Review of Aegean Prehistory I: The Islands of the Aegean Author(s): Jack L. Davis Reviewed work(s): Source: American Journal of Archaeology, Vol. 96, No. 4 (Oct., 1992), pp. 699-756 Published by: Archaeological Institute of America Stable URL: http://www.jstor.org/stable/505192 . Accessed: 02/05/2012 08:26

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Archaeological Institute of America is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Archaeology.

http://www.jstor.org Review of Aegean Prehistory I: The Islands of the Aegean JACK L. DAVIS

INTRODUCTION of the , and it is no surprise that its se- formed the basis for a tripartite Cycladic chro- Not so long ago the islands of the Aegean (fig. 1) quence established to Helladic and Minoan were considered by many to be the backwater of Greek nology, parallel on the Greek mainland and . The exis- prehistory.' Any synthesis of the field had perforce phases tence of a in the islands, on Keos, to base its conclusions almost exclusively upon data particularly and , had been demonstrated but in collected before the turn of the century. The entire , no instance had been documented. prehistory of the islands received fewer than 16 pages fully A the situation has been al- of discussion in Emily Vermeule's in the Bronze quarter century later, tered Age (Chicago 1964), almost all of this concerned with drastically. Particularly in the last decade prog- ress has been the art and graves of the Early Bronze Age ; very rapid, hampering any attempt to a other parts of the were generally men- produce totally up-to-date synthesis of new data, as tioned only in passing. Vermeule had no choice but does the frequent appearance of important studies in to write that for the Cyclades "only three village sites new Greek periodicals of limited circulation. Indeed, [ on Melos, Kastri on , and Thera] publication has been so voluminous and diverse that have been excavated in a manner one could call in- it is difficult even for specialist Aegean prehistorians formative, in contrast to nearly two thousand known to stay abreast of new developments. The essential or suspected graves" (p. 47). Of these only Phylakopi annual reviews of new work in Greece, published by also offered a deep stratigraphy covering all phases successive directors of the British and French schools

S The publicationof this review was made possible by a The followingspecial abbreviations are used in this paper: generous subvention from the Institute for Aegean Prehis- AEMT To AQXatoA)oytm6vEeyov orri Maxe- tory. bovia XatIOepri. I am particularlygrateful to those friends and colleagues "Chronique" "Chroniquedes fouilles,"BCH. who responded to requestsfor informationconcerning their Cycladica L. Fitton ed., Cycladica: Studies in Mem- recent research,and in particularto those who provided me ory ofN.P. Goulandris (London 1984). with offprints, preprints,or photographs.The more specific L. Marangou ed., Cycladic Culture: contributionsof several colleagues are acknowledgedas ap- in the 3rd Millennium BC (Ath- propriate later in this review. Here I should like to express ens 1990). my thanks to Robert Arnott, Robin Barber,Cyprian Brood- S. Dietz ed., Archaeology in the Dode- bank, Tristan Carter,John Cherry, Christopher Chippin- canese(Copenhagen 1988). dale, John Coleman, MichaelCosmopoulos, Tracey Cullen, Emporio S. Hood, Excavations in Chios 1938- Scren Dietz, Christos Doumas, Angelia Douzougli, Noel 1955: Prehistoric Emporio and Ayio Gale, David Gill, David Hardy, Carol Hershenson, Donald Gala, 1-2 (BSASuppl. 15-16, London Keller,Sandy MacGillivray,Sturt Manning, Lila Marangou, 1981-1982). MarizaMarthari, Dimitris Matsas,Lyvia Morgan, Christine "First Coloniza- J.F. Cherry, "The First Colonizationof Morris,John Overbeck, Mehmet Ozdogan, Ernst Pernicka, tion" the MediterraneanIslands: A Review , Efi Sakellaraki,Diamantis Sampson, Eliza- of Recent Research,"JMA 3 (1990) beth Schofield, Zophia Stos-Gale, Rene Treuil, Sarah 145-221. Vaughan, David Wilson, and KostasZachos. Tracey Cullen, Gazetteer R. Hope Simpson and O.T.P.K. Dickin- Fred Kleiner,and Jerry Rutterconceived of this projectand son, A Gazetteer of Aegean Civilisation encouraged me to undertake it. Over the past year and a in the Bronze Age 1: The Mainland and half I have at times regretted acceptingthe assignment,and Islands (SIMA 52, Goteborg 1979). in moments of panic have cursed them severallyand collec- GreekPrehistory E.B. Frenchand K.A. Wardleeds., Prob- tively,but I am in the end thankful that they convinced me lems in Greek Prehistory (Bristol 1988). to do it. I am also grateful to Shari Stockerfor help with the , Saros M. Melas, The Islands of Karpathos, Saros illustrationsand proofreading, to Bill Parkinsonfor compil- and and Kasos in the Neolithic and Bronze ing references, and to John Bennet, Cyprian Broodbank, Age (SIMA68, G6teborg 1985). John Cherry, Mihalis Fotiadis, Donald Keller, Sandy Mac- Kastro Tigani R.C.S. Felsch, Das Kastro Tigani: Die Sturt Curtis Gillivray, Manning, Runnels,Jerry Rutter,and spiitneolithischeund chalkolithischeSied- DavidWilson for their prompt commentson my penultimate lung ( II, Bonn 1988). draft.

699 AmericanJournal of Archaeology96 (1992) 700 JACK L. DAVIS [AJA 96 in Archaeological Reports (AR) and in BCH's "Chro- BEYOND THE CYCLADES TO A PAN-AEGEAN nique des fouilles" ("Chronique") are, of course, es- PERSPECTIVE sential points of departure for both scholars and students, but can be patchy in their coverage and, in The decision to include virtually all islands of the any case, are not intended to be synthetic.2 The bible Aegean (excluding only Crete and those of the Sa- of Aegean prehistory, R. Hope Simpson and O.T.P.K. ronic Gulf) reflects trends in recent scholarship that Dickinson's A Gazetteer of Aegean Civilisation in the have defined problems demanding a canvas much Bronze Age 1: The Mainland and Islands (Goteborg broader than that offered by the Cyclades alone.5 Two 1979, completed 1977; hereafter, Gazetteer), is now examples of topics that require a pan-Aegean focus nearly 15 years out of date, and in any case omitted may suffice: the Neolithic colonization of the Aegean coverage of the islands of the Aegean that lie north islands; and the Minoanization of the Aegean in the of and Chios.3 Robin Barber's The Cyclades in Middle Bronze Age and early part of the Late Bronze the Bronze Age (Iowa City 1987) provides an excellent Age. overview of the results of research in the southern In the past decade there have been, for the first Aegean, but publications relevant to Cycladic prehis- time, systematic attempts to describe initial settlement tory have been so prolific in the past six years that an of the islands of the Aegean in terms of principles update is also desirable.4 drawn from island biogeography.6 The general idea

Les Cyclades G. Rougemont ed., Les Cyclades:Mater- TAWIII D.A. Hardy et al. eds., Thera and the iaux pour une etude de geographiehis- AegeanWorld III. 1: Archaeology(Lon- torique(Paris 1983). don 1990); 2: Earth Sciences(London Manika I-II A. Sampson, Mdtvtxa I: Mia z#wroEA-- 1990); 3: Chronology(London 1990). Aa6txIj r6Aqro'nrj XaAxiba ( 2 For an index to the sites describedin AR between 1976 1985); Mdtvtxa II: O rtwroEliali- and 1986, see AR 33 (1987) 78-87. All references below to x6; oxtoa6 g xal To VEQoTa/Efio AR without further specificationof title or author are to the (Athens 1988). annual reports, "Archaeologyin Greece,"compiled by H.W. Minoan A.G. Papagiannopoulou,The Influence Catling and, since AR 36 (1990), by E.B. French. References Influence of Middle Minoan Potteryon the Cyc- to "Chronique"are to "Chroniquedes fouilles,"BCH, com- lades (SIMA-PB96, G6teborg 1991). piled by G. Touchais and, since BCH 114 (1989), by A. Minoan Thalas- R. Higg and N. Marinatos eds., The Mi- Pariente. socracy noan Thalassocracy:Myth and Reality 3 The islands of the northern Aegean are included in D. (SkrAth32, Stockholm 1984). Leekleyand R. Noyes,Archaeological Excavations in theGreek Neolithic and A. Sampson, H NEOlAtOij) xat 77 Islands (ParkRidge, N.J. 1975), which is, however, further Protohelladic HQwtroEAlabit I aIrlvEofpota (AQ- out of date than the Gazetteerand is not nearly so authori- tativeor exhaustivein its coverage. XEiov EvfloixdOv MEAE)•bv, Hap- darqoTLarov KA' T6Mov, Athens 4 With reservations(see my reviewin AJA93 [1989] 293- 1981). 94), W. Ekschmitt,Kunst und Kultur der KykladenI: Neoli- Neolithic A. Sampson, H NEOAtOtXjrjrEQobog tikumund Bronzezeit (Mainz 1986) mayalso be recommended Dodecanese oaraAwmbExdivrua (Athens 1987). as a reasonablycurrent review of Cycladicprehistory; the "Perspectives" J.L. Davis, "Perspectiveson the Prehis- extensiveillustrations, many in color,are particularlyworthy. toric Cyclades:An ArchaeologicalIn- See also "Perspectives"for a recent brief overview of the troduction,"in P. Getz-Preziosi,Early EarlyCycladic period. CycladicArt in NorthAmerican Collec- 5 Islands of the Saronic Gulf include all those that cur- tions(Richmond 1987) 4-45. rently belong to the administrativedistrict of Peiraeus, in- Prehistoric J.A. MacGillivray and R.L.N. Barber, cluding , , Kythera,and .Covered Cyclades The PrehistoricCyclades (Edinburgh in this review,in wholeor part,are territoriesof the following 1984). Ephorates of Prehistoric and Classical Antiquities: 1st Silber, Blei und G.A. Wagner and G. Weisgerber, Silber, (Keos); 11th (Euboiaand Skyros);13th (northernSporades); Gold Blei und Goldauf :Prdihistorische 18th (); 19th (); 20th (, Chios, und antikeMetallproduction (Der An- ,and Limnos);21st (Cyclades,Samos, and Ikaria, ex- schnitt,Beiheft 3, Bochum 1985). cluding Keos and ); and 22nd (Dodecanese and "Sourcesof Z.A. Stos-Gale and C.F. Macdonald, Amorgos). For a comprehensivedescription of the admin- Metals" "Sourcesof Metals and Trade in the istrative districts that comprise the Greek Archaeological Bronze Age Aegean," in N.H. Gale Service, see AR 36 (1990) 4. The order of presentationfor ed., BronzeAge Tradein the Mediter- yearly reports both in AR and in ArchDeltmirrors this ad- ranean (SIMA 90, Jonsered 1991) ministrative structure, and it will also be followed here. 249-88. 6 J.F. Cherry, "Pattern and Process in the EarliestColo- TAWI C. Doumas ed., Thera and the Aegean nization of the MediterraneanIslands," PPS 47 (1981) 41- WorldI (London 1978). 68, has considered these issues most extensively. 1992] THE ISLANDS OF THE AEGEAN 701

MACEDONIA THRACE

SKALA SOTIROS

THASOS SAMOTHRACE MIKROVOUNI

IMBROS THEOTHESSALYLOGOSOR LIMNOS ; POLIOCHNI TR MYRINAOS

NORTHERN THESSALY YOURA KYRAPANAYIA K AYIOSEFSTRATIOS MITHYMNA PETROS ALONNISOSSERYIOS NAXOS LTHERMI PAAMAR KALYMNOSLESBOS KOUMBI SKYROSISCHORA

EUBOIA

MANIKASTKYMETAYRALANSKOTEINICAVE

ARGOLID 9 AYIAIRINI IKARIAHERMION KEOS KAS FRANCHTHICAVE KYTHNOSSYROS OS KASOSIASOS GROTTANAXOS SERIPHOSIOS SOSTISKOUKOUNARIE SARIAOS SIPHNOSPAROS`) CAVEOFZAS KALYMNOSl SERAGLIO PHYLAKOPIAKKAKARYSTOSRPATHOSANDRSKARKOI AMORGOS YIALIKNIDOS MELOSIOS ASTYPALAIANISYROSQ THRTHERA TILOSTRIANDA AKROTIRIT>O ALIMNIAKALYTHIESCAVE CYCLADES DODECANESE HALKIKOUMELO CAVE SARIA KARPATHOS

KAOSRAPEZA

CRETE

Fig. 1. PrincipalAegean archaeologicalsites discussedin this review of these studies is that geography has played a major adjacent mainland coasts, its absolute size, and the role in determining the date, extent, and rate of set- presence or absence of stepping stone islands between tlement in the Aegean. Examination of the way in it and a mainland were important factors in deter- which both animal and plant species have come to mining the likelihood that the island will have been inhabit island groups in other parts of the world settled at a particular time in the past. This is not of suggested that certain patterns of colonization might course to suggest that social or political factors should also be recognizable in Greece. For example, it has be eliminated from the equation, but rather that been hypothesized that the distance of an island from biogeography can provide an initial investigative 702 JACK L. DAVIS [AJA 96

framework within which the importance of cultural unlikely to make a great difference to present gen- determinants of settlement can be more explicitly eralizations. While it is certainly true that some settle- defined. ments of Palaeolithic and earlier Neolithic date may Hand in glove with these analyses have come at- be lost to fluctuations in coastlines or may yet lie tempts to explain why extensive colonization of the undetected, the fact remains that no clear evidence of occurred so long after settlement of a Palaeolithic presence has yet been recognized in the adjacent mainlands. It has been realized for some islands, at least not on any island that was then sepa- years that the islands of the Aegean cannot have been rated from an adjacent mainland by an appreciable inaccessible to potential human colonizers, inasmuch gulf. as obsidian from Melos was reaching the Greek main- The evidence for earliest colonization of the Ae- land already in the later Palaeolithic and Mesolithic; gean islands has recently been updated.9 Most of the finds are well documented in strata excavated at islands of the Aegean appear to have been first in- in the southern Argolid and are habited during the EBA, although traces of Neolithic clearly Melian in origin.' The recognition that the habitation are more plentiful today than even a year Aegean was being navigated long before the intro- or so ago. A Neolithic presence has now been docu- duction of agriculture to Greece has obvious and mented in the Cyclades (Keos, Naxos, Thera, Amor- important repercussions for how the process by which gos, , Saliagos, and Siphnos), in the northern agriculture was spread from the Near East to Greece Sporades (Kyra Panayia and Youra), on Chios, Psara, is viewed: clearly an absence of evidence for settle- Samos, in most of the Dodecanesian islands, and on ment in the earlier phases of the Neolithic in the Limnos, Lesbos, Samothrace, and Thasos. In all cases Greek islands no longer requires us to postulate the the earliest material yet recognized appears to be later existence of a more northern route of migration for in date than that from adjacent mainlands; in general, Neolithic immigrants, for which there has been pre- the final period of the Neolithic (FN) seems to have cious little evidence. The Aegean sea of the later been the time of maximum expansion in settlement. Palaeolithic was navigable and navigated.8 The earliest well-documented Neolithic settlement is Of equal importance is the realization that inhab- that of Ayios Petros on Kyra Panayia, situated near itants from the adjacent mainlands had the capability the eastern end of a string of islands that leads out of establishing settlements in the Aegean islands long into the Aegean from Thessaly, itself probably the before they actually did so. Documentation in the most densely settled part of the Greek mainland dur- future of any evidence for earlier transient activity is ing the Early and Middle Neolithic.1o Certainly it

See C. Perlks, Les industries lithiques taille'esde Franchthi frew, M. Gimbutas, and E.S. Elster eds., Excavations at Si- (Argolide, Grace) 1: Prisentation generale et industries paldo- tagroi: A Prehistoric Village in Northwest Greece (Los Angeles lithiques (Franchthi 3, Bloomington 1987) 142-45; 2: Les 1986) 477-85; J.E. Coleman, "Greeceand the Aegean," in industries du Misolithique et du Niolithique initial (Franchthi R.W. Ehrich ed., Chronologies in Old World Archaeology" For 5, Bloomington 1990). further discussion of Melos and (Chicago, in press); and D. Grammenos, NEOALtOLXgQEUvvE FranchthiCave, see now also C. Renfrew and A. MaxE6ovia Aspinall, or7v XEVTtQLX)xatL avaToALtx?1 (Ltfl•AtoOfl•X7Trl "AegeanObsidian and FranchthiCave," in Perliks1990, 257- Ev AOrlvatg ApXatoAoyltxr EratQEiag 117, Athens 1991) 70. These authors also note that postulated linkages between 29. Recent excavations and explorations in Turkish Thrace tunny fishing and collection of obsidian on Melos by main- are of particular interest. At the site of Hoca Qegme in 1990- landers are not supported by the evidence from Franchthi 1991 (phase IV), ceramics similar in character to those of Cave, since obsidian has been found there in levels earlier central , with parallels in Hacilar IX-VI, have been than those in which evidence for deep sea fishing first ap- excavated. Preliminary notices of these results will appear in pears. Araftzrma Sonu(larz, and Germania. See also M.J. Mellink, 8 Recent research in Thessaly has tended to support a "Archaeology in Anatolia," AJA 96 (1992) 125; and M. Oz- hypothesis that agriculture was introduced to Greece from dogan, Y. Miyake, and N. OzbaSaran Dede, "An Interim the Near East through migration, rather than through trade Report on Excavations at Yarimburgaz and Toptepe in East- between Near Easterners and an existing Mesolithic Greek ern Thrace," Anatolica 17 (1991) 81-82. population. For a current review of the evidence, see J. 9 "First Colonization" 145-221. Hansen, "The Introduction of Agriculture into Greece: The 10 Palaeolithic finds have been reported from both Thasos Near Eastern Evidence," AJA 96 (1992) 340-41 (abstract). and Euboia; both islands would then have been attached to See also J.M. Hansen, The Palaeoethnobotany of Franchthi adjacent mainlands. Palaeolithic finds have also been re- Cave (Franchthi 7, Bloomington 1991) 174-81; C.N. Run- ported from various sites on , Kyra Panayia, and nels, "Trade Models in the Study of Agricultural Origins Skyros, all in the northern Sporades, but are not yet well and Dispersion," JMA 2 (1989) 149-55. On the scarcity of documented; see "First Colonization" 167. Stretches of open Neolithic in early sites Eastern and Thrace, see sea between these islands and the mainland of Thessaly may C. Renfrew, "Sitagroi in European Prehistory," in C. Ren- also have been negligible. 1992] THE ISLANDS OF THE AEGEAN 703 cannot be doubted that Early Neolithic populations land bridge that joined Paros and ; the sec- were capable of organizing and launching large-scale ond contemporary with the fourth-millennium FN colonizations across open sea. The settlement of the site of Kephala on Keos; and the third represented island of Crete in the later eighth or earlier seventh by the expansion of population at the transition be- millennium is an obvious example of a successful tween the Neolithic and EBA that resulted in the enterprise of that sort, one that "indicates an exoge- creation of the Early Cycladic culture.'3 Notional col- nous introduction of farming and farmers through a onizations of the Cycladic islands have been modeled: purposive, planned and comparatively long-range one beginning from the Greek mainland (specifically colonization." That island, like others of the Aegean, from Attica and Euboia), and another starting from thus far lacks any clear evidence for pre-Neolithic the Dodecanese and progressing via Ikaria and As- activity."I typalaia. On the basis of these models it has been It is equally important that those conditions be suggested that in the Saliagos phase, it is most likely determined that eventually did result in successful that colonists came to the Cyclades from the southeast colonization of the Aegean islands and in extensive, Aegean alone, but that later colonists probably pro- widespread, and long-lived settlement. A Neolithic ceeded from both directions at once. In all cases, the agricultural package of domesticated crops and ani- survival of settlements on Naxos would have been mals would have offered potential colonists a better especially favored by circumstances of natural geog- chance for long-term survival in areas where wild raphy. Indeed, only on Naxos have traces of all three resources were restricted. But although a Neolithic phases of settlement been recognized and only on way of life was an obvious precondition for the via- Naxos can a case for continuity in settlement through- bility of long-term occupations on the tiny and re- out the Late Neolithic and Final Neolithic periods be source-poor islands of the Aegean, it cannot have made. been the only factor in play. Indeed, there is a gap of A clearer understanding of the factors that led to some two to four millennia between the colonization expansion of settlement into the islands of the Aegean of Crete and that of the Cyclades, northern Sporades, will also require a careful examination of the character or the Dodecanese, despite the fact that some of these of societies in those areas likely to have provided islands lie directly along potential lines of migration colonists. Although a topic well beyond the scope of between Crete and Anatolia. Principles of island bio- this review, it is worth noting that there appears to be geography also suggest that the survival rate of settle- a close correlation between an expansion of settlement ments close to colonizing mainlands will be higher, on the island of Euboia and the establishment of since their populations will be able to look to the settlements in many of the Cycladic islands during the neighboring mainlands for support, whether for mar- Final Neolithic. In particular, the habitation of mar- riage partners, breeding stock, or, in crisis, even for ginal parts of southern Euboia could be seen as part seed grain.12 The colonization of Crete should thus of a general movement of populations into marginal be seen as a purposeful effort to found a new settle- areas, the beginning of a trend that would lead to a ment in an especially favorable island environment, prodigious expansion of settlement in the EBA.'4 So- and not as the result of a gradual expansion of pop- cial, as well as geographical, problems would have ulation through the islands to its north and east. been faced by early colonists in the Cyclades. Several Three general horizons of Neolithic colonization in of these have recently been examined in a study that the Cycladic islands have recently been defined: the seeks to explain the differential development of Crete first contemporary with the site on Saliagos, an islet and the Cyclades in the course of the EBA.'15 The role that in the fifth millennium B.C. was situated on a of exogamy in linking relatively tiny communities of

I C. Broodbank and T.F. Strasser, "Migrant Farmers and 15 S.W.Manning, "The Emergenceof Divergence:Devel- the Neolithic Colonization of Crete," Antiquity65 (1991) opment and Declineon BronzeAge Crete and the Cyclades," 233-45. in C. Mathers and S. Stoddart, Development and Decline in 12 Broodbank and Strasser (supra n. 11) 238-39. the Mediterranean Bronze Age (in press). There has as yet, 13 See C. Broodbank, "Colonization and Culture in the however, been no serious attempt to use ceramic character- Neolithic and Early Bronze Age Cyclades," AJA 96 (1992) ization analyses to distinguish between a scenario in which 341 (abstract).I am extremely grateful to Cyprian Brood- items were produced on two different islands in the Neolithic bank for a copy of the text of his paper as read, and for within similar cultural traditions, and one in which items permission to summarize his argument here. were traded between the same two islands. See J.E. Cole- 14 On the relationship between settlement in southern man's (supra n. 8) discussion of evidence for Neolithic trade Euboia and FN/EBA colonization of the Cyclades, see Keller in the Aegean. I am grateful to both authors for preprints and Cullen (infra n. 61). of their forthcomingpublications. 704 JACK L. DAVIS [AJA 96 the islands is emphasized. Exogamy, it is suggested, literature has been summarized in the volume Bronze also promoted a stylistic homogeneity in artifacts Age Trade in the Mediterranean.17 In general a clear within the Aegean, of a sort recognizable already in pattern appears to have emerged from recent analy- the earliest Neolithic Saliagos culture of the Cyclades. ses. In the EBA Cyclades, Siphnos and served By the second phase of the EBA, however, exchanged as major sources for lead, silver, and copper. Both of goods had come into the hands of specialist produc- these sources were replaced in the Middle Bronze ers, partly because opportunities for the intensifica- Age by Laurion, which came to dominate the Aegean tion of agricultural production were so limited, and as the principal supplier of metals.18 It is clear too the exchange of agricultural goods on a large scale that exchange in marble vessels and figurines played infeasible. a role in EBA exchange, but slight progress has been It is clear that exchange played a crucial role in made in this arena. There is little agreement among enabling permanent settlement in the islands of the scholars as to the validity of attributing marble figu- Aegean, even if it is not yet possible to quantify the rines to individual sculptural hands or workshops, extent of this trade. In the EBA, the similarities in and there is little basis for assigning such personalities the formal characteristics of ceramics, marble vessels, or production centers to particular locations within figurines, and metal objects that typify the Early Cy- the Cyclades, because the vast majority of marbles cladic culture are indicative of social and economic lack verifiable archaeological contexts.'" For these rea- ties maintained among the settlements of the islands; sons and because of their rarity, marble figurines are these relationships may also be regarded as necessary not likely ever to provide us with more than the adaptations that would have provided access to addi- crudest measurements of intra-Aegean exchange. tional resources of food and manpower in times of Fewer than 2,000 figurines are known, and these were crisis.16 Exchange may, therefore, be seen not as the produced over some 600-700 years! incentive for colonization of the islands of the Aegean, The study of ceramic fabrics promises in the long but as an indispensable enabling mechanism that pro- run to be of greater value in reconstructing patterns moted the survival of groups once established, partic- of exchange. Given the diversity of geology among ularly on smaller, more remote, and impoverished the islands of the Aegean, it is in many cases possible landfalls such as the so-called Amorgian islets of to recognize imported products on the basis of both Epano and Kato Koufonisi, , Schinousa, or visual inspection and petrological analysis. The results . of a long-term project conducted under the auspices Our understanding of trade in metals has been of the Fitch Laboratory of the dramatically improved in recent years. Much current should soon be available. This study has focused spe-

16 See J.F. Cherry, "Island Origins: The Early Prehistoric 389-406. Evidence from the analysis of the small number Cyclades," in B.W. Cunliffe ed., Origins, the Roots of Euro- of actualsilver artifacts (as opposed to litharge samples)that pean Civilisation (London 1987) 25-26; "Perspectives" 26- have yet been examined is more equivocal: Laurion and 30. Siphnos do not emerge so obviously as dominant sources. 17 N.H. Gale ed., Bronze Age Trade in the Mediterranean 19 On the definition of sculptural hands, see P. Getz- (SIMA 90, Jonsered 1991). Of particular importance is the Preziosi, Sculptors of the Cyclades: Individual and Tradition paper by Z.A. Stos-Gale and C.F. Macdonald, "Sources of in the Third Millennium B.C. (Ann Arbor 1987); see ch. 7 Metals." Fig. 1, p. 254, provides a map of ore deposits for patterns of distribution, such as they are, given the lack sampled for lead-isotopic composition through 1990. On the of a properexcavated context for the vast majorityof figures. subject of direct evidence for Bronze Age Cycladic metal- For doubts about the use of a canon in the production of lurgy, see also Z.A. Stos-Gale, "Cycladic Copper Metallurgy," Cycladic marble figures and about the correctness of pro- in A. Hauptman, E. Pernicka, and G.A. Wagner eds., Old cedures employed to assign sculptures to hands, see C. Ren- World Archaeometallurgy (Der Anschnitt, Beiheft 7, Bochum frew, The Cycladic Spirit: Masterpieces from the Nicholas P. 1989) 279-92. These papers should be read in conjunction Goulandris Collection (New York 1991) ch. 9 and pp. 137- with N.H. Gale and Z.A. Stos-Gale, "Cycladic Lead and Silver 41. Soon see also C. Chippindale and D. Gill, "Material and Metallurgy," BSA 76 (1981) 169-224. See also E. Pernicka, Intellectual Consequences of Esteem for Cycladic Figures" "Erlagerstatten in der Agais und ihre Ausbeutung im Alter- (in prep.) for discussion of the magnitude of the problem tum: Geochemische Untersuchungen zur Herkunftsbestim- faced by archaeologists because such a substantial part of mung archaologischer Metallobjekte," JRGZM 34 (1987) the corpus of marble figurines has been acquired through 607-714. For the role of specific islands in prehistoricmet- illegal excavations; Chippindale and Gill also adduce sub- alworking and trade, see discussion of Karpathos, Keos, stantial art historical grounds for doubting that individual Thera, Chios, and Melos in the following section of this hands of prehistoric sculptors can be defined. J.F. Cherry, review. "The Individual in Prehistory: Reflections on Attribution 18 Possible explanations for the replacement of Siphnos Studies in the Bronze Age Aegean," in J.C. Crowley and R. Laurion as by dominant supplier to Keos have been discussed Laffineur eds., EIKON. Aegean Bronze Age Iconography: by N.H. Gale, Z.A. Stos-Gale, and J.L. Davis, "The Prove- Shaping a Methodology (Aegeum 10, in press). nance of Lead Used at Ayia Irini, Keos,"Hesperia 53 (1984) 1992] THE ISLANDS OF THE AEGEAN 705 cifically on the EBA and has included material from of the extent of changes observable in the material the sites of Keos Ayia Irini; Naxos Cave of Zas, Grotta, culture of the islands. Political dominance by Crete and Palati; Thera Akrotiri; Melos Phylakopi; los Skar- may well have encouraged the adoption of Minoan ; Amorgos Markiani; and Keros Kavos. It is antic- ways in its overseas colonies. But conquest is, in and ipated that the final product will be presented in the of itself, no real explanation for change in material form of a handbook that can be used in the field for culture: one can too easily point to a plethora of fabric indentification.20 historical examples in which imperial control resulted A second central issue in prehistory that benefits in relatively few changes in the everyday life of the from a pan-Aegean perspective is the relationship majority of the conquered. between Crete and the islands of the Aegean in the Second, there is an enormous diversity in the settle- MBA and earlier phases of the LBA, the time of the ment history of those island centers that were pulled so-called Minoan Thalassocracy. Recent research on within the Minoan orbit. The entirety of the southern many of the Aegean islands leaves no doubt that Aegean was strongly affected by contact with Crete; contacts between Crete and the Cyclades became es- the northern Aegean appears to have lain largely pecially intense after the construction of the Old Pal- outside the sphere of Cretan cultural influence. In the aces.21 Exchange appears to have played a major role west, no evidence for a heavily Minoanized settlement in motivating Cretan involvement in the affairs of the has been recognized north of Keos; in the east, Kos islands to its north. On many of these islands, local is the dividing line. In the central Cyclades, Naxos industries were deeply affected by the contact. Local marks the limit. None of this is to say, of course, that styles of decoration, particularly for pottery, and local evidence for contact with Crete cannot be recognized manufacturing techniques were abandoned in favor farther to the north. Minoan ceramic imports have of Minoan-inspired prototypes. A number of these long been acknowledged at Pefkakia in Thessaly, and changes may well have resulted from elite emulation on Samos, while Minoan artifacts appear to have made of the status goods produced by the Minoan civiliza- their way even to Troy, albeit in small quantities.23 tion. Others, for instance the widespread adoption in The recent discovery of a Minoan sealing and roun- the islands of Minoan forms of loomweights, may del at the site of Mikro Vouni on Samothrace must be reflect subtler variation in the structure of Cycladic viewed within this general context.24 The situation of social and economic organization.22 these discoveries, however remarkable, should be dis- Two items have been clarified by extensive recent tinguished from that of Cretan finds in the heavily debate. First, it is not possible to measure the extent Minoanized zone of the southern Aegean. Minoan of political control by the Cretans as a direct reflection finds north of an east-west line drawn between Attica

20 This project is under the direction of S.J. Vaughan, to 21 Evidencefor interactionbetween Crete and the islands, whom I am extremely grateful for the informationsumma- as understood in 1983, is summarizedby papers in Minoan rized here. Her work will soon be publishedin S.J. Vaughan, Thalassocracy.The current state of affairs is outlined in two Early Bronze Age Cycladic Pottery Fabrics: Studies in Materials papers by M.H. Wiener,"The Nature and Controlof Minoan and Technology (Fitch Laboratory Occasional Paper 5, in Foreign Trade," in Gale (supra n. 17) 325-50; and "The prep.). Vaughan emphasizes that the evidence supports her Isles of Crete? The Minoan Thalassocracy Revisited,"in hypothesis that "alreadyin the EBA there were well-estab- TAWIII.1, 128-60. lished trade and exchange routes throughout the Cyclades 22 See, e.g., the arguments made in J.L. Davis, "Cultural and beyond" and that there was "very specialized ceramic Innovationand the MinoanThalassocracy," in Minoan Thal- production and trade reflecting skilled exploitation of the assocracy 159-66. best raw materialsavailable on each island, with connections 23 For Minoan finds in the eastern Aegean, see W.D. perhapsfor exchange with other items such as ores."Various Niemeier, "The End of the Minoan Thalassocracy,"in Mi- special studies in press or in preparation examine the so- noan Thalassocracy205-14 and discussionon p. 215; for the called EBA talc ware (S. Vaughan and D. Wilson, "Interre- western coast of the Aegean, see J.B. Rutter and C.W. Zer- gional Contactsin the EB II Aegean: The Talc Ware Con- ner, "EarlyHellado-Minoan Contacts," in Minoan Thalasso- nection," in C.W. Zerner and E.B. French eds., Wace and cracy75-82. For an as yet unverifiedreport of MM I pottery Blegen: Pottery as Evidence for Trade in the Aegean Bronze at Troy, see D.F. Easton,"The Chronologyof West Anatolia Age 1939-1989, in press), and ceramic fabrics from Amor- in the EarlyBronze Age,"BICS 35 (1988) 180-81; for stone gos Markiani, Melos Phylakopi, and Keros Kavos. Recent vases of Minoan type at Troy, see P.M. Warren, Minoan petrologicalstudies of ceramicsfrom Akrotirion Thera and Stone Vases (Cambridge 1969) 17; "Minoan Stone Vases as from Mikri Vigla on Naxos are described in the following Evidence for Minoan Foreign Connexions,"PPS 33 (1967) section of this review. In addition to these, see also S.J. 37-56. Vaughan, "Bronze Age CycladicWhite Wares:A Question 24 See the following section of this review and infra n. of Materialsand ManufacturingTechniques," AJA 96 (1992) 100. For earlier finds in the islands, see T.G. 342 (abstract).I thank D.E. Wilson for a preprint of his and Palaima,"Linear A in the Cyclades:The Trade and Travel Vaughan'spaper on talc ware. of a Script,"TUAS 7 (1982) 15-22. 706 JACK L. DAVIS [AJA 96 and Samos remain sporadic, and the material cultural apparently established de novo on virgin soil, and assemblages in which they have been discovered are there is very little evidence for preexisting non- predominantly non-Minoan in character. The archae- Minoan populations. The settlements of Trianda on ological picture does not, however, entirely conform Rhodes and Seraglio on Kos appear to have been of to predictions that "the strength of Minoan contacts substantial size from the time of their foundation. is proportional to proximity to Crete."25Rather, there Such a pattern, and the fact that the material culture appears to have been a dramatic falloff in Minoan of these sites was almost entirely Minoan in character, influence between the north and , and may well reflect purposeful large-scale colonization. even within the southern Aegean there are marked But such an interpretation will not explain the Minoan differences in the nature of Cretan influence. presence in the Cyclades. More than a decade ago I argued that "there existed For example, it is now obvious that the settlement a zone in the Western Cyclades (encompassing at least of Akrotiri on Thera was ancient when a Minoan the islands of Thera, Melos, and Keos) in which there presence was first felt; its history began already in the was regular exchange between Cycladic settlements Neolithic period. Even at the time of most intense and Crete."26 My argument in 1979 was that these Cretan presence in the earlier stages of the Late islands had been preferentially supplied with Minoan Bronze Age, elements of the preexisting, non- products and that Minoan trading activities in the Minoan, Cycladic material culture survived at Akrotiri southern Aegean were to some extent directional (and and at contemporary centers like Phylakopi on Melos thus purposeful). The principal settlements on Thera, and Ayia Irini on Keos. There can be no doubt that Melos, and Keos were interpreted as three important local non-Minoan populations continued to occupy ports along a "Western String" exchange route be- these sites, even if we allow for the possibility that one tween Crete and the mainland. The subsequent defi- or all may have been administered by a Cretan over- nition of a similar "enriched" zone in the Dodecanese lord or that Minoan elements were present in the local lends, I think, support to my proposition that one of population. The case for actual Cretan settlement the most important motives for such exchange was seems strongest for Thera, partly because of the re- the acquisition of metals by Crete. The extent to which markable divergence of its settlement pattern from eastern Attica supplied the Minoan world with lead what was apparently the Cycladic norm. On both Keos and silver has become much more clear in the past and Melos, recent systematic surface surveys have decade; copper now too can be added to the list of reinforced a picture of islands dominated by a single mineral products that reached Crete from Laurion. "primate" center; the pattern of settlement on Thera The sharp drop-off in evidence for Minoan contact appears to have been radically different and perhaps to the north of Samos may suggest that Cretans in the approximated that of New Palace Crete, with its dense eastern Aegean were also interested in the acquisition array of towns, villages, and villas. of a particular product, either one locally produced Such evidence does not suggest that the motivation or that could be acquired through the medium of for Minoan involvement as far north as Melos and secondary distribution centers.27 Keos was primarily oriented toward conquest and This much is clear. The desire to colonize (i.e., to colonization. Neither Ayia Irini nor Phylakopi seems establish new Minoan settlements abroad) cannot have to have expanded remarkably in size during the New supplied the only motivation for Cretan activity in the Palace period and the countryside of both islands Aegean, although it may explain archaeological evi- remained relatively vacant. Nor does evidence from dence produced by recent excavations and surveys in the central and northern Cyclades attest to an expan- the Dodecanese. There a remarkable increase both in sion of Minoan population, although here the quality the size and number of settlements occurred dur- of our evidence is less adequate than in the western ing the Minoan New Palace period, particularly on Cyclades.28 At present, Delos (at approximately the Rhodes, Kos, and Karpathos. These settlements were same latitude as Keos and Samos) is the island farthest

25 P.M. Warren, "The Thalassocracy of Minos," Proceed- suggested that the acquisition of metals may have been a ings of the Classical Association, London 67 (1970) 64. motivation in eastern Aegean-Cretan exchange. 26 J.L. Davis, "Minos and Dexithea: Crete and the Cyc- 28 E. Schofield, "The Western Cyclades and Crete: A 'Spe- lades in the Later Bronze Age," in J.L. Davis and J.F. Cherry cial Relationship'?" OJA 1 (1982) 9-25, for the earlier evi- eds., Papers in Cycladic Prehistory (UCLAMon 14, Los An- dence; for pertinent recent data, see the discussions of Naxos geles 1979) 143-57. and Paros in the following section of this review. 27 See Niemeier (supra n. 23) 206 n. 18, where it is also 1992] THE ISLANDS OF THE AEGEAN 707 north on which Minoan imports have been recog- social units and the diversityof environments, should nized.29Andros, Tinos, Syros, and Mykonos remain in the future permit us to explain in much greater virtual blanks; nowhere in the Aegean is a systematic detail the materialconsequences of the Thalassocracy program of surface exploration and excavationmore of Minos. necessary. The Late CycladicIII period in the Aegean offers The past decade has, however, produced new evi- a similar opportunity for a synthetic comparative dence from the central Cyclades.Finds from surface study of the processes by which Mycenaeanmaterial investigations at Mikri Vigla in western Naxos, al- culture was adopted in the islands of the Aegean, a though difficult to date with precision, document question largely forgotten in the recent rush to doc- contact with Crete in the MBA and early LBA, and ument the Minoanpresence overseas. Mycenaeanma- reinforce the picture offered by older discoveries at terial culture was perhaps even more all-pervasivein Alla in the southeast. Discoveries both at Grotta (the the islands than was Minoan.3 But is this to be taken capital, or Chora, of Naxos) and, more surprisingly, as evidence for the settlement of mainlandersin the on the islet of Kato Koufonisi have shown that at a islands?Many of the same issues need to be faced that time contemporary with LM I on Crete, a range of are being confronted in discussion of the Minoan Minoanand mainlandceramics similar to those found Thalassocracy.32The only good stratigraphical se- in the so-called "Western String" was also reaching quence that bridges the earlier and later parts of the the central islands. What is missing at present is any LBA, that reconstructedat Ayia Irini on Keos, does quantified information. Thus, although we now may not suggest any obvious gap in the settlement se- have a somewhat better idea of the "content"and quence at the time when Mycenaeanmaterial culture spatial distribution of the trade network that served came to dominate Minoan. Indeed, at that site, as in the central Cyclades, we still lack any measure of its other Cycladic settlements of the early LBA, there magnitude. Nor is there much evidence that contact had alwaysexisted considerable evidence for contact with Crete had there the same profound effect on the with the Greek mainland. The principal change that material culture of native non-Minoan peoples as at had occurred by the time of LH III was not the Akrotiri,Phylakopi, and Ayia Irini. replacementof Minoan and Minoanizedby mainland Other arguments posed recently in support of Mi- traditions, but rather the subtraction of Crete as a noan political control of the Aegean purport to find significantinfluence on local life-styles. their justification in historical Egyptian documents: As in the case of the Minoan Thalassocracy,in the namely, the description of Crete and the "Islandsin Mycenaeanempire a great amount of diversitycan be the Middle of the Great Green" as a single entity is recognized in the nature of the responses of individ- seen as providing evidence for Minoan hegemony.30 ual islands to contact with Mycenaean mainlanders. Ultimately,however, the conflict that has resulted in In the far north on Thasos, Mycenaeanceramic tra- the division of many Aegean prehistorians into two ditions were adapted and incorporated into local ce- camps-one of Cycladic "nationalists,"the other of ramic repertories, but standard Mycenaean wares Minoan "imperialists"-may be unresolvable.Indeed appear to have reached the island only infrequently. one might well debate the wisdom of using prehistoric Within the Aegean there are differences too in burial archaeologicaldata for the reconstructionof political customs. Most obvious are the large cists used for events. Nonetheless, the islandsof the Aegean present burial in LB III Chios and Psara, a sharp contrast to an extraordinaryopportunity for prehistoriansinter- the chambertombs of the Dodecanese. In Euboiaand ested in the processes of cultural contact and change the northern Sporades,the chambertomb appears to to investigatea wide range of case studies in a variety have been standard.The chambertomb was also pres- of settings. The laboratory-likebenefits of archaeol- ent on Naxos, while on Paros and Tinos, a variety of ogy on islands, in particularthe well-boundednessof tholos tomb was introduced. On Karpathosthe larnax

29 See discussion in J.L. Davis, "'Ener oi5'otOrt Ce Keftiu and the Minoan Thalassocracy," in Minoan Thalas- An'oi6tbag:Thoughts on Prehistoric and Archaic Delos," socracy 197-203. TUAS 7 (1982) 23-33. 31 M. Marthari,"The MycenaeanExpansion in the Cyc- 30 See W.-D. Niemeier, "Mycenaean Elements in the Min- lades," in K. Demakopoulou ed., The Mycenaean World: Five iature Fresco from Thera?" in TAW III.1, 267-82; "Creta, Centuries of Early Greek Culture 1600-1100 BC (Athens Egeo e Mediterraneo agli inizi di bronzo tardo," in M. Mar- 1988) 56-57. azzi, S. Tusa, and L. Vagnetti eds., Traffici micenei nel Med- 32 Perhaps the most insightful discussion of the problem iterraneo: problemi storici e documentazione archeologica (Ta- is that by J.C. Wright, "Umpiring the Mycenaean Empire," ranto 1986) 245-70; and esp. Y. and E. Sakellarakis,"The TUAS 9 (1984) 58-70. 708 JACK L. DAVIS [AJA 96

was employed and burial customs appear to have been have summarizedmost extensively those studies that related to those of Crete. The forms of burial prac- will be least accessibleto readers in North America. ticed elsewhere, including the well-investigated island Presentationin this section follows the current or- of Keos, remain a mystery. More generally, the ab- ganizationof administrativedistricts within the Greek sence of evidence that large numbers of intact Myce- Archaeological Service, beginning with Keos in the naean vases have been looted over the years by first Ephoreiaand concludingwith the 22nd Ephoreia perspicacious archaiokapiloi may in itself suggest that of the Dodecanese and Amorgos. burial in chamber tombs never the acquired promi- Keos nence in many islands that it had in the Dodecanese. Work on Keos in the past decade has concentrated on publishing results of earlier excavationsat the site A REVIEW OF RECENT WORK of Ayia Irini. Seven volumes of the final excavation My specific goal in this section is to summarize report have appeared, all since 1983, giving Ayia Irini recent scholarship that pertains to the islands of the the longest well-documented prehistoric sequence in Aegean. Much of it will be pertinent to issues already the Cyclades.In addition, no less than three separate discussed above. Readers will notice that my definition surface survey projects are providing the details of of "recent" varies from one part of the Aegean to the overall prehistoricsettlement system of the island. another. Even in an extensive format, there remain Habitation at Ayia Irini appears to have begun certain absolute limitations on space, and it makes somewhatlater than at the FN site of Kephalanearby. little sense to consider in detail literature that has EBA levels at Ayia Irini were extensive but remain already passed into more general syntheses and is largelyunpublished.35 Pottery from period I deposits, readily accessible. I, therefore, assume that the reader the earliest at Ayia Irini, is distinctly different from will have ready access to the Gazetteer. For the Bronze that from Kephala and lacks several distinctive fea- Age Cyclades, I have not generally taken into account tures characteristic of that site, including pattern- those publications encompassed by Barber's compre- burnishing and scoops, although there are general hensive synthesis of Cycladic prehistory, The Cyclades similaritiesin shapes, and the wares are still within a in the Bronze Age. For other islands, I have attempted Neolithic tradition. After period I there appears to to provide an overview of significant research pub- be a break in the sequence, inasmuch as period II lished since 1980.33 I have paid most attention to new represents a fully developed EB II phase of occupa- fieldwork, relatively little to art historical matters. In tion, distinguished by close relations with eastern At- the case of Early or the Thera frescoes, tica.36Period III is characterizedby the introduction the literature is so vast that it warrants separate re- of Anatolianizingceramic shapes (constituting about views.34 My references to secondary literature attempt 10% of the pottery in assemblages),at least some of only to epitomize a portion of recent scholarship, but which were locally produced; their introduction was I hope that I have provided readers with sufficient not marked by discontinuitiesin the life of the settle- clues to allow them to find their way into more spe- ment, and continuitywith previous traditionswas the cialized publications, should they desire to do so. I rule.37 In neither period is there evidence for any

3" Since manypublications in Greece,particularly jour- 35 All finds earlier than the MBA will be published by nals,have often been publishedmany years later than their D.E. Wilsonand M.E. Eliot in AyiaIrini: PeriodsI-III (Keos coverdate, in composingthis reportI decidedto consider IX, in prep.). I am grateful to David Wilson for the of oppor- reports fieldworkfor inclusionif theywere summarized tunity to read and to summarizebriefly parts of his working in AR 27 (1981)or "Chronique"1981 or later,even in cases manuscriptfor the volume. wherethe originalreport abstracted had been composed 36 D.E. Wilson, "Keaand East Attica in Early Bronze II: manyyears prior. Beyond PotteryTypology," in J.M. Fossey ed., ZvvELtUoQa 34 Fortunately,excellent recent syntheses are readily avail- McGill: Papers in GreekArchaeology and Historyin Memory able.See esp. n. 19, supra,as wellas P. Getz-Preziosi,Early of ColinD. Gordon(Leiden 1987) 35-49. Artin NorthAmerican Collections Cycladic (Richmond1987); 37 Wilsonargues for contemporaneitybetween Ayia Irini Getz-Preziosi,Early Cycladic Sculpture: An Introduction (Mal- II and the "green"period at Poliochni; and between ibu Ayia 1985);J.L. Fitton, Cycladic Art (Cambridge, Mass. 1990); Irini III, the "red"and "yellow"phases of Poliochni,and, in J.E. Coleman,"'' of the EarlyBronze Age Ae- part, Emborio I. Wilson demonstratesthat the 89 convincingly gean,"AJA (1985)191-219; C. Renfrew, "The Goulandris closest parallels for the Anatolian-stylepottery from Ayia Museumof Cycladicand AncientArt," AR 32 (1986)134- Irini III are to be found in southwesternAsia Minor. The and L. 41; Goodison,Death, Women, and the Sun (BICS most complete published summary of evidence relevant to Suppl.53, London1989) ch. 1. Thesepublications are rich later EBA Keosis now: D.E. Wilsonand M. Eliot, Irini, with referencesto "Ayia earlierstudies of Cycladicart. For the Period III: The Last Phase of Occupation at the E.B.A. Therafrescoes and other contemporarywall paintings from Settlement,"in PrehistoricCyclades 78-87; see also Vaughan the islandsof the Aegean,see discussionsof Keos,Thera, and Wilson (supra n. 20). Melos,and Rhodes,below. 1992] THE ISLANDS OF THE AEGEAN 709

(of large roughly squared limestone blocks with rec- tangular towers) marks the beginning of period V, which must be on the basis of Minoan imports roughly contemporary with MM IIB/MM IIIA. As in the EBA, MBA Keos was closely linked through exchange to other areas of the Aegean. Ties to Crete are evident already from the beginning of period IV, although 4')0l local ceramic traditions are dominant (fig. 3). In pe- riod V, there is evidence for the local use of the Cretan linear script and for Cretan-style administrative prac- tice, as represented by a roundel of Minoan type and an inscribed tablet fragment. Aspects of material cul- ture became increasingly Minoanized in the course of the MBA (and in the earlier stages of the LBA) as local potters copied Cretan shapes and many elements 2. Irini, Keos. Stone and beads as a Fig. Ayia gold strung of Cretan technology were introduced to the island, necklacefrom MBA grave 24. (CourtesyJohn C. Overbeck) including a Minoan system of metrology. Crete influ- enced local religious practices, and Minoan-style wall direct contact with the wares ; imported paintings were executed locally (figs. 4-5).39 are recognized from period I, including probable Melian products, and quantities of imported ceramics reached as high as 30% in period II. It is indeed increasingly clear that exchange played a significant 'N role in the life of the settlement from the time of its ...I. foundation. The Middle Bronze Age at Ayia Irini has been divided into two principal periods, IV and V, the former subdivided into three subphases. The evi- dence for both periods IV and V has now been fully presented.38 There was a break in occupation at Ayia Irini after period III. The earliest period IV deposits i iiiiMA, at the site are appreciably later than the beginning of the MH period on the Greek mainland and it is clear that a gap exists in the Ayia Irini sequence roughly contemporary with EH III and the earlier stages of the Middle Bronze Age. The cemeteries of period IV are among the very few cemeteries known in the MBA Cyclades (fig. 2). From the beginning of period IV :IN, the settlement was fortified, its principal gateway guarded by a horseshoe-shaped tower. This earliest defensive system was destroyed and eventually re- placed by a system set farther north on the neck of Fig. 3. Ayia Irini, Keos. Fragmentof an MBAbarrel-jar with the peninsula to encompass a somewhat larger area red and black matt-painteddecoration. (CourtesyJohn C. within the town. The construction of this new system Overbeck)

38 J.C. Overbeck, Ayia Irini: Period IV: The Stratigraphy spread; see A.H. Bikaki, Ayia Irini: The Potters' Marks (Keos and the Find Deposits (Keos VII.1, Mainz 1989); J.L. Davis, IV, Mainz 1984). For the full corpus of lead weights (from Ayia Irini: Period V (Keos V, Mainz 1986). Further discussion whichit has been deduced that a Minoanmetrological system and documentation of the material of period IV will be was used locally),see K.M. Petruso,Ayia Irini: The Balance included in the second part of KeosVII, now in preparation Weights. An Analysis of Weight Measurement in Prehistoric byJ.C. Overbeck. Crete and the Cycladic Islands (Keos VIII, Mainz 1992). Tech- 39 The context of the tablet with Linear A signs found in nological innovationsof Minoan type include the so-called a stratumof period V has been discussedby Davis (supran. "fireboxes,"which arguablywere employed for the produc- 38) 99 (also GORILAI, KE 1); for the roundel, GORILAII, tion of aromatics,and potter'swheel disks. These and other KEWc 2. Signs adopted from the Cretanscript replaced the standardizedceramic shapes have been discussed by H.S. older systemof marksthat had been used sincethe beginning Georgiou, Ayia Irini: Specialized Domestic and Industrial Pot- of period IV when marking pottery had first become wide- tery (KeosVI, Mainz 1986). Other evidence of metals and 710 JACK L. DAVIS [AJA 96

t.-??LL??-- 1 ,?..,.,...... ,...~...... ~~~~~

i ii ! '.~-- ! i I! i:i ' - i!...... !

y I -. '- ...... ' ' " . " . "' • ...... ! • ...... '...... • ...... L, . ., ...... " .. . -..!.-...... ,,. ....,k .

Fig. 4. Ayia Irini, Keos. Detail of a miniaturepainting: ships, men, and cauldrons. (CourtesyLyvia Morgan)

As yet no overview of the settlement in any period divisions of period VI are followed by the three sub- of the LBA has appeared in the series of final exca- divisions distinguishable in period VII. An early vation reports. Deposits from house A, published in phase (VIIa), which bridges the gap between periods 1983, continue to provide the most detailed picture VI and VII, is marked by the first importation of LM of the nature of the material culture at the site in IB/LH II ceramics to the site, and is earlier in date periods VI and VII.40 The stratigraphy of these pe- than the main destruction deposits of house A. These riods is now, however, understood in considerable are in turn assigned to a middle phase of period VII detail as is the architectural development of the site. (VIIb) and are characterized by the appearance of the Two principal subgroups of material have been rec- Cretan Marine Style and the so-called Alternating ognized within period VI: an earlier group that con- Style. A late phase follows the destruction of house A tains LM IA types but in which the LH I style is not but is still contemporary with the latest part of LH II represented, and a later one in which the LH I style, (VIIc). In the early LBA, the area of houses C, F, and the LM IA style, and mainland matt-painted types EJ seems to have consisted of a few independent contemporary with LH I are represented.41 These establishments with large open spaces between them, metalworkingat both Ayia Irini and Kephalahas been sum- 41 J.L. Davis and J.F. Cherry,"Spatial and Temporal Uni- marized recently by Z.A. Stos-Gale,"Lead Isotope Evidence formitarianismin Late Cycladic I: Perspectivesfrom for Trade in Copper from during the Late Bronze and Miloson the Prehistoryof Akrotiri,"in TAW111.1, 185- Age," in GreekPrehistory 265-82, esp. 276, 282, and fig. 13; 200. The latest of the two period VI subphases,best repre- and by Z.A. Stos-Galeand N.H. Gale, "The Role of Thera sented in deposit A of room 18 in house A, cannot,as recently in the Bronze Age Trade in Metals,"in TAWIII.1, 77, and has been argued by Warren(infra n. 152), be the result of a figs. 6-7 on 78, 84. For a general overview of the contexts destruction that occurred contemporarilywith the Seismic of industrial activities at Ayia Irini, see E. Schofield, "Evi- DestructionLevel (SDL) at Akrotiri (on these deposits, see dence for Household Industrieson Thera and Kea,"in TAW the discussion of Thera infra). The relevant deposits at III.1, 201-11. The more than 30 near-life-size terracotta Akrotiri produced no trace of the Late Helladic I style and figures of Minoan style found in the temple at Ayia Irini were presumablydeposited beforeits inception on the Greek have now received their definitive publicationin M.E. Cas- mainland,while the LH I style is clearly represented in the key, The Temple at Ayia Irini: The Statues (Keos II.1, Prince- deposit in house A. The house A deposit must, therefore, ton 1986). On the wall paintings from Ayia Irini, see E.N. be later in date than the SDL at Akrotiri;its overallcharacter Davis, "The CycladicStyle of the Thera Frescoes,"in TAW is, in fact, more similarto that of the Volcanic Destruction III.1, 214-27, and L. Morgan,"Island Iconography: Thera, Level (VDL)at Akrotiri(particularly in terms of the suite of Kea, ,"in TAWIII.1, 252-65. Of particularinterest is mainlandimports represented)than it is to the SDL. There Morgan's reconstructionof fragments from the northeast can be little question that pottery of the LH I style and of bastion in the fortificationsof Ayia Irini as parts of a min- contemporarymatt-painted styles reached the southernCyc- iature fresco, thematicallyrelated to the miniaturesof the lades since mainland imports are well represented in LC I West House at Akrotirion Thera. levels at Phylakopion Melos, and are found already in the 40 W.W. Cummer and E. Schofield, Ayia Irini: House A SDL at Akrotiri (infra n. 153). Once again, the danger of (KeosIII, Mainz 1983). drawing suppositiouscorrelations needs to be emphasized. 1992] THE ISLANDS OF THE AEGEAN 711

--...... ??------?------?------?---

."-......

...... - J .r• .. r

; i

. - .. . - ;.; ...... " ......

dip

.....r4•'~rr -"-"'':...:• ...... • .. .-..

------

crn

Fig. 5. Ayia Irini, Keos. Detail of a miniature painting: town by a river. (Courtesy LyviaMorgan) in contrast to the impression given by published plans different parts of the site indicate continuation of of Ayia Irini, which show walls of all phases of the settlement in LH IIIA:2 as do remains of structures LBA, whether contemporary or not.42 built outside the fortifications near the former (period Systematic investigation of the LH III settlement VII) main gateway to the town. Continued use of has emphasized that reoccupation following the main burnished and matt-painted wares in LH IIIA: 1, and period VII destruction was substantial.43 There seems of conical cups and tripod cooking pots, attests to the to have been no gap in occupation at the site after survival of these types into LH III alongside Myce- period VII and a sequence of deposits following the naean types proper; there is strong continuity in local general destruction of the town in period VIIb and ceramic traditions. contemporary with LH IIB and LH IIIA:1 on the One objective of a recent systematic surface survey Greek mainland can be distinguished stratigraphi- (1983-1984) in northwest Keos was to determine the cally. LH IIIA: 1 material from most parts of the site character of settlement and land use in the immediate appears to have been laid down as the result of an- hinterland of Ayia Irini.44 Results suggest that Keos other widespread destruction. Deposits from several may have followed a rather different pattern of de-

42 For preliminaryobservations on the characterof period Cyclades241-54. My summaryhere is drawn from the text VII subphases, see E. Schofield, "Ayia Irini, Keos, in Late of an unpublished paper entitled "Ayia Irini, Keos: Late CycladicII," BICS 32 (1985) 155; and E. Schofield,"Destruc- HelladicIII," by C. Morrisand C. Hershenson.I am grateful tion Depositsof the EarlierLate Bronze Age from Ayia Irini, to both of them for a copy of it and for allowing me to Keos," in Prehistoric Cyclades 179-83. summarizetheir conclusions. Remains of VIII the LH III 43 period (i.e., settlement) 44 J.F. Cherry,J.L. Davis, and E. Mantzourani,Landscape are numerous. No general discussion has yet appeared in Archaeology as Long-Term History: Northern Keos in the Cy- print, although the sequence of LH IIIC phases from the cladic Islands (Monumenta Archaeologica 16, Los Angeles temple has been described and some representativepottery 1991). Chs. 6-9 discussresults of the survey pertinent to the illustrated;see M.E.Caskey, "The Temple at Ayia Irini, Kea: prehistoryof the island and review evidence resulting from Evidence for the Late Helladic IIIC Phases,"in Prehistoric all earlier work on the island. 712 JACK L. DAVIS [AJA 96 velopment in the EBA than did many other Cycladic while Kephala was certainly a village of considerably islands. The population of the island (at least of its smaller scale. Two new small aceramic sites in the northwest part) appears to have been concentrated at northwest part of the island near Kephala have been the site of Ayia Irini alone and there is no Keian assigned to the later Neolithic period on the basis of parallel to the scatter of small EC settlements and lithic attributes.46 cemeteries recognized on so many other islands. Evi- Mineralogical field study has succeeded in relocat- dence available from two other surveys that have ing sources of lead ore, long known to exist on the ranged more extensively supports our hypothesis that island and mined for a short time in the later 19th the settlement at Ayia Irini dominated the settlement and early 20th centuries. The deposits are distinguish- system of Keos as a whole.45 Other sites are both tiny able from those of Laurion on the basis of their lead- in size and lack evidence for the imported ceramic isotope ratios and are low in silver content. There is wares that bear such vivid testimony to the interaction as yet no firm evidence that they were exploited in between Ayia Irini and the outside world. No pottery prehistoric times.47 earlier than that at and Paoura has been Kephala The Northern Sporades nor has located material that recognized, survey might The principal development in the prehistory of the be to close in the Irini employed gaps Ayia sequence northern Sporades has been the publication of results between I and II, and between III periods periods of excavations at the site of Ayios Petros.48 The site and IV. Finds of the late Middle and Late early Cy- lies inside a large bay on the southwest side of the cladic are most but these are almost periods plentiful, islet of Kyra Panayia. In the Neolithic Ayios Petros the remains of and other exclusively large storage jars was attached to the adjacent mainland of Kyra Pan- coarse wares that were in rural perhaps employed ayia, but even during the Palaeolithic there appears settings by a population largely resident at Ayia Irini to have existed a sea lane between it and the island of itself. LB III finds have been especially scarce. Alonnisos to the southwest.49 Most of the site is now results have cast most on Neolithic Survey light submerged but even originally it seems to have been Keos. Detailed collection of surface materials at less than two-tenths of a hectare in area. Stratification Paoura and Kephala has allowed the size of those sites was shallow (only a little more than a meter in depth) to be estimated with greater precision. Analysis of and much eroded but nonetheless excavation pro- lithics from both suggests that Paoura's closest ties are duced conclusive evidence that the principal period with Saliagos; the assemblage of Kephala is distinct of occupation was contemporary with the latter part from that of Paoura and its lithics have strong affini- of the EN period and the beginning of the MN period ties with those of mainland sites. Paoura may well in Thessaly; 14C dates of 6740 + 120 B.P. and 5860 + have supported a population of 75-130 individuals, 400 B.P. have been published.5? Some LN, EBA (of

45 Thus far the results of extensive survey by H.S. Geor- eds., The Engineering Geology of Ancient Works, Monuments giou and N. Faraklasin the northern part of the islandhave and Historical Sites: Preservation and Protection (Rotterdam been published;see H.S. Georgiouand N. Faraklas,"Ancient 1988) 1739-45. HabitationPatterns of Keos: Locationsand Nature of Sites 48 Most recently Neolithic remains have also been re- on the Northwest Part of the Island,"Ariadni 3 (1985) 207- ported in the KyklopiCave on Yioura (the most remote of 66. Survey teams from the Universityof Athens have oper- the northern Sporades); see "Travel Guide," Archaeology ated principallyin the territories of the Classicalcity-states 45:2 (1992) 7. of Poieessa and Karthaia.For specificallyprehistoric finds, 49 N. Efstratiou, Ayios Petros: A Neolithic Site in the North- see G. Galani, L. Mendoni, and H. Papayeoryiadou,"EitL- ern Sporades(BAR-IS 241, Oxford 1985). Survey of Kyra aLveClaxl EQUvva oTilv K~ca,"Arhaiognosia 3 (1982-1984) Panayia by Efstratiou revealed no additional prehistoric 237-44. finds; see also "FirstColonization" 167, for further discus- 46 L.E. Talalay,"Body Imagery of the Ancient Aegean," sion of sea level fluctuationsand of the possibilitythat the Archaeology44:4 (1991) 46-49, has also discussed the char- northern Sporadeswere tied to each other and the adjacent acterof settlementat Kephala,the controlof local resources, mainlandduring the Palaeolithic. and the relationshipbetween cemeteriesand territorialdef- 50 Efstratiou(supra n. 49) 167, appendix V; S.G.E. Bow- inition. For the most authoritativerecent discussionof Ke- man,J.C. Ambers,and M.N. Leese, "Re-evaluationof British phala and nearby sites, see R. Torrence, "The Chipped Museum Radiocarbon Dates Issued between 1980 and Stone," 173-98, and T.M. Whitelaw,"Investigations at the 1984,"Radiocarbon 32 (1990) 59-79. Efstratiouargues that Neolithic Sites of Kephalaand Paoura,"199-216, in Cherry there are parallelsat Ayios Petroswith ceramics of the second et al. (supra n. 44). phase of Thessalian EN (Achilleionphase). These elements 47 "Sourcesof Metals"257-59; Stos-Galeand Gale (supra could not, however, be defined at Ayios Petros as a distinct n. 39) 86; and Pernicka (supra n. 17). These assessments stage in the life of the settlement, but were found together contrastwith views expressed in M.E. Caskeyet al., "Metals with ceramicelements of the sort that characterizethe classic in Keos:A FirstApproach," in P.G.Marinos and G.C.Koukis MN Sesklo phase in Thessaly. 1992] THE ISLANDS OF THE AEGEAN 713

Troy I character), and MBA (Gray Minyan) pottery manent, rather than a seasonal, occupation; if this in surface levels and in a pit reflect later use.51 Pa- interpretation is correct, Ayios Petros would be the laeolithic artifacts have been reported at Ayios Petros earliest published permanent settlement yet known in and at several locations on Alonnisos. the Aegean islands. Stone foundations for structures with both rectan- gular and curving walls appear to have supported Euboia and Skyros mudbrick superstructures. Two child burials were Euboia. Our picture of earliest settlement on Euboia excavated amidst the habitation levels while other has been greatly illuminated by the publication of the human bones were found scattered. Artifactual evi- results of recent surveys and excavations designed to dence does not suggest that those who used the site supplement older investigations by Theoharis and by of Ayios Petros were particularly isolated from adja- members of the British and American Schools at Ath- cent mainlands. Nearly all of the pottery appears to ens.52 More than 60 Neolithic and EB I sites are now have been produced locally, but stylistic affinities in- known, for the most part small settlements of limited dicate continuing contact with the Thessalian main- duration. Analysis particularly of finds from the sites land. Most distinctive are carinated bowls with red- of Psahna Varka, Psahna Glyfa, Psahna Votsika, patterned decoration on a creamy ground. Imported Seinen Mnima, Politika Spilaio Marmara, Karystos materials include Melian obsidian and flint that ap- Plakari, and Chalkis Manika have allowed the earlier pears to have been Thessalian in origin. The stone prehistory of the island to be divided into six tentative used for axes, on the other hand, was probably of phases: an Older (AQXat6lTQr) Neolithic I and II local derivation, quarried on the island of , (roughly coincident with EN on the mainland); a to the north of The extensive collection Newer Neolithic I, approximately equiva- Kyra Panayia. (NE•rrEQrg) of terracotta figurines (50 in number) from the site is lent with the earlier phases of the mainland LN; two noteworthy given the fact that most are from secure stages of Final Neolithic (the latest equivalent to the contexts; they include male as well as female and Attic-Kephala culture); and EH I.53 No Euboian phase animal types. Features are for the most part rendered corresponding in character to mainland MN has yet with deep incisions, some with white filling. Animal been noted. Each of these stages is defined with ref- bones include a full range of domesticates. It is con- erence to characteristic ceramic and lithic artifacts, vincingly argued that the site was most likely a per- and to patterns in settlement.54 Surface survey has

51 Efstratiou(supra n. 49) 166, appendix IV, reports pro- phalaon Keos,and Neolithicfinds from the AthenianAgora. visionalresults of spectroscopicanalyses of samples of Gray Mostof these Euboiansites appear to have been settlements, Minyanpottery that suggest that, in their composition,they although human bones from PolitikaSpilaio Marmaraare resemble most closely the of Volos in of the Older Neolithic II phase and presumablyderive from Thessaly. Y. Liritzis, L. Orphanidis-Georgiadis,and N. Ef- a burial. Older Neolithic I sites are poorly known and the stratiou,"Neolithic Thessaly and the Sporades:Remarks on characterof the ceramicssuggests isolation from mainland CulturalContacts between Sesklo,Dimini, and Aghios Petros traditions. Based on Trace Element Analysis and ArchaeologicalEvi- Pre-Neolithic finds have also been reported on Euboia; dence,"OJA 10 (1991) 307-13 found, however,no evidence see Neolithicand Protohelladic23; "FirstColonization" 165- that Neolithic pottery from Ayios Petros had been imported 67. For other recent reports of pre-Neolithicfinds from the from Thessaly. island, see A. Sampson,"a-IclcEica AyLwvvag," ArchDelt 38 52 A. Sampson,Neolithic and Protohelladic;see also Samp- B' (1983) 154 (Middle Palaeolithicat AiyiannaSarakeniko); son, ArchDelt 37 B' (1982) 174 ArchDelt38 B' (1983) 155 stone "Q•eol: TactLdXYlg," (pre- "KeQcao(La," (pre-Neolithic historic remains reported at the site of Panayitsa, near tools at Kerasia Panayia and Kerasia Nero); E. Sarantea, Oreoi); P. Kalligas, "EQgTQLa(Ev~giUCQ Ma- tEQLtoXfl): "EQyaoT'iltLC xcTccraoxe••Ig aIItXaLOXLOLXc0V gyckXetLv ArchDelt 36 B' (1981) 201 (a prehistoric settle- XchLovE4~Taov MEaxQ XacXWvag," orTl JTEQLtoxflorTl XoV Vxd;og otrlV ment at Plakakia in the vicinity of Eretria); A. Sampson, EV'3otLa,"AAA 18 (1985) 81-85; and E. Sarantea, I00o- "Kcaortt-hoh JLoVa," ArchDelt38 B' (1983) 153 (Neolithic io'roptxad Evptyara Ngag; Aprdtxrq E2fiotag (Athens finds from excavationsat KastelliPisonas, on a summitover- 1986). looking the Lelantine Plain); "ZovyxXhLxaAytvvrag," 54 Representativespecies of shell in each phase are de- ArchDelt38 B' (1983) 154 (Neolithic axes and terraces at scribed (Neolithicand Protohelladic47-49): two pierced pec- ArchDelt38 B' 154 ten from Politika Marmara were used as Ayianna Zougkleika);"KaLt~utg," (1983) jacobeus Spilaio (Neolithic pottery from caves). jewelry. Murex shells were found at VasilikoLinovrohi. Flint 53 Sampson, Neolithic and Protohelladic,draws parallels tools are abundant, but obsidian relativelyrare before FN. between the materialfrom these phases and specificgroups Other artifactsof special interest include Newer Neolithic I of finds from the Greek mainland or elsewhere in the Ae- fragments of schematic human figurines and animal figu- gean: Older Neolithic II is compared with material from rines from PsahnaVarka (Neolithic and Protohelladic84-85) Nea Makri;Newer Neolithic I, with the Arapi and Tsangli and FN bases with matt and cloth impressions from Thar- phases of Thessaly, with the LN I of Corinth, and with rounia SkoteiniCave (Neolithicand Protohelladic144). Saliagos on Antiparos; and Final Neolithic (FN), with Ke- 714 JACK L. DAVIS [AJA 96 been most extensive in central Euboia and it is not mains have been discovered in several rock-shelters. surprising that the area around Eretria and Chalkis West of Karystos, FN finds have been recorded at two is best represented by newly discovered sites. The locations: a small site on the summit of the Kazara greatest numbers of sites appear to belong to the Final ridge and a larger settlement on the ridge of Plakari, Neolithic and EH I periods; these seem to have been both in defensible positions a short distance from the times of settlement expansion. All parts of the island sea. Limited salvage excavations at Plakari have un- were inhabited, including inland and coastal locations covered architectural remains with pattern-burnished in the north and south.55 For the EBA as a whole pottery similar in character to that from Kephala on more than a hundred sites are now known from ex- Keos.59 In contrast, several EBA sites were located on cavation or surface investigations.56 As for the Neo- the shore of the peninsula: Ayia Pelayia on lithic, these are concentrated in the central parts of the northeast side,60 Ayia Paraskevi at the south end, the island. and Akri Rozos on the northwest. On the east side of The most intensively surveyed part of Euboia is the the bay of Karystos another EBA site was found at area around the bay of Karystos in the south.57 The Ayia Irini. entirety of the Paximadi peninsula southwest of Kar- Finds of the MBA and LBA in contrast are relatively ystos has now been investigated, and 19 concentra- scarce here in the extreme south of Euboia. A large tions of prehistoric material have been recognized. site was established in the MBA to the northeast of All belong to the Final Neolithic and early part of the Karystos at Ayios Nikolaos; finds include plentiful Early Bronze Age. The focus of settlements appears evidence for metalworking, Gray Minyan and matt- to have been in the north of the peninsula, where flat painted pottery, bronze artifacts, and a lead pottery land and water are more accessible. The earliest re- clamp. The settlement was apparently not inhabited mains yet located in southern Euboia belong to the in the LBA and as yet there have been no indisputable Late Neolithic and come from the cave of Ayia Triada, LBA finds reported in southern Euboia.61 north of Karystos; they include obsidian and black- In addition to the preceding general surveys and burnished pottery.58 Fragments of human bone sug- trials, more extensive excavations have been pursued gest that the cave may have been used as a place of at several locations. The most important of these have burial, while just below it LN and FN habitation re- been at the apparently massive EBA site of Chalkis

55 The distribution of sites by date (as summarized in 1985). I am grateful to Donald Keller for a copy of his Neolithic and Protohelladic) is as follows: Older Neolithic I, dissertation. 9; Older Neolithic II, 8; Newer Neolithic I, 15; FN I, 4; FN 58 See also Neolithic and Protohelladic 92, 145. II, 18; EH I, 23. These figures are now supplemented by 59 Pottery from the excavations at Plakari has been pub- more recent data, particularlythat from survey in southern lished by D.R. Keller, "Final Neolithic Pottery from Plakari, Euboia(infra ns. 57, 61). Karystos," in P. Spitaels ed., Studies in South Attica 1 (Mis- 56 Manika I, 334-76 summarizes much of the evidence cellanea Graeca 5, Ghent 1982) 47-67. The presence of and includescommentary on a catalogueof EBA sites;many copper ore from the site is noted by Keller and Cullen (infra new sites have been identified since earlier investigationsby n. 61). All obsidian probablyoriginated on Melos and the Theoharis and by members of the British and American raw material does not seem to have been used in an econom- Schools. See also a report of EH II remains at Drosia Gai- ical fashion. Unworked nodules of obsidian appear to have daros: A. Sampson, "AQoomd,"ArchDelt 38 B' (1983) 154. been imported directly from Melos and are most plentiful Kalligas has provided a detailed review of prehistoric re- on Cape Mnima,at the southern tip of the peninsula. search on the island before the 1980s: P.G. Kalligas, " 60 Highly diagnostic EB II types, including sauceboats, and the Cyclades," in Cycladica 88-98, and publishes a num- found at Ayia Pelayia, represent the latest stage of Bronze ber of older finds including a marble figurine from Makro- Age occupation yet recognized on the Paximadi peninsula horifo, just south of Chalkis; he also discusses various (Keller and Cullen, infra n. 61). Neolithic and EBA artifacts thought to come from the vicin- 61 Since 1984, the Canadian Karystia Expedition has com- ity of Makrikapa, in east-central Euboia (including gold and pleted survey of the Paximadi peninsula and parts of the silver vessels); most are now in the collections of the Chalkis southeastern coast of the island. Prehistoric finds have been Museum, the Metropolitan Museum of Art, and the Benaki reported but not yet described in detail. See D.R. Keller and Museum (see E. Davis, The Vapheio Cups and Aegean Gold M. Wallace, "The Canadian Karystia Project, 1986," EchCl and Silver Ware [New York 1977] 63-65). For general trends 31 (1987) 227; Keller (supra n. 57) 237 n. 5.6; and Keller in the history of MBA and LBA settlement on Euboia, see and T. Cullen, "Prehistoric Occupation of the Paximadhi Manika I, 336, 342. Peninsula, Southern Euboea," AJA 96 (1992) 341 (abstract). 57 See D.R. Keller, Archaeological Survey in Southern Eu- I am grateful to Keller and Cullen for a copy of the full text boea, Greece: A Reconstruction of Human Activity from Neo- of their paper. lithic Times through the Byzantine Period (Diss. Indiana Univ. 1992] THE ISLANDS OF THE AEGEAN 715

- -i-~l-_---'---:':-m ::-m::

Fig. 6. Skoteini Cave, Tharrounia, Euboia. Fragments of figurines. (Courtesy AdamantiosSampson)

Manika, at Tharrounia Skoteini Cave, the largest known cave on the island, and at Classical Eretria. The existence of prehistoric remains in the Skoteini Cave has long been known. Excavations since 1986 have, however, uncovered stratified deposits, which, when published, promise to clarify considerably the Late and Final Neolithic sequence of Euboia.62 The earliest traces of human presence can be dated to Newer Neolithic I in the Euboian sequence (equiva- lent to early LN on the mainland); higher strata are of Final Neolithic, Early Helladic, and LH III date. Finds of exceptional interest include an EH clay seal- 4'I ing and a pithos handle that had been stamped re- peatedly with a seal; various EH metal tools and items of jewelry; copper tools and needles from the FN levels; Neolithic stone and clay figurines, including steatopygous females and animal figurines (fig. 6); an image of an ithyphallic man accompanied by a naked woman, both rendered in relief on a Neolithic pithos; and well-stratified deposits of Neolithic pottery (fig. 7). Human bones were mixed within the Neolithic Fig. 7. Skoteini Cave, Tharrounia, Euboia. Neolithic vase levels; the discovery of a skull separated from the rest with incised decoration.(Courtesy Adamantios Sampson)

62 Few detailsconcerning the excavationshave as yet been report, kindlyprovided by the excavator,Adamantios Samp- published; see AR 33 (1987) 15; "Chronique"1986, 562; son. 1987, 672. My description is summarized from a longer 716 JACK L. DAVIS [AJA 96

?r:--*

Ni?:-:_o-I__l:- ?s.i ?:::::-

...... ::i~:::-::::::_:~!i~i,;';l:-i~ii:-___--:-_ :::;40i'--: ::6J,- _: _

:?::oz

Fig. 8. Manika,Euboia. Foundations of EBA buildingsin Zousi plot. (CourtesyEfi Sakellaraki) of the skeleton seems to point to the practice of sec- investigations have focused on the EBA community ondary burial. Some 400 m from the cave the remains although it is clear that the settlement continued to of a badly eroded Neolithic cemetery has been exca- be inhabited in the MBA and LBA (fig. 8). The earliest vated, including eight slab-built graves (four intact); strata are dated to a stage late in EH I. Most architec- multiple burial of skulls suggests the practice of sec- tural remains are contemporary with EH II on the ondary burial in two of these. Remains of an open- Greek mainland. The final major stage of EBA oc- air FN settlement have also been excavated on a pla- cupation, less widespread than the preceding EH II teau above the cave. phase, belongs to the so-called Lefkandi I phase and The site of Manika, north of the Euripos near is characterized by Anatolian elements in ceramics Chalkis, has been the target of especially intensive and metallurgy; after this the settlement appears to investigation. Although long known, only in the have been further reduced in size. The cause of this course of the past decade has the true extent and diminution is unclear: no obvious evidence for de- importance of Manika begun to be clarified.63 Recent struction either by human or natural forces has yet

63 For earlier research see Kalligas (supra n. 56). The in 1984). Manika II contains an overview of recent work at main recent bibliographyincludes Manika I-II, principally the site and detailed plansshowing the locationsof the many the results of research in 1982-1986; fieldwork in 1983- test excavations.There is also an extensive discussionof the 1984 has also been briefly describedby A. Sampson and E. ground stone industryin Manika II, 80-104. Sakellaraki,"Mdvtxa," ArchDelt 38 B' (1983) 139-40, and The extremelydetailed presentationof excavationsin the by Sakellaraki,"MdXvtxa," ArchDelt 39 B' (1984) 120-23. Zousis plot describescomplexes of rectangularand apsidal For more detailed presentationsof results, see: E. Sakellar- rooms, some of which appear to have had two stories (122- aki, "NWotLT0ot onl Mdvtxa Xakxl&ag (otx6~6bo fIaMa- 24, 131-32). The pottery has been extensively described. ozractCTLov)," ApXEiovEvfloi•xiv MEAEdv 27 (1986-1987) Extraordinaryfinds included a bronze ring with incised 5-21 (the excavationof tombsin 1986-1987); E. Sakellaraki, decorationon its bezel (109, 224-26, no. 425), a terracotta T "teddy-bear"(140, no. 9), stone vessels (217-18, nos. 388- "Mdvtxa XakCX &ag. :ITQWtaLoyQDtaLxl eQvva oroTO OLXLto~6:OLX636Tbo Zoorlq," ArchDelt 41 B' (1986) 101-270 92), stone pestles (218-19, nos. 393-94), a stone pebble (a report on severalstructures and a discussionof the settle- figurine (141, no. 9), a terracottaquadruped figurine (140, ment in toto); "Ano6nI1 Ep3ota xcLatLIn x6Qo," AAA 19 no. 10), two bronze chisels (224, nos. 423-24), a stamp seal (1986) 30-35 (excavationof graves and soundings within (266, no. 427), and a leg from a terracottaanthropomorphic the settlementin 1984-1985); and "NewEvidence from the figurine (266, no. 428). Few terracottawhorls and no loom- Early Bronze Age Cemetery at Manika, Chalkis,"BSA 82 weights were found. Ground stone and obsidian chipped (1987) 233-64 (a full publicationof eight tombs excavated stone tools were common (220-24, 226-34). 1992] THE ISLANDS OF THE AEGEAN 717

Fig. 9. Manika,Euboia. Fragment of with impresseddecoration. (Cour- tesy Efi Sakellaraki) been identified. There had previously been a continu- 1985, its maximum extent was estimated at 45-50 ha, ity in architecture, ceramics, and burial customs at but this estimate has recently been revised upward to Manika; the settlement of the Lefkandi I phase did about 80 ha.64 From the results of augering and re- not represent a major break with previous traditions. mote sensing, it has been suggested that areas situated Indeed, Anatolian types are most often found in fu- between the various soundings that define the maxi- nerary contexts (and even there are not common); mum extent of the settlement were compactly settled. they, like Cycladic imports in earlier stages of the Even if we may assume, as seems likely, that settlement EBA, appear to have been selectively chosen as grave was not contemporary in all parts of the site, Manika gifts because of their prestige value as exotica. More must have been one of the largest settlements in the common imported goods include Melian obsidian, Aegean in the EBA; it is certainly the only pre-palatial and andesite, which was used for the manufacture of island settlement with a claim to rival contemporary a substantial portion of all ground stone tools. Anatolian and Near Eastern centers in its scale.65 It The biggest surprise of recent years has been the has also been argued that the settlement from its initial suggestion of the enormity of the size of Manika. In stages was, in some sense of the word, "organized":

64 See H. Marukian,E. Kambouroglou,and A. Sampson, mainland sites are generallysmaller than 4 ha (see D. Kon- "CoastalEvolution and Archaeology North and South of sola, H aOrtmozroolo0 rod)tt,1kP oarovg wro7oeAAa6txOog Chalkisin the Last 5000 Years,"in A. Rabaned., Archaeology otxtaroPgo[Athens 1984] 94-102). In the Cyclades, settle- of CoastalChanges (BAR-IS 404, Oxford 1988) 71-79. See ments were still smaller(C. Broodbank,"The Longboatand also Sakellaraki,ArchDelt 41 (supra n. 63) 103, regarding Societyin the Cycladesin the Keros-SyrosCulture," AJA 93 the extent of the settlement. [1989] 319-37). Nor were there habitationsin the northern 65 Estimatesfor sizes of manyprincipal EBA sites on Crete Aegean that approached Manikain size (C. Renfrew, The are an order of magnitude smaller (T.M. Whitelaw,"The Emergenceof Civilisation[London 1972] 236-44). For esti- Settlementat Fournou Korifi, Myrtosand Aspects of Early mates of sizes for Near Easternand Anatoliansites, see also Minoan Social Organization,"in O. Krzyszkowskaand L. Renfrew. Nixon eds., Minoan Society[Bristol 1983] 339). The largest 718 JACK L. DAVIS [AJA 96 that all architectural remains thus far explored are oriented to the cardinal points and take their cue from similarly aligned streets.66 The town may also have been fortified.67 The cemetery of Manika lay to the south, adjacent to the settlement; its maximum extent has not yet been determined with certainty, but it appears to have covered at least 6 ha.68 Within this area, the density of tombs varies; many appear to be grouped in clus- ters. Approximately 200 tombs with inhumation bur- ials have thus far been investigated: burial chambers of varying sizes cut into the bedrock were entered via shaft-like antechambers, some with steps leading down from ground level; slabs or built walls blocked passage into the main chamber. The state of preser- vation of most of the human remains is poor but adequate for certain generalizations about the nature of burial customs. Bodies were placed into the tombs in contracted positions. It is possible that some tombs never held more than a single burial, but most con- tained multiple burials. The earlier burials were swept aside when necessary to make room for additions; sand or pebbles were sometimes used to separate 4 iijiii layers of older bone from newer. Other graves appear to have been employed as ossuaries. Traces of burning inside some tombs reflect burial ritual. Cycladic and Anatolian types of pottery are both found with some frequency: frying pans are especially common (fig. 9), recovered in over 20 graves.69 Types such as sauce- Fig. 10. Manika,Euboia. Marble figurine. (CourtesyEfi Sak- boats, especially common in settlement levels, are ellaraki) rarely found in the cemetery. These newly excavated graves and buildings pro- 11); terracotta figurines, and both terracotta and vide clearer contexts for a number of EBA artifact stone zoomorphic vessels are also represented. The types that had not been well documented in excavated marble anthropomorphic figures include schematic contexts. Scientific analyses have shown that bone types, examples reminiscent of Plastiras and Louros tubes of Cycladic type with incised decoration once features, canonical folded-arm figurines, and two contained azurite, presumably used as a paint.70 More seated figures. This corpus is an important addition than 10 anthropomorphic figurines of marble and of to the still relatively small body of EBA marble figures bone or shell have thus far been reported (figs. 10- that have been found in the Aegean in secure undis-

66 Cf. Sakellaraki,ArchDelt 41 (supra n. 63) 104, 122, but EarlyHelladic Graves of Manika:Contribution to the Socio- see also Sakellaraki,BSA (supra n. 63) 236 where there is economic Conditionsof the EarlyBronze Age," in R. Laffi- said to be no evidence for town planning at the site. neur ed., Thanatos:Les coutumesfundraires en Egie a l'fAge 67 On fortifications,see Sakellaraki,ArchDelt 41 (supran. du Bronze(Aegeum 1, Liege 1987) 19-28; M. Fountoulakis, 63) 101, 134, and 266. Characteristicsof EBA architecture "Some Unusual Burial Practicesin the Early Helladic Ne- at Manikaand elsewhere in Euboia have been reviewed in cropolis of Manika,"in Thanatos29-33. A. Sampson, "The Type of the Early Helladic House in 69 Importantnew evidencesuggests that these vesselswere Euboea,"in ManikaII, 120-23; and Sampson,"Architecture at least sometimes brought to the grave containing food, and Urbanizationin Manika,Chalkis," in R. Hagg and D. since animalbones have been found in them. See Sakellaraki, Konsola eds., Early Helladic Architectureand Urbanization BSA (supran. 63) 240, 264, and Coleman(supra n. 34) 202- (SIMA76, G6teborg 1986) 47-50. 204. 68 On tomb form and burial ritual especiallysee Manika 70 Cf. Sakellaraki,BSA (supra n. 63) 251, 264 on the use II; Sakellaraki,BSA (supran. 63) 256-64; A. Sampson,"The of these and other vessels as containersfor paint. 1992] THE ISLANDS OF THE AEGEAN 719

one has a spiral, another, a leaf-shape imprint, and the third, a geometric design.73 Rescue excavations at Classical Eretria in the Vour- atsas plot, near the northeast corner of the later , have recovered deep prehistoric deposits extending well beneath the level of the water table. Finds re- portedly date to the later phases of EH and to the beginning of the MBA, and there is a thin overlying stratum containing Minyan, Minoan, LH IIIB, and LH IIIC types. The greater part of the pottery may be dated to EH II, EH III, and early MH; proto- Minyan handmade bowls are well represented. The earlier finds were associated with architectural ;::: re- mains, including a storeroom that still contained a 15- cm thick stratum of carbonized grains and pulses on its floor. A well-preserved potter's kiln has been now moved to the Eretria Museum.74 The exact date of the kiln is not entirely certain, since excavation be- neath the water table made it impossible to record detailed stratigraphy, but it is clear that it is earlier than the Geometric period and much of the associated pottery was of EH types. The prehistoric settlement explored by tests in the Vouratsas plot probably lay on a low headland by the sea. Excavations on the of Eretria have also yielded prehistoric finds. Most of the prehistoric pot- tery recovered there may, however, have been incor- porated within mudbricks, subsequently decayed, and Fig. 11. Manika,Euboia. Bone figurine. Efi Sak- (Courtesy reused after the Bronze Age for fill, particularly for ellaraki) the core of the northeast tower of the fortifications. Finds from the acropolis belong mainly to the later turbed archaeological contexts.71 Metal artifacts are stages of the Middle Helladic period and the Late common. The especially 50-odd artifacts consist of Bronze Age (only LH IIIA-C is definitely attested). both silver and bronze, and include a curved knife, Elsewhere, in an area close to the western circuit of chisels, pins with pyramidal, conical, spherical, dou- the historical city wall, Neolithic stone axes have been ble-spiral, and roll-tops, tweezers, a cosmetic scraper, reported.75 and razors, rings, necklaces, bangles. A number of Several other sites on Euboia have been explored these find closest types typological parallels in the through more limited excavation; results have not yet eastern Aegean and in Anatolia.72 Of three stone seals, been published in detail. In the mountains of central

71 The variousfigures found in excavationsat Manikaare 44, esp. 138-39. On study of the material from the recent reviewed and discussed by E. Sakellaraki,"Nouvelles figu- excavations,see P. Ducrey, "Les activites de l'tcole suisse rines et cycladiques petite glyptiquedu bronze ancien d'Eu- d'arch6ologieen Grece,"AntK 28 (1985) 143. For the most AntK34 bee," (1991) 3-12. See also Manika II, 70-71. detailed description of prehistoric Eretria, see P. Ducrey, For the 72 typology of metal finds, see V. McGeehanLir- une cite de la Grece antique,"DossPar 94 (1985) "The "ter6trie, itzis, EarlyBronze Age Metalsfrom Manika,Euboea," 8-10, and S. "Des N olithiques aux in Miiller, Myc6niens," ManikaII, 105-19; also E. Sakellaraki,BSA (supran. 63) DossPar 94 (1985) 12-16. I.R. Metzer, "Die Keramik von 250-51. Eretria," Aexeiov EvfloixCbv 26 (1984-1985) Manika also MeAetrhv 73 II, 72; Manika I, 459-60, and Sakellaraki, 221-52 also reviewsthe evidence for prehistoricoccupation ArchDelt41 (supra n. 63) 266. and for Dark Age settlement. 74 See A. Tuor, "Die in G prihistorische Grabung 10," 75 AR 35 (1989) 23. The characterof Neolithicoccupation 24 "Zur AntK (1981) 83-84; bronzezeitlichenSiedlung in G remainsunclear. Miiller (supra n. 74) 15 has suggested that 10,"AntK 25 (1982) 158-60. On the overall distributionof CretanLN imports (!) are present among finds from Tuor's material prehistoric at Eretria, see also C. Krause, "Zur excavationsin the area of the later Agora. staidtbaulichenEntwicklung Eretrias," AntK 25 (1982) 137- 720 JACK L. DAVIS [AJA 96

--:j::0 7

$!!i:i::::?i4ilo

4j.

Fig. 12. Excavationsat Kaloyerovrysi,Euboia. (Courtesy Adamantios Sampson)

Euboia east of Chalkis, most of a small EH-MH site had been covered with slabs and contained parts of at Fylla Kaloyerovrysihas been explored (fig. 12). Both two to three dozen disarticulated incomplete human rectangular and apsidal buildings were associated with skeletons; among the grave goods was a marble bowl EH levels; obsidian was notably rare as was fine Ur- of Cycladic type. Remains of a pillaged Mycenaean firnis pottery, surprising given the proximity of the chamber tomb were also investigated nearby.78 At site to Chalkis Manika. MH buildings of two different Mourteri on the south side of the gulf of Kyme, two phases lay above the EH levels; the highest phase is buildings have been excavated, one rectangular, the reported to date to the transitional period between other apsidal; the pottery was exclusively of EH II, the Middle and Late Bronze Age. MH cist graves and and obsidian was noticeably scarce among the non- the remains of a plundered shaft grave of early LH ceramic finds.79 At Amarynthos on the coast in the date were excavated among the MH buildings.76 west central part of the island, excavation has pro- In the northwest part of the island a second MH duced Neolithic, EBA, MBA, and LBA finds; parts of settlement was investigated at Aidipsou Koumbi. buildings and a section of a possible fortification wall Foundations of two separate late MH buildings were of the Early Helladic period have been explored.80 found stratified over an earlier MH structure; on the Most recently, abundant EB, MB, and LB finds were floor of the latter was a thick burnt layer. Beneath it recovered in renewed tests.81 At Linovrochi, a coastal was a cist grave, also of MH date." Within the limits site east of Lefkandi, the construction of irrigation of the inland mountain village of Avlonari Ayios Yior- facilities uncovered structures of EH I and II; subse- yios in the east-central part of the island, south of quent archaeological investigations revealed part of a ancient Kyme, excavations have uncovered an EH II probable apsidal building.82 Elsewhere on the island, ossuary, consisting of a rectangular built tomb that a figurine and vase were found in a grave at Eretria

76 See A. Sampson,"KaXoyEQ61oQvouP ( '1u 6ov," ArchDelt prehistoricsettlement in the area of Kyme, including brief 38 B' (1983) 154; Manika II, 122, fig. 115. My description descriptionsof severalnew Neolithic, EBA, MBA, and LBA of this site is summarizedfrom a text kindlyprovided by the findspots. excavator,Adamantios Sampson. 80 L. Parlama, "MLxQ( avaoxaWlLxO'OV QEurvacV TQO- 77 A. Kovnti," ArchDelt37 B' X64o AAA 12 (1979) 3-14. Sampson, "Atbblapo6g: LoToQLx6 lqgAtaQ1avOov," (1983) 175-76; "Kov Rl Ab&rlaoou,"ArchDelt 38 B' (1983) 81 "Chronique" 1988, 668. 141. See GazetteerG87. 82 Sakellaraki,AAA 19 (supra n. 63) 36-37; Manika II, 78 Manika I, 332, 366, figs. 124-26; "AyLogFEd)QyLog; 123, pl. 161; Manika I, 361-63, fig. 76. E. Sakellaraki,"I8L- AvukvaQfov," ArchDelt38 B' (1983) 141, 154-55. otxFg ovrkkoyFg,"ArchDelt 38 B' (1983) 152 has described 79 A. Sampson,Evflo rj Kitrl7(Athens 1981) 56-58; 31- various prehistoricfinds of EH and LH IIIC date that are 32 for FN finds from surface collections and from excava- in the care of the community of Vasiliko;these include a tions at the site of PotamiaKastri near the northern end of Phi-figurinefrom LefkandiXeropolis and a Mycenaeanstir- the bay; and 47-52 for general discussion of patterns of rupjar from Linovrochi. 1992] THE ISLANDS OF THE AEGEAN 721

Magoula;s3 prehistoric remains have been investi- In addition to this program of surface reconnais- gated at Petries; a Mycenaean tholos tomb previously sance and recapitulation, excavations have been excavated at Aliveri Katakalos has been studied and started at the most promising of the EBA sites, a low restored;84 and a previously looted Mycenaean cham- peninsula called Palamari, which lies on the northeast ber tomb has been excavated at Limni Fasouli in the coast of the island and faces toward the islands of northwest.85 Ayios Efstratios and Limnos. Reports have summa- Skyros. The prehistoric antiquities of the island of rized results of field campaigns in 1981, 1985, and Skyros have been extensively investigated in recent 1986.89 Remains of architecture observable on the years. All archaeological bibliography relevant to the surface suggest that the site was quite large, on the island has been synthesized, and artifacts from un- order of 10 ha; of this, an area of ca. 260 m2 has thus published excavations have been described and illus- far been excavated. Some traces of possible defenses trated as have various prehistoric chance finds in the on the neck of the peninsula are of uncertain date. collection of the local museum.86 Skyros seems to have Within the town, the buildings appear to have been been most extensively settled in the EBA; 10 sites with built entirely of stone with roofs of schist slabs. characteristic EBA pottery have been identified. All A series of important stratified habitation deposits are in the more fertile northwest part of the island, appear to range in date from EB II through the three along a coastal strip near the modern capital, beginning of the MBA, and EH I pottery has been the only part of Skyros where there is evidence for observed in excavated surface levels and in surface continuity in habitation between the Neolithic and collections.90 The earliest architectural remains yet Bronze Age.s7 Parlama postulates two separate foci of investigated are associated with pottery that contains residence at the Chora in the EBA, one on the acrop- a mixture of Helladic and Cycladic elements, as at olis above the modern town, the other north along Ayia Irini on Keos farther south. There are glazed the coast; complete vases turned over to museum sauceboats and saucers, but also sherds with Kerb- authorities suggest that there were EBA cemeteries schnitt and other impressed patterns. In the following in at least four nearby locations. Several vases find phase of occupation, the sauceboats continue but An- close parallels in the Cyclades, as does a pair of bone atolian types are present, as in phase I of Lefkandi tubes with incised triangles and herringbone pattern. and in period III at Ayia Irini. In one instance the A fragment of a schematic marble figurine of "violin" Anatolianizing deposits are overlain by house Alpha, type comes from the tiny island of Atsitsa, just off the which itself contained gray-burnished wares and part northwest coast.88 of a possible duck-vase with incised decoration that

83 Manika II, 123, pl. 162; Manika I, 364; Sakellaraki pos Group," in Prehistoric Cyclades 33, fig. Ig. The pyxis lid (supra n. 71) 3, 5, fig. 2, 11, no. 2, pl. 2.2 and pl. 1.3 for a from Skyros together with the two incised bone tubes have schematicfigurine and a vase found with it. more recently been republished, with reference to parallels 84 For Petries,see Manika II, 120, pl. 160. For Katakalos, at Manika on Euboia, by E. Sapouna-Sakellaraki, "Kvxha- see A. ArchDelt39 B' (1984) 125, and f in I Sampson, "AXtl3PFt," 8tx& rlg Exgov," OPItta ErIm (BtfiAtoOlrxr7r)T; ev GazetteerF91. A0r7vatg ApXatoA)oytxjgEratZpiag 103, Athens 1986) 85 E. Touloupa, "Altvrl," ArchDelt 33 B' (1978) 130; see 293-99. GazetteerG98. 89 M.D. Theochari and L. Parlama, "Palamari, an Early 86 L. Parlama, H x~dpog or?7v eJroxqTrov Xa)hxov(Diss. Bronze Age Settlement at Skyros," in Higg and Konsola Univ. of Ioannina 1984). See also A. Sampson, "AQTe[tLot," (supra n. 67) 51-55; L. Parlama, "NcO6cQactcToTXE•a aCr6 ArchDelt 38 B' (1983) 155, where abundant Neolithic pottery TqyVcVa(voxacT 0j TO rtQO'i(YTOQoXO) otxtoot (o1oHO nat•tagtc and obsidian are at Artemisi in the southern in MovauEo Kvx- reported part TrlgExtVoU," I6pv/paN.H. Tovd)av6pj, of the island. Aa6txjg Tdxvrjg, Ata)dEtLg 1986-1989 (Athens 1990) 87 Parlama dates another 13 sites to the EBA on the basis 125-34; ns. 11-12 briefly summarize the results of the 1987 of finds of obsidian alone; while such an assignment is pos- season. In addition, see A. Sampson, "Ht•aXtpaQLtEx? QOU," sible, it is not demonstrable given the frequency with which ArchDelt 38 B' (1983) 141, and E. Sapouna-Sakellaraki, obsidian was used in all phases of the Aegean Neolithic and "Ex1Otog,"ArchDelt 38 B' (1983) 150 for results of emer- Bronze Age. gency excavations in 1983; EH III pottery is there reported. 88 Parlama (supra n. 86) publishes a fragmentary EC pyxis 90 See Parlama (supra n. 86) 91-94 for the most complete lid (107-108, pl. 53, bottom right) and a small EBA bottle published description of the 1981 excavation season, includ- from Chalandriani on Syros (105, pl. 53, bottom left); the ing the fullest description of the character of the EB II and former has a close parallel in a more complete lid from Lefkandi I ceramics from the site, with several photographs Skyros (pls. 49-50), and the latter has parallels not only on and drawings. Among other pottery that characterizes these Skyros (pl. 48.10-11), but also on Euboia at Kyme Kastri phases are sizable quantities of the talc ware found at Ayia Potamias (supra n. 79), and among finds from the EBA Irini and other Cycladic sites (see Vaughan and Wilson, cemetery on Epano Koufonisi. For the material from Kou- supra n. 20). fonisi, see F. Zafeiropoulou, "The Chronology of the Kam- 722 JACK L. DAVIS [AJA 96 finds its closest parallels in Phylakopi 1.2. Red-bur- dard Mycenaean type.92 Evidence of a Mycenaean nished bowls, which can be paralleled both in Troy V presence elsewhere on Skyros is meager. and in the Middle were recovered Cycladic period, Thasos from a pit dug into Anatolianizing deposits before the A review of the prehistory of the island with a construction of house Alpha.91 Immediately to the gazetteer of prehistoric sites has recently been pub- west, in house Gamma, Anatolianizing types were lished.93 The Neolithic of the island remains poorly excavated on a floor beneath the main deposit of the known and unrepresented in excavations. Late Neo- house, which contained ajug of Phylakopi I type, with lithic ceramic types of the sort characteristic in Mac- pottery, including a relief-decorated "face-pot" that is edonia have been recognized as have white-painted said to find its closest parallels in Troy IV-V and the dark-ground types evocative of the southern Aegean. "brown" of Poliochni. phase Links with Macedonia continued in the EBA, but the House Gamma is also of interest for its architecture MBA is a blank. The most significant Bronze Age and non-ceramic contents. Its one fully explored center on Thasos is at Kastri, a room has stone-built interior fixtures that include a Theologos naturally fortified plateau in the southwestern part of the raised a bin, and a hearth; various platform, storage island. Some 300 m2 of the settlement have been jars, querns, and handstones, all in a good state of excavated; contemporary stone-built tombs in ceme- preservation, suggest that the building was aban- teries nearby at Tsiganadika, Vrysoudes, and Kentria doned in haste. Metal artifacts include a piece of lead contained local imitation Mycenaean pottery and and a of bronze; metals also sheeting roll-top pin weapons, knives, and tools that also find antecedents to have been worked at the site. appear in the south. The burials for the most part contained There are few traces of Middle Bronze Age activity multiple inhumations although are also later in date than the abandonment of the settlement represented. The analysis of finds from these graves at Palamari: a handful of worn sherds from only badly suggests local mining and metalworking already by Atsitsa have possible parallels among the Minoan and the end of the LBA.94 of Irini V, and can Minoanizing shapes Ayia period Since 1986 the small site of Skala Sotiros Profitis be added to the and sherds Minyan matt-painted Elias, located on a small hill in the northwest part of noted at tou Houma. Documentation years ago Papa the island, within the village of Skala Sotiros, has been is in short for the earlier of similarly supply phases explored. There, immediately beneath building levels the LBA. Of 100 vases now in the museum Mycenaean of the sixth to fourth century B.C., are remains of a of Skyros (all but one with some recorded findspot), small (ca. 1400 m2 in extent) habitation of the EBA, none is in LH earlier date than IIIA:2. Virtually all surrounded by a fortification wall, the earliest phase of these vases were found in the vicinity of the modern of which was built in part with masonry of herring- Chora and probably derive from tombs in cemeteries bone style.95 Excavations inside the fortifications have that existed north, south, and east of its acropolis. No yielded the remains of several buildings. Floors asso- unplundered tomb has yet been excavated but like ciated with a structure destroyed in a conflagration the looted tombs investigated by Theoharis at Krokos, contained numerous complete pots, some filled with most are likely to have been chamber tombs of stan- the remains of carbonized seeds (mostly legumes),

91 For the fullest discussionof the stratigraphyof houses possibilitythat local iron and copper were also being worked A and B, see Parlama (supra n. 89) 124-26. The exact at the end of the Bronze Age on Thasos. Matsas(infra n. chronology of these post-LefkandiI groups remains to be 100) mentions gold objects, probablyof local metal, from sorted out. It is important to note that thus far no EH III latest LBA graves at Kastrias well as evidence for the work- painted wares have been published from the site. ing of local copper in the Chalcolithicphases of the settle- 92 Parlama(supra n. 86) 136, also reports unverifiedlocal ment at Kastri.Analysis of lead artifactsfrom gravesat Kastri testimonythat suggests the existence on Skyrosof large cist points to compositions compatible with those of Thasian graves of the variety found on Psara in the Mycenaean ores: see E. Pernicka,G.A. Wagner, and W. Todt, "Chem- period. ische und isotopischeZusammensetzung friiheisenzeitlicher 93 C. Koukouli-Chrysanthaki,"Die fruihe Eisenzeit auf Bleiartefakte aus Thasos," in C. Koukouli-Chrysanthaki, zwischen1600 und 1000 v. Thasos,"Siidosteuropa Chr.(Ber- HI•oibzootxj' O96og (in press). lin H. 1982) 119-43; Matthaus,"Thasos im Altertum,"in 95 C. Koukouli-Chrysanthaki, "OLtxtoT6Lg t g HQWotlli G.A. G. Wagnerand Weisgerbereds., AntikeEdel- undBunt- EjToX"igTov Xakxoi, orCrExda TorilQogO;e6oo," AEMT Thasos Kou- 1 metallgewinnungauf (Bochum 1988) 13-39; (1987) 389-406; [IQIooLdgEooX ig XaX- ArchDelt35 B' xoi "Otx"to'g6;qg kouli-Chrysanthaki,"OEoX6yog," (1980) 422. ormExd6I ca Zo1wrqogoS oov (II)," AEMT 2 (1988) 94 E. Pernicka and G.A. Wagner, "Thasos als Roh- 421-31. For reports in the Greek press on later seasons, see stoffquellefuir Bunt- und Edelmetalleim Altertum,"in Wag- AR 35 (1989) 97. ner and Weisgerber (supra n. 93) 224-31, suggest the 1992] THE ISLANDS OF THE AEGEAN 723 part of a hearth, and various small finds, among them Neolithic, and the absence of bones from domestic chipped and ground stone, bone pins, needles, and species suggest that the mines were principally spindle whorls. worked in the Palaeolithic, and dates of 15,000- These deposits appear to mark the end of the EBA 12,000 B.C. have been advanced.96 If correct, these settlement; by that time the circuit was no longer would be the oldest underground mines yet docu- serving a defensive function. From an older phase of mented in Europe; Thasos would still have been at- habitation, two fragments of a stele in local schist and tached to the adjacent mainland at this time. decorated in relief with the figure of a warrior have been retrieved, both reused as building material in Samothrace the earliest phase of the western peribolos of the Investigation of prehistoric Samothrace began in settlement. The head of the warrior was engraved 1974 following rescue excavations on the coast south- with shallow incision; he wears a necklace, holds his west of modern Chora, at the site of Mikro Vouni; a right hand raised to his breast, and grasps a dagger wider program of regional survey, geological investi- in his left; set obliquely to his chest is a spear and a gation, and ethnography has followed.97 Systematic double-bladed axe is stuck through his broad belt. survey elsewhere on Samothrace suggests that the The closest parallels for this stele are in Troy I. Sev- settlement pattern in the Bronze Age was character- eral fragmentary anthropomorphic schematic marble ized by a high degree of nucleation. Mikro Vouni figurines, and parts of two others that were reused as itself appears to have been a narrow-necked peninsula building material within the circuit wall of the settle- in prehistoric times. Excavations have thus far inves- ment, also find their closest parallels at Troy. These tigated levels of the FN, the earlier phases of the EBA, finds clearly predate construction of the peribolos but and of LH III; one sounding is nearly 8 m deep.98 thus far soundings have failed to recognize earlier Nine phases of construction have been recognized, habitation levels. Parallels for ceramics from the ear- and surface collection suggests that the settlement was lier levels of Skala Sotiros are with Troy I, Sitagroi approximately 1 ha in extent. The earliest material Va, and Emborio IV-V, from the later levels with from Mikro Vouni is compared with that of Kum Troy II-IV and later phases of the Macedonian EBA. Tepe la and Ib, Poliochni "black," and Emborio VII- Finds other than pottery from the excavations include VI; the latest, to the middle phases of Troy VI.99 obsidian, several metal objects, and a gold button- The very recent announcement of Minoan docu- shaped ornament. ments, found amid destruction debris (including frag- Elsewhere, in the southwest part of the island, at ments of white plaster with red decoration) in the the site of Limenaria Tsines, two hematite mines (of penultimate architectural phase of the site, is partic- some 15-20 in the vicinity) are thought to be prehis- ularly exciting.'00 These include a roundel and a nod- toric in date: handmade pottery, some of it Neolithic, ule of types well known in Crete and generally found has been found at the mouth of one mine and appears in administrative contexts. The Samothracian roundel to postdate the mining. Bone and stone tools have may be of local clay, and was stamped repeatedly been found in both mines and appear to have been around its edges with a Minoan "cushion" sealstone employed for the extraction of minerals; the presence engraved with Cretan hieroglyphic signs; it perhaps of horn from the Saiga antelope, extinct before the had Linear A signs written in ink on one side. The

96 C. Koukouli-Chrysanthakiand G. Weisgerber,"OEo r 98 D. Matsas,"XcaotOQ96'x' 1987: AgQXotoXoyLXtgxaLt EO- ToivEg:IHQOLOtQLx 6 oQUXELo," ArchDelt 37 B' (1982) 322- voagQaLoXoyLXEg EQgyacrEg,"AEMT 1 (1987) 499-502. 24; "ALLEVa6QLaE6oov," ArchDelt 38 B' (1983) 319; Views of architecturalremains are included; an LBA "spit ArchDelt39 B' rest" and an FN vessel are illustrated. "OvxtELo(xQatg oTrI 0BOrI Tv;VEg," (1984) anthropomorphic 268. See also C. Koukouli-Chrysanthakiet al., "Prdihisto- Elsewhere, prehistoricremains have been reported at the rischerundjunger Bergbauauf Eisenpigmenteauf Thasos," site of Mantroudain the southern foothills of the central in Wagnerand n. 93) 241-44, where '4C massifof the island(E. Skarlatidou, Weisgerber(supra "IHavaylarl MavdX0W," dates of ca. 6400 B.C. on artifacts of deer antler used for ArchDelt35 B' [1980] 434) and on the acropolis of Vrihou, mining tools have been reported. The dates are based on west of Chora, where early remains have been the calcium in the antler, not collagen, and thus offer only recently investigated. I thank D. Matsas for information a terminusante quem for its age. about this site. X* 97 D. Matsas,"MLxQ6 Bovvw aoO96xr'lg: MiLa to'lo- 99 "Chronique"1989 reports that in 1988 lower levels of o' BA were with remains earlier TOQLXI?Ixov6•o•Ta otvavrlotaLrLxo o6trl~a tov At- occupation reached, substantially yaov," Anthropologika6 (1984) 73-94; in date (ca. 5500-5000 B.C.) than previouslyreported. Arhaiologia13 (Sept.-Nov. 1984) 35-43, esp."X•catoOQaxlq," 36. A selection 100D. Matsas,"Samothrace and the NortheasternAegean: of pottery from surface collections at the site is catalogued The MinoanConnection," Studia Troica 1 (1991) 159-79. and illustrated. 724 JACK L. DAVIS [AJA 96 nodule had been pressed against a peg wound round with iron (the well-known "tin bangle" from period with string and had clearly been used to seal a wooden IV, the earliest Aegean artifact in which tin is the container; afterward it had been stamped, leaving a principal constituent). In several instances at Thermi, leaf-shaped imprint. The discovery of a mudbrick brasses were also identified. marked with a linear sign amid the destruction debris For the most part, the isotopic signatures of the suggests that writing was practiced locally. The asso- lead contained in the copper from which unalloyed ciated pottery has been described in some detail and copper artifacts and arsenic bronzes were fashioned is compared to that from early subphases of Troy VI are the same as those from contemporary artifacts and with the Poliochni "brown" phase. Carbon-14 found in the Troad and at Yortan. Deposits in north- dates for two samples of associated bone have thus west Anatolia may be the most probable candidates far been reported (with a combined calibrated date of for the source of the copper. Lead-isotope ratios ap- 2030-1785 at a probability of two standard devia- pear to rule out Laurion as the source for even a tions). single artifact of those from the "yellow" phase at Poliochni. The majority of copper-based artifacts that Lesbos, Chios, Psara, and Limnos 101 have an isotopic signature compatible with those of Lesbos and Limnos. A sounding beneath the Helle- Aegean ores have incompatible chemical composi- nistic cemetery of ancient Mithymna on Lesbos tions. Chemical and lead isotopic analyses of artifacts reached prehistoric levels for the first time.102 Also of from both Thermi and Poliochni suggest, however, interest is the publication of a single Neolithic sherd, that tin bronze was not invented or produced in the perhaps from the site of Halakies in the south of the north Aegean. Tin bronze appears to have been ex- island.'03 On Limnos, continuing excavations, begun ported to the northeast Aegean from a source as yet at the site of Myrina Riha Nera in 1986, have uncov- unknown. Imported tin does not appear to have been ered in several soundings parts of a complex of EBA added to copper from sources that had previously buildings with at least three phases of habitation; been used either or with associated finds are said to be contemporary with the locally unalloyed alloyed arsenic. The fact that these "local" sources continued "yellow" and "green" phases of Poliochni.104 At Po- liochni itself restoration work has clarified some de- to be exploited after tin bronze had become wide- that the EB II was a time when tails of the architectural history of the prehistoric spread suggests period became from settlement. 105 copper available a broader geographical Of more interest than the results of new excavations range of deposits. Of the five silver from Poliochni that have are those offered by recent metallurgical analyses of objects EBA finds from older excavations at Poliochni on been examined (including a double spiral-headed pin from the "blue" the earliest well- Limnos and at Thermi on Lesbos.o06 Of 33 objects period, probably stratified silver artifact from the all of available for analysis from Thermi (phases III-V), Aegean), are only three included tin as a constituent. In the longer- pure metal; they are not from the same ore and lived settlement of Poliochni, among 74 samples an- appear to have been extracted by cupellation.'07 Lev- alyzed (from the "blue" to "yellow" phases), the quan- els of silver were found to be so high in lead objects tity of tin bronzes was found to increase throughout that it is clear that the lead had not been desilvered. the life of the settlement, reaching more than 50% by The isotope ratio of the silver pin suggests that the the "yellow" phase. Arsenic was used alternatively silver from which it was made came from outside the with tin as an alloy to produce bronze; tin in one Aegean area; of the other objects, some are compati- example from Thermi appears to have been alloyed ble with Laurion and Siphnian isotopic fingerprints.

'01 I am grateful to M. Ozdogan for the informationthat prdhistorique(Paris 1990) 321-30, for an overview of work there has been no relevant recent archaeologicalwork con- at the site and for references to other prehistoric sites on ducted under Turkish auspiceson the islandsof or the island. See also AR 35 (1989) 91-92; 36 (1990) 62. . 106 F. Begemann, D. Schmitt-Strecker, and E. Pernicka, 102 A. Arhontidou-Argyri,"XQovLtx K' EQoQE(ctgAQX- "The Metal Finds from Thermi III-IV: A Chemical and Catorill'v Eroig 1986,"Lesviaka 12 (1986) 72. Lead-IsotopeStudy," Studia Troica2 (in press); E. Pernicka 103 Kastro Tigani 105, n. 413. et al., "Onthe Compositionand Provenanceof MetalArte- 104 Arhontidou-Argyri (supra n. 102) 56, 68-69. A paper facts from Poliochni on ," OJA 9 (1990) 263-97. entitled "MycenaeanEvents from Lemnos," was delivered Three new 14Cdates for EBA levels at Thermi and one from by Arhontidou-Argyriat the InternationalCongress of My- Poliochni have been published: see Begemann et al. (supra); cenology in Rome-, October 1991. and M. Korfmann ed., Demircihiiyiik II: Naturwissenschaft- 105 M. Ricciardiand S. Tine, "Poliochni1986-1987: Inter- liche Untersuchungen (Mainz 1987) xviii, fig. 4. venti di restauroconservativo e ASAtene64- For valorizzazione," 107 the pin, see K. Branigan, Aegean Metalwork of the 65 (1986-1987) 389-401; see also P. Belli, "L'abitatodi Early and Middle Bronze Age (Oxford 1974) no. 2064. Poliochni,"in P. Darcque and R. Treuil eds., L'habitat6geen 1992] THE ISLANDS OF THE AEGEAN 725

A particularly exciting development is the discovery Architectural remains were for the most part that in the northeast Aegean isotopic and chemical scrappy, although there were substantial stone walls compositions of copper-based metals vary so uni- even in the lowest stratum (period X) reached in area formly through time that their composition can itself A; human skulls were found buried beneath one floor be used as a means to assign poorly stratified finds to of period X. An adolescent had also been buried a place in the Trojan stratigraphical sequence. Thus beneath a wall of a room that dated to period VIII. the types of metal used for bronze artifacts of Ana- A spring in area A was later converted into a well, tolian types at the site of Kastri on Syros can be and was finally abandoned only in period II. At the tentatively dated to Troy IIg, providing support for end of period IV, the settlement appears to have been the contemporaneity of the Kastri I phase in the destroyed by fire at a time approximately contempo- Cyclades with EB II at Troy. rary with the earlier stages of the Aegean EBA; this Chios. The final publication of the results of exca- is the only evidence of violent destruction recognized vations at the prehistoric site of Emborio in south- at Emborio. After the destruction, the ruins of the eastern Chios (1952-1955), the best anchorage destroyed houses were deliberately filled with stones between the Chora of the island and Kato Fana, and to build a level platform, which ultimately supported at Ayio Gala in the northwest (1938) has provided for the foundations of period II houses. the eastern Aegean a well-documented sequence of For the most part, finds later than the EBA were levels spanning the Later Neolithic and EBA.s08 Ten found only in mixed levels. Burnished gray wares and periods, numbered from top to bottom, were defined matt-painted types were both represented. Most of on the basis of excavated deposits: period I, the latest the gray ware and matt-painted ware belongs to the of the EBA periods, is contemporary with Troy II; LBA, but there are also types that find parallels in period VII ceramics have close parallels in FN mate- Troy III-V. The matt-painted (most often in bi- rial at Kephala on Keos;'09 and period X corresponds chrome red and black) pottery, though apparently of to an early phase of the mainland Greek Late Neo- local production, looks ultimately to Crete for inspi- lithic. Still earlier material lay beneath the water table ration, and the limited range of motifs represented and was not investigated. Wheelmade pottery was first (e.g., spirals, foliate bands, wavy bands) is similar to introduced in period I, along with an Anatolian tan- the local Minoanizing pottery of the Dodecanese. The kard shape that is closely paralleled in the Cyclades absence of light-on-dark imitations distinguishes it among ceramics of the Kastri Group. MBA and LBA from the local Minoanizing wares of Rhodes and Kos. finds were not assigned to periods. There is also no evidence that a full range of Minoan The deepest and earliest stratified sequence of de- specialty implements (e.g., fireboxes) or tools (e.g., posits at Emborio was excavated in area A, on flat discoid loomweights) was introduced at Emborio; the ground northwest of the acropolis between it and the lack of obvious Cretan imports also appears to distin- historical Greek sanctuary. EBA, as well as LBA, de- guish Chios from its neighbors to the south."' posits were also explored at the northern edge of the The most substantial Mycenaean remains come acropolis, and on its western ascent. In the EBA the from the north edge of the acropolis in area F. Build- minimum extent of the settlement has been estimated ing levels of the LBA for the most part were dated to at approximately 3 ha. Before the EBA the settlement LH IIIC, although two floors in area F contained non- may not have included the acropolis itself. Slopes Mycenaean style pottery that may belong to earlier southwest of the acropolis served as a cemetery: a phases of the LBA.l"2 The pottery in LH IIIC levels rock-cut EBA tomb and two Mycenaean cist graves has its closest stylistic ties to Attica and Euboia; and were found in trials.11"0 the bulk of it appears to be Mycenaean rather than

110 108 In addition to the British investigations,rescue exca- See S. Hood, "Mycenaeansin Chios,"in J. Boardman vations in the Baha plot (on the south shore of the harbor and C.E. Vaphopoulou-Richardsoneds., Chios:A Conference at Emborio) have explored what appears to be an undis- at the Homereion in Chios (Oxford 1986) 169, for discussion turbed Neolithic stratum: see L. Ahilara, of the use of cists for burialsand a "AQXaLOkoyLXtd Mycenaean comparison Xov," Hiaka Hronika 17 (1985) 77-80. The between on Chios and those in the at Ar- XQovLXtdtrg graves cemetery relative chronology of Emborio has recently been reeval- hontiki on Psara.Similar cist graves may have been used on uated in Kastro Tigani 72-83. See also the analysis of a Skyros(supra n. 92). copper ingot from Emborio,in Stos-Galeand Gale (supran. I'l Note, however,the presence of a strainer-vase(Emporio 39) 81, fig. 10, and 82. 575), fig. 256.2675), and a single discoid loomweight of 109 Hood's view that period X of Emborio was contempo- Minoantype (Emporio633). rary with the beginning of EN on the Greek mainland has 112 Several sherds may date stylisticallyto LH IIIB, as do not been widely accepted;it is unfortunate that only a single the four vases found in a cist grave (Emporio582-83). '4C date (from period IV) is availablefor the site as a whole. 726 JACK L. DAVIS [AJA 96

Anatolian in character, although several local shapes cylindrical in shape, from a Mycenaean level. Three of Anatolian derivation are represented."3 The LH fragments of amber, also of Mycenaean date, are IIIC town on the acropolis was destroyed by burning; Baltic in origin.115 The animal bones from both Neo- on the west slope, a contemporary destruction level lithic and Bronze Age levels included characteristic was recognized, with a later poorly preserved phase Neolithic domesticates."6 of LH IIIC stratified above it. The dating of finds from the two caves excavated at Gala before World War has been contro- A variety of small finds from Emborio has been Ayio II versial. No stratification was in the lower published: most belong to the Neolithic and EBA. recognized cave; excavated material to have washed into They include terracotta spoons, one (and possibly appears it a hole in its roof. In the cave, two) terracotta stamp seal(s), a lump of clay used for through nearby upper two distinct strata were recorded, the lower of which sealing, possible crucibles, stone molds (one, of LBA contained ceramics of a character similar to the bulk date, for a butterfly pendant), loomweights, spindle of the finds from the lower cave. The character of the whorls, terracotta figurines (mostly of Mycenaean finds from the lower cave is not closely paralleled types), an ox protome with incised decoration of EBA either at Emborio or at Tigani on Samos. Most schol- date, fragments of several stone vases, including a ars view the finds from Ayio Gala as largely earlier lamp of a type characteristic of the Cretan New Palace than those of Emborio period X and partly contem- period, and an assortment of ground stone tools, porary with the Middle Neolithic of the Greek main- stone, and bone tools."114 the most chipped Among land."7 Small finds from Ayio Gala include terracotta distinctive of the various or bronze finds are copper human heads from figurines or attachments to vessels, of familiar EBA varieties, a flat axe, a pins Aegean bowls of steatite, part of a stone schematic figurine, and a of a complete Mycenaean knife, fragment cop- bone tools, stone and shell pendants, celts, and stone per ingot of LBA type. Metal finds that are clearly bracelets. Neolithic in date are few; they include the tip of a Finds from elsewhere on the island echo the se- knife or dagger blade and a ring pendant with an quence of Emborio. At least 14 sites with traces of attached suspension loop. A phallic pendant of spon- definite or possible prehistoric activity can be added dylus shell has close parallels in the Cyclades. One to those listed in the Gazetteer.118 They are dominantly faience bead comes from an EBA context, another, of EBA and LH III date.19

113 For a fuller discussion of the Mycenaean finds from and are dated still earlier. Since the earliest types found in Emborio and elsewhere in Chios, see Hood (supra n. 110) the upper cave are not represented either in the sequence 169-80. of Emborioor Tigani, Felschsuggests that both the Emborio 114 For the chipped stone, see S. Hood and P.G. Bialor, and Tigani sequences began later than the Arapi phase of "The Chipped Stone and Obsidian Industries of Emporio the Greek mainland Late Neolithic. See also J.B. Rutter, and Ayio Gala," in Emporio699-713. Obsidian is notably who, in a reviewof the Emboriopublication (AJA 88 [1984] scarcein all levels at Emborioand in both cavesat Ayio Gala; 410-11), dated the earliest levels at Emborio to the Final local flint and even limestone was used in its place. Certain Neolithic of the western Aegean, considerably later than other features of the assemblage(e.g., the small and totally materialfrom the Ayio Gala caves. Hood preferred to view expended cores) suggested to Bialor that obsidian was in the materialfrom both caves as representingcultural tradi- short supply. Somewhat counterintuitively,a higher per- tions contemporarywith, but different from, those of Em- centage of obsidian was found in earlier levels than in the borio. In his reconstruction,habitation at Ayio Gala began later levels at Emborio; there are only two tanged arrow- later than Emborio period X and overlapped in date with heads, both from post-Neolithicstrata. periods VIII/IX-VI/V; the earliest phases of occupation at 115 For the analysisof the amber, see C.W. Beck and C.A. both sites were viewed as contemporarywith the later stages Shustak,"Amber from Emborio,"in Emporio727-30. of the mainlandGreek Early Neolithic. 116 J. Clutton-Brock, "The Animal Bones," in Emporio 118 See Emporio 2-9; also, E. Yalouris, "Notes on the To- 678-97. pography of Chios,"in Boardmanand Vaphopoulou-Rich- "I Felsch, Kastro Tigani 96-98, argues that the stratigra- ardson (supra n. 110) 141-68, where detail is added to the phy in the upper cave was inverted,with characteristicforms discussionof some sites-particularly, a map of the (proba- of the EBA in the lower deposits and Neolithic types in the bly) fortified EBA site of Petranosin the southwest part of upper. Parallelswith ceramics characteristicof the Larisa, the island (147, 148, fig. 2). Arapi,and OtzakiA phasesof the Greekmainland suggested 119 Among finds from them, a marble dagger pommel to him that the upper cave was alreadyoccupied at an early comes from Kato Fana (Gazetteer370), two marble handles phase of the mainlandGreek Late Neolithic.Complete vases from Dotia,and a stone axe from the town of Chios (Emporio from the lower cave are compared to types of Hacilar VI 6, fig. 3). 1992] THE ISLANDS OF THE AEGEAN 727

Psara. Excavations in the LH III cemetery at Ar- cent years from the northern islands of the Cyc- hontiki were resumed in 1983 after a hiatus of more lades.121 than 20 years.120 In addition, the northwest and south- Syros. Of particular importance is an ongoing proj- west coasts of the island have been systematically ex- ect that seeks to prepare a complete final publication plored. The cemetery is located on the west coast of for the results of 19th-century excavations on Syros, the island within the bay of Ayios Nikolaos, 3 km at the site of Halandriani; the vast majority of graves north of the modern town of Psara. Most of the and grave goods from investigations in 1861, in the graves, cists with built entrances, appear to have been 1870s, and in 1898 remain unpublished. One prelim- used for multiple burials or as ossuaries; in one large inary report has appeared thus far in which a score grave excavated in 1984, the first articulated inhu- of ceramic and marble vessels are described and illus- mation was discovered amid the remains of earlier trated, most for the first time.'22 Other recent re- interments. the were beads of Among grave goods search has focused on the analysis of metal artifacts and clay, glass paste, faience, gold, semiprecious from the nearby site of Kastri; much of the copper stones such as carnelian, a bronze cylinder seal, pins, appears to have come from outside the Aegean and tools and and whorls of both weapons, spindle clay is similar in composition to that in use at Troy.123 and stone. West of the are remains of walls cemetery Delos. Recent excavations in the sanctuary of Apollo that to a settlement. Excavation in levels belong upper on Delos have explored Mycenaean strata just outside ceramics of date, produced Mycenaean contemporary the fifth-century B.C. prytaneion and beneath the with those from the surface finds, however, cemetery; prodomos of the prytaneion. Mycenaean material had suggest also Neolithic and EBA occupation. been employed to fill cavities in the bedrock, a leveling On the tiny island of Daskaleio, just off the coast of operation apparently conducted in conjunction with Arhontiki in the bay of Ayios Nikolaos and perhaps the establishment, beneath the prytaneion itself, of a once joined to the adjacent mainland, are remains of round cutting in the bedrock (arguably a fountain), similar character and date; 300 m east of Arhontiki, 5 m in diameter.124 These remains are in a sector of EBA levels have been found buried be- occupation the where activities had neath 5 m of alluvium. sanctuary Mycenaean not previously been recorded. Investigations in 1990 The Northern Cyclades launched a systematic attempt to document more Little new information has become available in re- thoroughly all pre-Archaic architecture.

120 Full reports have thus far appeared only in the local on the islandis included. I thank Mr.Hekman for providing journal of the island: A. Tsaravopouloset al., "AQXaLtoo- me with informationabout his project.Cyprian Broodbank Ta Psara 49-51 (1984) 5-11; N. will also include statistics the total number of ytLXi•Q•vva crra WagQ," summarizing Zafeiriou et al., "AQXaoLooytxLiT'evva PWacQv,"Ta Psara ceramicand marbleobjects from the Halandrianicemetery 49-51 (1984) 2-4; L. Ahilara, "Avaoxa)txil 6Qaornl- in his forthcomingpublication of the potteryfrom the recent to Ta QLO6T?rTa W 'xaTdd 1985," Psara 67-69 (1986) Britishinvestigations at Kavoson Keros.Various finds from Craorz 10-11; A. Papadopoulouet al., "Avaoxa)txiij?Q'Eva o ra the island of Syros have also been recently illustratedby F. Ta WaQd1986," Psara 73-75 (1986) 2-7 (the most complete Zafeiropoulou,To MovuEoi n; TZiPov(Athens 1988). description of settlement remains). See also Ahilara (supra 123 Most recently,see "Sourcesof Metals"267. The inter- n. 108) 73-75. A paper by L. Ahilara, entitled "Mycenaean pretationof these data is another matter:see "Perspectives." Events from Psara,"was delivered at the InternationalCon- It is not a foregone conclusionthat bronzes from Troy were gress of Mycenology in Rome-Naples, October 1991. See manufactured from northwest Anatolian copper; see the also Gazetteer 371. discussionof metallurgicalanalyses on Lesbos and Limnos In additionto 121 the new investigationson Syrosand Delos summarizedin this review.For variantopinions, see recently described here, a find from one of the EC graves excavated C. Doumas, "The EBA in the Cyclades:Continuity or Dis- at Diakoftison Mykonos(Gazetteer 308) has been published continuity,"in GreekPrehistory 21-29; and J.A. MacGillivray, n. (Hekman [infra 122] 24); see also E.-M. Bossert, "Zu "CycladicSociety at the End of the EarlyBronze Age," to be einigen Figurgefildenvon den Kykladenund aus Westklein- published in the proceedings of the 6th InternationalCol- asien," in Beitrdge zur Altertumskunde Kleinasiens. Festschrift loquium on Aegean Prehistory,Athens 1987. KurtBittel fiir (Mainz 1983) 121-38, esp. 127 fig. 2.3 where 124 R. Etienne and A. Farnoux, "Delos. 1. Le prytanee. A. an ECvase from Mykonos,stamped with a seal, is illustrated. Sondages,"BCH 112 (1988) 746-52; A. Farnoux,"Delos. 2. 122 J.J. Hekman, "Chalandrianion Syros.An EarlyBronze Nettoyage des murs de l'habitatpre-archaique," BCH 114 Age Cemetery in the Cyclades: Report of the Research (1990) 897-900. For a summaryof recent work, see AR 36 Undertaken in 1990," Netherlands Institute at Athens News- (1990) 66. letter3 (1990) 19-30. A detailed review of earlier research 728 JACK L. DAVIS [AJA 96

The Western Cyclades125 manufactured was being produced on Kythnos. All Kythnos. Archaeological investigations of Bronze but two objects in this hoard now appear, however, to Age Kythnos in 1984-1985 included surface survey have been found on Naxos, perhaps in the Cave of and the excavation of a previously unknown site at Zas, along with several similar tools now in Copen- Skouries, east of the modern Chora in the northeast hagen.'27 Kythnos, nonetheless, appears to have been part of the island.126 Amid fragments of baked clay a major source for the copper used to manufacture from furnaces, pottery of the EC II period was found copper-based artifacts in the southern Aegean in the with obsidian tools and slag. Some 20 round buildings EBA.128 of schist slabs had apparently been erected to shelter Siphnos. Full publication of the results of interdis- smaller furnaces built inside them. Two have been ciplinary research on Siphnos has greatly transformed excavated. Just to the southeast of Skouries at Yeron- our impression of the importance of this island as a imos possible LBA pottery was excavated in associa- supplier of metals in prehistoric times.129 Wide-rang- tion with a rectangular structure on a hilltop; there ing investigations of traces of ancient mining on the was reason to suspect that an EBA settlement associ- island have demonstrated that already in the earlier ated with the mining operation at Skouries would third millennium Siphnos served as a significant have been located on the nearby bay of Ayios loannis. source of silver and lead in the Aegean.'30 A consid- At Tsoulis, 2 km to the south, traces of an ancient erable amount of EBA pottery, and at least a small surface mine suggest prehistoric copper mining; a amount of FN, has been found on the surface in the fragment of an EC II jar came from surface deposits vicinity of the mines at Ayios Sostis, in the northeast as did stone and obsidian tools. part of the island; most of the ceramics were mixed Isotopic compositions of copper samples from the with debris from the mining, but a number of datable mines, furnaces, and various copper deposits on the pieces come from within a deposit of mining debris island match those previously determined for the so- (Versatz) inside one mine (no. 2) itself. Artifacts from called "Kythnos Hoard," which is typologically of EC Ayios Sostis have now been published in detail; they II date. The chemical composition of copper from include objects directly connected with metallurgy, slags and ores suggests that arsenical copper of the such as crucibles and tuyeres, as well as domestic sort from which objects in the "Kythnos Hoard" were vessels.'31 Several other lead and silver mines in other

128 125 Of all the islands of the western Cyclades,the archae- "Sourcesof Metals"266-67. Presentevidence suggests ological resources of Seriphos remain the most poorly that Kythnoswas a principalsource not only for the Cyclades known.Lately an EC sherd on a copper slag heap at Alivassos but also for Crete in the EBA, although data for Crete is still has been mentioned; see G. Weisgerber,"Bemerkungen zur exiguous. prihistorischenund antiken Bergbautechnik,"in Silber,Blei 129 H. Matthius, "Sifnosim Altertum,"in Silber, Blei und und Gold 112, n. 28. Gold 17-58, has summarizedcomprehensively the evidence 126 Z. Stos-Gale,N. Gale, and A. Papastamataki,"An Early for prehistoricarchaeological discoveries on the island be- Bronze Age Copper Smelting Site on the Aegean Island of fore the mid-1980s. Various finds from older excavations Kythnos,"in J. EllisJones ed., Aspectsof AncientMining and are reillustrated;a selection of obsidian tools and debitage Metallurgy(Bangor 1988) 23-30, and O. Hadjianastasiou recoveredat the sites of Voriniand PlatyYialos in the course and S. MacGillivray,"An Early Bronze Age Copper Smelting of the recent investigationsby the Max Planck Institute is Site on the Aegean Island of Kythnos.Part II: The Archae- illustratedin fig. 9, p. 29. On artifactsfrom Platy Yialo, see ological Evidence," in Aspects of Ancient Mining and Metal- also E. Pernickaet al., "Alte Blei-Silber-Verhiittungauf Sif- lurgy 31-34. Two '4C dates are associated with charcoal nos," in Silber,Blei und Gold 197; for Vorini, see also G.A. inclusionsin copper slags; see R.E.M.Hedges et al., "Radio- Wagnerand G. Weisgerber,"Andere Blei-Silbergruben auf carbonDates from OxfordAMS System: Archaeometry Date- Sifnos," in Silber,Blei und Gold 168-69. For EBA pottery list 11,"Archaeometry 32 (1990) 226. from Ayios Andreas, see Vaughanand Wilson(supra n. 20). Also of interest is the recent reconsideration(and rejec- Matthausreports finds that are possiblyof FN date at several tion) of claims of pre-Neolithic occupation at Maroulaon sites, including Ayios Sostis (p. 30); Gropengiesser 1987 Kythnos: see Perkls 1990 (supra n. 7) 125-26; finds may (infra n. 131) 13, 35, n. 291, also mentions materialthat may point instead to Neolithic or EBA activity.Earlier analysis by be typologicallyof the later Neolithic from the site of Ak- J.F. Cherry,"Four Problems in CycladicPrehistory," in Davis rotiraki (Gazetteer312) in the southeast part of the island. and Cherry (supra n. 26) 25-32 reached the same conclu- With it were terracottaimplements probably associated with sion. metallurgy. On the 127 "KythnosHoard," see J.L. Fitton, "Esse Quam 130 The extent to which its gold deposits were exploited, if Videri:A Reconsiderationof the Kythnos Hoard of Early at all, in prehistorictimes remains unclear. CycladicTools," AJA 93 (1989) 31-39. The discoveryof a 131 H. Gropengiesser, "Siphnos, Kap Agios Sostis: Ker- significant number of metal tools in recent excavations at amische prahistorische Zeugnisse aus dem Gruben- und the Caveof Zaswould tend to support Fitton'sand Renfrew's Huittenrevier,"AM 101 (1986) 1-39; AM 102 (1987) 1-54. suggestion that the Cave of Zas was the original findspot of These papers fully describe material found at Ayios Sostis this hoard (but also see infra n. 177 for bronzes from the both on the surface and inside mine no. 2; for commentary KoronasCave on Naxos). on prehistoricartifactual material presented in Silber,Blei 1992] THE ISLANDS OF THE AEGEAN 729

parts of the island may also have been in use in A street and a courtyard (the latter with a possible prehistoric times.132 sacred stone or baetyl in its corner) offered access to The scarcity of slags has raised the possibility that both shrines.'35 Within the sanctuary, benches and ore from the mines of Siphnos was smelted elsewhere, platforms served for the display of objects. A number in areas where fuel supplies were more abundant. of drain-tile fragments may derive from terracotta Whatever the case, lead-isotope analyses of lead and channels that were used to conduct water from the silver finds from the Cyclades suggest that a substan- roof of the east shrine; interior walls of the sanctuary tial amount of the Siphnian metals was being used in were extensively coated with white plaster, perhaps in the EBA, even at Ayia Irini, which was located near a part painted red; and coarser plain plaster is likely to competing source at Laurion.'33 Continued use of the have been employed to seal its roof. A few fragments mines of Siphnos in the LBA has been postulated on of plaster with painted designs almost certainly derive the basis of thermoluminescence dates of slag samples ultimately from levels earlier in date than the building at Ayios Sostis, but no lead-isotope analyses of LBA of the sanctuary. The buildings as a whole were well artifacts have yet indicated a Siphnian provenance. constructed, but there is no good reason to believe Melos. The new British School excavations at Phy- that the cult center was directly controlled by those lakopi and an intensive survey of parts of the island who resided in the contemporary Mycenaean mega- (1974-1977) resulted in a burst of publication in the ron at Phylakopi. early and mid-1980s.'34 Pottery ranged in date from LH I/LM I to LH IIIC, A report on the results of excavations in the My- although pre-LH III material is present only in small cenaean sanctuary at Phylakopi represents the most quantities as stray finds in later contexts.'36 The as- recent contribution to the final publication of that sortment of terracotta figures and figurines from the campaign. The sanctuary consists of two parts, an east sanctuary, probably both votives and cult images, is and a west shrine: the latter constructed first, during unparalleled elsewhere in prehistoric Greece: not only a time contemporary with LH IIIA on the Greek female, animal, and chariot-group figurines of well- mainland; the former added in LH IIIB just inside known Mycenaean varieties were recovered, but also an extension to the main fortification wall of the female figures with bell-skirts of Cretan types, wheel- settlement. The entire complex was severely damaged made bovine figurines, a possible fish rhyton, bird- at a time when LH IIIC styles were current (in phase askoi, a crudely fashioned large female figure with 2b), approximately contemporary with the destruc- explicitly represented genitalia, and a large (45 cm), tion of the citadel at Koukounaries on Paros; parts of exquisitely modeled and decorated figure, the so- the complex were reused, only to be abandoned finally called "Lady of Phylakopi." Most unusual are several in a later stage of LH IIIC. capped male figures with elongated bodies and ex-

und Gold, see Gropengiesser 1987, 53-54. Among the pot- 200-11; and Gale, Stos-Gale,and Davis (supra n. 18) 389- are tery numerous examples of bases with matt, or leaf, 406. The process by which Laurionafter the EBA replaced impressions and a possible fragment of a Cycladic frying Siphnos as the dominant source of lead and silver in the A pan. pattern-burnishedsherd and a fragmentof a "cheese- Aegean is discussedat length in the latter paper. are the clearest of FN. For an obsidian pot" examples "Sali- 134 See C. Renfrew and M. Wagstaff eds., An Island Polity: from see G.A. agos point" Ayios Sostis, Wagneret al., "Early The Archaeology of Exploitation in Melos (Cambridge 1982); Bronze Lead-Silver and Age Mining Metallurgyin the Ae- C. Renfrew, The Archaeology of Cult: The Sanctuary at Phy- gean: The Ancient Workingson Siphnos,"in P.T. Craddock lakopi(London 1985); see also specializedstudies in Prehis- ed., Scientific Studies in Early Mining and Extractive Metal- toric Cyclades,and, for analyses of lead and litharge, Stos- lurgy (BMOP 20, London 1980) 81, pl. 1. See also H. Gro- Gale and Gale (supra n. 39) 87, fig. 20. A third volume, pengiesser,"Prihistorische Keramik von Siphnos,"in Kollo- reporting results of excavationsin the town elsewhere than quium zur aigischen Vorgeschichte (Schriften des deutschen in the sanctuary,is in preparation. Meanwhile,for a sum- Archdologen-Verbandes 9, Mannheim 1987) 63, where fur- mary of pertinent results, see C. Renfrew, "Phylakopiand ther reference is made to surface finds from the site of the Late Bronze I Period in the Cyclades,"in TAWI, 403- Akrotiraki;the oldest materialthere is compared to the FN 21. of Kephalaon Keos. Impressed wares from Ayios loannis, Claimsof Palaeolithicand Neolithic tools of obsidianhave also in the southeast, are compared to ThessalianEN types, been reported by the popular press but remain unsubstan- and it is claimed that they represent the earliest Neolithic tiated ("Chronique"1985, 841). ceramics in yet recognized the Cyclades,a claimthat remains 135 See P. Warren,"Of Baetyls,"OpAth 18 (1990) 203-204. to be substantiated. 136 Several objects from earlier excavations at Phylakopi 132 Near a mining shaft at Ayios loannis, the outline of a are publishedfor the first time or republished.Most notable ship pecked on the bedrockfinds a parallelon Naxos among are an ivory ring (CMS I, 410; see Renfrew 1985 n. the EC [supra probable rock-carvingsfrom Korphi t'Aroniou;see 134] 295-96) and various figurine fragments (276). On the n. Weisgerber(supra 125) 107, 109, fig. 102. small finds in general, see C. Renfrewand J.F. Cherry,"The 133 See E. Pernickaand G.A. Wagner,"Die metallurgische Other Finds,"in Renfrew 1985 (supra n. 134) 299-359. Bedeutung von Sifnos im Altertum,"in Silber,Blei und Gold 730 JACK L. DAVIS [AJA 96 plicitly rendered genitalia. There is some parallelism stone assemblages from inside and outside the sanc- between types found in the two shrines. Nearly iden- tuary delineated certain patterns: in the course of LB tical pairs are represented in several instances, and III, as the frequency of domestic use and discard of within the west shrine, the distribution of figures and obsidian declined within the settlement as a whole, figurines on platforms in its corners may indicate a more obsidian was brought to the shrines and its use male-female dichotomy in cult practices. was increasingly incorporated into ritual activities. Probable votive gifts comprise sealstones, a scarab Analysis of faunal remains, however, did not detect of a Syro-Palestinian type, a small human head patterns in consumption that would point to special worked in sheet gold by repousse, two bronze figu- sacrificial practices. rines of a "smiting god" or "Reshef," their best par- Additional fieldwork sponsored by the British allels in Syria and Palestine, and a terracotta mold for School in 1989 intensively surveyed a number of sites the manufacture of flat chisels. Other metal objects (most of them already known from earlier survey), included projectile points, knives, an attachment in including several of the Early Cycladic period.'37 New the form of a small bird, rings, awls, pins, and lead fragments of frescoes retrieved in the recent excava- clamps. Beads, mostly of glass paste, were common in tions have been described, and new reconstructions the sanctuary but rare elsewhere; two were of faience. have been proposed for wall paintings previously dis- One fragment of ivory probably belonged to a handle, covered at Phylakopi (fig. 13). Finally, a considerable another to a sword or dagger pommel. Numerous quantity of unpublished pottery of Cycladic and Mi- fragments of ostrich shell may all derive from a rhy- noan MBA types from older excavations on Melos has ton, fabricated from a single egg. Pieces of tortoise also been catalogued and illustrated.'38 shells preserve drill holes and served as sound-boxes for a lyre. Two triton shells may have provided mu- The Southern Cyclades'39 sical accompaniment for ritual activities. Various spin- los. Explorations on the island of los since 1983 ning and weaving equipment was probably not have made the first substantial contributions to our associated with worship. understanding of the prehistory of the island since Among stone finds are fragments of vessels of Cre- early in this century; new sites have been located and tan New Palatial types, pedestaled lamps or censers new information gathered about those previously in local tuff with chevron decoration, and a pendant charted.140 Of greatest significance has been the dis- in semiprecious stone of a reclining lion or dog, pos- covery in 1984 of the site of , north of ancient sibly of Egyptian manufacture. Ground stone tools los near a large natural harbor on the west coast of were manufactured primarily from local volcanics; the island (fig. 14).'41 Excavations were started in 1986 marble slabs were imported. A comparison of chipped following systematic surface collection and have con-

137 On these surface collections, see AR 36 (1990) 67, and related to the frescoes of Xesti 3 at Akrotiri, may have Annual Reportof the Managing Committee,the BritishSchool adorned the wallsof the PillarCrypt at Phylakopiand nearby at Athens1989-1990, 26. The results will be publishedby R. rooms. and G. Sanders. In associationwith this a Catling project, 139 Elsewherein the southern Cyclades,the publicationof republicationby R. Arnott of the finds from the 1897 exca- an EBA bracelet from is of interest; the ar- vationsat Pelos is in preparation.I am grateful to Mr.Arnott chaeological resources of the island itself remain largely for this information; a summary of his paper, "The Early unexplored. See R. Arnott, "AnEarly Cycladic Bracelet from CycladicI Cemeteryof Pelos, Melos:Approaching a Repub- Pholegandrosin the FitzwilliamMuseum," BICS 36 (1989) licationof the Excavation,"will be publishedin BICS as part 117-26. of the minutes of the MycenaeanSeminar for 1992. Another 140 See also AR 33 (1987) 49 for a report of a prehistoric member of this team has recently describedthe distribution settlement and EBA cemetery located at Halara Mangana- of Keros-Syrossites on the island and the characteristicsof riou in the south of the island. R. Arnott, "EarlyCycladic lithics from various Melian sites, within the context of a Objectsfrom los Formerlyin the FinlayCollection," BSA 85 reconsiderationof the causes that lay behind the emergence (1990) 3-14, has recentlypublished several marble figurines, of a prismaticobsidian blade technologyin the Aegean EBA. marblevases, obsidian blades, and a lead figurine (of dubious See T. Carter,"Blood and Tears: A CycladicCase Study in authenticity),formerly in the collectionof George Finlayand Microwear Analysis. The Use of Obsidian Blades from now in the collectionof the British School at Athens. Gravesas to in the of the Sixth M. Razors," appear proceedings 141 Marthari,"Ix&Qxog: Evacg7rToroxvxhcaLx6g otx- InternationalFlint Symposium,October 1991, Madrid. Lato6gonly Io," in I6evya (supran. 89) 97-100. Myaccount 138 For PhylakopiMBA pottery,see MinoanInfluence 358- is based on this publication,supplemented by information 75. For frescoes, see Morgan (supra n. 39) 252-66. Morgan concerning the results of the 1990 and 1991 excavation suggests that a fresco composition involving presentationof seasons, kindly provided by MarizaMarthari. cloth to a goddess with associated monkeys, thematically 1992] THE ISLANDS OF THE AEGEAN 731

r-LI-

I' / /"/

I --S ~

-> II

rr'~ P \\ / / ,*'- ..- \ N-

r r\ -

Fig. 13. Phylakopi, Melos. Presentation scene. Wall painting from the Pillar Crypt. (CourtesyLyvia Morgan) tinued through 1991. Squares, a road, and parts of and chipped stone, bronze and lead objects, and two contemporary building complexes have thus far been marble figurines similar to the Apeiranthos type. The exposed. Beneath surface levels was a deep destruc- walls of the structures are preserved to an extraordi- tion layer, which rested on the floors of the buildings nary height, in places nearly 3 m; most of the build- and dates both their destruction and final period of ings had a second story. All architectural remains use to the EBA (fig. 15). Finds include EC II sauce- excavated thus far appear to belong to the Keros- boats (fig. 16) and saucers, loomweights, a large num- Syros phase of EC II; no pottery diagnostic of the ber of spindle whorls, many implements of ground Kampos group or of the succeeding Kastri group has

Z . . . . . 4w ...... oli o

low

14 ON OR a 'a Q-1 N': ''0 -X;

OZ,]?, omb? [Im io

Fig. 14. Skarkos,los. Generalview. (CourtesyMariza Marthari) 732 JACK L. DAVIS [AJA 96 ,: -_::::::- ~--~.?;;::::~:s: -

--:~i:i-'::::::~

- --:Ci:::::~iii-iiii:-~_::: ::::~.: i. :i~i' .:- i-: :_:.r:.: : -:: :::?:-::-:: '::'' :::::::?:~ ::::::; :i~L~~~~~-_~8~99~1 III ~rp~ -rs~?ns~~?:~ge~a~k~li~%gpge~~~:~g i-Zi~~:__i : : ii:iii:

:?--illi~--i;i-;-::~:~_:i::-:::::::_;i:-_~i :-i:-::~:------: :: i:-'::lil-ixi-i-i-i-i.i~ii:i---~ :i- .-.:::::::-:_:,?,:os-: ?:::::-:_:::ii~~r:;-::: ~i8~i~iia-i~ii~i::i-:-i:~i,_~i~::~ ~?liBa~?lssgsh*-3*~:.:::?~~xlapaa*8a~.r? : ;:i:::::-::::-:-::::; :-::;:::~:::::?

i-i9~i-o:i

-i:-:- ":i:::: :

-~~-:--

Fig. 15. Skarkos,los. Wallsof EBA buildings.(Courtesy Mariza Marthari)

been recognized. After the abandonment of the set- tlement, several cist and pithos graves of transitional MC/LC I date were dug into the EC debris. Thera and . In recent years much progress has been made toward setting the site of Akrotiri within the context of the overall settlement history of the island of Thera.142 Additional prehistoric settle- ments have been located both on Thera and Thera- sia.'43 It is now clear that these islands were densely settled in the centuries before the final eruption of the volcano and that their settlement pat- terns at the time of the eruption were more similar to those of Crete in the New Palace period than to those of other Cycladic islands, such as Melos and Keos.144 Limited excavation has continued at Akrotiri. Room 7 of the West House has been investigated and there have been stratigraphical tests north, south, and east of the West House. The western facade of Xesti (Ash- Fig. 16. Skarkos,los. Fragmentsof EC II sauceboats.(Cour- lar) 5 has been cleared as has a terraced platform to tesy MarizaMarthari) its west. Excavation has also progressed in Xesti 3 (fig.

142 Resultsof the first 20 years of work at Akrotiriwill be many of the papers included in TAW III, where a current summarizedin C. Doumas ed., Axpwnrlpt&O ag: Eixoat overviewof the stratigraphyof the site as a whole is offered Cpevvag. by C. Doumas,"Archaeological Observations at AkrotiriRe- XOvtEa EZvMxreptorpara-HIpo3Ajipara-Hpooo7rrt- xg; (in press). lating to the VolcanicDestruction," in TAW111.3, 48-50. 143 Resultsof researchprior to 1980 are convenientlysum- 144 Evidencefor the overallpattern of settlementon Thera marizedwith relevantbibliography in C.G. Doumas, Thera: and Therasiahas been discussedby Davisand Cherry(supra Pompeiiof the AncientAegean (London 1983). For a later n. 41), with a gazetteerof sites describedin print up to 1989 summaryincluding the most importantbibliography of the on pp. 190-91. This gazetteer should be supplemented by early 1980s, see R. Barber, The Cycladesin the BronzeAge the addition of newly located sites describedby M.A. Aston (Iowa City 1987) ch. 8. More recent work is discussed in and P.G. Hardy, "The Pre-MinoanLandscape of Thera: A 1992] THE ISLANDS OF THE AEGEAN 733

.?... ?.

I, I '-', i

: 'I

~~-Jl

? / " ' .... . ?I .: .. I/ / ." . l . ,.. , •,???~ ,/ . •,,< . "4, ' '7 ' ' .

Fig. 17. Akrotiri, Thera. Isometric drawing of the eastern part of Xeste 3. (After C. Palyvou,in TAWIII, 52 fig. 6)

17), within rooms 3B and 14.145 Other subsidiary House of the Ladies, where investigation was resumed explorations have demonstrated that the level of the in 1987 after a hiatus of some 15 years. Many con- roads within the settlement was raised prior to the structional details of the building are now clarified; final volcanic destruction of the site; the removal of the original interpretation of room 9 as a lightwell (a volcanic debris west of room Delta 15 (fig. 18) revealed unique architectural feature at Akrotiri) is supported. hollows left by wooden furniture embedded in the Of special interest among recent finds is part of a ash.146 Excavation has been most extensive in the horns of consecration, cut from volcanic stone.'47 Re-

PreliminaryStatement," in TAW 111.2, 348-60. Places in 1990, 108-109. Beneath the plateia west of Xesti 5, several which the pre-eruption surface of the island is visible have bedrock-hewnchambers, perhaps originally chamber tombs, begun to be mapped and the morphology of the pre-erup- appear to have been reused by Middle Cycladicpotters for tion island reconstructed; in several places, pre-eruption cleaning and storing clay. Stone vases and other EC artifacts artifactshave been reported for the first time on the surface were recovered in the course of their excavation, some of of the pre-eruptionsoils, notablyin a quarryat Megalohori them perhaps intentionallyremoved from the chambersby and on the calderaslopes at Megalo Vouno, but the material the potters: see Prakt 1985, 171-75. For Xesti 3, see Prakt has not yet been adequatelyexamined or closelydated. Finds 1987, 244-45; Ergon 1990, 113. of later MC and LC date from the Mavromatisquarry near 146 For roads, see Prakt 1985 (supra n. 145) 175; for ex- Akrotirihave now been describedin more detail and appear cavationwest of room Delta 15, see Ergon 1988, 129. to extend over an area of at least 5 ha. On Mavromatis,see 147Prakt 1987 (supra n. 145) 245-54; Ergon 1990, 109- C. Televantou, ArchDelt37 B' 11. of earlierrelief and other frescohave "OguvXicEMcvaQopt•trTl," Fragments painted (1982) 358-59; and Minoan Influence 358. been found embedded in the floors of the upper story,and 145 For tests outside the West House, see C. Doumas, "Av- seem to have been recycled from older ruined structures. Prakt 1985, 169-70; "Av- See also T. "The of the House of the aoxa(l Oil~~ag(AxQ••T•iQL)," Sali-Axioti, Lightwell Prakt 1987, 241-44; Ladiesand Its Structural in TAW 437-40. aoxa(flnOiljag (AxQ(r•Tv)lov)," Ergon Behaviour," III.1, 734 JACK L. DAVIS [AJA 96

?II;

I I /~/ I I I II I/:-) > ~ `~U/b/ IIIII1I/ I IMr~?~

=-\

/IX\ \\ \I I ~"

- \\

r;-

Fig. 18. Akrotiri,Thera. Isometricdrawing of the southern entrance of sector Delta, rooms 15 and 16. (After C. Palyvou,in PrehistoricCyclades, 141 fig. 6) cently published frescoes include depictions of three More than 75% of the total EBA pottery comes, in youths and an adult from Xesti 3, room 3B, which fact, from the final destruction levels of the settle- may represent a male initiation ritual.148 ment, presumably the result of recycling of various Akrotiri was settled already in the Neolithic and kinds (e.g., through incorporation in mudbricks or was inhabited throughout the EBA.'49 The EBA set- roofing). All major phases of the EC period appear tlement appears to have been of substantial size. Evi- to be present, and there are types characteristic of the dence for Neolithic occupation remains slim; 16 Kampos and Kastri assemblages. The overall distri- sherds have shapes and decoration paralleled in finds bution pattern of the Neolithic and EC sherds sug- from Saliagos. Sherds of the EBA are plentiful. Al- gests that the focus of the earliest occupation at most all of the material is highly fragmentary and its Akrotiri was in the southwestern part of the excavated dating, for the most part, depends on stylistic criteria. area, in the vicinity of Xesti 3.

148 For illustrations,see C. Doumas, "Avaoxacl) Oeigag 149 P. Sotirakopoulou,"Early Cycladic Pottery from Akro- (AxQOprijLov),"Prakt 1986, 208-11; Ergon 1988, 130, fig. tiri,"BSA 81 (1986) 297-312. The extent of the EBA occu- 103; Ergon 1989, 116, figs. 109-11. Final study of the fres- pation had been describedbefore this but the evidence had coes from the West House has now been completed; a vol- not been set forth in detail. See also Sotirakopoulou,"The ume will be published shortly by the Greek Archaeological Earliest History of Akrotiri: The Late Neolithic and Early Society: C.A. Televantou, AxQwigrj(t Orjag: OL TotXoyga- Bronze Age Phases,"in TAW111.3, 41-50. These two papers A Otxiag. are and should be read icE9trrg vrLtxrj complementary together. 1992] THE ISLANDS OF THE AEGEAN 735

Akrotiri was destroyed twice in the initial stages of Both sites were discovered (and badly disturbed) dur- the LBA.'15 A crisp description of the stratigraphy of ing modern quarrying. the West House and of soundings beneath its floors There should no longer be any debate over the has provided valuable details concerning the later dating of the final destruction of Akrotiri or concern- history of Akrotiri.'151' On the floors of the West House ing its relative position in the Minoan chronological lay deposits of the Volcanic Destruction Level (com- sequence.'56 Akrotiri was abandoned in LM IA at a monly now abbreviated VDL); beneath them were time before the LM IB style on Crete had developed. homogeneous deposits that had been used as fill. The explosion of the Santorini volcano appears to These deposits appear to derive from debris accu- have occurred as a single event; there was no appre- mulated in a destruction that wracked the site at a ciable delay between the depositions of the various very early stage of the LBA. The fact that so many strata of volcanic ejaculates from its eruption. It is buildings were damaged and subsequently required highly unlikely that the LM IB style developed in the extensive reconstruction suggests that the destruction interval between the abandonment of the settlement was caused by an earthquake. The deposits from the and the final eruption of the volcano. West House and others of similar character elsewhere Evidence from elsewhere in the Aegean leads to are commonly said to belong to the Seismic Destruc- these same conclusions. It has been established for tion Level (SDL) of Akrotiri.'52 some years that the eruption of the Theran volcano, Still further below the rooms of the West House as represented by volcanic ash deposited in soil at intact vases and sherds of late MC types have been Phylakopi on Melos, occurred at a time when LM IA found in pits of uncertain function that were cut into and LC I styles were current, and not even at the very the bedrock.'53 Earlier stages of the MC period, in- end of the LC I period.'57 Recent excavations both in cluding types characteristic of Phylakopi I, are rep- East Crete and in the Dodecanese have revealed layers resented, but not in well-stratified deposits.'54 This of Theran ash stratified over deposits of LM IA and phase in the island's history has, however, been more beneath LM IB floors, with pieces of pumice from the systematically explored in excavations in two other eruption in LM IA contexts. The cumulative evidence locations, Ftellos and the Karayeoryis quarries.'55 in all cases suggests that the fall of ash occurred before

150 M. Marthari,"The Destructionof the Town at Akrotiri, also A. Papagiannopoulou, "Some Changes in the BA Pot- Thera, at the Beginning of LC I: Definition and Chronol- tery Production at Akrotiri," in TAW III.1, 57-66. Most of ogy," in Prehistoric Cyclades 119-33; and C. Palyvou, "The the material comes from deposits of the SDL in trench A Destruction of the Town at Akrotiri, Thera at the Beginning between Xesti 2 and room Delta 20, as described by Marthari of LC I: Rebuilding Activities," in Prehistoric Cyclades 134- (supra n. 150), but included are finds of Phylakopi I style 47. from the so-called Sacrificial Fire Deposit (S. Marinatos, '51 M. Marthari,"The Chronology of the Last Phases of Thera III, 19-24) and from soundings beneath the West Occupation at Akrotiri in the Light of the Evidence from House. Papagiannopoulou also catalogues the small amount the West House Pottery Groups," in TAW 111.3, 57-70. of mainland Gray Minyan pottery that has thus far been 152 It was only in the course of the rebuilding of the settle- recognized in pre-VDL deposits at Akrotiri, describes a se- ment in the wake of this SDL that the walls of the houses lection of Minoan imports of MM II and later date, and were decorated with pictorial frescoes. The SDL and the discusses in some detail the process of Minoanization of the general rebuilding of the town that followed it must not be local ceramic industry. confused with the earthquake damage and more ad hoc 154 A selection of material has been recently discussed by reconstruction that occurred soon before the final volcanic Sotirakopoulou (supra n. 149). destruction of the settlement. On this point, see E.N. Davis 155 For a preliminarypresentation of finds from the Ka- (supra n. 39) 226, and comments by C. Renfrew, in TAW rayeoryis quarry, see Minoan Influence 321-23; a full pub- III.1, 70. Most recently, P.M. Warren, "A New Minoan De- lication of this material by M. Marthari will appear in BSA. posit from , c. 1600 B.C., and Its Wider Relations," A report on the 1981 season of excavation at Ftellos has now BSA 86 (1991) 339, has drawn attention to stylistic similari- appeared: M. Marthari, "Oil)a: rXIkkog,"ArchDelt 36 B' ties between material from the SDL and pottery in a deposit (1981) 373. from Knossos that he dates to a newly defined transitional 156 See C. Renfrew, "Summaryof the Progress in Chro- MM III/LM IA phase. His suggestion that the same earth- nology,"in TAW111.3, 242. quake was responsible for the destruction of Knossos, Ak- 157 See A.C. Renfrew, "Phylakopiand the Late Bronze I rotiri, and other southern Aegean sites, such as Ayia Irini Period in the Cyclades,"in TAWI, 412-16; J.L. Davis and on Keos, cannot be entirely correct (supra n. 41). J.F. Cherry, "Phylakopi in Late Cycladic I: A Pottery Seria- 153 Some pottery of MC character from Akrotiri is analyzed tion Study," in Prehistoric Cyclades 148-61; and Davis and by A.G. Papagiannopoulou, in Minoan Influence 26-69; for Cherry (supra n. 41) 198. additional observations on the technology of production, see 736 JACK L. DAVIS [AJA 96 the introduction of the LM IB style. Mainland Greek volcanic events. Volcanic fallout may also affect the imports to Thera support this chronology. The stylis- growth rings of trees. The study of cores and of tree- tic character of the LH I pottery from Akrotiri places rings (both in the United States and Ireland) suggests the abandonment of the settlement some time before that a major volcanic event occurred in the later 17th the beginning of LH II, which in turn appears to have century, and neither the ice cores nor tree-rings in- begun earlier than LM IB.1'58The eruption of the dicate a major eruption ca. 1500. Recently reported Theran volcano can consequently not be held directly '4C dates of samples from short-lived plant matter responsible for the well-known and widespread ho- retrieved from the VDL at Akrotiri in the main seem rizon of LM IB destructions on Crete.159 to support a date in the 17th century for the eruption. The absolute date of the Theran eruption continues There are very few artifactual synchronisms between to be far more controversial, and is bound up with Greece and Egypt during the early LBA. The dates arguments over the acceptability of a proposed "high" and contexts of relevant imports and exports are in chronology for the Aegean as a whole. For a number almost all cases disputable, and the evidence on which of years it has been obvious that many '4C dates from the traditional Aegean chronology has been based Akrotiri were too early to fit traditional chronologies seems capable of accommodating a higher chronol- that placed the volcanic destruction of the site ca. ogy.160 1500 B.C. Recently arguments have been proffered Many specialized studies of artifacts from Akrotiri for moving its destruction into the later 17th century have recently appeared. Metallurgical analyses have B.C. Supporting data can be divided into three cate- shed light on the nature of local industries and their gories: ice cores, tree-rings, and '4C dates. Fluctua- sources for raw material.161 Chipped stone tools were tions in acidity levels within ice cores from Greenland manufactured almost totally of imported Melian ob- in many cases are demonstrably correlated with major sidian. Yiali sources are hardly represented, but char-

158 See Y.G. Lolos, "On the Late Helladic I of Akrotiri, raphy through 1989 can be found in the two reviews by Thera,"in TAW111.3, 51-56. See P. Warrenand V. Hankey, Manning. Aegean Bronze Age Chronology (Bristol 1989) 97-98, and esp. Most recently,evidence pertainingto the date of the erup- 214 on the date of the Theran destructionrelative to LH II tion of the Thera volcanohas been criticallyexamined within and the relationshipof LH II to LM IB. the context of a general overview of the evidence for the 159 See J.S. Soles and C. Davaras,"Theran Ash in Minoan relativeand absolutechronologies of the Aegean, Anatolia, Crete: New Excavationson ,"in TAW111.3, 82-95; the Levant,and Egypt during the Bronze Age in S.W. Man- P.P. Betancourtet al., "Excavationsat :The Evidence ning, The Absolute Chronology of the Aegean Early Bronze for the Theran Eruption," in TAW 111.3, 96-99. From a Age: Archaeology, Radiocarbon, and History (Sheffield 1992). Cycladicperspective, it never did appear plausible that the Manning here proposes and discusses a revised absolute final desertion of Akrotiri had actually occurred at a time chronology for the entire Cretan Bronze Age and, in addi- contemporarywith LM IB in Crete. It has been difficult to tion to radiocarbondates from Akrotiri, reconsiders those imagine that the absence of LM IB styles at Akrotiricould availablefor Ayia Irini on Keos. Warrenand Hankey have be explained by a stylistictime lag between the inception of continued to defend a lower chronology, based largely on LM IB on Crete and its introduction to Thera, despite the artifactualsynchronisms, and would place the eruption of undeniablyclose relationshipbetween LM IB vase painting the Santorinivolcano in the later 16th century (supran. 158, and the Thera frescoes noted, among others, by L. Morgan, 215); see also P.M. Warren, "The Minoan Civilisationof "Morphology,Syntax, and the Issue of Chronology,"in Pre- Crete and the Volcano of Thera,"Journal of the Ancient historicCyclades 165-78. The massesof LM IB potteryfound ChronologyForum 4 (1990-1991) 29-39. Muhlyhas recently both at Phylakopion Melos and at Ayia Irini on Keos, even claimed that the artifactualsynchronisms between Greece farthernorth than Thera, suggest that commercewith Crete and the Near East support a low chronology;his arguments was frequent and that certain settlements in the Cyclades have been countered by Manning:J.D. Muhly, "Egypt,the were au courantwith the latest Minoanfashions both before Aegean, and the Late Bronze Age Chronologyin the Eastern and after the Theran eruption. Mediterranean:A Review Article,"JMA 4 (1991) 235-47; 160 The bibliographyon this subjectis far too extensive to S.W. Manning, "Response to J.D. Muhly on Problems of be summarizedhere. For a reviewof the state of the contro- Chronology in the Aegean Late Bronze Age,"JMA 4 (1991) versy before TAWIII, see S.W. Manning,"The Bronze Age 249-62. See also Manning, "Thera, Sulphur, and Climatic Eruption of Thera: Absolute Dating, Aegean Chronology, Anomalies,"OJA 11 (in press). and MediterraneanCultural Interactions,"JMA 1 (1988) 161 Stos-Gale(supra n. 39) 276, 281, fig. 12 (copper);Stos- 17-82. For a balanced picture of the results of TAWIII as Gale and Gale (supra n. 39) 85-88 (lead and litharge). See they pertain to this issue, see S.W. Manning, "The Thera also Y. Bassiakos et al., "Provenance Studies of Theran Eruption: The Third Congress and the Problem of the Lead,"in TAW111.2, 337-45, who report lead ore (cerous- Date,"Archaeometry 32 (1990) 91-100; "The SantoriniErup- site) from the excavationsat Akrotiri. From the description tion: An Up-date,"JMA 2 (1989) 303-13; S. Hood, "The of the samples and their composition it seems more likely Third InternationalCongress on Santorini(Thera)," Kadmos that they are of litharge; see Stos-Galeand Gale (supra n. 29 (1990) 84-86; and S. Sherratt,"Fallout from the Aegean 39) 85-88. Big Bang,"Antiquity 65 (1991) 998-1001. All basic bibliog- 1992] THE ISLANDS OF THE AEGEAN 737

acteristically MBA/early LBA rhyolite denticulated success in defining the local products of Akrotiri.'67 bifaces are present.162 Ostrich eggs were fashioned The local character of the LC I ceramic industry at into rhyta.163There have been attempts to reconstruct Akrotiri emerges distinct from that of Crete and other the metrological system in use at Akrotiri on the basis islands, despite the fact that many aspects of Minoan of ceramic container capacity.'64 Artifacts associated technology were blended with earlier Cycladic pro- with cloth production imply that weaving at Akrotiri duction techniques.168 Even in the EBA, ceramic as- had, as at other Cycladic sites, become Minoanized semblages at Akrotiri were complex mixtures of through the introduction of the warp-weighted loom; imports from several locations with geologically dis- a peculiar lack of spindle whorls ties Akrotiri more tinct clays outside Thera.'69 closely to Crete than to her island neighbors.'65 The A tentative reconstruction of some aspects of the number of lead weights from the site has nearly dou- agricultural economy of prehistoric Thera is now also bled in recent years. It has been argued that the possible. Pre-eruption soils in the vicinity of Akrotiri fractional units within the system of measure on which were more mature in their development than previ- these weights are scaled find parallels both in Linear ously thought.'70 Analysis of palaeobotanical remains A and . It has also been suggested that at has begun, in particular of Spanish vetchling (Lathyrus Akrotiri, such weights, as well as rarer weights of clymenum),a crop well represented in the West House. stone, were used for the measurement of heavy goods, Examination of these pulses and the contaminants wool in particular.'66 and weed seeds found with them suggests that the Imported ceramics attest to considerable interac- various deposits of vetchling in the West House came tion between Akrotiri and other areas of the Aegean. from different fields, probably of quite small size. Chemical and petrological analyses have had some Informed speculation about the economic and social

162 T.D. Devetzi, "The Stone Industry at Akrotiri: A The- and Pigment Analysis," in TAW III.1, 459-69, offer tech- oretical Approach," in TAW III.1, 19-23; A. Moundrea- nological observations on locally produced ceramics, includ- Agrafioti, "Akrotiri: The Chipped Stone Industry," in TAW ing analyses of clay bodies, pigments, and firing temper- III.1, 390-406. Rhyolite bifaces are of a type previously atures. On deposits of talc and other minerals locally avail- recognized on Keos and Melos (see R. Torrence, "Other able to the residents of Thera before the eruption, see W.L. Silaceous Materials," in Davis [supra n. 38] 95-96. Thus far Friedrich and C.G. Doumas, "Was There Local Access to analyses of chipped stone appear to support Torrence's Certain Ores/Minerals for the Thera People before the Mi- position that obsidian trade at this time was not highly reg- noan Eruption? An Addendum," in TAW III.1, 502-503. ulated by central authority or in the hands of specialist 168 M. Marthari, "Investigation of the Technology of Man- craftsmen (R. Torrence, Production and Exchange of Stone ufacture of the Local LBA Theran Pottery:Archaeological Tools: Prehistoric Obsidian in the Aegean, Cambridge 1986); Consideration," in TAW III.1, 449-58; "The Local Pottery contra Barber (supra n. 143) 117-19, who has been dismis- Wares with Painted Decoration from the Volcanic Destruc- sive of Torrence'sarguments. tion Level of Akrotiri, Thera: A Preliminary Report," AA 163 J.A. Sakellarakis, "The Fashioning of Ostrich-Egg 1987, 359-79. Several specific shapes from the VDL have Rhytain the Creto-MycenaeanAegean," in TAWIII.1, 285- also been the subject of special investigations. See C. Gillis, 308, where eggs from Akrotiriand from Phylakopion Melos "Statistical Analyses and Conical Cups: A Preliminary Report are considered. from Akrotiri, Thera," OpAth 18 (1990) 63-93, and R.B. 164 L. Katsa-Tomara, "The Pottery-Producing System at Koehl, "The Rhyta from Akrotiri and Some Preliminary Akrotiri: An Index of Exchange and Social Activity," in TAW Observations on Their Functions in Selected Contexts," in III.1, 31-40; C. Doumas and A.G. Constantinides, "Pithoi, TAW III.1, 350-60. Size, and Symbols: Some Preliminary Considerations on the 169 S.J. Vaughan, "Petrographic Analysis of the Early Cy- Akrotiri Evidence," in TAW III.1, 41-43. cladic Wares from Akrotiri, Thera," in TAW III.1, 470-87. 165 I. Tzachili, "All Important Yet Elusive: Looking for See also other studies by Vaughan (supra n. 20). Evidence of Cloth-Making at Akrotiri," in TAW III.1, 380- 170 S. Limbrey, "Soil Studies at Akrotiri," in TAW 111.2, 89. 377-82; but see also O. Rackham's skepticism concerning 166 A. Michailidou, "METQLX6o'oT'TXM xtI oCxCoeLgT(XlQ- the potential of agricultural production on the island to ay0oyilg aTo Atya(lo, OTarVYoTEQTI EtoXoil TOUXakXxo3," explain its prosperity in the Bronze Age and in particular Meletimata tou K.E.R.A. 10 (1990) 65-96; and "The Lead its density of settlement: "Observations on the Historical from Akrotiri: Weights The Archaeological Record," in Ecology of Santorini," in TAW 111.2, 384-91. On in-progress TAW III.1, 407-19. Michailidou also notes her forthcoming analyses of additional faunal remains from Akrotiri and publication of a potsherd from Akrotiri that has been incised preliminary data from the West House, see C. Trantalidou, with a record of commodities in Linear A; this is the first "Animals and Human Diet in the Prehistoric Aegean," in such document from the site. On the weights from Akrotiri, TAW 111.2, 392-405; on shell, including triton shells and see also Petruso (supra n. 39). murex, see L. Karali-Yannacopoulou, "Sea Shells, Land 167 V. Kilikoglou et al., "A Study of Middle and Late Cy- Snails, and Other Marine Remains from Akrotiri," in TAW cladic from in TAW Pottery Akrotiri," III.1, 441-48; and E. 111.2, 410-15; E. Aloupi et al., "Analysis of a Purple Material Aloupi and Y. Maniatis, "Investigation of the Technology of Found at Akrotiri," in TAW III.1, 488-90. Manufacture of the Local LBA Theran Pottery: The Body 738 JACK L. DAVIS [AJA 96

organization of landholding in prehistoric Thera may be analyzed in terms of general architectural mod- be possible in the near future."'7 els.'73 In this way a basic architectural vocabulary for Only about 1 ha of the site of Akrotiri has been the site can be compiled, one that serves not only to excavated. Local topography and test excavations sug- permit generalization about the architectural idiom gest that the site originally covered some 20 ha, and of Akrotiri but also encourages attempts to explain that its harbor probably lay in what is now the valley any variation from the expected.174 of Ayios Nikolaos, some 200-300 m west of the cur- The art of Thera continues to attract considerable rent excavations."72 Means of access, lighting and ven- attention from art historians, and justifiably so. But tilation, and drainage all appear to have influenced even the recent literature is already so voluminous the plan of the town and the overall density of its that only a separate review could do it justice.'75 Sev- habitations, as did a lack of open courts within the eral papers delivered at the most recent Thera and houses. One principal purpose of narrow alleys was the Aegean World Congress are of particular impor- to provision the interiors of buildings with light and tance.176 Paintings from the West House have been ventilation. The design of individual structures can newly reconstructed in light of architectural analyses.

171 See A. Sarpaki, "'SmallFields or Big Fields?'That Is 44 examines the relationshipbetween Cycladicpottery dec- the Question,"in TAW 111.2,422-31; also A. Sarpakiand oration and fresco painting. G. Jones, "Ancient and Modern Cultivation of Lathyrus C. Televantou,"New Light on the West House Wall-Paint- ClymenumL. in the Greek Islands,"BSA 85 (1990) 363-68. ings," in TAWIII.1, 309-24, offers new reconstructionsfor Recent entomological studies of material from the West the paintingsof the West House. The interpretationof the House are also of interest and have revealedspecies not now iconographyof the West House paintingsremains a popular native to Greece but common in the Near East. See Ergon topic. See especially L. Morgan, The Miniature Wall Paint- 1989, 117. ings of Thera: A Study in Aegean Culture and Iconography 172 C. Palyvou,"Notes on the Town Plan of Late Cycladic (Cambridge 1988). S. Hiller has seen the miniature West Akrotiri,Thera," BSA 81 (1986) 179-94. Palyvouis cautious House frescoes as a reflection of a vanished Minoan epic in estimating the size of the town, noting only that it was poetry:"The MiniatureFrieze in the WestHouse-Evidence large "by the standardsof the time." This becomes clearer for Minoan Poetry,"in TAW III.1, 229-34. The localities when the likely area of the site is compared with that of representedin these same frescoeswas reconsideredby J.W. contemporaryAegean centers: see discussionby Wiener in Shaw, "BronzeAge Harboursides,"in TAW III.1, 420-36, TAWIII.1 (supra n. 21) 129-31. by J.A. MacGillivray,"The Therans and Dikta," in TAW 173 C. Palyvou,"Architectural Design at Late CycladicAk- III. 1, 363-69, and by G. Heiken, F. McCoy,and M. Sheridan, in TAW and of rotiri," III.1, 45-56; Axporjlpt OjeIa: Otxo6otmxt~ "Palaeotopographic Palaeogeologic Reconstruction Minoan in TAW rtXVrlxat oQpoAoytxadorotXEita OrTj YOrEpoxvxAa6tx7 Thera," 111.2, 370-76. W.-D. Niemeier, aXLTExtlroxvt (Diss. Athens PolytechnicUniv. 1988). "MycenaeanElements in the MiniatureFresco from Thera," 174 In this regard, Palyvouhas emphasizedthe specialchar- in TAWIII.1, 267-82 argues strongly against the presence acteristicsof Xesti 4, includingits ashlarfacades with courses of Mycenaeaniconography in the miniature frescoes of the diminishing regularly in height from bottom to top; most West House, and suggests instead that they illustrate the masons' marks found at Akrotiri come from this building MinoanThalassocracy. On the interpretationof the Xesti 3 alone (see discussion in TAW III.1, 56). For other recent frescoes, see especiallyMorgan (supra n. 39), and N. Mari- discussionsof specific features of Theran architecture,see natos, "Minoan-CycladicSyncretism," in TAW III.1, 370- C. Palyvou, "Observationssur 85 fenetres du cycladique 76. J. Vanschoonwinkel,"Animal Representations in Theran recent ~ Thera," 123-39, and A. Michailidou,"The Settle- and Other Aegean Arts,"in TAW111.1, 327-47 has surveyed ment of Akrotiri (Thera): A Theoretical Approach to the animal representationsat Akrotiri. Function of the Upper Storey,"293-306 in Darcque and Other recent studies of the iconography of the Thera Treuil (supra n. 105). frescoes include: N. Marinatos,"Role and Sex Division in 175 For a catalogue of the various frescoes and a review of Ritual Scenes of Aegean Art,"Journal of Prehistoric Religion major studies of them through 1988, see S.A. Immerwahr, 1 (1987) 23-34 (West House and Xesti 3); "A Puberty Rite Aegean Painting in the Bronze Age (University Park, Pa. 1990) at Thera: Evidence from New Frescoes,"Journal of Prehis- 185-88, and Immerwahr'sown discussionin ch. 4. An atlas toricReligion 3-4 (1989-1990) 49-51 (Xesti 3); C. Doumas, illustratingthe Theran frescoes accompaniedby brief infor- "Conventionsartistiques 'a Thera et dans la Mediterrande mativetexts by C. Doumas has been promised by the Thera orientale Al''poque prehistorique,"in P. Darcqueand J.-C. Foundation.I thank D.A. Hardy for this information. Poursat eds., L'iconographie minoenne (BCH Suppl. 11, Ath- 176 Several papers in TAW III considered the local char- ens 1985) 29-34; N. Marinatos,"An Offering of Saffron to acterof the Theran style of wall paintingin comparisonwith the Minoan Goddess of Nature: The Role of the Monkey Cretanfrescoes, those from other Cycladicislands, and from and the Importanceof Saffron,"in T. Lindersand G. Nord- Rhodes: among them, see E.N. Davis (supra n. 39); R. Laf- quist eds., Gifts to the Gods:Proceedings of the UppsalaSym- fineur, "Compositionand Perspectivein Theran Wall-Paint- posium 1985 (Boreas 15, Uppsala 1987) 123-32; and S. ings," in TAW III.1, 246-50. S.A. Immerwahr,"Swallows Morris, "A Tale of Two Cities: The MiniatureFrescoes of and Dolphins at Akrotiri:Some Thoughts on the Relation- Thera and the Originsof GreekPoetry," AJA 93 (1989) 511- ship of Vase-Paintingto Wall-Painting,"in TAWIII.1, 237- 35. C. Televantou, "Tca xot?oo Lt•t[TXr wr6 TlyVqTo'oroTLxT• 1992] THE ISLANDS OF THE AEGEAN 739

The walls of this house had been decorated at least Chora of the island), and concerted efforts to explore twice with wall paintings: of the earlier series only a Neolithic remains.'77 Only a decade ago the very ex- few aniconic elements have been recognized. Three istence of permanent Neolithic settlement could be new cities have been added to the miniature fresco in doubted.'78 Today, there are well-stratified finds from room 5, for a total of five in all. Fragments of a two locations: the western edge of the shore at Grotta previously unrecognized town belong to the west wall; and the Cave of Zas, high on the central massif of the on the north wall the well-known fragments that de- island. The sequence at Grotta appears to begin ear- pict a "meeting on the hill" and a "shipwreck" are lier than that in the Cave of Zas, and is in part con- associated with previously unpublished ships and a temporary with material from Saliagos.179 town on a hill; the "Nilotic landscape" of the east Finds from Grotta come from a rescue excavation frieze appears to have provided landscape context for near a place where more than half a century ago "sub- a third town; and the more completely preserved Neolithic" material was reported and from a sounding frieze of the south wall has been altered slightly at nearby Kokkinovrachos.8so These discoveries cou- through additions and repositionings. pled with the presence of "sub-Neolithic" material on Palati suggest that the Neolithic site was extensive. The Central Cyclades Pottery from the lowest levels inside the main cham- Naxos. Recent fieldwork on Naxos has included ber of the Cave of Zas shares features with that from systematic site survey, continuing excavations at the the Saliagos culture, notably white-painted patterns, important prehistoric center of Grotta-Palati (the but crusted and pattern-burnished wares suggest that

Oeiol," ArchEph1984, 14-54, discussesjewelry from Ak- In 1989-1990, a large EC cemeterywas explored at Kato rotiri in detail, with illustrationsboth of actualfinds (mostly Sangri in the western part of Naxos: Lambrinoudakisin bronze and bone pins and stone beads) and of their depic- CycladicCulture 26. The excavationof a grave at Panormos tions in the wall paintings,and, in "The Theran Wall-paint- Korfari tonAmygdalion has also been described:0. Hadjian- ing: Artistic Tendencies and Painters," in Crowley and astasiou,"HI&voQto," ArchDelt 36 B' (1981) 378. In the east Laffineur (supra n. 19), has defined hands of individual of the island a new prehistoricsite, fortified with rounded painters. towers, has been reported at Zas Kastelli;see "Chronique" 177 For the results of general survey, see R. Treuil, "Pro- 1989, 818; also C. Doumas, "CycladicCulture," in Cycladic spection archeologique a Naxos en 1981," in Les Cyclades Culture20; "Weaponsand Fortifications,"in CycladicCulture 59-65; for work (1982-1984) in the region of Kinidaros 92. Akrotiri and the Phaneromeni Monastery, see AR 30 (1984) Aside from systematicfieldwork a number of individual 53; "Chronique"1985, 839; and R. Treuil, "Naxos,"ArchDelt prehistoricartifacts from the island have been publishedfor 38 B' (1983) 350. I thank R. Treuil for kindly providing the first time. These include EC stone and terracottaobjects informationabout the project. A Bronze Age site has been from graves in the region of Panormos (Bossert, supra n. reported at Kalamadikou,and prehistoricsites at a number 121), among them a terracotta vessel of the "teddy-bear" of other locations have been noted; one is probablya cem- type. Many previously unillustratedEC finds from Naxos etery. A source of flint and associateddebitage at Stelida to have been illustrated in CycladicCulture, several of which the southwestof the Chora of Naxos has been studied, but are also illustratedin F. Zafeiropoulou,Naxos: Monuments its exploitationin prehistorictimes has not yet been conclu- and Museum(Athens 1988). These include: marblepalettes, sivelydemonstrated (see M. S6f6riades,"Un centre industriel stone vessels, querns, pestles, hammers, chipped stone, a pr'historique dans les Cyclades:Les ateliersde d'bitage du sealstone,stone beads,figurines, sling stones, terracottavases silex de Stelida," in Les Cyclades67-80). The results of (includinga bell-shapedcup and a spoutedjar of the Kastri palynologicalinvestigations at Grottaand two other locations group), jugs and a tankardfrom Panormos,a bone spindle on the island are described in J. Josette Renault-Miskovsky, whorl,bronzes (some from the KoronasCave in the northern "Lesconnaissances actuelles sur les v6g6tationset les climats part of the island), lead rivets, and a silver-platedbronze quaternairesen Gr6ce, d'apres les donnees de l'analysepol- dagger. linique,"in Les Cyclades99-109. 178 See Les Cyclades 64-65. The current state of Neolithic and EBA research on the 179 The evidence from both sites has been summarizedby islandhas recentlybeen summarizedin CycladicCulture. For K. Zachos, "The Neolithic Period in Naxos," in Cycladic maps showing the locationsof excavationsconducted at the Culture 29-32. A selection of metal, ceramic, and stone capitalof Naxos under the auspices of the Greek Archaeo- artifactsfrom recent excavationsat the Cave of Zas is pre- logical Society, see V.K. Lambrinoudakis,"Avcaoxac•i N&- sented. ovu,"Prakt 1985, suppl. pls. 6.1 and 6.2, facing p. 160. The 180 See O. Hadjianastasiou,"A Late Neolithic Settlement results of investigationsprior to the 1980s have been sum- at Grotta, Naxos," in GreekPrehistory 11-20, with more marizedwith extensive bibliographyand a gazetteerof pre- general comments on the distributionof Neolithic pottery historicsites by V. Fotou, "Lessites de l'Cpoqueneolithique in the Cyclades. Also Hadjianastasiou1989 (infra n. 185) et de du bronze a Naxos," in Les 15-57. For 209, n. 17, and A Prakt l'age Cyclades "OLx6tE6bolClTZox&dXXl," 1985, the history of research on Naxos, see also V. Lambrinou- 153-57, where the excavation in the Dimitrokalliplot and dakis,"Archaeological Research on the EarlyCycladic Period the context of this early materialis describedin more detail. in Naxos," in Cycladic Culture 25-26. 740 JACK L. DAVIS [AJA 96 the cave was first inhabited somewhat later than the chrome matt-painted varieties of probable Melian or- abandonment of the settlement at Saliagos itself.'81 igin, Middle Minoan imports, and Gray Minyan are Higher levels share features with FN Kephala on represented-types very scarce in Naxian contexts."83 Keos. The sequence appears to continue unbroken Later pots and terracotta lamps imitate shapes of into the EBA, providing an important bridge between Minoan origin. the later Neolithic and transitional EBA sequences of Approximately 140 fragments of figurines were the Greek mainland and the eastern Aegean. collected, for the most part small (some minute), The later EBA stratum is characterized by Kastri standing, and anthropomorphic. They are diverse in Group types, and contained a substantial number of style and as a group quite unique in the Cyclades; bronze tools and ornaments. Around a hearth were general parallels with Cycladic marble sculpture and scattered pieces of unbaked clay with 15 seal impres- Cretan coroplastic suggest that their production be- sions, made by an estimated five different seals bear- gan in the EBA. Other small finds include discoid ing linear motifs. The highest levels contained finds loomweights of Minoan type, fragments of emery, and of the historical periods mixed with Middle and Late marble objects (including a bowl and probably a Cycladic wares, including LH IIIC, and much Minyan folded-arm figurine). Petrographical analyses and sty- and MC matt-painted. A strip of gold from a Neolithic listic observations indicate that substantial quantities stratum is the earliest gold object from the Cyclades; of pottery were being imported, both from Melos and copper axes, awls, pins, and spatulas are also repre- Crete.'184 sented in Neolithic contexts. Domesticated crops were Elsewhere on the island, evidence for settlement in similar to those exploited on the Greek mainland; the later Bronze Age has been recognized in the barley played a prominent role in the diet. Two large Kalandos area (in the extreme south of Naxos), at leaf-shaped, bifacially retouched, obsidian spearheads Sangri, at Rizokastellia, and at Grotta itself.'85 Rescue are exceptional components in the lithic assemblage; excavations in a building plot (the Dimitrokalli plot) as on the Greek mainland, there appears to have been at the west end of Grotta, immediately south of the a shift from percussion to pressure-flaking at the end causeway that leads to Palati, have produced well- of the Neolithic. Other small finds of particular inter- stratified deposits of the early LBA, the first published est include a marble bowl and two bird-heads carved from the site. Excavations included the reexamination from bone. of a 10-m stretch of paved road cleared before World Surface investigations in 1985 at the site of Mikri War II and subsequently backfilled. On the floor of a Vigla and the publication of Middle and early Late building to the south of the road were local and Cycladic finds from Grotta have shed light on the imported pots of LC II and LM IB types, covered by later prehistory of Naxos. Mikri Vigla is a promontory destruction debris of rubble and mudbrick; these in- on the west coast, some 8 km south of the Chora.'82 cluded a jug, probably imported from the Dode- Artifacts and traces of architecture are abundant on canese, Marine Style sherds, and fragments of a vase the surface, and have been mapped by the Greek by the so-called Reed Painter. This was a site well Archaeological Service and the British School at Ath- situated within the Minoan orbit and in close contact ens. Fragments of monochrome painted wall plaster with centers of the southern and western Cyclades. were associated with the most impressive building Large-scale excavations at the Chora of Naxos were (structure 7): to the east of it a rescue excavation completed in 1985. Work had been resumed in 1978, yielded fragments of an MC storage jar. Local and the principal goal being to complete excavation of imported pottery from surface collections ranges in various parts of the LH IIIA to LH IIIC settlement date from EC through LC III; polychrome and mono- explored in earlier campaigns (1949-1974), but not

181 In addition to Zachos (supra n. 179), finds from the A Bronze Age Settlement on Naxos," BSA 84 (1989) 63- Cave of Zas are discussedin Zachos,"AvcowxcaXi Yilckaov 162. Zdg Nad6og:Kat'torejro 1987,"ArchDelt 42 B' (in press), a 183 Middle Minoan dark-ground styles appear best repre- report on excavation seasons in 1985 and 1986; and in sented, but one sherd of the LMIB marinestyle was present. Zachos,"Late Neolithic Origins of CycladicMetallurgy," and 184 S. Vaughan, "Appendix 2: Petrographic Analysis of A. Douzougli, "The 'Attic-Kephala'Culture: A New Ap- Mikre Vigla Wares,"in Barber and Hadjianastasiou(supra proach to an Aegean Culture,"both to be published in the n. 182) 150-59. proceedingsof the 6th InternationalColloquium on Aegean 185 0. Hadjianastasiou,"Naxian External Connections in Prehistory,Athens 1987. I am extremely grateful to Dou- the Late Bronze Age," BSA 84 (1989) 205-15. The excava- zougli and Zachos for allowing me to read and make refer- tions themselves are more fully described with a plan in ence to these papers in advance of their publication. Hadjianastasiou,Prakt 1985 (supra n. 180) 153-57, and 182 R.L.N. Barber and 0. Hadjianastasiou,"Mikre Vigla: suppl. pl. 5, facing p. 152. 1992] THE ISLANDS OF THE AEGEAN 741 fully published. In 1982-1985, however, new exca- explored, and to the east the Mycenaean fortifications vations in the Plateia Mitropoleos explored a segment have also been located.s89 On the Lower Plateau, east of the LH IIIC fortification wall of the town, nearly of the upper acropolis, extensive and deep EBA strata 18 m long, and buildings just inside it. Foundations (containing EC I material) have been explored be- for the wall are of unworked stones; on these lay a neath the LH IIIC levels; a tanged point of Saliagos superstructure of mudbricks. Buildings of LH IIIB type, a marble pendant in female steatopygous form, date outside the fortification show that the town had and ceramics suggest that Koukounaries was inhab- been larger before its construction.'86 Two lumps of ited already in the Neolithic period.'90 raw caolinite and montmorillonite clay were found in Excavations beneath the temenos of the Athena an LH IIIC context, and appear to have been im- temple on the Middle Plateau have revealed, at the ported to Naxos, perhaps from Melos. lowest levels, remains of building materials and arti- Paros.'87 Important new discoveries continue to be facts of LH IIIC (including a pierced triton shell), made at Koukounaries, on the southwest side of the contemporary with destruction levels of the megaron bay of Naoussa in northern Paros. Excavations at on the Upper Plateau; on the bedrock itself are EBA Koukounaries began in 1976. Yearly campaigns have layers. Ash mixed with animal bones and shell suggest uncovered the remains of a long-lived settlement that that from Geometric times this area was the focus for had begun already in the Saliagos phase of the Cy- rituals. The cemetery of the Mycenaean settlement cladic Neolithic. Early Cycladic artifacts attest to oc- was located in a valley northwest of Koukounaries. cupation, but the settlement appears to have been Three excavated tombs resemble mainland tholos deserted between the EBA and LH IIIC, when habi- tombs and have chambers (rectangular with rounded tation was established anew.'88 In LH IIIC the site corners) with stone walls and corbeled roofs built in was well fortified with Cyclopean walls and the plateau Cyclopean style; all were robbed.191 on top was occupied by a "mansion"; in its ruins were All finds from excavations ca. 1900 on the citadel large numbers of ceramic, stone, bronze, lead, and of Paroikia (the modern Chora of the island) have ivory artifacts. The complex was destroyed in a mas- been recently reexamined.'92 The principal architec- sive conflagration, promoted, no doubt, by a hostile tural remains should be dated to the final phase of attack: the body of an adult, killed by a wound to the Phylakopi I and are contemporary with the earlier head, had been hastily buried before the wound part of period IV at Ayia Irini. A small amount of healed; the bodies of other humans and livestock were pottery is both earlier and later in date than main left amid the debris. deposits. One of the earliest finds is a large fragment More recent investigations have defined the extent from a burnished rolled-rim FN bowl. Several sherds of Mycenaean settlement more accurately. Defenses from burnished tankards are contemporary with Ayia on the southern approaches to the acropolis have been Irini period III; a fragment of a hat-vase belongs to

186 See Lambrinoudakis,Prakt 1985 (supra n. 177) 144- in PrehistoricCyclades 184-206, and "The Decline of the Lambrinoudakisand F. Geometric 61; Zafeiropoulou, "Avaouxaq? Settlementof Koukounariesat Paros,"in R. Higg Prakt 1985, 162-67, with earlier ed., The Greek Renaissance the B.C.: Tra- MlqrQoO'•Xcwg;Nad'ov," of Eighth Century references. Subsequent to completion of the excavationsin dition and Innovation (Stockholm 1983) 173-83. Surface the PlateiaMitropoleos, the most significantstructures were reconnaissancein the Naoussa area has resulted in the dis- roofed in situ to form a subterraneanmuseum. Subsidiary covery of new EBA sites: see Schilardi,in PrehistoricCyclades investigationsin conjunction with the building of the roof 184, with references. There is limited evidence for rehabi- providedadditional details concerning the prehistoricsettle- tation in later LH IIIC after the destruction of the citadel; ment. See Ergon 1989, 122-25. The study of Mycenaean a substantialProtogeometric settlement followed. pottery from ArchaeologicalSociety excavations at Grottais 189 D. Schilardi, "Avaoxc•i oatvl 1H o," Prakt 1985, currentlybeing undertaken by M.B. Cosmopoulos,who de- 111-12; Ergon 1990, 105-107. 190 livered a paper entitled "MycenaeanNaxos: An Overviewof Ergon 1988, 134; Ergon 1989, 121. the Evidence,"at the InternationalCongress of Mycenology 191 For the Mycenaeancemetery, see D. Schilardi,"Avao- in October 1991. Rome-Naples, xafi fdkxov," Prakt 1986, 169, 203-204; "AvacoxacTj~I H- 187 Aside from excavated finds, EBA stone (including a Qov," Prakt 1987, 236-40. For soundings on the Middle schematic figurine) and ceramic artifacts, perhaps from Plateauwithin and around the Athena temple precinct,see southwestern Paros, have been published; one vessel has Prakt 1985 (supran. 189) 130-31, 142; Prakt 1986, 191-97, parallels in the Kastri Group. See O. Hadjianastasiou, 203; Prakt 1987, 235-36; Ergon 1988, 134; Ergon 1989, "o0wroxvxhablcd&EVtitrCa anwt6q krx1o," ArchEph 118-19. 1-4. t• 1983, appendix 192 J.C. Overbeck, The Bronze Age Pottery from the Kastro 188 The results of campaignsin the 1970s and early 1980s at Paros (SIMA-PB 78, Jonsered 1989); several additional have been summarized in recent papers by D.U. Schilardi, finds from Rubensohn'sexcavations are illustrated by Pa- "TheLH IIIC Period at the KoukounariesAcropolis, Paros," pagiannopoulou, Minoan Influence 336-37. 742 JACK L. DAVIS [AJA 96

Fig. 20. Paroikia,Paros. Beakedjug with matt-painteddec- oration. (CourtesyJohn C. Overbeck)

export from Greece of a great many EC marble fig- urines and vases. Some 500 objects have been attrib- uted to the so-called "Keros Hoard," which was reputedly recovered at Kavos in the southwest part of Fig. 19. Paroikia,Paros. MBA barrel-jarwith light-on-dark the island. The presence of so many marble objects decoration.(Courtesy John C. Overbeck) on so small an island is astonishing, and the nature of prehistoric activity on Keros was the center of much an EC type best known in funerary contexts. Later discussion in 1983 at the British Museum conference material includes familiar Melian types such as the on EC art; several participants suggested that Kavos paneled cup, a single bell-cup of LM IB date, and had served as a prehistoric ritual center in the EBA deep bowls of LH IIIC. In general there appears to Aegean.'94 Investigations at the site were resumed in have been less Minoan influence on the development 1987, and included a systematic collection of surface of MBA pottery at Paroikia and less contact with the artifacts."95Preliminary analysis of finds suggests that Greek mainland than at Ayia Irini or Phylakopi. The Kavos and, in the later stages of EC II, a settlement main deposits are characterized by Gray Minyan of on the adjacent islet of Daskaleio were parts of a single developed mainland varieties, MC burnished wares, large settlement.'96 Adjacent to this settlement was a and matt-painted barrel-jars and jugs (figs. 19-20); deposit of special character in which marble artifacts duck-vases and the extensive use of potters' marks are and certain ceramic types were discarded in greater features in common with the earlier MBA at Phylak- quantities than elsewhere on the site. opi and Ayia Irini. The earliest pottery is of Keros-Syros EC II types, Elsewhere on Paros at Tripiti, Mycenaean pithos but Kastri varieties are also represented.'97 Occasional sherds have been reported, and on Antiparos at Ayia sherds reminiscent of Phylakopi I are found but are Kyriaki, part of a much damaged cist grave containing extremely rare; their date is uncertain. Visual and an EC figurine was excavated in 1983.193 petrological analysis suggests that an exceptionally The Amorgian Islands. Extensive illicit excavation on high percentage (50-60%) of pottery from the site Keros since the late 1950s has resulted in the illegal was imported from elsewhere in the Cyclades. Ce-

193 0. Hadjianastasiou,"Hdoog" and "AvTrlCaQog,"Arch- 197 I am grateful to Cyprian Broodbankfor sharing with Delt 38 B' (1983) 350. me a draft copy of his report on the ceramicsfound in the 194 See discussionin Cycladica16, 27-29, 33-35. recent explorationsof Kavos, and for allowing me to sum- 195 Annual Reportof the Managing Committee,the British marizeseveral of his conclusionshere. Broodbankalso notes Schoolat Athens1986-1987, 32-34. a pyxis of KamposGroup type in the Naxos Museum, as- 196 Gazetteer337-38. signed to Keros,but not definitelyto the site of Kavos. 1992] THE ISLANDS OF THE AEGEAN 743 ramic analysis demonstrates that the forms prominent 2 m deep. Part of a fortification wall appears to have in the special deposit are all compatible with funerary sported a rounded tower, and was apparently a pre- material, and may also suggest that imported goods decessor of a previously investigated EBA enceinte. were brought to the island by its own inhabitants. The fortifications rest on earlier habitation levels that While a full interpretation of the site's function has lay beneath the water table and could not be investi- not yet been offered, these observations do not ac- gated. The fourth and latest phase of occupation tively support the hypothesis that Kavos was an inter- seems to have been contemporary with the EBA set- national sanctuary.'98 tlement already known at the Heraion. By then the Elsewhere in this group of tiny and sparsely popu- fortifications investigated in this sounding were out lated islets, excavations on Kato Koufonisi have ex- of use and a rectangular building, part of a larger plored part of a settlement of LC II date.'99 complex, extended the area of habitation beyond that enclosed within the settlement. The Samos previously new evidence from the Heraion also to the existence Neolithic finds were recovered in several stratified points of a substantially larger prehistoric settlement than soundings and in various isolated pits in the bedrock previously suspected. At Tigani, the overall distribu- that were explored in the late 1960s while examining tion of finds also suggests that estimates for the size the Hellenistic settlement at Tigani. The publication of the settlement in the later of the of these finds has forced a revision of the older relative stages Neolithic require revision. chronology established for Tigani. The various pits Newly published and republished small finds from have been assigned to four chronological phases (I- Tigani include 11 stone vases, and several figurines, IV) based on stratification in the soundings; several both terracotta and marble.201 In the lithic industry, phases have been subdivided on the basis of stylistic blades of obsidian are dominant, with a few obsidian criteria. The latest deposits appear to correspond to arrowheads in Tigani III, a situation that contrasts Emborio phases VII-VI in date (i.e., they antedate remarkably with that at Emborio and Ayio Gala on the beginning of the EBA); the possibility has been Chios, where obsidian played a palpably subordinate raised that phase I of Tigani began prior to the be- role.202 Despite close stylistic relationships recogniz- ginning of the Late Neolithic on the Greek mainland. able in the ceramic styles of the two islands, it is not No indisputable stylistic similarities with Troy I or the clear that there was much direct exchange in pottery. EBA sequence from Thermi on Lesbos have been It has, in fact, been suggested that in the Neolithic a noted; it has been suggested, however, that the earliest cultural dividing line should be drawn between them, ("black") phase of settlement at Poliochni on Lemnos with Samos looking away from Chios toward the Cyc- may be only slightly later than Tigani period IV. lades and the southeast Aegean.203 Several artifacts from Tigani do, however, find par- allels in EBA levels elsewhere in the Aegean area and The Dodecanese habitation probably continued at the site, despite the The past decade has witnessed a veritable explosion absence of an undisturbed EBA stratum. of new information relevant to the prehistory of the Recently published finds from the excavations in southeast Aegean, although as yet excavation has been the Heraion on Samos may close the gap between its limited and survey non-intensive. In the early 1980s, sequence of EBA settlements and the Tigani se- the absence of clear evidence for pre-LBA settlement quence.200 There, at some remove from EBA levels on most of these islands was particularly odd, as was previously uncovered, four stages of construction the paucity of documentation for contact between the were identified within prehistoric levels more than Dodecanese and Crete, the Cyclades, or the Greek

198 Contrast the opinion expressed in Renfrew (supra n. 200 See H. Kyrieleis, H.J. Kienast, and H.-J. Wei8haar, 19) 99-100. "Ausgrabungenim Heraion von Samos 1980/81,"AA 1985, 199Hadjianastasiou 1989 (supra n. 185) 206, 215, n. 59. 409-18; Felsch (KastroTigani 39, n. 92) sees no evidence Partsof two vessels from the settlementare illustratedon pl. for overlapbetween stratified deposits from Tigani and from 40d-e. The sherd on pl. 40e may well derive from a bur- the Heraion. nished kraterof "Aiginetan"type and, if so, would point to 201 Cf. the head of a small marble figurine from the Her- mainland, as well as Minoan, imports at the site. Several aion (Kyrieleiset al. [supra n. 200] fig. 42.1). previously unillustratedterracotta vases from Apano Kou- 202 Emporio 699-712. fonisi have now been published in CycladicCulture 172-73, 203 Cf. A. Sampson,"The Neolithicof the Dodecaneseand nos. additional 178-80; EC terracottavessels and stone fig- Aegean Neolithic Culture,"BSA 79 (1984) 246, fig. 6. Local urines from both Apano and Kato Koufonisi are illustrated sources of flint do seem to have been availableon Chios;see in Zafeiropoulou(supra n. 177) 30, 32-33, 37, 40. Hood and Bialor (supra n. 114). 744 JACK L. DAVIS [AJA 96

4 12 3 16171 / 3 20 121 22

5 6 Q4 2 7

024 25 26 27

28 2

3 9 8c- 29 30 337 334 35 V/36 (10 1 10 31 32

9a 4- ) I 13 15 40 I 14 41 42 1

Fig. 21. CharacteristicNeolithic pottery types in the Dodecanese. (CourtesyAdamantios Sampson)

mainland in pre-Mycenaean times--despite attested 4), broadly contemporary with the Late Chalcolithic exchange between Crete and the Anatolian coast, no- of Anatolia; each phase has been defined on the basis tably with Knidos and Iasos.204 of stratified deposits from recent excavations and Neolithic. The picture now emerging for later from extensive surface finds (fig. 21).206 All Neolithic phases of the Neolithic is one of a relatively dense sites thus far investigated in the Dodecanese are small distribution of sites, with similarities in material cul- in comparison to those on either the Anatolian or ture to Late Chalcolithic centers of Western Anatolia. Greek mainlands. Caves were frequently chosen for The general characteristics of Neolithic settlement occupation. can be outlined.205 Earliest settlement is represented Final reports on the results of recent excavations at best by the lower phase of occupation in the Kalythies five Neolithic sites in the Dodecanese have already Cave on Rhodes, and is contemporary with the begin- appeared: the cave at Kalythies Ayios Yioryios on ning of the Late Neolithic period on the Greek main- Rhodes; the cave at Arhangelos Koumelo on Rhodes; land. The Late Neolithic of the Dodecanese has been and open-air sites at Partheni on Leros and at Kastro divided into four phases (Late Aegean Neolithic 1- on Alimnia.207 These and sites on Yiali (below) were

204 Evidence prior to 1982 is summarized in J.L. Davis, Samos. An earlier find from , published by A. "The Earliest Minoans in the Southeast Aegean," AnatSt 32 Furness ("Some Early Pottery of Samos, Kalimnos and (1982) 33-41. Chios," PPS 22 [1956] pl. 19), is there republished and 205 For discussion, see Sampson (supra n. 203) 239-49; identified correctly as a "scoop" of FN type (pp. 112-13, n. Neolithic Dodecanese; "NEohlt0tx6tEv&iZaTa Wtr6Tr P680," 487). AAA 12 (1979) 24-39; "To N.A. Atyalo ora NrohltO6tx 207 Neolithic Dodecanese. The existence of the site of Par- theni on Leros was known from XQ6voa," ArchEph 1983, appendix 5-13; "Topographical previously investigations by Survey of Prehistoric Sites in the Dodecanese," in D.R. Keller Hope Simpson and Lazenby (Gazetteer 367). Material from and D.W. Rupp eds., Archaeological Survey in the Mediter- Kalythies and Partheni is also illustrated in Sampson (supra ranean Area (BAR-IS 155, Oxford 1983) 283-85; and "Pe- n. 203) 241, figs. 2-3, and 243, fig. 4; material from the riodic and Seasonal Usage of Two Neolithic Caves in Koumelo Cave in Sampson 1979 (supra n. 205). For Par- Rhodes," in Dodecanese 11-16. theni, see also S. Marketou, "HTa@OEvt,Ono• KovtaQ[ba," 206 See Kastro Tigani 136-37 for an initial evaluation of ArchDelt 35 B' (1980) 557; for preliminary reports on the the chronological system developed by Sampson and its Koumelo Cave, the Ayios Yeoryios Cave, the Kastro of Alym- interconnections with the Neolithic sequence proposed for nia, Partheni on Leros, and for prehistoric material of less 1992] THE ISLANDSOF THE AEGEAN 745

Fig.22. KalythiesCave, Rhodes. Neolithic pottery. (Courtesy Adamantios Sampson) initially identified or examined as part of a general recovered by sieving and impressions of plants were survey of prehistoric archaeologicalresources in all found on pottery. Plentiful human and faunal re- islands of the Dodecanese (with the exception of As- mains include a full range of basic Neolithic domes- typalaia) during 1977-1980. The purpose of that ticates (cow, pig, sheep, goat, and dog), all but dog project was to supplement and clarify earlier reports apparentlyraised for meat rather than for secondary summarizedin the Gazetteer:60 new prehistoricsites products, as well as various species of birds, fish (in- were recorded, 35 of them on Rhodes. Several of cluding a tunny-sized specimen), and crustaceans. these have, however, yielded only finds of chipped Only selected joints of deer appear to have been stone and are not conclusively of pre-Mycenaeanor brought to the site. Fox, hare, and marten may have even prehistoric date.208Thus far little evidence for been deliberate live imports to the island, and the occupation anywhere in the southeast Aegean prior earliestdomesticated chicken and blackrat in Europe to the Late Neolithic has been recognized.209 may also be attested.210Human bones belong to at Three phases of the Late Neolithic have been rec- least four infants and eight older individuals;juve- ognized in the Kalythies Cave excavations (fig. 22). niles and adults are mainly represented only by hand The most recent corresponds to the second phase of and foot bones and front teeth. The cave may, there- occupation at Tigani on Samos. Here as elsewhere in fore, have served as a site of permanent burial only the Neolithic Dodecanese, Melianobsidian is common for infants, while the bones of juveniles and adults (80% of the chipped stone), but obsidian from Yiali were transferred to another location for secondary and Anatolia is also represented. Plant remains were burial.

closelydeterminable date at PylonasCave of Halil and at see C. Doumas, "II((Q66laDL," ArchDelt 29 B' (1973-1974) Arhangelos Limani Papakonstanti,both on Rhodes, see 981. For the site of Harkadioon , see C. Doumas, Sampson, "AQXdyyExkog"and "Aylog Fe6qylog Kak- "Ti?og,"ArchDelt 30 B' (1975)369-72. vUO~wv,"ArchDelt 34 B' (1979) 448-49; "Avaoxw?alxg eQ- 209 See "FirstColonization" 170-71 with reference to ex- the results yaoi eg: P6bog," ArchDelt35 B' (1980) 558-59. cavationsin the HarkadioCave on Tilos. See also 208 NeolithicDodecanese 96-115 contains a gazetteer that of investigationsin a cave at Erimokastroon Rhodes where includes Neolithic, EBA, and some later findspots on dwarf elephant bones and possible traces of human occu- Rhodes, Kos, Karpathos, , Alymnia, Leros, , pation have been noted, but in unstratifiedcontexts; and A. ,and Tilos. For a more complete list of pre-Mycenaean Sampson,H NEo)tOt)l xarotilxra7 oToFvaAil rg Niuvov sites on Kos, see NeolithicDodecanese 229-32. Four of the (Athens 1988) 210 where the possibilityof Middle Palaeo- five sites catalogued for Karpathosare discussed in more lithic materialfrom Yiali is raised. detail in Karpathos,Saros and Kasos(see under A4-A6, E39, 210P. Halsteadand G. Jones, "BioarchaeologicalRemains E40, andJ56). For finds of CycladicKeros-Syros type pottery from KalythiesCave, Rhodes,"in NeolithicDodecanese 135- at the site of Vathy Elliniko on (Gazetteer 365), 45. 746 JACK L. DAVIS [AJA 96

apsidal and rectangular structures. Habitation was contemporary with the latest phase of occupation at the Koumelo Cave, as were badly eroded buildings at Partheni on Leros. Neolithic settlements have also been investigated on the tiny island of Yiali (fig. 23).212 Prehistoric remains are rare in the northeast where deposits of obsidian are located. On the neck that joins the northeast to the southwest part of the island, pottery and ground stone finds of FN and EBA date have been excavated (fig. 24); only here has EBA pottery been recog- nized.213 FN finds are widespread in the south, and include terrace walls and the foundations for a three- room house with curving walls (fig. 25). Many vessels were associated with its destruction deposits as were two crucibles for melting copper. Remnants of over 70 rectangular cist graves have been excavated. No bones or grave goods were found in any, despite sieving, but from the Neolithic pottery around them there can be little doubt about their date. Ground SHellenistic stone tools were widespread. All appear to have been INeolithic manufactured from local materials, with handstones 0 60 7Obsidian sources of sandstone, limestone, and andesite, and querns of andesite. Axes were not common. Fig. 23. Yiali. Obsidian sources and archaeologicalsites. mostly (CourtesyAdamantios Sampson) Recent fieldwork in the Dodecanese does not sug- gest that Yiali ever served as a major supplier of Excavations at the Koumelo Cave on Rhodes un- obsidian to the Aegean; organized quarries have not covered two superimposed surfaces: the lower of been recognized. Although irregular flakes found these was contemporary with the uppermost level of both on Yiali and Alimnia were employed as tools for the Kalythies Cave, the top with the Final Neolithic expedient purposes, regular types of chipped stone of the Cyclades. Above Neolithic levels the cave was tools even on Yiali itself were manufactured of Melian filled with a deep deposit of ash, apparently ejected (and perhaps Anatolian) obsidian and of flint.214 by the of the Santorini volcano.21' Early Bronze Age. Sites appear to have decreased in On the tiny currently uninhabited island of Alimnia number between the Neolithic and the EBA. Contact between Rhodes and Halki, excavations also uncov- with the adjacent Anatolian mainland seems to have ered parts of a settlement, including walls from both been frequent. The earlier parts of the EBA are the

211 Problemsin assigningvolcanic ash found outside Thera ently not the availabilityof obsidian that attracted early to the Minoan eruption of Santorini are discussed by R.B. inhabitantsto the island. See also H.G. Buchholz and E. Gallowayet al., "Radio-isotopeAnalyses of Aegean Tephras: Althaus, , Giali, Kos: Ein Vorberichtiiber archiiologisch- Contributionto the Dating of Santorini Volcano,"in TAW mineralogische Forschungen auf griechischen Inseln (Mainz 111.3,135-44. There the likelihoodof a post-Neolithicerup- 1982); and R. Torrence and J.F. Cherry,Archaeological Sur- tion of the Yiali volcano, virtually contemporaneous with vey of the Obsidian Source on Giali in the Dodecanese (unpubl. that of Santorini,is asserted-a suggestion that found little ms. in British School of Archaeologyat Athens). support from others in attendance at the TAWconference; 213 Sampson (supra n. 209) 23 has suggested that some see the discussion of this paper on pp. 144-45. More gen- lithicsmay be earlierin date than the Neolithic.Matt-impres- erallyon the identificationof volcanicash likelyto be Theran sions are present on the bases of so-called "cheese-pots" in origin, see D.G. Sullivan, "MinoanTephra in Lake Sedi- (p. 101). Neolithic pottery and obsidian from both Melos ments in WesternTurkey: Dating the Eruption and Assess- and Yialihave also been found on the islet of Pergousa,just ing the Atmospheric Dispersal of the Ash," in TAW111.3, to the west of Nisyros (Neolithic Dodecanese 252). 114-18, and the following discussionon p. 119. 214 Although Melian obsidian is not at all rare in the Do- 212 For the distribution in the Aegean of obsidian from decanese, the quantitiesin which it is found and a general Yiali,see Sampson (supra n. 209) 216-18, supplementedby scarcityof evidence for other kinds of Cycladicimports may referencesin J.L. Daviset al., "Keosand the EasternAegean: be indicativeof much more limited interactionbetween the The Cretan Connection,"Hesperia 52 (1983) 365-66. This Cycladesand Dodecanese than among the Cyclades them- obsidian does not appear to have been extensively used selves,or betweenthe Cycladesand the Greek mainland(see becauseof its poor fracturingproperties. It was thus appar- also Melas 1988 [infra n. 237] 290). 1992] THE ISLANDS OF THE AEGEAN 747

Fig. 24. Yiali. Excavationsin sector A. (CourtesyAdamantios Sampson) least well documented.215 It has been suggested that finds from the Aspri Petra Cave on Kos should be dated to EB II because one-handled cups and spouted jugs are present; other sherds from surface collections at Troulli on Kos and on Astypalaia may also belong to EB II. In recent excavations material of similar character has been found at Ayios Fokas on Kos and on the island of Yiali. Finds from Askloupi and Mes- saria on Kos, from Muskebi on the Anatolian main- land, from Lindos on Rhodes, and from Nisyros have sometimes been dated to EB III, but it has most recently been argued that the material from Askloupi and Messaria is of EB II date. EB II parallels from Tsilimbiri on Kos and from new excavations of graves at Tavla near Antimaheia on Kos have been ad- duced.216 Rescue excavations at Seraglio on Kos and Asoma- tos on Rhodes should clarify the EBA sequence con- siderably.217At Seraglio, a fortified settlement of EB III has been revealed in deep soundings beneath the later Bronze Age town; it was itself built on top of earlier levels of EB III, including the remains of a potter's kiln. Part of a single long house, possibly with one "ellipsoidal" wall, has been cleared. Ceramics

215 For general discussionof the distributionand character of EBA materialin the southeastAegean, see NeolithicDo- decanese118-19; Sampson 1983 (supra n. 205) 12. For an update, see T. Marketou, "Asomatosand Seraglio: EBA Productionand Interconnections,"Hydra: WorkingPapers in MiddleBronze Age Studies7 (1990) 40. 216 For see H. EX. Tavla, Kantzia,"AvT•td•'XtLa: Ktltpa Zapou',"ArchDelt 39 B' (1984) 335. Fig. 25. Yiali. Foundationsof a Neolithic house. (Courtesy 217 Marketou (supra n. 215). AdamantiosSampson) 748 JACK L. DAVIS [AJA 96

(wheelmade red-burnished bowls, duck-vases, and fied group of finds (excavated in 1925) comes from a tankards) find close parallels both in Anatolia and in floor, not clearly associated with other architectural the western Aegean. At Asomatos near Kremasti, west features, southwest of the later Athena temple; finds of Trianda on the north coast of Rhodes, two long of similar character were probably excavated in a fill "megaron-like" buildings have been uncovered, and beneath the temple itself and in other parts of the were associated with similar pottery. They belong to acropolis. Only undecorated pots were found on the the latest phase of the site and overlay parts of houses floor; most have their closest parallels in MBA Crete. on a different orientation. Fortifications are surmised Small finds are both of MBA and LBA varieties and because of the density of architectural remains.218 include a stone lid of MM type, a spindle whorl, a Middle Bronze Age. The Middle Bronze Age in the terracotta scuttle, a discoid loomweight, stone vases, Dodecanese remains a mystery on which recent ex- marble pommels, and a bronze mirror. plorations have shed little light.219 The settlement at Late Bronze I and II. LB I patterns of settlement Asomatos came to an abrupt end in EB III. On the contrast radically with their predecessors. In northern basis of finds (mainly carinated bowls220) from com- Rhodes, the focus of activity shifted from Mt. Fileri- pressed levels found stratified in several soundings mos to Trianda on the coastal plain below, where there between EB III and LB I layers,221it has been argued is, for the first time, evidence for widespread habita- that occupation at Seraglio was continuous from the tion. Recent excavations have filled out the picture of end of the EBA to the LBA, but it is far from clear occupation sketched by extensive Italian excavations that all phases of the MBA are represented. A kiln before World War II.226 The LB I settlement at for firing the bowls has also been investigated.222 Mi- Trianda was large by island standards, and has been noanizing pottery with light-on-dark decoration, estimated at over 12 ha. Moreover, the thoroughly widespread in the Dodecanese and often assigned to Cretan character of the settlement has been rein- the MBA, is more likely to belong to the earlier phases forced by new discoveries. Buildings had plastered of the LBA, both in the southeast Aegean and in the floors and walls; characteristically Minoan architec- Cyclades where it was imported.223 As yet, there is tural features such as the polythyron and ashlar ma- little evidence for earlier versions of this style.224 sonry were employed; and new fresco fragments, Later phases of the MBA are best represented by including one with a painted lily and another with a finds from Mt. Filerimos on Rhodes, high above the double-axe and sacral knot are in the Minoan idiom. northern coastal plain, southeast of Trianda.225 At Horns of consecration in stone suggest Cretan reli- Profitis Elias on the most eastern spur of this massif, gious practices, as do new bronze figurines of Minoan survey and excavation have yielded unstratified finds types. and traces of architecture. Jugs more Anatolian than Results from recent excavations at Trianda in the Cretan in appearance were found with carinated and Theoharis plot have been fully published. Excavations conical cups; some pottery belongs to LM IA. The there in 1975, 1978, and 1980 uncovered structures existence of MBA material on Mt. Filerimos itself has within an area of approximately 360 m2.227 Three now been clearly demonstrated. The only well-strati- main constructional phases were recognized, Trianda

218 On Asomatos, see Marketou(supra n. 215) 40-47. Benzi, "Evidencefor a Middle Minoan Settlement on the 219 NeolithicDodecanese 119-20. Acropolisof Ialysos (Mt. Philerimos),"in Minoan Thalasso- 220 See Marketou(supra n. 215) fig. 5b. cracy 93-104 (including p. 102, for the possibilityof MM 221 Marketou (infra n. 230) 101-104. See also the report material from the region of Ayios Isidoros in the central of MBA pottery, including fragments of polychrome bird- western part of Rhodes). jugs (infra n. 236). 226 For older see Gazetteer348. The recent 222 bibliography, Marthariet al. (infra n. 223) 175. bibliographyis already extensive. In addition to references 223 Marketou(infra n. 230) 103; M. Marthari,T. Marketou, already cited (supra n. 225), see Doumas (infra n. 230) and and R.E.Jones, "LB I CeramicConnections between Thera Marketou(infra n. 230). Papagiannopoulou,Minoan Influ- and Kos,"in TAWIII.1, 171-84; Davis et al. (supra n. 212) ence218-22, remarkson the characterof pottery from both 361-66. older and recent excavationsat Trianda and at Seraglio on 224 Recently published results of excavationsin the Theo- Kos. at Trianda on haris plot Rhodes (infra n. 227) provide the 227 L. Papazoglou-Manioudaki, "Avaoxa)il tov well-documentedevidence in MLVtL- only support of an early date oLXLtooto ora P60ov," ArchDelt 37 A' for Dodecanesian xo0 ToQ1VT-carl•g light-on-darkwares. Only a single sherd (1982) 139-87. Initial results of this excavation had previ- from a pre-LMIA context has been publishedand the bulk ously been describedin a preliminaryreport by C. Doumas of the ware clearly belongs to the LBA. and Papazoglou,"Santorini Ash from Rhodes,"Nature 287 225 T. Marketou,"New Evidence on the Topography and (1980) 322-24. Site Historyof Prehistoriclalysos," in Dodecanese27-28; M. 1992] THE ISLANDS OF THE AEGEAN 749

I, II, and III, and two architectural subphases distin- walls are quite substantial and ashlar masonry was guished within the third phase.228 Both Trianda I and sparsely employed; in the western part of the house, II were explored mainly in soundings beneath the a second story was reached by wooden stairs within courtyard of a large building of Trianda III. No trace stairwells. After a destruction (perhaps caused by an of the LM IA style was recognized in the pottery of earthquake), lighter less consequential walls, which Trianda I. It is necessary, therefore, to redate the infringed on the area of the courtyard, were added earliest occupation at Trianda to the later MBA, ap- to the complex in Trianda IIIB. Individual strati- proximately contemporary with finds from Mt. Fil- graphical levels could not be assigned to phase IIIA; erimos. Trianda I appears to be the earliest stratum the final abandonment levels of phase IIIB are dom- at the site; in places, excavation reached bedrock and inated by LM IB/LH II ceramic types.231 Trianda should still be viewed as a new foundation in Other recent soundings at Trianda have provided the New Palace period of Crete. important details concerning the extent and duration The character of the shapes and decoration of pot- of the prehistoric settlement. Excavations since 1982 tery in Trianda I is overwhelmingly Minoan, and in the Markos plot (300 m west of the Theoharis plot) Cretan features such as discoid and cylindrical loom- have exposed the most western part of the Minoan/ weights, a bull figurine, a stone vase, potters' disks, Mycenaean settlement, including parts of an LM IA fragments of faience, and frescoes decorated with building with a polythyron.232In the Kattavenos plot bands or curvilinear motifs are all represented.229 (300 m south of the center of the Italian excavations Other small finds include fragments of small gold at Trianda), LM IB remains lay above volcanic ash, bands, a bronze sickle, bronze needles, and a lead rod. with LM IA remains beneath.233 These and other The cause for the destruction of Trianda I is not recent soundings suggest that the southern part of readily apparent. Phase II, on the other hand, showed the LM IA settlement had been abandoned after the obvious traces of destruction by fire, perhaps precip- eruption of the Thera volcano; in LM IB, new struc- itated by an earthquake. The damaged remains of the tures, including a possible defensive system, were built settlement were soon partly covered by a layer of above the ash layer in northern parts of the site, and volcanic ash from the Santorini eruption; the latest the settlement decreased in size. Imports from Crete pottery sealed beneath the ash belongs to the LM IA continued in LM IB and Cypriot wares were also style. The Minoan LM IB style is only represented in present. strata above the ash layer. Elsewhere on Rhodes, lay- Evidence of expansion of settlement in LB I is also ers of ash from the Santorini eruption have been evident in areas east and west of Trianda: contem- investigated: near the airport at Paradeisi and on the porary remains have been explored in five other lo- east coast at Kolymbia the layer exceeded a half meter cations. These are spread over an area of more than in thickness.230 14 km along the coast, and from the village of Trianda In Trianda IIIA, a large house with rooms around to the foot of Mt. Filerimos. In addition, on the east a central court was built in the Theoharis plot. Many coast of Rhodes LB I artifacts have been excavated in

228 Phases defined by the new excavationscan be approx- from Rhodes and Kos: Some Chronological Remarks Based imately equated with those defined by A. Furumark("The on the Stratigraphy,"in TAW 111.3, 100-12; C. Doumas, Settlement at Ialysos and Aegean History c. 1550-1400 "The Prehistoric Eruption of Thera and Its Effects," in B.C.," OpArch 6 [1950] 150-271) as follows: Trianda Dodecanese34-38. II= FurumarkTrianda I; TriandaIIIA and B= Trianda IIA 231 Later Mycenaean finds include a bronze arrowhead, and B, respectively.Trianda I, on the other hand, represents and a phi-figurine was recovered nearby at the so-called an initial stage in the settlement not defined by earlier in- Pyrgos of Paraskevas; see Papazoglou-Manioudaki (supra n. vestigators. 227) 181. 229 There are several strikingdifferences between this Mi- 232 See S. Marketou, "086g IEQo3 A6X'ou (otLx6OE60 noanized assemblage from Trianda and contemporaryde- Kcwvor. MaQ'xov, KT. MeQ. 551," ArchDelt 39 B' (1984) in posits the Cyclades.Obsidian from Meloswas present but 325-26, where long parallel walls perhaps belonging to a scarce; mainland matt-painted and Minyan wares do not drainagework of LMIB date are reportedas well as frescoes, appear to be represented at all; and there is little evidence Cypriot vases, and limestone horns of consecration. This is for the importation of Cycladic pottery. Not surprisingly, the plot described as plot 3 and marked on T. Marketou's from imports Western Anatolia are indicative of exchange map in Dodecanese (supra n. 225) 29, fig. 4; there a possible betweenRhodes and the adjacentmainland; see Papazoglou- defensive function is suggested for the . Manioudaki n. (supra 227) 171-72. 233 S. Marketou, "-IdQooog 9%gMatov (P6OQog 230 On Theran oLXOJTo- ash, see T. Marketou, "SantoriniTephra 8ov KaTrcapevo6)," ArchDelt 39 B' (1989) 327. 750 JACK L. DAVIS [AJA 96

the cave of Koumelo234 and are represented among parallel either in the Cyclades or on Crete. EBA and surface finds from Kolimbia-Theotokos.235 earlier MBA assemblages have not yet been well de- Results from recent excavations at Seraglio on Kos fined. echo those from Trianda.236 The LM IA town at By the later MBA, it seems clear that Karpathos, Seraglio also increased markedly in size, and two ar- Kasos, and Saria had been drawn into the Minoan chitectural phases have been recognized. After the orbit. The finest Minoan Kamares types characteristi- first was destroyed by earthquake, there was a major cally found at Phylakopi or Ayia Irini already in the reorganization of the town and a main street was earlier MBA, however, have not yet been reported established that led from the harbor up the hill to the and there is only slight evidence for the importation main focus of the settlement. The second phase of of Cycladic wares.238 Definite or possible finds have the settlement was destroyed before the end of LM been published from more than 40 separate locations. IA (probably also by earthquake); remains of human Accumulating data suggest the existence of an exten- victims were found in its rubble. As at Trianda, a layer sive Minoanized center at Pigadia, capital of modern of volcanic ash covered this debris in part, but the ash Karpathos. This settlement lay in the southeast part appears to have fallen measurably later than the of the island on the south side of a large bay that earthquake destruction. faced the southern tip of Rhodes.239 The focus of Recent research on Karpathos, Kasos, and Saria prehistoric occupation appears to have been the paints a picture similar to that being established for coastal plain, bordered on the west by the hill of Skopi Rhodes and Kos.237 Here too there is considerable and on the east by the acropolis of Sissimos; chamber question about the date and extent of earlier (i.e., pre- tombs, some as early as MM III/LM IA, were dug into Mycenaean) settlement. Surface investigations have the slopes of both. documented considerable activity in the later Neo- The existence of prehistoric settlement at the Xe- lithic period and during transitional Neolithic/EBA nona site has long been recognized, but it is now phases, but finds are extremely worn and difficult to obvious that occupation in the Pigadia area was much

234 See Sampson, in Dodecanese (supra n. 205) 13-14; in IIIC; two potter's kilns of MM/LM IA date; EBA sherds); the Koumelo Cave, early LBA pottery (said to be of LM/LC "O66; An•Tpkov xla KokoxotLW•Vq (otx6~e6bo A. and Cypriot types, including a base-ringjug) was found to Mvukvmx),"ArchDelt 38 B' (1983) 396 (reached the top of be stratified both above and below the ash layer (Neolithic Mycenaean levels); "066g EX. BEvt?kXov Xat Koko- Dodecanese 214). See also Neolithic Dodecanese 74. Layers of xoTrg&vq (otx6bE6ro E. Ataxavao'rdorq)," ArchDelt 38 B' ash associated with archaeological finds have also been rec- (1983) 396 (excavation of part of the Mycenaean settlement); ognized on Tilos, Karpathos,and Halki. For further remarks H. Kantzia, "066g KokoxoTr;bvYq 25 (otx6rbE6o Avt. see Melas 1983 (infra n. 237) 58. OacXaoLvoV)," ArchDelt 39 B' (1984) 329-30 (LH III above 235 Melas 1988 (infra n. 237) 300-302. MM/LM IA; stratified EBA including part of a wall and a 236 T. See Marketou, "Marine Style Pottery from the Ser- partly intact potter's kiln); "Fwvta WTv o6&v Iw. Ocok6yov aglio in Kos," BSA 82 (1987) 165-69, and brief reports on xat AnFkkov (otx6no6co FTQC. HLLdvta)," ArchDelt 39 B' recent rescue excavations in the Seraglio, including I.H. (1984) 330 (LH III, MM/LM IA, with EBAjust above virgin Papahristodoulou,"066g BEgototnokoXov40 (otx6rbE6o EX. soil); "O66g H(caIoFov (otx6n'6co Aeon. Kakoy-ilov)," XaTr•ratGoryov)," ArchDelt 34 B' (1979) 452-54 (pure My- ArchDelt 39 B' (1984) 331 (reached top of LH IIIC stratum). cenaean finds, including a phi-figurine, from soundings near 237 For recent workon these islands,see M. Melas,"Minoan the temple of Demeter); "066g EX. BEvL~XEov 30 (otx6nEbo and Mycenaean Settlement in Kasos and Karpathos," BICS N. TovrESil)," ArchDelt 34 B' (1979) 456-57 (two prehis- 30 (1983) 53-61; "Survey of Karpathos, Kasos and Saria, toric building levels, equivalent to Morricone's first and sec- Dodecanese,"in Keller and Rupp (supra n. 205) 287-89; ond city, in a sounding near the northeast corner of Karpathos, Saros and Kasos; "Exploration in the Dodecanese: Morricone's zone I; from the foundation levels for the walls New Prehistoric and Mycenaean Finds," BSA 83 (1988) 283- of the older phase came sherds with representations of birds, 311. See also "2jil~XLatoEXXqvoxcaitCag," ArchDelt 37 B' apparently in the Cycladic Black and Red style; in higher (1982) 417 for explorations by the Ephoreia of Caves in the levels LM I and LH II-IIIB types were represented, along cave of Ellinokamaras on Kasos (for earlier bibliography see with jugs of the familiar Dodecanesian light-on-dark style); Gazetteer 359). For copper in Karpathos in the LBA, see "O66g TocXa&bdXl at 25rg MaCL'ov (otx6Orc6o ELQ. Xo- Stos-Gale (supra n. 39) 276. (ov')," ArchDelt 35 B' (1980) 552-53 (in the Pizzoli zone of 238 Karpathos, Saros and Kasos no. 558, from the site of earlier Italian excavations; four major stratified levels were Palio Mitato (D23), appears to be a sherd from a Melian jug. recognized, dating from later MM to LH IIIB/C); "066g M. 239 A large limestone figure found there was presented by AhXE&vbgov xLatEX. BEvtEkXov (otx6bE60o Xcaql- J.T. Bent to the British Museum in 1886. Melas (Karpathos, ArchDelt 35 B' Caq(v odCpa)," (1980) 553 (LH architectural re- Saros and Kasos 147-48) has suggested that it may have mains farther west than the previously postulated limits of served as a cult statue; cf. Renfrew's remarks concerning the settlement); "066g EhXEveOiLovBEvL~Xov a L Ay. similarly "colossal" figures from the southeastern Cyclades HacgaoxvFvlS (otx6nbE6o ahx6E•6v BaclOhnlov)," ArchDelt 36 ("Speculations on the Use of Early Cycladic Sculpture," in B' (1981) 409 (opposite zone I of the Italian excavations; Cycladica 24-30). five architectural phases recognized, dating from MM to LH 1992] THE ISLANDS OF THE AEGEAN 751 more extensive; some 150 m to the east of Xenona, a Since thousands of tombs have been looted and principal focus of LB I settlement has been identified their contents transferred to museum collections out- at Vroulidia. Here recent construction work afforded side Greece, a great many artifacts both on Rhodes opportunities to examine stratigraphy. Finds are of and abroad have little or no context. A pleasant ex- types characteristically found at Minoan and Minoan- ception are tombs in the Lindos area excavated in izing centers elsewhere in the southern Aegean and 1925 in a salvage operation designed partly to com- include plentiful conical cups, red wall-plaster, pottery pensate for damage done by tomb-robbers. The re- of MM III/LM I varieties, and extensive deposits of sults of these excavations have recently been murex shells. published, and many tomb-groups now divided be- Elsewhere on Karpathos, on Saros, and on Kasos, tween Turkish and Danish museums have been reas- similar ceramics are widespread, if not so closely dat- sembled on paper.245It has also been possible, in most able. Sites are generally small with undistinguished cases, to relocate the sites of the original excavations. finds (occasional loomweights of Minoan type are not- The restudy of Mycenaean finds from Italian ex- able); most are located on, or near, coasts. Pottery cavations at Trianda (lalysos) has clarified the later actually imported from Crete appears to be excep- history of northern Rhodes. There the identity of tional outside Pigadia. Trapeza (L60) on Kasos is an those who used the numerous chamber tombs (of LH exception; Minoan finds were plentiful there and ash- IIB-IIIC date) at Makra Vounari and Moschou lar masonry was possibly employed. On Kasos the Vounari has long been at issue. Very little pottery primary focus of occupation seems to have shifted later than LH IIIA: 1 was published in the report of from the southwest area, that part of the island closest Italian excavations at Trianda, nor have excavations to Crete, to the north coast at some time between LB on Mt. Filerimos yielded substantial numbers of LBA I and LB III. finds.246Furumark, in his reanalysis of the excavations Minoan finds have also been recognized on many of Trianda, all but ignored the scanty Mycenaean of the smaller islands of the southeast Aegean, e.g., evidence and assumed that the settlement was aban- at Garipa on Telos,240 and at Pontamos Ayioi Anargyroi doned in LH IIIA:1. on Halki.241 It is now clear, however, that substantial numbers Late Bronze III. Study of the Mycenaean period in of LH IIIA:2 sherds were found in the older exca- the southeast Aegean continues to be hampered by a vations; these were distributed over so large an area lack of properly excavated and completely published that it is unlikely that they were transported from settlement sites; fullest information comes from ex- elsewhere, despite a lack of associated architectural cavations conducted in the 1920s and '30s when these remains.247 Current excavations at Trianda may re- islands were under Italian occupation.242 Recent solve the matter soon. Thus far they have produced syntheses of previous archaeological investigations on some evidence for occupation later than LH IIIA: 1, Rhodes provide convenient points of departure for including a few sherds of LH IIIB; LH IIIC is possibly any discussion of the complex history of settlement in also represented. the Dodecanese during the Mycenaean period.243 The first chamber tombs at Ialysos were built in LH Findspots are concentrated in the more fertile areas IIB/IIIA: 1; weapons were prominent in these earliest of the north, particularly around Trianda, but cham- burials.248 Minoan influence and imports had almost ber-tomb cemeteries have been recognized in most totally disappeared. Elsewhere on Rhodes, contem- other parts of the Rhodian periphery and at several porary finds are thinly distributed. In LH IIIA:2, sites in the interior.244 however, an island-wide pattern in the geographical

240 246 A. Sampson, "ML0vwLXda6r TrlvT•Xo," AAA 13 (1980) For a current summaryof the evidence, see M. Benzi 68-73; "Fta Ua," ArchDelt 34 B' (1979) 461. (supra n. 225) 100-102; and M. Martelli, "La stipe votiva 241 Melas 1988 (supra n. 237) 307. dell'Athenaion di Jalysos: Un primo bilancio," in Dodecanese 242 See M. Benzi, "Mycenaean Rhodes: A Summary," in 104-20, esp. 115, n. 1. Dodecanese 59-72. 247 M. Benzi, "Mycenaean Pottery Later than LH IIIA:I 243 C. Mee, Rhodes in the Bronze Age: An Archaeological from the Italian Excavations at Trianda on Rhodes," in Survey (Warminster 1982), subsumes bibliography and dis- Dodecanese 39-55. cussion included in the Gazetteer, and reevaluates much un- 248 Benzi (supra n. 242); see also "Tombe micenee di Rodi published material from earlier excavations; see also Benzi riutilizzate nel TE IIIC," SMEA 23 (1982) 323-44 (where (supra n. 242). contexts are clarified for chamber tombs at Ialysos, Kala- 244 More recently Sampson, in Neolithic Dodecanese 101, varda, and Mandriko); and "Rhodes in the LH IIIC Period," has reported the presence of Mycenaean pottery in a cave in Greek Prehistory 253-62. My summary is based largely on at Kakoskali near Arhangelos on Rhodes. Benzi's account in Dodecanese, the most comprehensive and 245 S. Dietz, Excavations and Surveys in Southern Rhodes: authoritative recent conspectus of the evidence. The Mycenaean Period (Lindos IV. 1, Copenhagen 1984). 752 JACK L. DAVIS [AJA 96 distribution of tombs was established that persisted Trianda on Rhodes. In LH III, there appear to have until LH IIIC. LH IIIB finds are, however, strangely been far fewer sites than in MM III/LM IA and none uncommon, and it has been argued that this situation has yet been located on Saria; most settlements and reflects an actual decrease in settlement.249 In LH cemeteries were in naturally fortifiable locations IIIB, there was a substantial decrease in the number where no evidence of earlier LB I occupation has of tombs built or in use at Ialysos and perhaps more been recognized. LM/LH IIIA and earlier LH IIIB generally in northwest Rhodes. In other parts of the types of ceramics are well represented; the later island, the number of tombs in use also appears to phases of LM/LH IIIB and IIIC are curiously absent. have decreased substantially, except in the southeast Imports from the mainland have been recognized but where there are slightly more IIIB finds than in pre- Cretan types remained dominant, in contrast to the vious periods. situation on Rhodes and Kos.251 Cemeteries with LH IIIC, on the other hand, witnessed a sudden chamber tombs of Mycenaean type became the rule increase in the number of burials at Ialysos; many on Kos, Kalymnos, and Astypalaia.252 older tombs were reused and more Mycenaean figu- Excavations at Kardamaina (ancient Halasarna) on rines were deposited in burials. Most interments ap- the south coast of Kos may help sort out details of pear to date early in LH IIIC. A "continuous but thin Dodecanesian relative chronology in LB III; a strati- thread" seems to have been maintained between the graphical sequence of deposits is said to run unbroken Argolid and Rhodes throughout the IIIC period, but from LH IIIA: 1 through Greek and Roman times.253 at most sites there is no pottery of the later stages of Mycenaean finds have now also been recognized even LH IIIC. It has been suggested that there may have on smaller islands of the southeast Aegean such as been internal migration in early IIIC from other parts Leros.254 of the island into the Ialysos area.250 Recent research on Kos, Karpathos, Kasos, and Amorgos Saria has also clarified the pattern of Mycenaean set- Many new prehistoric sites have been reported on tlement. In Karpathos, Kasos, and Saria there remains Amorgos and the locations of older finds identified an unexplained gap in the local sequence of habitation with greater certainty through recent surface explo- during LM IB and LM II, a peculiar absence given rations. Some 18 sites can now be dated to the Early the presence of substantial amounts of LM IB at Cycladic period.255 A fortified EC stronghold at Mar-

249 C. Mee, "The LH IIIB Period in the Dodecanese,"in ristodoulou, "MEoagt&,"ArchDelt 34 B' (1979) 457-58; also Dodecanese 56-58. Papazoglou 65-66, n. 9. For LH/LM chamber tombs on 250 On the issue of internal migrationand more generally Kalymnos, see AR 30 (1984) 70. For the investigation of two concerningregional variationin materialculture during LH chamber tombs on Astypalaia at Synkairos, located on the IIIC, see C. Macdonald,"Problems of the Twelfth Century north shore of the island between Steno and Trito Marmari, B.C. in the Dodecanese," BSA 81 (1986) 125-51; on LH see C. Doumas, "Aotv•nXCuLa,"ArchDelt 30 B' (1975) 372 IIIC pottery in the eastern Aegean, see S. Sheratt, "The (finds included a bronze spearhead, a fishhook, lead weights, Development of Late Helladic IIIC," BICS 32 (1985) 161. chisels, and spatulas; bones in one tomb were fully disarti- See also J.H. Crouwel, "Fragmentsof Another Octopus culated and in both tombs they showed strong signs of from Kalymnos in Amsterdam," BABesch 59 burning); for Mycenaean Astypalaia in general, see Macdon- (1984) 63-68 (two fragments of an unpublished LH/LM IIIC ald (supran. 250) 148. vessel said to have been found on G. Aleura et Kalymnos). 253 al., "Avaoxa~il orlv KaGcatcalva (AQ- 251 See Saros and Kasos but also R. Karpathos, 180; Jones XaLa Aka

263 On the relevant analyses from Kastri, see Z.A. Stos- 25-29; and J.B. Rutter, "The 'Early Cycladic III Gap': What Gale, N.H. Gale, and G.R. Gilmore, "Early Bronze Age It Is and How to Go about Filling It without Making It Go Trojan Metal Sources and Anatolians in the Cyclades," OJA Away," in Prehistoric Cyclades 95-107. Arguments are sum- 3 (1984) 23-43. On the southwestern Anatolian origin of marized and discussed at length in Manning 1992 (supra n. Kastri Group types, see M.J. Mellink, "The Early Bronze 160). in West Age Anatolia: Aegean and Asiatic Correlations," in 265 Recent excavations at Phylakopi have not produced G. Cadogan ed., The End of the Early Bronze Age in the detailed stratigraphy for the MBA. For this and other infor- Aegean (Leiden 1986) 139-52, and Wilson and Eliot (supra mation concerning the second city of Phylakopi, I am grate- n. 35). ful to Robin Barber, who has allowed me to study his For 264 arguments pro and con, see "Perspectives"; refer- unpublished report on newly excavated Middle Cycladic ences in n. 123 (supra); Warren and Hankey (supra n. 158) pottery. 1992] THE ISLANDS OF THE AEGEAN 755 ning of MM III. The earlier stage of period VI has stratigraphical sequences of each and every site and been considered more or less contemporary with the island need first to be considered individually and Seismic Destruction Level at Akrotiri on Thera, and only then related to those of other islands: it is partic- both have been dated in Cretan terms to the very ularly important that stratified groups of material beginning of LM IA. This is also the time when the from Cycladic settlements continue to be published, Second City of Phylakopi was destroyed. There rather than selections of finds arranged by phases would, therefore, appear to be a gap in the Ayia Irini established at some other site such as Phylakopi or sequence. On Crete, the transitional MM/LM phases Ayia Irini. Such a procedure is complicated, but the have never been closely defined; refinements there complexity is only reflective of the truly intricate na- should allow the nature and duration of this gap to ture of the archaeological record itself. Our purpose be described in greater detail. after all is not to establish a chronological system of The beginning of the LBA in the Cyclades is clearly mnemonic simplicity for students. represented by deposits of early period VI at Ayia Enormous progress is also being made in sorting Irini and from the Seismic Destruction Level at Ak- out the relative chronology of other islands and island rotiri. The publication of these deposits will provide groups of the Aegean. More detailed publication of definitions for early LC I that are practicable within finds from excavations in the prehistoric settlements the Cyclades. It would be unwise to redefine the be- at Manika should clarify the material cultural se- ginning of LC I in reaction to revisions in Cretan quence of Euboia for Bronze Age periods earlier than terminology. Subtle definitions of MM III and LM I, the first phase of Lefkandi. Our understanding of of the sort currently being proposed by Minoan spe- later EBA and MH phases will profit from the com- cialists, could only be distinguished in the islands of plete publication of the results of excavations in the the Aegean were substantial quantities of Cretan im- Vouratsas plot at Eretria, and from Palamari on Sky- ports present. This is rarely the case, even in the ros. In the northern Aegean, stratigraphy from Skala Cyclades, where Minoan pottery never amounts to Sotiros and Kastri on Thasos will be of great impor- more than a fraction of the total in an assemblage. tance in linking the cultural sequences of the northern For the remainder of the LBA, recent publications Aegean to those of the south, both in the EBA and have clarified many details of relative chronology. The LBA. Emborio on Chios and Tigani and the Heraion eruption of the Santorini volcano represents a pivotal on Samos also emerge as pivotal sites that tie the later stage and it is now obvious that the material in the Neolithic and EBA Cyclades to the eastern Aegean Volcanic Destruction Deposits at Akrotiri is contem- and Anatolia. porary with a later, but not the latest, stage of LM IA Recent developments in the Dodecanese have al- on Crete. Stratigraphy at Phylakopi on Melos suggests ready made important contributions to our under- that only some time after the eruption of the Santorini standing of the cultural history of the latest MBA and volcano did the first LM IB imports reach that site.266 earlier LBA. The systematic inventory of archaeolog- Times contemporary with LM IB on Crete and LH ical remains and the identification of new sites have II on the Greek mainland are well represented by radically changed our picture of the density and dis- Ayia Irini period VII, which can itself now be subdi- tribution of settlements, particularly in the Middle vided. Plentiful finds of the LB III period have been and Late Bronze Ages. The earlier MBA is, however, illustrated, including destruction deposits at Kou- still poorly understood, as is the EBA. The investiga- kounaries on Paros (dated to an early stage of LH tion of EB III levels at Asomatos on Kos is thus of IIIC), a long sequence of ceramic assemblages in the great importance. We are still a long way from un- sanctuary at Phylakopi (LH IIIA-C), and groups in derstanding the significance of substantial gaps in the temple at Ayia Irini (LH IIIC). Publication of LH habitation sequences present on many of the islands IIIA-B deposits from Ayia Irini will also be invalu- of the Dodecanese. able. It should also be evident from this review that We thus seem well on our way to establishing a relative chronology has not been the only, or perhaps relative chronology for the Cyclades that reflects the even the principal, emphasis of recent fieldwork in full range of variation within these islands as a whole. the Aegean islands. Excavations such as those of house If this goal is to be accomplished we must resist the A on Keos or the sanctuary on Melos have been temptation to impose rigid chronological frameworks published with an opulence of detail that makes it on a pan-Cycladic or pan-Aegean scale. Rather the now possible to begin to reconstruct the social and

266 See Davis and Cherry (supra n. 41). 756 JACK L. DAVIS economic organization of prehistoric island centers much better known than it was only a decade ago, for the first time. The full documentation of se- and, perhaps of greater importance, excavations of quences at Akrotiri, Melos, and Keos permits us to settlement sites have established local stratigraphical hypothesize about the processes responsible for sequences in all major parts of the Aegean. The pub- changes in material culture there. Intensive surveys lication of current excavations and advances in the on Keos and Melos, and systematic efforts also to study of absolute chronology promise to resolve long- inventory prehistoric resources in parts of Naxos, los, standing problems in relative chronology and should Siphnos, Keros, Amorgos, Euboia, Skyros, Chios, in make it possible for archaeologists to shift at last the the Sporades, and in the Dodecanese are giving us focus of their efforts from matters of chronology to the first reliable pictures of patterns in settlement the reconstruction of the societies and economies of within entire prehistoric island landscapes. the islands of the prehistoric Aegean.268 Indeed, the It seems fair to say that the Aegean islands should next decade promises to be exciting. no longer be viewed as a backwater but as one of the DEPARTMENT OF (M/C 129) best-documented areas of prehistoric Greece, or even UNIVERSITY OF ILLINOIS AT CHICAGO of prehistoric southern Europe.267 Great progress has BOx 4348 been made. The relative chronology of settlement is CHICAGO, ILLINOIS 6o68o

267 Indeed, the prehistoricarchaeology of Melos has been 268 Two importantsyntheses of availableevidence for ab- employed as one principal case in a textbook now widely solute and relative chronology in the Aegean are in press used in archaeologyclasses in North America: see C. Ren- and deserve special notice: see Coleman (supra n. 8) and frew and P. Bahn, Archaeology: Theories, Methods, and Prac- especiallythe detailed treatmentof the islands provided by tice (London 1991) 438-45. Manning 1992 (supra n. 160).