Chapter 5
Since the initial discovery of Antioch the ruins connected with the semicircular rock-cut area at the centre of the site have attracted the greatest attention and have been interpreted as the remains of a temple by almost all observers since Arundell (PLATE 72).1
The temple
The excavation of the temple in 1913 and 1924
During Ramsay’s excavations at Antioch between 1912 and 1914, attention was directed to the area in June and July of 1913, under the supervision of Callander. In an unpublished preliminary account Callander reported that the ‘Augusteum’ (he took the identification of the temple as an imperial sanctuary for granted) consisted of three parts, the cella, the pronaos and a platform at the west front nineteen feet wide. He thought that this platform might have been reached by a flight of seven steps, each around nine inches high, and might have stood on the same level as the pronaos which stood about six feet above the surrounding piazza. The rock walls of the cella were still about seven feet high and the only exit was a break, four feet wide, on the west side. The rock chamber measured 26' 3" by 18' 6" with walls about six feet thick, and was surrounded by acolonnade of Doric pillars (diameters measured at 2' 9", 2'; 11", 3'; 1" with fluting 4⅛" wide). The capitals were Roman Corinthian, and Callander noted a base, as well as architrave blocks, portions of frieze and cornice and two corner stones from the pediment, which indicated the general effect of the temple.2
72. Ruins of the imperial temple and temenos from the north-east (Ballance).
Fig. 21. Foundation plan of the imperial temple.
In 1924 excavation of the temple area was renewed after the arrival of Robinson’s architect, F.J. Woodbridge, on July 4 and continued until July 10. Robinson’s journal mainly relates the uncovering of blocks from the entablature and other decorated pieces. Some useful deductions can be made from the find spots of these blocks. The fragments of the pediment from the east end contained parts of a door or window frame, and at the bottom of the foundations at the east end the excavators found several pieces of a very elaborate central acroterion, which they were later able to reconstruct, as well as parts of the corner acroterion at the north-east corner. Many of the blocks carried masons’ marks, for instance IH, LI, IE, apparently therefore in Latin lettering.3 Inside the temple Robinson remarked on a long bench of two vertical and two horizontal courses, the upper with a projecting moulding, which at 75 cm was too high to have been a seat, but seemed to have been a table or a shelf which continued on the west and south sides, although no moulded course was preserved there. After the excavation was complete, blocks were rearranged for photography and to facilitate the work of the architect. In particular the blocks in the north-east corner were built up again, to illustrate the architectural order of the building. Woodbridge’s drawings pay close attention to these pieces and when we began work in 1982 it was clear that the north-east corner blocks were still close to one another, where they had been assembled. During further excavation on 11 August the cella was cleared down to the rock, revealing a deep break running diagonally across the floor from the north-east. At the end of the month a complete marble roof tile and part of the roof cornice with an intact lion’s head waterspout were recovered about three metres south of the building. On the north side two items of sculpture came to light, a fragment of a marble tree trunk, presumably a statue support, 25 cm high and part of a marble base with the colossal left foot of a statue (base: 18 cm high and 41 cm deep; foot: 17 cm across). These may have come from the cult statue.
The following discussion of the temple begins with an examination of the material remains which we surveyed in 1982/3, namely the foundations which are still in situ and the blocks from the upper parts of the building, including the majority which were lying around the foundations at the time of our survey, and the smaller number which were taken to Yalvaç or to Afyon Museum after Robinson’s excavations (Fig. 21).
Fig. 22. Sections across the imperial temple.
Pronaos foundations
The surviving visible foundations of the temple, mostly cut from the rock, but including some blocks of the original substructure, occupy a central position in front of the semicircular portico (PLATE 73; Fig. 21). At the time of our field work in 1982/3 material from the temple’s superstructure was scattered all round the foundations, much as it appears in photographs taken at the end of Robinson’s excavations in 1924. In the last few years, however, many of the architectural pieces have been arranged in neater fashion by the staff of the Yalvaç Museum, in connection with the cleaning and excavation programme carried out by Mehmet Taşhalan.
Sections of the euthynteria of the temple, the uppermost foundation course, are to be seen around the south, west and north sides. On the west side this appears as a rock-cut step, whose upper surface was finished with the point of apick or a stone chisel, and which projects between 34 and 46 cm in front of the first built course of stone. The euthynteria appears as a stone step, 6–11 cm high on the south and 15–16 cm high on the north side. The euthynteria on the north side ran at a different angle from the podium of the temple itself. At the west end of the foundation of the northern parastas it is 70 cm beyond the projected line of the north wall of the podium; further east the distance is only 50 cm. There may have been slippage or movement in the wake of earthquake activity.
73. The temple foundations from the south-west.
Two courses of grey limestone blocks remain from the foundations of the pronaos. The blocks of the west wall of the podium, which are set on the natural rock at the south-west corner, were finished with the point of a pick on top but have rough backs and sides. Many of the blocks on the south and all those on the north side have irregular drafted edges with margins of 6–12 cm. Their height varies producing irregular courses, a feature to be noted in late hellenistic and early Imperial construction elsewhere in the region, for instance in the walls of the sanctuary of Apollo Clarius and around the Augustan Doric temple at Sagalassus.4 Irregular rectangular ashlar gave more stability to the foundations than even coursing.
The second layer of foundation blocks was laid 1.75 m behind the first. These also have drafted margins and were finished in similar fashion on the front edge. The four largest rectangular blocks of the wall, evenly spaced, supported the bases of the four columns of the prostylos. The fill behind these foundations for the front colonnade appears to have consisted of small rubble and irregular pieces of grey limestone, which were probably embedded in clay, as in the foundation fill of the roughly contemporary imperial cult temples at Ancyra and Pessinus.5 This differs from the technique of setting rubble in lime-mortar, which had already been introduced at Ephesus by the mid-first century BC.6 It is also found in Flavian buildings at Ephesus and in the pseudo-dipteral temple of the mid-first century AD excavated at Sardis.7 This technique is first found at Pisidian Antioch in the foundations of the Nymphaeum, which probably dates to the later first century AD (see below).
Cella foundations (Figs. 21 and 22)
The west foundation wall of the cella is of heavily weathered rock (PLATE 74). The northern part was clearly lower and thinner than the southern and was deliberately cut back where blocks were inserted to make up the full width. These blocks were finished with a pointed chisel at the front and roughly marked with the pick above. The northernmost of them is set in line with the north edge of the second row of pronaos foundations.
74. Foundations of west cella wall.
The eastern foundation walls are much thicker and the visible cuttings allow conclusions to be drawn about the appearance of the cella wall and the position of the socle mouldings. The rock foundation is best preserved at the north end of the east side. At the highest level beside the central chamber there is a ledge 80 cm wide, on which the floor paving of the cella itself must have been laid. Next to this, and at a level 19.5 cm lower there is a second ledge 1.12 m wide, which must have supported the blocks with the socle mouldings which were visible at the outside of the cella wall (PLATE 75). Six blocks of the socle survive, including the north-east corner piece, whose lower width measures 1.10, fitting neatly on to the available space. East of the platform for the socle another cut in the rock is visible, 29.3 cm below it, and below this there is a further drop of 32.7 cm to the flat rock platform, approximately 1.52 m wide, which extended round the three enclosed sides of the temple. This implies that the gap between the socle moulding and the top of the podium was filled by two steps, both probably 29.5 cm high, and by a thin paving strip, 3.7 cm thick and certainly with a moulded edge on the outside, which would have been laid on the flat platform.8 One step (height 29.9 cm, length 1.23, depth 1.015) was identified among the blocks on the south side of the south cella wall. It must have belonged to the lower course, for it had a peg hole 0.355 behind the front face and the first 30 cm of the top of the stone were smooth and meant to be seen, while the rest was finished with a fine pick to receive a stone on top. The treads of the steps round the temple were accordingly approximately 30 cm wide, roughly equal to their height.
75. Cella foundations from the east.
76. Imposts for vault below cella.
The rock-cut foundations of the podium were completed with built stonework at the south-east corner. They would have been faced, as they were topped, with finely finished cladding stone (about 25 cm thick), none of which has survived.
The interior of the cella foundations contained a vaulted basement, measuring 7.95 x 5.54 m at the base. Around the north, east and south sides there was arock-cut ledge, varying between 48 and 62.5 cm in width and standing 33–36 cm above the rock floor, which slopes slightly down from west to east. The basement room was no deeper than this. On the north and south sides the ledge supported an impost course, of which two blocks survived in situ before they were disturbed by local people between the 1982 and 1983 seasons (PLATE 76). They were 1.06 m from front to back, with a moulding on the front side, and provided the base for a vaulted roof, whose curve can be seen as a mark on the west interior wall of the cella foundations (Fig. 22). The blocks of the vault were supported by a curved ledge, between 3 and 6 cm deep, which was cut into the natural rock at the north side of the east end. There are less clear traces of asimilar cutting at the south end of the west wall. The rock wall of the chamber was undercut on both sides by between 10 and 25 cm, most notably on the south, to accommodate the stone packing of the vault (PLATE 77). The maximum height of the vault in the middle of the chamber cannot have been above 2.50 m. This was not designed as a cult room, as Tuchelt suggested in his 1981 study,9 but was simply an underground cellar, comparable to the subterranean vaulted room located asymmetrically beneath the cella of the Maison Carrée at Nimes or that below the temple of Rome and Augustus at Caesarea Maritima.10 The height of the vault and of the floor which was laid above it, however, was such as to require that the interior cella walls had their foundations about 20 cm higher than the socle blocks of the exterior (see above).
77. Undercut north wall of the vault below the cella.
Reconstruction of the cella walls
The walls of the cella measured 12.50 x 10.65 (maximum external measurements). Many blocks from the walls were noted, although none survives in situ. What follows is a simple catalogue, beginning from the lowest courses. The positions of the blocks are those we noted in 1982/3. In the following lists the measurements are given in the order height, depth (from front to back), length. The + symbol is added to dimensions which could only be incompletely measured. The bracketed compass directions refer to the sides which the pilasters or the adjoining sections of the cella walls faced.
Socle mouldings
The socle had an Attic–Ionic profile (see PLATE 78). Most of those we observed had clamp holes for lateral attachment to their neighbours and dowel holes for vertical attachment 10 cm from the front face. The socle blocks at the north-east corner have now been set in place on the temple, but they rest on the broad stone platform which served as the foundation for the steps around the cella, not in their original locations.11
1. On west side, lying on the pronaos. 0.562; 0.592; 1.187.
2. Against north side of podium. 0.582; depth not recorded; 1.237.
3. North of cella in front of the central part. Height not recorded; 0.521; 1.231.
4. North-east of north-east corner of cella. 0.571; depth not recorded; 1.195.
78. Socle moulding.
5. Behind temple to north-east. Height 0.562; 0.509 + ; 1.055.
6. North-east corner piece, fashioned as a pilaster which projects 0.067 (N) and 0.065 (E) from the wall. Height 0.572; depth 0.865 (N) and 0.98 (E); width of pilaster 0.93 (N) and 0.935 (E); overall width (E) 1.10 (78).
Orthostats
7. Near south-east corner. Corner piece fashioned as a pilaster. The pilaster is smooth, but the orthostat blocks from the adjoining wall had drafted margins except where they joined the pilasters. Height 0.913; width 0.917 x 0.909. A pry hole on top at the edge of the pilaster shows that blocks from the rest of the cella wall interlocked with those of the pilaster, to give greater stability.
8. East of cella near south-east corner. 0.92; 0.639; 1.25. The drafted edges measure 0.022–0.025, giving total margin widths of 0.05. Upper margin 0.036 (PLATE 79).
9. Next to 8. Height 0.914; 0.542; 1.31.
10. East side. Height 0.907 + ; 0.48–0.596; 1.316.
11. Block including the north-east pilaster (fits on to 6). Height 0.939; width 0.907 (E) x 0.927 (N); overall width 1.57.
There are holes for the attachment of veneer on the inside surface of the cella wall attached to the corner pilasters. The full width of the cella wall, with veneer, was approximately 0.70
Cella wall blocks
North-east pilaster – fitting above 11
12. B14. 0.437; depth 0.564 (E), 0.716 (N); width 1.453 (E). Pilaster sides 0.897 (E) x 0.91 (N).
13. B21. 0.431; depth not recorded; pilaster side 0.896 (N).
14. B12. 0.431; width 1.456. Pilaster sides 0.903 (N), 0.896 (E).
15. B11. 0.455; depth 0.87 (E); width 0.85 (N). These are pilaster sides.
16. B16. 0.414; width 0.842 (E); width 1.179 (N) Pilaster sides 0.842 (E), 0.839 (N).
79. Orthostat (no. 8).
South-east pilaster
17. B3. 0.416; depth 0.67; width 1.452. Pilaster sides 0.891 (E), 0.894 (S).
18. B4. 0.429; depth 0.61 (E); width 1.42+ (S). Pilaster sides 0.891 (E), 0.893 (S).
South-west anta and south pilaster blocks
19. B6 (from the anta). 0.419; depth 0.637; width 1.423 (S). Pilaster sides 0.86 (W), 0.845 (S).
20. Be9, south wall. 0.421; pilaster sides 0.771+, 0.837.
21. Be8, south wall. 0.439; width 1.067.
North-west anta and north pilaster blocks
22. B22 (from the anta). Height 0.445; 0.66; 1.05 (anta).
23. B23. Pilaster side 0.859.
24. Be10, north wall. 0.458; width 1.10.
Other cella wall blocks (arranged by heights)
25. B5. 0.306 + ; 0.64; 1.304.
26. B9, east wall. 0.397.
27. B13. 0.42.
28. B7, east wall. 0.425; 0.61; 1.315.
29. B7a, east wall. 0.429; 0.602; 1.317.
30. B25. 0.431.
31. B24, north wall. 0.436; depth 0.72.
32. B15. 0.436.
33. B17, north wall. 0.438; 0.635; 1.309.
34. B26. 0.441; width 1.177.
35. B20, north wall. 0.445.
36. B2 (corner block). 0.448; depth 1.439; width 1.227.
37. B19, north wall. 0.456; 0.64; 1.439.
38. B8, east wall. Depth 0.667.
The top course of the cella walls was decorated with a row of elongated bead-and-reel mouldings.12 The following blocks were noted:
39. Bup2, south side. 0.428; depth 0.445–0.472; width 0.855+ (broken). See Taşhalan 1995, 270 pl. 8. 1.
40. Bup3, north side. 0.437; c.0.60; 1.257.
41. Bup4, north-east corner. 0.434; width 1.21.
From these figures, and in particular from the measurements of the pilaster at the north-east corner, we may infer that the pilasters themselves tapered from awidth of c. 0.93 at the bottom (no. 6) to c. 0.84 at the top. (no. 16). The heights of the courses at the north-east corner, from the bottom, were approximately as follows:
Orthostat c. 0.94; course 1, 0.44; course 2, 0.43; course 3, 0.43; course 4, 0.455; course 5, 0.415; course 6 with bead and reel, 0.43. This gives a total of 3.55 metres, which falls some two metres short of the full height which we may calculate for the cella walls (see below p. 140). The height differences between the courses are insignificant. This is almost perfect isodomic masonry.
Tendril frieze
A frieze with acanthus plants (Rankenfries) ran along the top of the cella wall. This location is definitively demonstrated by the blocks which contain both sections of the frieze and adjoining parts of the pilaster capitals (42, 56, 71, 75). One block illustrates a winged female figure emerging from the foliage, and this presum-ably occupied a central position on one of the short sides (PLATE 83). Tuchelt and others have voiced doubts about the reconstruction of the temple with two friezes, one at the top of the cella wall and the other above the architraves, but the link between the acanthus frieze and the pilaster capitals at the four corners of the cella proves that the former is correctly located in Woodbridge’s reconstruction, while the siting of the garland frieze above the architrave has never been called into question.13 Woodbridge proposed that the acanthus frieze was two courses in height.14 Two arguments support this hypothesis. Firstly, some of the surviving frieze blocks (the upper according to Woodbridge) have a bead-and-reel moulding at the top, while others do not. Secondly, the pilaster capitals at the corners of the cella, up to which the frieze ran, were certainly two blocks high. Mehmet Taşhalan has lately proposed an alternative, namely that the frieze was only a single course high, and that the sections with bead-and-reel decoration rested on the short front and rear walls, while the other slabs ran along the long sides of the cella. This meets an objection to Woodbridge’s reconstruction, that at no point is it possible to observe where the supposed upper and lower courses of the frieze in fact joined,15 and it is supported by the observation that all the surviving parts of pilasters which have sections of frieze joined to them are upper blocks (42, 56, 71, 75). However, it is anomalous that the friezes along the sides of the temple should lack the astragals found on their counterparts at the west and east ends. Moreover a frieze two courses high appropriately matches the pilasters which were also made out of two blocks. Woodbridge’s reconstruction should be preferred.
Frieze fragments with bead and reel moulding
42. TF1, from the west end of portico. 0.56; 0.73; 0.81. Broken on the left. This is a crucial piece for it includes at the left side part of the top of a pilaster capital, proving that the tendril frieze ran along from the pilaster capitals at the top of the cella wall.
43. TF3, from the south-west corner. 0.52; 0.634; 0.912. (PLATE 80).
44. TF5, east end. 0.50; 0.627; 1.172. Illustrated by Robinson, The Art Bulletin 9 (1926), Fig. 9 top.
45. TF6, east end. 0.517; 0.542; 0.915. Illustrated by Robinson, op. cit., Fig. 12 bottom and by Taşhalan 1995, 273 pl. 10.1.
80. Frieze block (no. 43).
81. Frieze block (no. 46).
82. Frieze block (no. 49).
83. Central frieze block from west end of the temple (Afyon Museum).
46. TF9, broken right. 0.491; 0.635; 0.965 (PLATE 81).
47. TF11, by north wall. O.501; 0.576; 0.822. Illustrated by Robinson, op. cit., Fig. 10 left.
48. TF12, west end. 0.44+; 0.57; 0.892. Illustrated by Robinson, op. cit., Fig.8 top.
49. TF13, at north-west, by the podium wall. 0.345 + ; 0.434; 0.934 (PLATE 82)
50. Yalvaç Mus. inv. 1233. 0.50; 0.415; 0.92. Robinson, op. cit., Fig. 6.
51. W a). 0.50; width 1.05.
52. Afyon Mus. E.1484. W b), middle block with female Rankenfigur. Width 0.92. Illustrated by Robinson, op. cit., Fig. 4.
53. Yalvaç Mus. Inv. 1824. W c), middle block with foliage. 0.51; 0.62; 0.93. Illustrated by Robinson, op. cit., Fig. 8 bottom.
(Total surviving length of fragments 10.428.)
Frieze fragments without bead and reel moulding.
54. TF2, broken left. 0.492; 0.65; 0.603.
55. TF4, near east wall at back. 0.352 + ; 0.631; 0.52.
56. TF7, attached to the pilaster at the north-east corner. 0.506; 0.545; 0.74. This fragment, like no. 42, shows that the frieze adjoined the pilasters. [ = 75. Pb2].
57. TF8. 0.507; 0.573; 0.92 (PLATE 84).
58. TF10, broken right. 0.36 + ; width 0.27 + .
59. TF11a, on the west side. 0.50; 0.62; 0.499. Illustrated by Robinson, op. cit., Fig. 10 right.
60. TF14, at north-west corner. 0.501; 0.643; 0.375.
61. W d). Width 0.55.
62. W e). Width 1.20.
63. Yalvaç Mus. Inv. 1825. W f) 0.50; 0.322; 1.11. Illustrated by Robinson, op. cit., Fig. 5, and by Taşhalan 1995, 273 pl. 10. 2.
64. W g). Depth 0.29; width 0.80.
84. Frieze block (no. 57).
85. South-east pilaster capital lower block (no. 69).
86. North-west pilaster capital, upper block west side (no. 71).
65. W h). Height 0.50.
66. Yalvaç Mus. A. 1232. 0.51; 0.30; 1.20.
67. Yalvaç Mus. A. 109. 0.49; 0.31; 0.78.
68. A. 107. 0.38+; 0.47; 0.82.
(Total length of surviving fragments 10.177.)
Pilaster capitals
69. Po1, south-east corner lower part. 0.497; width: 0.975 (E side bottom measurement); length: 1.03 (south side bottom measurement). The block has two lateral clamp holes to connect it with the walls beside it but no dowel hole on top. Illustrated by Taşhalan 1995, 272 Plate 9. 2 (PLATE 85).
70. Po2, north-west corner lower part. 0.507; west side measures 0.76 at the bottom and 0.932 top. The block has a large Lewis hole (8.3 x 2.8 x 13.2 cm deep for lifting).
71. Pb3, north-west corner top part, lying north of pronaos. 0.465 (0.385 to the abacus); west side width 1.391 (includes part of frieze); north side width at bottom 0.865 (PLATE 86 and 87).
72. Po3 = Yalvaç Mus. Inv. 1657. 0.442; depth 0.77 + ; width at bottom 1.03. There was a large dowel or a Lewis hole on top. The top and bottom of the block were claw chiselled.
73. Pb1, north-east corner top block, includes beginning of tendril frieze on north side. 0.516 (0.401 to abacus); width on north side 1.07. There were no dowel holes on top, so the architrave was not fixed to the capital.
74. Pb2 = TF7 (section of frieze adjoining edge of capital). Illustrated by Taşhalan 1995, 272 Plate 9. 1.
75. Pb4 in Yalvaç Mus. 0.522; 0.99 (bottom); 0.90 (top).
The construction technique using two blocks to form large capitals has been taken as an indicator of a date before about AD 50.16
87. North-west pilaster capital, upper block north side (no. 71).
Architraves
The heights of the fasciae are indicated in brackets after the overall height.
77. A1, south-west corner. 0.612 (0.184, 0.143, 0.106); width 1.356. Illustrated by Robinson, op. cit., Fig. 25 and by Taşhalan 1995, 272 P. 11. 2 (right).
78. A2, to east of 77. 0.436+.
79. A3, near south-east corner. 0.646 (0.187, 0.142, 0.099); 0.50; 1.561 + .
80. A4, near south-east corner. 0.656 (0.18, 0.143, 0.102); 0.60 + ; 2.484 (PLATE 88).
81. A5, north-east corner. 0.649 (0.181, 0.143, 0.101); 0.38; 2.09 + . The back is pick-marked but straight, the bottom is claw chiselled.
82. A6, north side. 0.652 (0.181, 0.146, 0.096); 0.386 below, 0.61 above; 2.677 (PLATE 89).
83. A7, north-west of cella. 0.651 (0.184, 0.146, 0.097); width 1.545 + .
84. A8, by A7. 0.628 (0.187, 0.145, 0.102); 0.382 below; 0. 964. Claw chiselled below; a wall fragment with no soffit panel.
88. Architrave (no. 80).
89. Architrave (no. 82).
85. Yalvaç Mus. A.426–1. 0.511 + ; 0.47; 0.735.
86. Yalvaç Mus. A.426–3. 0.645 (0.18, 0.142, 0.08); 0.40+ ; 0.39+.
One architrave section is illustrated by Robinson, op. cit., Fig. 25.
Frieze with garlands and bulls’ heads
87. GF1, south side. 0.611; 0.665 + ; 1.37. Op. cit., 12 n. 15a, Fig. 13.
88. GF2, by 1. 0.52 + ; 0.78 (bottom); 1.10. Op. cit., 12 n. 15d, Fig. 16.
89. GF3, south-east corner. 0.595; 1.69 (south side); 1.94 (east side). The block is decorated on two sides; there was no bull’s head at the corner. Robinson, op. cit., 12 n. 15c, Fig. 15.
90. GF4, east side. 0.604; 0.574 (bottom); 1.02 + .
91. GF5, east side. 0.595; 0.375+; 0.856+.
92. GF6, north-east corner. 0.604; 1.221 (north side); 1.03 (east side). Op. cit., 12 n. 15k, Fig. 9, 22, 23. There is a dowel hole measuring 4 x 1.6 cm at the outer corner for attaching the cornice block.
93. GF7, north side. 0.47+; 0.575+; 0.98+.
94. GF8, north side. 0.512+; 0.71; 0.544+.
95. GF9, north of pronaos. 0.516+; 0.827; 1.435.
96. GF10, from pronaos. 0.508 + ; 0.778 (bottom); 1.284 +.The back is rough; the right hand edge is cut obliquely to give a good join with the next block (PLATE 90).
97. GF11, north side. 0.603; 0.872; 1.432 Oblique cut at both ends, back rough.
98. GF12, north-west corner fragment, buried beneath Corinthian capital. Op. cit., 12 n. 15g, Fig. 16. The find spots of the blocks still on the site show that this frieze ran round both the cella and the pronaos of the temple.
99. Afyon Mus. E. 1491. 0.594; 0.36, 0.53; 0.925. Robinson, op. cit., 12 n. 15f, Fig. 17.
100. Yalvaç Mus. A.1842. 0.614; 0.61 + ; 2.54.
101. Yalvaç Mus. A.1843. 0.61; c. 0.70; 1.25 (PLATE 91).
102. Yalvaç Mus. A.1781. 0.60; c. 0.79; 1.34 (PLATE 92).
90. Garland frieze (no. 96).
103. Yalvaç Mus. A.1841. 0.605; 0.75; 1.41.
Illustrated by Robinson, op. cit., Fig. 14; Taşhalan 1995, 275 pl. 12. 2; Mitchell, Anatolia I (1993) 106 Fig. 15.
All the garlands and the bulls’ heads are carved differently. A single team may have worked on the south and east sides. These garlands never rise above the bulls’ heads and the ribbons hanging from the heads are very finely carved (PLATE 90). The blocks on the north side have the garlands knotted above the bulls’ heads, and have much thicker ribbons (PLATE 91 and 92).
Cornice
The row of dentils was visible only on a single fragment from the horizontal cornice, now in Yalvaç Museum.
104. Yalvaç Mus. Inv. A.1866. Height 0.325; width 0.874; length 1.175 (PLATE 93).
91. Garland frieze (no. 101).
92. Garland frieze (no. 102).
Other cornice fragments
105. south-west of pronaos. 0.386; 1.10; 0.615.
106. south of cella. 0.33+; 0.496+.
107. south side. Depth 0.49+; 0.901. Clamp holes for lateral attachment at each end.
108. South side. 0.375; 1.118; 0.448. The decoration has alternating acanthus leaves with open and closed palmettes.
109. south side. 0.364; 0.77; 0.512. Open and closed palmettes with acanthus leaves.
110. 0.374; 1.05+; 0.65. Closed palmette and acanthus leaf.
111. On south wall. 0.436; 1.00; 1.326. Lion’s head between two half acanthus leaves from which closed palmettes spring.
112. By 111. 0.347; 0.94; 0.928. Clamp holes for lateral attachment at each end.
113. By 112. 0.356; 0.99; 0.912. Acanthus leaf and closed palmettes.
114. East side. 0.368; 1.119; 1.027. Complete decoration of open and closed palmettes.
115. East side. 0.395; 1.125; 0.807. Open palmette between two closed palmettes.
116. North-east corner. 0.49; 0.865+ (north); 1.385 (east). East side decoration: closed, open, and closed palmette, acanthus leaf at the corner, with palmettes on the raking cornice. There is a hole (22 x 42.5 x 14 cm deep) and two dowel holes for fixing the acroterion. Also clamp holes for fixing adjacent front section and the next block of the raking cornice to the left. Robinson, op. cit., Fig. 28.
117. Tiny frag.
118. South-west corner. 0.923; 1.561; 1.32.
119. small frag.
120. North-west of pronaos. 0.376; 1.08; 0.504. The letters EI have been carved 9 cm high on the bottom.
93. Dentils (no. 104).
94. Gable end from south-east corner of the temple (no. 122).
121. North-west corner piece. Not measurable, but the hole for fixing the acroterion is visible.
122. South-east corner. Yalvaç Mus. A 1875. Robinson, op. cit., Fig. 27 (PLATE 94). The regular design of the horizontal cornice was a succession of open palmette, acanthus, closed palmette. The raking cornice had only open and closed palmettes.
Gable
123. Yalvaç Museum A.1831. The crown of the gable from the west end. Height 0.61 width 0.55-0.71; length 1.20.
124. Yalvaç Museum A.112. Fragment of false window from the west gable. Height 0.585; width 0.76, length 0.87. Clamp holes on top for fixing to block behind and to left. Illustrated by Robinson, op. cit., Fig. 26; Taşhalan 1995, 27 PLATE 14.
125. Acroterion from the west end of the gable (PLATE 95; Fig. 26).
95. Acroterion from the centre of west gable end (Kelsey archive 7. 1668).
Fragments from cella door?
126. Yalvaç Mus. A.1836. Part of console, perhaps from main door. Height 0.60; length (top) 0.78.
127. Yalvaç Mus. A.1875. Part of door moulding? Height 0.695; length 1.10; depth 1.30. Illustrated by Robinson, op. cit., Fig. 24; and by Vandeput, Architectural Decoration, 304, pl. 71.3.
96. Colum n drums (Ballance).
Columns
Several small fragments were observed near the north-west corner of the pronaos, not necessarily from a single column. The columns were made of relatively small drums (PLATE 96).17
128. Height 1.15 + ; diam. c. 0.944; one surface smooth, no dowel hole. Taşhalan 1995, 268 pl. 5. 2.
129. Height 1.73, diam. 0.896; one smooth surface, no dowel hole. Taşhalan 1995, 268 pl. 5. 2.
130. Height 1.063, diam. 0.894; one smooth surface, no dowel hole.
131. Height 0.545+, diam 0.908; one smooth surface, no dowel hole.
132. Height 0.79+.
133. Height 0.90, diam c. 0.89 (at south-west corner).
Base
Yalvaç Museum. Plinth 1.20 square. Column diameter 0.93. Taşhalan 1995, 269 pl. 53 (PLATE 97).
Capital
135. Height 0.987; lower diam c. 0.82; width of abacus 1.10. There is a Lewis hole (29 x 29 x 12 cm deep) in the top for the architrave attachment. Taşhalan 1995, 268 pl. 5. 1; Mitchell, Anatolia I (1993) 106 Fig. 16 (PLATE 98).
Roof
136. Marble roof tile. 83 cm long.18
97. Column base (no. 134).
98. Column capital (no. 135).
The reconstructed ground plan and elevation (Figs. 23 and 25)
The dimensions of the temple are not easy to establish, since everything has been reduced to the foundations, and not all of these are visible. The maximum visible dimensions from the foundations of the north-west parastas to the rear of the podium, and from the north to the south side of the podium footings are 26 x 15 m. The length of the temple itself from the front of the prostylos to the rear wall of the cella is 19 m and the width is 10.10 m, but this excludes the two steps which ran round the outside of the cella walls, as well as the staircase which began between the parastades and led up to the prostylos at the top of the podium. The rock foundations of the podium itself are 13.56 m wide at the east end. The dimensions of the cella were approximately 12 x 10.10 m and the pronaos was 7.70 m deep. The basement room beneath the cella is almost arectangle (north side 7.93, south side 7.95, west side 5.6, east side 5.54).19
A comparison between the rock-cut foundations, especially at the north-east corner of the cella, and the blocks of the socle course at the foot of the cella wall (see below) indicates that the socle, with a maximum lower width of 1.10, rested on a foundation 1.12 m wide. The walls of the cella itself, including a veneer cladding inside, whose presence is implied by fixing holes but which does not survive itself, were probably about 70 cm thick. Two steps, the lower with a tread 30 cm wide, ran round the cella. The width of podium on which the temple stood is not precisely known, for none of the cladding blocks which would have hidden the foundations survives to show their thickness. But the foundations themselves extend c. 1.20 m beyond the putative line of the lower step at the east and north sides of the cella, and the cladding blocks, certainly including amoulding at the top of the podium, would have added at least another 25 cm to this figure.20 We may therefore imagine a platform about 1.50 wide running round the whole building.
Fig. 23. Reconstructed ground plan of the imperial temple.
The arrangement of the columns in the pronaos appears to be an insoluble problem. Clearly there were four columns at the front resting on the foundation course visible today (although there would have been some 2 m more of the foundation courses before the visible column bases of the prostylos). One Attic–Ionic base, now in the Yalvaç Museum but almost certainly from the temple (catalogue no. 134, Pl. 97), has a plinth 1.20 square and supported columns with a base diameter of c. 93 cm. Taşhalan in his discussion of the plan assumes that four such columns were placed with interaxial spaces of 2.90, 3.20 and 2.90, leaving alarger gap between the middle pair of columns, but his reconstruction drawings, like those of Woodbridge, indicate equal spacing. Without positive evidence to the contrary, equal spacing would be the safer assumption. We would thus have four plinths 1.20 wide with three spaces between them of 1.765, giving a total width of 10.095 m. The interaxial spaces on this reconstruction would be 2.96.
Three alternative arrangements for the columns behind the prostylos have been proposed (Fig. 24). The distance between the front of the temple and the front of the cella was 7.70. Taşhalan assumes three columns along the sides; Tuchelt has two, while Woodbridge, in an unpublished plan, indicates a double prostyle of four columns. Taşhalan assumes interaxial spacing of 2.90 between the columns on the side, leaving a gap of 1.80 between the plinth of the rear column and the pilaster at the front of the cella, but his figures are based on a false measurement of 8.80 for the length of the pronaos. This is 1.10 too long. His three-column arrangement would only work with much smaller spacing between the plinths, about 1.37 (interaxial spacing 2.57). This seems cramped, and we prefer Tuchelt’s and Woodbridge’s two-column arrangement, allowing 2.70 gaps between the plinths and interaxial spacing of 3.90. Neither of these arrangements allows for the same spacing between the columns as at the front of the temple. The matter would be decided if we could identify an intact architrave from the pronaos, but this has not so far been possible. It is also difficult to decide whether this second row comprised four columns behind the first, or merely the two columns of the peristyle. Tuchelt, preferring the latter arrangement, concludes that the roof of the pronaos, which covered a wide unsupported span, must have been made of wood. A width of 10 m could readily be spanned by awooden roof and double front colonnades are unusual in Roman podium temples of the Augustan period. Tuchelt’s proposal is more plausible than Woodbridge’s double colonnade.21
There was evidently a staircase leading from the paved square in front of the temple to the level of the prostyle bases. The latter must have been at the level of the lower of the two steps on which the socle of the temple stood. When we recorded levels on the temple area, the socle foundations were measured at between 3.53 and 3.59 above datum at the east end of the cella, and the steps were 0.592 below this, say 3.00 m above datum. The top step of the euthynteria and the foundation of the parastas at the front of the temple were measured at 0.625 and 0.60 above datum. Since then Taşhalan’s excavation has revealed another step of the euthynteria, some 20 cm lower than the one we measured (Taşhalan 1995, 269 pl. 6). So the height difference covered by the staircase should have been at least 2.60 and probably close to 3 m. We may assume that the first riser would have been inside the line of the north and south parastades, which were 4.35 m in front of the prostylos. This suggests that we should reconstruct twelve steps, with treads of c. 36.5 and risers of c. 26.5 cm.22
There will have been another staircase leading from the floor of the pronaos up to the floor of the cella. The foundation of the latter, measured in the north-east corner, was at 3.725 above datum, to which we should add c. 10 cm for the floor thickness. The level in the pronaos in front of the cella was about 80 cm lower (allowing that the socle was placed around 3.60 above datum and that there were two step courses, around 30 cm high, around the cella walls, see below). The full width of the socle was about 1.10, and we may therefore suggest three steps with treads about 36.5 wide and risers of 26.5, dimensions with correspond with those proposed for the front of the podium.
Fig. 24. Alternative reconstructions of imperial temple plan.
Fig. 25. Imperial temple elevation (after Woodbridge).
So much may be concluded about the ground plan. The elevation of the cella may be reconstructed on the basis of the numerous surviving elements from all parts of the decorative order (Fig. 25). These have been listed in catalogue form above, with notes appended where they affect the question of the reconstruction. The question of the height of the cella may only be resolved approximately.
Fig. 26. Acroterion statue of the imperial temple (Woodbridge).
Tuchelt suggests that the column shafts would have been eight times as high as their lower diameter of 92 cm, so 7.36 or 25 Roman feet.23 Base and capital bring this figure up to 8.792 or 30 Roman feet. This would have corresponded to the following courses on the cella walls: two steps, 59 cm; socle, 56 cm; orthostats, 92 cm; tendril frieze, 2 x 50 cm; archi-rave, 65 cm; giving a total of 3.72 m and leaving a height difference of 5.07 m to be made up by the blocks of the cella wall. These average a little less than 43 cm high, so there would have been 12 courses. Woodbridge (The Art Bulletin 9 (1926) Fig. 3) has fourteen courses, while the reconstructed elevation by Taşhalan, 1995, 262 drawing 6, assuming only a single course of frieze, shows fifteen courses. The reconstruction of the courses of the cella wall rests in any case on the unproven but reasonable hy-pothesis that the column height was eight times the lower diameter.
99. Julio–Claudian inscribed architrave.
There were apparently no inscrip- Julio-Claudian inscribed architrave. tions on the temple. An unpublished reconstruction of the front elevation by Woodbridge shows a dedication DIVO AUGUSTO carved on the door lintel, but nothing was found by the excavation to warrant the proposal and it was excluded in the published version of the elevation of the temple façade (The Art Bulletin 9 (1926) Fig. 3). In JRS 2 (1912) 104 no. 41 W.M. Calder published a Latin text reading [pontifex m]ax. cos. IIII trib. [pot. . . ] in letters six inches high carved on what he described as a cornice found in a street near the market in Yalvaç. The same inscription is reproduced in one of Ramsay’s 1912/13 notebooks and this shows the text on the upper fascia of an architrave with mouldings carrying egg-and-dart and palmette designs, whose profile is virtually identical to that of the architraves of the imperial temple. There is a strong temptation to assign the inscription to the temple, but the attribution will not work. The block survived until the 1980s in the location where it was seen by Calder and Ramsay, serving as the doorpost of a house. In 1995 it had been moved to a petrol station on the road running west out of Yalvaç, where it served as one of the sides of a vehicle inspection bay. The photograph shows that the decoration, unlike that on the temple, was left incomplete, with no bead-and-reel design carved on the fillets above the three fasciae (PLATE 99). At 73 cm, it was also taller than the temple architrave. The style of the carving nevertheless suggests a Julio–Claudian date and the titles would suit Tiberius between AD 21 and 31. It presumably comes from another very imposing early imperial building which has yet to be identified.
The porticos and the Tiberia platea
The semicircular portico which forms the backdrop to the imperial temple at Antioch, partly constructed but largely cut from the natural rock of the hillside, is one of the most striking features of the site and attracted the attention of all its early visitors (Fig. 27). G. Weber published a sketch plan,24 but it remained for T. Callander to establish further details during Ramsay’s excavations of 1913, which are supposed to have cleared the whole area. In a letter to Kelsey written in 1924 he provided an account of his work, which had remained unpublished. The central rock wall reached a height of 21' and was about 360' long. The level rock pavement in front of the rear wall, on which the stoa rested, was 19' 4" wide and columns and architectural pieces were located along the line of the stylobate. The stoa was two storeys high, with a bottom storey of 12’ and an upper one which was probably lower. In the northern section of the semicircle, where the natural rock is lower, the sweep of the crescent was made from stone blocks, and the rock face was marked by horizontal bedding lines, which show where the courses of masonry ran. Patches of smooth, hard stucco were also noted on the rock face (they had disappeared by 1982) to indicate how the rear wall of the stoa was finished. At the front of the stoa there was a rock-cut step 1' 1½" wide and 11" high leading down to the level of the pavement round the temple.25
The rock-cut features of the portico are still visible and Callander’s account can be checked against our own observations. At a–a’ on the plan (Fig. 28) the vertical face is 7.14 high. The line of large beam holes, 40 cm high and 25 cm deep, which supported the wooden beams of the second storey floor runs at c. 3.90 above ground level. Below this, at 3.57, is a ledge 5 cm wide which supported the ceiling of the first storey and hid the wooden beams from view. The drawn section also shows three smaller rectangular holes cut into the rock face, two in the first and one in the second storey. Many more such holes are visible in the rock face. Their size varies.26 They are too large for fixing veneer to the wall and were probably designed to support wooden scaffolding set up during the construction of the stoa.
The beam holes run round the southern section of the portico from a point immediately east of the north-east corner of the temple. They are mostly 30 cm high, 30–55 cm wide and about 70 cm apart. Their spacing corresponds approximately to the spacing of the triglyphs in the Doric frieze of the lower order, which were c. 65 cm apart. The greater width at the rear is explained by the fact that the stoa was curved and the beams splayed out like the spokes of a large wheel. The beam holes in the southern section are interrupted between U1 and R. Here the rock is cracked and flawed, and there are traces of working, where it was cut and levelled to take a masonry cladding (PLATE 100). At this point the beams were supported by this masonry, which has entirely disappeared.
Fig. 27. Plan of curved portico.
Fig. 28. Cross sections of the curved portico.
The same phenomenon is to be observed in the whole northern section of the portico. PLATE 101 shows the point at which the vertical rock face with beam holes was taken up by a built section. Here the rock is cut in a series of horizontal ledges, corresponding to the horizontal courses of the masonry set in front of the rock face. No doubt stone was in fact systematically quarried from this area, for use in the construction of the temple itself and the surrounding porticos. In this section the quarried face was not smoothed to serve as the rear of the curved portico, but was faced with stone cladding up to 95 cm thick. A fragment of the ancient construction survives at the north-west corner, where a white limestone block with a 6 cm projecting rim serves as the foundation for the grey limestone block from the lowest course of the wall above it.
100. Portico. Rock face without beam holes.
101. Portico. The junction of the vertical rock face with the built section of the portico
For the construction of the built section of the portico we are entirely dependent on the observations in Callander’s letter, on Robinson’s 1924 excavation diary, and the recent clearance work of Mehmet Taşhalan. Robinson’s team uncovered the lower part of fourteen Doric columns in situ (PLATE 102), several displaced Doric columns and capitals, architraves, frieze blocks with triglyphs and a tiny piece of Doric cornice. Woodbridge’s unpublished plan notes that the positions of sixteen columns could be determined. These finds allow the entablature of the lower order to be reconstructed with two triglyphs between each inter-columniation, an arrangement which is found as early as the fifth century BC and is paralleled locally.27 The columns, with a lower diameter of around 54 cm, had lightly reeded lower sections and the regular twenty flutes above (PLATE 103).
Very little was recovered by Robinson’s excavations from the upper storey apart from four Ionic column shafts and a single volute from an Ionic capital.28 Taşhalan in addition illustrates a small Attic–Ionic column base and reports fragments of a tendril frieze, and this allows him to reconstruct the order.29 Two of the Ionic column shafts have rectangular holes cut in them, perhaps to fit awooden balustrade. The design of the second storey departs from the commoner practice of installing half-columns attached to rectangular piers, which were abutted by a stone balustrade.
The upper storey was reached by a staircase which ran directly up from the east end of the north stoa alongside a rectangular ground floor chamber.30 It is also likely that the upper storey could be entered from the rear through entrances on the hill slope.
102. Doric columns uncovered by the Michigan excavation (Kelsey archive 7. 1592).
103. Doric column from portico (Kelsey archive 7. 1554).
Robinson’s excavation diary contains information about other finds from the area. Behind the temple towards the south there were remains of later brick constructions. Robinson thought that the bricks themselves might originally have been used for flooring the upper storey of the colonnade. Small finds included coins ranging from a bronze of Augustus, recovered from the pavement of the colonnade itself, to later specimens of Constantius, Theodosius (2) and ten or eleven Byzantine coins including one of Justinian. Three large pithoi were found in what appeared to be an artificially enlarged fissure in the rock at the back of the colonnade directly behind the temple. Sculptural finds included a small female figure with her hand to her breast, a headless bust and a small altar. There were several lamps, glass bottles, items of pottery and bronze, and a bronze tube about an inch long containing a tightly rolled silver leaf (9.5 x 3.5 cm) inscribed with a thirteen-line magical Greek text, which Robinson was to publish many years later.31
The recent clearing work has clarified the relationship of the semicircular portico to the north and south stoas. The corner between the south stoa and the short perpendicular stretch of the east stoa which led to the two-storeyed semi-circular recess behind the temple, was marked by a heart-shaped Doric column.32 There is no evidence either for a double colonnade on either side of the main square or for rooms behind the side stoa.33
Robinson’s excavation diary reveals nothing about the architectural decoration of this large square in front of the temple. Virtually all the paving and other stonework has been removed and it is only possible to trace the north and south edges from a few surviving foundation blocks. At the time of our survey our measurements appeared to show that the two side stoas were not parallel but opened out from one another as they receded from the temple. However, further examination of our own plans, and the measurements made by Tuchelt and by Taşhalan suggest that this was not the case. The phenomenon is mainly apparent along the foundations of the south wall, while the north stoa appears to be laid out at right angles to the perpendicular axis of the temple (Fig. 27). The slight deviation from 90o on the south side may have been rectified by the stones of the facing wall, or our measurements may simply have been in error.
The area in front of the temple was therefore a rectangle measuring 83 x 66 m, and faced by stoas 4.80 m deep. The temple itself stood at the centre of a semicircle with a radius of 16.5 m, whose centre was exactly in the middle of the prostyle columns. On the opposite, east side of the square stood the three-arched propylon, whose four piers occupied an area roughly equal in depth to the stoas.
The propylon
The entrance to the sanctuary was through a grandiose, triple-arched gateway, lavishly decorated with sculptures which alluded to imperial victories. Robinson discussed the latter at length in his article on ‘Roman Sculptures from Colonia Caesarea (Pisidian Antioch)’ in which he also published a reconstructed elevation of the propylon by Woodbridge.34 So little of the foundations could be made out on the spot in 1982/3 (see Fig. 29) that it appears fruitless, on the basis of observations made then, to attempt to control Woodbridge’s reconstruction in detail or to publish a ground plan. The area has been re-examined recently by Mehmet Taşhalan, and he has already suggested an alteration to the ground plan proposed in one of Woodbridge’s unpublished drawings. Woodbridge suggested double piers facing east and west, with a gap between them. Taşhalan prefers monolithic piers with a half Corinthian column projecting at the front (as shown in our Fig. 31). This is better suited to the likely arrangement of the text of the Res Gestae, if, as seems likely, that was inscribed in ten columns on the inner faces of the two central piers.35 A fresh, detailed study of the architectural elements, taking into account the material in Woodbridge’s notebooks, is being undertaken by Ingeborg Kader (Munich). It would also be beyond the scope of this volume to attempt a fresh study of the sculpture, which is now housed in the museums of Yalvaç, Afyon and Istanbul. Its significance for the interpretation of the sanctuary is discussed below on pp. 163–4.
A comment is essential, however, on the date of the propylon, which can now be fixed from epigraphic evidence. Robinson’s excavations unearthed six broken sections of the west-facing architrave of the propylon with holes for the attachment of bronze letters, three of which can be traced on the site today.36 The whole inscription, without any differentiation between the parts that had been restored and those that were preserved, was published on Woodbridge’s reconstruction of the façade in the following form: Imp. Caes. Augusto pontifex max. / tribunicia potestate xxii cos. xii. (sic). When we examined the one surviving portion of the architrave in 1982, we believed that after the reference to [tribunic. ] potest. x in the second line we could make out the first two letters of the imperial title Germanicus, which would imply that the inscription did not refer to Augustus but either to Claudius in AD 50 or to Nero in AD 62, both of whom took the title Germanicus and held tribunician power for the tenth time in these years.37 However, the stone itself and another fragment of the architrave has now been re-examined by T. Drew-Bear, who argues for the attribution to Augustus,38 and a fresh examination of our own drawing shows that the reference to Germanicus is an illusion. All six fragments have now been taken into account in a new reconstruction by Dr Maurice Byrne, to be published in detail elsewhere. The following text results:
IMP • CAES[ARI• DI]VI • [F• A]VGVSTO • PONTI[F]ICI • M[AXIM]O COS• X[III• TRIB]VN[ICIAE] POTESTATIS • XXII • [IM]P • XIIII • P[ • P •]
For the emperor Caesar Augustus, son of a god, pontifex maximus, consul for the 13th time, with tribunician power for the 22nd time, imperator for the 14th time, father of the country.
By 1982 the right hand edge of the fragment referring to tribunician power had been broken off, and the number could be interpreted as XXI, XXII, or XXIII. The 1924 photograph resolves this ambiguity for it clearly shows holes appropriate for attaching the figure XXII followed by a punctuation sign, 2/1 BC by our reckoning. The figure for the consulate should accordingly be XIII and for the imperial salutation XIIII.39 The propylon, accordingly, was dedicated to the living emperor Augustus, soon after he received the title pater patriae on 5February 2 BC. It may profitably be compared with an almost contemporary inscription from the western part of the empire, the dedication on the architrave of the imperial temple at Pola in Italy, a building itself which was similar in appearance to the imperial temple at Antioch, Romae et Augusto Caesari divi f. patri patriae (ILS 110).40 The establishment of this date is of critical importance for the interpretation of the entire sanctuary.
The Tiberia platea
The broad, colonnaded street west of the propylon was identified as the Tiberia platea from the discovery in 1924 of one of Antioch’s best-known inscriptions, containing the edict against grain hoarding, issued in AD 93 by the then governor of Galatia–Cappadocia, L. Antistius Rusticus.41 The edict is in essence an appendix to an honorific statue set up for the imperial legate and for the procurator by the Tiberia platea, that is to say the owners of the shops that stood behind the colonnade. According to Robinson the inscribed block was found reused ‘in one of the later north shops which encroached on the square’, and it is so large that it cannot be far from its original position.42 Almost immediately after the end of the 1924 excavation season the inhabitants of Yalvaç, in search of building material, removed the paving stones of the street and the staircase. Very little is left today (Fig. 29), and more than anywhere else on the site we are reliant on the notes, drawings and photographs of the Michigan excavators. There is some room for argument about the street to which the name Tiberia platea was applied (see appendix 1 no. 1), but we have adopted the term for convenience and applied it to the paved stretch between the bottom of the propylon steps and the intersection with the cardo maximus. According to the measurements made by Woodbridge in 1924 the platea was 22.9 metres (about 77 Roman feet) wide and the distance to the cardo maximus was around 85 metres.
Fig. 29. Plan of the Tiberia platea in 1982.
Robinson’s excavation diary shows that digging began on 18 May and continued to 14 June. The coin finds he notes include a coin of Alexander, a Ptolemaic coin depicting an eagle, coins of Augustus, Maximinus and Philippus, the last with a statue of Fides on the reverse, and several issues from the late empire (Diocletian, Constantine, Licinius) and of the Byzantine period with the legend BAΣIɅEΥΣ BAΣIɅEΩN. His entry for 14 June notes the departure of Ramsay for Constantinople taking 200 Roman and Byzantine coins found in the excavation ‘to be cleaned by the British Museum’.43 Small finds included glass fragments, especially drinking vessels from the shops beside the street, one bronze and several terracotta lamps, a bronze crater, a ladle and two bronze spits. Among the pottery was good Roman ‘Samian’ and black and red glazed ware, and fragments of ribbed pithoi, large storage jars which belonged to the shops. There were nails, locks, bronze rings and a bronze saw blade. The nature of these finds suggests strongly that the shops beside the street were small bars and restaurants for the sale of food and drink. That would be entirely appropriate to their city-centre location, at the entrance to the imperial sanctuary. Customers whiled away their time gaming: nearly a dozen scratched circles or rectangles, divided up for dice games,44 as well as two Latin crosses, were noted when the paving stones were cleaned for photography at the end of the season. An unfluted column with a Latin funerary inscription (Septimia Tertia, published by Robinson, TAPA 57 (1926) 236 no. 73), another inscription which Robinson wrongly took to be a hellenistic text (TAPA 57 (1926) 233 no. 66; see above p. 16 n. 28), an altar depicting Cybele and Helios found in one of the shops, and three small domestic altars were recovered. The most important architectural and epigraphic finds were blocks that had fallen from the Propylon, and very numerous broken fragments of the Antioch copy of the Res Gestae of Augustus, the Monumentum Antiochenum.45 Their shattered condition leaves no doubt that the inscription was deliberately broken in antiquity.
Various architectural peculiarities of the street were noted as the excavation proceeded. Almost immediately the Michigan team identified the foundation made of hard concrete or greenish stone of a circular building on the south side of the street about 20 metres from the foot of the propylon steps. Several architectural fragments from this tholos included a curved architrave with an inscription for an emperor Antoninus, probably Caracalla. Immediately east of this, two shallow steps ran across the entire width of the street. These are plainly visible on a photograph taken at the end of the excavation (PLATE 104). At the foot of the propylon steps and set on the central axis of the whole sanctuary, was a raised boss or shield, set on a block 1.70 m square, carrying an inscription in bronze letters: T. Baebius T.f. Ser. Asiaticus aed. III mil pedum d. s. p. stravit (see appendix 1 no. 3). Because this was too large to be easily removed and inconveniently shaped for reuse as a building stone, most of it has survived in place today. Baebius Asiaticus was presumably responsible for paving the two main streets of the colony, whose combined length by today’s measurements approximates very closely to 3000 Roman feet. The other conspicuous block which survives today is a large marble base about 1.15 m square and 32 cm high with a hole in the middle and a rim round the edge, clearly part of a small fountain which stood directly in front of the second pier from the north of the propylon (PLATE 105). Underneath were remains of terracotta pipe-work. The excavators noted a similar arrangement in front of the third pier of the propylon. When the stones were removed from this they found a section of Roman bronze pipe, covered with cement, with a vertical piece at the east end to project the water. Removal of stones at the north and south ends showed that there were also fountains in these positions fed by terracotta pipes. We may conclude that there had originally been four fountains at the east end of the Tiberia platea, symmetrically placed opposite the piers of the monumental entrance to the imperial sanctuary. It may with difficulty have been possible to feed water to these from the pressure system of the aqueduct (see below p. 195).
104. The east end of the Tiberia platea showing two intermediate steps and the staircase leading up to the propylon (Kelsey archive 7. 1113).
105. Part of fountain at east end of the Tiberia platea (Kelsey archive 7. 1476).
106. Late Roman shops with beginning of staircase and column section on the north edge of the Tiberia platea. The foundations of the tholos are in the background (Kelsey archive 7. 1477).
By the time of the excavations little trace remained of the colonnades on either side of the street, but column fragments, two with late graffiti perhaps indicating the owners of the shops (one for Nicaetas son of Petros), were found in the shops on the north side, as well as two column bases, in place at the north edge of the street, but probably not dating to the original period of construction (PLATE 106). Finer details, such as three fragments of small Ionic capitals, a marble gargoyle with an animal’s head, and some minor items of statuary may have come from smaller monuments on or near the street. The area excavated by Robinson on either side of the street has recently been re-exposed by Mehmet Taşhalan, and he observes that the crude construction techniques to be seen in the walls of the shops suggest that in their visible form they are of late date.46
Fig. 30. Details of the tholos in the Tiberia platea (Woodbridge).
Fig. 31. Schematic ground plan of the imperial sanctuary and the Tiberia platea.
On the south side of the street, Robinson’s trenching revealed a well con-structed wall with a cornice extending west from the corner of a side street for 9.56 metres. A water conduit ran from east to west in front of this. The wall may have served as a platform for another important public building.
The tholos
Robinson observed in his preliminary report that enough architectural frag-ments were recovered from the circular building to warrant a restoration. In Woodbridge’s reconstructed plan it appears as a small rotunda with eight slen-der columns set on a square plinth surrounded on three sides by stone benches, the whole thing about 5.20 square (PLATE 107). It would be hazardous to attempt more detail on the basis of the drawings that have been preserved, but by combining Woodbridge’s notebooks with Swain’s photographs, there is a little more to be said. One drawing shows a section of the seating 88 cm long and 40 cm high with a carefully carved lion’s foot at the corner. The block was 1.52 deep with an anathyrosis at 25 cm from the front, thereby indicating the width of the seat (Fig. 30). This level supported the square plinth, 58 cm high, of which two drawings survive. They show a wall 36 cm thick to which was attached the internal curved section of stone work which must have supported the columns. The diameter of the plain columns was also 36 cm and they must have rested directly on the plinth, without an intervening base. Two delicately carved capitals are depicted, one Ionic, 17 cm high, diameter 36 cm, the other a composite capital, 48 cm high, lower diameter 36 cm, combining a bottom row of acanthus leaves with the ‘tongue’ element typical of a ‘Pfeifenfries’ below the abacus which is decorated with a rosette. It is not clear how these might have been combined on one building. Curved, double-sided combined architrave and frieze blocks rested on the capitals. On the outside the three fasciae of the architrave were divided by bead-and-reel decoration, and capped by the usual combination of astragal, ovolo and palmette mouldings. A narrow fillet separated the palmettes from the frieze which was decorated with garlands of fruit and flowers between bulls’ heads (PLATE 108). The mouldings of the inner face of the architrave were simpler: three bands of astragals above each of the fasciae followed by a row of palmettes. The internal frieze has been broken (PLATE 109). The lower face of the architrave had a decorated soffit panel with a Lesbian cymation framing a pattern of overlapping acanthus leaves. The cornice fragment (height 0.51, width 1.10, depth 0.87, letters 0.05) carried the inscription on the geison above a row of dentils and projecting consoles which are framed by ovolo moulding and alternate with downward-facing rosettes. The sima has been broken off. It was preserved in Yalvaç Museum but has now been restored to the site and reads -]i Antonini Aug. This is best restored as a reference in the genitive to Marcus Aurelius Antoninus Augustus, that is to Caracalla who became Augustus in AD 198 and died in 217 (PLATE 110). The internal ceiling of the tholos was probably made from a single circular piece of limestone decorated on the bottom with a circular band of leaf decoration, similar to that on the architrave soffit panels around flutings (Fig. 30). The building had a high conical roof made from several sections which are decorated with broad, plain overlapping leaf designs (PLATE 111). There is some resemblance between this structure and the circular temple of Fortuna, set on a high podium with four possibly Corinthian columns, which is depicted on two colonial coins issued under Gordian III (see appendix 1 no. 14).
107. Tholos foundations and propylon steps (Kelsey archive 7.1471).
108. Tholos. Architrave, exterior (Kelsey archive 7. 1380).
109. Tholos. Architrave, interior (Kelsey archive 7. 1379).
110. Tholos. Inscribed cornice.
111. Tholos. Roof panel (Kelsey archive 7. 1326).
Small circular decorative buildings had been a staple of Greek and Graeco-Ro-man architecture since the fourth century Bc.47 The example from Antioch may be compared with monuments built on a PLATE 111 Tholos Roof panel similar scale at Pergamum and Ephesus, (Kelsey archive 7. 1326^ which date respectively to the second and fourth quarters of the second century AD.48
Apart from the intrinsic interest of the building, the tholos is important as an example of epigraphically dated decorated public architecture.49 The design incorporates the conventional ornament of second– and third-century imperial architecture. The quality of execution here is of a good standard, as may be judged from the careful and even carving of the astragal and ovolo mouldings, and from Woodbridge’s elegant drawings of the two types of capital. On the other hand the execution of the leaves of the palmettes and of the stirrups and stems of the Lesbian cymation is drab. They are shown as flat and fleshy, without internal detail. In both cases the use of the drill is particularly evident.
The interpretation of the sanctuary
The question of the dedication of the temple at Antioch has been a matter for discussion since its discovery. Arundell took it to be the temple of Lunus or Mên Arcaeus. He was followed by Hamilton, likewise depending on the account of Antioch in Strabo, who suggested that it ‘may have been connected with the worship of Mên Arcaeus’.50 Weber, who made the first sketch plan of the sanctu-ary, offered no view on the matter. Ramsay, who had doubted whether the building was a temple at all in 1907, firmly identified it as an Augusteum during the excavations of 1912–14, swayed in particular by the discovery of fragments of the Monumentum Antiochenum during his final season.51 Robinson adopted acompromise position in his publication of the sculptures from the site in 1926, suggesting a joint dedication to Augustus and to Mên,52 but in his own publication of the fragments of the Monumentum Antiochenum he referred simply to a temple of Augustus.53 Ramsay’s own view meanwhile changed. In 1924 he wrote of ‘the temple of Mannes-Mên (identified with the reigning emperor in Roman times)’,54 and in 1926 of a late hellenistic temple, built after 189 BC, which was never dedicated to Augustus.55 In Ramsay’s joint publication with A. von Premerstein of the Monumentum Antiochenum the sanctuary is said to be ‘wahrscheinlich des kleinasiatischen Mondgottes Mên (Manes)’, although the propylon was ‘das zweifellos dem Augustus selbst gewidmetes Prunktor’.56 The tendency of more recent scholars has been to favour the ‘Roman’ interpretation of the cult. A. Kryzyzanowska, who produced a corpus of Antioch’s coinage, took the dedicatee to be Augustus.57 Magie and Barbara Levick suggested the possibility of a dedication to Iuppiter Optimus Maximus, who had a priest in the Flavian period at the colony,58 while F. Rakob proposed without supporting arguments that it was a Capitolium.59 H. Hänlein-Schäfer, in her full study of Augustan imperial temples, opted firmly for Augustus as the dedicatee, but left open the possibility that he shared the temple with Mên or Cybele.60 Simon Price in his comprehensive study of the imperial cult in Asia Minor left open the question both of the dedication and of the date of the temple.61
Against the recent trend can be set the lone voice of K. Tuchelt, who argued that the underground chamber was designed for an Anatolian deity with chthonic associations, and concluded that the temple was probably jointly dedicated to a twin cult of Mên and Cybele.62 The following arguments were mustered to support his theory. (1) The central chamber must have played a key role in the cult; ‘die Kammer war nicht etwa das Resultat einer Substruktion für das Podium, sondern umgekehrt: das ‘Podium’ ergab sich aus der Notwendigkeit der Schachtkammer’ (p. 507). (2) The west orientation for the temple placed it in the company of a series of Anatolian temples, notably the archaic Aeolian temples of Neandria and Larisa, and the Artemisia of Sardis, Magnesia on the Maeander and Ephesus, which were dedicated to various hellenised forms of the Anatolian mother goddess. (3) The windows or doorways in each of the gables at Antioch were appropriate to a cult in which epiphanies were important, for instance the worship of heavenly divinities, again such as Artemis at Magnesia or Artemis at Ephesus, where she had the characteristics of a moon goddess. (4) The iconography of the female figure of the middle acroterion of the west gable, wearing a high cylindrical polos for head-gear which supported a circular disc, and holding in her hands a branch in the form of a crescent moon, is appropriate for an Anatolian mother-goddess, with cosmic associations. (5) The acanthus frieze and also the garlands between the bulls’ heads of the upper frieze unusually contain poppy heads among the fruit and leaves of the decoration, and the poppy is a symbol of the goddess of fertility, most commonly identified with Demeter. (6) A bust of Demeter features prominently on the propylon of the sanctuary. Another bust, which may have come from the same structure, had features resembling those of Mên. Tuchelt’s conclusion, drawn from these arguments, is that the temple served the cult of a divinity or divinities with both chthonic and cosmic associations. Both the ‘cave’ beneath the cella and the epiphanies in the gable would be appropriate for the Anatolian mother goddess (Demeter/Cybele). Since, however, Mên was associated with Cybele in various contexts at Antioch, and a difficult passage of Strabo appears to refer to two sanctuaries of Mên at Antioch,63 there is a strong possibility that the cult of the two was linked at the temple, which may have served both divinities.
These arguments are of uneven weight and force. (1) Tuchelt laid greatest emphasis on the significance of the underground chamber, which is taken to be the architectural equivalent of a cave, often in Anatolia the location of the cult of a mother goddess.64 This proves to be a false trail. The two parallels adduced are the well known Hadrianic temple of Zeus at Aezani, and the recently excavated temple at Pessinus (p. 517). However, the supposed cave at the latter, as an extensive publication clearly shows, is a pure illusion: the temple at Pessinus had a more or less conventional 6 x 11 peripteral plan, enclosing a cella and a distyle pronaos, without any underground chamber. It was also almost certainly dedicated to the imperial cult.65 The Aezani temple indisputably has a large underground room beneath the cella, which has been interpreted as an architectural representation of the cave where Zeus was born. The cult of the mother of the gods has plausibly been located in this underground vault, while Zeus was worshipped in the cella.66 But the character of the underground chambers at Aezani and at Antioch is utterly different. The former is a vast, roomy chamber, reached by a staircase from the opisthodomos, and lighted by windows.67 This is not the case at Antioch. The vault of the chamber can be reconstructed from the architectural remains, and indeed its outline may be seen on the rock face, especially at the east end of the chamber.68 The height at the centre did not exceed 2.5 metres, and it would have been impossible to stand upright except at the centre of the room. What Tuchelt took for a bench around the sides of the room is in fact the low moulded cornice, from which the vault sprang. As he himself admitted there was no staircase from above and the only access (if any) must have been by a wooden ladder. The reconstruction of the cella walls simply leaves no place for a proper entrance. It would be fair to assume that there was atrapdoor through the cella floor and the chamber may have served some practical purpose, for instance as a treasury, but it cannot have served any cult.69
The remaining arguments are of less consequence. (2) The orientation of ancient temples had little cultic significance.70 Indeed the temple of Mên on Karakuyu itself disobeyed the supposed rule invoked by Tuchelt and faced east. Moreover, the two other main imperial temples known from central Asia Minor, the temple of Rome and Augustus at Ancyra and the temple at Pessinus, both faced west, exactly like their counterpart at Antioch. So also did the Sebasteion at Aphrodisias.71 (3) The windows in the gable may well have been designed at least symbolically to suggest the epiphany of the deity who occupied the temple, but divine epiphanies were a universal phenomenon, not a monopoly of Demeter/ Cybele. The divinised emperors, like their hellenistic royal forebears, were often characterised as epiphanes, and the temple of Rome and Augustus at Mylasa appears actually to have had a window in its gable, exactly as in the Antioch building.72 (4) The appearance of the female figure in the central acroterion of the west gable is remarkable, but she is not readily identifiable as a mother goddess (PLATE 94; Fig. 26). What Tuchelt took to be crescent shapes in her two hands are shown in Woodbridge’s drawing as the tendrils of the acanthus foliage, and the disc which is supported by her polos resembles a shield or, at least for the frontal viewer, a globe. She could be interpreted as Fortuna or even as Nike, with the shield (the clupeus virtutis) or globe both being familiar features of Augustan iconography.73 On the other hand, it is perhaps safer to interpret not simply this female figure, but her much more damaged counterpart from the east pediment, with arms turned downwards rather than raised aloft (see Taşhalan 1995, 277 pl. 15), and the female figure who appears from the foliage on the central block of the east side of the acanthus frieze (PLATE 83), as simply a Rankenfigur, a feature of the ornamental tradition on Asia Minor buildings which has been traced back to the classical period.74 (5) The lush fruits and foliage of both friezes are typical of the exuberant decoration which is to be found on Augustan public architecture, especially associated with the imperial cult and the promotion of the new Augustan world order, a golden age of fertility and prosperity.75 The poppies were a local touch. The opium poppy has always been one of the main crops of central Anatolia, and has given its name to the important market town of Afyonkarahisar (Opium-Black-Castle). It appears regularly as a decorative motif on Phrygian tombstones in the second and third centuries AD.76 Poppies were grown in the territory of Antioch, and the first western visitor to Yalvaç in modern times, Otto Friedrich von Richter, as he travelled north from lake Beysehir in 1816, noted, ‘in diesen Gegenden fängt die Opiumcultur an, wodurch diese Provinz berühmt, und dessen Hauptmarkt Afiomkarahissar ist.’77 (6) A bust of Demeter indeed stood on the propylon at Antioch, but the context in which she appears is one which represents Augustan victory and well-being, as Tuchelt himself concedes.78 The only deity with whom she is certainly matched is Poseidon, bearing a trident, alluding to Augustus’ naval victories.79 The bust of Mên, which was acquired by Konya Museum without a firm provenance fifteen years before the propylon was excavated, is only conjecturally, although not implausibly, assigned to the building. However, he would be perfectly appropriate to the context by virtue of being the patrios theos of Antioch, the old protector of the city, taken over by the Augustan colonists from his temple on Karakuyu.
The attribution of the temple at Antioch to an Anatolian divinity should be firmly rejected. The only viable alternative interpretation is that it was dedicated to a Roman cult, conceivably to Iuppiter Optimus Maximus, but much more plausibly to Augustus himself, or to Rome and the divine Augustus, who were worshipped together at the other main imperial temple in Galatia at Ancyra, as they also were at the provincial temples of Nicomedia in Bithynia and at Pergamum in Asia, and at the municipal imperial temple of Mylasa.80 Indeed Tuchelt himself in his study, which represented the first thorough attempt to make sense of the whole complex of buildings at Antioch, provided most of the more important arguments to support this view.
The argument needs to take account of the sanctuary at Antioch at three levels: as an architectural ensemble; in the identification of its individual buildings; and in terms of decorative and sculptural detail. In the first instance, the layout of the temple, the porticos, the propylon and the Tiberia platea should be considered as an architectural unit (Fig. 31). All these elements are welded together by rigorously symmetrical, axial planning which is one of the hallmarks of large-scale early imperial architecture. The arrangement of the approaching colonnaded street, the propylon and the vast rectangular square, combined with the rising flights of steps, two in the Tiberia platea itself, twelve to the foot of the propylon and twelve more to the top of the podium of the temple, sharpened the focus on the building at the eye of the design, the temple itself. Tuchelt himself emphasised that the parallels stem almost exclusively from the urbanised parts of the Roman West: the sanctuary of Fortuna at Praeneste, the Hadrianic sanctuary adjoining the theatre at Arausio in Gallia Narbonensis, and, on an even grander scale, the Forum of Trajan at Rome. Closer in date to the Antioch sanctuary than any of these is the Forum of Augustus completed by the dedication of the temple of Mars Ultor at Rome in 2 BC, with its parallel porticos and great semicircular exedrae framing on either side the high podium and huge façade of the temple itself.81 Comparable too is the planning of the central forum of a city which can be regarded almost as Antioch’s sister colony in the western empire, Colonia Iulia Augusta Nemausus (Nîmes), where the Maison Carrée is the counterpart of the Antioch temple. Recent study suggests that the designs for the decoration of the temple were directly inspired by those of the temple of Mars Ultor at Rome.82
The individual elements of the architectural mix at Antioch are also separately characteristic of specifically Roman imperial design. The triple arched propylon, completed in 2/1 BC, is only the second building of this type to be built in Asia Minor. It followed hard on the heels of the gateway erected at Ephesus in 4/3 BC for Augustus, Livia, the now-dead Agrippa and Julia by the imperial slaves Mazaeus and Mithridates.83 This combined a ground plan already familiar from large scale hellenistic architecture in Asia Minor (for instance the gates to the hellenistic agoras in Miletus and Ephesus) with a style of elevation which had developed in Roman monumental gateways.84 The Roman and imperial element is made clearer at Antioch than at Ephesus by the sculptural programme. This was fully demonstrated in Robinson’s original discussion, which elaborated on Woodbridge’s published reconstruction, and can be summarised here.85 Although points of detail are still open to question, the overall impression conveyed by the building is unambiguous. The elaborate, outward-facing elevation was dominated by huge figures in the spandrels of the arches, symbolising Rome’s victory over the barbarians and the fruits of peace which this had brought. Woodbridge’s drawing features draped winged Victories (PLATE 112); and naked Genii (PLATE 114) holding clusters of grapes supported giant garlands, containing flowers, fruits and bedecked with ribbons, which hung above each of the two side arches.86 On either side of the central arch were naked or half-draped figures of two barbarians, one kneeling (PLATE 113), the other apparently attempting to flee, each with his hands bound behind his back. They are generally taken to be Pisidian victims of the Augustan wars in southern Anatolia, or perhaps the Homonadensians, the Isaurian tribe which had barely been subdued by the time when the arch was dedicated.87 Robinson rightly points out that they are carved in a tradition of barbarian representation which goes back to the sculptures of the defeated Gauls at Pergamum, and it is probably better to treat them not as representatives of a specific tribe but as generalised barbarian enemies of Roman order.88 Above the spandrels ran an architrave, carrying the inscription with its dedication of 2/1 BC to Augustus, and then an elaborate frieze. At the four points where this projected above the three-quarter columns of the façade, there appear to have been portraits of divinities: Ceres/ Demeter (Cybele) with flowers and grapes bound into her hair and a vase of fruit on the side of the block (The Art Bulletin 9 (1926) Fig. 40), Poseidon with his trident, a Dolphin at the side (ibid., Figs. 38–9), and, if Tuchelt’s identification of the block in Konya Museum is correct, a youthful Mên wearing a Phrygian cap decked with laurel, depicted as a warrior with a sword belt across the right shoulder and breast.89 The side of the block contains a composition, typical for hellenistic Pisidia, of a circular shield with lance and greaves (Festschrift Bittel pl. 106, 1–3). The fourth sculpture from this position has not been recovered.
112. Propylon. Victory figure with garland (on site).
113. Propylon. Figure of captive (Afyon Museum).
114. Propylon. Genius with garland (Yalvaç Museum).
The remainder of the frieze contained at least two and perhaps three groups of two Tritons, each placed heraldically on either side of an armoured trophy, and perhaps located above the centre of the archways. The monsters, which surely allude to Augustus’ naval victories, have human heads and torsos, with long and bulky spiral tails, decorated with scales resembling acanthus leaves. Each of the two preserved suits of armour is of a different type (The Art Bulletin 9 (1926) Figs. 43, 44). Most of the rest of the space was taken up with emblems of warfare. The find spots suggest that four fragments illustrating sections of warships (ibid., Figs. 46–8; Taşhalan 1995, 282 pl. 25. 3) and in one case the sidus Iulium (The Art Bulletin 9 (1926) Fig. 45), originally came from the area above the south arch, while a group depicting an elaborately ornamented shield (ibid., Fig. 51; Taşhalan 1995, 282 pl. 25. 1), a quiver and baldrics (The Art Bulletin 9 (1926) Fig. 52), a cuirass with a Medusa head (ibid., Fig. 53), a spear and sword hilt (ibid., Fig. 49) and another suit of armour beside a fragmentary shield with two spears (ibid., Fig. 50) belonged at the north end. As a whole the composition presented in abstracto an image of Augustus’ naval and land victories. As such it recalls the frieze of weapons combined with sacrificial and religious equipment, probably from the porticus Octaviae at Rome,90 just as the remainder of the decoration readily finds Roman parallels. The link to Augustus is clinched by another block, excavated by Ramsay in 1914 but first identified in Konya Mu-seum by Robinson, which depicts the emperor’s birth sign, the Capricorn (PLATE 115; The Art Bulletin 9 (1926) Fig. 54).
The sculpture of the propylon was completed by several over-life-size statues which were placed on the roof of the building. One of the central positions was taken by a draped female figure, taken by Robinson to be Livia, the mother of Tiberius (The Art Bulletin 9 (1926) Fig. 55). From the south end came the base and lower part of a second female statue (ibid., Fig. 60). At the north side were two male figures, one clothed but with the right breast bare (ibid., Fig. 56–7), the other, also draped, with traces of a captive kneeling before him (ibid., Fig. 58). The lower part of another draped male figure was also recovered from the north end of the stairs in front of the building (ibid., Fig. 59). Unfortunately all the heads and faces of these statues have been lost, and their identification is a matter of conjecture. However, it is extremely likely that they represent leading members of the imperial family at the time of the erection of the gateway.91
115 Propylon. Capricorn (Yalvaç Museum).
Two other sculptural finds should be mentioned in association with the propylon, a broken but very attractive Victory, which was reused in one of the walls of the late Roman shops on the north side of the Tiberia platea,92 and the monumental head of Augustus found close to the intersection of the Tiberia platea with the cardo maximus, which could have belonged to one of the statues on the roof.93
The decoration of the propylon confirmed the message inherent in its dedicatory inscription, that it was erected to honour Augustus and members of his dynasty, to commemorate the victories he had achieved and the peace which these had brought to the Roman world. Since this was also the central message of Augustus’ propagandistic summary of his own achievements, compiled in 2 BC at the time when the arch was built but brought up to date and published at the time of his death in AD 14, it is wholly appropriate that the side walls of the central arch were chosen to carry a Latin version of the Res Gestae, as a pendant to the honours rendered by the propylon’s design and sculptural programme.
The same message was transmitted by the buildings inside the sanctuary itself. The earliest examples of curved porticos, usually designed to provide a monumental background for a central feature such as a fountain or a temple, have been identified at Sybaris in south Italy and at Vernegues-Alvernicum in Gaul, whose tetrastyle prostyle temple in front of a full semicircular rear wall (although without a portico) provides a ground plan strikingly reminiscent of Antioch.94 These are dated to the first century BC, but the design became widely popular under the Empire, because the optical effect of the columns arranged in asemicircle enhanced the impressive perspective within a relatively enclosed area. The fashion for curved monuments was also developed in the preference for semicircular exedrae which is so marked in imperial architecture.95 A detail of the poorly preserved decoration of the portico also hints at an imperial touch. The two storeys were supported below by Doric and above by Ionic columns and capitals, two of the three classic architectural orders. The set was completed by the richer and more elaborate Corinthian columns and capitals of the temple itself. The combination of all three styles within a single architectural design (although not here in a single building) is paralleled by the exterior decoration of the Colosseum at Rome, which has Doric at the first, Ionic at the second and Corinthian at the third and highest level, an arrangement which has been held to symbolise the fusion of Greek architectural fashions under the uniting influence of Roman power.96 Precisely this arrangement was also favoured for the three storeys of the porticos of the Sebasteion at Aphrodisias.97
The temple itself remained the building which dominated the entire complex. Its position, close to the summit of the city hill of Antioch, commands a vast view over the territory of the colony to the west. Conversely its lofty bulk, only half hidden from the west by the propylon, would have been visible viewed from virtually any angle within and beyond the walls of the city.98 Despite Tuchelt’s hesitations, his own analysis of the architecture, like Woodbridge’s and our own, shows it to have been an essentially orthodox Roman or Italic podium structure. The temple, open at the front but enclosed at the rear, stood on a two metre high platform and was approached by a broad flight of steps.99 Placing a temple at the head of an imposing stairway was a design feature already familiar in the hellenistic age, and best illustrated by the second century BC temple of Dionysus on the theatre terrace at Pergamum,100 but the arrangement with broad parastades, as at Antioch, was a western import which occurs frequently, and always in contexts with strong Roman associations, in the cities of imperial Asia Minor: the imperial temple at Mylasa,101 the temple adjoining the theatre at Side,102 and the twin temples at Ephesus dedicated respectively to the cults of the divine Julius Caesar and to the city of Rome.103
Even more emphatically than the basic ground plan, the elaborate decoration of the Antioch building stamps it as an imperial temple. Augustus and wealthy contemporaries closely associated with him had introduced new styles and standards of temple construction to Rome itself, above all to buildings such as the temples of Apollo Sosianus and Apollo on the Palatine, and the temple of Mars Ultor in the Augustan Forum. The style has been characterised most effectively by Paul Zanker: ‘The style of the new temples in Augustan Rome was intentionally a mixtum compositum. Podium, porch and tall heavy pediment [emphasized by the unique double frieze at Antioch] all belong to the Italic religious tradition, while the height of the columns, the capitals, and the organization of the façade are derived from hellenistic models. The lavish materials, however, including even the gilding of individual elements, outdid anything the past had to offer.’104 The designs, Zanker goes on to point out, were taken up with appropriate modifications, in the provinces: ‘No matter what differences there may be in detail, the overall impression is always the same, always bearing the stamp of the aurea templa built by the princeps in Rome. These temples are always set on apodium, with grand staircase, tall Corinthian columns, richly ornamented entablature, and elaborate roof cornice’ (p. 312). These words could have been written to describe the Antioch temple. In fact, Zanker chooses to illustrate his argument with a photograph of a building closely similar to it, the temple of Roma and Augustus at Pola, dedicated, exactly like the Antioch propylon, to Augustus as pater patriae, between 2 BC and AD 14 (p. 312 Fig. 243: ‘the rich architectural ornament is typical of the buildings of the imperial cult in East and West’).
Even in antiquity the only completely clear indication of the temple’s owner may have been the cult statue or statues which were housed in the cella. The 1924 excavation unearthed a broken fragment of a giant, sandalled foot immediately to the north of the temple foundations. It may have belonged to the cult statue of Augustus. The architectural decoration of the temple was, clearly by design, less explicit than that of the propylon. If we discount the doubtful exception of the female figure carrying a shield on the west acroterion, it contained no direct allusions to Augustus’ victories or achievements and there was no dedicatory inscription. Instead the décor combined the familiar motifs of fruitful growth and animal sacrifice in the two richly detailed and densely packed friezes. The intention was not to convey a specific symbolic message but rather to invoke an era of abundance and security, guaranteed by the Augustan peace.
The dating of the temple, despite the occasional voicing of doubts, has been securely placed in the Augustan–Tiberian period by analyses of the architectural ornament.105 Although Robinson never completed a full study of the architectural decoration and adduced no detailed parallels to support his first impressions, he was in no doubt that the temple was Augustan and that the decoration was typical of this period. He observed in passing that the temple of Rome and Augustus at Ancyra was probably later, and that the temple of Mars Ultor in Rome ‘is similar, although somewhat better’.106 In his 1937 review, which established an Augustan or Tiberian date for the temple of Rome and Augustus at Ancyra, E. Weigand made detailed observations on the relationship between the Lesbian cymatia and the ovolo and astragal mouldings of the Antioch temple, which he took to confirm the Augustan date.107 These views have been broadly supported with more detailed argumentation by W. Heilmeyer in his study of the decoration of Corinthian capitals and related architectural elements,108 and by Alzinger in his work on Augustan architecture from Ephesus.109 Tuchelt, while accepting that the temple belongs to the late Augustan or Tiberian period, emphasised that much of the ornament appears unfinished, that several hands, some rather unskilled, were clearly at work on the sculpture, and that ornamental details did not receive similar treatment wherever they occurred on the building. Hence an exact date is hard to fix.110 Waelkens in his discussion of the date of the imperial temple at Pessinus, examined some of the parallels with the Antioch carving. The use of pilasters formed from two courses of blocks in the cella wall is taken as evidence for a date before AD 50. Moreover they demonstrate Augustan features: both rows of acanthus leaves spring from the bottom of the block and the volutes of the helices remain independent of the caulis bush below them. In contrast the same features of the one surviving capital from the prostylos were handled rather differently, in that the base of the upper row of acanthus leaves did not reach the bottom of the capital and the volutes of the helices touch the elements below them. This might imply that the column capitals were completed later than the pilaster capitals (as would be normal) but it may simply illustrate the latitude for stylistic variation on the same building at the same period.111 The palmettes on the architrave and sima at Antioch are compared to middle Augustan examples from Ephesus, and the Lesbian cymation, which is particularly strikingly carved on the door-frame fragment (The Art Bulletin 9 (1926), Fig. 24), combines the very light, delicate design of late hellenistic work with a central stem for the intermediate flower which is typical of the Augustan period.112 Weigand had noted of the same decoration that the use of the drill to make small holes or eyelets in the leaves of the plants marked the decisive step from the hellenistic to the imperial form of the Lesbian cymation.113 More important than any of the individual details is the overall effect achieved by the lush but highly detailed ornamentation, which is reminiscent not only of early imperial architectural ornament elsewhere in western and central Asia Minor, but of the decoration of Augustan architecture at Rome and in the western provinces.114 This is particularly evident in the acanthus frieze, whose heavy and complicated shapes dominate the frieze zone in a way that becomes characteristic of this decoration in the Augustan period. It differs from the lighter, more rhythmic and naturalistic representation of the vegetation on earlier structures.115 The upper frieze of garlands and bulls’ heads has also been placed within the development of this decorative scheme. The design is in general found only on altars or on religious buildings, where the an to the ritual of sacrifice was particularly appropriate. Robinson and other commentators have remarked on the originality and individuality of the carving, in which all the bulls’ heads are clearly differentiated from one another. They may be compared with the garland frieze of the portico of Tiberius at Aphrodisias.116
The generally unanimous acceptance of an Augustan or at latest early Tiberian date should not disguise the fact that it is hazardous to attempt a more precise chronology on stylistic grounds alone. Stylistic dating without reference to chronological fixed points provided by epigraphic evidence is an inexact science. The firmest basis for dating the sanctuary at Antioch remains the inscription of 2/1BC on the propylon, and this recommends an Augustan date for the temple with more authority than any detailed analysis of the ornament. Although both formed part of a single architectural design, the question whether the temple was completed before or after the gateway remains open. If we combine the consensus of the stylistic analyses with the implication of the epigraphic date, it is difficult to believe that the building postdates Augustus’ death in AD 14.117 The temple, in other words, was erected for the cult of Augustus during his lifetime. That would not be an unnatural honour either in the whole context of the development of the imperial cult in Asia Minor, or in the precise circumstances of Antioch, whose inhabitants looked upon Augustus as the founder of their colony.118 On balance it would make ergonomic sense if the temple had been completed before the gateway, for that would have allowed easier access for building gangs and heavy masonry to the construction site. But the argument is no more than suggestive, for the site of the temple could easily have been reached from the sides, through gaps deliberately left in the porticos, and it is tempting to think that the temple is later than that of Mars Ultor in Rome, which seems to have provided at least some of the inspiration for the overall conception of the design.
Notes
1 A significant exception was W.M. Ramsay, The Cities of Saint Paul (1908) 250, who briefly had the idea that the semicircular feature was ‘no more than an odeum or small theatre’. L. de Laborde, one of the first group to identify the site in 1827, thought that the ruins belonged to a grave monument (see above p. 23). This may have been a false inference from local information. O. von Richter, who had seen the ruins a decade earlier but failed to identify them as the site of Antioch, had taken them for a Turkish burial ground.
2 Paraphrased from Callander’s report which is preserved in the Michigan archive; see appendix 2 (l).
3 D.M. Robinson in his diary took them to be Greek, and this may have led him to over-hasty conclusions about the nationality of the masons who worked on the temple. See p. 172 n. 106.
4 M. Waelkens, Sagalassos I (1993) 45 with pl. 36.
5 M. Waelkens, Epigraphica Anatolica 7 (1984) 39–40, 42 with refs.
6 See W. Alzinger, Augusteische Architektur in Ephesos (1974) I, 55; E. Fossel, JÖAI 50 (1972–5) 212–14; M. Waelkens, ‘The adoption of Roman building techniques in the architecture of Asia Minor’, in F.H. Thompson and S. Macready, Roman Architecture in the Greek World (1987) 95–6.
7 C. Ratté et al., AJA 90 (1986) 55–7.
8 Taşhalan 1995, 271 pl. 8. 2 illustrates a flat stone with a moulding at the top which could have been one of these paving blocks.
9 K. Tuchelt, Festschrift Bittel, 514–22; see pp. 157–9.
10 R. Amy and ?. Gros, La Maison Carrée de Nimes, Gallia, Suppl. XXXVIII (1979) I, 21 Fig. 3. See p. 170 n. 69.
11 Taşhalan 1995, 271 pl. 8. 3.
12 The shape of the elongated beads resembles the mouldings on the temple cornice at Pessinus, Epigraphica Anatolica 7 (1984) Taf. 12, on the Mazaeus gate at Ephesus, Alzinger, Augusteische Architektur, Taf. 135, and on the gate to the lower agora at Sagalassus. L. Vandeput, The Architectural Decoration in Roman Asia Minor. Sagalassos: a Case Study (1997) 59–60.
13 ‘Bemerkungen zum Tempelbezirk von Antiochia ad Pisidiam’, Beiträge zur Altertumskunde Kleinasiens. Festschrift für Kurt Bittel (Mainz, 1983) 501–22 at 507–8.
14 See the drawing in D.M. Robinson, The Art Bulletin 9 (1926) 5 ff. Fig. 3.
15 Taşhalan, 1995, 245–84 at 249 with figs. 4–6.
16 D.E. Strong and J.B. Ward-Perkins, PBSR 30 (1962) 12; although caution is advised by W. Heilmeyer, Korinthische Normalkapitelle, 27 with n. 108. See Waelkens, Epigraphica Anatolica 7 (1984) 48–9.
17 Compare first century AD columns made of limestone from Palmyra, and contrast the preference for monolithic columns of marble at a later date.
18 L. Vandeput, The Architectural Decoration in Roman Asia Minor (1993) PLATE 69.2–71.4 publishes photographs of the following items from this catalogue: 50, 57, 69, 71, 84, 118, 127, 135.
19 It must be said that these dimensions do not exactly coincide with those of other publications. Taşhalan 1995, 255 plan 2 gives maximum foundation dimensions of 27.95 x 14.20, a cella 12.50 x 10.65, a pronaos of 8.80, and the internal basement 7.75 x 5.60. Tuchelt, Festschrift Bittel, 504–6 gives the width of the podium foundations as 13.60, the width of the west wall of the cella as 9.70, and the dimensions of the internal basement as 7.90 x 5.65. Direct comparison between these measurements is impossible, as it is not clear what measuring points were used. Unfortunately there is no record of a detailed plan of the foundations made by F.W. Woodbridge during the 1924 excavations, although his careful reconstruction of the elevation (The Art Bulletin 9 (1926) Fig. 3) implies that one must have been made.
20 We noted one broken block that might conceivably have come from this cladding. It measures 76 x 89 x 46.5 cm deep, and was finished with a claw chisel on the front and the sides.
21 A.T. Hodge, The Woodwork of Greek Roofs (1960); R. Martin, Manuel d’architecture grecque I (1965) 2–46, esp. 16–18; R. Meiggs, Trees and Timber in the Ancient Mediterranean World (1982) 218–27.
22 Tuchelt, Festschrift Bittel, 507 arrived at similar figures: twelve steps with risers of 25 and treads of 35 cm, covering 3.85 m from west to east. These figures are taken over by Taşhalan 1995, 248, with 255 Plan 2.
23 See also M. Wilson Jones, ‘Designing the Roman Corinthian Order’,JRA 2 (1989) 35–69.
24 G. Weber, Jdl 19 (1904) 96–101 Taf. 9.
25 Letter in Michigan archive, see appendix 2, II.B.3 (73.12).
26 One group between Z1 and Z3 on the plan were 1.80–1.90 above ground, measuring 14–17 cm high, 15–17 cm wide and are spaced 40–50 cm apart. Another between R and Q3 are 1.75–1.90 above the ground, 12–14 cm high, 10–12 cm wide and 6–7 cm deep. The largest 24 x 23 cm; another is 13 x 27 x 11 cm deep; their depths vary from 8 to 14 cm.
27 J.J. Coulton, The Architectural Development of the Greek Stoa, 114–17.
28 Illustrated by Taşhalan 1995, 278 pl. 17.
29 Taşhalan 1995, 250, 263 drawing 7, 278 pl. 18. Woodbridge’s handsome reconstructed elevation, The Art Bulletin 9 (1926) Fig. 3, shows only a plain frieze in the upper order.
30 Taşhalan 1995, 251, 278 pl. 16. This stairway has not been shown in the reconstructed plan Fig. 32.
31 D.M. Robinson, Hesperia 22 (1953) 172–4.
32 See J.J. Coulton, The Architectural Development of the Greek Stoa, 136–7.
33 As are indicated in Woodbridge’s unpublished plan of the temple area.
34 The Art Bulletin 9 (1926) 21–41 with Figs. 31–54.
35 Taşhalan 1995, 251–3, 256 plan 3. He also illustrates several pieces of the architectural order and decorative sculpture, PLATEs 22–8. For the positioning of the Res Gestae see the discussion of Ramsay and von Premerstein, Das Monumentum Antiochenum, 13–16, who suggest as the most likely arrangement, that the inscription was carved in ten columns on the internal walls of the central arch of the propylon.
36 The first of these is published by T. Drew-Bear, XII Aras., 17 pl. 1.
37 In early reports we published this late dating of the propylon: S. Mitchell, Anchor Bible Dictionary I (1992) 264; Anatolia. Land, Men, and Gods in Asia Minor I (1993) 107, cf. II, 7; The Dictionary of Art 2, 159–60.
38 XII Aras., 14; Taşhalan 1995, 251–2.
39 Augustus had been imp. XIV since 8 BC (cf. Cassius Dio LV.6.4; Reynolds and Ward-Perkins, The Inscriptions of Roman Tripolitania no. 319). T.D. Barnes, JRS 63 (1973) 23 argued that Augustus received his fifteenth imperatorial salutation on the occasion of military successes won by Gaius Caesar in Arabia during his consulate of AD 1, and R. Syme, Roman Papers III, 1198–1219 at 1205 endorses the suggestion. The chronologi-cal table compiled by R. Cagnat, Cours ďépigraphie latine, 179 anticipates this conclusion.
40 For the temple at Pola, see below p. 165.
41 It was published virtually simultaneously by W.M. Ramsay, JRS 14 (1924) 179 no. 6 and by Robinson, TAPA 55 (1924) 5–20, each clearly believing he had the publication rights.
42 AJA 1924, 440. Robinson’s excavation diary indicates more precisely that the wall was about eight metres from the foot of the steps to the propylon. The block, now in Afyon Museum, is 90.5 cm high, 1.26 broad and 45.5 cm deep. On the front there are three panels, 51.5, 23 and 51.5 cm wide, and the middle one, which carries the text of the edict, is slightly recessed.
43 Coins from Ramsay’s earlier excavation at Antioch were published by G.F. Hill, Num. Chron. 14 (1914) 299. A memorandum in the Kelsey archive, which apparently related to these earlier excavations, reads, ‘Roman coins from the first to third centuries AD were rare. The most interesting were some four or more specimens of "second brass" of the provincial Asiatic series of Augustus. An As of Vespasian and another of Domitian were also noted. From the period of Diocletian and Maximian coins became more frequent and were plentiful for the later fourth century; they were mainly from Eastern mints. Byzantine coins were common: the great majority were of the sixth and seventh centuries but the series extended as late as the tenth. The mints chiefly represented were Constan-tinople, Nicomedia and Cyzicus.’
44 For bibliography on these see MAMA X 330 note and C. Roueché, Aphrodisias in Late Antiquity (1989) 111–12, and index s.v. gameboards.
45 See W.M. Ramsay, JRS 6 (1916) 108–28 esp. 109 ; D.M. Robinson, AJPhil 47 (1926) 1–54, esp. 3; W.M. Ramsay and A. von Premerstein, Das Monumentum Antiochenum, Klio Beiheft 1927, esp. 1–5.
46 Taşhalan 1994, 268.
47 F. Seiler, Die Griechische Tholos (Mainz, 1986).
48 W. Koenigs and W. Radt, ‘Ein kaiserzeitlicher Rundbau (Monopteros) in Pergamon’, Ist. Mitt. 19 (1979) 317–54; J. Keil, JÖAI 26 (1930) 45 ff. (Ephesus).
49 See the list compiled by Vandeput, Architectural Decoration, 33–5. Note that the date of AD 50 or 62 attributed to the Antioch propylon in that list should be changed to 2/1 BC, in accordance with the discussion above.
50 Arundell, Discoveries I, 275; Hamilton, Researches I, 474.
51 JRS 6 (1916) 107–8.
52 The Art Bulletin 9 (1926) 18; compare also the report of A.M. Woodward, JHS 45 (1925) 213, derived from information provided by Robinson. Followed by J.B. Ward-Perkins, Roman Imperial Architecture, 280.
53 AJPhil. 47 (1926) 25.
54 JRS 14 (1924) 201.
55 JRS 16 (1926) 111; cf. JHS 50 (1930) 277 (date in second century AD).
56 Ramsay and von Premerstein, Das Monumentum Antiochenum, 1, 3.
57 Mélanges Michalowski (Warsaw 1966) 520 (non vidimus); Monnaies coloniales d’Antioche de Pisidie (1970).
58 D. Magie, Roman Rule in Asia Minor (1950) I, 460, II, 1320; B. Levick, Roman Colonies in Southern Asia Minor (1967), and RE Suppl. XI, 52.
59 Röm. Mitt. 81 (1974) 83.
60 Veneratio Augusti. Eine Studie zu den Tempeln des ersten römischen Kaisers (1981) 191–6.
61 Rituals and Power. The Roman Imperial Cult in Asia Minor (1984) 269–70.
62 Festschrift Bittel, 501–22.
63 Strabo XII.3.31, 557; cf. XII.8.14, 577. On the problems of interpretation see now R. Syme, Anatolica. Studies in Strabo (1995, but written in the 1940s) 344–7.
64 Festschrift Bittel, 517–18; cf. MAMA IX, xxxv n. 9.
65 M. Waelkens, ‘The Imperial Sanctuary at Pessinus: archaeological, epigraphical and numismatic evidence for its date and identification’, EA 7 (1986) 37–73; cf. S. Mitchell, Chiron 16 (1986), 31–3.
66 L. Robert, BCH 105 (1981) 331–60; MAMA IX, xxxiii–xxxv. These conclusions, in any case, have been disputed by the current excavator of Aezani, K. Rheidt, Ant. Welt 28. 6 (1997) 493–4, although his arguments are not decisive.
67 See the detailed publication by R. Naumann, Der Zeustempel zu Aezanoi (1979); compare the photograph in S. Mitchell, Anatolia, II 17 Fig. 5.
68 Pace Tuchelt, Festschrift Bittel, 505 (‘Spuren, die auf eine Einwölbung hinwiesen, wurden nicht beobachtet’).
69 Note too the vault under the cella of the temple of Rome and Augustus at Caesarea Maritima, K.G. Holum, King Herod’s Dream. Caesarea on the Sea (1986) 88 ff.
70 See S.R.F. Price, Rituals and Power, 167–8 citing H. Nissen, Orientation. Studien zur Geschichte der Religion (1906–10) esp. 243 ff.
71 Tuchelt tried unsuccessfully to show that the Ancyra temple was also dedicated to an Anatolian deity (Arch. Anz. 1985, 317–9), but see Mitchell, Chiron 16 (1986) 28–9; cf. also H. Halfmann, Chiron 16 (1986) 41–2.
72 For Mylasa, see the eighteenth century illustration of Pococke, reproduced in Price, Rituals and Power, 167 Fig. 10. In general, Hänlein-Schäfer, Veneratio Augusti, 72–3; also figs. 34–7 for the Mylasa temple.
73 Robinson, The Art Bulletin 9 (1926) 17–18, observes, ‘The round disc seems to be adistinctively Roman addition. It may represent the round clypeus, or shield, which was consecrated to Augustus by the Roman senate and people, and which adorned his house on the Palatine. To the oriental the disc would symbolise the sun as the crescent would the moon.’ Neither interpretation supports the theory that the temple was associated with Cybele or with Mên. The globe was, in fact, an ubiquitous feature of Augustan victory imagery.
74 Hänlein-Schäfer, Veneratio Augusti, 93, 193. For Rankenfiguren, see K. Schauenberg, Rh. Mus. 64 (1957) 217 ff. Robinson, The Art Bulletin 9 (1926) 11 n. 9 points out that this was a particular feature of buildings at Antioch’s mother city, Magnesia, which may have still been a source of influence on styles at Antioch; see C. Humann, Magnesia am Mäander, 72, 75, 77, 147. For the acroteria figures see Robinson, The Art Bulletin 9 (1926) 17–18 with figs. 29–30. He regards the circular object as deliberately ambiguous, both shield and solar disc.
75 P. Zanker, The Power of Images in the Age of Augustus (1988) esp. 172–92.
76 See M. Waelkens, Die Kleinasiatischen Türsteine (1986) 320, index s.v. Mohnkopf, in Ranken, and the indexes to MAMA IX and X, citing numerous examples.
77 Wallfahrten im Morgenlande (1822) 356.
78 Festschrift Bittel, 513–4, following Robinson’s full description and analysis. See further below, p. 162.
79 Robinson, The Art Bulletin 9 (1926) figs. 39 and 40.
80 Price, Rituals and Power, 249 ff. nos. 19, 70, 100, 108.
81 P. Zanker, Forum Augustum (1968).
82 P. Gros, Princeton Encyclopedia of Classical Sites, 616–7 with bibliography.
83 E. Wilberg, Forschungen in Ephesos 3 (1923) 40 ff.; W. Alzinger, Augusteische Architektur in Ephesos, 9 ff.
84 Cf. W. Alzinger, ANRW II.7.2 (1981) 816.
85 The Art Bulletin 9 (1926) 21–41 with figs. 31–54.
86 Woodbridge shows a Victory standing opposite a Genius on the west-facing spandrels of the north arch, and two Genii in the equivalent position above the south arch. Robinson rightly states that this arrangement is only exempli gratia and deduces from the find spots of two of the Genii (illustrated in his figs. 35 and 36) that they were placed above the north arch (outer face). Since one of the victory figures was found at the top of the stairs behind the propylon, it is most likely that the Genii were on the west and the Victories on the inner, east face of the propylon.
87 For the Homonadensians, see Levick, Roman Colonies, 203 ff.; Syme, Anatolica, 257–69, who both tend to date hostilities to a period shortly after 6 BC.
88 The Art Bulletin 9 (1926) 25–6; compare the appearance of defeated barbarians, also derived from the Pergamene prototypes, on a late republican grave monument for a Roman general which was erected on the via Appia, W. von Sydow, ‘Die Grabexedra eines römischen Feldherren’, JdI 89 (1974) 187–216. Compare especially his 203 Fig. 12 with Robinson figs. 41–2. Von Sydow argues that these reliefs may also depict Asiatic enemies of Rome, despite their ‘Gallic’ appearance.
89 The three blocks are illustrated by Taşhalan 1995, 283 pl. 26, 1–3.
90 Zanker, Power of Images, 125–7 with Fig. 102.
91 The whole sculptural programme should obviously be compared with the slightly later sculptural decoration of the Sebasteion at Aphrodisias: R.R.R. Smith, JRS 77 (1987) 88–138 and 78 (1988) 50–77.
92 D.M. Robinson, ‘Eine Nike aus Antiochia in Pisidien’, Antike Plastik. W. Amelung zum 65. Geburtstag (Leipzig 1928) 200–5.
93 D.M. Robinson, AJA 30 (1926) 124–36.
94 Illustrated by Tuchelt, Festschrift Bittel, 510 Fig. 4 nos. 2 and 3; see also F. Rakob, Röm. Mitt. 81 (1974) 81–6 esp. 82 n. 56., referring to P.H. v. Blanckenhagen, Rh. Mus. 70 (1963) 100 ff.; and Hänlein-Schäfer, Veneratio Augusti, 47.
95 S. Settis, ‘Esedra e ninfeo nella terminologica architettonica del Mondo Romano’, ANRW I. 4 661–745.
96 J.J. Onians, Bearers of Meaning. The Classical Orders in Antiquity, the Middle Ages and the Renaissance (1988) 41–8.
97 R.R.R. Smith, JRS 77 (1987) 88–138. Full publication of the architecture of the Aphrodisias Sebasteion is awaited.
98 Hänlein-Schäfer, Veneratio Augusti, 34 holds that the temple would have been relatively invisible, inside the sanctuary, but the terracing of the hill to the west ensured its prominence.
99 See Tuchelt’s axonometric drawing, Festschrift Bittel, 506 Fig. 3, in which the only significant error is to omit the floor of the cella and the steps leading to it from the pronaos.
100 W. Radt, Pergamon, 218–22.
101 H. Hänlein-Schäfer, Veneratio Augusti. Eine Studie zu den Tempeln des ersten römischen Kaisers (1981) pls. 34–37.
102 A.M. Mansel, Die Ruinen von Side, 90–2.
103 Price, Rituals and Power, 254 no. 27 for literature.
104 The Power of Images in the Age of Augustus, 256.
105 See Vandeput, Architectural Decoration, 227 index s.v. Antiochia pros Pisidian: Temple of Augustus, with references to very numerous sections of her book which allude to the individual elements of the decoration of the temple. A late Augustan-early Tiberian date is consistently favoured.
106 The Art Bulletin 9 (1926) 11 with n. 8, where he contests the view of Ramsay père, that the temple was late hellenistic, and fille, that it belonged to the second century AD. He suggested construction between 21 BC and AD 14, ‘probably by Ionian workman’. The final observation may stem from his belief that the temple blocks carried Greek masons’ marks. In fact, the ones he records are clearly in Latin letters, which may well be significant.
107 Gnomon 13 (1937) 420–1.
108 Korinthische Normellkapitelle (1970) 46 (acanthus frieze), 81 ff., 84 (capitals).
109 Augusteische Architektur in Ephesos, 103 (ovolo and Lesbian cymation), 125–6 (comparing the Lesbian cymatia at Antioch with those on the middle Augustan Memmius monument at Ephesus); 130 (acroterion, held to be Augustan).
110 Festschrift Bittel, 508–9.
111 EA 7 (1986) 48–9.
112 EA 7 (1986) 50–1.
113 Gnomon 13 (1937) 422.
114 Zanker, Power of Images, figs. 96, 238–40 (garlands on Roman altars); Fig. 243 b) (Rankenfries and capitals of the temple at Pola) Fig. 252 (an elaborate door-frame moulding at Pompeii).
115 See W. von Sydow, JDAI 89 (1974) 206–10 for a detailed analysis of this development between the late republican and Augustan period.
116 K. Tuchelt, Ist. Mitt. 25 (1975) 122–5 on a roughly contemporary garland frieze on the imperial altar placed outside the bouleuterion at Miletus, which belongs to a different tradition of carving. For Aphrodisias, see M. Waelkens, ‘Notes de l’architecture sur l’agora et le portique de Tibère à Aphrodisias de Carie’, in J. de la Genière and K. Erim (eds.) Aphrodisias de Carie Colloque de l’Université de Lille III (1987) 123–34.
117 Against the assertions of Mitchell, Anatolia I, 104 and figs. 15–16 that it was Tiberian.
118 Compare the cults of Rome and Augustus initiated by Roman emigrants among the Milyadeis, on the western boundaries of Pisidia, and of Rome archegetis together with members of the Augustan dynasty, by colonists at Attaleia, both apparently during Augustus’ lifetime; Mitchell, Anatolia I, 102–3.