Chapter 4
The defence system
The Roman colony occupied a natural hill, rising to about 1235 m above sea level. The site has in the past been described as oval or rectangular,1 but in reality it is rather heart-shaped. Its overall dimensions are 785 x 990 metres and the area within its walls covers approximately 46.5 hectares (Fig. 18). It is clear that the site was chosen because it combined relative accessibility from the plain to the west with excellent natural protection.2 The hill is not a level plateau but rises gradually from fifteen metres on the west to sixty metres on the east side above the plain, so that the city is literally tilted towards Pisidia, Antiochia AD Pisidiam (PLATE 55). On the west the hill slopes gently, while the sides are steeper on the north and genuinely precipitous on the south-east side, where the river Anthius forms a deep canyon separating the site from the outlying massif of Sultan Daǧ and thus offers additional protection.
Very little archaeological investigation has taken place outside the walls of the city and accordingly little is known about extra-mural buildings. The aqueduct system, which approached the city from the north, is described in a later chapter. Recently a stadium, 30 m wide and 190 m long and horseshoe shaped at the south end, has been identified immediately to the west of the modern track which runs by the western city wall. The contours of the seating and the central spina can still be made out.3 It is unlikely that there was much private housing outside the city wall. The fact that even in the time of Hadrian the newly erected city gate remained well within the line of the city wall, suggests that nothing substantial had been built down the slopes. On the other hand virtually nothing is known about the necropoleis of the city. A trial dig in 1924 on the hill north-west of the city and west of the aqueduct proved unfruitful. According to Robinson’s excavation journal, only a Byzantine burial was found there inside asquare structure described as a sentinel tower. Other Christian burials were discovered below the floor of the large basilica inside the city. The steep cliffs south-east of the city beyond the Anthius are unlikely to have served as a graveyard, although from a distance they appear to have been some of the main quarries for building stone used at Antioch.4 Many doorstones, sarcophagus fragments, parts of built tombs and funerary inscriptions have been noted in the buildings of Yalvaç. Most have naturally been removed from their find-spots, but we noted the remains of massive built tombs in the Kizilca Mahalle, south of the site, with representations of a door carved in low rectangular stone walls, carved from enormous blocks, which cannot be far from their original positions.5
Fig. 18. City map of Pisidian Antioch.
55. Antioch from the north. The black shadow in the centre of the hill marks the imperial sanctuary and the ridge around the edge indicates the city walls. The aqueduct piers are visible to the right.
The whole city was protected by a wall which followed the edge of the city hill. In 1982 several sections, usually consisting of a single row of blocks, were visible and have been shown by an uninterrupted line on the city plan (PLATE 56). Elsewhere the wall line has been inferred from ridges in the ground, rock cuttings or other indications. Without excavation it was not possible to determine the wall’s width. In the description of his 1826 visit Léon de Laborde gives the impression that more could be seen then than now. He suggests that some wall sections were made from well dressed ashlar, some from rusticated masonry, and others from a patchwork of reused material.6 These observations can now be confirmed by the clearing work carried out by Yalvaç Museum, which has exposed short sections of wall built from reused masonry as well as good ashlar blocks of substantial dimensions (typically 1–1.1 by 0.75–1.75 m) (PLATE 57). As yet there is no evidence for fortifications at Antioch before the late Roman period.7 The ashlar masonry may derive from the walls of the original Augustan colony, or even from its hellenistic precursor, but the use of spolia points to extensive rebuilding in the late Roman period, either in the mid-third century AD, as at Antioch’s sister colony Cremna,8 or around AD 400, as at the Pisidian metropolis Sagalassus.9 The walls may also be compared with the fortifications of Aphrodisias. The situation and size of the two sites are not dissimilar, although Antioch stands on a hill, while Aphrodisias is in a plain. Both sets of fortifications were built at previously undefended sites; both may have enlarged the previously built-up area; and both were well constructed, largely from spolia. The Aphrodisias walls have been dated around the middle of the fourth century AD. 10
56. Section of the city wall on the west, made from reused ashlar masonry.
In the north-west corner of the site the Roman bath house seems to have functioned as a colossal bastion (PLATE 58). Its massive walls were integrated into the late Roman circuit, just as the Doric temple and Augustan heroon at Sagalassus functioned as towers in the city wall which was built up against them. Arundell assumed that the defence line terminated in the north-west, where the precipitous hillside required no artificial defence, but we observed traces of wall foundations even along the steepest cliffs, and there is no doubt that the circuit was complete.11
According to a letter written by Callander to Kelsey, W.M. Ramsay had cleared some minor sections of the city wall at the ‘North Gate’ and on the west side in 1913. This ‘North Gate’ has not been identified, although it is extremely likely that one existed at the north end of the cardo maximus, probably to the west of the nymphaeum building which stood at the head of the street. Robinson’s excavationdiary also refers to a trial trench started 100 yards south of this ‘North Gate’, where the excavators found a long architrave and parts of a cornice, but not enough to induce them to continue working there. Today the only entrance on the north side of the site is an archway which can be observed in the north wall of the bath house. This is too small to be interpreted even as a minor entrance to the city, and Mehmet Taşhalan’s recent excavations show that it was a structural feature of the bath house, an external door by which fire-wood could perhaps have been brought into the building without carting it through the streets of the city.12 Just as the cardo maximus almost certainly led to a gate at the north end of the city, so we would expect to find another entrance at the south end, following the projected line of the north-south street to a point where it would have intersected with the fortification wall. In fact nothing is visible above ground at this point, but a minor entrance to the city can be observed about 150 metres to the south east, the southernmost point of the wall circuit and virtually the lowest part of the city, where a stretch of roadway, with grooves cut in the rock to make its surface less slippery, appears to lead through the line of the city wall to the river Anthius. This would have been used by the inhabitants of the lower part of the city, probably one of its poor residential quarters, to gain access to the stream. However, the only genuine city gate identified so far is the triple archway excavated by Robinson, which is situated in a recessed position about half way along the west wall and at the end of a 90 metre stretch of street leading to the decumanus maximus.
57. Spolia and mortared rubble in the city wall.
58. The bath house from the south.
There are no traces of towers in the city wall, apart from the bastion-like projection formed by the bath house. Because the edge of the city wall constantly changed direction, especially on the north and south sides, many wall sections were indented, and this helped to make the use of towers superfluous.13 Arundell mentions an ‘elevated spot with foundations, perhaps the acropolis of the city’, about 300 feet to the south of the sanctuary of the imperial cult.14 The spot indicated is the summit of the site, where there is now an amorphous ruin of the Byzantine period, which may have been a strong point dating to the final phase of the occupation of Antioch. Nothing visible on the surface indicates that there was an acropolis here at an earlier date.
The city gate
The most important structure associated with the city walls of Antioch is the monumental city gate at the west side of the site, which takes the form of a triumphal arch with three entrances. This was excavated in 1924 by the Michigan team, but in 1982/3 only half of the eastern and the two central piers were still visible. We reproduce our ground plan of these, for no equivalent was published in 1924 (Fig. 19a). Since our work the area has been cleared again by Taşhalan and the architectural blocks from the superstructure assembled beside the foundations. The elements of the building are relatively well preserved and Woodbridge was able to prepare reconstructions of both the inner and the outer façade of the building which appear to be correct at the essential points.15 Many of the sculptural elements of the decoration were also reproduced and discussed by Robinson, making this the best illustrated of all Antioch’s monuments.16
Fig. 19a. City gate. Ground plan of remains visible in 1982.
The arch had four piers set on high bases with pilasters at each of the four corners of the piers and round-topped niches on both the interior and exterior faces of the piers to accommodate statues. Corinthian pilaster capitals supported architraves with three fasciae and archivolts above the three entrances. The panels between the arches and the upper architrave of the building contained adecorative scheme which deliberately recalled the Augustan propylon. At each end there were pilasters decorated on the front panel with an acanthus design and surmounted by further Corinthian capitals. Two central pilasters, with athyrsus in place of the acanthus pattern, were placed directly above the niches of the middle piers (The Art Bulletin 9 (1926) figs. 79–80). Above the right and left hand arches the space between the pilasters was filled by victories or winged geniuses standing on high pedestals, each holding the end of a garland which was supported in the centre, above the apex of the arch, by a bucranion (ibid., figs. 772–8, 81). The panel above the central arch contained two kneeling bearded figures on pedestals, reminiscent of the captive Pisidians depicted on the propylon, each holding a standard on either side of a trophy, placed in the centre above the middle arch (ibid., figs. 70–2). The entablature which completed the decoration consisted of an architrave which carried inscriptions in bronze letters in the upper of its three fasciae. The design of the frieze combined a pattern of acanthus and palmettes (ibid., figs. 90–7) with tritons, hippocamps and weapons, including shields, scabbards and baldrics, quivers and spears (ibid., figs. 82–9). Above the frieze there was an elaborate cornice with lion-head spouts on the sima (ibid., figs. 98–101).
The gateway carried two inscriptions of bronze letters attached to the upper fascia of the upper architrave, one facing into and the other out of the city. The inward facing text illustrated in Woodbridge’s drawing (Fig. 19b) began with the letters C. IVL. ASP. and this is confirmed by one of the Michigan photographs showing blocks from the architrave including the original letters · IVL · A attached at this point (Kelsey archive negative 8.2). Robinson noted that it had also been possible to reconstruct the reading TRIB on another block of the architrave, and several bronze letters were recovered in the excavation.17 On the basis of this Robinson concluded that the gate had been constructed by or under the auspices of the consul of AD 212, C. Iulius Asper, and that the building accordingly should be dated to the early third century. Even if the identification of the person named on the gate with the consul of 212 were correct, there is no need to assume that the building belongs in this year. In fact the most recent study of the evidence for the origin and career of C. Iulius Asper suggests no connection at all with Antioch.18 Waelkens, examining the architectural decoration in 1982/3, argued on grounds of style that it fitted better into the later second century.19
Re-examination of the evidence for the inscriptions by Dr Maurice Byrne suggests that the gate is earlier still. Scrutiny of Woodbridge’s drawing of the outer façade of the gateway shows the text reconstructed as beginning with an imperial title. Byrne, after study of the architrave blocks on the site, which have been re-exhumed by Mehmet Taşhalan, has succeeded in differentiating the inward and outward facing blocks. In a preliminary paper he suggests that the inward facing text began C. Iul. Asp(er) Pansini[an]us IIvir V trib and ended with the verb ornavit, while the outward facing inscription contained the imperial titles [trib. p]ot. XIII cos. III, which are applicable only to Hadrian in 129, to Septimius Severus in 206 and to Caracalla in 210.20 Further work on the notebooks and photographs of the Michigan excavators has now shown that this text also contains references to [H]adriano Au[g.] and to Sabinae Au[g.].
Fig. 19b. City Gate (Woodbridge)
A provisional text reads:
(exterior)
IMP. CAESARI [divi Nervae nep.] DIVI [Traiani fil. Traiano H]ADRIANO AU[g. pont. max. trib. p]ot. XIII cos. III P.P. ET SABINAE AU[g.– –] COL[onia – –]
(interior)
C. IVL. ASP(er) PANSINI[an]VS IIvir V TRIB[unus militum – – –] D(e) S(ua) P(ecunia) F(ecit) ET ORNAVIT
For the emperor Caesar Traianus Hadrianus Augustus, grandson of the deified Nerva, son of the deified Traianus, pontifex maximus, with tribunician power for the 13th time, consul for the 3rd time, and for Sabina Augusta ... the colony.
C. Iulius Asper Pansinianus, duovir quinquennalis, military tribune, made and decorated (the gate) from his own money.
The gate was dedicated to Hadrian and Sabina in AD 129, the year in which the emperor visited the interior of Asia Minor. He doubtless included Antioch in his itinerary.21
Interior divisions of the city and the street pattern
The origin of the street system at Antioch is uncertain. The hellenistic strata remain buried beneath the level reached by the excavations of Ramsay and Robinson, or have been removed as the foundations of the Augustan colony were laid down. However, the excavations of Mehmet Taşhalan have uncovered what appears to be hellenistic masonry in the edge of the side street running south from the Tiberia platea, and this hints at an early origin for the lay-out of the street system (PLATE 59). In fact the block to which this stretch of wall belongs measures approximately 28 by 56 metres, proportions which conform to those of insulae found in other Seleucid cities, notably at Syrian Antioch.22 So far this observation cannot be confirmed by the measurements of other insulae. The side streets on either side of the decumanus maximus do not appear to occur at regular intervals, and further excavation will be needed to establish whether the long, thin blocks typical of hellenistic city lay-outs occur elsewhere across the site. We may, however, safely exclude the possibility that the earlier city occupied a different site altogether from the Roman colony.23
As already mentioned, the site of Antioch is uneven and rises towards the east. It was divided into districts called vici, some named after the vici of the city of Rome itself (see appendix 1 no. 4). Their boundaries were presumably determined by the street pattern, but they cannot now be identified on the ground. For us the city is articulated into three areas by its two main axes, the cardo and decumanus maximus. The cardo maximus ran from NNE to SSW for much of its course at or around the 1210 m contour line, which can be seen as the mid point between the lower, western and higher, eastern part of the town. The decumanus maximus, in contrast, climbed steadily from the level of the western wall to the height of the enclosed square in front of the imperial temple (PLATE 60). The two streets meet about 70 m south of the Tiberia platea. The cardo maximus was the larger of the two streets. The section which has recently been excavated between the intersection with the decumanus maximus and the west end of the Tiberia platea is 6 m wide, narrowing to 5.1 m for a short stretch close to the platea. Its full length was a little over 400 metres long.24 The street edge can be followed northwards from the intersection for about 260 metres until it opens out into an elongated square, measuring roughly 30 x 120 metres, which was closed off at the north end by the large nymphaeum which collected water from the aqueduct (see chapter 6). It is not impossible that this area served as the civic forum of the Augustan colony, as appears to have been the case with the elongated fora of Syrian Apamea and Gerasa in the Decapolis.25 However, several other open areas of the city lay-out could have taken on some of the functions of the forum at various periods in Antioch’s history, including the square in front of the imperial temple within the imperial cult sanctuary and the Tiberia platea below it to the west, the rectangular square to the west of the theatre, which was constructed in the early fourth century AD, and the large, roughly rectangular depression, as yet unexcavated, which lay to the north of the city church.26 One of the main streets of Antioch was known as the Augusta platea. Although Ramsay and Robinson regularly applied the term to the square in front of the imperial temple, the term platea is appropriate to a broad street, usually lined with colonnades and shops, not to an open plaza, and it is not unlikely that this was the name of the cardo maximus in antiquity.27
59. Section of hellenistic wall in a side street of the Tiberia platea.
60. The decumanus maximus from the west in 1995 (Tasçlialan).
There was no straightforward decumanus maximus, but the main street from the city gate had a more complicated arrangement. It did not follow a simple course from WNW to ESE across the site, but included two dog-legs in its passage to the Tiberia platea. Inside the gate a wide paved street runs a little east of north uphill for 90 metres to the intersection with the main stretch of the decumanus maximus. The paving has recently been uncovered by Mehmet Taşhalan, and this operation has revealed that the semicircular fountain, which Robinson’s excavation had discovered in the middle of the street behind the gate, and which could still be seen in 1982/3 (PLATE 61), was in fact situated at the lower end of a series of stepped stone basins, which brought water down the middle of the street in a series of decorative controlled waterfalls, doubtless arranged for their aesthetic effect.28
61. Semicircular fountain basin inside the city gate.
The decumanus maximus itself ran for 320 metres from the intersection at the top of this street to the cardo maximus. Its width varied between 6 and 7.5 m, and its paving stones covered one of the city’s main drains. Four side streets led off at right angles on the north side, respectively 2.5, 3, 3.2 and 1.2 metres wide, but set at irregular intervals. The first is 77.5 metres from the west end; the gaps between the next three side streets are respectively 49, 68 and 62.5 metres, and there is a further 42 metres until the cardo maximus. There is a six-metre-wide side street on the south side, about two-thirds of the way from the western end to the intersection with the cardo.
The most remarkable feature of the decumanus maximus, the tunnel through which it passed beneath the south part of the cavea of the theatre, is discussed below.
As was usual with the plateae of the cities of the Roman East, sections of the main axes at Antioch may have been changed into colonnaded streets during the imperial period.29 During the 1924 excavation several trenches were dug across them. One was around the middle of the cardo, about half-way between the ‘North Gate’ and the city church west of the imperial sanctuary. One of the photographs taken of this trench (Kelsey archive 7. 1483) clearly shows sections of architraves and of a corner cornice which belong to the monumental Hadrianic building described below.
A second excavation was started in the south-west of the city, where two trenches were opened respectively 4.5 and 7.5 metres north of the city gate. Their aim was to cut the street running north from the gate at right angles. The diary mentions only the discovery of a pavement in the southern trench, but the second one was chosen because it extended in front of a ‘pile of native stone seemingly cut from the podium of a temple’ and led to the finding of two rock-cut steps, obviously the foundations of a stairway, on the east side of the street before it reached the decumanus maximus. West of the steps there were two pavement slabs and a pedestal (1 x 1.16 x 0.56 m high) which probably sup-ported a section of a column (diam. 0.58, ht. 0.375) found in the same trench. This part of the street was evidently lined with columns standing on pedestals.30
Mehmet Taşhalan’s excavation has shown that the decumanus maximus was mostly without colonnades, but he has identified four Doric columns and other parts of the entablature from a building close to the intersection with the cardo on the south side. He suggests tentatively that they belonged to a small fountain house, which should probably be dated to the early imperial period.31
As in most Roman cities the main streets were gradually monumentalised, as public or religious monuments were built along them. There is evidence for several such buildings along the cardo maximus. The large nymphaeum which closed off the vista to the north will be discussed in chapter 6, but the survey revealed traces of three other buildings along the higher ground at the east side of the street. Elements of the white limestone entablature of a tall building, including two parts of an architrave (PLATE 62) and a corner cornice (PLATE 63) were found about half-way along the street, evidently the location of Robinson’s trial trench. The architrave, whose longer fragment is 3.8 m long, 0.537 high and 0.875 wide, is simply decorated and has a broken upper edge. There are three fasciae below a badly damaged moulding consisting of an astragal with elongated bead-and-reel design (of second century type) and a Lesbian cymation of which only the lower stems are visible. The cornice (1.306 x 0.878 x 0.244 high) has consoles shaped as rectangular blocks with two fasciae on their sides and a plain soffit below. These are all framed by an egg-and-dart motif and alternate with cassettes, decorated with a variety of rosettes. Above this another egg-and-dart moulding leads to a palmette anthemion. The arrangement with double-fasciated block consoles decorated with soffits was popular in the time of Hadrian.32 This date is confirmed by the anthemion, whose two outer palmettes, and especially the open one with strongly bent leaves, is comparable to the palmettes on the late Hadrianic cornice of the temple of Zeus at Aezani. The small, apparently palmette-decorated leaves covering the lower part of some of the palmettes also occurs in a section of the cornice anthemion of the late Hadrianic gate at Attaleia.33 A Hadrianic date for this building at Antioch seems likely.
62. Architrave from building on the cardo maximus.
63. Corner cornice from building on the cardo maximus.
A fragment of good quality tendril frieze in white limestone (1.52 x 0.56 x 0.48 high) was found closer to the Tiberia platea. It comes from the middle of a frieze since it shows an acanthus bush with tendrils extending in both directions (PLATE 64). The acanthus itself appears later than one on a middle Augustan sarcophagus lid with tendril decoration found at Ephesus,34 and may be compared with the acanthus bush on the tendril frieze of a mid-first century AD sarcophagus from Hierapolis.35 However, the bush on the Antioch fragment better retains the old closed shape than examples of the late Flavian period or the second century AD, which have a more open arrangement.36 The tendrils themselves anticipate those of the Domitianic nymphaea at Miletus and Ephesus, and the treatment of the flowers more closely resembles that of the early Domitianic nymphaea built by Trajan’s father at Miletus and by Laecanius Bassus at Ephesus than that on the late Domitianic nymphaeum of Pollio at Ephesus.37 A date in the third quarter of the first century AD seems most likely.
Part of a grey limestone Doric entablature (width 1.22, depth 0.285 + , height 0.45 + ) with a Latin inscription reading [– –]s L. f. Ser. Frug[i] was found immediately to the north of the junction of the cardo with the Tiberia platea. At least one of the triglyphs has a plain ledge below the taenia instead of a regula with guttae (Fig. 20), a sign that the decoration of the building was never completed. It is impossible to suggest a precise date. The Doric order was used in south-west Asia Minor at least until the reign of Hadrian.38 The partially preserved name of the dedicator of the building, with mention of tribe and filiation, would fit in the first century AD. It appears overall from the possible dates of these three buildings on the east side of the cardo that the monumentalisation of the street spread north-wards during the first and early second centuries AD.39
Besides the public buildings there were also shops or workshops along this street, at least in late antiquity. A small section of a commercial quarter was excavated in 1924 between the city church and the intersection of the cardo with the Tiberia platea. Robinson found at least five late rooms, constructed from older architectural elements including a Doric column from the semicircular portico as well as fragments of Latin inscriptions (PLATE 65). The excavation diary indicates that several of these mention the name of Augustus and they too must have come from the sanctuary and its surroundings. The walls of some of these rooms may still be made out. They are constructed from reused pillars connected by sections of irregular stonework, broken brick and tiles, all set in a whiteish mortar with grey limestone chippings. The excavators found at least five late coins (1 Licinius, 1 Justinian, 3 ‘Byzantine’) and some ‘Roman’ glass bottles. The diary also mentions that the houses were built on monumental Roman foundations, and a single course of good ashlar blocks, still in situ to judge from the position of the clamp holes and belonging to the corner of a tall structure which lay east of the apse of the city church, can still be seen on the spot. The 1924 excavation of the road intersection also produced several fragments of large statues, including two heads of Augustus.40 This statuary must once have adorned an important city centre location and indeed further examples have been brought to light in the new excavations of Mehmet Taşhalan.41
64. Frieze block from building on the cardo maximus.
Fig. 20. Doric frieze with inscription from the cardo maximus.
65. Row of shops at the intersection of the cardo maximus and the Tiberia platea (Kelsey archive 7. 1303).
The theatre and its surroundings
The most notable feature of Antioch’s theatre, its situation astride the decumanus maximus, has already been mentioned. About half-way between the city walls and the cardo maximus, the decumanus maximus reached the southern section of the theatre. The theatre is in a poor state of preservation. When Arundell identified the site in 1833 all the theatre seats had already been removed, leading him to the mistaken belief that the diameter of the cavea did not exceed 150 feet.42 However, the shape of the hill as well as some of the analemma blocks from the south side clearly show that at least in its final phase the theatre was approximately 95 metres wide at the front, dimensions comparable to the theatre at Aspendus in Pamphylia, and larger than those of Sagalassus, Termessus and Selge, the leading cities of Pisidia (PLATE 66). They relate, however, to an enlargement of the original ground plan, for the theatre thereby was actually built above the existing decumanus maximus, which ran beneath it in a tunnel some 54 metres long. In this section the street, which had an entrance on the north side leading into the cavea of the theatre, was little more than five metres wide and spanned by vaulted masonry. At the entrance of the tunnel on the west stood an archway, built at the instigation of the governor of the recently created province of Pisidia, Valerius Diogenes, and dedicated to the emperors Maximinus (Daia), Constantine and Licinnius between AD 311 and 313. This remarkable arrangement is reminiscent of the two-storey passage under the west part of the mid-second century theatre at Myra in Lycia but is essentially without parallel in the cities of Asia Minor or elsewhere in the Roman Empire.
66. The theatre. The masonry in the centre marks the entrance of the tunnel which ran beneath the structure (Ballance).
The original building may date back to the foundation of the colony, or even to the hellenistic period, although no precise date can be suggested without evidence from excavation. The inscription mentioning the construction of atemporary wooden amphitheatre, tentatively assigned to the late first century AD (appendix 1 no. 7), does not imply that there was then no permanent theatre in the colony; it may simply have been unsuitable for the venationes and gladiatorial games which were staged on that occasion. The first epigraphic evidence for a real theatre is dated to the first half of the third century AD (appendix 1 no. 6).
The surface remains show that when the original building was enlarged, the upper seats of the refurbished cavea were constructed over the line of the decumanus maximus. The foundations of the southern part of the cavea can be seen extending south of the street, above what we may presume to have been avaulted tunnel. Such an arrangement can hardly be attributed to the original building phase, when the theatre would surely have been designed in conformity with the street plan. When the theatre was extended above the old street the original facing of the street edge was cut away to give a better hold for the filling of mortared rubble which was inserted between the old wall and a new ashlar facing. The original street was narrowed to just over 5 metres to make the vaulting easier, and the new edge was provided with buttresses to support the vault (PLATE 67). There was also an entrance to a stairway on the north side of the narrowed and vaulted stretch of the decumanus, which led up to the cavea of the theatre. However, the main entrance was now on the south side of the street at ahigher level, where its foundations can still be seen. On the north side of the street there was presumably a vaulted parodos between the cavea and the theatre façade, which has been cleared in the recent excavation.
At its western end the tunnel underneath the vaulted cavea was faced with amonumental arch, with a span of a little over 5 metres, several fragments of which are lying around in the area (PLATE 68). They had been unearthed in 1913 by an exploratory trench dug along the wall of the proscenium and at the corner of the theatral area. Callander’s letter to Kelsey, which refers to this trench, also mentions finding a cistern which contained ‘a lot of sculptured fragments (some good pieces)’. The arch was made of white limestone and had a simple archivolt composed of three inscribed fasciae below a quarter round and a cyme recta. It was crowned by a simple cornice, the central part of which contained an inscribed dedication Concordiae Augg(ustorum) (PLATE 69). The text on the archivolts, which has been provisionally restored from the available fragments by T. Drew-Bear and M. Christol, mentions the construction of the arch itself and some other building at the express initiative of the governor of Pisidia, Valerius Diogenes.
67. The decumanus maximus running beneath the theatre viewed from the east.
68. Fallen blocks from the arch (Ballance).
A provisional version of their reconstruction of the text is offered in Taşhalan 1997, 289:
Concordiae Aug(ustorum)
Felicissim[is temporibus DD NN] Gal. Val. M[aximini et Fl. Val. Const]antini [et Val. Liciniani Lic]inni pissimorum August(o)rum / arcum [– – – por]ticibus e[t – – – orn]atu a fund[ – – M. Val. Diogenes v.] p. praes. instantia sua facere curavit / d. n. m. q. eor[um].
To the concord of the emperors.
For the most happy times of our masters Galerius Valerius Maximinus and Flavius
69. Inscribed cornice block from the arch.
Valerius Constantinus and Valerius Licinianus Licinnius, the most devout emperors, Marcus Valerius Diogenes, a most perfect man, governor, devoted to their spirit and majesty, at his own initiative took charge of the construction of the arch (with the) porticos (and all the) decoration from the foundation.
The imperial names and titles point to a date between 311 and 313 AD.43 Several other fragments of horizontal architraves, one found near the theatre, appear to belong to porticos which were constructed during the same building programme (see appendix 1 no. 10). These porticos almost certainly surrounded the large square (width about 48 metres) which lay west of the theatre and north of the decumanus. It seems likely that this square functioned as a new agora or forum, like the agora at Side or the late fourth-century tetrastoon east of the stage building at Aphrodisias.44 The extension of the seating capacity of the theatre was evidently contemporary with this development. Inside the tunnel we noted marble fragments which probably belong to a restoration or extension of the stage building. They include unfluted columns of grey marble and the fragment of a small archivolt in high quality white marble with grey veins. This had a downward-facing soffit panel decorated with a drilled leaf pattern on the inside (PLATE 70), and has three fasciae on the outside divided by astragals (with late, very elongated reels) below and egg-and-dart moulding and a cavetto with an anthemion, whose palmettes have degenerated to a row of parallel S-shaped leaves. Above the archivolt there was a tendril frieze, whose leaves also suggest avery late date and whose linear style of carving is reminiscent of woodwork, rather than the plasticity of a stone relief (PLATE 71). The tendrils find parallels in some stage building fragments from Phrygian Hierapolis, which almost certainly belong to the restoration of the scaena during the reign of Constantius II (AD 337–61).45 The Antioch fragment may be slightly earlier and belongs to the first half of the fourth century. It thus confirms that the vaulting of the decumanus maximus, the enlargement of the theatre’s seating capacity, and the construction of the porticoed square to the west of it were part of a building programme which also included important alterations to the stage building. These activities occurred shortly after Antioch became metropolis of the newly created province of Pisidia under Diocletian. They provided the city with a larger theatre and a new public square, befitting its new rank. In this respect, the ambitious renovation programme may be compared with what had already happened more than three centuries before when the colony was founded. It may also be compared with the new building activities at Aphrodisias under Diocletian, when that city received new buildings to match its status as the metropolis of Phrygia and Caria.46
70. Soffit panel from the archivolt of the theatre stage building.
71. Archivolt from the theatre stage building.
Mehmet Taşhalan excavated the theatre in 1996 and his illustrated report confirms and elaborates on the interpretation offered here. He suggests that the building had a hellenistic origin and was enlarged to span the street in the late third and fourth centuries. Its use continued into the fifth and sixth centuries. He publishes several newly discovered fragments of the decoration of the stage building as well as a reconstruction of the inscription on the arch.47
Notes
1 F.W. Kelsey, The Michigan Alumnus 31, 18 (1925) 404 (oval); W.M. Ramsay, The Cities of Saint Paul (1907) 248; E.H. Hudson, Theology 5. 37 (1938) 229 (rectangular).
2 Compare Levick, Roman Colonies, 42.
3 M. Taşhalan, Pisidian Antioch. The Journeys of St Paul to Antioch (Istanbul 1991) 33. Taşhalan suggests on the basis of incomplete excavations here that the stadium was built in the hellenistic period, restored in the second century AD, and destroyed by Christians in the late empire. In connection with the stadium we should note the inscription which mentions the building of a wooden amphitheatre at Antioch, appendix 1 no. 7. Stadia were a feature of the larger cities of Asia Minor: Aphrodisias, Cibyra, Laodicea, Aezani and Perge, for instance.
4 For the region’s limestone deposits, see A. Philippson, Handbuch der regionalen Geologie V. 2 (Kleinasien) (1918) pl. 1. The quarries on Karakuyu, identified by Sterrett, Wolfe Expedition, p. 218, as a source of building stone for the city, were probably only used for the buildings of the Mên sanctuary.
5 See the rough map of Yalvaç published in JRS 16 (1926) Fig. 48.
6 Laborde, Voyage II, 113–4.
7 Of the six Augustan colonies discussed by Levick in Roman Colonies, only two, Cremna and Olbasa, certainly had fortifications. Comama and Lystra were apparently unfortified.
8 S. Mitchell, Cremna in Pisidia (1995) 188–94.
9 M. Waelkens, Sagalassos I (1993) 48; II (1993) 15–16.
10 For a recent and much improved plan of Aphrodisias, see R.R.R. Smith and C. Ratté, AJA 100 (1996) 5–33. The walls of the city are discussed by C. Roueché, Aphrodisias in Late Antiquity (1989) 42–3.
11 Arundell, Discoveries II, 269.
12 Taşhalan 1993, 264, see plan on 273. He suggests that the two rooms VI and VII with exterior doors, were connected with the heating system of the bath house.
13 See R. Martin, L’Urbanisme dans la Grèce antique (2nd edn, 1983) 196; F.E. Winter, Greek Fortifications, 103–4, 116, 118–19. Roueché, Aphrodisias in Late Antiquity, 170–1 observes that the bastions at Aphrodisias were probably added to the fortification wall in the mid-fifth century.
14 Researches II, 275
15 See The Art Bulletin 9 (1926) Fig. 1. (inner façade), Fig. 67 (outer façade; with a fanciful reconstruction of buildings in the background).
16 The Art Bulletin 9 (1926) Figs. 70–101.
17 AJA 1924, 435–44; The Art Bulletin 9 (1926) 45–6. The actual letters that survive are now in the Istanbul Museum.
18 K.-H. Dietz, ‘Iulius Asper, Verteidiger der Provinzen unter Septimius Severus’, Chiron 27 (1997) 483–522. His tribe is now known to be the Galeria, while most Antiochenes were enrolled in the Sergia. The main villa owned by the family was eleven miles outside Rome along the via Latina.
19 See AS 34 (1984) 9; accepted by L. Vandeput, The Architectural Decoration in Roman Asia Minor (1997) 39, cf. 89 (acanthus decoration on consoles), 92 (anthemia), 94 (capitals). These elements are all thought to be later than the decoration of the Hadrianic nymphaeum at Sagalassus.
20 M. Byrne, ‘Starting to read the inscriptions on the City Gate of Pisidian Antioch’, First International Colloquium on Pisidian Antioch, Yalvaç 1997 (forthcoming).
21 See H. Halfmann, Itinera Principum (1986) 204–6, who already suggested that Hadrian will have passed through Antioch in the summer of 129.
22 For Syrian Antioch, see D. Feissel, Syria 62 (1985) 77–103.
23 Tuchelt, Festschrift Bittel, 502.
24 Taşhalan 1995, 290–1, with pl. 18–21.
25 The Hadrianic forum of Syrian Apamea measures 110 x 45 m, see J. C. Balty, Guide d’Apamée (1981) 61 and plan at end. At Gerasa, whose plan dates to the third quarter of the first century AD, the oval forum was situated at the extreme south end of the cardo maximus; see J.B. Ward-Perkins, Cities of Ancient Greece and Italy: Planning in Classical Antiquity (1974) 115 figs. 26–7.
26 The contours and size of this depression, measuring about 45 x 30 m, are reminiscent of the shape of the Hadrianic forum at Cremna; S. Mitchell, Festschrift für Jale Inan (1989) 229–45 and Cremna in Pisidia, 56–69.
27 See appendix 1 nos. 1 and 2. The suggestion was made by Tuchelt, Festschrift Bittel, 514–15.
28 A central water channel, beginning at the Hadrianic fountain house below the acropolis, is one of the striking features of the main colonnaded street of Perge in Pamphylia.
29 Inscription 4 in appendix 1 may belong to such a colonnade. Ramsay’s proposed early first century AD date may be too early, and colonnaded streets only became common during the late first and second centuries (Waelkens, Pessinonte I, 123–4). However, acolonnaded street was laid out at Sardis after the great earthquake of AD 17 (G. Hanfmann, Sardis from Prehistoric to Roman Times (1982) 142), and the colonnaded street at Perge in Pamphylia is reported to be of Julio-Claudian date (cf. the Claudian arch from Perge, published by J. Inan, Ist. Mitt. 39 (1989) 237–44. These monumental streets, which appear to have originated in the cities of Syria, including Syrian Antioch, during the Augustan period, may have spread to Asia Minor earlier than is generally appreciated, and Pisidian Antioch, with its Augusta platea and Tiberia platea, could have played an important role in the diffusion of this architectural fashion. See in general K. Lehmann-Hartleben, RE IIIa (1929) 2109 ff. and W.L. MacDonald, The Architecture of the Roman Empire II (1986) 33, 41 ff. For discussion of two examples in Pisidia see Sarah Cormack, in S. Mitchell, Cremna in Pisidia (1995) 123–39, and M. Waelkens, AS 40 (1990) 193 with Fig. 6 (Sagalassus).
30 Compare Waelkens, Pessinonte I, 124, 131.
31 Taşhalan 1995, 288–9.
32 D.E. Strong, PBSR 21 (1953) 124, 126, 128 Fig. 3, 131–3 Fig. 5, pl. XXXIb and XXXVa-b.
33 R. Naumann, Der Zeustempel zu Aizanoi (1978) pl. 56a; V.M. Strocka, Markttor, 43 Fig. 49.
34 W. Alzinger, Augusteische Architektur I, 112; II, 100 Fig. 164; V.M. Strocka, Girlandensarkophage, 898–9 Fig. 9.
35 Strocka, Girlandensarkophage, 900–4 figs. 12–13; Waelkens, Dokimeion, 17 pl. 1, Fig. 2.
36 See, for instance, Strocka, Markttor 26 Fig. 18 (Hadrianic); 27 Fig. 21 (Hadrianic); 40 Fig. 44 (late Domitianic); 43 Fig. 29 (Hadrianic).
37 Strocka, Markttor, 39–40 figs. 43–4.
38 See J.J. Coulton, AS 32 (1982) 55–7; 38 (1988) 139, 143 Fig. 8; S. Mitchell, AS 38 (1988) 54 with n. 5; Festschrift für Jale Inan, 238; and Cremna in Pisidia, 62–3
39 A similar gradual spread of monumental building from one end of a main thoroughfare to another can be followed along the main colonnaded street of Syrian Apamea, which was rebuilt after a major earthquake during the reign of Trajan; see Balty, Guide d’Apamée; J.-P. Rey-Cocquais, Princeton Encyclopedia of Classical Sites, 66–7.
40 W.M. Ramsay, JHS 50 (1930) 253 (now in Konya Museum). The other, finer head from a more than life-size statue was published by D.M. Robinson, AJA 30 (1926) 125–36, and taken to Istanbul Museum.
41 Taşhalan 1995, 291 with figs. 22–6.
42 Arundell, Discoveries II, 273.
43 T. Drew-Bear, XII Aras., 13–17; Taşhalan 1995, 289. Taşhalan’s article illustrates several details of the tunnel area more clearly than they appeared to us in 1982/3. His pl. 7 shows the entrance to the theatre through the parodos and pl. 13 shows that there was a row of small shops on the north side of the tunnel in the vaults beneath the cavea. See further Taşhalan 1998, 348 pl. 7a and b.
44 A.M. Mansel, Die Ruinen von Side (1963) 97–101; K. Erim, Aphrodisias. City of Venus Aphrodite (1986) 88–9; C. Roueché, Aphrodisias in Late Antiquity (1989) 39–42 no. 20.
45 T. Ritti, Hierapolis. Scavi e ricerche I (1985) 107 for the restoration.
46 Erim, Aphrodisias, 32; Roueché, Aphrodisias in Late Antiquity, 15–34.
47 Taşhalan 1998; note pp. 352–5 figs. 11–14 for the decoration, and p. 341 fig. 3 for the arch text.