HESPERIA 72 (2003) H ELLEN15TIC Pages I2I-I45 Dl SCOVE Rl ES AT TE L DOR, I S RAE L

ABSTRACT

This articleis a preliminarypublication of a senes of findsmade in 2000 at , ,during excavations sponsored jointly by the HebrewUni- versityof Jerusalem and the University of Californiaat Berkeley.A limestone Nike anda groupof architecturalfragments are conjectured to comefrom a 3rd-or early-2nd-centuryDoric temple or propylon.Fragments of a superb theatricalmosaic or mosaicsin the opusvermiculatum technique are attrib- uted to an andronor oecusand are comparedwith mosaicsfrom late-3rd- centuryAlexandriaand 2nd-century Delos, Pergamon? Rhodes, and Pompeii. The findssuggest the presenceof a sophisticatedHellenized community at HellenisticDor.

INTRODUCTI ON

Twentyyears of excavationby an internationalconsortium at the harbor townof Dor(ancient Dora: Figs. 1, 2) haveyielded significant remains of the Hellenisticcity. These include stretches of thetown wall and its main gate(areas A, B, andC); an arsenalof catapultballs (area B); numerous houses(areas A, B, C, D2, F,G, andH); olive presses (areas A, D2, andF); andmasses of smallfinds, particularly terracottas, pottery, and coins. Evi- dencefor both orthogonal plans and plans relating to contourlines (areas A, B, C, F,G, andH) hasalso been found. Furthermore, these campaigns haverefuted a numberof long-heldbeliefs about the site.Chief among theseis the contentionof its firstexcavator, John Garstang, that the im- pressiveashlar foundations and associated 10-m-high Ionic columns on thewestern, seaward side of themound (areas F andH) arethe remains of theearliest Hellenistic temple(s) in theMiddle East. On thecontrary, it is nowclear that they are Roman and date to thelater 2nd century C.E.1 In the summerof 2000,a teamfrom the Universityof Californiaat 1. See Garstang1924; cf. Stern Berkeleyopened five squares at Dor on the southernside of the mound 1995,2000. Unless otherwise noted, all (areaD1) in orderto clarifythe chronologyand form of a largebuilding datesin this articleare B.C.E. of the LatePersian/Early Hellenistic period, the so-calledPersian Palace

American School of Classical Studies at Athens is collaborating with JSTOR to digitize, preserve, and extend access to Hesperia ® www.jstor.org ANDREW STEWART AND S. I22 REBECCA MARTIN

MCX \0 Figure1. Mapofthe eastern \ fS ' km 200 I \ Mediterraneanin the Hellenistic ( ^ period.E.Dintino

(Fig.3). Two large pits were discovered on the periphery of the building: oneon the north side and one on the east (Fig. 3: squaresAS 14 and AQ 12).Pit 1 (Fig. 4) in squareAS 14 dates to the Late Hellenistic period. It produceda limestone statue of a Nike (Fig. 5) and two Doric capitals (Fig. 6:a-b) carved in the local , or kurkar.A nearby wall and fills belowtwo nearby floors produced three more architecturalmembers at- tributableto the same building, including an Ionic anta capital (Fig. 6:c). Pit2 in squareAQ12 was Roman, and yielded many fragments of one or moremosaic floors most spectacularly,pieces of a superb composition in opusvermiculatum of theater masks and assorted flora (Figs. 8-10). These finds are described and illustrated in the following pages. Since excavationof the area and its surroundingshas not been completed, how- ever,it is likely that additional fragments and perhaps the buildings from whichthese fragments originate lie beneath the surface of the adjacent squares.This publication must therefore be regarded as preliminary.

SCULPTURE AND ARCHITECTURE

SQUAREAS 14: PIT 1

Pit1 included loci L 26111 and L 26171 in square AS 14 (Fig. 4). L26171 (sealed by a Late Hellenistic or Roman wall, W 16150) yielded theNike (1) and L 26111 two Doric column capitals (2, 3). J

HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I23

- 49 i o 1Za TEL DOR2

b l

o so iZh

D

- - l

@ ! t l

G i l i

+ ! +

Xa-ew**2wa@@wV-s

\w@* >

\* *;X--*^

Figure2. Tel Dor 2000, site plan. Areasexcavated in 1980-2000 are markedin black.J. Berg I24 ANDREW STEWART AND S. REBECCA MARTIN

AU AT AS AQ AR

15

14

13

Figure3 (lefi). Schematicplan ofthe easternpart of areaD1 showingthe so-calledPersian Palace, the build- ingsabove, and the locations of pits1 and2. J.Berg

Persian:1 HellenisticE RomanE Figure4 (aboqwe).Area D1, view of the Nike (1) and capitals(2, 3) in pit 1 duringexcavation. Diagnosticpottery and lamps recovered from the pit comprisethe PhotoI. Hirschberg following: 261206 wheelmadelamp (3rd century) 261211 importedAttic plate (4th century) 261212 bowlwith outcurved rim decorated with rouletting andpalmettes (3rd century) 261372 "Megarian"relief bowl fragment; unguentarium fragment 261374 brazierfragment 261375 "Megarian"relief bowl fragment; unguentarium fragment;Eastern-type relief bowl fragment (2ndcentury) 2. Forthe introductionof ES(A)see 261418 importedAttic lamp with handle missing (4th century) Slane1997, pp. 269-282 (TelAnafa), 261446 unguentariumwith tall foot and teardrop-shaped body whereit firstappears sealed under a (late3rd-early 2nd century) buildingconstructed ca. 125; cf. Berlin 1997, p. 24; atTel Kedeshin the Gali- Thepit contained no EasternTerraSigillata A, orES(A), a fabricthat lee, ES(A)is absentfrom the destruc- wasintroduced into northernIsrael around 140-130. At Dor it firstap- tionlevel of 145 but appearsin the pearsin areasA andC in stratumIII, the earliestpart of whichdates to reoccupationlevel of the lastquarter of the century(Sharon Herbert, pers. around150-100 (though the chronologyis imprecise).2The pitwas thus comm.,2002). ForDor see Stern1995, probablyclosed in themid- to late2nd century, before ES(A) reached the I:A,pp. 43-44, 233-234; I:B,pp. 218- site,and almost certainly no laterthan ca.100. 221. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I25

Figure5. Limestonefigure of Nike 1 Headlessstatue of a wingedNike Fig. 5 (1).Front, left profile, back. Nach- sholim,Center of Nauticaland Inv.261429. Limestone,no tracesof plaster.H. 0.61; p.H. of figure0.54; RegionalArchaeology at Dor. W. 0.27; D. 0.23 m. Missinghead, both arms,right breast, wings, right Scaleca. 1:6.Photos G. Laron. shoulder,and muchof rightside abovewaist; battered and weathered. Draped in a long,V-necked peplos with overfold,girdled below the breasts,and a cloak crossingthe backdiagonally from lower right to upperleft. The figurestrides forwardwith rightleg advancedand left armextended sideways, perhaps to hold up the cloak.A roughlypunched, rectangular tenon, 0.15 x 0.07 x 0.07 m thick,protrudes below the feet. On the frontthe recessesof the folds are roughlypunched; the limbsand draperyare modeled with a fine claw chisel (threeteeth/6 mm) overlaidby muchflat chiseling;the backis sketchedonly, with a punchand coarseclaw (five teeth/25 mm).

2 Doric capital Fig. 6:a Inv.261430. Kurkar,with smallfragments of plaster(0.015 m thick) adhering.H. 0.270;Diam. (column)0.460; abacus 0.61 x 0.62; H. (shaft)0.080; H. (echinus)0.080; H. (abacus)0.110 m. Upperedges of abacusbeveled in a band0.07 m wide, reducingheight of sidesof abacusto 0.08 m. Tracesof three neckingrings at top of shaft.

3 Doriccapital Fig. 6:b Inv.261985. Kurkar,with some plaster(0.005-0.015 m thick)adhering. H. 0.370;Diam. (column)0.460; abacus0.62 x 0.62; H. (shaft)0.160; H. (echinus)0.090; H. (abacus)0.120 m. Upperedges of abacusbeveled in a band0.10 m wide, reducingheight of sides of abacusto 0.095 m including plastercoating. No neckingrings visible. ANDREW STEWART AND S. I26 REBECCA MARTIN

a b c

Figure6. (a, b) Doric capitals2, 3; kurkar.(c) Ionic antacapital 6; S QUAR E A S/AT 1 4: WAL L W 1 6 3 6 0 limestone.Nachsholim, Center of Nauiicaland WallW 16360, RegionalArchaeology constructedto thesouth of pit 1 (Fig.3),is LateHellenis- at Dor. Scale1:10. A. Adams, E. Dintino. ticlEarlyRoman; no diagnosticpottery was recovered from it. In secondaryuse as part of the foundationof thewall was: 4 Doric capital Inv.260202. Kurkar,with some plaster(0.005-0.015 m thick)adhering. H. 0.320;Diam. (column)0.440; abacus0.62 x 0.63; H. (shaft)0.150; H. (echinus)0.065; H. (abacus)0.105 m. Upperedges of abacusbeveled in a band0.03 m wide, reducingheight of sidesof abacusto 0.10 m, including plastercoating. A singlethick (0.015 m) neckingring visible.

SQUAREAR IS: FILL LOCI L 262I2, L 26I83 SquareAR 15 includesfill locus L 26212,1m belowa fragmentaryfloor, F26076; and fill locusL 26183,immediately below fragmentary floor F 26134,which itself underlies floor F 26076.Floor F 26076is Roman;the latestpottery in thefill below it, L 26212,was typical of theEarly Roman period,including fragments of ridgedstorage jars. Lyingin the fillwas: 5 Columndrum Inv.262260. Limestone,no plasteradhering. H. 0.530;lower Diam. 0.585; upperDiam. 0.570 m. A beveledgroove runs up its side,0.06 wide narrowingto 0.04wide in the trough,and 0.04 m deep;in its bedding,an empolioncutting 0.05x 0.07 x 0.03 m deep.

FloorF 26134is alsoEarly Roman, dated by the latest pottery from its fill locusL 26183. Lyingin thefill was:

6 Ionicantacapital Fig. 6:c Inv.261744. Limestone,with smallfragments of plaster(0.002-0.005 m thick)adhering to cavettomolding only. H. 0.310;below, W. 0.237 x L.0.525; above,0.295 x 0.575 m. On uppersurface, a dovetailclamp hole: L.0.10, W.0.075/0.055,D.0.03 m. H. (wallsection) 0.135; H. (moldings)0.175 m: frombottom, half round;ovolo; half round;cavetto; fascia. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I27

DISCUS SION Althoughthe actualbuilding to whichthese fragments belong has yet to appear,the architecturalfragments 2-5 couldbe froma Dorictemple, propylon,stoa, porch, or colonnaded court. As willbecome clear, however, if thesefragments, the antacapital (6), andthe Nike (1) all belongto- gether,then some of thesepossibilities can be excludeda priori. Despitetheir different materials (limestone and the local sandstone), theseobjects probably were all associated with a singlestructure. Not only werethey deposited near each other and at approximatelythe samelevel, butthe diameter of thelimestone column drum (5) neatlycorresponds to thatof theDoric capitals (2-4) whenone allows for the obligatorydimi- nutionof the shaft(Fig. 7). The empolioncutting in the lowersurface onlyand the materialand size of the drumindicate that 5 wasthe lowest element,doweled to thebuilding's stylobate for greater stability. Above it, the naturalfriction of the stonewas evidently deemed sufficient, as was usualat Dorwhere clamps and dowels were rarely used. The needto pro- - - - tectthe columns from damage by passing foot traffic, together with a con- victionthat a toughermaterial than kurkar was needed to carrythe heavy weightof the colonnade,could easily account for the substitution.The plastercoating, still visible on 2-4 and6, wouldhave covered any differ- encesin colorand texture. The antacapital (6) hasroughly the sameproportions as the column capitals(especially 4) andmay also belong to thesame structure. Here the choiceof stonewas perhaps determined by the complicated molding, which wouldhave been difficult to carvein kurkar.If the capitalbelongs to the ensemble,the orderincorporated some Ionic elements a commonfea- tureof LateClassical and Hellenistic Doric. If not,we havecomplemen- taryfragments of twobuildings of almostexactly the samescale, a kurkar andlimestone Doric one and a limestoneIonic one, both destroyed at the sametime anunlikely alternative. Thebuilding apparentlywas securedwith a grille (fitted into the groove on5) andembellished with Nike akroteria (as 1). No remainsof its entab- laturecame to lightand perhaps none should be expected.Kurkar is too fF weaka stoneto sustaina lintelof anylength, and of thehundreds of miscel- \ lb laneousGraeco-Roman architectural fragments recovered at Dor, not one (to ourknowledge) is fromthis kind of entablature.The building'ssuper- ag'\ structurewas probably of timber.Garstang's Ionic "temple" on the west j i u \ l \ sideof the mound(area F) furnishesa preciseparallel. As statedabove, these Doric fragments could theoretically belong to a numberof buildingtypes. Any freestanding structure must have stood to the northof the excavatedarea (area D1: Figs.2, 3);the possibilityof a porchimmediately points to thestill incompletely excavated "Persian Pal- ace."But the porchesof Hellenisticpalaces and palatial mansions never 0 0.5 1 METER l l l l seemto havebeen embellished with akroteria, and an interior colonnade Figure7. Restoredelevation of Doric affordsno placefor sculptural refinements of thiskind. Furthermore, the columnincorporating 2 and 5. dimensionsof the Dor columnsexceed all but thoseof the verylargest E. Dintino templeand palatial courtyard colonnades in theHellenistic Near East (see ANDREW STEWART AND S. REBECCA I28 MARTIN below).Nor is a freestandingstoa a likelypossibility; these were not popu- larin the ,and Hellenistic examples built elsewhere apparently es- chewedsculptured akroteria completely.3 Therefore,unless this Doric building was uncanonical (always a pos- sibility,especially in the East),or the Nike stood on anotherstructure en- tirely,a templeor propylon is themost likely source for these architectural fragments.Yet, to ourknowledge, no pre-Herodian/EarlyRoman Doric templesor propylaiahave been discovered in Israeland only three such templesare known in Syriaand Jordan:4 1.Tell Jebel Khalid (Amphipolis/Tourmeda/Nikatoris?), North .Hexastyle amphiprostyle limestone Doric temple, ca.13 x 20 m, surroundedby altars.Discovered in the summer of 2000.Published by G. Clarkeet al.5Date: 3rd century. 2. TellNebi-Mend (Laodicea ad Libanum), South Syria. Tetrastyle prostylelimestone Doric temple, ca. 9 x 20 m. In area10 of the lowertown. Excavated in the early1990s. Unpublished. Date: possiblylate 3rd-early 2nd century. 3. Pella,Jordan. Stylobate and part of the facadeof a Dorictemple in pinkmarble/limestone. On Tell Husn (S. Hill), University of Sydney,area 34. Discoveredin 1993.Unpublished. Date: 1st century? Unfortunately,the column heights of allthree temples are uncertain. Nev- ertheless,as discussedbelow, the templeat TellNebi-Mend might be a usefillguide to thepossible appearance of thebuilding at Dor; its publica- tionis eagerlyawaited. The Dorcolumns are easy to reconstructon paper thoughthe mea- gerremains do makethe resultlook somewhat comical (Fig. 7). If 5 was indeedthe lowest drum, simple extrapolation from its twodiameters and thediameter of the columnstump on 2-4 indicatesa shaftabout 4.06 m highand a totalcolumn height of ca.4.27 m, or about12 Ionicfeet. As notedabove, these dimensions exceed those of mostNear Eastern sanctu- aryand palace colonnades, but theyare perfectly acceptable for a small templeor sizeablepropylon (see Table 1).6 The lowerdiameter/height

3. Synopses: Coulton1976, pp. 55- houses);Stucchi 1987, p. 258, fig. 14; locationof the site,see Gawlikowski 56;Nielsen 1994. See also Netzer2001. Fedak1990, pp. 148-150,fig. 221 1996,p. 128;Talbert 2000, map67 4. We thankGraeme Clarke, (Es-Suweida,Syria: Tomb of Ham- (squareG4). directorof theTell Jebel Khalid rath);Herbert 1994, pp. 37-42, fig.2:7, 6. In addition,at Dura,the citadel excavations,for alertingus to these 8,pls. 3-14 (TelAnafa, Israel: mansion, palace'scourtyard columns had a lower andfor providing us with plansand drumsonly); Netzer 2001, pp. 88-91, diameterof 0.61 m andan upper reconstructedelevations of his newly 103,304-305, figs. 127, 128,453 diameterof 0.51 m (Dura II, p. 14); discoveredDoric temple. (Jericho:Hasmonean palace, pavilion atNippur, the palace'scourtyard ForDoric in othercontexts see, andporticoes). The two templesat columnswere a massive0.84 m thick e.g.,Bliss and Macalister 1902,p. 57, Hammonare Ionic prostyle and butonly 4.2 m high,an idiosyncratic pl.19:7, 8 (Maresha:houses); Avigad amphiprostyle,and lack canonical 1:5ratio evidently occasioned by the 1954,p. 95, fig. 57;Fedak 1990,p. 142, propylaia:see Dunandand Duru 1962, useof brick(Fisher 1904, p. 422);and figs.203, 204 (Jerusalem: Tombof p.48,fig.lO,andp.76,fig. 17. atJericho, the Hasmoneanpalace's BeneHezir); Dunand and Duru1962, 5. Clarkeet al.2000, pp. 123-126, pavilioncolumns (ca. 100-80) were pp.31-34, pls.22, 23, 98, 99 (Ham- fig.2; Clarke,forthcoming, with re- about5 m high andthose of the garden mon:hypostyle hall); Ploug1985, visedelevation of the facade.For the colonnadeabout 4 m high (Netzer p.128, nos. 1, 2, fig.21 (Hama,Syria: conjecturedancient name and exact 2001,pp. 304-305, fig. 453). HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I29

TABLE 1. EXAMPLES OF DORIC LATE CLASSICAL AND HELLENISTIC TEMPLES, PROPYLAIA, AND PORTICOES

LouerDiam. ColumnH. Interaxial LouerDiam. Interaxialto Struchlre Date of Columns(m) (m) (m) to ColumnH. to ColumnH.

Epidauros,Tholos ca.360 1.0 6.88 2.37 1:6.9 1:2.9 Tegea,Temple of AthenaAlea ca.340 1.55 9.56 3.58 1:6.2 1:2.6 JebelKhalid, Dorictemple 3rdcentury 0.88 5.47 or 4.55 2.40 1:6.2or 5.1 1:2.3or 1.9 Pergamon,Temple of AthenaNikephoros ca.250 0.75 5.25 2.37 1:7.0 1:2.2 Pergamon,Doric Templeof Asklepios ca.220-190 0.69 4.78 2.14 1:7.0 1:2.2

Pergamon,Propylon to Athena Nikephorossanctuary ca. 160 0.68 5.0 2.49 1:7.3 1:2.0 DuraEuropos, Bicolumnar Monument(Propylon?) 2nd century? 0.90 6.3 2.9 1:7.0 1:2.1 Delos,Propylon to Kyntheionsanctuary 95/94 0.53 4.0 1.8 1:7.5 1:2.2

JebelKhalid, Portico of Governor'spalace 3rdcentury 0.70 3.6 2.1 1:5.2 1:1.7 Delos,Stoa of AntigonosGonatas ca.250 0.70 ca.4.5 2.53 1:6.4 1:1.8 Hammon,Porticoes of 222/221 0.56 3.92 2.16 (North) 1:7.0 1:1.8 Milk'Ashtartsanctuary 2.42 (East) 1:7.0 1:1.6 Priene,Agora, North Stoa ca. 150 0.70 5.2 2.32 1:7.4 1:2.2 Athens,Stoa of AttalosII ca. 150 0.74 5.23 2.43 1:7.1 1:2.2

ratioofthe columnwas about 1:7.3; a l-cm-thickplastercoating (see 2-4) wouldhave reduced this to about1:7.0. As such,it is canonicallyEarly to Mid-Hellenistic,as Table 1 shows.7 Furthermore,minus its beveled top, one of theDor capitals (2) neatly conformsto the Vitruvian(i.e., Hellenistic) division of the Doriccapital into threeequal parts (Vitr. 4.3.4). All threeof the Doriccapitals (2-4), althoughidiosyncratically proportioned in otherrespects, roughly echo the High andLate Hellenistic ratio of echinusheight over abacus width thatcharacterizes the capitalsof laterHellenistic buildings from Lindos andDelos, and the capitalsfrom Hammon (Umm El'Amed) in southern a Phoeniciancult site only 62 krnup the coastfrom Dor. The capitalsfrom Hammon are as unevenly proportioned and finished as those

7. Sincesome of the measurements fig. 18;2001b; and forthcoming (Jebel indeed7.3 timesas highas theirlower uponwhich these ratios are based are Khalid);AvP II, pp. 11, 50 (Pergamon, diameter);Delos X;l, pp. 98-99 (Kyn- estimated(but must be correctto with- Athenatemple; propylon);AvP£.2, theion);Delos V, pp. 17-18 (Anagonos in a coupleof centimeters),they have pp. 19-25 (Asklepiostemple); Downey stoa);Dunand and Duru 1962, p. 37 beenrounded off to one decimalplace. 1988,p. 83, fig. 38 (Dura;Downey (Hammon,with p. 187 forthe date); Sources:Roux 1961, pp. 177-178 2003 showsthat the 2nd-century Wiegandand Schrader 1904, p. 193 (Tegea;Epidauros); Pakkanen 1998, ensembleof herfig. 35 is a fiction, (Priene);Travlos, p. 513, fig. 645 (Atta- p. 73 (Tegea);Clarke 2001a, p. 223, thoughthe columnsare real and are los stoa). I30 ANDREW STEWART AND S. REBECCA MARTIN fromDor, and the site has also produced some Ionic anta capitals that are verysimilar to 6.8 This marksthe limitof whatcan be plausiblyextrapolated from the remains.But the combinationof the Nike (1) andthe columndrum (5) withits ca.0.60-mdiameter prompts a further,purely speculative, conjec- ture.The statueis nowheadless, but when complete it stoodaround 0.67- 0.70m (2 Ionicfeet) high. If it too conformedto Vitruvius'sprescriptions (Vitr.3.5.12)and equaled the height ofthe building'stympanon, then the latterwould also be around0.67-0.70 m.9 Hellenistic Doric tympana are typicallyeight to tentimes as wide as they are high and Hellenistic Doric horizontalcornices are aroundnine to thirteentimes as wide as the tympanonheight. So hypothetically the tympanon should measure ca.5.3- 7 m wideand its horizontalcornice ca. 6.0-9 m wide. As for the colonnade,the exampleslisted in Table1 indicatethat a columndiameter of 0.60 m wouldproduce interaxials of around1.8-2.1 m, andthus a tetrastylefacade of modestwidth (6-7 m),on a foundation about8-9 m wide.As a crosscheck,the Doriccapitals and drums tenta- tivelyattributed to the 7.5 x 3.5 m foundationof the Romanpropylon to TempleH at Dor areabout 10% smaller than 2-4 and alsoindicate a 8. C£ Coulton1979, p. 81, figs.2-4 tetrastylefacade; the propylonto the Kyntheionat Delosoffers a rough (group10); Dunand and Duru 1962, parallel.l°A hexastylefacade for 2-4, however,would yield a muchmore pp. 102, 104, 117,133,figs.23,26,35, substantialwidth (9.6-10.6 m), requiring a foundation around 11.5-12.5 52, pls.22,23; cf. Shoe 1936,pp. 174- mwide, and producing a tympanon considerably higher than the Nike (1). 176 and,e.g., pl. 17:8,31 (Delos). 9. This rulewas generally observed The excavation'snext priority, therefore, is to searchfor a foundationthat fromthe 5th centuryonward: King is ca.8-9m across,tailored for a tetrastyleprostyle or amphiprostyle temple 2000,p. 104. or,perhaps, a propylon. 10. Inventorynumbers: the foun- What of the Nike?The pose,tooling, and unevenfinish indicate dationis W 20270/20280;the capitals thatthe figure stood on thebuilding's left-hand corner vis-a-vis the spec- are202086,202205, and 203855; and the drumsare 204409 and 203955. The tator.The materialcould suggest Cypriot manufacture, though the indif- capitalsand drums are from late (phase ferentquality perhaps militates against this possibility. The poseis a stock I or later,i.e., Crusader)fills. For the one,exemplified most famously in the Hellenisticperiod by the Nikeof Kyntheion,see DelosX;l, pp. 98-99. Samothrace. 11. On theseNikai in general,see Typologicallythe figureseems to fit betweena collectionof Late King2000,pp.104-116;LIMCVI, ClassicalNikai from 4th-century Megara, Epidauros, and Delos, on the 1992,pp. 881-883,nos.381,388,401- 406, s.v.Nike (U. Grote).For illustra- one hand,and a seriesof flamboyantlybaroque ones from 2nd-century tionsof the late-4th-centuryNikai, see Pergamon,Samothrace, and Halikarnassos, on theother.1l The restrained Purgold1881 (Megara);Marcade 1951 poseand drapery align 1 squarelywith the formergroup, and the figure (Delos);Alscher 1957, pl. 4 (Megara); showslittle sign of the stronglytapering proportions, sprung rhythms, Yalouris1967 (Epidauros);Gulaki andwild, frothy draperies of the latter.In the Aegeanat least,this high 1981,figs.35-41 (all);and Webb 1996, fig. 115 (Delos);and for the 2nd-cen- baroque-evenrococo-fashion began to emergearound 200, to judge by turyexamples, see Marcade1951, p. 84, a fineterracotta Nike in Parisfrom a well-datedgrave at Myrina.l2 Unlike fig. ll:b (Delos);Schober 1951, pls. 90, theNike from Dor, these 2nd-century Nikai wear their girdles hiked up so 98 (Pergamon);Grote 1992, pl. 15 farthat their breasts jut provocatively,and sometimes one breastis even (Pergamon);Webb 1996, fig. 135 leftbare. (Samothrace);Poulsen 1997, pp. 77-78, Unfortunately,the hundredyears between these two groupsof figs.97-100 (Halikarnassos). 12. Paris,Louvre MYR 171:Mol- Nikai(ca. 300-200) is somethingof a blackhole: certifiably 3rd-century lard-Besques1963, p. 67, pl. 80:d;the freestandingNikai are all but nonexistent. The onlyviable candidate is the gravecontained autonomous coins of impressivestatue, perhaps Athena Nike, from the shipmonument in the Myrinadatable to 196-190. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I3I

TABLE 2. DOR CHRONOLOGY: LATE 4TH THROUGH 1 ST CENTURIES

Date Etvent

332 Alexanderthe Greatpasses by Dor on his marchfrom Tyre to Gaza andEgypt. 323 Deathof Alexander;Phoenicia (including Dor) soondisputed between AntigonosOne-Eye (satrap of Syria)and Ptolemy (satrap of Egypt). 301 Antigonosis defeatedand killed at Ipsos;Ptolemy annexes Judea, Phoenicia,and Coele-Syria. 219 Dor,now a Ptolemaicfortress, withstands a siegeby AntiochosIII of Syria(Polyb.5.66.1), who thenmarches on to defeatat Raphiain 217 (the so-calledFourth Syrian War). 202-199 AntiochosIII returns,destroys the Ptolemaicarmy at Paneion(Banyas, NorthGalilee), and annexes Coele-Syria, Phoenicia (including Dor), andJudea (the so-calledFifth SyrianWar). 139/138 Dor,now a Seleukiddependency occupied by the pretenderTryphon, is besiegedby AntiochosVII Sidetesand Simon Maccabee. Tryphon, however,manages to escape(I Maccabees15:11-37;Joseph.AJ 13.223-224;BJ 1.50). 102-99 Dor,now ruled along with Gazaand Strato's Tower by the tyrant Zoilos,is takenby the Hasmoneanking Alexander Jannaeus, who treatsthe townharshly (Joseph. AJ 13.324(335); 14.76; Syncellus558). 63 Pompeyabolishes the Seleukidmonarchy, takes Jerusalem, detaches Coele-Syriaand southern Phoenicia from the Hasmoneankingdom, andgives Dor andthe othercities their freedom (Joseph. AJ 14.4.4; BJ1.7.7). Dor restartsits calendarat Year1 andmints coins dated fromthis year. 34 MarkAntony gives Kleopatra VII of Egyptthe coastalcities of Phoeniciaand Judea, with the exceptionsof Tyreand (Joseph.AJ 15.4.1;BJ 1.18.5). 31-30 Battleof Actium;Octavian marches south from Syria, invades Egypt, andtakes Alexandria.

Agoraat Cyrene.Possibly erected in connectionwith the ThirdSyrian Warof 246-241,the figurehas a torsosimilarly proportioned to thatof the Nikeat Dorand achieves a similarrelationship between clothing and body,but has far longer legs and a differentlydraped himation. Finally, the symmetricallyarranged swallowtail folds on the Dor Nike add an archaistic touchthat is unparalleledin thegenre until the Roman period, when hints of it occuron a numberof Nikaifrom Palmyra and Jordan.13 If the templeor propylonwas indeed built at Dor in the 3rdto early 2ndcentury, then it lastedbarely a hundredyears: the pottery in pit 1 dates 13. See Ermeti1981 (Cyrene); its demolitionto ca.150-100. Although any number offactors could have LIMC VIII,1997, pp. 879-881, occasionedits constructionand destruction, the eventsoutlined above in nos. 15, 26b,36, s.v.Nike (P.Linant de Bellefonds). Table2 areworth recalling.l4 14. Forsources and discussion, see A naturaltemptation would be to connectthe Nikefrom Dor and its Dahl 1915,pp. 65-78. templeor propylonwith the Raphiacampaign of 219-217,and (in the I32 ANDREWSTEWART AND S. REBECCAMARTIN

absenceof recordedearthquakes or othernatural disasters) their destruc- tioneither with Sidetes' siege in 139/138or with that byJannaeus in 102- 99.According to Josephus, Sidetes besieged Dor both by land and sea, and catapultballs found on the city'sseaward side demonstratethat it was indeedbombarded from that direction. As forJannaeus,his egregious bru- talityand relentless destruction of paganshrines in thecities that resisted himwere legendary. Indeed, Syncellus even lists Dor amongthe coastal townsthat he destroyedand whose inhabitants he massacred.15But so far the sitehas yielded no evidenceof widespreaddestruction in thisperiod, andto makethese or anyother connections, more excavation and much morehard evidence are needed.

THE MOSAIC

Thesuperb fragmentary mask-and-garland mosaic also unearthed in 2000 (Figs.8-10)is thefirst major example of Hellenisticopus vermiculatum in theregion. Unfortunately, it was not found in situ,but had been broken up andtossed into a Romanpit (pit2) in area-D1 on thesouthwest side of the tel (Figs.2,3). Giventhe qualityof thismosaic, it is hardto believethat someonedestroyed it on a whim.Earthquake damage or urbanrenewal cometo mindas possible explanations, but until the original context of the mosaicis found,no firmconclusions are possible. The fragmentsvary in size fromsmall clusters of tesseraeto larger sectionsmeasuring over 0.40 x 0.30m. Thanks to theefforts of colleagues fromHebrew University and at the sitemuseum at Nachsholim,the dis- paratepieces of one maskand the areaadjacent have been successfillly restored;a fragment of a secondmask and a separategeometric zone await supplementationfrom filrther excavation and restoration. The masksand garlandsare almost certainly part of a borderfrieze; the central emblema, if therewas one, has yet to be identified.

SQUAREAQI2: PIT 2 Pit2 in squareAQ12 includedloci L 26053,26081-82,26121-22,2615F 52,26164-65,26169,26195,26234,26236-37,and 26248. The pit con- tainedmany mosaic fragments, apparendy discarded at random. Unfortu- nately,the diagnosticpottery from the pit was extremely limited: 260674 Romanlamp (2nd century C.E.) 260909 Atticblack-glaze fragment; Roman Western Terra Sigillata The majormosaic fragments are:

7 Sectionof a mask-and-garlandfrieze Fig.8:a-c Inv.260885.Recomposed from several fragments (A-I) foundin L 26081. Stone,ceramic, and glass set into shelly mortar. H.0.511; H. (field)0.428; 15.E.g., at Gaza shortly afterward W. 0.747;Th.0.020-0.204 m. Extensivedamage on top,bottom, and sides and (Joseph.AJ 13.364): see most recently to noseand right eye of mask;heavily encrusted before restoration. Youthful Bar-Kochva1996, pp.127,132-133. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I33 male maskfacing to its rightwith speira (or coil of hairover the templesand forehead),hat, andwool fillet tied by a blue tainia, set againsta backgroundof fruitsand flowers.Bordered below by a redband of threerows of tesseraeand aboveby two similarred bands, separated by a stripof white.

8 Fragmentof a bouquet Fig.9 Inv.261241. FromL 26122. Stone,ceramic, and glassset into shelly mortar.H.0.475; H. (field)0.269; W.0.294; Th.0.204 m. Extensivedamage on top, bottom,and sides;heavy encrustation. Red and gold cloth wrappedaround a fruitand floralbouquet. Bordered above by two red strips,separated by a strip of white.

9 Fragmentwith perspectivalmeander Fig.10, left Inv.261718. FromL 26169. Stone, ceramic,and glassset into shelly mortar.H.0.168; H. (field)0.120; W.0.177; Th.0.133 m. Damageto sides; verylittle encrustation.Perspectival meander set into blue fieldwith rosettes framedby a partiallypreserved red andwhite border.

10 Fragmentwith perspectivalmeander Fig.10, right Inv.261718. FromL 26169. Stone, ceramic,and glass set into shelly mortar.H.0.135; H. (field)0.122; W.0.124; Th.0.126 m. Damageto sides; verylittle encrustation.Perspectival meander set into blue fieldwith rosettes.

The techniquerepresented in thesemosaics is trueopus vermiculatum, using3-5 mm2tesserae in the fieldin a widerange of reds,blues, and yellows.There is alsoextensive use of glassin tonesof blueand green, characteristicof Hellenistic mosaics.l6 To datethere is no evidencefor the useof leadstrips.l7 The white limestone field is framedat topand bottom by redbands using larger tesserae up to 4 x 6 mm in size.Most of the tesseraeused for this white background and adjacent sections of floorare rectilinearand laid horizontally. But as theyapproach the decorated areas theybegin to curveand include tiny chips,often only 1 mm across,in orderto followthe contoursof the fruits,flowers, and mask. This tech- niqueis particularlyclear where the top of themask approaches the upper redborder. The workis set intoa bedof finemortar 0.019 m thick,sup- portedby a heavybacking of coarsermortar.l8

16. Guimier-Sorbetsand Nenna floors.Cf. Dunbabin1979. 1992.Westgate (2000) attests to the 18. Preservedup to 0.204 m thick popularityof the colorscheme. Its in the bouquetfragment. In PalaceV basicpalette of redand yellow stone at Pergamonthe tesseraeare set into andbright blue and green glass was a thin mortarbacked by a thicker especiallypopular with easternHellen- (0.015)one, a coarseaggregate mortar isticmosaicists because of its visual (0.03-0.04thick), and rubble (ca. 0.20 kinshipwith opus sectile. thick);AvPV.l,pp. 53-54. At Delos 17.Joyce (1979) sees the use of the tesseraeare generally set into leadstrips in Delos as one indication mortar0.015-0.02 thick, backed by a of theirorigins in the pebblemosaics coarsermortar (0.03 thick) and then of mainlandGreece, whereas in Punic- mortaredrubble (ca. 0.04 thick);Delos influencedPompeii, lead strips are XXIX,pp. 26-27. unknownin the earlyopus signinum I34 ANDREW STEWART AND S. REBECCA MARTIN

TH E MASK FRAGMENT The maskand its associatedfloral frieze (7) dominatethe white limestone floor(Fig. 8:a-c). The maskoccupies the rightside of the fragmentand is turnedslightly to the spectator'sleft. The youthfillface with its heavy eyelids,deep-set, large eyes, full mouth,and partedlips is typicalof New Comedymasks (see below).Topping the maskis an extravagant, trefoil-shapedhat. The speirais held in placeby a rolled,brown wool filletsecured by anX-shaped tainia. It is furtherembellished with sprays of broad-leafedand variegated ivy (kittos)and its fruits(korymboi). The blueribbons that billow out on eitherside of the faceare the endsof the tainia,implying that it servesa doublefunction: it holdsthe whole con- coctionof hat,fillet, and fruits together, and it alsosecures the maskto thegarland. The rich floral frieze enhances this luxuriant Dionysiac atmo- sphere.It featuresivy sprays (kittos and korymboi), pinecones (konoi or stro- biloi),wild olives (agrielaiai), pomegranates (rhoai), and wild roses (rhoda agria).l9 The maskitself is 0.361m highand 0.369 m wide.The faceis ren- deredin stones(nine to sixteentesserae per square centimeter) laid along contourlines and curvingaround features to suggesttheir volume. No glassis usedon the face.A wide rangeof colorsis employed:pinkish- brownand light gray for skin;pinkish-red for roundedfeatures like the cleftchin, ear, left cheek,and nostrils; and white mixed with soft yellow forhighlights. The properright side of the faceis shadeddark brown to emphasizeits recession.Various tones of redare used for the partedlips, withlighter red for highlights. The openmouth is brown-blackand ends abruptlyat its properleft in a verticalline, showing that it is cutthrough the fabricof the mask.The extantportions of the noseare prominently outlinedin brown.The fullright cheek, seen almost in profile,contrasts pointedlywith the smoothleft one,emphasizing the rotation of theface. The eyesare heavy-lidded and are hooded by thick, dark eyebrows. There areno eyelashes.The enormouspupils occupy three-quarters of the brown irises.Here the tesserae are truly tiny- theleft iris and pupil use no fewer thantwenty-eight of them. The faceis framedat the sidesby brownwavy hair and ringlets, and aboveby a thick,banded fillet. This fillet, rendered in brownsand blues, is outlinedat its bottomwith blue glass. Four small clusters of lightyellow- greenivy fruits stud the trefoil-shaped red and gold hat; each casts a shadow renderedin the darkhues of the backgroundstones, either red or blue. The ivyleaves are rendered in olivegreen and dark green glass and con- trastsharply with the turquoisestone. The blue-greenribbon that ties themall together is renderedin darkblue, bright blue, and light blue glass; greenglass; and stones that are nearly turquoise. These variations create a strikingchiaroscuro and depth. The verysmall, tightly laid tesserae mea- surebetween 2 and3 mm2. Likemany other Hellenistic theater masks, this one is eclectic.It seems to combinefeatures of twocomic masks described in Pollux'sOnomasticon 19. Foridentification of floraltypes, (4.147):mask 13, the DelicateYoung Man, and mask 16, the episeistos, or seeAbbe 1965, pp. 100, 147, 156-160; SecondWavy-Haired (or perhaps Floppy-Haired) Young Man. These two Harrison1962. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I35

a

b

Figure8. (a) Mosaicfragment 7 with comicmask and floraldesign; (b)detail of ribbon,flowers, and pomegranates;(c) detail of mask. Nachsholim,Center of Nautical and RegionalArchaeology at Dor. PhotosG. Laron c /d , ,, e - >\ Cs t C^ \ e ,

I36 ANDREW STEWART AND S. REBECCA MARTIN

E2igure9 (lefi). Mosaicfragment 8 withbouquet of flowersand fruits. Nachsholim,Center of Nauticaland RegionalArchaeology at Dor. PhotoG. Laron

Figure 10 (above). Mosaic fragments youthsspend too muchtime indoorsenjoying parties. Pollux describes 9 and 10 with perspectivalmeanders. themas follows: Nachsholim,Center of Nautical and RegionalArchaeology at Dor. (13)The DelicateYoung Man has hair like the Admirable and is PhotoDor Excavations theyoungest of all,white, reared in the shade,suggesting softness.

(15)His hairis wavy,as is thatof (16)the SecondWavy-Haired, who is moredelicate and fair-haired.20 A terracottasuspension mask from Amisos (Fig. 11) alsocombines aspectsof themasks described by Pollux.2l The maskdates to themid-2nd century,the beginning of Webster,Green, and Seeberg's period 3 (ca.150- 50).22Like the mask from Dor, it hasa soft,round face, large, heavy-lidded eyes,and wavy hair. Wool fillets and ribbons bind the hair and hang down on eitherside of theface; loose ringlets (only partially preserved on the left sideof the mask)also frame the face.

20. Poll.4.147: (13) o 8 °Cs°CA°S vol. 1, pp. 19-22. Webster,Green, and

, , \ \ \ , V£0tV6AXOg. 1t0CX£5 £V XAlA IOV WAy- Seeberg(1995, vol. 2, p. 210) notethat

Xpov, savzv 8£ V£XTT0g, £OXOg, the Louvremask should be reinter- AXtTpO96Ag, sAA0q OWOONA0V. pretedas a femaleowing to its parted (15) £st£60VTL L Tt0CX£g, 0aW£0 XL hair;but sinceno partingis visiblein t10 TZ 0£UT£t00 £st£L0, sAA0- the publishedimage the ribboncovers T£t00 OVIt XL tAVUX QV XOpNV. oree c the spotwhere it shouldbe we con- textand translations are from Webster, tinueto identifyit as a youngman (see Green,and Seeberg 1995, vol. 1, Webster1961, p. 89, maskZT 5; and pp. 19, 21. Webster1969, p. 89, maskZT 5). 21. Paris,Louvre inv. D 510:Mol- 22. Webster,Green, and Seeberg lard-Besques1972, p. 87, pl. lll:d; 1995,vol. 1, pp. 60-64. cf.Webster, Green, and Seeberg 1995, FipreLaD i1011. Amisos,Cn, IlEariondeyterracotta coc- i^t^0 0 00 : 0 : 00-: 1--_ gS 8

HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I37

masleMid-2nd century. Paris, i X--^ t-::: 0t 0:40: f::- 00 __

The Dor maskXsfloral fringe consists primarily of ivy,with two olive leavesat the lowerleft. But in the friezeproper, olives and flowers soon takeover, then pomegranates and oak leaves. The five- andsix-petaled wildroses appear to be randomlyplaced, yet areusually located near the olivesprays; occasionally the flowers are shovwn in profile.With the excep- tionofthree glass tesserae in thecenters of someflowers, the tesserae used forthe flowers are of stone.Their petals are darker at the center and shade to whiteat theiredges. The two pomegranates(Fig. 8:a-b)are largely stone,with darkercolors in firedclay. The uppermostis modeledvfith goldenand pinkish-brovwn stones like those used on theface of themask. Theolive berries are predominantly stone: brovwns, olive green, and opaque white.Most of thegreen leaves employ only green glass. The oliveleaves, however,use both blue and blue-green glass and opaque white stone, and the oakleaves are white and light blue, vfith blue and blue-green glass highlights. Mostof the decoratedarea is set againsta darkbackground of blue- brownstone. At timesthis background serves as a trueshadow-as vfith the oliveleaves below the chin of the maskand the loweredge of the flutteringribbon. But in generalit seemsto functionmore as a darkback- dropagainst which the lighter glass and stone tesserae stand out. Many of the garland'sleaves and flowers (as well as the top of the mask)extend beyondthis dark ground.

THE BOUQUETFRAGMENT The borderzone and white backgroundof the bouquetfragment (8, Fig. 9) areidentical to thoseof 7. A clothbinding of deepgold vxrith a redhem holdsa bouquetthat extends to the right.The bouquet'sdark I38 ANDREW STEWART AND S. REBECCA MARTIN backgroundis also identical to thatof 7, andits floraare very similar, with twopomegranates, wild roses, and olive leaves. Novelties include one yel- low-greenfruit and a six-petaledflower of gold andorange with white highlightsand a redcenter. To theleft of the floweris a partialblue-glass leafwith white highlights and central veins. The pomegranatesare larger andslightly cruder than those of 7.This is mostevident in thelower of the two fruits,where the openingcrown is awkwardlyplaced and clumsily shaped.Clearly, at leasttwo craftsmen produced this mosaic. Extrapolatingfrom 7, a 16-cm-highsection of the fieldis missing. The fragmentshows a clearcontinuation of the tessellatedfloor beyond the frieze,with large white tesserae at its top laidperpendicularly to the bandedborder. There is no suggestionthat the meander(see below) was laidin thisdirection.

THE MEANDERFRAGMENTS Two largesections of the mosaic'sperspectival meander survive (9, 10; Fig.10). The white keys are 0.073-0.079 m high,and the width of a com- pletesequence can be estimatedat 0.14 m. The numberof tesseraeranges fromseven to nineper square centimeter. Each perspectival"box" created by the doublemeander contains a simple,geometrically rendered rosette; at leastfive more rosettes are preserved in othersmall fragments. A wide rangeof colorsis representedin the manysmall and two large fragments: gold,several blues, green, white, and red.23 Halfofthetesserae used in themeander are glass (light blue and green), andthe stones are white, two shades of red,and dark blue. The doublekey of thesmaller fragment (10) is formedby a singlerow of whitestones, and the illusionof depthis renderedin multiplecolors: the "outside"of one exteriorkey is darkred, while the "inside"is lightred. The centerof the keyon bothsides is lightblue, and the framed rosette is whitewith a blue center,set againstthe darkblue background. In the samefragment, the nextsequence shows a variationon thiscolor scheme, with dark and light blueglass for the exteriorof thekey, and shades of redfor its interior.The largerof the twoprincipal fragments (9) carriesthe samearrangement of colors,but herea morecomplete picture of the designcan be seen,par- ticularlythe reversalof colorsfrom exterior to interior.In this case,the dominantsequence is bluefor the exteriorand red for the interior.All of the rosettescontain the samecolors. The meanderzone is thusquite varied. It is difficultto determine how,if at all,this part of themosaic related to thegarland frieze, although theirborders hint at a certaincontinuity. Perhaps it framedthe mosaic's centralemblema. The workmanship,materials (particularly the heavyuse 23. The doubleperspective meander of mosaic25 in niche37 of the Agora of glass),and care in designindicate that this sectionwas madeby the of the Italianson Delos,in whichthe samecraftsmen who produced 7 and8. Both9 and10 havevertical seams distributionof colorsis irregular,sug- thatextend to the bottomof theirpreserved mortar bedding (0.133 and gestshow the disparatecolors of the 0.126 m thick,respectively). This pattern,a conventionalone, was most Dor meanderwould have worked certainlylaid in situ.24Three other geometrical fragments (not meanders) together(DelosXXIX, pp. 136-139, figs.29-31,pl.A:l). fromDor have these vertical seams. Possibly some or all of themwere laid 24. Forthe mosaic-layingtechnique, at the edgeof the pavementor encloseda putativecentral emblema. No see Ling 1998,pp. 14-15;Westgate seamsappear on the twolarge garland fragments (7, 8). 2000,pp. 272-273. HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I39

Figure12. Thmuis, mosaic with per- sonificationof QueenBerenike II(?) bySophilos. Late 3rd century. Alexandria,Graeco-Roman Museum,inv. 21739. Courtesy DeutschesArchaologisches Institut, Cairo

DISCUSSION Ongoingrestoration of the dozensof remainingmosaic fragments and furtherexcavation should yield a morecomplete picture of the composi- tion.In particular,a second mask is suggestedby a fragmentof aneyebrow andadjacent wavy strands of hair.The colorsused in thispiece are identi- calto thoseof 7, andthe hairis treatedsimilarly. It seemsto comefrom theproper left side of a now-missingmask. Since the eyebrow terminates closeto the edgeof the face,this mask was also slightly averted from the viewerand foreshortened, facing in the oppositedirection as the restored mask.Its thineyebrow indicates a differentcharacter type. Manyof the otherexcavated fragments may come from different ar- eas of the mosaicor fromanother floor entirely. They feature garlands, meanders,and other patterns that, in theircurrent state, cannot be cer- tainlyconnected with 7-10. Of particularinterest are several fragments thatshow two pomegranates together with red and blue fruits and flowers; grapesclustered among leaves and flowers; and two pinecones hanging by greenglass pine needles. The possibilityof a secondfloor is alsoraised by a decorativefragment that may represent part of Pan'sthrowing-stick, or lagobolon,and with a borderconsisting of onlytwo rows of redtesserae. TheDor mosaics strongly recall the technique, palette, and verisimili- 25. Graeco-RomanMuseum, tudeof Hellenisticworks such as Sophilos'slate-3rd-century personifi- inv.21739: Daszewski 1985, pp. 142- 158,pls. A, 32, 42a;Grimm 1998, cationfrom Thmuis now in the Graeco-RomanMuseum at Alexandria pp.79-81, fig. 81:a,c; Dunbabin1999, (Fig.12).25 Sometimes identified as a portraitof BerenikeII (reigned246- pp.24-26, fig. 25, pl. 4. 221),this figure also displays a fullface, animated countenance, and subtle I40 ANDREW STEWART AND S. REBECCA MARTIN

chiaroscurorendered in a virtuosoopus vermiculatum. The mid-2ndcen- Figure13. Pergamon, garland turygarland of theHephaistion mosaic from Palace V at Pergamon(now mosaicfrom Palace V. Mid-2nd century.Berlin, Pergamonmuseum, in Berlin)is set againsta similardark ground and contains a varietyof inv.70. AfterAvPV.1, pl. 18 scrollingvines from which fruits and flowers spring (Fig. 13). It alsocon- tainsinsects and playfill erotes but remains more delicate and less lush and vibrantthan the mosaicsfrom Dor.26 The opustessellatum and vermiculatummask-and-garland mosaics fromDelos may provide parallels for the entirecomposition. The closest comefrom the mosaic borders of thereception/dining rooms (andrones or oeci)of someof the2nd-century houses. For example, on friezeM of mo- saic68, a (damaged)garland links bull-heads to theatricalmasks.27 Origi- 26. Berlin,Pergamonmuseum, nally,there were three masks on eachlong side, two on eachshort side, and inv.70: AvP V.1,pp.53-61, pls.17-19, oneangled bull-head at each corner.The eight preserved masks are spaced figs.27-38;Kriseleit 2000, pp.17-23, 1.10-2.40m apartand vary in size,but aregenerally shorter and much figs.8-15. narrowerthan the mask on fragment7 fromDor; the young man (mask I) 27. DelosXXIX, pp.156-169, measures0.30 (H.) x 0.20 (W.) m.28The friezeis 0.35 m highand en- figs.55-79, pl. A:3-4. 28. DelosXLJX, pp.160-163, closesseveral concentric geometric patterns and a verydamaged central fig. 70, pl. A:3. emblemawith a sceneof Athena,Hermes, and an unidentified central fig- 29. C£ DelosXXJX, pp.245-251, ure.The mosaicis framedat the top andbottom by a thickblack trim no.215, figs.184-195;Siebert 1971; 0.06-0.09m wide(with seven to tenrows of tesserae).29 alsoAvPV.l,pp.53-74, figs.67-74, Thesegarland-and-mask mosaics are the visual equivalent of theflo- pls.5,12-15,26-39; Radt1999, ral"garlands" (stephanoi) of sympoticpoetry "woven" for the Muses by the pp.63-78, figs.18-22 (Pergamon: PalacesIV andV); andKonstantino- Hellenisticepigrammatists. They are the successorsto the garlandsthat poulos1986, pp.148-150, pl.27; commonlyembellish Classical and Early Hellenistic symposion-kraters, at Papachristodoulou1993, p.37, pl.22 least two of whichhave been found at Dor.By garlandingthe roomlike (Rhodes). =: :

HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I4I

|

I_81u-- I_w_I I_w - - N I NwB,N'wNwN'w l w z l z NNw-MMNw - l I p 1 Figure14. Pompeii,mask-and- garlandmosaic from thefauces of the banquetersthemselves, the mosaicsdefine it as a hospitablespace the House ofthe Faun.Late 2nd markedby Dionysiacenthousiasmos and the alteredstate of consciousness century.Museo Nazionaledi Napoli, it creates.30 inv.9994. Aftera l9th-centurywatercolor Furthercomparanda come from the Houseof the Faunat Pompeii, by G. Marsigli;Norman Neuerberg Archive, whosemosaics date to thelate 2nd century.31 The mask-and-garland frieze Universityof Californiaat Berkeley thatborders the opussectile pavement in the house'sfauces(Fig. 14)32 dis- playsmany similarities to theDor mosaic. The fieldmeasures 0.49 x 2.81 m andcontains two symmetricallyplaced, female theatrical masks facing awayfrom each other, connected by a longgarland. The garlandis bound togetherwith a spiralingribbon that is yellowon one sideand red and whiteon the other.It encirclesthe garland three times and is tiedin two largebows, one at either end.The garland contains a rangeof florasimilar to thatin the Dor frieze: pomegranates invarious stages of ripening,quinces, 30. SeeAnth.Pal. 4.1-2 (Meleager pinecones,a shootof grain,an acorn, ivy, and olives. These cover and over- andPhilip) with Gutzwiller1998, pp.276-291; cf., e.g.,Lissarrague lapthe edgesof a rectilineardark background set into a whitefield. The l990a, p. 197,pls. 17, 19-22;l990b, shadingis moredramatic than in fragment7 from Dor, but the light source pp.26-29; Stern2000, color pl. 2:1 is notconsistent. Room 34 of thehouse (probably its triclinium) was deco- (Dorkraters). ratedwith the so-calledTiger-Rider Mosaic framed by a mask-and-gar- 31. Baldassare1994, pp. 94-96, landfrieze.33 This friezeshares with the Dor mosaicthe varietyand op- figs.12-14. 32. MuseoNazionale di Napoli, posingorientation of its masks;the multicoloredribbon that binds its inv.9994. garlands;and its darkbackdrop.34 33. MuseoNazionale di Napoli, Althoughthese parallels are suggestive, the extant fragments from Dor inv.9991: Baldassare 1994, pp. 104- do not asyet entitleus to reconstructthe programof the mosaicfloor or 105,fig. 28; Dunbabin 1999, pp. 43-44, floorsin anydetail. It is likely,however, to haveresembled the Delian- fig. 43. Pergameneandron/oecus type rather than the westernfauces type. The tech- 34. Its pervasivedark background hereis akinto easternHellenistic nique,quality, and parallels (stylistic and iconographic) of the fragments mosaics,like those from Pergamon: allowus to datethe mosaicat Dorwith some confidence to the mid-late Dunbabin1999, p. 44. 2ndcentury. I42 ANDREW STEWART AND S. REBECCA MARTIN

CODA

Althoughthe Nikefound in recentexcavations at Tel Dorwas probably carvedby a localcraftsman or perhapsa Cypriot,the mosaicist(s)must havebeen trained in one of the maincenters of Hellenisticart, such as Pergamonoperhaps morelikely-Alexandria. Yet the mosaicitself is almostcertainly not an import.Both the shellymortar used for the thin beddingfor the tesserae and the coarser mortar below look very much like the same,local material, although laboratory analysis of the supportis necessaryto confirmthis observation.The makersof the mosaicsurely traveledto Dorto executeit. There is goodevidence that Hellenistic crafts- menof allkinds traveled widely and there is someevidence that mosaicists themselvesdid so. Thus [Askle]piades from Arados in Syriasigned a mo- saicon Delosand Dionysios from Alexandria signed another at Segestain Sicily.3sThe lattersuggests an awarenessof anddesire for work of a par- ticularstyle and quality asdoes the newlydiscovered mosaic at Dor. The eagernessof Dor'sinhabitants to identi withthings Hellenic is evincedin theirfoundation myth.36 The name"Doros" or "Dora"for the sitecame from a Hellenizationofthe Semitic"Dor."37 This name is tiedto a mythicalgenealogy of the city,in whichDoros, son of Poseidon,was its eponymousfounder.38 Although the groupidentity to whichthis geneal- ogybears witness need not be legible as such in thearchaeological record,39 the newlydiscovered architectural fragments, Nike, and mosaic suggest thatat leastsome residents of Dor consideredthemselves Hellenized or evenethnically Greek. Herean immediate question arises over the unusual popularity of the Doricorder at Dor. In additionto thefragments published here and those attributedto the propylonof TempleH, the Roman"bouleuterion/syna- gogue"in areaB wasDoric, and many stray Doric capitals have been re- coveredin areasF and H. Did the inhabitantsof Graeco-RomanDor perhapsfeel that the Doricorder was somehow uniquely theirs? For if a Romanscholar could blandly conflate Doros son of Poseidonwith Doros son of Hellen,renowned ancestor and eponymous hero of the Dorians andthe Doricorder's inventor, afortiori so, surely,could they.40 The mosaicat Dor is an indexof an individual'sHellenization that mayowe its genesisto andperhaps in turneven reinforced the city's

35. Westgate 2000, p. 273. form(/Q) on silvertetradrachms of pp. 16-19, 91-95; Stern1995, I:A, p. 2; 36. For discussionsof Hellenic and PtolemyV mintedin the cityin 205, 2000,p. 201. Hellenistic identity,see Hall 1997, andin fullin 139/138on Tryphon's 39. SeeWaldbaum 1997; Hall 1997; 2002; Morris 2000. leadsling bullet; from 64/63 it occurs cf. Stern1989; Berlin 1997. 37. "Dor":see, e.g., Joshua17:1 1; regularlyon the city'sRoman-period 40. Serv.ad Aen. 2.27:sane Dorus Kings4:11; "Dora/Doros":see, e.g., coins.See Dahl 1915,pp. 16-20, 62- Neptunif liusfuit, undeDori originem Polyb.5.66.1; Joseph.AJ 13.223-224; 63, 94-95; Stern1995, I:B, p. 469, ducunt;cf. Vitr. 4.1.3 forDoros son of BJ1.50; cf. Dahl 1915, pp. 16-20. The nos.90-94 (coins),pp. 491-496 Hellenand the Doricorder. On the Hellenized versionwas apparentlycur- (bullet);2000, pp. 211-213, 264-267, originsand development of Dorian rent as early as ca. 500 (Hekataios of figs.142, 182. self-consciousness,see Hall2002, Miletos ap. Steph. Byz., s.v. Aupog; 38. ClaudiusIolaus ap. Steph.Byz pp. 82-89. FGrHist 1 F275); it occursin shortened s.v.Aupog; FHG p. 363;Dahl 1915, HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I43

self-styledHellenic filiation. This stunningcomposition reaffirms that patronsliving outside the majorcenters of artwere at timessophisticated andresourceful enough to turnto thesecenters to satisfytheir tastes. To- getherwith the architectural fragments described above, the mosaic offers a glimpseof theways in whichself- and group identities could be subjec- tivelyformed, not onlythrough myth, but also through material culture.

ACKNOWLED GMENTS

Althoughthis articleis a jointventure, Andrew Stewart is primarilyre- sponsiblefor the sectionson architectureand sculpture, and S. Rebecca Martin,who excavated the mosaic, for the sections relating to it.We thank EphraimStern of the Instituteof Archaeologyat HebrewUniversity of Jerusalem,director of theTel Dor excavations from 1980 to 2000,for per- missionto publishthis material; his formerdeputy and the currentdirec- tor,Ilan Sharon, for his enthusiastic assistance with the project both in the fieldand out of it;and Bracha Guz-Zilberstein for analyzing the ceramic evidenceand, along with the othermuseum staff at the "Glasshouse"at Nachsholim,for offering unstinting help and encouragement. The Universityof Californiaat Berkeleyexcavation staff in areaD1 wereAllen Estes (assistant director); Catherine McGowan (area supervi- sor);Bruce Redwine (recorder); and Sarah Kamp, Rebecca Martin,Jessica Nager,Annie Smiley, Sarah Stroup, and Martin Wells (unit supervisors). Wells and Kampexcavated the Nike and the architecturalfragments. AmandaAdams, John Berg, Erin Dintino, and Svetlana Matskevich cre- atedthe plansand architectural drawings. Tali Goldman and Matskevich assistedwith research in Jerusalem;in Athens,Camilla MacKay and the staffof the BlegenLibrary at the AmericanSchool of ClassicalStudies greatlyfacilitated our work, and Dorothy King kindly shared her expertise in figuralakroteria and tombs;in Berkeley,John Ceballosand David Sullivanof theU.C. Berkeley library kindly located and specially ordered keyitems for us. The mosaicwas expertly conserved and mounted by Orna Cohen and photographedby GabiLaron; all fieldphotographs were taken by Israel Hirschberg.Ezra Marcus of the HebrewUniversity of Jerusalemidenti- fiedthe stone used for the Nike and architectural fragments. Graeme Clarke of theAustralian National University at Canberra alerted us to comparanda forthe architectureand unstintingly furnished information, plans, and el- evationsof his newdiscoveries at Tell Jebel Khalid. Susan Downey of the Universityof California,Los Angeles,kindly communicated her current viewson the ZeusMegistos sanctuary at Dura.For generous guidance on the mosaic'stechnical, stylistic, and iconographicproblems, we thank KatherineDunbabin of McMasterUniversity; Richard Green of theUni- versityof Sydney;Ze'ev Weiss of theHebrew University of Jerusalem; and RuthWestgate of CardiffUniversity. Finally, we aregrateful to the three anonymousreviewers for Hesperia, who savedus fromseveral errors and misstatements. . , . .

I44 ANDREW STEWART AND S. REBECCA MARTIN

REFERENCES

Abbe,E. 1965. ThePlan* of Virgil's Daszewski,W. 1985.Corpus of Mosaics Guinzier-Sorbets,A.-M., andM.-D. Georgics,Ithaca. fromEgypt 1: Hellenistic and Early Nenna.1992. aL'emploi du verre, Alscher,L. 1957. GriechischePlastik 4, RomanPeriod, Mainz. de la faience,et de la peinturedans Berlin. De'losV = F. Courby,Le Portique d Anti- les mosaiquesde Delos,"BCH116, Avigad,N. 1954. AncientMonuments in goneou du nord-est et lesconstructions pp.607431. theKidronValley,Jerusalem. voisines(Delos V), Paris1912. Gulaki,A. 1981.Klassische und klassi- AvP II = R. Bohn,Das Heiligtum der DelosXl = A. Plassart,Les sanctuaires et zistischeNikedarstellungen,Bonn. AthenaPolias Nikephoros (AvP II), lescultes du Mont Cynthe (Delos X]), Gutzwiller,K.J. 1998. Poetic Garlands: Berlin1885. Paris1928. HellenisticEpigrams in Context, AvP V.1 = G. Kawerauand T. Wie- DelosXXIX = P.Bruneau, Les mosaiques Berkeley. gand,Die Palasteder Hochburg (DelosXXE), Paris1972. Hall,J.M. 1997.EthnicIdentity in (AvPV.1),Berlin 1930. Downey,S. B. 1988.Mesopotamian GreekAntiquity,Cambridge. AvPX].2 = O. Ziegenausand G. de ReligiousArchitecture: Alexander . 2002.Hellenicity: Betfween Luca,DasAsklepieion (AvPX].2), throughthe Parthians, Princeton. Ethnicityand Culture, Chicago. Berlin1975. . 2003.MesopotamianReli- Harrison,R. 1962.aPlants," in The Baldassare,I. 1994.Pompei: Pitture e giousArchitecture after the Greek NewBible Dictionary, J. Douglas, mosaiciV: Regio KI, pt. 2, Rome. Conquest:New Evidenceand New ed., GrandRapids, pp. 1005- Bar-Kochva,B. 1996.Pseudo-Hecataeus, Ideas,"ArchaeologicalInstitute of 1()06. On theJews: Legitimizing the Jetwish America,104thAnnualMeeting: Herbert,S. C. 1994.TelAnafa I.1-2: Diaspora,Berkeley. Abstracts26, pp.119-120. FinalReport on Ten Years of Exca- Berlin,A. 1997. "BetweenLarge Dunand,M., andR. Duru.1962. vationaf a Hellenistisand Roman Forces:Palestine in the Hellenistic OummElAmed{, Paris. Settlementin NorthernIsrael (JRA Period"BiblArch 60, pp.3-51. Dunbabin,K. 1979 "Techniqueand Suppl.10.1), Ann Arbor. Bliss,F., and R. Macalister.1902. Exca- Materialsof HellenisticMosaics," Joyce,H. 1979.aForm, Function, and vationsin Palestineduring the Years AJA83, pp.265-277. Techniquein the Pavementsof 1898-1900,London. . 1999.Mosaics of the Greek and Delos andPompeii,"AJA 83, Clarke,G. W.2001a.aThe Governor's RomanWorld, Cambridge. pp.253-263. Palace,Acropolis,Jebel Khalid," DuraII = P.V. C. Baurand M. I. Ros- King,D. L. V. 2000.aThe Sculptural in TheRoyal Palace Institution in tovtzeS,eds., The Excavations at Decorationof the DoricOrder, theFirst Millennium B.C.: Regional DuraEuropos: Preliminary Reports ca.375-31 B.C." (diss.King's Col- Developmentand Cultural Inter- II, New Haven1931. lege,London). changebetween East and West Ermeti,A. L. 1981.LAgora di Cirene Konstantinopoulos,G. 1986.Aeoxacoa (Monographsof the DanishInsti- III.1:Il monumentonavale, Rome. Podos,Athens. tuteat Athens4), I. Nielsen,ed., Fedak,J.l990. MonumentalTombs of Kriseleit,I. 2000.Antike Mosaiken, Aarhus,pp.215-247. theHellenistic Era:A Study of Selected Mainz. . 2001b.The Governor'sPal- Tombspromthe Pre-Classical to the Ling,R. 1998.Ancient Mosaics, Lon- ace,Akropolis," in Jebel Khalid on EarlyImperial Era, Toronto. don. theEuphrates: Report on Excavations Fisher,C. S. 1904."The Mycenaean Lissarrague,F. 1990a.UArolmdthe 1986-1996(MeditArch Suppl.5), Palaceat Nippur,"AJA8, pp. 403- Krater,"in Sympotica:ASymposium Sydney,pp.25-48. 432. onthe Symposion, O. Murray,ed., . Forthcoming.JebelKhalid, Garstang,J.1924.Tanturah (Dora)," Odord,pp. 19S208. 2002 Season,"MeditArch 16. Bulletinof the British School of Ar- . 1990b.TheAesthetics ofthe Clarke,G. W., et al.2000.aJebel chaeologyinJerusalem 6, pp. 65-75. GreekBanquet: Images of Wineand Khalidon the Euphrates," Gawlikowski,M. 1996. aThapsacus Ritual,A. Szegedy-Maszak,trans., MeditArch13, pp. 123-148. andZeugma: The Crossingof the Princeton. Coulton,J. J. 1976.TheArchitectural Euphratesin Antiquity,"Iraq 58, Marcade,J.1951.Les sculpturesdeco- Developmentofthe Greek Stoa, pp. 123-133. rativesdu Monumentdes Taureaux Oxford. Grimm,G. 1998. Alexandria:Die erste a Delos,"BCH 75, pp.55-89. . 1979.aDoric Capitals: A Pro- Konigsstadtder hellenistischen Welt, Mollard-Besques,S. 1963.Catalogue portionalAnalysis," BS 74, pp. 81- Mainz. raisonnedesfigurines et reliefsen 153. Grote,U. 1992. aBauplastikaus Per- terre-cuitegresques,etrusques, et Dahl,G. 1915.aThe Materials for the gamon,"in Mousikosaner: Festschrift romains2: Myrina,Paris. Historyof Dor,"Transactions ofthe fiir Max Wegnerzum 90. Geburtstag, . 1972.Catalogue raisonne des ConnecticutAcademyofArts and O. Brehmand S. Klie,eds., Bonn, Jtgurtneset reZzeysenterre-cutte Sciences20,pp. 1-131. pp. 179-187. gresques,etrusques, et romains 3: HELLENISTIC DISCOVERIES AT TEL DOR, ISRAEL I45

Epoaqueshellenistique et romaine, Roux,G. 1961.L'architecture de IArgo- Talbert,R. A., ed.2000. TheBarrington GreveetAsieMineure,Paris. lide aux IZe etIIIe sieclesavantJ.-C. Atlasof the Greek and Roman World, Morris,I. 2000.Archaeology as Cultural (BEFAR199), Paris. Princeton. History:Words and Things in IronAge Schober,A. 1951.Die Kunstvon Perga- Waldbaum,J.C. 1997.Greeksin the Greece,Malden, Mass. mon,Innsbruck. Eastor Greeksandthe East? Prob- Netzer,E. 2001.Hasmonean and Hero- Shoe,L.T.1936.ProflesofGreekMold- lemsin the Definitionand Recog- dianPalaces at Jericho. Final Reports ings, Cambridge,Mass. nitionof Presence,"BASOR 305, ofthe 1973-1987Excavations 1: Siebert,G. 1971.aSur le mosaiquede pp.l-18. StratigraphyandArchitecture, Jeru- l'habitationV de l'£lotdes bijoux a Webb,P. 1996.HellenisticArchitectural silem. Delos,"BCH95, pp. 147-165. Sculpture:Figural Motifs in Western Nielsen,I. 1994.Hellenistic Palaces: Slane,K. 1997.TelAnafa 2.1: The Anatoliaand theAegean Islands, Traditionand Renemsal, Hellenisticand RomanFine Wares Madison. Aarhus. ( JRA Suppl.10.2), Ann Arbor. Webster,T. B. L. 1961.Monuments Pakkanen,J.1998. The Temple of Stern,E. 1989.aThe Beginning of IllustratingNemu Comedy (BICS AthenaAleaat Tegea:AReconstruc- GreekSetdement in Palestinein Suppl.11), London. tionof the Peristyle Column, Hel- the Lightof the Excavationsat Tel . 1969.Monuments Illustrating

. , . slnt. Dor,"in RecentExcavations in Israel: NemuComedy (BICS Suppl. 24), Papachristodoulou,I. 1993. Aeoxac'a Studiesin IronAgeArchaeology 2nd ed., London. Podo: 2,400xpd;vea, Athens. (AASOR49), S. Gitinand W. G. Webster,T.B. L.,J. R. Green,and Ploug,G. 1985.Hama 3.1: TheGraeco- Dever,eds., Winona Lake, Ind., A. Seeberg.1995. Monuments RomanTomun, Copenhagen. pp. 107-124. IllustratingNemu Comedy (BICS Poulsen,B. 1997.The Sculpturefrom . 1995.Excavations atDor: Final Suppl.50.1-2), 3rded., London. a LateRoman Villa in Halicarnas- ReportI:A-B (QedemReports 1-2), Westgate,R. 2000.aPavimenta atque sus,"in Sculptorsand Sculpture of Jerusalem. emblematavermiculata: Regional Cariaand the Dodecanese, I. Jenkins . 2000.Dor: Rulerof the Seas, StyIesin HeHenisticMosaics and andG. B. Waywell,eds., London, 2nd ed.,Jerusalem. the FirstMosaics at Pompeii," pp.74-83. Stucchi,S. 1987.aL'architettura funera- AJA104, pp. 255-275. Purgold,K. 1881.Nike ausMegara," riasuburbana cirenaica in rapporto Wiegand,T.,and H. Schrader.1904. AM 6, pp.257-282. a quelladella chora vicinore ed a Priene,Berlin. Radt,W. 1999.Pergamon: Geschichte quellalibya ulteriore, con speciale Yalouris,N. 1967.aTa axporripLa Toi undBauten einer antiken Metropole, riguardoall'eta ellenistica," QAL 12, vaoi)rrlS 'Agr£lltAot,"ArchDelt 22 Darmstadt. pp.249-377. A', pp.25-37.

AndrewSfewart UNIVERSITYOF CALIFORNIAAT BERKELEY DEPARTMENTOF HISTORYOF ART BERKELEY,CALIFORNIA 94720-6020 astewart@socrates .berkeley.edu

S. RebeccaMartin UNIVERSITYOF CALIFORNIAAT BERKELEY DEPARTMENTOF HISTORYOF ART BERKELEY,CALIFORNIA 94720-6020 [email protected]. edu