YR 6/4, 5YR 6/3, slip 2.5YR 7/4, colors 7.5R 4/4 (red), 5YR 3/1 (black). H. 0.17, W. 0.205, L. 0.345 m. Date: 510-500 B.C. Bibliography: TdA II, p. 1, fig. 52, pl. 4 (Type VIII). Travlos, p. 62, no. 77. (Cf. an example from vase painting: CVA, Netherlands 5 [Leiden 3], pp. 13-14, pl. 113.)

3

31

32

33. Acr. Mus. 9574 (Kl 53). Fragment of a geison tile decorated with a painted bichrome maeander. It preserves part of the flange and the anathyrosison the surface of the left side. Clay 7.5YR 6/4, slip 5Y 7/4, colors 10YR 3/1 (black), 1OR4/2 (red). H. 0.056-0.065, W. 0.09, L. 0.124 m. Date: 510-500 B.C. Bibliography:TdA II, pp. 12-13, figs. 15-16 (Type XII). Cf. E. Curtius and F. Adler, Olympia, II, Die Baudenkmdler,Berlin 1892, p. 194, pl. 119:5;H. Payne, Necrocorinthia,Oxford 1931, fig. 112:b. 34. Acr. Mus. 9576 (K154). Fragment of a geison tile decoratedwith a painted bichrome maeander.The face forms an obtuse angle at the center. Clay 1OYR 6/3, slip 2.5Y 7/4, colors 5YR 3/1 (black), 5YR 4/2 (red). H. 0.06, W. 0.136, L. 0.075 m. Date: 510-500 B.C. Bibliography:See 33. 35. Acr. Mus. 85 (K294). Antefix of Ionic type preserving the lower half of a molded gorgoneion.The features of the face are sensitively modeled. The ear-caps, the beardedsnake coiled at the left corner,and the type of the face indicate an Eastern-Ionic influence. Traces of red and blue. Clay 5YR 6/4, slip 5Y 8/3, colors 7.5R 3/6 (red), 5B 6/1 (blue). H. 0.12, W. 0.24, D. 0.12 m.

35a

Date: 510-500 B.C. Bibliography:TdA II, pp. 40-41, pl. 5 (Type IX); J. Floren, Studien zur Typologiedes Gorgoneion,Munster 1977, p. 61, pl. 6:3; J. Belson, The Gorgoneionin Greek Architecture, diss. Bryn Mawr College, 1981, pp. 5-6. For the Ionic elements,see ATK II, pls. 30:3 and 53:2. 35a. Acr. Mus. 78 (K292). Antefix of the same type as 35. It preservesthe face of the Gorgon almost intact and the beginning of the cover tile as well. The hair consists of spiral curls painted black, and the lips and the tongue in the open mouth are painted red. The forehead, the outer edges of the eyes, and the nose are wrinkled. Clay 5YR 6/4, slip 5Y 8/3, colors lOR 4/3 (red), lOYR 2/1 (black). H. 0.195, W. 0.20, L. 0.16 m. Date: 510-500 B.C. Bibliography:See 35 and S. Casson and D. Brooke, Catalogueof the AcropolisMuseum, Cambridge1921, II, pp. 289-290. 35b. Acr. Mus. 79 (K293). Antefix of the same type as 35 and 35a. It preservesthe face of the Gorgon almost intact and the beginningof the covertile. Clay 5YR 6/4, slip 5Y 8/3. H. 0.215, W. 0.20, L. 0.15 m. Date: 510-500 B.C. Bibliography:See 35.

35b

36. Acr. Mus. 9714 (K298). Fragment of a molded palmette set on a semicircularplaque which belongs to an antefix of Laconiantype. Clay 7.5YR 6/4. H. 0.08, W. 0.095 m. Date: 500 B.C. Bibliography: TdA II, pp. 37-38, figs. 49-51, including restoration(Type VII); Travlos, p. 62, no. 76. 37. Acr. Mus. 9577 (K152). Fragment of a geison tile decoratedwith a painted bichrome maeander against a black background. It may belong with antefixes 38 and 39. Clay 5YR 5/4, slip lOYR 7/4, colors lOYR 3/1 (black), 1OR4/2 (red). H. 0.053, W. 0.187, L. 0.095 m. Date: 500 B.C. Bibliography:TdA II, p. 12, fig. 14 (Type XI). 38. Acr. Mus. 104 (K302). Antefixof Attic-Ionictype decoratedwith a tall moldedpalmette.Seventhick leaves,alternately red and black, fan out from a diamond-shaped heart;below, two voluteswith red-eyeddoublespiralsare tied with a red band from which hangs a red leaf. The beginningof the pentagonalcovertile is also preserved.

Clay 5YR 5/4, slip lOYR 7/4, colors lOYR 2/1 (black), 5R 3/4 (red). H. 0.23, W. 0.186, L. 0.12 m. Date: 500-480 B.C. Bibliography: TdA II, pp. 43-44, fig. 57, pl. 6 (Type XI). Similar painted design occurs on ridge palmettes at Brauron(inv. nos. 54 and 58, from the Temple of Artemis). For the Ionic style, see ATK II, p. 99, figs. 30-31 (antefix from Samos, dated 525-500 B.C.). 39. Acr. Mus. 10253 (K301). Plate A. Antefix of the same type as 38. Clay lOYR 7/4, slip 2.5Y 7/4, colors lOYR 2/1 (black), 5R 3/3 (red). H. 0.23, W. 0.186, L. 0.08 m. Date: 500-480 B.C. Bibliography:See 38. 40a-b. Acr. Mus. 10037 (K300) and 10093 (K299). Two fragments of a corner floral akroterion of AtticIonic type. It preserves the bottom central part of two volutestied with a band and parts of the palmette.Under the band is a three-leavedpendent palmette. The akroterion is painted in the light-on-darkstyle.

38 39

_

_

377

ai

7

H. 0.19, W. 0.62, L. 0.565, D. 0.07 (flange 0.06) m. Date: Ca. 480 B.C. Bibliography:Wiegand, p. 185, fig. 195, no. 109; TdA I, p. 22, fig. 23 (Type X, Roof 10).

Clay 5YR 5/6, lOYR 7/4, color 1OYR3/1 (black). H. 0.11 m. Date: 490-480 B.C. Bibliography: TdA II, pp. 41-43, figs. 54-56, including restoration (Type X). For the Ionic style see ATK II, p. 99, figs. 30, 31 (antefix from Samos dated 525-500 B.C.).

42. Acr. Mus. 76 (K47). Part of the same sima as 41, from the right slope of the pediment. Clay lOYR 7/4, slip 1OYR7/4, colors 1OYR3/1 (black), 7.5R 5/2-4/2 (red).

41. Acr. Mus. 9482 (K46). Plate B. Part of a light-ondark raking sima with a cyma reversa profile, from the right slope of the pediment. The top fascia has an olive branch with pairs of leaves pointing left. The cymation has a double chain of palmettes, alternatelyfree and inscribedin volutes. The volutes form chains at the base of the palmettes. Additional ornaments such as small palmettes, "eyes", and lozenges fill the black-glazed background. Clay lOYR 7/3, lOYR 7/3-6/3, colors 5YR /1 (black), lOR 5/4, lOR 3/4 (red).

H. 0.19, W. 0.60, L. 0.23, D. 0.08 m.

Date: Ca. 480 B.C. Bibliography: Wiegand, p. 186, fig. 196; TdA I, pp. 22, 72, 75 (Type X, Roof 10). For the motif cf. CVA, Deutschland 13 (Mannheim), pl. 32:5 (dated 480470 B.C.); CVA, Great Britain 7 (British Museum 5), pl. 75; Athens, Nat. Mus., Acr. 703 and Acr. 1518a (Makron, 480-470 B.C.); CVA, Great Britain 8 (British Museum 6), pl. 83 (480-460 B.C.).

41

V.tw~~~~~~~~~~~~Xl

42

46 43

43. Acr. Mus. 113 (K62). Part of a light-on-darkraking sima from the right slope of the pediment. Decoration similar to that of 41 and 42, but different in details, e.g. the shape of the palmettes and the absenceof "eyes"and lozenges. Clay 1OYR6/4, slip lOYR 7/4, colors 1OYR2/1 (black), 5R 4/3 (red). L. 0.115, H. 0.19, W. 0.225, D. 0.08 m.

Date: Ca. 480 B.C. Bibliography:TdA I, pp. 23-24, fig. 21 (Type XI). 44. Acr. Mus. (K156). Fragment of an eaves tile with light-on-darkdecoration:an olive branch with leaves in groups of three and pairs of fruits on delicate stems, pointing left; red bands on the upper and bottom surfaces. It preservesthe whole face and two blind holes for the attachmentof the antefixes.

47 45

48 44

Clay 2.5Y 7/4, slip 2.5Y 7/4, colors 2.5Y N3/ (black), 5R 3/2 (red). W. 0.66, L. 0.16, H. 0.062 m. Date: Ca. 480 B.C. Bibliography:TdA II, pp. 14-15 (Type XIII). Cf. the decorationin red-figure vase painting (e.g. Athens, National Museum 18602, pyxis).

48. Acr. Mus. 103 (K386). Light-on-darkantefix of the same type as 47. It preserves the base with its composition of volutes and small palmettes. Clay 5YR 5/4, slip 1OYR 7/4, colors 1OYR 3/1 (black), lOR 3/6 (red). H. 0.25, W. 0.16 m. Date: Ca. 480 B.C. Bibliography:See 47.

45. Acr. Mus. 9615 (K183). Cornerfragmentof an eaves tile of the same type as 44. It preservestwo holes for attachmentto the roof and the antefix. On the right side is anathyrosismade after firing. Clay 5YR 6/3, slip 2.5Y 7/4, colors 2.5YR /0 (black), 5R 4/4 (red). W. 0.15-0.18, L. 0.21, H. 0.06-0.075 m. Date: Ca. 480 B.C. Bibliography:TdA II, pp. 14-15 (Type XIII). 46. Acr. Mus. 106 (K413). Light-on-dark antefix decorated with an eleven-leaved palmette rising above two volutes which form three pairs of spirals. Three small, shell-shaped palmettes are in the center of the palmette and at either side of the volutes. A five-leaved pendent palmette is under the upper pair of spirals. This piece probablybelongs to sima 43. Clay 10YR 6/4, slip 10YR 7/4, colors 10YR 2/1 (black), 5R 4/3 (red). H. 0.26, W. 0.17, D. 0.095 m. Date: Ca. 480 B.C. Bibliography:TdA II, pp. 45-46, figs. 58, 59, pl. 7 (Type XII). Cf. Le Roy, pp. 139-140, A.94, A.182, pl. 52 (from the Athenian Stoa, Delphi, 480-477 B.C.) and pl. 125 (plan); G. Hubner, "Dachterrakottenaus dem Kerameikos von Athen," Athenische Mitteilungen 88, 1973 (pp. 67-143), p. 73, figs. 1, 2 (dated 470 B.C.). 47. Acr. Mus. 101 (K389). Light-on-darkantefix similar to 46. The palmette has a shell-shaped heart. Small, seven-leavedpalmettes are placed between and to either side of the volutes. At the base, two small, thin volutes grow from the lower spirals. The piece preservesa dowel hole for attachmentto the eaves tile. Clay 7.5YR 7/4-6/4, slip 10YR 7/4, colors 10YR 3/1 (black), 1OR3/6 (red). H. 0.25, W. 0.19 m. Date: Ca. 480 B.C. Bibliography:TdA II, pp. 45-46, figs. 58, 59, pl. 7 (Type XII). Cf. Hubner (46 above), p. 90, fig. 56 (Brauron). The shape of the plaque and the decorationin part recall the antefixes from Rhamnous (B. Petrakos, <
49 49. Acr. Mus. 105 (K447). Ridge tile with light-on-dark decoration.Half the covertile is preserved,as is the ridge palmette,which is paintedon both faces. The top surface of the covertile is decoratedwith pairs of opposedvolutes against a triangular black backgroundto either side of the palmette.The piece is probablyassociatedwith antefixes 47 and 48. Clay 10YR 8/4-7/4, slip 5Y 7/4, colors 5Y 2.5/1 (black), 10YR 5/6 (red). H. 0.32, W. 0.20, L. 0.30 m. Date: Ca. 480 B.C. Bibliography:TdA II, pp. 63-65, figs. 87-89 (Type 1I (33 above), p. 197, fig. 17. II). Cf. Olympia

50

50. Acr. Mus. 10233 (K448). Ridge tile of the same type as 49, preservingonly the cover tile, which is decorated with a volute patternon the top surface. Clay 7.5YR 6/4, slip 2.5Y 7/4, colors 5Y 2.5/1 (black). H. 0.19, L. 0.446, W. 0.19 m. Bibliography: TdA II, pp. 63-65, figs. 88-89 (Type II).

Ap~

~ ~ ~

~~J

55~~~~5 xxvi

51. Acr. Mus. 9539 (K67). Cornerof a raking sima with a lion's-headwaterspout,from the left slope of the pediment. Light-on-darkdecorationof an olive branchon the top fascia and a lotus-and-palmettechain on the cymation (cyma reversa). The palmettes are inscribed in volutes which form spirals at their base. The lion's head preservespart of the mane with thin locks incisedin clusters. This sima is the smallest found on the Acropolis.

54. Acr. Mus. 9649 (K206). Right cornerof an eaves tile with light-on-darkdecoration:an olive branchwith pairs of leaves pointing right and a fruit at the base of each pair. Red bands on the top and the bottomsurfaces.The tile preservesa socket with a dowel hole for the attachment of the antefix. Clay 10YR 7/3, slip 2.5Y 7/4, 5Y 7/3, colors 7.5R 3/2 (red), 5YR 3/1 (black). L. 0.25, W. 0.31, D. 0.04-0.05m. Date: Ca. 470 B.C. Bibliography:TdA II, pp. 15-16, figs. 19-21, pl. 8 (Type XIV); Hubner (46 above), p. 71, pl. 57:2. For the ornament,cf. CVA, Italia 37 (Ferrara 1) 8; CVA, Great Britain 4 (British Museum 3), 44, 5a and Ic, pl. 89 (Great Britain 364).

Clay 5YR 5/4, 2.5Y 7/4, slip 10YR 7/4, colors 10YR 3/1 (black), 7.5R 5/2 (red). H. 0.105, W. 0.15-0.40 m. Date: Ca. 470 B.C. Bibliography:TdA I, p. 24, figs. 26, 27, pl. 2 (Type XII). For the design, cf. Athens, Nat. Mus. 9683, pelike by the Pan Painter (J. D. Beazley, Attic Red-figure Vase Painting, 2nd ed., Oxford 1963, p. 550, no. 1; 470 B.C.). 52. Acr. Mus. 9552 (K68). Part of the same raking sima as 51, also fromthe left side of the pediment.An incisedA on the bottom surface indicates the position of the piece on the roof. Clay 5YR 5/4, slip 2.5Y 7/4, colors 10YR 3/1 (black), 7.5R 3.2 (red). H. 0.015, W. 0.41, L. 0.38 m. Date: Ca. 470 B.C. Bibliography: TdA I, p. 28, fig. 29, pl. 2 (Type XII).

55. Acr. Mus. 115 (K435). Antefix with a light-on-dark palmette above two volutes which form three pairs of spirals. Small palmettes with a central pointed leaf are placedbetweenthe spirals. The piece belongs, with eaves tile 54, to the same roof as sima 51-53. Clay 10YR 7/4, slip 10YR 7/4, colors 10YR 3/1 (black), 7.5R 3/2 (red). H. 0.15, L. 0.22, W. 0.113 m. Date: Ca. 470 B.C. Bibliography: TdA II, p. 47, fig. 60, pl. 8 (Type XIII); cf. Bulletin de correspondancehelle'nique 103, 1979, Chron. p. 565, fig. 96 (Olympia).

53. Acr. Mus. 9532 (K71). Corner of the same raking sima as 51 and 52, but from the right slope of the pediment. It preservesthe remainsof a lion's head and a rectangular channel for the waterspouton the top of the tile backer. Clay 10YR 7/4-7/6, slip 10YR 7/4, colors 10YR 3/1 (black), 7.5R 5/2 (red). H. 0.105, W. 0.30, L. 0.16 m. Date: Ca. 470 B.C. Bibliography:TdA I, p. 25, pl. 2 (Type XII).

XXVil~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

52

56 56. Acr. Mus. 75 (K76). Plate B. Tile of a light-on-dark raking sima with a cyma reversa profile, preservingthe face almost intact. The decoration consists of an olive branch pointing right on the top fascia and a double lotus-and-palmettechain on the cymation. The palmettes have a shell-shaped heart. The lotuses show a pointed petal between two recurved leaves against a red backgroundwith tonguedborder.The palmettesare inscribed in volutes which form chains at their base. The back is paintedred and has an incised F indicatingthe positionof the piece on the roof. Clay 5YR 5/3, slip lOYR 7/4, colors 1OYR 3/1 (black), 5R 4/2 (red). H. 0.165, W. 0.673, L. 0.244 m. Date: Ca. 470-460 B.C. Bibliography: TdA I, pp. 27-28, figs. 30 and 34, pl. 3 (Type XIII). For the ornament,cf. sima 51-53. 57. Acr. Mus. 4931 (K117). The upper part of a handmade lion's head. Its mane consistsof two ranksof pointed, wavy lockswith inciseddetails. It probablybelongsto the same roof as sima 56. Clay 2.5Y 7/3, slip lOYR 7/4. H. 0.165, W. 0.24 m. Date: Ca. 470-460 B.C. Bibliography:TdA I, p. 51, fig. 61; cf. Hubner (46 above), p. 138, fig. 14, pl. 68:1 (S.Z255, Kat. E12); F. Willemsen, Olympische Forschungen IV, Berlin 1959, p. 33, pl. 37 (Zeustempel 72, 472-456 B.C.); M. Mertens-Horn, Die Lowenkopf-Wasserspeier des

griechischen Westensim 6. und 5. Jahrhundertsv. Chr. (Romische Mitteilungen, Erganzungsheft 28), Mainz 1988, p. 39. 58. Acr. Mus. (K216). Fragment of an eaves tile with light-on-dark decoration:a double guilloche of red and yellow strands, having a central row of pointed leaves with triangularhearts, pointing right. On the top surface is a red band, and on the soffit, an olive branch. Clay 7.5YR 6/4, slip 10YR 7/4, colors 7.5R 3/2 (red), 5YR 3/1 (black). H. 0.075, L. 0.135, W. 0.345 m. Date: Ca. 470-460 B.C. Bibliography: TdA II, p. 17, figs. 21, 24, 25 (Type XVII). 59. Acr. Mus. 10158 (K381). Antefix with light-on-dark decoration.A nine-leaved, semicircularpalmette with a shell-shaped heart rises above two recurved volutes which emerge from a pendent lotus. The volutes are tied with a red band. Four red eyes with white contours fill the spirals to either side of the lotus. Clay 5YR 5/4, slip 2.5Y 8/4, colors7.5R 3/4 (red), 7.5YR 3/2 (black). H. 0.17, W. 0.18, L. 0.13 m. Date: 470-460 B.C. Bibliography:TdA II, p. 49, figs. 61-63, pl. 9 (Type XIV). The closest parallel is found at Brauron (inv. no. 12). The type derives from the Archaic Corinthian light-on-darklotus-and-palmetteantefix (cf. CorinthIV, i, pI. 1).

60. Acr. Mus. 10159 (K382). Antefix of the same type as 59, preservingonly the base with the lotus. Clay 5YR 5/4, slip 2.5Y 8/4, colors7.5R 3/4 (red), 7.5YR 3/2 (black). H. 0.11, W. 0.18, L. 0.13 m. Date: 470-460 B.C. Bibliography:See 59. 61. Acr. Mus. 9656 (K218). Fragmentof a light-on-dark eaves tile decoratedwith a double guilloche; its central

59

60

^

'~~~~~1

row of shell-shapedpalmettespoints right. Clay 7.5YR 6/4, slip 1OYR7/4, colors 2.5YR 3/4 (red), 5YR 3/1 (black). H. 0.075, W. 0.235, L. 0.105 m. Date: 470-460 B.C. Bibliography:TdA II, p. 18, fig. 21, pi. 10 (Type XIX). Cf. the eaves tiles of the Tholos in the Athenian Agora:J. M. Camp, The AthenianAgora,London 1986, p. 77, fig. 52 (dated470-460 B.C.).

62. Acr. Mus. 9534 (K85). Left-hand corner of a lighton-dark raking sima preserving part of the lion's head from the eaves side. The top fascia carries a bead-andreel pattern. The cymation (cyma reversa) is decorated with a compositionof palmettesand lotuses in pairs, separated by lyre-shapedvolutes. The right soffit has beadand-reel and lesbian cymationpatterns.Incised e on the back. Clay 7.5YR 6/4, slip lOYR 7/4, colors lOR 3/4 (red), 7.5YR 3/2 (black). H. 0.15, W. 0.37, L. 0.15 m. Date: 470 B.C. Bibliography:TdA I, pp. 35-36, figs. 38-40 and 42, pl. 6 (Type XVI); Hubner (46 above), p. 110, pl. 57:4.

by lyre-shapedvolutes. Small palmettesfill the spacesbetween the volutes. Clay 10YR 7/3, slip 5Y 8/3, colors 10YR 4/3 (black), 1OR4/4 (red). H. 0.21, W. 0.38, D. 0.155 m. Date: Ca. 450 B.C. Bibliography:TdA I, p. 36, figs. 41-43, pl. 7 (Type XVII). Cf. Hubner (46 above), pp. 111-112, note 146, pl. 66:1 (Sounion sima). 64. Acr. Mus. 9644 (K95). Part of a light-on-darkraking sima from the right slope of the pediment. It preserves four spirals, two small palmettes,and two lilies growing from the spirals. On the soffit is a bead-and-reelpattern against a red band. Clay 1OR 7/3, slip 5Y 8/3, colors 10YR 4/3 (black), 1OR4/4 (red). H. 0.15, W. 0.10 m.

63. Acr. Mus. 9538 (K86). Part of a light-on-darkraking sima decoratedwith olive branchesrunning to the right on the top fascia and the soffit. The cymation (cyma reversa) has a compositionof alternating pairs of opposed palmettes and lotuses;the paired elements are separated

Date:420-405 B.C. Bibliography: TdA I, p. 41, fig. 46, pl. 9 (Type XX); Hubner (46 above), pp. 116, 119.

i,&~~~~~~~~~~~~~~~~w

X1. ..X

63

V

Clay 10YR 8/3, slip 10YR 8/2, colors 10YR 2/1 (black), 7.5R 4/4-4/6 (red). H. 0.15, L. 0.24 m.

65. Acr. Mus. 77 (K99). Plate B. An entire tile from a lateral sima with light-on-dark decoration.A composition of palmettesand lyre-shapedvolutes flankedby lilies which grow fromthe spirals is repeatedon either side of a waterspout;part of the lion's mane is preserved.On the back is an incised e indicatingthe positionof the piece on the roof. On the soffit a bead-and-reel pattern appears against a red band. Clay 10YR 7/4, slip 2.5Y 8/4-7/4, color 1OR4/4 (red). H. 0.15, W. 0.165, L. 0.10 m.

Date:420-405 B.C. Bibliography: TdA I, p. 43, fig. 48, pl. 10 (Type XXI). 68. Acr. Mus. (Kl 18). Handmade lion's-head waterspout from a sima, preservingpart of the akroterionbase. The surfaceof the face and the muzzle is damaged.The mane is arrangedin two layersof incisedwavy locks.The mouth is open at the sides to expose the teeth of the animal. Traces of black glaze are preserved on the eyes, lips, and mane. The soffitcarriesa red band. Clay 7.5YR 6/4, slip 10YR 7/4, colors 10YR 3/1 (black), 5R 3/3 (red). H. 0.256, W. 0.26, D. 0.22 m. Date: 430 B.C. Bibliography:TdA I, pp. 51-53, figs. 62, 63; Mertens-Horn (57 above), p. 58 (Acropolis Kl 18) and pp. 67-68 (Olympia, LK-WS 38, 39); A. Mallwitz, OlympischeForschungenV, Berlin 1964, p. 113, pls. 39 and 14 (dated430 B.C.).

Date:420-405 B.C. Bibliography: TdA I, pp. 43-44, figs. 34 and 51 (Type XXI). 66. Acr. Mus. 108 (K98). Right half of a lateral sima preservingpart of the lion's head. The decorationis of the same type as 65. Clay 7.5YR 7/4, slip 2.5Y 8/4, colors 5Y 2.5/1 (black), 5R 4/3-4/4 (red). H. 0.15, W. 0.265, L. 0.155 m.

Date:420-405 B.C. Bibliography:TdA I, p. 43, figs. 34 and 47, pl. 10 (Type XXI). 67. Acr. Mus. 9551 (K101). Right half of lateral sima preserving part of the lion's head. The decorationis of the same type as 65. 62-65 belong to the same roof.

S~_E

--

'

i-

65

THIS CATALOGUE WAS DESIGNED AND TYPESET BY THE OFFICE OF PUBLICATIONS AMERICAN SCHOOL OF CLASSICAL STUDIES PRINCETON, NEW JERSEY

THE ATHENIAN AGORA VOLUME XXIV LATE ANTIQUITY:

A.D. 267-700

By ALISON FRANTZ with contributions by Homer A. Thompson and John Travlos The Athenian Agora has long been recognized as a crucial site for the archaeology of Athens from the earliest habitation to the devastating Herulian destruction of A.D. 267. No systematic study of the subsequent centuries has appeared since Gregorovius' of 1889, when archaeological evidence was virtually unstudied. This book, the most recent in the Agora series, collects for the first time the archaeological and historical evidence for the area of the Agora in Late Antiquity, a period which spans the last flourishing of the great philosophical schools, the defeat of classical paganism by Christianity, and the collapse of the late Roman Empire. The half-century of excavation in the Athenian Agora by the American School of Classical Studies has yielded the only substantial body of evidence for the Late Antique city. By that time, the Agora had lost most of its significance as a civic center and can be understood only as part of Athens as a whole. Therefore, although the primary focus of this volume is the material uncovered by the Agora excavations, the study also takes into account past and current discoveries elsewhere in the city. Alison Frantz discusses fortifications, streets, houses, temples, baths, shops, industrial establishments, and systems for water and sanitation, together with their associated finds, correlating archaeological, epigraphical, and literary evidence to present as comprehensive an account as available information now permits of the history and topography of the city in the years before A.D. 700. The course of Athenian construction and destruction is traced from the mid-3rd century through the Herulian invasion, the period of recovery in the 3rd and 4th centuries ending with the invasion of the Visigoth, Alaric, in A.D. 396, the 5th century, which saw the closing of the schools of philosophy by Justinian and the first Christian churches, and the gradual decline of the city until the Slavic invasion of the 580's, when Athens began an accelerated slide intGsoblivion. Special attention is paid to questions surrounding the history of the philosophical and rhetorical schools, the establishment of Christianity, and the removal of works of art from Athens to Constantinople. A separate chapter by Homer A. Thompson offers the first detailed treatment of the Palace of the Giants and suggests a new interpretation of its purpose. The appendix by John Travlos provides a meticulous description of the Post-Herulian Wall, its course, and its construction. The book is fully illustrated with plans, drawings, and photographs and contains an index of literary and epigraphical sources in addition to a general index. Published February 1989. xxii +

156 pp., 76 pls., frontispiece. Quarto. Cloth. $65.00.

HESPERIASUPPLEMENTXXII ATTIC GRAVE RELIEFS

THAT REPRESENT

WOMEN IN THE DRESS OF ISIS

By ELIZABETH J. WALTERS This book treats the group of 106 Attic grave reliefs of Roman date (from the end of the 1st century after Christ to the beginning of the 4th) which show women dressed in the costume of the goddess Isis. They constitute about a third of the published grave reliefs of the period. A significant body of material is the series of thirty-four fragmentary reliefs discovered in the Athenian Agora, which are published as a group for the first time. Walters discusses the iconography, Egyptian and Ptolemaic sources, production, and style of the Attic reliefs, with comparisons drawn to other sculpture, both in relief and in the round. New proposals regarding the chronology and meaning of the "Isis" reliefs are followed by a catalogue of the Agora fragments. Appendixes present the types of dress given to Isis in Roman times and a chronological list of the Attic reliefs. A general index, indexes of sites and museums and inscriptions, and a concordance to Conze's corpus of the Attic grave reliefs complete the volume. Published October 1988. xvi +

136 pp., 2 figs. in text, 52 pls. Quarto. Paper. $40.00.

ORDERS SHOULD BE PLACED WITH THE AMERICAN SCHOOL O:F CLASSICAL STUDIES AT ATHENS c/o THE INSTITUTE FOR ADVANCED STUDY, PRINCETON, NEW JERSEY, 08543-0631, UNITED STATES OF AMERICA.

NEW PUBLICATIONS OF THE AMERICAN

SCHOOL OF CLASSICAL

STUDIES

AT ATHENS

CORINTH VOLUME XVIII, PART I THE GREEK POTTERY By

ELIZABETH

G.

PEMBERTON

The final publication of the results of the American School excavations from 1961 through 1973 in the Sanctuary of Demeter and Kore on Acrocorinth begins with the presentation of the pottery of the Greek period. In this volume, Elizabeth G. Pemberton publishes the pottery used in the Sanctuary from the Protocorinthian period through 146 B.C. A glossary of descriptive terminology is followed by twenty-eight shape studies. These studies not only trace the formal development of the types of vessels which are present but also consider the significance of patterns in the occurrence of Corinthian and imported wares and in the popularity of specific shapes with respect to the history and development of the Sanctuary and the activities carried on there. Over six hundred pieces, both whole vessels and fragments, have been selected for inclusion in two catalogues. Catalogue I presents eleven context groups consisting of material from votive pits, deposits of votive discards, and building fills which spans the Greek history of the Sanctuary. These groups reflect the architectural development of the complex and the types of votive and domestic pottery used in all periods, and at the same time they shed light on the cult activities at the Sanctuary. Catalogue II includes nearly five hundred pieces arranged by fabric and decoration. Fine and coarse wares in a wide range of Corinthian and imported fabrics are discussed. Examples of post-Classical phialai are the subject of a contribution by Kathleen W. Slane. The pottery is fully illustrated with photographs and drawings of profiles and decoration. A concordance and lot list are included, as well as a bibliography for Corinthian findspots outside the Sanctuary and an index of findspots and proveniences. Indexes of decorative schemes, dipinti and graffiti, and painters supplement the general index. Published December, 1989. xx + 235 pp., 38 figs. in text, 61 pls., 2 plans. Quarto. Cloth. $65.00.

HESPERIASUPPLEMENTXXIII HELLENISTIC

RELIEF MOLDS FROM THE ATHENIAN By

completed by

AGORA

CLAIREVE GRANDJOUAN

EILEEN MARKSON

and

SUSAN

I.

ROTROFF

This volume treats an unusual group of terracotta molds found in the Athenian Agora. Similar molds are known from other sites in the Greek world, but the group in the Agora, consisting of over one hundred fragments, is by far the largest. The molds were used to produce small, rectangular relief plaques, but it is not known in what material or for what purpose they were made, since no finished plaque has ever been found. Grandjouan's study fixes the date of the Agora molds in the 4th and 3rd centuries B.C. and establishes their context in the world of Hellenistic decorative arts, drawing attention to the especially close connection of their iconographic repertory with the Atticizing luxury arts of the Black Sea region. The catalogue presents 110 pieces arranged by iconographic type. Photographs illustrate the fragments, frequently both mold and cast, and sketches by the author suggest reconstructions for several extremely fragmentary types. An appendix presents the evidence for an ancient kind of cake, commonly called the plakoun6ta,which is known from literary sources and which appears in banquet reliefs and in other cultic contexts. Published January, 1990. xii; + 73 pp., 2 figs. in text, 26 pls. Quarto. Paper. $25.00.