d--- Histoire de Tart A FRESH LOOK AT THE IONIC BUILDING AT GARNI - Despite the archaeological activity of recent years at sites such as Artaxata, Armavir and Dvin, Garni remains the most striking collec- ion of monuments in Armenia dating from the Graeco-Roman period. The fame of Gafni has been enhanced recently by the anastylosis of the peripteral Ionic building which occupies the apex of the triangular tongue of land, jutting out above the gorge of the Azat river, on which he ancient fortress is situated. The beauty of the site, the harmonious ines of the rebuilt Ionic building, and the convenient location some 30 km from Erevan on the road leading to the famous monastery of he Holy Lance at Gelard, have served to make Gafni one of the most famous tourist attractions in all Armenia. _ Garni was celebrated in antiquity as a virtually impregnable fortress. The first historical reference to it is found in Tacitus, who calls it by he name ‘“‘Castellum Gorneae” !. From antiquity throughout the middle ges it preserved its importance as a station for the royal garrison and a powerful fortress*. After the Mongol invasions the fortress was gradually abandoned though the village outside the walls has been continuously occupied until the present day. Garni was fairly frequently visited by early travellers, whose attention was drawn to the remains of the walls and above all to the ruins of a small columned building of the Ionic order >. It was this building that first attracted archaeologists .o Garni : then, in 1949 systematic excavation of the site was undertaken by a team of archaeologists under B.N. Arak‘elyan. The enceinte walls _were cleared down to their foundations, and within the fortress were | Tacrrus, Annals XU, 44-48. * Among references to Garni in Classical Armenian writers are: MOSES OF CHORENE » 12, 11, 90; Faustus II, 8; Exisasus III; SeBeos X. ~° Among the more famous travellers to report on Gafni were: J. CHARDIN, Voyages n Perse, Amsterdam, 1686, who on p. 312 gives an exaggerated description he had heard tom the natives; R. KER PORTER, Travels in Georgia etc. London 1822, Vol Hp, 624- 29; DuBOIs DE MONTPEREUX, Voyage autour du Caucase, Paris 1839-43, Vol. HT p. 386- _ 402. Dubois attempted to reconstruct the original appearance of the Ionic building, but he result is rather inaccurate. 222 R.D. WILKINSON discovered the remains of a small bath-building, of a complex 6 storage and other rooms, and of a hall*. In addition some furthe, stones from the Ionic building were discovered which led to sligh modifications in the earlier reconstruction °. In any discussion of the civilisation of ancient Armenia the Ion; building from Garni must occupy a central place. It is still the surviyip, building in Armenia most unmistakably Graeco-Roman in style; ; its architectural technique and decoration it is immeasurably superjo to any other building of this period whose remains have been found in Armenia; and so much of it has survived that anastylosis proved relatively straight-forward. The impression is nowadays sometime: given that the date, function and archaeological significance of thy Ionic building are as easily determined as the details of its architecture It now seems to be generally accepted that the Ionic building was : temple, dedicated to the God Mithras, and built by the Arsacid Kin. Tiridates I in about 70 A.D. and in a style which, though basicalh Graeco-Roman, incorporated a number of more specifically Armenian decorative motifs®. The Ionic building is therefore, according to thi interpretation, a convincing demonstration of the sophistication 0 Armenian civilisation in the early Arsacid period, capable of produci a blend of Graeco-Roman and local architectural traditions. : This article looks again at the evidence and challenges the generally accepted interpretation. In particular it looks at: | (a) the building’s purpose. What are the reasons for assuming it to by a temple? Is there anything else it could have been? * B.N. ARAK’ELYAN, Garni J, Erevan 1952, and Gafni IJ Erevan 1957 (both in Russian). ° The basic publication of the building remains N. BUNIATYAN, Het‘anosakan taéar Trdati palatin kic’ Garnii amroc'wn (The pagan temple next to the Palace of Tiridate in the fortress of Garni), Erevan 1933, A.A. Sahinyan was the member of the excavating team most closely concerned with the Ionic building, and was subsequently in charge of the anastylosis that was undertaken in the early 1970s. His theories are expounded in K‘asali basilikayi €artarapetut’ yuna (The Architecture of the Basilica of K‘asaf), Erevan 1955, chapter 8; Garni ew Gelard (Garni and Geghard), (whose section on Garni has been published with some modifications and some good photographs in ‘“*L’architecture des constructions antiques de Garni”, REArm N.S. VI, 1969, p. 180-200); Antik darasrjani k‘arakert t‘al Garnium». (The stone vault of the antique period at Garni PBH 1976, 1, and «Garni antik taéari karuc‘man zamanaka” (The date of construction of the antique temple of Garni), PBH 1979, 3. © The conclusion reached by SAHINIAN (‘The date of construction...”). See also C. Burney and D.M. LanG, The Peoples of the Hills, London 1971, p. 251. THE IONIC BUILDING AT GARNI Dak he style and decoration. What do these tell us about the men that built it and the architectural and artistic context in which they did 50? | -) the building’s date. What is the evidence from history, from circum- stance and from the stones themselves? _ pESCRIPTION OF THE BUILDING Although the Ionic building at Garni 1s now well known, its architec- e has never been fully and scientifically described in a Western - Buropean language. The following description is intended to fill that gap (and this section may be omitted by the non-specialist or by those already familiar with the building). It also serves to demonstrate hat the Ionic building can be described in its entirety without recourse ny architectural terms that are not standard to Graeco-Roman archaeology, and attempts to clear up a few architectural points on which there may still be room for doubt. (For illustrations see works eferred to in note 5, in particular the Revue des Etudes Arméniennes, 1969), plates LIV-LXIX). The stone is local basalt. The building, which is hexastyle peripteral, rises on a podium placed on two crepis steps. The podium is formed of tall rectangular blocks between a base moulding and a crowning moulding which though made up of various plain fillets and roundels achieve an overall profile of 45°. The podium continues beyond the level of the front colonnade to form two buttresses flanking the approach which consists of nine steep steps. These buttresses are faced with two sculptured bas-reliefs of naked youths kneeling on one knee, their hands and head in a position of support. The podium is 1.77m high from the crepis and is 11.70m wide. Its length is 15.85m excluding the buttresses, 17.8 with them. From the top of the podium which forms the stylobate rise the Ionic columns, six along the ends, eight along the sides, unfluted, their attic bases resting on plinths almost exactly one metre square and one metre apart. The shafts, made up of four drums, are 5.8m high and taper gently towards the top where they are 0.65m in diameter as opposed to 0.66m just above the apophyge. The lower part of the capital, which effects the transition from the shaft, is formed of the upper apophyge, then a fillet, a smooth convex roundel, and an astragal. The echinus consists of an egg and tongue moulding of which only one egg is fully visible owing to the palmettes. The pulvinus is flat, and the spiral is flat and deeply cut. The eye is just above the level of the astragal and outside the line of the top of the column. The volutes are decorated with a variety 224 R.D. WILKINSON of floral motifs. The corner capitals appear to be of the usual typ The arrangement is such as to permit the echinus to face outwards on the two sides, and the two volutes to point along the colonnades. The outs; de corner therefore, has a small volute pinched out at 45°; the inside corne where the two volutes net is squared off into an incised right- ~angled cut in the usual manner ’ The architrave consists of three fasciae, each higher than the one below. and each crowned with a moulding. Between the first and second fasciae is an astragal, between the second and third a rope decoration, whj above the third come in ascending order a flat fillet, an astragal, a cyma recta decorated with a spiky leaf-and-dart (Lesbian cyma), and a wide, _ flat fillet, sloping slightly back. The architrave is 0.595m high, rather higher than the frieze which with the crowning moulding is only 0.43; high. The frieze is decorated with a deeply incised acanthus scroll, lacking in variety of relief, but rich and elaborate in design. This bulges out towards the bottom to present what is virtually a cyma recta profile. Above is a flat fillet, an egg-and-tongue, and then a wider fillet. The dentils the cornice block are surmounted by another leaf-and-dart. Then instea, of the consoles often found in late Ionic and Corinthian comes a band, projecting above the leaf-and-dart, and slightly hollowed out under- neath. An astragal effects the junction between this band and the sima or gutter, which is of the usual cyma recta profile, decorated with a leaf pattern and with projecting lions’ heads. These lions’ heads are purely decorative and do not in fact serve as waterspouts. The raking cornice bears the decoration not only of the horizontal cornice but of the upper part of the frieze as well. Thus below the dentils there is a flat fillet and an egg-and-tongue. The tympanum is of flat polished stone. The rests for the acroteria are visible on the corresponding stones of the pediment, and a large stone, decorated with an acanthus — has been found. It was probably one of the central acroteria °. The cella walls are placed so that the pilasters marking the four corners are exactly behind the second column on each fagade. These pilasters have attic bases similar to those of the portico, and the capitals are similar to the corner capitals of the portico, except of course that those of the antae pilasters have four corner volutes, the rear pilaster capitals havin; only three. Round the cella walls is carved a continuation of the egg- 7 BUNIATYAN (op. cit. fig. 26 p. 53), whose drawings are on the whole most accurat gives a picture of a corner capital in which not only the outermost corner but also the two other outside corners of the volutes are pinched out. Such a capital, halfway between the normal and the foursided type, would be most unusual. In fact the corner capitals are not as Buniat‘yan drew them, but are of normal type. : 8 BUNIATYAN (op. cit. p. 9) says that this stone cannot have been an acroterion owing to its size. He gives no details of its dimensions however. Sahinyan has reversed this judgment. Certainly the stone resembles other acanthus leaf acroteria, eg. that fron the Trajaneum at Pergamum (Altertiimer von Pergamon, Berlin 1895, Vol. V. 2, plate XIV). THE IONIC BUILDING AT GARNI 225 and-tongue of the pilaster capital echinus at the same level. Above comes a simplified architrave with the mouldings reduced to fillets or unde- corated curves. Above that is a flat level corresponding to the frieze, and on that were laid the stone slabs forming the ceiling of the portico galleries. The pronaos is only 1'/, times the length of the pilasters. The wall that divides the pronaos from the cella proper is pierced by a large fine doorway, some 5m high and 2m wide. The elaborately decorated doorposts are made up, from the inside going outwards, of : fascia, rope decoration, fascia, leaf-and-dart, fascia, astragal egg-and-tongue, fillet, concave moulding with thin-leaved decoration, and fascia. The lintel is even more elaborate. It consists of a lower part decorated in the same way as the doorposts, and an upper part decorated with a succession of mouldings that is virtually a repetition of the exterior order from frieze to sima. The part of this lintel that corresponds with the exterior cornice projects at each end beyond the line of the doorposts, and is supported by two consoles decorated like an Jonic volute member with spirals and oculus, with the upper volute considerably fuller than the lower one. The soffits of the architrave are decorated with strips ending in a concave semicircle. These strips contain a wide variety of motifs: vine tendrils, palm leaves, scaly leaf-patterns etc. The blocks making up the ceiling of the porticoes and pronaos have shallow coffers of various shapes, each bounded by an egg-and-tongue moulding within which is a rich and varied decorative pattern of botanical inspiration. The shapes include squares, lozenges and octagons inscribed in squares. ‘he roof of the cella and of the building in general is one of the few doubtful aspects of the architecture. It used to be assumed that the roof must have consisted of beams and rafters supporting les°, That is the usual way in which peripteral buildings of this kind ere roofed. But it is now believed that the roof of the cella was a arrel-vault of stone. Between the stone ceiling and the basalt roof slabs there would have been a filling of rubble and mortar made from a light volcanic ash. On that the tiles of the roof would have rested 1°, The idea that a peripteral building of this kind should have had a arrel-vaulted roof at first arouses scepticism. A wooden roof would have been more normal. However, the architectural elements discovered, lough perhaps not wholly conclusive, do strongly suggest that there was a barrel vault'’. And there are sufficient parallels from other BUNIATYAN, op. cit, p. 13. © Saninyan, REArm N.S. VI, 1969, p. 193. -'! Sahinyan’s arguments (“Antik daraSrjani”) when taken in conjunction with the itallels adduced here are persuasive. 226 R.D. WILKINSON parts of the Roman Empire for it to be perfectly plausible, archaeolog}. cally speaking, that the Ionic building at Garni possessed such a roof The most famous type of Roman barrel vault is of course tha; of concrete, which was one of their greatest contributions to the development of architecture. But stone barrel-vaults are fairly wide. spread in Imperial times and are especially common in Asia Mino; and the East. In conception such a vault is probably merely an elongated arch. Examples where this is particularly easy to comprehend include the well preserved theatres at Myra and Side where the cavea, insteaq of being hollowed out of a hillside as was the Greek practice, is built up on a series of arches and barrel-vaulted passages!*. The seats g the famous stadium at Perge are also supported on a succession 9 barrel vaults, shaped like half-cones, which taper inwards from exterior arches '*. From this it is but a short step to using a barrel-vault for roofing a room. The room however must be fairly small and the walls fairly solid or the weight of the stone and the technical difficulties 6 construction become too great. Four examples of such stone barre vaulted chambers may be cited from South-West Asia Minor. 1) Th Augusteum from Sidyma, a town in North-West Lycia, some 10 miles West of Xanthus **. This is a tetrastyle prostyle temple of a simplifiec Doric with high widely spaced columns and a plain frieze with no triglyphs. It dates from the reign of Tiberius. 2) A fine tomb between Sura and Myra in Lycia'®. This is also a tetrastyle prostyle building. with Corinthian columns and a cella on a high podium not unlike that at Garni. In the podium is built a second chamber which would have held the sarcophagus. It is certainly of Roman date. 3) A building of similar design, though with no burial chamber, is known from Patara'®. There the order had palm-leaf capitals, and the building is probably of the IInd or IIrd centuries A.D. 4) And from Termessu: ‘2 For Myra, v. CH. TexieR, Description de l’'Asie Mineure, Paris 1849, vol. UZ plate 216. For Side v. CounT LANCKORONSKI, Stddte Pamphyliens und Pisidiens, Vienn 1890, Vol. /, fig. 109, p. 149. 13 LANCKORONSKI, op. cit. Vol. I fig. 40, p. 55. ‘4 BENNDORF and NIEMANN, Reisen in Lykien und Karien, Vienna 1884, Vol. p. 61, fig. 42. 15 Antiquities of Ionia (published by Society of Dilettanti) Vol V (London 1915) Myra, plates YXVIT and XXVIII; BENNDORF and NIEMANN, op. cit., Vol. II, p. 4 fig. 32 and plate ¥7, CH. Texter, Description de l’Asie Mineure, Vol. JI, plate 213 ff. ‘© TexiER, L’Asie Mineure Vol. JIT plate 189. THE IONIC BUILDING AT GARNI 22i we have a somewhat similar prostyle tomb with four Corinthian columns *’. It is however not on a podium. All these buildings had barrel-vaulted roofs, and so it is clear that _ parrel-vaults were regularly used for roofing small temples and temple- tombs in South-West Asia Minor from Imperial times. That such a practice was not limited to this area is shown by the strange building nown as the Temple of Diana at Nimes, where although the general 2 plan and design are very different from the examples given above the method of roofing is very similar '8. And a Syrian example is provided by a barrel-vault from the temple precincts at Baalbek *”. : So we may conclude that although it is not proved beyond all — doubt that the Ionic building from Garni had a stone barrel-vault : over its cella, both the evidence of the stones themselves and comparison _ with architectural practice in other countries suggests that it did. For although even in Roman times wooden rooves remained the usual form ___ of roofing for the cellas of peripteral buildings where the span was too wide for any form of solid stone roof, on smaller buildings, especially _in Asia Minor, stone barrel-vaults seem to have been regularly used ~ to cover rectangular cellas. FUNCTION Such was the architecture of the Ionic building from Garni. But if ere are few doubts concerning its architecture, its function 1s harder establish. Even in the early middle ages there was uncertainty, and its _ purpose had been forgotten. Moses of Chorene, speaking of Tiridates the Great, says : “About that time (immediately after the Council of Nicaea) Tiridates completed the building of the fortress of Garni, with hard stones carved with the hammer, joined together with iron and with lead. He also erected a summer residence decked with columns and magnificient bas-reliefs for his sister Xosroviduxt, and had carved _ there a Greek inscription in memory of her’’*°. This description of the site, in particular the mention of a building decked with columns and ‘7 LANCKORONSKI, Stddte Pamphyliens und Pisidiens, Vol. I, figs. 88 and 89; te XX, 18 D.S. RoBertson, Greek and Roman architecture, Cambridge 1929, p. 237 and plate XV. ‘° Baalbek I, Berlin and Leipzig 1921, plate 32 and 33. 20 MOSES OF CHORENE, II, 90. 228 R.D. WILKINSON bas-reliefs and of a Greek inscription mentioning Tiridates’ queen ; uncannily accurate. But the Ionic building was probably never itsejf a summer residence. What may have happened is that when the palace buildings were ruined, the memory that there had been a palace lingered on, and naturally attached itself to the only ancient building s still standing, ie. the Ionic building. The main hall of the palace is known to have been ruined by the 7th century since the church of that periog _ was built partly on its site. It has in fact nearly always been assumed that the building was g pagan temple. An early theory was that it was a temple to the deified Augusti?’, but in recent years the opinion of Kamilla Trever that it was dedicated to the god Mihr or Mithras has become widely accepted. This however is only a supposition; there is no direct evidence. The _ Greek inscription of a King Tiridates probably refers to the fortress _ and not to the Ionic building, and anyway gives no indication that th building was a temple or that it was connected with the cult 0 Mithras ?*. The basic argument is that if the building was a temple, and if it was built during the reign of Tiridates I, then it could well have been dedicated to Mithras since we know that Tiridates I was.a faithful worshipper of that God 7°. To support this theory it has been argued that the archaeologic remains of the vaulted roof show there to have been a hole. According to this theory the purpose of this hole was not only to augment the light coming through the door, but to enable the cult-statue of the 21 _ K.K. Romanov, “Razbaliny xrama rimskogo tipa v Bas-Garni” (The ruins the temple of Roman style at Bash-Garni) in /z istorii. dokapitalistiéeskix formats (On the history of pre-capitalistic formations), Moscow-Leningrad 1933, p. 639-660. ?2 The inscription has attracted much discussion and comment, principally in: H.A. MANANDYAN, GaFnii hunaren arjanagrut’yuna ew Garnii het°anosakan taéari — karuc’man zamanaka (The Greek inscription of Garni and the date of construction of the pagan temple of Garni), Erevan 1946, and ““Novje zametki o Grecheskoi nadpisi i jazyéeskom Xrame Garni” (New observations concerning the Greek inscription and the pagan temple of Garni), Tgir 1951, 4, p. 9-35; TREvER, Oéerki, p. 174-211; Atak‘elyan, Garni 1, p. 59-68; L. Moretri, “Due note epigrafiche”, Athenaeum 33 (1955), p. 32-46; J. and L. Ropert, Revue des Etudes Grecques 1956, p. 183-4; G.X. SARGSYAN, “Garnii hunaren arjanagrut‘yan masin’’ (On the Greek inscription. — of Garni), Tgir 1956, 3; S.M. KRKYASARYAN and H. BARTIKYAN, “Evs mi angam_ Garnii hunaren arjanagrut‘yan masin” (Once more on the Greek inscription from. Garni), PBH 1965, 3, F. Feyory in M.L. Cuaumont, Recherches sur l'histoire de l’Arménie de l'avénement des Sassanides & la conversion du Royaume, Paris 1969, appendix Iil, p. 177-182. a 23, K.V. TREVER, Ocerki po istorii kultury drevnei Armenii (Studies in the history of the culture of ancient Armenia), Moscow-Leningrad 1953, p. 59ff. This is by no means intended as criticism of the admirable section in her book on the cult of Mithras. in Armenia and the East: it is merely wished to point out the lack of evidence showing the Ionic building of Garni to have been connected with that cult. THE IONIC BUILDING AT GARNI 229 sun-god Mithras to be bathed in his own light?*, But the actual re-building shows this to be fanciful. The building faces North, a most ~ ynsuitable orientation if the light of the sun is to have any significance, and which does not permit a shaft of light falling through the hole the roof to illuminate a statue unless this was placed well forward in e cella. Then the face of the statue would not be lit up, merely its : back. In spite of extensive and thorough excavations around the building, no inscriptions or votive offerings or anything that must necessarily be connected with a temple or its cult were found. The only possible exception is a sculptured fragment which could be a bull’s horn. What was uncovered however was a large number of graves, as well as buildings belonging to the palace or barracks complex. The graves are rather earlier than the palace buildings, and appear to date from 7 e Ist and 2nd centuries A.D. Is it not possible that the Ionic building was in some way connected with the graves, that it was in fact a From early in the history of Classical Architecture there are examples ~tombs which have many of the features of a temple. The Nereid onument from Xanthus, believed to date from the end of the Sth ntury B.C. is perhaps the earliest. On a high podium measuring 7 x 10m stands a tetrastyle Ionic peripteros, six columns along the de, enclosing pronaos, naos and opisthodomos*°, This monument as probably a prototype for the larger and more elaborate Mau- soleum at Halicarnassus. Here the peripteros was probably of nine columns by eleven, and the stylobate c. 27 x 33m?°. Another arge tomb of the same type was found at Belevi near Ephesus. Here the peripteros was square, 8 x 8 columns, and stood on a podium carved in part from solid rock?’. The Mausoleum dates from the 4th, the Belevi tomb from the 3rd century B.C. Tombs based on designs and motifs borrowed from temple architec- 24 SaHINYAN, REArm N\S. VI, p. 194. The building has been restored with such a hole, but all the stonework appears to be new. 25 G, NIEMANN, Das Nereiden Monument in Xanthus, Vienna 1921. 26 Exactly how the Mausoleum should be reconstructed has long been in doubt. € D.S. RoBertson, Greek and Roman Architecture, p. 150ff. and fig. 65, and H.W. Law in Journal of Hellenic Studies LYX, 1939, p. 92 ff. 27 T. Fyre, Hellenistic Architecture, Cambridge 1936, p. 50ff. and Josef KeIL, Jahreshefte des Osterreichischen Archaeologischen Institutes, XXVIU, Beiblatt 25ff.; XXIX, Beiblatt 105 ff.; XXX, Beiblatt 175ff. 230 R.D. WILKINSON ture were frequently built in Roman times as well. Numerous examples are known from Termessus in Lycia?’, (sometimes considered to be ; in Pisidia). The sarcophagus was usually not placed in a chamber jp the podium, as at the Mausoleum and Belevi, but in full view on the | stylobate. The designs are of great variety. The tomb of T.CI. Agripping — for example is like a prostyle tetrastyle temple on a podium, with engaged half columns continuing round the cella. The cella walls between the columns have been removed however, so the impression is of g peripteros*°. Another prominent tomb consists of two rows of four columns along the edges of a podium; between are two thick square pillars, with pilasters at the corners, which support a thick wide arch acting as baldachino to the sarcophagus°*°. These tombs, though elaborate, are quite small. The stylobate of the first measures appro- ximately 5 x 7m that of the second 4 x 5m. They are thought to date from the 2nd century A.D. | Other tombs from Asia Minor are closely modelled on prostyle temples. Two of these have been mentioned already because they had stone barrel-vaults : the tomb between Sura and Myra, and the tomb with palm-leaf capitals from Patara*'. Buildings from Termessus show how little difference there sometimes was between tombs and temples. Among the temples is numbered a tetrastyle prostyle Corinthian building on a podium approached by steps flanked by buttresses, as at Garni*?. Another tetrastyle prostyle Corinthian building, not on a podium, is numbered among the tombs*? And from Palmyra there are several temple-tombs, a particularly grand example being a hexastyle prostyle building of the Corinthian order with a stylobate measuring 14.5 x 18.5m°*. At Iasos in Caria the Italian excavators have discovered a temple-tomb, tetrastyle prostyle on a podium with ten approach steps flanked by buttresses*>. Its size is rather smaller than the Ionic building at Garni. : 28 R. HEBERDEY and W. WILBERG, ““Grabbauten von Termessos in Pisidien”’. Jahres- hefte des Osterreichischen Archdologischen Institutes 111, 1900, pp. 177 ff. 29 HEBERDEY and WILBERG, op. cit., p. 180, figs. 52-55. 30 HEBERDEY and WILBERG, op. cit., p. 187, figs 16 and 62. *1 See notes 15 and 16. 32 LANCKORONSKI! : Vol I], p. 88, figs 43-45; Plate IV. 33 LANCKORONSKI : Vol II, p. 118, figs. 88 and 89. 34 Temple-tomb no 86 in. Palmyra I, Berlin 1932, p. 71-76; II plates 38-44. 35 -4nnuario della scuola archeologica di Atene, Vol. 43-44 (New Series 27-28), 1965 66, p. 469-479; Vol. 45-46 (N.S. 29-31), 1967-68, p. 579. THE IONIC BUILDING AT GARNI 231 ‘Thus it is not always easy to distinguish between tombs and temples f the Roman period. The architecture of Garni could in nearly every espect belong equally to a tomb or to a temple. The podium for xample is of a type found both in temples (the Corinthian prostyle »mple from Termessus; the Ionic temple from the theatre terrace at Pergamum) °°, and in tombs (the tomb of Claudia Antonia Sabina from ardis)*’. The Ionic order was used not only at Belevi, the Mausoleum nd the Nereid monument, but also for the tomb of T.CI. Agrippina t Termessus, while the great temple from Aphrodisias shows the order as still being used for temples in the 2nd century A.D.*°. The vaulted roof, as we have seen, was used both for tombs and temples provided they were sufficiently small. A full peripteros on a building of the size of Garni is unusual, and there are no exact parallels. It is however no more surprising that Garni should be peripteral if the building was a lavish tomb than if it was a humble temple. A larger mb than the Ionic building at Garni is that of Claudia Antonia Sabina om Sardis*°; a smaller temple is that of Artemis from Termessus*°. either, however, is peripteral. One feature of the architecture of the Ionic building at Garni which closer to a tomb than it is to Graeco-Roman temple architecture the proportions of the ground plan. There exists no peripteral smple which is hexastyle with only eight columns along the sides : the sual number is at least eleven*!. The Trajaneum at Pergamum with n exceptionally square ground plan has six columns by nine*+*. Tombs owever have a much squarer ground plan: the Mausoleum was eleven olumns by nine, and the Belevi monument eight by eight. The tomb of T.CI. Agrippina, virtually peripteral, was four by six. At Garni the proportion of the front of the stylobate to the side is approximately :1.4. This is even squarer than prostyle temples such as the Corinthian rostyle temple from Termessus (1:1.45), or the Augusteum from 3° Altertiimer von Pergamon, IV. 37 Sardis, Vol I, part 1, Leyden 1922, pp. 170-174. °8 Society of Dilettanti, Antiquities of Ionia, London 1840, part ITT, p. 45ff. and hapter JI, Plate IIT ff. -3° See note 37. ~*° LANCKORONSKI, Vol. II, p. 90-91, fig. 47, 48, 49 and Plate IX. = For general confirmation of this see the table of principle temples at the end ~Dinsmoor, The Architecture of Ancient Greece, 3rd edition 1950, facing p. 340. “2 Altertiimer von Pergamon, Vol. Vii. D2 R.D. WILKINSON Sidyma** (1:1.7). From Syria the two prostyle temples from Niha have stylobate proportions of 1:1.9 (the larger temple) and 1:2 (the — smaller) **. Prostyle tombs however are squarer : the Corinthian prostyle _ tomb from Termessus has stylobate proportions of approximately — Lele, : In addition to what can be deduced from architectural comparisons, __ there are other reasons for supposing that the Ionic building at Garni could have been a tomb and not a temple. During the course of _ the excavations various fragments of white marble were found not more than 50 yards away from the building. These fragments are either figure sculpture or else pieces of architectural carving. The principal piece of sculpture is a draped torso, preserved from the neck to the knees, and some 50cm high. The back and left side are very roughly worked which shows that the figure was originally up against a wall*®. There are two other fragments of marble figures: one a block of marble bearing traces of a pair of feet*’; the other a second torso, preserved from neck to waist, some 15cm high. The arms here are folded beneath the drapery, the left forearm crossing the chest to support the right elbow, the right forearm rising towards the chin which it must originally _ have been supporting *®. Other insignificant fragments of drapery have © also been found. The architectural fragments include a section of | stylised cornice, with sima, dentils and egg-and-tongue, 10cm high and 7cm long*?, and a length of spiral-fluted column, whose propor- tions match the fragment of cornice®®. The column is rough and unfluted in one place which shows that it too must once have been standing close or lightly attached to a back wall. These marble fragments, both the figures and the architectural decoration, clearly came from one or more sarcophagi of the type known as Asiatic or Sidamara*’. The size of the fragments of sculpture, 43 Termessus, v. note 32; Sidyma v, note 14. 44D. KRENCKER and W. ZSCHWEITZSCHMANN, Rémische Tempel in Syrien, Leipzig 1938, p. 1OSff. and plate 58. 45 See note 33. *© B.N. ARAK’ELYAN, “Kandakagorcut’yun hin Hayastanum” (Sculpture in ancient Armenia), PBH 1969, no. 1, pp. 63-64, fig. 9. 47 ARAK’ELYAN, loc. cit. 48 ARAK’ELYAN, Garni, II, p. 72, fig. 39. a? Joid., Dx 74, Tie, 40. 5° Ibid., plate XIX. >! For a full study of sarcophagi of the Asiatic or Sidamara type see C.R. MOREY, _ THE IONIC BUILDING AT GARNI pee jn particular that of the larger torso, the attitude of the other torso with its hand raised to its chin °*, the twisted column, the stylised carving _ of the cornice, and the fact that both the figure and the column were close to a wall, or even lightly attached at the back all show that the fragments are the remains of a marble sarcophagus, or possibly more than one marble sarcophagus of that type. Although so little of ‘the Garni sarcophagus remains, a very close parallel can be found for the cornice fragment. Two sarcophagi from Denizli (on the borders of Caria and Lydia) now in the Louvre have their cornice carved in jdentical manner to that from Garni°*. The sima is covered with a flat Jeaf decoration, the space between the leaves being deeply gouged out with the drill with the result that at intervals the leaves are pock- marked with round holes. Below the sima come two flat fillets, then the dentils, square and close together, and below that an egg-and-tongue ‘whose eggs are carved so deeply that they are almost in the round. e two sarcophagi from Denizli are dated to c. 175 A.D.°* and the remains of sarcophagus from Garni must be almost exactly contempo- rary with them. __.. Therefore in addition to a number of humbler graves of the Ist-2nd century A.D. remains of at least one elaborate Asiatic sarcophagus have been found. Such sarcophagi were not buried in this ground; they were placed in tomb-buildings where they could be seen. The Tonic building at Garni would have provided an ideal receptacle for such a sarcophagus, and the temptation to consider the two connected very strong. One rather surprising fact about the Ionic building at Garni is that it has survived at all. What is more it remained virtually intact until the seventeenth century, and not in some remote spot where it _ might have escaped notice but within the walls of one of the most important fortresses of the land. But the pagan temples were deliberately __ destroyed after the conversion to Christianity °°. Those who consider the Ionic building to have been a temple are therefore compelled to say The tomb of Claudia Antonia Sabina and the Asiatic sarcophagi, Sardis V, part I, Princeton For a figure in this pose v Morey, op. cit., plate 28. Ibid. plates 26 and 27. Ibid. p. 81-82. a For discussion of whether churches were sometimes built in their place see _ A. KHATCHATRIAN, L’architecture arménienne du IV au VI siede, Paris 1971, p. 32ff. 234 R.D, WILKINSON that it was a happy stroke of luck that the temple at Garni was spared while all the others were destroyed >*°. But if the Ionic building was a tomb and not a temple, its survival then appears as much less of q _ miracle. Moreover it is perhaps significant that Moses of Chorene in his description of Garni, which appears to reflect someone’s persona] knowledge if not necessarily his own, does not speak of the Ionic © building as a temple but as a summer residence*’. But on numerous _ occasions he talks of Armenian temples, and certainly gives the impres- _ sion that he knew what they had looked like>®. Yet it does not occur to him that the Ionic building at Garni might have been one of them, That may be because pagan temples in Armenia were not hexastyle peripteral buildings of Graeco-Roman type >? | Thus the proportions of the ground plan, the fact that remains of q © late 2nd century sarcophagus were found nearby, and the circum-._ stances of its survival all suggest that the Ionic building at Gaini was not necessarily, as is generally supposed, a temple, but could have been a tomb. THE BUILDING’S STYLE AND DECORATION It is clear that, whatever its function, the Ionic building from Gafni_ is of basically Graeco-Roman style. It has however been argued that — in the details of its architecture it shows itself to be a truly Armenian building, albeit based on Graeco-Roman models, and that it incorporates a number of purely Armenian features which distinguish it from th real Graeco-Roman buildings of neighbouring lands to the West an which mark it out as a true monument of Armenian architecture °° °© Eg. SAHINYAN, Kasali basilikayi, chapter 8. 57 MOSES OF CHORENE II, 90. i 58 MOSES OF CHORENE II, 9. Of course if Moses was an 8th century — author as — _ is now widely believed, this point loses most of its force. e 59 Neither of course as a general rule were Armenian tombs. The intention is to” show that the Jonic building at Garni could have been a tomb of what was for Armenia though not for the Roman Empire, an unusual type. °° The beginning of that theory can be seen in K.V. TREvER, Nadpis o postroennit armjanskoi kreposti Garni (The inscription concerning the construction of the Armenian fortress of Garni), Leningrad 1946, p. 6 and Ocerki, 73-75, where an attempt is made to show the influence of Urartian traditions. The theory is developed in full by SAHINYAN, - K‘asali basilikayi, chapter 8, p. 222 ff., and has most recently been restated by ARAK ELYAN in his Aknarkner hin Hayastani arvesti patmut'yan (A look at the history of ancien Armenian art), Erevan 1976, p. 27ff. THE IONIC BUILDING AT GARNI 2a5 .. This theory if justified would mean that in Armenia in Roman ‘times architectural and constructional techniques had reached such high level of skill that the Armenians were capable of taking a -pasically Graeco-Roman design and constructing a building which was not only sound and attractive in itself, but also bore the unmistakable ‘marks of their own distinctive treatment of their Graeco-Roman model, The presence of such a building would indeed be a powerful argument for the existence in Armenia at that time of a sophisticated civilisation, deeply influenced by the Graeco-Roman world and impregnated with jts culture, but nonetheless preserving ancient building traditions, and ilready giving intimations of the glories and individuality of Armenian mediaeval architecture. _It is clear therefore that these supposedly distinctive Armenian haracteristics of the Ionic building must be examined in detail. Fhey are said to be seen in: 1) the proportions of the ground plan; the use of iron clamps; 3) the quality of the masonry; 4) the barrel-vault yf stone; 5) the details of the order; 6) the technique and motifs f the carving. These arguments will be discussed in that order. 1) The proportions of the ground-plan are indeed too square for a sraeco-Roman temple. It is true that they are much closer to the sroportions of Urartian temples and, probably, other pagan temples the Graeco-Roman period. But the proportions of the stylobate ind the number and arrangement of the columns would be perfectly ceptable for a Graeco-Roman tomb®?. 2) The use of iron clamps bedded in lead is a widespread feature ancient architecture, being used by the Achaemenids, the Greeks d the Romans as well as the Urartians. Thus the fact that such mps are used on the Ionic building at Garni does not prove that in any way reflects the survival of Urartian building technique °? In fact the clamps here are used exactly as they were used in Graeco- Roman work to hold the stones of the cella wall together and to join *! Sahinyan points out the proportions of the cella of the Ionic building are almost lentical to those of the Urartian temple discovered at Arin-berd (v. SAHINYAN, “L’archi- ccture des constructions antiques de Garni”, REArm, VI, 1969, pp. 195-195 and L. OGANESYAN, Arin-berd I, Erevan 1961, p. 34 (in Russian). But since the architecture the two buildings was in other respects quite different and there was at least en hundred years between their dates of erection this is surely a coincidence. °2 Trever’s argument concern