<<

HESPERIA 70 (200I) ICONOG RAPHY Pages46z-492 AND THE DYNAMIGS O F PATRO NAG E A SARCO P HAGU S FROM T H E FAMILYOF HERODESATTICUS

ABSTRACT

A sarcophagusfiom the estateof HerodesAtticus in Kephisiacommemo- ratesthe intimate connections ofthe familywiththe cityof Sparta,the Battle of Marathon,and the cultstatue of Nemesisat .Theiconographic allusionsto Marathonalso reflect the prioritiesof the SecondSophistic, an intellectualmovement that appealed to the pastto establishcultural and po- liticalsuperiority. The unusualand meaningfill decorative program suggests thatthe family commissioned this sarcophagus. The earlierview that the more unusualAttic sarcophagiwere prefabricated, but that theirthemes simply provedunpopular, should be modifiedin lightof this study.

INTRODUCTION

In Septemberof 1866,during the construction of a housein the Kephisia suburbof ,workers discovered a marbleburial chamber, roughly squarein plan.1Robbers had long since plundered the chamber, removing thedeceased and most ofthe portable possessions. In 1866,the significant remainingartifacts included four carved marble sarcophagi and only a handfillof smallobjects.2 Otto Benndorf,who wrotethe firstcomplete descriptionof the chamberand its contents,was also the firstto suggest

1. I thankthe Universityof Mich- andNeel Smithfor discussing with me this study,still stand in the tomb iganfor support that allowed me to someof the ideaspresented here. Photo- today.Cramped space in the tomb undertakepreliminary research in graphswere kindly provided byJan San- makesit difficultto providea complete Athensfor this article;and the College ders;the BritishMuseum; the Deutsches photographicrecord. I thereforerefer of the Holy Crossboth for funds to ArchaologischesInstitut, ; the in manyinstances to the line drawings purchasephotographs and for a leaveof GreekArchaeological Service; and the of the Ledasarcophagus produced by absencethat allowed me to continue KunsthistorischesMuseum, . Robert(1890, pl. III andp. 9), which mywork. I amgrateful to Elaine Permissionto reproducedrawings was havebeen used in mostsubsequent Gazda,Mark Landon, Kenneth La- generouslygranted by MarinaBelozer- scholarship(Figs. 1 and5 here,respec- patin,and Miranda Marvin for their skaya(Fig. 15) andthe DeutschesArchao- tively).Benndorf (1868, p. 40) mentions commentson earlydrafts of this article logischesInstitut, Athens (Fig. 16). thata greenglass vessel (Gefass) without andfor their graciously offered insights 2. All foursarcophagi, including the handlesand a possiblebronze mirror andcriticisms; and to ThomasMartin so-calledLeda sarcophagus considered in werealso found in the tomb.

American School of Classical Studies at Athens is collaborating with JSTOR to digitize, preserve, and extend access to Hesperia ® www.jstor.org X, , , X, \ -N ,>

462 ELLEN E. PERRY thatthe structurebelonged to the familyof the famous2nd- A.C. Figure1 (opposite).Drawing of the -millionaire,Herodes Atticus. At the timeof Benndorf'swriting, Leda sarcophagusfrom Kephisia. theevidence to supportthis attribution was meager. and Aulus Front(top), left and right sides Gelliusboth indicate that Herodes owned a villain Kephisia(Philostr. VS (center),back (bottom). Robert 1890, 2.562;Aul. Gell. 1.2.1-2), but these testimonia and a fewadditional arti- pl. III factsonly placed Herodes in thegeneral neighborhood. Since then, how- ever,a numberof archaeologicaldiscoveries, including portraits and in- scriptions,have confirmed not onlyHerodes' presence in Kephisiabut, morespecifically, his ownership of landin theimmediate neighborhood of the tomb.3 An inscriptionfound built into the nearby Church of HagiaParaskevi identifiesthis tombwith the family.It recordsthe recentloss of an un- namedinfant child of Herodes.The textof the inscriptionindicates that otherchildren of his hadalready died:

'Hto8Ns, o' v8r xotuNv,ov wavTa rvtavTov ovr xo,uNvutorv vas ovr or waoa (ptov ,uNvt\ oXx , xroas,, vsof \ xrvrcet, n xaToa, yas,, 'Htocl)8Ng8ruceas axtoa xos Aaxtoucet qZ- n rTvZovXO waocl)v o N v )vXag otcetv,o Cl)C,c, wor cecl)a- 0rXaU^, an rvn Urlxatc,a, v,urxrtootof , waooc,. ,

Herodesset in the depthsof the earththis his lock of hair, Havingdampened the tipsof the hairwith his tears, Whenfor less than the cycleof a He hadneither grown his hairnor reared you, dear son, Forhe cutthis lock in the thirdmonth. Mayit be a truetoken to youthree children's souls Thatyou will someday receive among the coffinsthe body of your father.4

JenniferTobin notes that the block on whichthis inscription was found is 3. Tobin(1997, pp.211-239, fig. 42) similarin thicknessand treatment to the blocksfrom the dromosof the hasthoroughly documented the arch- Kephisiatomb.5 This observation supports the identification of the struc- aeologicalevidence for Herodes' presence tureas the restingplace for several of Herodes'children. in Kephisia. 4. SEG XXVI290 (= Ameling1983, Thefour sarcophagi in theKephisia tomb date to theAntonine period II, pp. 143-146,no. 140).See Tobin andare therefore consistent with Herodes' dates.6 One sarcophagus has no 1997,p. 225,where this translationis figuraldecoration, while commonsepulchral motifs adorn two others: creditedto M. B. Richardson. garlandson one,and Erotes on the other.The fourthbox, known as the 5. Tobin1997, p.225. Ledasarcophagus, depicts a themethat is unusualon sarcophagi, the family 6. Benndorf1868, p. 40;Tschira 1948-1949,col. 86. of HelenofTroy (Fig. l).This particularcoffin presents a rareopportunity 7. Kochand Sichtermann 1982, to examinethe dynamicsof artpatronage in the ,precisely pp.46S461, whereKoch also cites becauseit hasbeen convincingly attributed to a"private" (i.e., non-imperial) (ns.28-30,32) otherexamples of anoma- butfamous family. louspieces from this period: a Dionysiac GuntramKoch suggests that this sarcophagus, like many, was prefab- sarcophagusnear the Hephaisteionin ricated.Because the theme is unusual,however, he alsoconjectures that its Athens;one with centaurs in the National Museum(NM 1184);one with Erotes decorativeprogram was a marketfailure, and that this and other anoma- in Thessaloniki(Thessaloniki Museum lousthemes were quickly eliminated from the filneraryrepertoire when 1248);and two cineraryurns, one in theyproved unpopular.7 The iconographyof thiscoffin, however, alludes Athensand one in Patras. =

r WL

I I

l 464 ELLEN E. PERRY

Figure2. Right side of sarcophagus, showingdetail of Eros.The gabled lid does not belong to the sarcoph- agus.Photo author, reproduced with permissionof the GreekArchaeological zervlce

to specificconnections that the familyof Herodesenjoyed both with the cityof Spartaand with the deme of Marathon.In thiscase, the sarcopha- gus was clearlynot prefabricated,but commissioned,and its decorative programwas part of a self-consciousmythmaking that celebrated family identity.8The decoration may, more specifically, commemorate the coffinXs occupants HerodesXdaughter, Elpinike, whose very name is an allusion to Marathon;and perhaps her husband, L. VibulliusHipparchus, about whomwe knowlittle, but who is likelyto havebeen a kinsmanof Herodes.

THE LEDA SARCOPHAGUS

The Ledasarcophagus is of thekline type, with its originallid in theform of one or,more probably, two figuresreclining on a couch.The pitched- rooflid thatrests on thebox at present (Fig. 2) doesnot, therefore, belong 8. SimilarlyXEwald (1999> esp. to it, butto one of the othersarcophagi in the tomb.That this is a kline pp.79> 129-130)> in a studyof intel- lectualactivities as representedon sarcophagusis clear from a long,narrow rectangle set into the rightside sarcophagiXsuggests that unusual of theupper molding on thefront (Fig. 1, upperright; Fig. 3, upperleft). iconographiesreflect the particular A correspondinginset probably appeared on theleft sideof the molding, interestsof the deceased. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 465

Figure3. Caryatidon right front cornerof sarcophagusand detailof inset rectangle(upper left). Photo author,reproduced with permission of the GreekArchaeological Service

but damageto the stonein this areahas destroyedany trace of it. Such rectangularinsets often decorated the frames of ancientcouches, and thus it is not surprisingto findthem represented on klinesarcophagi and kline monuments(e.g., Fig. 4).9 These insets appear especially frequently on the earliestAttic klinesarcophagi, supporting an earlydate for the example fromKephisia.l° On the left end of the boxin question,Leda is depictedstruggling with a swan(Fig. 1). She is nude,in profile,and faces left. Shewears a "melonhairstyle" (melonenfrisur), and the bun in backsits high on the crownof herhead. She clutches a clothwith her right hand, apparently in an attemptto coverherself. She bends slightly at the knees,perhaps with 9. Robert1890, p. 9;Wrede 1977, the effortof strugglingagainst the swan,and raises her left heel off the p. 428. Forexamples of suchinsets on ground.Her rightleg is not visible;we arepresumably to thinkof it as bronzecouch frames, see Richter1966, obscuredby the left leg.With one armLeda holds the swanoff, but his p. 106,as well as fig. 530 (a couchfrom sinuousneck writhes toward her mouth in anattempted kiss. The swanis Boscoreale,now in the Staatliche enormous,with a torsoslightly larger than Leda's own. He is shownin Museen,, inv. 890J) and fig. 532 (a couchfrom Pompeii, now in the midair,with his wings fillly outspread, the left wing extending behind Leda NationalMuseum, Naples, inv. 78614). asif to embraceher. He hasthe talons of a birdof preyrather than webbed 10.Wiegartz 1975, p. 188. feet, and with them he clutchesthe cloth.These talons,presumably 466 ELLEN E. PERRY

Figure4. Klinemonument showing inset rectangles(below mattress at thoseof an eagle,serve as a reminderthat this is no ordinaryswan, but head and foot). Rome,Museo Zeushimself. Nazionaledelle Terme, NM 72879. On theright side of thesarcophagus, Eros strings his bow (Figs. 1-2). CourtesyDeutsches Archaologisches He is clearlya thematiccompanion to Ledaand the swan.llHis bowand Institut,Rome, neg. 80.2707 wings,the outlines of whichare still apparent, were removed with a chisel orsome similar implement, possibly because they were gilded or otherwise completedin metal.The tombrobbers, at anyrate, appear to havere- movedmetal attachments from the sarcophagusand damaged it in other ways 12 On the frontof the box standthe Dioskouroi,flanking their sister, Helen.She andher brother Polydeukes were, of course,the offspringof Leda'sunion with and so, taken together, the frontand sides form a coherentprogram related to theroyal house of Sparta.The Dioskouroi are nude,except for the chlamysthat each wears pinned at the shoulder.The left armof theleft figurehangs down, grasping the hiltof a shortsword, whilehe raiseshis rightarm, which is bentat the elbow.A hollowin his righthand indicates that he onceheld a metalspear. He restshis weight on his rightleg andhis headturns to the properleft, that is, towardthe 11. Indeed,several known versions centerof thecomposition, where Helen stands. His twinon theright side of Ledastruggling with the swandepict of the compositionis almosta mirrorimage. The onlydivergence from Eroseither encouraging or physically helpingthe swan:Wiegartz 1983, strict,bilateral symmetry is thefact that both Dioskouroi wear the chlamys pp. 173-174. pinnedat theright shoulder. Helen wears a sleevedchiton and a himation 12. Forremoval of the spearsfrom looselydraped around her shoulders, crossing in frontand then wrapped the handsof the Dioskouroi,see below. ICONOGRAPHY AND THE DYNAMlCS OF PATRONAGE 467

aroundher left arm. In herleft hand she holds a fruit,possibly an apple or a pomegranate.Her posture is generallyfrontal but her head inclines very slightlyto theleft. The frontand sides of thissarcophagus have been interpreted as de- pictionsin reliefofindividualClassical sculptures in theround, or as having beenstrongly influenced by suchsculptures.13 This impression derives, in part,from the formal character of thedecoration. There is no indicationof architecturalorlandscape setting on the frontor sides;the background is blank.The individualfigures line up on a frontalplane and stand isolated fromone another.They do not makeeye contactor acknowledgeone another'spresence in anyway. Iconographic analysis of the antecedents forthese figures (see below, pp. 471-483) demonstrates that most do not, in fact,reproduce sculptures in the round,although they certainly appear to do so to moderneyes. Sketchy,roughly carved figures of a Tritonand Nereid decorate the back(Fig.l).The Nereidrides on the tailoftheTriton, and the two together occupymost of therepresentational field. As a result,these figures do not havethe statuesquequality of thoseon the otherthree sides. The Triton, whosesnaky tail spreadsout overalmost the entirelength of the relief, holdsa skyphosin his left handand an oarin his right.The seminude Nereidgrasps a garmentthat billows over her head in an arc.The entire compositionis surroundedby a simple,rectangular molding that resembles a modernpicture frame. This is anodd scene, which seems to havelittle to do thematicallywith the rest of thedecorative program. It is not,however, uncommonto findapparently incongruous themes on thebacks of Attic sarcophagi.l4 Marinethiasoi often appear on the mattressesof Attickline lids, and probablyallude to the afterlife.15Nereids, in particular,served in ancient artas escorts through the liminal phases of life,including the passage from thisworld to the next.16Nevertheless, the locationof thismarine thiasos, on thebody of the coffinitself, is unusualin theAttic repertoire. Caryatidsstand at the fourcorners of the sarcophagus(Figs. 1-3). Theirgarments, particularly the stacked folds of thehimatia, are standard on archaizingsculpture of the Romanperiod.17 These figures stand with theirfeet together,one handhanging at the side and the otherresting acrossthe breast.(Two of the caryatidshave their right arms bent while theother two bend their left arms.) Their hair is partedin themiddle and corkscrewlocks fall to theirshoulders. Their kalathoi touch the upper moldingand emphasizethe figures'filaction as architectontcsupports. The soclesor statue bases that project at eachof thefour corners act in the sameway.

13. Benndorf1868, p. 38: "zum sarcophagusas "strongly influenced by 15. Formarine thiasoi on the lids of Theil nochnachweisbare Repro- freestandingsculpture." klinesarcophagi, see Kochand Sichter- ductionenvon einzelnenStatuen sind"; 14. Kochand Sichtermann 1982, mann1982, p. 422. Robert1890, p. 10: "sollenwohl gerade pp.376-377. Brilliant(1984, p. 126), 16. Barringer(1995) analyzesthis alsCopien beruhmter Werke wirken." writingof sarcophagiin general,posits iconographyin detailand dates it back Tobin(1997, p. 223) rightlystates the thatthe rearrelief often "fluctuated to the Archaicand Classical periods. natureof the relationshipmore tenta- betweenan implieddependency on the 17. Fullerton1990; Zagdoun1989, tively,characterizing the figureson the frontand an apparentautonomy." passim. 468 ELLEN E. PERRY

Figure5. Drawingof a marbletorso of a female,found in the Kephisia tomb. Robert1890, p. 9

Whenthe workers cleared the tomb in 1866,they discovered the marble torsoof a recliningfigure, now missing.18 This figurewas clearly part of 18. We do not haveany specific the originallid. Carl Robert described the torsoas representing a female, informationas to wherethe torsowas andillustrated it in his publicationof the piece(Fig. 5).19 He alsosug- found.Benndorf (1868, p.39) states gestedthat the coffin was large enough to holda marriedcouple.20 Indeed, explicitlyonly that it wasdiscovered duringthe clearingof the tomb.There earlyin the developmentof the Attickline sarcophagus, when this piece is no evidencefor Tobin's statement wasapparently produced, portraits of recliningmarried couples were stan- (1997, p.224) thatthe torsowas found dardlid decoration,at least on sarcophagiintended for adults; other com- in the sarcophagusitself. Benndorf positions forexample, an individual reclining alone, or a pairof women- arguesfor the associationof torsoand wereexceptional.21 The evidencesuggests, therefore, that one woman or, sarcophaguson the basisof technical moreprobably, a married couple, occupied the Kephisiasarcophagus. features,such as the insetrectangle mentionedabove. 19. Robert1890, p.9. 20. Robert1890, pp.9-10, where HISTORICAL CONTEXT: THE CITY OF SPARTA he alsoreports the dimensionsof the AND THE FAMILY OF HERODES sarcophagus:L. 2.27 m (front)and 2.33 m (back);H. 1.12 m;W. 0.94 m The historyand aspirations of Herodes'family provide the bestexplana- (leftend) and 0.89 m (rightend). 21. In somecases the malefigure of tionfor the decorativeprogram of the threefinished sides. In particular, a couplehas been chiseled away and the thisprogram commemorates the family's association both with the cityof femaletransformed into a male.This Spartaand with the demeand Battle of Marathon.Although Herodes' factis of interestbecause, as Wrede tiesto Marathonhave longbeen known, the close association ofthis Athe- (1977, p.428) pointsout, it indicates nianfamily with Sparta has only been fillly recognized in recentyears. A thatthe productionof Attickline sar- cophaguslids was geared toward the seriesof articlesby AntonySpawforth has now demonstrated the exist- representationof marriedcouples. Mar- enceofthis significant relationship, attested largelyin inscriptions.22 Some riedcouples were the ruleonly for Attic of the evidenceis as follows:an inscription from Sparta records the erec- sarcophagi;other sarcophagi are more tion in thatcity of a statuein honorof Herodes'grandfather, who was variedin thisrespect. See Goette1993, apparentlyabenefactor oftwo brothers from Sparta.23 In addition, Herodes' p.108; Kochand Sichtermann 1982, pp.371-373. Fora discussionof the fatherwas almost certainly a citizenof Sparta;at anyrate, he seemsrather datewhen the klinesarcophagus was unusuallyto haveenrolled as a Spartanephebe. Near the endof his life, introducedto Attica,see below, sometimein themid-, he evenoccupied the eponymous patronomate pp.484-487. of Sparta,the highest magistracy of thecity at that time; and just before he 22. Spawforth1978,1985, andesp. died,the city appointedhim to serveas Kytherodikes,governor of the 1980. islandof Kythera,which had granted as a gift to thecity.24 Herodes 23. IG V i 516. 24. IG V i 380. He appearsto have himselfapparently enrolled as a Spartanephebe, and Spawforth has even diedbefore he couldserve his term; suggestedthat one Claudia Tisamenis, who was commemorated in Sparta Cartledgeand Spawforth 1989, bya statuarygroup that included her husband and son, was his sister. If so, pp.109-111. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 469

it seemsespecially significant that her parents named her afterSpartan familyfriends, and that she eventually married a Spartan.25 In additionto the epigraphicevidence cited by Spawforth,the very nameof Herodes'foster child, Polydeukion or Polydeukes,may serve as furtherevidence of a connectionbetween the family and the city of Sparta. A numberof scholarsbelieve that Polydeukion was not just a fosterchild but a bloodrelative of Herodes(see below,Fig. 17).While the precise natureof thiskinship has not yet been firmly established, the supposition appearsto be supportedboth by the boy's"nomen," Vibullius, which he sharedwith Herodes, and by two inscriptions that pair the name of Herodes' mother,Vibullia Alkia, with Polydeukion's.26 Indeed, the many inscriptions andmonuments that Herodes erected on variousfamily estates suggest thatthis relationship was more significant than the philanthropy through whichHerodes supported other youths.27 Polydeukion,also called Polydeukes, shares his namewith one of the Dioskouroi.There is significant,compelling evidence to suggestthat, in the 2nd centuryA.C., suchdivine and heroic names often functioned as claimsof descent.Thus the stemmaof the Spartanaristocrat Eurycles includesthe nameRhadamanthys because his familyintermarried with Cretanaristocrats; and Tyndares, the Spartanpriest of the Dioskouroi, mayhave claimed descent from the godswhom he served.28Indeed, an- otherfoster child of Herodes,Memnon, offers a strikingparallel.29 Mem- nonhas long been associated with a portraitof a youngAfrican found on Herodes'estate at Loukou.30 His name wollld appear to alludeto hisorigins, sincehe sharesit withthe Ethiopianprince reputed to havefought in the TrojanWar.31

25. The evidencefor Herodes' Herodes'biographer, Philostratus. twenty-fiveknown portraits of Poly- tenureas a Spartanephebe is IG V i 45 26. Scholarswho believethat deukion,and contrasts these with the (= Ameling1983, II, p. 98, no. 70), in Polydeukionwas kinsman of Herodes singlepreserved portrait of anotherfos- whicha certainKorinthas identifies includeGraindor (1930, pp.116-117); ter child,Memnon, which is discussed himselfas a "fellowephebe of Herodes Gazda(1980, pp.3-4); Ameling (1983, below.There are, as yet,no securely Atticus"(auvr(pog'Arrexou Tou I, p. 114);Meyer (1985, p.393); and identifiedrepresentations of a thirdfos- 'Hto8ou).Spawforth (1980, pp. 208- Tobin(1997, p. 99).The two inscrip- ter child,named Achilles. Polydeukion 210) oncebelieved that this referred tionsin whichboth Polydeukion and is "themost frequently represented to Herodes'son. Ameling (1983, II, VibulliaAlkia are mentioned are youthof the 2nd centuryC.E. whowas pp.98-100), however, has argued IG II23972 andIG II23973 (Ameling not in someway connected with the thatthe dateswould only allow for 1983,II, p.171, nos.174 and175, Romanimperial house" (Gazda 1980, thisfigure to be the sophisthimself, respectively).Herodes' full name is p.2). Meyer(p.393) citesthirteen in- andSpawforth (Cartledge and renderedas L. VibulliusHipparchus scriptionsthat mention Polydeukion, Spawforth1989, pp. 167,261, n. 10) TiberiusClaudius Atticus Herodes in butonly one concerningMemnon and hassince come to agreewith this IG II23603 (= Ameling1983, II, two,or possiblythree, concerning position.As forthe reconstructionof p.109, no. 89);see alsoAmeling 1983, Achilles. the identityof ClaudiaTisamenis, II, p.105, no. 76. Most of the other 28. Spawforth1985, pp.197,200- . . . . . Spawforth(1980, p. 213) pointsout nscrlpeonst lat menhonnlS name 201. thatTisamenus is an attestedSpartan abbreviateit greatly(see Ameling 29.The fosterchildren of Herodes name,but not an attestedAthenian 1983,II, pp.236-237).Only one other arementioned by namein Philostr.VS name.The family'spenchant for evenmentions the "nomen,"Vibullius, 2.558. . . r namlngglr s Wltn t ze temlnlzlng-lS whichcame to him fromhis mother's 30. Graindor1915; Blumel 1933, endingis seenin the exampleof sideof the family(Ameling 1983, II, pp.30-31, pl. 45. Herodes'daughter, Athenais. Claudia p. 123,no.103). 31. Gantz1993, pp.36-37,622- r m. - lsamenls1S not menhonecDy 27. Meyer(1985, p.393) citessome 624. 47o ELLEN E. PERRY

Amongaristocrats of the2nd century, claims of descentfrom the im- mortals,whether or not nomenclature reflected those claims, seem gener- allyto havebeen "based on localmythologies."32 Though such claims ap- pearalready among the Romanaristocracy of the Late Republic,their widespreaduse was an example of the muchbroader tendency during the SecondSophistic to consolidatecultural and political authority through appealsto the past.33Under such circumstances, a claim of descentfrom a god or heroof particularlocal importance implied local prominence. Onceagain, Herodes' own familyprovides examples. The Roman-born Regilla,Herodes' wife, is praisedas the descendantof Anchisesand Aphroditebecause she belonged to the Romanaristocracy.34 Herodes, on theother hand, claimed descent from Kekrops, , Keryx, and Herse, on the one hand,and and , on the other,all of whom areassociated in mythor historywith Athens or with the familydeme, Marathon(Philostr. VS 2.546-547).35 Regilla and Herodes, in otherwords, claimedheroic and divine ancestors associated with their local origins- Regillafrom the ancestorsof Rome,Herodes from the mythologicalno- bilityof Athensas well as fromsome of its greatestgenerals.36 Thus the nameof Herodes'maternal kinsman, Polydeukion, which evokes an asso- ciationspecifically with the Dioskouroi, more generally points to thegeo- graphichome of the twingods, the cityof Sparta.37 The evidence,therefore, points to a remarkableand unusually close relationshipfor a familyof Athenianorigin to havehad with the cityof Sparta,even in the internationalclimate of the 2nd century.While for- eignerswere commonly enrolled for ephebic training at Athens during the Principate,Herodes and his fatherare the onlyknown foreigners to have enrolledas ephebesat Sparta.38At leasttwo differentexplanations have beenoffered for the remarkablyclose association of thisAthenian aristo- craticfamily with Sparta.Walter Ameling suggests that the relationship wasessentially economic in origin thatit wasonly natural for one of the wealthiestfamilies of RomanGreece to takean interestin one of the

32. Spawforth1985, p. 193. fromthe Dioskouroi.These include IG the Dioskouroiand Sparta may also be 33.The mostfamous example of V i 463, 471, 529, 530, 537, 559, 562, the subjectof a reliefin the Athens LateRepublican and Early Imperial 971, 1172, 1174, and1399. One need NationalMuseum, NM 1450 (Rhom- appealto divineand heroic ancestry is not necessarilyhave come from Sparta iopoulou 1997, p. 92 withbibliogra- theJulian claim of descentfrom Venus to haveclaimed descent from the Dios- phy).On it, Polydeukionis depictedin andAeneas (Zanker 1988, pp. 195- kouroi,but claims of divinedescent are "heroicnudity" and with his horse,as 215).For appeals to connectedwith local mythology often if he wereone of the Dioskouroi.This the pastas a strategyfor establishing enoughthat"intermarriages can some- basicformula may be foundon other culturaland political superiority, see timesbe inferredfrom the evidencefor funeraryreliefs of the period(e.g., Alcock1993, pp. 163-164;Swain 1996, the transferof thesepedigrees from AthensNM 1775, Rhomiopoulou pp.65-100. familyto familyand city to city" 1997, p. 102), but the sculptorof 34. IGXIV 1389,lines 3-4 (= Ame- (Spawforth1985, p. 193). This is why, NM 1450 appearsto haveemphasized ling 1983,II, pp. 153-160,no. 146). whenthe familyof the dynastEurycles the allusionto the Dioskouroiby 35. Forthe mythologicalfigures, see claimsdescent from the Cretanhero depicting,on a pillarin the back- Gantz1993, pp. 233-239. Rhadamanthys,this claimcan be taken ground,an amphorawith a conicallid, 36. Evidenceof thispractice outside as an indicationthat someone in this a typethat Sanders (1992;1993) has Herodes'family includes a numberof familymarried a memberof the Cretan associatedspecifically with Spartancult inscriptionsfrom the areaaround aristocracy:see above,p. 469. andart. Spartain whichcitizens claim descent 37. Polydeukion'sconnections with 38. Spawforth1980, p. 204. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 47I

wealthiestand most powerfil cities of thetime, and that, furthermore, the family'spossession of nearbyland implied a vestedinterest in theregion.39 Spawforthand Tobin, on theother hand, suggest that the connection was politicaland sentimental in origin,and specifically that it mayhave stemmed fromthe period when Domitian condemnedTiberius Hipparchus, Herodes'grandfather, on chargesof tyranny,presumably over his fellow Athenians(Philostr. VS 2.547; Suet. Vesp. 13). We knowthat, on this occasion,Hipparchus lost his property; he maywell have also lost his life. Whateverhis fate, the members of hisfamily probably became temporary exilesfrom Athens, either by theirown choiceor becauseof a formal, imperialdecision.40 This exile may have provided the occasion for Herodes' fatherto forgethe strongties with the citythat resulted later in his ap- pointmentto the two importantmagistracies mentioned above (p. 468). Whateverthe reasonsfor the family'sclose association with a city that wasnot theirhome, that association is preciselythe sortof biographical detailthat an aristocratof the Imperialperiod might have wanted to commemoratewith a commissionedwork of art.

ICONOGRAPHIC ANTECEDENTS

SPARTANRELI EFS The familyof Helenwas unusual subject matter for a sarcophagusof the Imperialperiod. Mythological sarcophagi of the RomanEmpire tended to employone of a fewtraditional themes, including the stories of Orestes, Alcestis,Meleager, Adonis, and Selene with Endymion. In Attica, the range of availablethemes was even narrower than elsewhere in the Empire.41It is not alwayseasy to understandwhy certainmyths were more popular thanothers, but the reasonsmust have had to do withworkshop reper- toire;they may also have had to do withgenerally recognized sepulchral symbolism.42Most patrons apparently purchased prefabricated sarcophagi

39. Ameling1983, I, p.29. Ameling (p. 16) suggeststhatTiberius Claudius termann1982, p. 376).For an overview is referringto the familyestate near the Hipparchuswas put to deathsince of someof the mostpopular mythological Monasteryof Loukouin theThyreatis, thereis no evidenceof activityon his sceneson sarcophagi,see Turcan 1978; thoughSpawforth (1980, p.210) partafter his condemnation,which Sichtermannand Koch 1979; and believesthat land ownership in this tookplace around A.D. 92-93. IG V i Koortbojian1995. areais morelikely to havecreated ties 516, whichhonors him by name,would 42.The scholarshipon whetherand withArgos than with Sparta.Recent seemto indicatethat the familyhad to whatextent such themes may be con- excavationson the siteunder the someassociation with Spartaeven nectedto the life of the deceasedis vast. directionof TheodorosSpyropoulos beforehis disgrace,though events in Scholarswho considerthe topicat length haveproduced a greatdeal of new the aftermathof the condemnationstill includeCumont (1942); Nock (1946); architectureand sculpture. Although serveas a plausibleexplanation for the Turcan(1978); and Koortbojian (1995). mostof this materialis not yet enduringstrength of thatassociation. Foroverviews, see Kochand Sichtermann published,brief discussions of someof 41.The sculptorsof Atticafavored 1982,pp. 583-617; Kleiner 1988. Because the sculpturesappear in Spyropoulos someof the themesthat were common I amconcerned in the presentstudy with 1993;Datsouli-Stavridis 1993; and in Rome,like the storiesof Hippolytus a uniquedecorative program, I haveraised Tobin1997. andMeleager; however, they did not issuesof interpretationthat are related, 40. Spawforth1980, pp.204-205; makeuse of manyother themes that butnot identical,to thosecentral to Tobin1997, pp. 15-16,323-324.Tobin werepopular in Rome(Koch and Sich- scholarshipon commonthemes. 472 ELLEN E. PERRY

Figure6. Documentrelieffrom Spartadepicting the Dioskouroi and Helen.Sparta Museum 201. Photo J. Sanders,reproduced with permission of the GreekArchaeological Service

adornedwith one of a handfulof themesreadily available from a particular workshop.The depictionof Helen'sfamily, however, does not belongto the stockrepertoire. The unparalleledappearance of this themeon the Kephisiasarcophagus requires some explanation.43 Earlieranalyses of theKephisia sarcophagus those that suggest that it depicts,or is stronglyinfluenced by, individual Classical sculptures in theround do notoffer a thematicinterpretation of its program. Yet most of thefigures represented cannot, in fact,be associatedwith actual, known sculpturesin theround. The identifiable iconographic antecedents suggest thata thematicinterpretation is the correctone, andthat the familyof Herodesused the sarcophagusto constructand maintain a particularleg- endaryand historical identity. The frontof the Kephisiasarcophagus, with its depictionof the 43.The Ledaand Eros types each in thisrespect. appearindividually on a few other Dioskouroiand Helen (Fig. 1), is particularlyinteresting sarcophagi(Wiegartz 1983). It is the Becauseof the mannerin whicha sarcophagusis configured, the viewer familyas a wholethat is unusualas a experiencesits various fields front,sides, back, and lid sequentially.This themein funeraryart. sequentialviewing usually gives thematic priority to thefront register, of- 44. Brilliant1984, p. 125. tenproviding it withan immediate, autonomous meaning, one which the 45. Examplesin Chapouthier1935. endsof the boxmight or mightnot amplify.44The clearantecedents for The earliestsurviving figural repre- sentationsof the Dioskouroiflanking a thefront register, in thisinstance, are not individual sculptures in theround, femalefigure appear to dateto the end butrelief sculptures on whichall three figures appear in preciselythis ar- of the 2nd centuryB.C. (Chapouthier rangement,with Helen in the centerand the Dioskouroion eitherside.45 1935,pp. 97-98). ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 473

Figure7. Document relieffrom Spartadepicting the Dioskouroiand Helen. SpartaMuseum 202. Photo J. Sanders,reproduced with permission of the GreekArchaeological Service

Threedocument reliefs of the EarlyPrincipate now in the SpartaMu- 46. SM 201, SM 202, andSM 203. seum,two of which are illustrated here, offer the closest extant comparanda Tod andWace 1906, pp.33-34,158; Sanders1992, p.205. (Figs. 6-7).46The texts of these documentreliefs list membersof ox 47.These reliefs are dated on the SeTr0£V£5, participants in anannual feast associated with the twin gods.47 basisof theirinscriptions, since they The reliefsthat appearabove the inscriptionsdepict the Dioskouroiin andother Dioskouroi reliefs from the heraldicsymmetry, as theyalso appear on the Kephisiasarcophagus. On SpartaMuseum are often crudely the Spartanreliefs, an archaizingfemale figure wearing a kalathosstands carvedand difficult to datebased on betweenthem. stylisticcriteria. Sanders (1993, pp.218-219)acknowledges this Thetheme of Helen'sfamily, which is so unusualon sarcophagi,or in- difficulty,but suggeststhat some of the deed,in any2nd-century Athenian context, was a familiarone in Sparta, otherDioskouroi reliefs in the museum the mythologicalhome of the Dioskouroi.During the Empire,the twin maydate to as lateas the 2nd or 3rd godswere honored at severaldifferent sanctuaries around the city, includ- centuryA.C. ing at Phobaeum,where, perhaps in the 2nd or 1st centuryB.C., Helen 48. Cartledgeand Spawforth 1989, pp.194-195. cameto sharethe cult with them.48 For several , beginning in the 49. Forcoin devices, see Grunauer- Hellenisticperiod, the Dioskouroiand their attributes were also favorite von Hoerschelmann1978, p. 5. For deviceson Spartancoins and reliefs, and during the , Spar- administrationof the cultsand claims tanfamilies of the highestpolitical standing participated in the adminis- of descentfrom the Dioskouroi,see trationof theircult. Several even claimed descent from the heroes.49 Cartledgeand Spawforth 1989, p.162; Otherplaces in the Mediterraneanworld did Sanders1993, p.219. producereliefs of 50. Forexample, a numberof re- the Dioskouroiflanking their sister (Fig. 8).5° On thoseexamples, Helen liefsfrom Pisidia depict the trio;see doesnot generallyappear as an archaizingcaryatid with a kalathos,as on Metzger1952, pp.22-27, pl. III. the reliefsfrom Sparta. Fernand Chapouthier has suggestedthat such 474 ELLEN E. PERRY

Figure8. Relieffrom Telmessos iconographywas particularly popular in southernAsia Minor because many depictingthe Dioskouroiand Helen. citiesin thatarea traced their origins to Spartancolonization or Spartan KunsthistorischesMuseum inv.I 702. Courtesy Kunsthistorisches ancestors.5lThis interpretation, if true, would be entirelyin keepingwith Museum, Vienna, neg. I 6969 thecommon Second Sophistic practice of, in SimonSwain's words,"cities reassertingor recreating the roles of foundersand civic myths."52 It would alsoillustrate Swain's recent contention that"the primacyofAthens, Sparta, andArgos in theGreek heritage was particularly important in thedemon- strationof . . . Greekness."53 On eachof thecult reliefs from Sparta itself, Helen wears archaizing, stackedfolds, and her kalathostouches the moldingabove her; on one example,she clearlystands on a statuebase (Fig. 6). Stylistically,these figuresshare much with the archaizingcaryatids on the cornersof the Kephisiasarcophagus. This similarity lends a meaningfulambiguity to the overalldecorative program of thelatter. On the onehand, we understand thefemale figure on thefront to be Helenbecause of herposition between the Dioskouroi.On the other,the styleof the cornercaryatids and their associationwith the rest of theprogram require us to understandthem as depictions or,more precisely, as statues of Helen,likethose on the Spar- tandocument reliefs. The ambiguityof Helen'srole as centralfigure and corner caryatid unitestwo very different sculptural traditions, the Spartancult reliefs and 51. Chapouthier(1935, pp. 100, the klinesarcophagus, into one,eclectic composition. Herms, caryatids, 232) furtherargues that the female andother support figures often appear at the cornersof Attickline sar- figureon thesereliefs is a syncretism of Helenand a localdivinity. cophagi.Since the kline sarcophagus is,by its very nature, styled as a funerary 52. Swain1996, p. 73. See also couch,corner support figures are analogous to thelegs on a genuinecouch. Alcock1993, pp. 163-164. The typeand styleof the supportfigures on klinesarcophagi can vary 53. Swain1996, p. 75. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 475

greatly,which surely indicates that the specificstylistic parallels between theKephisia caryatids and the Spartan Helen-caryatids are significant. By depictingHelen and the Dioskouroiin theircustomary alignment on the frontof thebox, but removing the archaizing, caryatid version of Helento thefour corners, the artist is ableto evokeSpartan cult without abandon- ing a decorativemotif common to manykline sarcophagi. Theremay be anotherreason for the dualrepresentation of Helen on thissarcophagus. The artisthas moved the Helen-statues, but not removed them,providing a crucial,conceptual clarification for the viewer.When thesubject of sculpturalrepresentation is,itself, a sculpture,the artist must indicatesomehow that this is the case.Otherwise, the viewer may misin- terpretthe subject as an actual person or divinity. The front of theKephisia sarcophaguspresents an inversionof thisdynamic. Since the Spartanre- liefs areits primaryreferent, a knowledgeableviewer might easily have assumedthat the centralfigure on the sarcophaguswas also a statue.The archaizingcaryatids at the cornerscreate the necessaryvisual context to contradictthis assumption.Their very presence, as well as the stylistic contrastthey offer, indicates to theviewer that the woman on the frontis a "real"(albeit idealized) woman. The transformationof the centralfigure from statue (on the Spartan reliefs)to woman(on the frontof the sarcophagus)may be motivatedby the desireto conveya moreintimate tribute to the deceased.If anyfigure on this boxrepresents the deceased,it is likelyto be the womanin this position,front and center. To place her in therole of Helen,while carefillly distinguishingher from recognizable representations of the Spartan statue, setsa personaltone that, to someextent, balances the multivalent allusions to familyhistory with a particularcommemoration of the deceased.

THE CONNECTIONWITH MARATHON Thefigure on theleft end of thesarcophagus, Leda, is no morea depiction of a famoussculpture in the roundthan are Helen and the Dioskouroi. Whileshe does give this general impression sheseems to havethe pro- portionsof a LateClassical female figure such as the Knidian Aphrodite in fact,the motif of Ledalocked in a viciousstruggle with a swandoes not 54.The relieffrom Brauron appearuntil the Hellenisticperiod. A handfillof Hellenisticreliefs, in- is in the AthensNational Museum (NM 1499);that from Argos is in the cludingone fromArgos (Fig. 9) andone fromBrauron, are sufficiently BritishMuseum (BM 2199).The similarto oneanother to be consideredreplicas of thesame model, and are prototypeof all of thesereliefs may arguablythe earliest-known depictions of thetheme.54 Classical representa- havebeen a reliefor painting,since tionsof Ledaand the swan, including the replica type that scholars attribute thereare no knownsculptures in the toTimotheos (Fig. 10), differ greatly from these Hellenistic struggles, since roundthat reproduce this type. Wiegartz(1983, pp. 171-174,fig. 2) theydepict Leda simply holding or embracingthe creature. Indeed, Clas- citesone version that appears to be a sicalvase paintings depicting mythological rape tend to emphasizepur- sculpturein the round,and then argues suit;they depictphysical contact less often,and consummationalmost convincinglythat it is in facta table never.55 foot whosecomposition is derivedfrom Hellenisticreliefs such as thosefrom Brauron and Argos represent a one of the manyrelief versions. widespreadtype that can be seenin differentmedia for centuries following 55. Shapiro1992, pp. 61, 64-65. 56. Wiegartz1983. For an itemiza- its firstappearance.56 As on the Kephisiasarcophagus, the enormous swan tion of gem representations,see depictedin theseexamples appears on the left, his nakedvictim on the Dierichs1990, pp. 46-47. right.There are, however, significant differences between these putative 476 ELLEN E. PERRY

Figure9. Reliefof Leda and the swan,from Argos. BritishMuseum 2199. CourtesyBritish Museum, neg.PS 215697 prototypesand the Kephisiaversion. The proportionsof the Ledafrom Argosare less classicizing in overallconception than those of the delicate Kephisiafigure. The torsoof the Argosfigure is longerand her buttocks andthighs are more prominent, features that combine to makeher seem heavierand less classically restrained than the Leda depicted on the Kephisia sarcophagus.Moreover, the reliefs from Brauron and Argos display a bru- talitythat is wellbeyond that seen on the Kephisiasarcophagus.57 On the former,Leda and the swan are closely locked. She does not have the strength to holdthe swan away from her, as she does on thesarcophagus. The swan's talonsclutch at herthigh and he useshis sinuousneck to gainaccess to her,but instead of tryingto kissher, he pinsthe back of herneck with his beakand pushes it down,forcing her to bendforvfard (Fig. 9). In response to thisviolent attack, Leda's left arm strains downward as she fights vainly 57. Wiegartz1983, p. 172. to keepherself covered. 58. Lacroix1956, pp. 22-23. By contrast,the relativerestraint exemplified by the figureson the 59. So alsoKampen 1988, pp. 15- sarcophagusbetrays a sensibilitythat was common in mythologicalrepre- 16, for a discussionof the Rapeof the sentationsof rapeduring the Roman Empire. A numberof Imperialpaint- SabineWomen on the friezeof the BasilicaAemilia. Kampen argues that ings,sculptures, and coins depicting figures such as Apollo and Daphne or the stylisticclassicism of this monu- Heraklesand Auge demonstratethis phenomenon.58The classicismof mentlends authority to the frieze's styleand calm demeanor of suchscenes tend to directthe viewer's atten- message,which encourages Roman tionaway from the brutal act of rapeitself in orderto promotesome other womento view Sabinewomen as message.59The decoration of theKephisia sarcophagus similarly promotes exempla becauseof theirimportant role an alternativereading, one thatestablishes family identity by evokinga as socialmediators. 60. Forthe employmentof legends legendarypast. Its programis thereforeless about the encounterof Leda aboutrape to fashionidentity in this withthe swanthan about the offspring,both legendary and historical, of casecultural rather than familial see herunion with Zeus.60 Joshel1992. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 477

Figure10. Leda and the swan. Replicaof a typecommonly attrib- utedto Timotheos.Capitoline Museum302. Courtesy Deutsches ArchaologischesInstitut, Rome, neg. 75.2228

61. Replicalists appear in Dohl A Hellenisticcomposition, then, inspiredthe representationof Leda 1968,pp. 10-13, 49-50, 62-63; and in LIMC III.1,1986, p. 881,s.v. Eros on the Kephisiasarcophagus. But the sarcophagusreinterprets the Helle- (A. Hermary,H. Cassimatis,and nisticcomposition by mutingthe violenceofthe scene,and providing Leda R. Vollkomer). with moreclassicizing bodily proportions. In otherwords, this figuredoes 62. Moreno(1995), in particular, not simply"copy"a famous, lost sculpture,as earlierscholarship suggested, attributesboth Eros types to Lysippos: but reinterpretsits predecessorsin a typicallyRoman, eclectic manner the morefamous type (e.g., Fig. 12) andthe chubbytype that is reproduced one that combinesclassicizing proportions with Hellenisticiconography. on the Kephisiasarcophagus. He as- The Eros on the rightside of the Leda sarcophagus(Fig. 2) holds the signsthe chubbyEros to Lysipposon top of his bow with his right handwhile his left handgrasps the middle. the basisof perceivedsimilarities with Unlike the other figureson the sarcophagus,he does reproducea wide- the so-calledLateran Poseidon sculp- spreadreplica type, known so far from seven sculptures,two reliefs,and pp.108-146), tures.Bartman (1992, a gem (see, e.g., Fig. 11).61He is a chubby,long-haired baby who leans however,throws into questionthe very existenceof a singleprototype for the forwardwith both legs bent at the knees, and his right leg advanced; many,and apparently quite varied, he looks awayfrom, ratherthan toward,his bow.These featuresare in Romanstatuettes traditionally called distinctcontrast to those of the Eros type traditionallyattributed to Ly- "LateranPoseidon." Dohl (1968,pp. sippos (Fig. 12), though some scholarshave attributedboth types to that 49-50), whileacknowledging that the artist.62 chubbyEros could be a LateHelle- nisticcreation, prefers to datethe Only one of the five figureson the front and sides of the Kephisia prototypeto the EarlyEmpire, in part sarcophaguscan, therefore,be identifiedwith a known sculpturaltype. becauseof its eclecticismof style. Nevertheless,the figuresappear to the viewerto reproducesculptures in 478 ELLEN E. PERRY

Figure11 (leJi). Eros.Museo the round.As notedearlier, this impressionresults from the mannerof ArcheologicoNazionale, Venice, theirrepresentation: the figures stand at a distancefrom one another, dis- inv.170. Courtesy Soprintendenza Archeologicaper il Veneto engaged.The fact thatthere is little depthto the composition,and no landscapeor architectural setting, enhances the impression further. Yet, to Figure 12 (right).Eros of the type 2nd-centuryeyes, these formal characteristics need not havesuggested attributedto Lysippos.Capitoline sculpturesin theround; other reliefs from this period are characterized by Museum410. CourtesyDeutsches preciselythese formal qualities and do notexplicitly replicate sculptures in ArchaologischesInstitut, Rome, theround. A famousexample is theProvince series from the Hadrianeum neg.79.1577 in theCampus Martius. These personifications of the provinces (Fig. 13), likethe figures on the Kephisia sarcophagus, adhere to a coherentthematic program,but each individual figure stands apart from the others, surrounded bya blank,undecorated background that enhances her statuesque isolation. As RichardBrilliant has pointed out, the endsof a sarcophagusoften "extendthe implication of thefront frieze but at a secondorder of impor- tance."63Inthis case, the ends take up a themefirst presented on thefront registerand, together with the klinelid, transform it intoan explicit allu- sionto a differentsort of monument,the Classicalcult-statue base. The blankbackground, spatial distance, and lack of engagementbetween fig- ures,while unusual for sarcophagusdecoration at anyperiod, is compa- rableto whatwe knowof cult-statuebases from the Classicalperiod, in- cluding,for example,the MantineiaBase attributed to the workshopof 63. Brilliant1984, p. 138;he is Praxiteles(Fig. 14), or the base of thecult statue of Nemesisat Rhamnous describingan Achillessarcophagus in (Fig.15). The impressionof a cult-statuebase would have been enhanced the CapitolineMuseum. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 479

Figure13. Province relief, Hadrianeum.Courtesy Deutsches ArchaologischesInstitut, Rome, neg. 845 by the fact that the Kephisiasarcophagus supported a sculpturein the round,the reclining figure of thedeceased, in a positionanalogous to that of a cultstatue on its base.64 Neitherof thesepoints is sufficientin itselfto demonstratethe allu- sionto a cult-statuebase, but the subject matter themiraculous concep- tionand the familyof Helenof Troy-also seemsto supportthe reading: miraculousbirths and family connections were popular themes on such basesduring the Classicalperiod.65 Since the Kephisiasarcophagus was producedvery early in the historyof Attickline sarcophagi, and the type wouldstill have been unfamiliar in thearea around Athens, a viewerfrom Atticamay have been less likely than a Romanto placethis work readily into its "correct"typological context, that of the klinesarcophagus, and thereforemore likely to perceivean allusion,based on theseformal and thematicparallels, to a Classicalcult-statue base.

64.The recliningposition seen in Polydeukionin a recliningposition was"neither temple nor tomb": the femaletorso from the Kephisia analogousto thatof a klineportrait. IGXIV 1389,line 43 (= Ameling tombis not commonamong cult Herodesis knownto haveheroized 1983,II, pp. 153-160,no. 146).It statues;it is, however,consonant with Polydeukion,even instituting games in seemspossible, therefore, to under- traditionalGreek depictions of the his honor,and this reliefrepresents the standthe femalefigure that once re- heroizeddead, and, in Atticaat this boyas a hero,reclining on a couchas at clinedon top of the Kephisialid as a time,might have been read as an a banquet.We knowthat Herodes also heroine,whose semidivine status would advertisementof the semidivinestatus heroizedat leastone of the womenin havebeen reinforced by the resem- of the sarcophagus'soccupant. A relief his family,his wife Regilla.A memorial blanceof hersarcophagus to a cult in the BrauronMuseum (Gazda 1980, inscriptionfound on the familyland statuebase. p. 4) providesan interestingcompari- outsideRome indicates that Herodes 65. Ridgway1981, p. 173;Lapatin son:it depictsHerodes' foster child constructeda restingplace for her that 1992,p. 113,n.20. 480 ELLEN E. PERRY

It shouldcome as no surprisethat a sculpturefrom the familyof Figure 14. Partof the Mantineia Base,Athens, NationalMuseum. HerodesAtticus alludes, in styleand content, to the basesof cultstatues Courtesy Alinari 24305 fromthe Classicalperiod. Herodes was one of the mostfamous figures of the SecondSophistic, the classicizingliterary, linguistic, and rhetorical movementthat was so popularduring his lifetime.Philostratus tells of Herodes'devotion to ancientrhetorical models, and in particularof his specialfondness for speaking in thestyle ofthe 5th-century oligarch Critias (Philostr.VS 2.1.34-35). The classicismand intellectual activity associ- atedwith this period were not, however, limited to literaryand rhetorical style.Rather, as Jas Elsner notes, the movementwas "a deep antiquarian examinationof the arts,rituals and myths of the canonicalGreek past" thatapplied to allrealms of elite,public life.66 Boththe styleand content of the visualarts during the SecondSo- phisticreflected these antiquarian and intellectual values.67 Herodes him- selfcommissioned other sculptures that might be deemedantiquarian or classicizing.His portraits,for example, like those of so manyintellectuals of the period,alluded generally to Late Classicalimages of important

66. Elsner1998, p. 169.For the 67. Elsner(1998, p. 181),in an Greeksculptor Kresilas. But the

. . P . P . . varlous manltestatlons ot antlquanan- analysisof the Spadareliefs, concludes: finishedpanels are a fundamentally ism in the SecondSophistic, see Bowie "Thusthe Spadarelief of Bellerophon Roman,Second Sophistic, product- 1974;Alcock 1993, pp.163-164, 189- bringsto mindPolyclitus' Doryphorus, a creativetransformation of the 198;Alcock 1994; Anderson 1993; forexample, while Diomedes in the givenmodels into idealizingantiquar- Zanker1995, pp.247-256; Swain reliefof the theftof the Palladiumis a ian compositionswith a romantic 1996. copyof a statue-typeattributed to the mood." ... . - ...... , )_ , , >...... _ ,,,,,., ,-^ os ... ,,,, ...... 6 _ .. 7 * § fl

tll

I - ?

t - - 8 - - -

-efi @t # !-< t X t

#F , , = t ? n

{ 2 2s 3 t

Figure 15. Reconstructionof the base of the statueof Nemesis at Rhamnous.Drawing M. Belozerskaya,in Lapatin1992, p. 109,fig. 1

oo H 482 ELLEN E. PERRY individuals.68We alsoknow of twocult statues commissioned by Herodes in ivoryand gold, an uncommonchoice of material,and one thatwill surelyhave reminded Athenian viewers of thegreat chryselephantine stat- uesof the Classicalperiod.69 One parallelbetween the Kephisiasarcophagus and the baseof the cultstatue of Nemesisat Rhamnous(Fig. 15) is particularlynoteworthy: bothdepict the same subject, the family of Helenof Troy, disposed around threesides of a four-sidedmonument.70 On thefront of theNemesis base, fourfemale figures stand in the center,flanked by two malefigures on eitherside. Kenneth Lapatin has identified three of thesefigures convinc- inglyas Helen, Leda, and Nemesis, while traces of piloi testifyto thepres- enceofthe Dioskouroi. (1.33.7-8) also indicates thatTyndareus "andhis children"were represented on the Nemesisbase.71 Sharedsubject matter is not, of itself,necessarily significant, espe- ciallyif one monumentis not a literalcopy of the other.In the present case,however, other evidence would seem to suggestthat this subject mat- ter is meaningful,and that the sarcophagusmay allude, deliberately and explicitly,to the Nemesisbase. The similaritiesof style,for example, and thevery rarity of the subjectmatter support this hypothesis. 68. Zanker(1995, pp.22>221, Perhapsmost important, however, is the historicaland mythological 243-244) emphasizesthe intellectual associationbetween the cult statue of Nemesisand Herodes' family deme, qualitiesof Herodes'portraits; Smith Marathon.According to Pausanias,the cult statue at Rhamnous was erected (1998,pp. 78-79) suggeststhat they as a resultof the Battleof Marathon,the very battle that was fought and borrowsome of the syntaxof the wonby Herodes'ancestor, Miltiades (Paus. 1.33.2-3; see alsoAnth. Pal. famousportrait of Demosthenes.For the desirabilityof representingoneself 16.221,222, 263 foranecdotal variants).72 The storywas that the Persians as an intellectualduring this period, see transporteda block of Parianmarble with themto Greecein 490 B.C., Ewald1999. with the intentionof usingit for a victorymonument, but the wrathof 69. On the chryselephantinestatues Nemesisfell upon them at Marathon, and the marble was used instead by commissionedby Herodes,see Lapatin Pheidiasto createthe cultstatue of Nemesis. 2001,pp. 127-128.I thankProfessor Lapatinfor allowing me to see partsof This 2nd-centuryversion and its variantsin the PalatineAnthology his manuscriptwhile the bookwas still differsignificantly from the storyrelated by Pliny(HN36.17), in which in press. thesculptor is Agorakritos,and the Battle of Marathonplays no role.The 70. See Lapatin1992, pp. 113-114. laterversion may, therefore, represent atypically Second Sophistic reworking 71. See Lapatin1992. Lapatin of the past,according to whicha traditionor legendis "recovered"and makesuse of Pausanias'sdescription, subsequentlyemployed to establisharistocratic or civicidentity and cul- but alsonotes some of the problems associatedwith it. turalsuperiority.73 It also exemplifies a topos of SecondSophistic literature, 72. Forthe familyclaim of descent andof Pausaniasin particular,the identityof imageand divinity.74 The fromMiltiades, see above,p. 470. marblecomes to Greecebecause of an actof hubris,but falls into Greek AlthoughHerodes owned estates all possessionbecause of anact of Nemesis.The very material therefore sym- overGreece and even in ,Mara- bolizesboth the goddessand the abstractidea for which she stands. This thonwas the familydeme, the putative locationof the family'sorigins. The marblewas Nemesis even before the sculptorundertook to transformit

. . considerablearchitectural remains of nto zer1mage. the estateat Marathonare discussed By the2nd century,then, the cultstatue of Nemesisserved both as a in Tobin1997, pp.241-283. trophyto Miltiades'victory, and as a symbolof thejust punishment im- 73. Forparallels, see Swain1996, posedby the goddess on thePersians. For this reason, Herodes and his kin pp.65-100. musthave honored the cultand its statuegreatly, a hypothesisthat finds 74. Elsner1996. 75. IG II23969 (= Ameling1983, II, supportin an inscriptionfrom Rhamnous indicating that Herodes was pp. 169-170,no. 173),which also indi- activein thatsanctuary on morethan one occasion.75 catesthat Polydeukion was present Naturally,the Kephisia sarcophagus does not simply copy the Nemesis c urlng. tnese acavltles. . . . 76. Thejob of reconstructingthis centuryA.D ...... as a primemeans for

ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 483 base.There were, for example,fourteen figures represented on the base, butonly five on the sarcophagus(or nine, if oneincludes the cornercary- atids).76More troubling, perhaps, is an iconographicdifference that has somebearing on the interpretation ofthe Kephisia piece. On the sarcopha- gus,Leda mates with the Zeus-swan,and is thereforethe naturalmother of Helen and the Dioskouroi;Nemesis does not appearat all. On the Rhamnousbase, by contrast,Pausanias indicates that Leda was merely a nurseand adoptivemother, while Nemesis was Helen'snatural mother (Paus.1.33.7). Farfrom being problematic, however, such a discrepancymay be a deliberatestrategy of self-consciousmythmaking on the partof Herodes' family.Graham Anderson, writing of literaryimitation during this period, hasobserved that, in additionto the moreobvious sorts of literaryimita- tionpracticed at the time (including stylistic imitation of a particularClas- sicalauthor's style and pastiches of severalworks of literature),"sometimes a classicalframework will serveto evokethe ethosof anauthor when the imitatoris obviouslysetting himself at variancewith his originalin some significantway"; other scholars have identified analogous transformations in the visualarts.77 If we applythis strategyto an interpretationof the Kephisiasarcophagus, we maysee that it evokesthe ethos ofthe cult statue at Rhamnous,but sets itselfat variancethrough a significantdeparture fromthe iconography ofthe "original."This difference-that Leda, on the sarcophagus,is undeniably the true mother of Helen-directsthe viewer's thoughtsback to Leda'shome, Sparta, and to the Spartanvotive reliefs thatinspired the frontregister. In otherwords, it is preciselythe icono- graphicdiscrepancy that servesto combinethe two verydifferent civic referencesinto one, possibly as a symbolicreflection of themarriage of the deceased. Takenas awhole, the decorative program ofthe Kephisiasarcophagus is a distillationof severaliconographic and typological antecedents, in- cludingnot onlythe klinesarcophagus and the herorelief, but also (and mostimportantly) Spartan cult reliefs and the cultstatue of Nemesisat Rhamnous.The combinationof theselast two typesaccords with a gen- eraltendency during this periodto employart, especially art associated with cult,in the establishmentof one or morecultural identities.78 The resultis an eclecticand multivalent creation of a sortthat is commonin sculpturalproduction of the RomanEmpire.

monumentfell to GeorgiosDespinis ethnicand religious self-assertion." andVasileios Petrakos: Petrakos 1981; Smith(1998, pp. 7(}78) discusses 1986.An identificationof the various severalwell-known monuments from figuresis suggestedby Lapatin(1992). thisperiod that negotiate the dual 77. Anderson1993, p. 72. For (Greekand Roman) identity of the severalanalogous case studies in the aristocratfrom the GreekEast. He visualarts of the RomanEmpire, see eveninterprets one of these,the Gazda,forthcoming. Monumentof Philopappos,as a display 78. Elsner(1997, p. 196) arguesthat of the honorand'striple identity as "theicons of paganpolytheism in the Macedonianking, Athenian archon, eastwere used during and after the 2nd andRoman consul. 484 ELLEN E. PERRY

DATE OF THE SARCOPHAGUS AND IDENTITY OF THE D ECEASED

A relativechronology was established by Arnold Tschira for the fourcof- finsfound in the Kephisiatomb. He concluded,on thebasis of thesize of eachbox, the availablespace, and the wearmarks on the floor,that the garlandsarcophagus (1) was placed in thetomb first, the Erotessarcopha- gus (2) second,and the undecoratedsarcophagus (3) third(Fig. 16). The Ledasarcophagus (4) was, according to hisreckoning, almost certainly the lastof the fourto be movedinto the tomb.79 If it is a fairlystraightforward matter to establishthe relative chronol- ogy of the foursarcophagi in the tomb,it has provensomewhat more difficultto establishreliable absolute dates. Most scholars agree that the tomband its contentsdate firmly to the 2nd centuryA.C., but detailed argumentshave failed to securea moreprecise date. Because both the style andthemes depicted on the Ledasarcophagus are unusual, Koch has sug- gestedthat it belongedto an early"experimental"phase in the develop- mentof Atticsarcophagi, which he datesto approximatelyA.D. 170-180. Whilemost Attic sarcophagi appear to havebeen produced by a narrowly definedcircle of sculptors,he argues,a few wereproduced during this decadeby other, less successful workshops. A numberof themesand mo- tifsappeared during this period that did not, ultimately, prove to be popu- larand so underwenta sort of"natural selection" and disappeared from the repertoire.80Koch places the Kephisia example in thisexperimental phase, andsuggests an evenmore precise date for it, about180, on the grounds thatthis is whenWiegartz dates the arrivalof the klinesarcophagus in Attica.8l The Kephisiasarcophagus was undoubtedly produced very early in theAttic kline series, as its rectangularinset implies.82 Attendant cultural circumstancesalso suggest that this coffin was one of thefirst examples of itstype to appearin .Kline sarcophagi were produced both in Rome andin AsiaMinor before they arrived in Attica,so it wouldnot be at all surprisingto discoverthat they were introduced from one of theseareas by a familyor familieswith the cosmopolitanlifestyle and connections that werede rigueur among the intellectual elite ofthe 2nd century.83 The wealthy andprestigious family of Herodescould plausibly have been among the firstin Atticato employthis new type.

79.Tschira 1948-1949, col. 86;see moreplausible than Benndorf's hypo- 1975,pp. 187-188,219; Wiegartz alsoTobin 1997, pp. 222-228. Benn- thesis,and has also been accepted by 1977,p. 386. dorf(1868, p. 36) originallysuggested Tobin.The inscriptiondiscussed above, 82. See above,p. 465 andn. 10. thatthe Ledasarcophagus was the p. 462, seemsto supportTschira, since 83. Forthe earlyappearance of the thirdto be placedin the tomb,and that it suggeststhat the undecoratedsarco- klinein Romanfunerary art, see Koch the undecoratedsarcophagus went in phagus,the one thatprobably held the andSichtermann 1982, pp. 58-61 last.This orderwas presumably based infant,was placed in the tombat a time (klinemonuments) and pp. 66-67 on the factthat the undecoratedsar- whenonly three children were buried (klinesarcophagi). For the expectation cophaguswas found closest to the door, there. thata sophistof thisperiod would be andthe Ledasarcophagus stood behind 80. See above,p. 462 andn. 7. well-traveledand function as a"cultural it. Tschira'sargument, based as it is on 81. Kochand Sichtermann 1982, travellerand cultural ambassador," see the wearmarks on the floor,seems pp.371, 415, 422, 460-461;Wiegartz Anderson1993, pp. 28-30. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 485

A B c Figure16. Reconstructionof the orderand mannerin which sarcophagiwere installed in the Wiegartz,however, offers a veryspecific date for the introductionof Kephisiatomb. Number 4 is the Attic klinesarcophagi, A.D. 180,virtually without argument.84 He does Leda sarcophagus.After Tschira assertin a footnote butagain without argument or appeal to monuments 1948-1949,cols. 87-88, fig.2, courtesy of fixeddate-that the introductionof the klinelid in Atticaoccurred DeutschesArchaologisches Institut, Athens "somewhatlater" than in AsiaMinor, where the firstexamples date to aroundA.D. 165-170.85He offersno compellingevidence for this date, nordoes he explainwhy the first Attic examples had to be produceda full decadeor moreafter those from Asia Minor. The argumentsoffered thus farcan thereforeonly safely secure a datefor the Kephisiasarcophagus between170 and180. Anotherapproach to datingthe sarcophagusmight be to identifyits occupants.Tobin has assignedall fourcoffins in the Kephisiatomb to childrenof Herodes,in partbecause of the inscriptiondiscussed above thatmentions three children who clearly predeceased him. She assigns the undecoratedbox, the third to be interred,to the infantmentioned in that inscription,speculating that since Regilla was eight months pregnant at thetime she died, her unborn child might have lived beyond her death and thenbeen interred at Kephisia.She then assigns the othertwo coffinsto Herodes'children Regillus and Athenais, who died in thelate . Tobin acknowledgesa potentialproblem with the attributionto Athenais: Philostratusthe Elderexplicitly states that she was buriedelsewhere.86 Finally,she hypothesizes that the Ledasarcophagus might have belonged to Herodes'daughter, Elpinike, who diedbefore her father.A funerary inscriptiondoes suggest that the family buried her in Kephisia;and Elpinike diedafter several other siblings, which corresponds well with Tschira's rela- tivechronology for the fourcoffins in the tomb.87

84.Wiegartz 1975, pp. 188,219. tion of the klinelid into theworkshops thatshe was buried in the cityby the He repeatsthe assertionin Wiegartz of AsiaMinor is basedon the stylistic Athenians,who alsodecreed that the 1977,p. 386, againwithout argument, evidenceof one of the earliestexam- dayof herdeath should be takenout of onlyciting his previouspublication. ples,the Melfisarcophagus, now in the the calendaras a consolationto her Koch(in Kochand Sichtermann 1982, MuseoNazionale del Melfese.The father. p. 371, n. 37) thenreasserts the date, portraitlid hasbeen dated to about170 87. Fora discussionof the inscrip- alsowithout argument, citing both of basedon its resemblanceto likenesses tion,which was found in Kephisia,see Wiegartz'searlier publications. of Faustinathe Younger. See Kleiner Tobin1997, p. 228. Forthe text,see 85.Wiegartz 1975, p. 188,n. 161. 1992,p. 306 withbibliography. IG II212568/9 (= Ameling1983, II, The traditionaldate for the introduc- 86. Philostr.VS 2.558. He writes p. 140,no. 136). 486 ELLEN E. PERRY

Specificattributions of thesefour coffins are problematic for two rea- sons.First, we cannotbe surehow many children were born to Herodes andhis wife,Regilla. Philostratus, for example,in spiteof the wealthof detailhe providesabout the family, does not mention a boywho was born to the coupleearly in theirmarriage while they were still in Rome,nor is thisinfant mentioned in the largecorpus of inscriptionsassociated with Herodes.We onlyknow of his existencebecause of a letterwritten by MarcusAurelius to hisold teacher, Fronto, requesting that the latter make an effortto consoleHerodes on the boy'sdeath (Fronto Ep. 1.6.10).88 A secondcomplication is that Herodes raised an unspecified number of fos- terchildren. We knowboth from inscriptions and from Philostratus that he treatedsome of thesechildren "as his own."Indeed, his fosterchild Polydeukionreceived better treatment in lifethan Herodes' own son, Atti- cusBradua. This being the case, we cannotentirely rule out the possibility thatone or more of Herodes'foster children may rest in theKephisia tomb. If it provesimpossible to assignall four coffins to knownpersonalities, however,the Ledasarcophagus may prove to be an exception.The now- missingfemale torso from the lid suggests that the deceased reached adult- hood;therefore, historical and epigraphic sources probably name her. The factthat early Attic kline sarcophagi tended to containand represent mar- riedcouples also narrows the field of possibilities:there is onlyone couple in Herodes'familywho might claim both a hereditaryinterest in theSpar- tan Dioskouroiand the rightto a burialplace in Kephisiaalong with Herodes'children: his daughter Elpinike and her husband, Lucius Vibullius Hipparchus.89If these areindeed the coffin'stwo occupants,then the familymust have commissioned the sarcophagusat thetime of Elpinike's death. Elpinikeof coursehad an undisputedright to a placein the family tombat Kephisia.Both the female torso that was found in excavationand theiconographic program of thebox itself argue strongly in favorof iden- tifyingher as the deceased. This is allthe more true because her very name, Elpinike,is anexplicit reference to thefamily connections with Marathon: it wasthe nameof Miltiades'daughter. The name"Elpinike" was an un- usualone in 2nd-centuryAthens, and thus it wouldbe difficult to miscon- struethe reference,which functioned, like the namesof Polydeukionand 88.The failureof ourother sources Memnon,as a recognizableallusion to familyorigins, in thiscase specifi- to mentionthe infantor his passingis callyto theBattle of Marathonand to thegeneral who won that battle.90 It all the moreastonishing because the shouldnot be at allsurprising, then, that Herodes buried his daughter in a emperor'sletter makes it clearthat coffinwhose decorative program and style alluded, at leastin part,to the Herodestook the deathhard (id Herodesn o n aequof ert an im o) . cult statueof Nemesis,who playedsuch an importantrole in the very 89. LuciusVibullius Hipparchus is samebattle. identifiedas Elpinike'shusband in The deathof Elpinikehas recentlybeen assigned to the mid-, IG II212568/9 (= Ameling1983, II, whichon thesurface does not seem to sitentirely well with the traditional p. 140,no. 136). datingof the Kephisiasarcophagus to aboutA.D. 170-180.However, the 90. Her fullname was Appia Annia securetermini for Elpinike'sdeath derive from the factthat she outlived RegillaAgrippina Elpinike Atria Polla(Ameling 1983, II, pp. 139- hermother, who died around 160, but not her father, who died in thelate 140,nos. 134, 135, and136 [=IG II2 .Scholars have offered two arguments for a datearound 165, neither 12568/9]),though Philostratus refers of whichis conclusive.Tobin believes that she might have died during the to hersimply as Elpinike. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 487

plaguethat struck around 165, which is indeeda possibility,but there is no otherevidence to supportor refilteit. Amelingclaims that Philostratus datesElpinike's death to beforethe deaths of Herodes'three most famous fosterchildren, Polydeukion, Memnon, and Achilles. There is, however, no indicationin Philostratus'stext of the chronologicalrelationship be- tweenher deathand theirs. It is truethat he discussesElpinike's death earlier,but only a fewlines earlier, and in anycase this does not prove that herdeath occurred first. In fact,Philostratus does not provideany clues thatmight help us determinewhen any of theseyoung people might have died.91Thus, as far as the surviving evidence is concerned,Elpinike might verywell have died as late as the mid-170s. It is possiblethat the Kephisiasarcophagus dates to aroundA.D. 170- 175,five or tenyears earlier than the datesuggested by Wiegartz for the introductionof thekline sarcophagus into Attica. This revision would cor- respondmore closely to a plausibledate for Elpinike'sdeath. If Elpinike wasindeed buried in theKephisia sarcophagus, its datemight serve as one of the best fixedchronological points for the introductionof the kline sarcophagusinto Attica, since Elpinike's death, though not pinpointed to theyear, can at leastbe datedto withina rangeof . ThatElpinike shared this resting place with her husband is anattrac- tiveidea for several reasons. As notedearlier, it wascommon for married couplesto shareearly Attic kline sarcophagi. Moreover, although we know verylittle about Lucius Vibullius Hipparchus his existenceand activity areonly attested in a few inscriptions-Amelingmakes the provocative suggestionthat because he shareda "nomen"with Herodes'foster child VibulliusPolydeukion, he mighthave been Polydeukion's brother. This, if true,would make both Lucius Vibullius Hipparchus and Polydeukion sons of Herodes'cousin, Publius Aelius Vibullius Rufils, Athenian archon of 143/4(Fig. 17).92 If Ameling'ssupposition is correct,or if Hipparchusand Polydeukionwere otherwise as closelyrelated as theirnames suggest, the allusionto Spartanvotive reliefs on the frontregister of the coffinwould

91.The dateof Polydeukion'sdeath mid-130s)to Athens(late 130s) to Meyer1985; 1989; Ameling 1988. hasbeen a matterof heateddispute, Rome(where he becameconsul 92. Ameling1983, I, p. 170;II, with suggestionsranging from 146/7 ordinariusin 143 andremained for pp. 147, 169-170.The Vibulliiwere to the early170s. Polydeukion's severalyears); and, rather improbably, a familywhose name is perhapsmost portraituresuggests that he musthave it wouldsuggest that he hadstarted to readilyassociated with the cityof beenat leastin his mid-teensat the raisePolydeukion as a fosterchild ,but they appear also to have timehe died,and Tobin points out that yearsbefore he wasmarried with hadboth Athenian and Spartan Herodes'association with the boy childrenof his own.We do knowthat associations.Spawforth (1978, p. 258, appearsto havebeen a long-standing Herodeshad pupils (o,uXxa) whilehe n. 68) suggeststhat one of the most one. She thereforeargues that a dateof wasin AsiaMinor, since Philostratus famousmembers from the Corinthian deatharound 147 doesnot seemto (VS2.568) saysthat he sentthem to branch,Lucius Vibullius Pius, may correspondto the careerdates of Pergamonto hearAristocles speak; but havebeen a brother(or perhaps cou- Herodes.Indeed, one mightadd that this is a farcry from raising foster sin)of Herodes'maternal grandfather, sucha dateimplies that Herodes was children(Xtoo(pt,uo). For a summary LuciusVibullius Rufus. The Vibullius caringfor Polydeukion while traveling of the recentdebate over Polydeu- namewould, therefore, have been all overthe Mediterraneanworld from kion'sdate of death,see Tobin 1997, passedon to Herodesby his mother, AsiaMinor (where he servedas pp.231-234. Essentialcontributions VibulliaAlkia (see above, n. 26). Correctorof the FreeCities in the to the debateinclude Follet 1977; oo oo

TiberiusClaudius HIPPARCHUS

l

L. VibulliusRuis = ClaudiaAthenais ClaudiaAlkia

l

L. VibulliusHipparchus I VibulliaAlkia = TiberiusClaudius ATTICUS Herodes

P.Aelius Vibullius Rufus ClaudiaTisamenis = TiberiusClaudius Aristocrates HERODESATTICUS = Regilla Herodianos

LuciusVibullius ? Vibullius LuciusVibullius = Elpinike Regillus Athenais Bradua Unnamed Unnamed ClaudiusHerodes Polydeukion HIPPARCHUSII

Athenais

Figure 17. Familytree of Herodes ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 489

be evenmore meaningfill.93 Hipparchus would have been able to makethe sameclaim of descentfrom the Dioskouroithat is positedabove for Polydeukion,and we mighttherefore understand the Kephisiasarcopha- gus as figuratively"marrying" Spartan iconography with allusionsto the cultof Nemesisat Rhamnous,in a symbolicreflection of theliteral mar- riageof Elpinikeand Lucius Vibullius Hipparchus.94 Thereis a thirdreason to believethat this was not only the repository of Elpinike'sremains. L. Vibullius Hipparchus valued his relationship with Herodesso muchthat he later"inserted" himself into another, much more public,family monument, the nymphaeum at Olympia.Almost half a cen- turyafter Herodes commissioned this building, with portraits of his im- mediatefamily in the niches,Hipparchus had the complexrenovated, at whichtime he reinscribedstatue bases with his own name and with that of his daughterby Elpinike,Athenais.95 It would be fitting,then, if the man who hadso publiclyand proudly inserted himself into the familymonu- mentat Olympiahad also been laid to restwith that same family.

CONCLUSIONS

The unusualiconography of the Kephisiasarcophagus appears to com- memoratethe intimateconnection that Herodes' family enjoyed with the cityof Sparta,as wellas with the demeof Marathon,the Battleof Mar- athon,and the cult statueof Nemesisat Rhamnous.The sarcophagus achievesthese allusions by meansof an eclecticsynthesis of models,and is, in this respect,typical of sculpturalproduction during the Roman _ . q mplre. Contextand iconography, moreover, supportTobin's attribution ofthe Kephisiatomb to the familyof HerodesAtticus. They also provide evi- dencein favorof a suggestionthat she makes in passing,that this was the restingplace of Herodes'eldest daughter, Elpinike. The femaleportrait fromthe klinelid, the approximatedate of the sarcophagus,and the epi- graphicevidence that Herodeslaid his daughterto rest in Kephisia all suggestthat this was the repositoryfor herremains. In addition,the stylisticand thematic allusions to the cultstatue of Nemesispoint to the samehistorical event as does Elpinike's name, the Battle of Marathon,one of the definingevents in thehistory of Herodes'family. This convergence of iconographyand nomenclature would seem to offerstill more evidence for the identityof the deceased.It is alsopossible, though not beyond

93. Polydeukionhimself cannot n. 75) suggeststhat, before the young madethis impossible. It is temptingto have been buried in the Kephisia man'sdeath, Herodes had already thinkthat it wasonly after Poldeukion's sarcophagus:the female torso provides formulatedsome idea of the goddess's deaththat Herodes turned to the young the most compelling evidence against valueas a symbolof the unificationof man'sbrother in searchof a marriage . . r n any suc.R supposlton. .. .zere 1S no evl- two branchesof the family.It evenseems partnerfor his daughter. dence that Polydeukionmarried before possible,in viewof bothPolydeukion's 95. Athenais,who was presumably he died; indeed, he may have been too privilegedposition and his activityat namedafter her aunt, is clearly young to do so. Rhamnous,that Herodes originally identifiedon herbase at Olympiaas 94. Indeed, Polydeukion'sactivity plannedfor Polydeukion to marry "daughterof VibulliusHipparchus." with Herodes at Rhamnous(above, Elpinike,but that his untimelydeath See Bol 1984,pp. 101-102,134-141. ELLEN E. PERRY 49o doubt,that Elpinike shared this restingplace with herhusband, Lucius VibulliusHipparchus. Perhapsmore important than the attributionof this sarcophagusto specifichistorical figures, though, is therealization that a particularfamily historycan explain its decorativeprogram. This understanding necessarily modifiesour view of the"experimental phase" in Atticsarcophagi as it was identifiedby Koch.Koch noted correctly that, if sarcophagiare prefabri- cated,then we mustfind an explanationfor anomalousdecorative pro- grams.His explanationwas that certain subsidiary workshops, not associ- atedwith the maincircle of Atticproducers, offered for sale themes that provedto be unpopular.According to thisscenario, these workshops were drivenout of themarket within a decade,and their unpopular motifs were notto be foundagain on Atticsarcophagi. The exampleof the Kephisiasarcophagus suggests, however, that the weakpoint of Koch'sexplanation is the presumptionthat all sarcophagi wereprefabricated. In this instance,the unusualiconography appears to resultfrom the fact that the workwas commissioned.The patron, Herodes, musthave collaborated closely with the workshop on a designthat would commemoratethose legends and alliances that were especially significant to the family.The entirepoint, it wouldseem, was to producea unique piece a workof artthat would represent this family and no other.This observation,in turn,suggests the possibilitythat other anomalous Attic sarcophagiand cinerary urns from this decade were special commissions, ratherthan market failures.

REFERENCES

Alcock,S. 1993.Graecia capta: The Benndorf,0.1868."Romisches Grab in Landscapesof RomanGreece, Kephisia,"AZ36, pp. 35-40. Cambridge. Blumel, C. 1933. StaatlicheMuseen zu . 1994."Minding the Gapin BerlinV1, Berlin. Hellenisticand Roman Greece," in Bol, R. 1984. Das Statuenprogrammdes Placingthe Gods:Sanctuaries and Herodes-AtticusNymphaums (OlForsch SacredSpace in AncientGreece, 15), Berlin. S. Alcockand R. Osborne,eds., Bowie, E. L. 1974. "The Greeks and Oxford,pp. 247-261. Their Past in the Second Sophistic," Ameling,W. 1983.Herodes Atticus I- in AncientSociety, M. I. Finley,ed., II (Subsidiaepigraphica 11), , pp. 166-209. Hildesheim. Brilliant,R. 1984. VisualRarratives: . 1988."Der Archon Dio- Storytellingin Etruscanand Roman nysios:Zur Datierung einiger Art, Ithaca. attischerPortraits der mittleren Cartledge,P., and A. Spawforth.1989. Kaiserzeit,"Boreas 11, pp.62-70. Hellenisticand RomanSparta:A Tale Anderson,G. 1993.The Second of Two Cities,London. Sophistic:ACultural Phenomenon in Chapouthier,F. 1935. LesDioscures au theRoman Empire, London. serviced'une de'esse, Paris. Barringer,J. 1995.Divine Escorts: Cumont, F. 1942. Recherchessur le symbo- Nereidsin Archaicand Classical lismefune'rairedes Romains, Paris. GreekArt, Ann Arbor. Datsouli-Stavridi,A. 1993. rivgrlor Bartman,E. 1992.Ancient Sculptural 0f710 O72 eV,0£a1C8 Kvvov,oc'as, Copiesin Miniature,Leiden. Athens. ICONOGRAPHY AND THE DYNAMICS OF PATRONAGE 49I

Dierichs,A. 1990."Leda-Schwann- Kampen,N. 1988. "The Muted Other," Rhomiopoulou,K. 1997.ESr7wo- Gruppenin derGlyptik und ihre ArtJ47, pp. 15-19. pco,zzacxayAvn-;a TOV E0vexov monumentalenVorbilder," Boreas 13, Kleiner,D. E. E. 1988. "Roman AgoxacoRoycxovMova£c'ov, Athens. pp.37-50. FuneraryArt and Architecture: Richlin,A., ed. 1992.Pornography and Dohl, H. 1968. 'I)erEros des Lysippn Observationson the Significanceof Representationin Greeceand Rome, (diss.Gottingen). Recent Studies,"JRA1, pp. 115- Oxford. Elsner,J.1996."Image and Ritual: 119. Richter,G. M. A. 1966. TheFurniture Reflectionson the ReligiousAppre- . 1992. RomanSculpture, New of the Greeks,Etruscans, and Romans, ciationof ClassicalArt," CQ 46, Haven. London. pp.515-531. Koch, G., and H. Sichtermann.1982. Ridgway,B. S. 1981.Fifth Century . 1997."The Origins of the Icon: RomischeSarkophage, Munich. Stylesin GreekSculpture, Princeton. Pilgrimage,Religion, and Visual Koortbojian,M. 1995. Myth,Meaning, Robert,C. 1890.Die antiken Culturein the RomanEast as 'Resis- and Memoryon RomanSarcophagi, SarkophagreliefsII, Berlin. tance'to the Centre,"in TheEarly Berkeley. Sanders,J. 1992."The Early Lakonian RomanEmpire in theEast, Lacroix,L. 1956. "Un aspect meconnu DioskouroiReliefs," in @eRoRofxv: S. Alcock,ed., Oxford,pp. 178-199. de la legende d'Heraclessur une LakonianStudies in Honourof . 1998.Imperial Rome and Chris- monnaie de Pergame,"RBN 102, HectorCatling, J. Sanders,ed., tian Triumph,Oxford. pp.5-30. London,pp. 205-210. Ewald,B. C. 1999.Der Philosophals Lapatin,K. D. S. 1992."A Family . 1993."TheDioscuri in Post- Leitbild:IkonographischeUnter- Gathering at Rhamnous?Who's ClassicalSparta," in Palagiaand suchungenan romischenSarkophag- Who on the Nemesis Base," Coulson1993, Oxford, pp. 217- reliefs,Mainz. Hesperia61, pp. 107-119. 224. Follet,S. 1977."La datation de l'Archonte . 2001. Chryselephantine Shapiro,H. A. 1992."Eros in Love: Dionysios(IG II23968): Ses Statuaryin theAncientMediterra- Pederastyand Pornography in consequencesarcheologiques, litteraires, nean World,Oxford. Greece,"in Richlin1992, pp. 53- et epigraphiques,"REG 90, pp.47-54. Metzger, H. 1952. Cataloguedes monu- 72. Fullerton,M. 1990.TheArchaistic Style in mentsvotifs du Museed Adalia, Sichtermann,H., andG. Koch.1979. RomanStatuary (Mnemosyne Suppl. Paris. GriechischeMythen auf romischen 110),Leiden. Meyer, H. 1985."VibulliusPoly- Sarkophagen,Tubingen. Gantz,T. 1993.Early GreekMyth: deukion:Ein archaologischepi- Smith,R. R. R. 1998."Cultural Choice A Guideto LiteraryandArtistic Sources, graphischerProblemEall," andPolitical Identity in Honorific Baltimore. AM 100, pp.393-404. PortraitStatues in the GreekEast Gazda,E. 1980."A Portrait of . 1989. "Zu Polydeukion,dem in the SecondCentury A.D.," Polydeukion,"Bulletin of theMuseums Archon Dionysios, und W. Ameling JRS 88, pp.56-93. ofArt andArchaeology, University of in Boreas11, 1988,"Boreas 12, Spawforth,A. J. S. 1978."Balbilla, the Michigan3, pp. 1-13. pp. 119-122. Euryclids,and Memorials for a , ed. Forthcoming.The Ancient Moreno, P. 1995. "Erosa Mindo,"in GreekMagnate," BSA 73, pp.249- Art of Emulation:Studies in Artistic Lisippo:L'arte e lafortuna, 261. Originalityand Traditionfromthe P. Moreno, ed., Rome, pp. 166- . 1980."Spartaand the Family Presentto ClassicalAntiquity,Ann 167. of HerodesAtticus: A Reconsidera- Arbor. Nock, A. D. 1946. "Sarcophagiand tionof the Evidence,"BSA 75, Goette,H. R. 1993."AttischeKlinen- Symbolism,"AJA50, pp. 140-170. pp.203-220. Riefel-Sarkophage,"in Grabeskunst Palagia,O., and W. Coulson, eds. 1993. . 1985."Familiesat Roman derromischen Kaiserzeit, G. Koch,ed., SculpturefromArcadiaand Laconia Spartaand Epidaurus: Some Mainz,pp. 107-110. (Oxbow Monographs30), Oxford. ProsopographicalNotes," BSA 80, Graindor,P. 1915. "Tetede negredu Petrakos,V. 1981. "La Base de la pp. 191-258. Museede Berlin,"BCH39, Nemesis d'Agoracrite(rapport Spyropoulos,T. 1993."N£a yXvTcTa pp.402-412. preliminaire),"BCH 105, pp. 227- asoxTryaxa Tov ApXatokoytxov . 1930. Un milliardaireantique: 253. Movortov Tpt1rokrxs," in Palagia HerodeAttiqueetsafamille,Cairo. . 1986. "rJpoF>FyaSrarn5 F>aQs andCoulson 1993, pp. 257-267. Grunauer-vonHoerschelmann, S. 1978. Tov ayaByaSrosrnes Nryrorxs," in Swain,S. 1996.Hellenism and Empire: Die Munzpragungder Lakedaimonier, Archaischeund klassische griechische Language,Classicism, and Power Berlin. Plastik:Aktendes internationalen in the GreekWorld, A.D. 50-250, Joshel,S. 1992. aTheBody Female and Kolloquiumsvon 22.-25. April 1985 Oxford. the BodyPolitic: Livy's Lucretia and in Athen II (Klassischegriechische Tobin,J. 1997.Herodes Attikos and the Verginia,"in Richlin1992, pp. 112- Plastik), H. Kyrieleis,ed., Mainz, City ofAthens: Patronage and Conflict 130. pp.89-107. undertheAntonines,Amsterdam. 492 ELLEN E. PERRY

Tod,M. N., andA. J. B. Wace.1906. . 1983."Leda und der Schwan in A Catalogueof the SpartaMuseum, derkaiserzeitlichen attischen Plastik," Oxford. Boreas6, pp.168-196. Tschira,A. 1948-1949."Eine romische Wrede,H. 1977."Stadtromische Grabkammerin Kephissia,"S Monumente,Urnen, und Sarkophage 1948-1949,cols. 83-98. des Klinentypus in den beideners ten Turcan,R. 1978."Lessarcophages Jahrhundertenn. Chr.," B 1977, romainset le problemedu pp.395-431. symbolismefuneraire,"ANRW Zagdoun,M. 1989.La sculpture II.16.2,pp. 1700-1735. archaisantedans l'art helle'nistiqueet Wiegartz,H. 1975."Kaiserzeitliche dansl'art romaindu haut-empire, Relief-Sarkophagein der Niko- Athens. laoskirche,"in Myra:Eine Iykische Zanker,P. 1988. The Power of Images Metropole,J. Borchhardt,ed., in theAgeof Augustus, trans. pp. 161-251. H. A. Shapiro,Ann Arbor. .1977."ZuProblemen einer . 1995.The Mask of Socrates: Chronologieder attischen TheImage of theIntellectual in Sarkophage,"S1977, pp. 383- Antiquity(Sather Classical Lectures 388. 59), trans.H. A. Shapiro,Berkeley.

Ellen E. Perry HOLYCROSS COLLEGE DEPARTMENTOF CLASSICS WORCESTER,MASSACHUSETTS OI6IO eperry@holycross .edu