Augustus to Jakob Munk Hjte AarhusUni versi t yPress Roman Imperial Statue Bases fromCommodus Aarhus Studies inMediterranean Antiquity(ASMA)VII ASMA is a series which will be published approximately once a year by Te Centre for the Study of Antiquity, University of Aarhus, Denmark. Te Centre is a network of cooperating departments: Greek and Latin,Classical Archaeology, History, and the Faculty of Teology. Te objective of the series is to advance the interdisciplinary study of Antiquity by publishing articles, e.g.,conferencepapers,orindependentmonographs,whichamongother things refect the current activities of the centre. J A KOBMUNKHJ TEROMANIMPERIALSTATUEBASESFROMAUGUSTUSTOCOMMODUSActa Jutlandica LXXX:2Humanities Series 78aAARHUSUNIVERSITYPRESSRoman Imperial Statue Bases Jakob Munk Hojte and Aarhus University Press 2003 Cover: Lotte Bruun Rasmussen Photo: Jakob Munk Hojte, Caserna dei Vigili, Ostia, Statue bases on a podium in the sacellum.Typeset with Trajan (cover) and Mignon (body)ISBN 87 7934 906 4Aarhus University PressLangelandsgade 177DK-8200 Aarhus NFax: (+43) 8942 338073 Lime WalkHeadington, Oxford OX3 7ADFax: (+44) 1863 730 079Box 311Oakville, CT 06779Fax: (+1) 860 943 9468

PrefaceTis book is an altered and revised version of my PhD dissertation defended at the University of Aarhus in May 2001. Due to other obligations and a certain fatigue and nausea at the thought of taking another swing at the large corpora (they do tend to have a rather musty odour), the manuscript was put in the drawer.Earlierthisyearitwouldremainhiddennolonger.Overthesum-mer and during a stay in Rome in the autumn, the tedious task of checking the entries in the catalogue and editing the text was carried out. Two major changes have been made. First of all, the three separate papers that formed part of the dissertation: Te Epigraphic Evidence Concerning Portrait Statues of Hadrians Heir L. Aelius Caesar, Imperial Visits as Occasion for the Erection of Portrait Statues?, and Te Statue Bases of Claudius. A Reassessment of Te PortraitureofClaudiusbyM.Stuarthavebeenpublishedelsewhere,and arethereforenotincludedhere(seebibliography).Tismeantthatagood number of cross-references had to be sorted out and text supplied where nec-essary. Secondly, the catalogue of statue bases has been revised and updated. A number of inscriptions included in the dissertation, as pointed out by the assessment committee, demonstrably did not belong to statue bases, and have consequently been excluded. In addition, a number of inscriptions that seemed too uncertain to include have also been removed. On the other hand, more material has been published in the intervening period. Te catalogue has been updated to include the bases mentioned in Supplementum Epigraphicum Grae-cum 2000 and LAnne pigraphique 2001. In the dissertation only a minimum of information about the individual bases went on paper, while the bulk was storedonaCD-ROM.HereIhavechosentoincludemoreinformationin theprintedtext,whichhasresultedinarathervoluminouscatalogue.Itis vvii.ci3my hope that the expanded format will improve its usefulness and encourage others to make use of the collected data, which I believe holds potential far beyond what has been covered in this book. Since my interest in Roman imperial statue bases was frst aroused by reading MeriwetherStuartsdissertationfrom1938ontheportraitureofClaudius, which includes the frst attempt at systematically compiling and analysing the epigraphicalevidencefromstatuebases,manyfriendsandcolleagueshave commentedandmadevaluablesuggestionsthathavegreatlyimprovedthe outcome.Somerequireparticularmention:IttaiGradelforinspirationand rewarding discussions. His encouragement is one of the primary reasons why the study has been brought to completion. Niels Hannestad and Lise Hanne-stad,mysupervisor,forvaluablehelpandadvicebothduringandafermy time as a PhD student. Niels interest in Roman sculpture and imperial por-traiture in particular initially got me started on this project. Robert Fleischer, my external supervisor, for making my much too short stay in Mainz pleasant and rewarding. More importantly for his comments on the part of my original project,whichwillunfortunatelyhavetostayinthedrawerforsometime yet, namely an unfnished study of the statue bases for the Hellenistic kings. TeexternalmembersoftheassessmentcommitteeGezaAlfldyandJane Fejfer, who gave precise criticism and good directions for both possible and required improvements far beyond the call of duty. I have tried as best I could to follow their recommendations. Finally and most dearly I want to thank my family, who have tirelessly accompanied me on countless journeys in (ofen futile) search of statue bases. TebookwasmadepossiblebygenerousfnancialsupportfromtheUni-versity of Aarhus, the University of Aarhus Research Foundation, Elizabeth Munksgaard Fonden, Landsdommer V. Gieses Legat, and the Danish Research Council for Humanities. Arhus, December 2004 Jakob Munk Hjte 6vom.imvivi.is1.10in.sisContentsPvii.ci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Lis1oiFic0vis.uT.niis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11I1vou0c1io. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13TvvisoiMo0mi1s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Identifcation of statue bases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Te language of the inscriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Types of statue base. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Literary testimony for imperial statue bases and inscriptions . . . . . . . 40Statue types and materials used for imperial statues . . . . . . . . . . . . . . . . 43Te cost of imperial statues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Damnatio memoriae and the reuse of statue bases. . . . . . . . . . . . . . . . . . 36D.1ic1uiIscviv1iosivomImvivi.iS1.10iB.sis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Imperial nomenclature and honorifc titles . . . . . . . . . . . . . . . . . . . . . . . . . 63Other dating criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70Dating by negative evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72Reliability of the dating criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74Dating accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Dates chosen for dedicating imperial statues . . . . . . . . . . . . . . . . . . . . . . . 78co1i1s7TuiAvviic.niii1voi1uiEviuicioi1uiS1.10iB.sis1o1uiEx1.1Pov1v.i1s .81TuiGiocv.vuic.iDis1vin01iooiImvivi.iPov1v.i1S1.10is . . . . . . . . . . . . . . . . . . . . . . . . . . 83Te geographical distribution of extant imperial portraits . . . . . . . . . . 86Te geographical distribution of statue bases . . . . . . . . . . . . . . . . . . . . . . . 88Te number of sites and the number of bases per site. . . . . . . . . . . . . . . 103Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109S1.10isDiuic.1iuBiiovi.uAi1iv.Riic .123Pre-accessional dedications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Posthumous dedications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132Occ.siosiovEvic1icImvivi.iS1.10is . . . . . . . 143Accession. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144Jubilees (decennalia and vicennalia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137Imperial visits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139Patterns of chronological distribution during a reign. . . . . . . . . . . . . . . 163Diuic.1ovsoiRom.Imvivi.iS1.10is . . . . . . . . . . . 167Statues dedicated by communities or their executive bodies . . . . . . . . 168Private dedicators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171Public or private: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179Corporations as dedicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181Military units as dedicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182Statue bases without dedicators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184Regional diferences and developments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183Coci0sio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189Biniiocv.vuv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1938vom.imvivi.is1.10in.sis C.1.ioc0i. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213I1vou0c1io1o1uiC.1.ioc0i. . . . . . . . . . . . . . . . . . . . . 217What is included:. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217Sorting system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218Geography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218Chronology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222Distribution maps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222Histograms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222Annvivi.1ios.uBiniiocv.vuv iovC.1.ioc0i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223C.1.ioc0ioiS1.10iB.sis Avv.ciuAccovuic1oEmvivov . . . . . . . . . . . . . . . . . . . . . . . . 229Augustus 229 Tiberius 263 Caligula 288 Claudius 294 Nero 319 Galba 319 Otho 330 Vitellius 330 Vespasian 330 Titus 344 Domitian 334 Nerva 363 Trajan 373 Hadrian 404 Antoninus Pius 466 Lucius Verus 309 Marcus Aurelius 331 Marcus Aurelius or Lucius Verus 369 Avidius Cassius 371 Commodus 371S1.1is1ic.iA.ivsis, Emvivovs(T.niisSE1-8) .391Augustus 391 Tiberius 392 Caligula 393 Claudius 394 Nero 393 Galba Otho and Vitellius 396 Vespasian 397 Titus 308 Domitian 309 Nerva 600 Trajan 601 Hadrian 602 Antoninus Pius 603 Lucius Verus 604 Marcus Aurelius 603 Commodus 606S1.1is1ic.iA.ivsis, Giocv.vuic.i (T.niisSG1-i) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 607Italy 607 Northern provinces 609 Gaul 611 Spain 613 Western North Africa 613 Greece 617 Asia Minor 619 Eastern provinces 621co1i1s9S1.1is1ic.iA.ivsis, C omv.v.1ivi (T.niisSC1-i). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623 Cuvooio cic.iDis1vin0 1io(uis1o cv.ms) (Fics. C1-io) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 633 Augustus 633 Tiberius 633 Caligula 634 Claudius 634 Nero 633 Vespasian 633 Titus 636 Domitian 636 Nerva 637 Trajan 638 Hadrian 638 Antoninus Pius 639 Lucius Verus 639 Marcus Aurelius 640 Commodus 641 All emperors 642 East-West comparison 642 Gio cv.vuic.iDis1vin0 1io(uis1vin0 1iom.vs) (Fics. G1-1o). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643 Augustus 633 Tiberius 644 Caligula 643 Claudius 46 Nero 647 Vespasian 648 Titus 649 Domitian 630 Nerva 631 Trajan 632 Hadrian 633 Antoninus Pius 634 Lucius Verus 633 Marcus Aurelius 636 Commodus 637 All emperors 638 10vom.imvivi.is1.10in.sisList of Figures and TablesFig. 1.Statue base in the National Museum in Athens with oval depression for a marble statue. Fig. 2.Base for a statue of Claudius (Claudius 87) in the Athenian Agora. Fig. 3.Statue base for Trajan in Delphi (Trajan 110). Fig. 4.Built-up bases in a building adjoining the forum in Lucus Feroniae. Fig. 3.Marble slab from a built-up base for Trajan (Trajan 37) in Lucus Feroniae. Fig. 6.Statue base for Marcus Aurelius (Marcus Aurelius 20) from Fidenae. Fig. 7.ExedraoppositethetheatreinEmeritaAugusta,with statues of the imperial family placed in niches in the wall. Fig. 8.Consoles carrying statues of Antoninus Pius, Marcus Aure-lius and Lucius Verus on the colonnaded street in Apamea ad Orontem. Fig. 9.Comparison between extant portraits and bases according to region. Fig. 10.Damnatio memoriae. Fig. 11.Te number of bases per year in Italy. Fig. 12.Te relative importance of the statue bases in Italy. Fig. 13.Te relative importance of the statue bases in the northern provinces. Fig. 14.Te relative importance of the statue bases in Gaul. Fig. 13.Te relative importance of the statue bases in Spain. iis1oiiic0vis.u1.niis11Fig. 16.Terelativeimportanceofthestatuebasesinwestern North Africa. Fig. 17.Te relative importance of the statue bases in Greece. Fig. 18.Te number of bases per year in Asia Minor. Fig. 19.TerelativeimportanceofthestatuebasesinAsia Minor. Fig. 20.Te number of sites according to region and the number of bases per site. Fig. 21.Tepercentageofsiteswithstatuebasesaccordingto region. Fig. 22.Te number of sites in relation to the number of bases for individual emperors. Fig. 23.Te number of precisely dated bases according to year of reign. Fig. 24.Public and private dedicators. Fig. 23.Map of the Roman Empire. Tables SE 1-SE 48.Statistical analysis according to emperor.Tables SG 1-SG 24. Statistical analysis according to region.Tables SC 1-SC 24.Comparison between emperors and regions.Figs. C 1-C 20.Chronological distribution. Histograms.Figs. G 1-G 16.Geographical distribution. Distribution maps.12vom.imvivi.is1.10in.sisIntroductionWhen the senate voted him [Didius Julianus] a statue of gold, he declined to accept it, saying: Give me a bronze one, so that it may last; for I observe that thegoldandsilverstatuesoftheemperorsthatruledbeforemehavebeen destroyed, whereas the bronze ones remain. In this he was mistaken, for it is virtue that preserves the memory of rulers; and in fact the bronze statue that was granted him was destroyed afer his own overthrow. DioC.ss. 74.14.2a In a short perspective the refections of Didius Julianus and Dio Cassius on the preservation of ones memory for posterity were to some extent correct. Until AD 193, the year Didius Julianus for a brief period succeeded in bribing his way to the purple by ofering a large sum of money to the praetorians, the Roman Empire had witnessed a long period of stability. Since the murder of Domitian in AD 96 the emperors, even if they were not equally liked, at least had the privileges of choosing their own heir, dying of natural causes and being elevated to divinity. Te murder of Commodus some months previously had ended this era and once again brought the Empire to the verge of civil war. It is not entirely clear whether Didius Julianus, in Dios rendering of the speech, is supposed to be referring to the statues of his two immediate predecessors, theunfortunateemperorsCommodusandPertinax,ortothoseofformer emperors in general; but being a virtuous ruler was apparently no guarantee against having ones statues made of precious metals ending up in the melting pot, and such images generally seem to have had a rather short existence.' Dio 1.Pekary 1983, 66-67 and below p. 47. i1vou0c1io13 Cassius and his audience, knowing the fate of Didius Julianus, could in hind-sight of course see the folly of his argument. In a longer perspective, however, it was not necessarily the kind of virtue advocated by Dio that would preserve the memory of an emperor. Nero, who had been hated with good reason by the senatorial aristocracy to which Dio belonged, seems to have been rather popular in the late fourth century to judge from the frequency of his portrait on the contorniats of the period.` When it comes to the preservation of the memory of a ruler by means of statues, which was evidently the intention of such monuments to judge from the speech by Didius Julianus, neither bronze nor virtue has proven efective. Instead,thesinglemostimportantfactorforpreservationofportraitsand statues for posterity was whether they were made of stone. Bronze seems to havebeenthepreferredmaterialforhonorarystatuesinmostpartsofthe RomanEmpire,butofthecountlessbronzestatuesofemperorsmadein antiquity only a minute fraction have been preserved because their value as commemorative monuments soon fell below the relatively high scrap value of bronze.` Consequently they were melted down for other uses, the same fate that has overcome practically all portraits in precious metal. To a large extent onlybronzestatuesplacedoutofhumanreachbyunusualconditions,like those aforded by the eruption of Vesuvius or landslides like that in Boubon (or statues lost at sea or in rivers during transport) have survived. Te scrap value of marble statues was much lower, and they thus stood a larger chance of survival, although lime kilns have taken their share of these too. Tis process of selection has profoundly infuenced the study of imperial portraits, which naturally must begin with the preserved specimens i.e. the marble portraits. Te issue of how these relate to those originally in existence has unfortunately not received the attention it deserves. Te primary aim of nearly all studies of imperial portraits has been to compile the genuine specimens, establish a typology,anddatetheintroductionofnewportraittypes.Tothisendthe numismatic evidence has proven especially useful, since the typology of the coin portraits regularly corresponds to that of the portraits in the round, and 2.Mittag 1999, 128-133. 3.Lahusen & Formigli (2001), in their recent monograph on bronze portraits, include 43 portraits of emperors from the frst two centuries AD. In contrast, more than 1,000 mar-ble portraits are known. 4.For the contexts in which bronze portraits have been, see Lahusen & Formigli 2001, 433-439. 3.Pfanner 1989, 162; Rose 1997b, 108-120; Pollini 1987, 8-17. Tis approach is prevalent in the series Das rmische Herrscherbild and in most museum catalogues. 14vom.imvivi.is1.10in.sisthe coin legends ofen supply an exact date. Because of the strong focus on the extant portraits, which ofen have no archaeological context, and because of the remarkable advances made within the feld of portrait studies during the last two generations, other archaeological, literary and epigraphical sources relatedtothededicationofimperialstatueshavebeenrelegatedtoavery subordinate position in nearly all studies of imperial portraits. It is the aim of this study of one of these documentary sources namely the epigraphical evidence from the statue bases to compensate for this lack of research and show that the study of statue bases is relevant if not crucial for the understand-ing of Roman imperial portrait statues. In his signifcant study Te Epigraphic Habit in the Roman Empire, Ramsey McMullen observed that Papyri and ostraca from Roman Egypt survive in sumcient numbers to invite statistical analysis and thus to teach us something out of the numbers themselves that is not evident in the body of any single text.' Tis applies to the statue bases of Roman emperors as well. By them-selves the inscriptions from each individual base yield information about the specifc statue once placed on the base, but since these almost invariably have beenlostormovedbeyondrecoveryoftheiroriginalcontext,theimmedi-ate testimony has little direct value for the study of imperial portraits. Of the limited number of portraits that have been found with their accompanying inscriptions, only the relief from Ostia dated AD 160 showing Lucius Verus in the portrait type, thought to have been created on his accession a year later, has modifed the chronology of an emperors portrait types during the period under consideration here." By systematically compiling the statue bases and using them as statistical data, however, they can reveal valuable information about where imperial statues were erected, when, by whom and for what rea-son; questions that cannot be answered by studying the extant portraits.6. Te term statue base will be applied below to all types of inscribed monuments intended to carry a sculpted representation of the emperor, see p. 19. In the text, statue bases are referred to as numbers in the catalogue. 7.MacMullen 1982, 234. 8.Fuhrmann 1939, 294-302 = (Lucius Verus 16). Other frst and second century portraits of emperors that have been found with accompanying inscriptions: Herculaneum (Tiberius 13 (theatre); Claudius 8 (basilica); Lanuvium (Claudius 9 uncertain); Misenum (Vespasian 8 [augusteum]); Neilly-le-Real (Augustus 73 [bronze bust]); Lepcis Magna (Augustus 114, Tiberius 76, Claudius 73 [Temple of Roma and Augustus]); Olympia (Antoninus Pius 201, Lucius Verus 108 [Nymphaeum of Herodes Atticus]); Aphrodisias (Claudius 113, Nero 40, 41, [Sebasteion]; Domitian 37 [theatre]); Boubon (Lucius Verus 113); Perge (Hadrian 373 uncertain); Philadelphia (Commodus 87 [relief ]); Cyrene (Tiberius 109 [Strategeion]). i1vou0c1io13Te epigraphical evidence from statue bases has been discussed before in connection with imperial portraits. Te frst person to systematically compile the statue bases for an emperor was Meriwether Stuart, who collected all the documentary sources relating to statues of Claudius in 1938.' Soon aferwards followed studies along similar lines concerning the other Julio-Claudian em-perors'" and the family of Augustus.'' Te statue bases of empresses have been systematicallycompiledforSabina'`andJuliaDomna,'`andsohavethose for the emperors from the mid-fourth century AD to the end of antiquity.' Although the results of these studies were noteworthy, they have had limited impact for two reasons. Firstly, a direct relationship between the chronological distribution of the statue bases and that of the extant portraits has not been established. Secondly, the studies have been too scattered chronologically to ofercomparativeevidencethatcouldrevealanyconsistentpatternsinfre-quency and geographical and chronological distribution of the statue bases. It is characteristic that the recent works on the portraits of Augustus,' Ca-ligula,' Hadrian'' and the Antonine princes'" that do include investigations of the epigraphical evidence make limited use of it for questions relating to theextantportraits.Otherstudieshavedealtmoreorlessthoroughlywith the statue bases relating to portraits in a certain setting or region. Examples of these are the excursus in Inan and Rosenbaums study of the portraits from Asia Minor,'' the regional studies of statue bases in Conventus Tarraconensis and in Venetia et Histria,`" and recent works concerning Julio-Claudian statue groups`' and imperial women in the Greek East.`` To overcome the obstacle of lack of comparative material, this study com-piles the imperial statue bases from a long continuous period. It includes all the emperors from Augustus to Commodus, a period of approximately 230 9.Stuart 1938. 10.Stuart 1939, 601-617. 11.Hanson & Johnson 1946, 389-400. 12.Carandini 1969. 13.Fejfer 1983, 129-138. 14.Stichel 1982. 13.Boschung 1993a. 16.Boschung 1989. 17.Evers 1994. 18.Fittschen 1999. 19.Inan & Rosenbaum 1966, 42-33. 20.Alfldy 1979, 177-273; Alfldy 1984. 21.Rose 1997a; Boschung 2002a. 22.Hahn 1994. 16vom.imvivi.is1.10in.sisyears that covers a large part of the principate including the second century AD,whentheproductionofportraitsofemperorsreacheditsheight.Te compiled corpus of statue bases comprises 2,300 monuments from nearly 800 diferent sites throughout the Roman Empire and beyond. Te broad chrono-logical perspective of this large sample shows that the statue bases provide a consistent and reliable picture of the geographical and chronological distribu-tion of imperial statues in antiquity which challenges previous assumptions regardingtheprinciplesthatgovernedtheerectionofimperialstatuesina number of ways. i1vou0c1io17Types of MonumentsPortraitstatuesinantiquitywerealmostinvariablyplacedonsomesort ofbasethatactedasasupportforthestatue,lifingitofthegroundor creatinganarchitecturalsettingforit.Tetermstatuebasetraditionally describesafree-standingmonumentconsistingofoneormoresquareor roundblocksofstonesumcientlylargetocarryastatue,anderectedfor thispurposealone.Inthefollowing,however,thetermwillbebroadened tosignifyanymonumentintendedtocarryathree-dimensionalrepresen-tation of approximately life size and larger. Tis wider defnition may seem awkward when applied to singular monuments, such as an arch or a niche holdingastatue,butitmayproveusefulindescribingthefunctionofthe variety of monuments, which all served the same purpose despite their dif-ferent appearance. Identifcation of statue bases All the monuments compiled in the catalogue of imperial statue bases have one common feature. Tey carry an inscription that identifes the emperor whosestatuewasplacedonthebase.Tiswas,naturally,notarequire-ment for a base, but merely the only means by which we can identify them today. Imperial statues could be placed on uninscribed bases, but since im-perial portrait statues have only very rarely been found together with their accompanyingbases,thesearepracticallyimpossibletoidentify.Tein-scriptionsasageneralrulefollowthestandardisedpatternfortitulihono-1 vvisoimo0mi1s19rarii,`` allowing us with a high degree of probability to identify monuments as statue bases from their inscriptions alone. Tis is useful because numer-ousinscribedmonumentsareinadequatelydescribedinpublications.In particular, the editors of the early corpora of inscriptions generally showed littleornointerestinthephysicalformofthemonumentsonwhichthe textswereinscribed,orinthecontextinwhichtheyhadbeenfound;and since many of the inscriptions have later disappeared, such information has been irrevocably lost. Even if satisfactorily published, the state of preserva-tion of the monument ofen does not allow unambiguous identifcation as a statue base from the physical properties alone; either because the inscription hasbeenremovedfromthemonumentandreusedinanothercontext,or because only a fragment of the monument without any recognisable features has survived. InhisstudyoftheportraitsofClaudius,Stuartestablishedtwocriteria for identifying statue bases: One, whenever an inscription employs a dative formula in Latin or an accusative in Greek and is cut on a stone reliably de-scribed as a statue base or as part of an arch or other monumental pedestal, there can be no doubt of the portrait character of the inscription. Two, when-ever a description of the stone on which an inscription is cut is not available, the dative case of the imperial name in Latin, or the accusative in Greek, is presumptiveproofoftheportraitcharacteroftheinscription.`Tefrst criterion encompasses approximately 1,300 inscriptions, or well over half of the monuments in the catalogue. Tis criterion is nearly foolproof. It should be noted, however, that for lack of a common terminology for diferent types of monument, descriptions may be misleading. One example, emphasised by Benjamin and Raubitschek, concerns a number of monuments from Athens describedasstatuebasesbytheeditorsofInscriptionesGraecae,whichon closer examination proved to be altars.` 913 inscriptions in the catalogue (40 of the total) have been identifed as being or pertaining to statue bases according to Stuarts second criterion, 23.ForGreektitulihonorarii,seeGerlach1908;Larfeld1914,432-436;Klafenbach1966, 63-69. For Latin see Cagnat 1914; Kajanto 1971, 3-19. Tituli honorarii consist of two ele-ments:thenameofthehonorandandthenameofthededicator.Tothisbasicscheme can be added a variety of information about the nature of the dedication, the date or the dedicators reason or motivation for erecting the monument. For an interesting view of the development of Latin honorifc inscriptions, see Salomies 1994, 63-106. 24.Stuart 1938, 13-14. 23.Benjamin & Raubitschek 1939, 63-83. 20 vom.imvivi.in.siswhich as he himself pointed out is not unfailingly accurate.` Tese inscriptions fall in two groups. Te frst and largest consists of about 600 monuments that lack description altogether, or are described in terms so vague as to preclude determination of the type of monument involved. Ofen this is no longer pos-siblebecauseofthepresentstateofthemonument;butinmanyinstances, especiallywithregardtotheinscriptionsintheoldercorpora,itissimple negligence on the part of the editors. Te other group consists of inscriptions cut on what is described as tabulae or plaques of stone. Tese may have been deliberatelysawnfromalargerblockofstoneforsecondaryuse,evenfor displayinmuseumsasisthecasefortheinscriptionsintheLapidariumof the Vatican Museum, but normally they were meant to be am xed to built-up bases or otherwise non-monolithic structures. Te problem with the second criterion is that monuments other than statue bases employ inscriptions that follow exactly the same schema. Tis is especially pronounced for Latin in-scriptions, where the dative case for the name of the emperor was used not only for statue bases, but also for altars, milestones and building inscriptions. StatuebaseswithGreekinscriptionscanmoreeasilybedetected,because tomyknowledgenoothertypesofmonumentusetheaccusativeformula. Whenever a reliable description of the stone is lacking, it is necessary to take into consideration all the available information concerning dimensions, layout of the inscription on the stone, size of the letters and content of the inscrip-tion;and,basedoncomparisonwithothermonumentsidentifedasstatue bases,ineachcasetojudgewhetherthemonumentcouldhaveservedasa statue base.`' Tis method, of course, is open to mistakes, and a number of entriesinthecatalogueundoubtedlyshouldnothavebeenincluded,while some statue bases probably have been unjustly excluded. Given the number of statue bases, however, this inaccuracy should not have any impact on the conclusions drawn from the material unless the fgures involved are exceed-ingly small, in which case caution at any rate should be taken. Aberrant formulations Whilst close to 90 of the inscriptions in the catalogue follow the standard pattern for honorifc inscriptions described above, Stuarts criteria do not take 26.Stuart 1938, 14. 27.For a discussion of the construction of statue bases, see Alfldy 1984 and Fabre, Mayer & Roda 1984, 11-21. 1 vvisoimo0mi1s21into account the remaining 10 that belong either with certainty, or with a high degree of probability, to monuments that served as statue bases but for avarietyofreasonsemployaberrantformulations.Sincetheseinscriptions have not been discussed before in the general context of imperial statue bases, it is necessary to present the various types and the reason for their inclusion in the catalogue. Te use of the dative case in Greek Te commonest deviation from the standard formula for honorifc inscrip-tionsonstatuebasesistheuseofthedativecaseinGreekforthename formula of the emperor. Te 133 examples of this can be divided into three categories. Firstly, the form of the Greek inscriptions could be favoured by the Latin practice of using the dative case. Tis is most obvious in the bilin-gual inscriptions, where both the Latin and Greek texts normally employ the same case (p. 27). It is also frequently found among dedications with Greek inscriptions erected in predominantly Latin-speaking areas in the West (An-toninus Pius 10, 13; Marcus Aurelius 1, 2, 9), as well as in cities in the East with a strong presence of Latin speakers. Secondly, buildings dedicated in the name of the emperor in Greek inscriptions take the dative case, and monu-ments that served as statue bases but whose physical form resembles that of buildings,suchasarchesandcitygates,thereforealwaysfollowthepattern for building inscriptions and employ the dative case. One monument, an ar-chitrave in Perge with an inscription for Claudius in the accusative case, was long believed to pertain to an arch (Claudius 143).`" New excavations on the site have shown that the arch according to the newly found inscription was insteaddedicatedtoDomitianandthedeifedVespasianandTitus(Vespa-sian70;Titus61;Domitian61).TenatureofthemonumentforClaudius remains obscure. Even monuments that were much closer in form to statue bases,andwhichservednootherpurpose,likethepillarontheAthenian Agora originally constructed as a monument for Attalos II of Pergamum but later re-dedicated to Tiberius (Tiberius 89),`' could employ the dative case. Finally, the dative case could be used to give the dedication religious overtones and connotations, since the dative case was generally reserved for statues of 28.Merkelbach & Sahin 1988, 110, no. 10. 29.Vanderpool 1939, 86-90. 22vom.imvivi.is1.10in.sis gods as distinguished from honorifc statues with inscriptions in the accusa-tive case.`" One particularly interesting example of this is found among the statues erected in the temple consecrated to Vespasian in Kestros in Cilicia. Te cult statue of Vespasian placed centrally against the back wall of the cella stood on a base with an inscription in the dative case, as would be expected for a cult statue (Vespasian 76). On both sides of the cult image stood statues of his sons, and later those of successive emperors lined the side walls of the cella;butunliketheoriginalcultstatue,allofthesewereaccompaniedby inscriptions in the accusative case. Inonlytwoinstancesdowefndtheinfuencereversedintheformof the accusative case used in Latin inscriptions (Augustus 194; Antoninus Pius 273). Both bases stem from the interior of Asia Minor, where neither Latin nor Greek inscriptions had long traditions. Te inscription for Augustus from Lystra is described as a pedestal, and although consecravit in l. 3-6 is unusual for a statue base, it has parallels in Asia Minor (Antoninus Pius 217). Te other inscription lacks description. Two further such monuments for Caracalla, Julia SoaemiasandJuliaMammaea,whichbeyonddoubtservedasstatuebases, are attested in Pergamum.`' Te use of the nominative case Te name of the emperor in the nominative case could be used both in Latin and in Greek as a label under a statue that formed part of a large ensemble of statues with a common dedicatory inscription. An illustrative example is the numerous bases in Eleusis, which may have been placed on the arches outside the entrance to the sanctuary or, in analogy to the Nymphaeum of HerodesAtticusinOlympia,couldhavebeenplacedonthenymphaeum identifed next to the southern arch (Hadrian 247; Marcus Aurelius 191).`` Otherexamplesarethegroupofstatuebasesfordeifedemperorserected inTuggainthethirdcenturyAD(Augustus118;Vespasian42-43;Tra-jan 91; Hadrian 160; Marcus Aurelius 170), and the labels under the reliefs placed between the columns in the two upper storeys on the two porticoes 30.Mitford 1947, 224; Veyne 1962, 49-98; Price 1984, 179. 31.Wiegand 1932, 34-33, no. 7 a-b. 32. Clinton(1989,36-68)proposestwoarcheswithimperialstatues.Fittschen(1999,122-126) suggests the nymphaeum as a possible location for the bases. For the nymphaeum at Olympia, see Bol 1984. 1 vvisoimo0mi1s23fankingtheprocessionalwayfromthepropylontotheimperialtemplein Aphrodisias (Claudius 113; Nero 40, 41, and possibly 39). 22 further monu-mentswiththeemperorsnameinthenominativehavebeenincludedin the catalogue. Tese have predominantly been described as statue bases or arches (Trajan 72). Te exceptions are inscriptions with the name of Tiberius inAenonaandChalkis(Tiberius48,91),whichentirelylackdescription. Te portrait character of the latter is implied by the presence of an identical monumentforGaiusCaesar.``Itisnotentirelyclearwhythenominative casewaschosenforthesebases,buttheymayoriginallyhavebelongedto agroupofbasessuchastheoneinEleusis.Naturally,thebasesmayhave supported objects other than statues of the emperors, and the emperor may in fact have been the dedicator. Tis might be the case for three bases found inthetheatreinLepcisMagna(Hadrian149-131).Exceptforthebasefor Tiberius in Iader (Tiberius 49), erected by the seventh and eleventh legions, noneofthestatuebaseswiththenameoftheemperorinthenominative case mention the name of the dedicator. Te use of the ablative case TenameoftheemperorintheablativecaseisfrequentlyusedinLatin building inscriptions to indicate the date of construction, and this must be its function in the inscription on the city gate erected by Sex. Iulius Frontinus in Hierapolis (Domitian 34). Te gate, however, probably also carried a portrait statueofDomitian.Teinexplicableuseoftheablativecaseonabasefor Caligula in Narbonensis (Caligula 6) might be a simple spelling error.` Four inscriptions from Tamugadi that employ the ablative case have been included in the catalogue (Antoninus Pius 143-147; Marcus Aurelius 131). All are de-scribed as bases, and have been accepted as such by Zimmer in his study of the statues bases from the forum of Tamugadi.` Te three inscriptions for Antoninus Pius mention a paved street constructed from public funds, and we cannot be absolutely certain that the statues placed on the bases were in fact those of the emperor. 33.IG XII, 9, 940. 34.Caesare for Caesari in l. 1. 33.Zimmer & Wesch-Klein 1989, 78, no. T 21; 82-83, T 43. 24vom.imvivi.is1.10in.sis Inscriptions with deviating compositions Te inscriptions normally begin with the name of either the emperor or the dedicator. Tere are, however, exceptions to this rule. Te most important are i tu the 38 Greek inscriptions headed by the formula Ayo0p pi. All except one of these a marble gable with a bust of Commodus that continues with ar u and the genitive case (Commodus 89)` otherwise follow the standard pattern for honorary inscriptions. Tar and the genitive case is found on six further monuments all described as statue bases (Augustus 173; Tiberius 124, 134; Caligula 26; Vespasian 62). In these instances the bases most probably carried imperial statues, but otherwise the formula most commonly appears inbuildinginscriptionsonbuildingsdedicatedonbehalfoftheemperor.`' ar o normally indicates altars,Pro salute or the Greek equivalent u omtpibutinafewcasestheformulationwasemployedonarches(Hadrian411; Marcus Aurelius 133; Lucius Verus 83). One monument described as a base, althoughitwasmorelikelyanaltar,carriedaninscriptionspecifyingthat the dedication consisted of both imagines and an ara (Marcus Aurelius 103; Lucius Verus 69). Te altar probably carried the inscription, and the statues stood in the immediate vicinity. In Stratonicaea an inscription begins by stat-ing the reason for the dedication apparently that Hadrian had carried out a successful hunt in the citys territory (Hadrian 360). Other bases begin with a dedication to a deity in the dative case: Aphrodite at Paphos (Tiberius 148-130) and Artemis at Ephesus (Trajan 144), or they are joint dedications to one or more gods and the emperor (Hadrian 362; Antoninus Pius 103). Te language of the inscriptions Latin was the omcial language of the administration of the Roman Empire. However, thanks to its wide use, Greek attained a status almost equal to that ofLatin.OtherlanguageswerespokeninvariouspartsoftheEmpire,and some,likeNeo-Punic,foundtheirwayintotheepigraphicalrecordofthe frst centuries AD. Hieroglyphs also continued to be used for religious writ-ings until late antiquity, but none of these languages are attested on imperial 36.Premerstein 1911, 43-48. 37.Tis type is particularly common in Egypt. For example Milne 1903, no. 176. 1 vvisoimo0mi1s23 statue bases.`" Te epigraphical evidence shows that Latin was the predomi-nant language, at least for writing on stone, north and west of a line running approximatelyfromthepointontheAdriaticcoastwheretheViaEgnatia begins to the mouth of the Danube and in the landscapes west of Cyrenaica in North Africa. Te province of Moesia Inferior is divided evenly between Latin and Greek inscriptions.`' StatuebaseswithLatininscriptions,atotalof1,309,completelydomi-nate in the western part of the Empire, and in addition 129 bases with Latin inscriptionshavebeenfoundinprovinceswhereGreekorotherlanguages were otherwise more common. Tese bases are concentrated in two settings: Roman colonies and areas with a strong military presence. In the province of Iudaea, for example, four out of fve attested imperial statue bases are in Latin. A military unit dedicated one in Samaria (Hadrian 408), and the decuriones in Colonia Aelia Capitolina another (Antoninus Pius 288). Likewise, the other provinces in the East had large percentages of statue bases with Latin inscrip-tions:Syria(34),Cappadocia(33),Galatia(43),andAegyptus(23). Apartfromthesepredominantlymilitarydedications,wefndstatuebases with Latin inscriptions in the provincial capitals of Ephesus and Gortyn, and in the Caesarean and Augustan colonies: Alexandria Troas, Ancyra, Antiochia en Pisidiae, Berytus, Knossos, Comana, Corinthus, Germa, Iconium, Lystra, Olbasa,Parium,andPhilippi.Inmanyofthesecolonies,Latinwaspersist-ently used for centuries afer the original infux of Latin colonists, at least for omcial documents." In Athens the colonies of Caesarea Antiochia and Iulia Diensium(Hadrian207,208)employedLatinfortheircontributiontothe statues of Hadrian in the Olympieion. To complete the picture, we fnd statue bases with Latin inscriptions in the civitas Stektorion in the province of Asia, and in the municipium of Stobi in Macedonia.Te use of Greek is almost completely confned to the area east of the lan-guage division line. Of the 973 inscriptions in Greek, the only exceptions are eleven bases in Rome and two more in Italy that were dedicated either by Greek cities or in two instances by Greek individuals (Claudius 13; Titus 14). 38.Hieroglyphs are frequently found on statues and bases of the Ptolemaic kings (Stanwick 2002), and also appear in connection with reliefs depicting the emperor as pharaoh, for example at Philae. 39.Marrou 1963, 377 shows a map with an indication of the approximate language bound-ary. 40.Corinth serves as an example. Here the preferred language gradually changes from Latin to Greek throughout the frst and second centuries AD (Kent 1966, 18-19). Among the imperial statue bases, 14 of 17 inscriptions are in Latin. 26vom.imvivi.is1.10in.sis23inscriptions,predominantlyfromAsiaMinorandthenortheastern provinces, are bilingual. In most instances the Greek text is an exact translation of a Latin original with the emperors name in the dative case in both Greek andLatin.'OnlyoneinscriptiononabasefoundinSagalassos(Claudius 146) translates the meaning of the text and employs the accusative case in the Greek text. Other inscriptions use the Latin formula for the emperors name, buthavethenameofthededicator,usuallyaGreekcity,inGreekaloneor in conjunction with a Latin translation.` Tis could indicate that the name formulaoftheemperorinLatinwasreadilyidentifableevenforaGreek-speaking audience,` while the other elements had to be in Greek to ensure comprehension. Types of statue base No typology exists for statue bases from the Roman imperial period like those devised for the statue bases from the Archaic, Classical and Hellenistic peri-ods. What follows is not an attempt to produce a typology for statue bases 41.Augustus 191; Tiberius 132; Caligula 20; Claudius 121; Nero 44, 32; Titus 33; Domitian 34 (ablative case in Latin), 37; Nerva 43; Trajan 173; Hadrian 84, 287, 288, 373; Antoni-nus Pius 66, 67. 42.Augustus 6, 167; Tiberius 7; Hadrian 196, 217. 43.Basic knowledge of the imperial nomenclature in Latin could be learned from coins with LatininscriptionsthatcirculatedfreelyinallpartsoftheEmpire.Evenifonlyafrac-tionofthepopulationinantiquitywastrulyliterate,thenumberofpeoplethatcould read simple formulaic inscriptions must have been many times larger. Te content of the inscriptionsonimperialstatuebasescertainlyhadanaudience,andcouldcertainlybe understood by the peers of the dedicator, who ofen use the bases for self-glorifcation, see Eck 1994, 630-662. On epigraphic literacy in general, see Harris 1983, 87-111; Harris 1989 and Franklin 1991, 77-98. 44.Bulle 1898; Jacob-Felsch 1969; Schmidt 1993. Bonnevilles (1984, 117-132) proposal for a system to achieve a uniform description of primarily Latin epigraphical monuments divides inscriptions on stone into 14 diferent categories, each with a number of subgroups. Tis typology focuses on the form of the monuments rather than their function, and the statue bases in this study fall within four of the fourteen diferent categories. Type 4 pidestaux, for example, only include monolithic bases and bases with a monolithic shaf, either cir-cular or polygonal with a moulded bottom and plinth. Built-up bases with am xed marble slabs can be found under type 8 and 9 Pierres et plaques moulures et non moulures which also include a broad range of other types of inscription. Te best discussion so far of western Latin statue bases is that of Alfldy (1984, 23-33) concerning the bases from 1 vvisoimo0mi1s27 of the Roman imperial period, being rather intended as an introduction to the variety of diferent types of base that were used for imperial statues, together with an evaluation of their frequency. Given the strict hierarchy that existed for dedications with regard to placement in towns, materials used and monu-ment size, it is reasonable to expect that statue bases for emperors deviated in a number of respects from those erected for persons of lower social standing, and a typology for imperial statue bases does not necessarily apply to statue bases in general or vice versa. Bases for life-size standing statues By far the most frequent type employed is a free-standing base for a life-size standing(orsitting)statue.Tesecouldbeconstructedinavarietyofways. Te simplest consisted of a square or cylindrical monolithic block of marble or sandstone, on which the plinth of the statue was attached for marble stat-ues (Fig. 1). Bronze statues were normally fastened directly into sockets on top of the base (Fig. 2). Some form of moulding was ofen carved at the top and bottomofthebase(frontpage).Temouldingswereveryfrequentlycarved separately and fastened with dowels to a monolithic shaf, in which case we are unfortunately ofen lef with nothing but the shaf, and no means of determin-ing the material of the statue (Fig. 3).' Te same is true of orthostat bases in the Hellenistic tradition, which were still used in the Roman period in Greece and Asia Minor. Te second large group of bases consists of built-up cores to which marble slabs were fastened (Figs. 4-3). Unless the slab with the dedica-tory inscription is found in situ or with great certainty can be associated with Venetia et Histria. Here a diferentiation between bases and altars with similar inscrip-tions is achieved on the basis of the cuttings on top of the monuments, and slabs from built-up bases are identifed by their dimensions and the layout of the inscription. 43.Tecatalogueincludesapproximately60cylindricalbases.Tisformwasparticularly popular in Asia Minor and in Achaea, where round bases had a long tradition, but they are also found in Southern Spain and in North Africa. Polygonal base shafs that are com-moninlateantiquityhaveonlybeenattestedinoneinstanceforimperialstatuebases from the fst two centuries AD (Commodus 96). 46.Techoiceofstonenaturallydependslargelyonwhatwasavailablelocally.Graniteis reported in Spain (Titus 23; Lucius Verus 39) and in Egypt (Caligula 28). In Syria basalt was used (Lucius Verus 133; Commodus 108). 47. For a number of drawings of the impressions on tops of base shafs without the top mould-ing, see Alfldy 1984, 167-169. 28vom.imvivi.is1.10in.sisFig. 1. Statue base in the National Museum in Athens with oval depression for the plinth of a marble statue (Authors photo). Fig. 2. Base for a statue of Claudius (Claudius 87) in the Athenian Agora with typical traces from a bronze statue (Authors photo). 1vvisoimo0mi1s29a base-like structure, identifcation of the type of monument to a large extent restsontheformulationandlayoutoftheinscription.However,thedimen-sions of the slab with the inscription ofen correspond to those of the front of the base, i.e. tall and narrow unlike building inscriptions, which are generally wider than they are tall." Obviously built-up bases were far more susceptible todamagethanmonolithicones,andtheslabsareofenratherfragmentary. Not all built-up bases were covered with marble slabs. Sometimes bronze sheets were used instead. In areas with a dearth of stone suitable for cutting inscrip-tions, this would be a particularly attractive solution, but the use may have been more widespread as shown by its occurrence in Rome on a base for Tiberius dedicatedbytheAenatorestubicinesliticinescornicinesRomani(Tiberius1). Because of the high scrap value of bronze, very few of these monuments have survived. Bronze sheets pertaining to imperial statue bases have been found in Augustomagus in Lugdunensis (Claudius 47),' Fodinae in Baetica (Nerva 19), and Herculaneum (Claudius 8). Other less durable media for inscriptions may have existed as well inscriptions painted on stucco, for example. However, in Pompeii and Herculaneum, where such inscriptions have been found in great numbers, none were painted on statue bases." In Italy built-up bases covered with stone slabs seem to have been more common than elsewhere judging from the large number of monuments found heredescribedmerelyastabulae.Tesebasesaregenerallytallerthanthey are wide, but depending on the setting, they may also be low and squat, such as the bases on the podium in the Caserma dei Vigili in Ostia (front page). TeheightofthebasesinItalyandthewesternprovincesgenerallyranges from 0.80 to 1.40 m. Te tallest complete bases for a standing statue in Italy, a base in Puteoli (Marcus Aurelius 33), measures 1.74 m, while there are more examples of signifcantly taller bases in western North Africa. A pair of bases forstatuesofMarcusAureliusandthedeifedLuciusVeruserectedinthe basilica in Cuicul (Marcus Aurelius 99; Lucius Verus 67) measure about 2.40 m. Bases over 2.00 metres tall are an exception, however. 48.Of course this is only a general rule. Built-up bases had the same variety of form as other bases, and in addition the plaque with the inscription did not necessarily cover the en-tire front of the base. Even busts could stand on built-up bases, as shown by the fnd of a bust of Cato in Volubilis (Lahusen & Formigli 2001, 42-44). Tis bronze bust with an inscription on its lower part stood on a tall, very narrow brick base that originally must have been covered in plaster. 49.Piganiol 1939, 430-437; Boon 1980, 117-132. 30.For inscriptions on bronze plaques, see France-Lanord 1960; Eck 1997, 193-207. For the unlikelihood of monumental inscriptions on wood, see Eck 1998, 203-217. 30vom.imvivi.is1.10in.sisFig. 3. Statue base for Trajan in Delphi (Trajan 110) with sepa-rately sculpted mouldings top and bot-tom (Authors photo). In the Greek East relatively few bases of the built-up type exist.' Instead, orthostat bases seem to have been used to save expencive building material, especially for large monuments. Troughout the period two diferent formats of base were in use: the taller narrow type common in the West (Fig. 3), and a much lower type with a height of between 40 and 63 cm (Fig. 2). Although there are a few monuments with a height above 2 m in Asia Minor (Augustus 31.ExamplesareattestedinPergamum(Trajan168),Ephesus(Nerva33)andinCorinth (Nerva 23); all places with a strong Italian presence. 1 vvisoimo0mi1s31Fig. 4. Build-up bases in a building (Augusteum?) adjoining the forum in Lucus Feroniae (Authors photo). 182; Antoninus Pius 263; Marcus Aurelius 248; Lucius Verus 118; Commodus 104), the tall and narrow bases tend to be slightly lower than those encoun-tered in the West. Common for all statue bases used for approximately life-size statues is that their width generally ranges from 30 to 73 cm. Most bases are square,butofenmaterialandtransportcostscouldbeloweredbymaking the bases slightly rectangular.Te letter size varies according to the language used. Of the 638 monuments described as statue bases with Greek inscriptions, only eight contain letters tallerthan7cm,andon83thelettersrangebetween2and3cm.While Greek bases almost always employ a homogeneous letter size throughout the inscription(Fig.2-3),baseswithLatininscriptionsofenuseverydiferent letter sizes to emphasise important elements in the inscription, typically the 32vom.imvivi.is1.10in.sisFig. 5. 10 cm. thick marble slab from the front of a build-up base for Trajan (Trajan 37) from the Augusteum? in Lucus Feroniae (Authors photo). names of the emperor and the dedicator (Fig. 6). Te less signifcant elements of the inscriptions are normally written with letters of approximately 4 to 3 cm; the name of the emperor normally appears in letters 6 to 8 cm high. 1 vvisoimo0mi1s33Fig.6.StatuebaseforMarcusAurelius(MarcusAurelius20)fromFidenae,nowin Museo Nazionale Romano, with varying letter sizes for the diferent parts of the inscrip-tion (Authors photo). 34vom.imvivi.is1.10in.sisEquestrian statue bases Equestrian statues ranked above standing statues in the hierarchy of honours,` and their majestic size and the fact that they could be viewed from all sides made them ideally suited to defne and dominate open spaces like fora. Te equestrianstatuesintheForumofAugustusandofTrajanandtheForum Romanum serve as excellent examples of this. Tus equestrian statues would seem an obvious medium for representations of the emperor. However, only 17 statue bases for equestrian statues of the emperors of the frst and second centuries AD have been identifed.` In one instance, the inscription mentions that the statue was an equestrian statue (Marcus Aurelius 133), but otherwise the size and especially the depth of the base or the marks of hooves on top of the base are the only criteria for identifcation. Tis of course means that if only the part of a monument with the inscription has been preserved, it is notnormallypossibletoidentifyitasanequestrianbase,andwemustex-pect the actual number of such bases to be somewhat higher than the extant examples indicate. It is worth pointing out that the percentage of statues for emperorsamongtheequestrianstatuesdoesnotexceedthepercentageof imperial statues among standing statues.Tere are three types of equestrian base. One consists of a built-up core on which marble or limestone slabs are amxed. Bases of this type were preferred intheWest,forexamplethetwobasesintheforuminVeleia(Claudius27; Vespasian14).IntheEasttheHellenistictraditionfororthostatbasescon-structedfromseveralblockscontinued(Tiberius99),butequestrianstatues with the horse standing on its hind legs only could also be placed on a single block (Claudius 94). Apart from this last example, which is rather shallow, the depths of the extant examples range from 1.44 m (Claudius 119) to 3.78 m (An-toninus Pius 110), the widths being from 0.60 m to 2.00 m. Like equestrian statues in general, the majority of the imperial equestrian statue bases belong 32.Cic. Phil. 9.13. 33.Augustus 32 (uncertain), 103, 192; Tiberius 71, 99, 101, 130; Claudius 27, 63 (uncertain), 94,119;Vespasian14;Trajan62(uncertain);Hadrian114(uncertain);AntoninusPius 110; Marcus Aurelius 133; Commodus 99 (uncertain). For equestrian statues, see Berge-mann 1990; 1992, 313-324. 34. Lessthanone-quarteroftheepigraphicalmaterialrelatingtoequestrianstatuesofthe imperial period concerns statues of the imperial family (Bergemann 1990, 119-133). In Venetia and Histria, imperial statues made up 20 of the total (Alfldy 1984, 36-38), and in the forum in Cuicul more than 60 (Zimmer & Wesch-Klein 1989, 33). 33.For the size of equestrian monuments, see also Bergemann 1990, 119. 1 vvisoimo0mi1s33 to the Julio-Claudian period (63). Afer the reign of Claudius we only have six examples, so the decline in the number of equestrian statues did not result from this honour becoming a prerogative of the emperor, although afer the second century AD hardly any examples of non-imperial equestrian statues are known. Tey seem to have gone out of use for other reasons. At the eastern end of the forum in Tamugadi, seven extremely large statue bases made of orthostat blocks have been identifed. Te largest, a monument for Antoninus Pius (Antoninus Pius 148), has a width of no less than 6.23 m andadepthof3.90m,andanotherforMarcusAurelius(MarcusAurelius 131) a width of 4.10 m. Zimmer proposes reasonably that the bases carried quadrigae. In Hierapolis a recently published inscription on a base constructed of at least four blocks mentions the erection of a o o, which in this connec-tion can only mean that the statue of the emperor stood in a quadriga. While such statue groups certainly stood on many arches, the bases in Tamugadi andHierapolisaretomyknowledgetheonlycertainlyidentifedbasesfor quadrigae.' Arches, columns and pillars According to Pliny the Elder, the function of columns and arches was to el-evatethestatueofthehonorandaboveothermortals."Whilecolumnsare rare included in the catalogue are only the well-known columns of Trajan and Antoninus Pius (Trajan 13; Antoninus Pius 17); a third column in Rome still to be seen today, that of Marcus Aurelius, does not have its base with the inscription preserved' triumphal and honorary arches and other types of gateandportalimitatingthearchitectureofthetriumphalarchareprolifc inallpartsoftheEmpire."Mostof(ifnotall)thearchesdedicatedtothe 36.Bergemann 1990, 3, n. 22. 37.Bergemann 1988, 113-128. 38.Pliny,HN34.27:Columnarumratioeratattollisuperceterosmortales,quodetarcus signifcant novicio invento. 39.Columns as statue bases have been discussed most recently by Jordan-Ruwe (1993). See also Settis 1988 and Vogel 1973. 60.Te catalogue of Khler (1939, 373-493) is still the most complete collection of Roman arches. De Maria (1988) updates and expands the Italian evidence, see also the review by Kleiner 1989b, 193-206. Other recent works on arches include Pensa et al. 1979; Kleiner 1983; Pfanner 1983; Hesberg 1992, 277-293; Wallace-Hadrill 1990, 143-181; Eck & Foer-ster 1999, 294-313; Kader 1996; Kpper-Bhm 1996; Roehmer 1997. 36vom.imvivi.is1.10in.sisemperor seem to have carried some form of sculptural representation of the emperor, and thus fall under the defnition of statue bases. Unlike most other arches, the one example of an arch where the presence of a statue of the em-peror is disputed, the arch of Hadrian in Athens, does not carry a dedication to the emperor but instead a statement about the two sections of the city that thearchseparates,andisthusexcludedfromthecatalogue.'Mostofthe 106 statues from 86 arches in the catalogue have been identifed by associa-tion with fnds of architectural fragments. 20 of the inscriptions pertaining to arches specify that the object was an arch. Six inscriptions also mention the statue placed on top.` To the 86 identifed arches should perhaps be added a small number of inscriptions described as lintels, architraves and epistyles that might have belonged to arches. Pillarswereafavouredformofpedestalforstatuesofkingsduringthe Hellenistic period, but they seem to have fallen out of use during the imperial period, and we know of no new pillars constructed for statues of emperors. However, existing pillars could be reused for imperial statues, like the pillar infrontoftheStoaofAttalusIIinAthens,whichwassuppliedduringthe reign of Tiberius with a new inscription and possibly a new or re-modelled quadriga on top.` Statues in architectural settings Many imperial statues were placed in an architectural setting. Sometimes they stoodonseparatelyworkedbases,ascanbeobservedontheNymphaeum of Herodes Atticus in Olympia (Antoninus Pius 201; Lucius Verus 108), the Nymphaeum Traiani in Ephesus (Nerva 33; Trajan 143), or in the scenae frons of the theatre in Aphrodisias (Domitian 37). Statues that stood in niches, on the other hand, were accompanied by an inscription on a marble plaque am xed 61.Adams 1989, 10-16; Post 1999, 179. 62.Arch and statue mentioned in the inscription: Trajan 86; Hadrian 142 (statue and quad-riga); Antoninus Pius 109, 179; Marcus Aurelius 132, 132. Arch but not statue mentioned in the inscription: Augustus 164; Tiberius 116, 79 (inscription on what is described as a lintel. Te inscription concerns the dedication a forum, a paved area, an altar of Augustus, a temple of Saturn and an arch. It seems reasonable to associate the inscription with the arch); Claudius 74, 73, 117; Vespasian 36; Hadrian 121, 147, 132; Marcus Aurelius 133, 132, 136; Lucius Verus 83, 91. 63.Tiberius 93. See, Vanderpool 1939, 86-90. 1 vvisoimo0mi1s37 Fig. 7. Exedra opposite the theatre in Emerita Augusta with statues of the imperial fam-ily placed in niches in the wall. Te fastening for the inscribed plaques are visible below the statues (Trillmich, W. et al. 1993. Hispania Antiqua. Mainz, taf. 49c). undertheniche.Anexcellentexampleofsuchanarrangementofimperial statues has been preserved in the theatre complex in Emerita. In a small room attachedtotheperistyleoppositethetheatre,sevenstatueswereplacedin 38vom.imvivi.is1.10in.sisniches. None of the dedicatory inscriptions are extant, but their positions below the niches can still clearly be discerned (Fig. 7). Te posthumous inscrip-tion for Vespasian in Misenum that was reused from a Domitianic dedication (Vespasian8;Domitian3)almostcertainlybelongedunderthenichethat held a nude statue of Vespasian in the back wall of the meeting house of the Augustales. Te inscription measures 30 by 93 cm, and its portrait character would thus not be immediately recognisable if it had not been found near the statue. Te same can be said of the 143 cm wide inscription for Tiberius from the theatre in Herculaneum (Tiberius 13), which may have been fastened to the wall under the bronze statue of Tiberius found there. In the South Bath in Perge, two wide panels found in the apodyterium were fastened below niches that held statues of Marcus Aurelius and Commodus (Marcus Aurelius 238; Commodus 93). One could suspect that many of the monuments described as tabulae whose measurements do not comply with the average size of statue bases may have been placed under niches holding statues. Other types of monument In Cilicia and Syria there was a local tradition for placing statues on consoles highabovethegroundalongthecolonnadedstreetsthatwereacommon feature in the area. Even though this position ofen rendered the inscriptions virtually unreadable, they still comply with the standard pattern for honorary inscriptions. Consoles with imperial statues are known in Olba (Tiberius 143), Pompeiopolis(Augustus196,197;Hadrian401;Commodus103),Palmyra (Hadrian 407), and Apamea ad Orontem (Antoninus Pius 282; Marcus Au-relius 231; Lucius Verus 134) (Fig. 8). Reliefs with portraits of emperors accompanied by dedicatory inscriptions havebeenincludedwhenthededicationconcernstheemperorportrayed. Tese are the reliefs from the sebasteion in Aphrodisias (Claudius 113; Nero 40, 41), the fragmented relief with a portrait of Lucius Verus in Ostia (Antoni-nus Pius 36; Marcus Aurelius 27; Lucius Verus 16), and the pediment with a portrait of Commodus in Philadelphia (Commodus 89). Generally the object of the dedicatory inscriptions in the catalogue is the statue of the emperor. However, a few inscriptions recording the dedication 64.Fuchs 1987, 167-169. 63.Franciscis 1991. 66.IGSK 34, 148-149. 1 vvisoimo0mi1s39 ofbuildingshavebeenincludedeitherbecausestatueswerecloselyasso-ciatedwiththebuildinginquestion,orbecauseastatueoftheemperoris mentionedintheinscriptiontogetherwiththestructurededicated.Tothe frst category belong two monuments described as aediculae with statues of VespasianandhistwosonsinSide'andLamosinCilicia"(Vespasian72, 77). To the second belong a macellum in Tolocaesarea in which the statue of Hadrian evidently had a prominent position since it is mentioned before the building in the inscription (Hadrian 362),' and two temples with statues in UzaaeandVolubilis(Trajan92;AntoninusPius102).Finallythehorreain Myra with busts of Hadrian and Sabina inserted over the inscription (Hadrian 371) and the Tropaeum Augusti (Augustus 80) have been included, although it is not absolutely certain that the second monument actually carried a rep-resentation of Augustus.'" Literary testimony for imperial statue bases and inscriptions While imperial statues are mentioned frequently in the literary sources,'' there arefarfewerreferencestostatuebasesandtheirinscriptions.'`Amongthe few examples that we do have, there is a wide variety of monument types. Te smallest is a bust of the young Octavian with an inscription in iron letters nam-ing him Turinus, seen by Suetonus;'` the largest the colossus of Nero in his Golden House,' which was later changed into a statue of Sol before being re-vamped as Commodus in the guise of Heracles. At least in this last instance an inscription following the normal pattern for dedicatory inscriptions must have 67.Mansel 1962, 198-208. 68.Bean and Mitford 1970, 208-209. 69.Hadrian 366: o [o]tpoov Tooxoioo v o vt]o xoi to [o vr r[i to voio 0ov xoi to ]oxr x[tm x]m v. Te most probable reconstruction, but of courseov r v tp p aom other objects than a statue could come into question. 70.Lamboglia 1938. 71.Lahusen 1984, 61-91, 111-120. 72.For inscriptions in literary sources, see Stein 1931. 73.Suet. Aug. 7.1, with comments by Gross 1980, 126-34. A pair of bronze busts of Augus-tus and Livia with inscriptions engraved on the bases (Augustus 73) have been found in Neuilly-le-Ral, now in Paris, Louvre Br 28 & 29. 74.Bergmann 1994. 40 vom.imvivi.is1.10in.sis Fig. 8. Two of the originally three consoles carrying statues of Antoninus Pius, Marcus Aurelius, and Lucius Verus on the colonnaded street in Apamea ad Orontem (Antoninus Pius 282; Marcus Aurelius 255) (Authors photo). been present, since Commodus, according to Herodian, had the title Germani-cus replaced by the number of his victories as a gladiator.' Another lost monu-ment known through literary sources and coins is the columna rostrata of Oc-tavian in the Forum Romanum, crowned by a golden statue voted by the senate afer his victory over Sex. Pompeius in 36 BC. Te base carried an inscription, part of which is repeated by Appian.' To the ancient spectator, the statue and base constituted a single inseparable entity; and like statues, which have ofen 73.Hdn. 1.13.9: u o tp ori ou ri 0ooi ooiixo io o aoyo i o touo m xoi aotm vti or Irovixou ovoo ou vixp ou ii oovto. Similarly in SHA Comm. 17.10: ac suum imposuit et titulum more solito subscripsit, ita ut illum Gladiatorium et Efemi-natum non praetermitteret. 1 vvisoimo0mi1s41only been recorded because of their exceptional size or material costs, inscrip-tions have been handed down because of their unusual content or their remark-able aferlife. Of particular interest is the inscription from a triumphal statue of Domitian, mentioned by Suetonius,'' which was torn of in a storm and fell into a nearby grave shortly before the death of the emperor. Even though the story is most probably an invention, it illustrates that to a contemporary Roman audience it was not an unfamiliar phenomenon for inscriptions to fall of their bases the point of the anecdote being where the inscription eventually ended up. Evidently the base must have been built up and covered with plaques, in this case most likely of bronze, since marble plaques would hardly have been car-ried away by a storm however thin they may have been cut. Another example of removal of an inscription is a base for a statue of Augustus on the Capitol that was struck by lightning in AD 14, damaging the inscription and leaving only aesari for Caesari, which meant god in Etruscan.'" Te missing C supposedly indicated the time of his death 100 days later. Other references are to the con-tent of inscriptions on statues in general. Inscriptions from statues of Galba, for example, evidently postulated a family relation to Q. Catulus,'' although this cannot be confrmed in the epigraphical record. Statues of Titus with inscrip-tions, again according to Suetonius,"" were erected in great numbers during his termasmilitarytribuneinGermaniaandBritannia.Tishasnotbeencon-frmed archaeologically either. Te only extant inscription from a statue base mentioned in the literary sources is the one from the Tropaeum Augusti near Nicaea, which Pliny the Elder reproduced in full (Augustus 80)."'aiyop ouoo,o vri vpvr vpvr 76.App.BCiv.3.130:r vr ti'tp p otooioor xaoou ouvr v xoi 0o otpor xoto tr yp oooov. See also Jordan-Ruwe 1993, 64-66. Te phrase bylandandseaoccursininscriptionsforSextusfatherPompeiusMagnusintheEast (see Amela Valverde 2001, 87-102), for example on statue bases in Mytilene (IG XII, 2, 202)andinMiletupolis(IGSK26,24),andwefnditinotherinscriptionsforAugus-tus(Augustus173:ao xoi0oo ao vao xoiao opyp oopr atpv,200:to opyp op oop o 0oo ovto, 180, 187). Te phrase is used once on a base for Tiberius, but then disappears until the reign of Trajan. 77.Suet. Dom. 13.2: E basi statuae triumphalis titulus excussus vi procellae in monimentum proximum decidit. 78.Suet. Aug. 97.2 and Dio Cass. 36.29.4. For a similar incident, this time concerning a Greek inscription, see Dio Cass. 76.11.2. 79.Suet. Galb. 2. 80.Suet. Tit. 4.1. 81.Lamboglia 1938. Pliny HN. 3.136-137. 42 vom.imvivi.is1.10in.sis Statue types and materials used for imperial statues Imperialimagesweretransmittedusingalmostanymaterialimaginable. Owingtopoorconditionsforpreservationoforganicmaterialsinmost areasoftheRomanEmpire,paintedimagesoneofthecommonesttypes ofportraitinantiquityhavebeenlostalmostentirely.Nopaintedimages of the emperors under consideration have been preserved. Te style may be glimpsedfromtheexampleinBerlinshowingSeptimiusSeveruswithhis wifeandchildren."`Teemperorsportraitappearedontheobverseofthe majority of the coins in circulation in the Empire. It could also be found on ceramics"`andsilverware,"lamps,"cameos"andmedallions,andinthe militarythestandardscarriedtheirportraits,asdidtheweaponryinsome instances."' Buildings and commemorative monuments carried reliefs show-ingtheexploitsoftheemperor,""orhaddecorativeelementswithimperial portraitslikeimaginesclipeatae."'Portraitsintheroundrangeinsizefrom miniature portraits'" and other transportable images'' to busts and life-size statues, or portraits of truly colossal proportions.'`Teportraitsoftheemperorthatwereaccompaniedbyinscribedbases represent only a fraction of these diferent representations, namely the life-size or larger standing, sitting or riding statue. Tere are three basic types: statues of the emperor wearing the toga (togata) or with his head covered by the toga (togatavelatocapite)inhiscapacityofpontifexmaximus,cuirassedstatues (loricata) stressing the aspects of the emperor as general, and nude or semi-nude statues of the emperor in a heroic pose or in the guise of a god.'` 82.McCann 1968, 79-80. See also Euzennat 1994, 111-113; Blanck 1969b. 83.Facsady 1996, 21-23. 84.Te Boscoreale Cup with a portrait of the triumphant Tiberius, Hron de Villefosse 1899. 83.Buchholz 1961, 173-187. 86.Megow 1987. 87.Exhibition Catalogue Berlin 1988, 338-360, no. 383-386. Among these Te sword of Ti-berius in the British Museum, inv. no. PS 107808. 88.Bonanno 1976. 89.Winkes 1969; Winkes 1999, 91-93; Budde 1963, 103-117. 90.Schneider 1976; Jucker 1964, 81-92. 91.Blanck 1971, 90-93. 92. Kreikenbom 1992. On the largest statue in antiquity the Colossus of Nero, see Bergmann 1994. 93. For the use of statue types in general, see Wegner 1939, 283-287 and Niemeyer 1968, 14-20. Togate statues: Goette 1990. Cuirassed statues: Stemmer 1978. Heroic statues: Maderma 1988. 1 vvisoimo0mi1s43Statues mentioned in the inscriptions from statue bases In the original context it was normally quite obvious what the object of the dedicationwas,andconsequentlythisinformationwassuperfuousinthe dedicatory inscription. If, however, the dedicator wished to specify the con-ditions under which the dedication took place, or if the inscription was not placedimmediatelybelowthestatue,theobjectcouldfnditswayintothe inscription. Against this background, we should not expect the examples to constitute a representative selection of the monuments in existence. Since the statue bases under discussion for the most part carried life-size standing statues, it is not surprising that the most commonly used term in the Latin inscriptions is statua, which roughly translates to statue.' In a few of the 43 occurrences, the type of statue dedicated is described in more detail. A statua triumphalis of Trajan was erected in Olvera in Baetica (Trajan 68), and a statue of Hadrian is specifed as being seven foot tall (Hadrian 113). One inscription mentionsanequestrianstatue(MarcusAurelius133)erectedinconnection with the building of the basilica in Tugga, four others mention statues placed onarches(Hadrian142;AntoninusPius109,179;MarcusAurelius132).In Tugga, Nanneia Instania Fida dedicated two colossal statues (Marcus Aure-lius169;LuciusVerus97).Anewlyelectedsevirdedicatedstatuaesacraeof Antoninus Pius and his two sons because of the honour of omce and out of gratitude for the citys new aqueduct (Antoninus Pius 93; Marcus Aurelius 92; Lucius Verus 61). Tree times we hear of statues that needed repair (Claudius 60; Tiberius 78 [aedem et statuas corruptas]; Marcus Aurelius 16 [vetustate cor-ruptis]). Imago was used to denote a wide range of diferent types of portrait.' In the ten instances where the word is mentioned on statue bases, however, the meaning seems to be bust or statuette, fve of which were made of silver (Trajan 92; Hadrian 123; Antoninus Pius 63; Marcus Aurelius 66, Lucius Verus 38). A diminutive base (0.143 x 0.233 x 0.21 m) mentions an aedicula with an imago of Nero (Nero 3). If this image was placed on the base, it must have been a bust orastatuette.Anotherinscription,evidentlyfromanaltardedicatedtothe wellbeing (pro salute) of Marcus Aurelius and Lucius Verus (Marcus Aurelius 103, Lucius Verus 69), mentions that the dedicated objects included both the altar and imagines. Since there was more than one imago, those of the emperors seem more probable than personifcations of Salus. Tese inscriptions imply 94.For Latin and Greek terminology for various types of portrait, see Roux 1962, 366-380; Daut 1973; Letta 1978, 3-19; Price 1984, 176-180. 93.Daut 1973. 44 vom.imvivi.is1.10in.sis that imago, at least when used on imperial statue bases, ofen referred to reli-gious objects. One inscription refers to statues as signa (Augustus 31; Tiberius 38) placed in a temple. Because the marble slab on which this inscription is cut has three separate dedications to Augustus, Gaius Caesar and Lucius Caesar (with later additions of Tiberius and Agrippa Posthumus), it was most prob-ably amxed to a large base carrying the statues. Two terms for statue frequently found in the literary sources, simulacrum and ef gies, are not attested on the statue bases in the catalogue.'voioand o Te two commonest Greek terms for statues, o yoo, both denote life-size statues. Te distinction between them seems to be one of con-text and function rather than one of type or material. Avoio, the slightly morefrequentterm(12examples),''referstohonorifcstatues,whereas o yoo(knownineightinscriptions)'"hasgenerallybeeninterpretedas meaning cult statue. Unfortunately the contexts of nearly all these bases have been lost, but the frequent use of the dative case in inscriptions referring to yo o oto supports the notion. However, the dative case was not restricted to cult statues as seen from its use on a base for an o voio in Lyktos (Trajan 131). Ei xmv, the vaguest term applied to imperial images, like imago denotes a likeness. It is only mentioned in one inscription (Trajan 188) for an image, probably a bust or miniature representation that was part of a larger dedica-tion of four futed columns, an iron door, three craters and three kantharoi. Bronze or marble Te vast majority of the extant portraits of emperors are made of marble. Tis does not refect the situation in antiquity marble portraits have simply sur-vived better than metal ones. During the Hellenistic period marble generally seemstohavebeenusedforstatuesofdeities,whilehonorifcstatueswere far more ofen made of bronze.'' From this it has been inferred that marble 96.Plenty of examples in Lahusen 1984, 61-91. 97. Augustus 168; Claudius 131; Nero 32; Vespasian 18; Titus 18; Nerva 34; Trajan 131; Hadrian 273, 341, 362; Antoninus Pius 212, 260. 98.Tiberius 148; Trajan 197; Hadrian 267, 423; Antoninus Pius 14, 298; Marcus Aurelius 9; Commodus 93. 99.Discussion by Tuchelt 1979, 70-90. Tuchelt (1979, 76) found that 26 out of 30 bases for promagistrates in the province of Asia erected during the last century BC carried bronze statues. Tis tendency is confrmed by the evidence from Kos collected by Hghammer 1993, 68-70. 1 vvisoimo0mi1s43was the more valuable material, and that honorifc statues in marble, because of their amnities with cult statues, also had religious signifcance.'"" Accord-ing to Lahusen the distinctive meaning of marble statues continued into the Roman period, and in his opinion this could explain why so many portraits of the imperial family in marble still exist compared with portraits of private individuals.Teformercouldbeerectedinpublic,whereasthelatterwere presumablyonlyerectedinprivatehousesandvillaswoesfrdasMate-rial und die Form der Bildnisreprsentation keine Einschrnkungen gab.'"' Countless examples prove the incorrectness of this assumption, of course, and during the Roman period there does not seem to have been any connection between the function of an image and its material. Both marble and bronze could evidently be used for public honorifc statues. Wecanobtainanideaofthefrequencyoflostbronzestatuesbyclose study of the statue bases. By investigating the marks on the top of fully pre-servedstatuebases,itisofenpossibletodiscernwhattypeofmaterialthe statue that stood on the base was made of.'"` Bases for marble statues some-times have a large round, oval or square depression into which the plinth of the statue could be lowered. Te plinth of the statue could also be placed on topofthebase,likethestatueofHadrianinthelibraryintheAsklepieion inPergamum,inwhichcasedetectionisrenderedmoredim cult.Basesfor bronzestatues,ontheotherhand,havefootprintsontoportwotofour dowel holes for fastening the feet of the statue (Fig. 2). Unfortunately very few publications of inscriptions record this type of information, and in order to obtain reliable fgures for the marble/bronze ratio all bases would have to be re-examined. Te only systematic investigation of the issue within the impe-rial period is that of Alfldy for the statues from Venetia et Histria,'"` which showed that bronze statues were slightly more common than marble ones in this area. However, the percentage of bases where anything defnite could be determined about the material of the statues was rather small, and as pointed out by the author, the choice of material may have varied depending on the rank of the honorand. A small sample of statue bases on Kos shows that the 100.Tuchelt 1979, 79-86. On the other hand, Smith (1988, 3) remarks about the portraits of the Hellenistic kings that: Our surviving portraits are mostly lower-grade works in mar-ble, ofen very generalised in character. 101.Lahusen 1992, 192. 102.For illustrations of the diferent traces lef by marble and bronze statues, see Alfldy 1984, 164-163. 103.Alfldy 1984, 37. 46vom.imvivi.is1.10in.sis preference for bronze statues extended into the Roman period,'" and in the fora in Cuicul and Tamugadi bronze statues predominate from the second to the fourth century.'" Another indicator of the frequency of bronze statues in an area is the ratio betweenextantportraitsandstatuebases.Sincebasessurviveregardlessof the material used for the statues, the ratio between extant bases and portraits shouldchangewiththepreferenceformarbleorbronzestatuary.InItaly thereareaboutthesamenumberofportraitsasbases(Fig.9),andinthe northwesternandHispanicprovincestheratioistwobasesforeachextant portrait.InGreeceandNorthAfricatheratiosare3:1and7:1respectively, and Asia Minor and the northeastern provinces top the list with around ten bases for each extant portrait. Although this by no means constitutes a very precise measurement, it indicates that large diferences in the choice of material prevailed within the Empire. In the East the Hellenistic tradition for bronze honorifc statues seems to have continued unchanged, while in the West and in Italy in particular marble was more popular.'" Statues in precious metals Tere is an extensive body of literature drawing on a large number of ancient literary sources on the topic of imperial statues in precious metals.'"' Te ques-tions of whether statues in silver and gold were a prerogative of the emperor, and whether they implied divine properties and had religious connotations, whichhasoccupiedmuchofthediscussion,neednotdetainusheresince the statue bases ofer no new insight regarding this question. Instead we can consider their frequency among the honorifc statues, as well as considering whichemperorshadthistypeofmonumentsdedicated,whoerectedthem and where. Unlike the ample literary evidence, the archaeological evidence is rather scarce. Hardly any examples of imperial statues in precious metals have sur-104.Hghammer 1993, 68-70.103.Zimmer & Wesch-Klein 1989, 34-84. 106.Te same tendency can be seen for the 130 attested Julio-Claudian statue groups collected by Rose (1997a). Among the western examples more than 30 are attested by sculptural evidence, while the fgure for the statue groups in the eastern part of the Empire is only 13. 107.Most importantly Scott 1931, 101-123; Pekary 1968, 144-148; Pekary 1983, 66-80; Lahusen 1978, 383-393; Lahusen 1999a, 231-266; Lahusen & Formigli 2001, 303-324. 1 vvisoimo0mi1s47 Fig. 9. Te number of extant approximately life size or larger imperial portraits (before slash) and the number of statue bases (afer slash) according to region.1 Notice that while the numbers are approximately even in Italy, there are far more bases than portraits in most other regions. EmvivovI1.ivNov1u- G.0iSv.iNov1uGviiciAsi.E.s1 ivAivic.Miov Augustus2 77/690/710/1113/226/1119/4014/418/2 Tiberius3 -/47-/3-/8-/14-/7-/30-/41-/1 Caligula4 14/22/01/92/02/03/63/100/1 Claudius3 28/381/72/114/102/106/312/461/1 Vespasian6 19/171/31/12/84/141/91/260/3 Titus7 23/170/20/01/72/141/81/161/2 Nerva8 14/120/30/10/20/30/70/180/1 Trajan9 60-70/413/143/36/102/248/433/373/11 Hadrian10 93/732/243/24/138/3216/12410/11310/11 Antoninus Pius11 ca. 40/630/133/17/183/826/313/701/10 Lucius Verus12 39/380/101/33/1317/367/104/222/7 Marcus Aurelius13 ca. 70/664/162/23/98/848/313/429/10 Commodus14 ca. 20/172/30/00/31/364/134/311/41Tisisbynomeansanallinclusiveillustrationofthegeographicaldistributionoftheextantimpe-rial portraits, and neither do the fgures lend themselves for comparison between diferent emperors. Since the chart was built on the works of others its accuracy depends entirely on these. Note that the date of publication efects the number of known portraits. A more correct chart of the geographical distribution could naturally be worked out by consulting the publication of each individual portrait. Tis would, however, require a considerable efort. In the most recent studies cited, Evers for Hadrian and Boschung for Augustus and Caligula, great efort has been exerted to identify the provenience ofas many portraits as possible, and these therefore ofer the most reliable evidence for the geographical distribution of the portraits.2Boschung 1993a, 227-230. vived until the present'"", and most probably had a very short lifespan.'"' Tose that have survived, although we would expect them to be the most outstand-ing examples of Roman art, are of rather mediocre quality at best, and also seem to deviate from the omcial portrait types in marble. Te evidence from 108.Fromtheperiodinquestion,theimperialportraitsinpreciousmetalare:agoldbust foundinAventicum,mostprobablyalikenessofMarcusAurelius(Jucker1981c,3-17; Witchel in Stemmer 1993, 237-262; Lahusen 2002, 46-63). Silver portraits have been at-tributed to Galba (Die Silberbste des Kaisers Galba 1993), and Lucius Verus (Bendinelli 1937) (from Marengo, now in Torino Mus). 109.Statues could even be melted down during an emperors lifetime (Tac. Ann. 3.70.1). Note also IGSK 11, 1, 23, on the reuse of old imperial portraits in Ephesus. 48vom.imvivi.is1.10in.sis3NocompletefguresexistforalltheportraitsofTiberius.However,alistof23specimensofTypus Kopenhagen NCG 623 can be found in Fittschen & Zanker 1983, no. 10. 9 of these derive from Italy, 9 from other parts of the empire, notably 4 portraits from Egypt, and 7 with no known provenience. Fittschen & Zanker 1983, no. 12, also lists 23 specimens of Typus Berlin-Neapel-Sorrent of which at least 13 have provenience in Italy. Only the portraits in Toulouse, Tripolis and Carthage have defnitely been found outside Italy. Boschung 1990, 363-367. and Boschung 1993b, 36-38 lists a total of approxi-mately 90 portraits of Tiberius.4Boschung 1989. 3Based on the 34 portraits appearing in Fittschen (1973, 33-38, cat. no. 17). Updated fgures will appear in Massner, forthcoming. Massner has kindly informed me that she has found approximately 80 life size portraits of Claudius in the round. 6Daltorp, Hausmann and Wegner 1966 and Bergmann and Zanker 1981, 332-349. 7Daltorp, Hausmann and Wegner 1966, Fittschen 1977, 63-67, cat. no. 21, and Bergmann and Zanker 1981, 332-349. 8Bergmann & Zanker (1981, 380-403) believed 14 life size portraits of Nerva in the round to be in ex-istence. Of these only two or three were new creations while the others were reworked mostly from portraits of Domitian. Most if not all the portraits seem to derive from Italy. 9Gross1940.Balty(1977/1978,43-61)enumeratesatotalof120portraitsbuthiscataloguedoesnot always include precise information about provenience. 10Evers 1994. 11Wegner 1979, 96-123. Many of the entries lack precise information about provenience. Very ofen the location of the museum gives a good indication of the fnd spot, especially in the case of local muse-ums, but for the large European and North American collections, whose portraits to a very large extent was bought in Italy, this method is useless and the proportion of portraits from Italy may therefore be even higher. Te portraits from the old Italian collections have been counted as having been found in Italy. 12Wegner 1980, 38-66. See also note 11. Fittschen (1999) suggests that 11 of the 17 known portraits of Lucius Verus erected prior to AD 161 come from Italy. For the distribution of the portraits of the fourth portrait type, see Fittschen and Zanker 1983, 79-81. 13Wegner 1979, 139-181. See also note 11. Fittschen (1999) suggests that 44 of 61 portraits of Marcus Aurelius with known provenience erected prior to AD 161 come from Italy. Te remaining portraits have been found in the Eastern provinces 4, North Africa 3, in the Northern provinces 4, Asia Minor 2, Spain 2, Greece 2. 14Wegner1980,76-181.Seealsonote11.NotealsoFittschen1999fortheportraitspriortoAD180. Tirteen portraits of Typus Capitol are known. Eleven have been found in Italy. thestatuebasesissomewhatricher,butitdoesnotseemthattheywereas common in antiquity as the volume of ancient literary references and modern scholarship devoted to them would suggest. Tis may well be a result of the tendency of ancient authors to give prominence to the unusual at the expense of the ordinary. Of course the nine inscriptions from statue bases mentioning thirteen portraits are a poor representation of the statues in precious metal once in existence, since our only way of knowing whether statues were made of gold or silver is if the material is specifed in the inscription. Te purpose of mentioning the material in the inscription was not to state the obvious, but to denote the weight of the metal used for the statue or bust. Tis measure-ment seems to have been of some importance, and occurs invariably except 1 vvisoimo0mi1s49when the statue is referred to indirectly (Trajan 92).''" We may therefore have a fairly representative segment of the inscribed statue bases for statues in pre-cious metals permanently amxed in one place. Transportable images, on the other hand, such as busts and statuettes in gold and silver used by the army and as cult objects, seldom had accompanying inscriptions and consequently do not appear in the epigraphical evidence.''' Imperial statues in gold referred to in inscriptions from statue bases are forthemostpartfoundinRome.Tiscouldexplaintheiroverrepresenta-tion in the literary sources, which to a large extent concern circumstances in the capital. In Rome, Gaius Geminius Atticus dedicated an image of Tiberius weighing3poundsofgoldand3poundsofsilver(Tiberius8).Tebaseis described as small, and probably carried a bust of the emperor. A centurion of the third Augusta, Gaius Papirius Aequos, s