Late Antique Walls: Aesthetics and Defence


The walls and gates imitated in mediaeval times are the obvious ones, the standard for embellishment in the West being Rome and, in the East, Constantinople. Here the Porta Aurea was beautified by statues and reliefs[1] (the statue of Theodosius the Great only fell in 739/40AD in an earthquake[2]) and formed an important element in Imperial ceremonial[3]. The flanking wall also appears to have been decorated with marble and granite column shafts, as Lafitte Clave observes[4] in his military reconnaissance of 1784: toute cette partie de l'enceinte a été autrefois réparée avec des colonnes et des pierres d'entablement d'anciens édifices grecs et romains que les Turcs n'ont pas respectés: on les voit sur le parement de la muraille et on peut lire sur quelques unes des restes des inscriptions grecques.  In Constantinople, up to the Sack of 1204, pagan statuary of the city was still considered to be an essential element of the city’s iconography, culturally and aesthetic ally.[5] When the Bulgarian Krousmmos, the new Senaccherab, sacked Constantinople in 812/813AD after having admired the walls of the city, he carted away bronzes and marbles. [6] To the disappointment and puzzlement of later travellers (as we have seen), almost nothing was left beyond the Hercules reliefs by the time Mehmet took the city.[7] Krousmmos was following a tradition much older than Julius Caesar at Corinth or Verres in Sicily, and one which the Turks also followed.[8]


A description of the Golden Gate is appropriate here, not only because it incorporated spolia, but precisely because its richness set a standard for others to follow. The Golden Gate was therefore lavish in materials and decoration, impressed medieval travellers greatly when it was perhaps but little decayed, and was coveted for its reliefs even when dilapidated (although, of course, there was little else in the City of similar date). We shall see the Golden Gate imitated in other cities in  Turkey, such as Nicaea and Thessaloniki with, in consequence, a lavish use of spolia.


The state entrance into the capital, the Golden Gate was architecturally splendid, and was built of large square blocks of polished marble fitted together without cement, and flanked by two great towers of the same material[9]. Van Millingen listed[10] amonst the other decoration a cross, a Victory, a crowned female figure (the city’s Tyche), a statue of Theodosius the Great, a bronze group of four elephants; and the gates of Monpseuesta, gilded and placed here by Nicephorus Phocas, as a trophy of his campaign in Cilicia; a Roman eagle at the SW angle of the N Tower; a laureated XP over the central archway on the city side; and twelve bas-reliefrs of classical subjects. It would be interesting to know how large a proportion of these adornments were spolia. Sir Thomas Roe, British Ambassador from 1621-8 (and with Earl of Arundel behind him), tried to get them, but after negociations the locals reckoned that the City would collapse if the reliefs were removed so, says Roe, they are like to stand, till they fall by tyme. And so they did: Spon admires various bas-reliefs in 1675, Sestini in 1778 notes that some reliefs are still in place; but Dallaway in 1795 states that those of Hercules have been brought down by an earthquake, and the Punishment of Prometheus mutilated by the Turks, who have walled up the gate. And the most exquisite fragments of finely sculptured columns, friezes, etc still littered the area in the early 19th century.[11] Other parts of the walls were also decorated, such as  near Bab Kapoussi along the Golden Horn, where one arched entrance had a bas-relief on either side, of which only the Nike on the E side has survived. The Golden Gate also had propylea, with columns of green marble, colonettes and pilasters, and was closed with gates of bronze (Macoudi) or iron covered with gold (Edressi)[12]


Unlike Constantinople, which kept her fourth-century size, most mediaeval enceintes are smaller than their antique counterparts, and built from spolia released by the shrinkage. An old idea is that such walls were built quickly as a response to immediate threat, so that the constructors, lacking the time to use newly cut stone (or perhaps lacking the necessary resources) have simply taken what came most conveniently to hand, and cleared away earlier monuments, possibly already in a state of collapse. This theory takes some believing, especially for anyone who has actually built anything, knows the ways of builders[13] or, indeed, has examined French enceintes such as Le Mans, Beauvais or Sens[14]. Shunning material already on or near the site, especially high-quality squared blocks, would have been perverse indeed, since every builder yearns for a straight edge and a square corner - qualities for which Roman materials were, as we have seen, much prized.[15] Even the apparently brutal reuse of funerary material (a practice at least as old as classical Athens) is for a good reason, since areas extra muros, where the cemeteries were located, needed to be cleared for military purposes, to ease troop movements. What is more, building brick or rubble walls would have been much easier, and surely equal to the scarce artillery and siege machinery of any likely invader. Thus we may conclude that admiration for the achievements of the past occasioned the use of spolia to produce smaller enceintes which yet retained the aura of the classical city about them, and proclaimed their heritage. That we must look for an aesthetic explanation (without necessarily imposing any concepts of decadence) for the appearance of spolia walls is underlined by parallel use in churches, where speed of construction did not matter.[16]  Even the argument that spolia were used for convenience does not negate an aesthetic appreciation of their qualities.


But the “rush” argument is not yet dead, the speed of marauding Arab cavalry[17] being paraded to explain what are perceived as badly-built walls in North Africa, with violence done to bas-reliefs – when it is obvious that these would be used face-inwards so as to present a smooth surface to the outside, with practicalities winning over aesthetics. As always, applying a different aesthetic viewpoint changes the nature of the argument. Thus at Thessaloniki, J.M. Spieser sees the walls as 249-250AD at the latest, and les remplois, provenant en grande partie des monuments funéraires, indiquent une reconstruction hative[18]. But his Plates XII & XIII show redan 86, with beautiful marble blocks in careful reuse, and scarcely hative. His Plate IX shows redan 97, south face, with hippodrome seats in re-use (they go as far south as Odos Egnatia). Since the hippodrome was located in the south corner of the walls, east-facing, we should ask why it is therefore that enormous quantities of its seats are re-used in the west walls, to a consistent height, and (by Odos Eirinis) carefully stepping down the hill so much so that the walls (partly dug out to their probable ground level) gleam white with marble? Surely because (by analogy with Nicaea and Constantinople), they flank the main gate – the Porta Aurea, and hence provide decoration in the most prestigiously classical manner possible. Similarly, the fort of the Heptapyrgion uses large quantities of column shafts, proud of the wall and angled, in careful courses alternating with tile; and at Odos Klavdianos, on the main west road, marble slabs, some of them fourteen feet in length, decorate the remaining complete bastion – another imitation or echo, perhaps, of the Golden Gate at Constantinople, and very splendid.


In Western Europe, such later defensive walls were usually dismantled in the 19th or early 20th centuries[19]; in Turkey, the excavators are in some cases still doing so, in order the better to study the monuments from which they were built. At the same time we can study how monuments became spolia. They might have been taken from complete as well as from ruinous monuments. Foss, for example[20], examining the seventh-century acropolis wall at Pergamum, and noting the use of the elements of the adjacent Temple of Faustina carefully laid in reverse order (architrave at the bottom, columns, and so on upward), suggests the whole structure was dismantled on the spot – in a spoliate version of “Chinese Towers”. This seems likely, for the same thing happened at Mytilene, where the mediaeval walls shelter at least seven identifiable structures, with the spolia grouped together[21]. In a slightly different case, at Athens, it has been pointed out that the architects of the Beulé Gate to the Acropolis consciously tried to imitate earlier works in their use of spolia, by dismantling a complete 4th century BC choragic monument: of Nikias, and incorporating in its new military setting much of the façade, with cornices, Doric frieze and epistyle blocks, complete with the identifying inscription. Nor is this chance for, in order to reassemble the pieces in the correct order, the workmen cut letters on some of the blocks while still in the original position, and then put them in the proper order to avoid unnecessary cutting and fitting[22]. This is most unusual, since the usual approach is a straight dismantling which (as at Pergamum) inverts the order of the elements; much the same happens with the complete monuments included in the Herulian Wall at Athens.[23] If the Beulé Gate was indeed built about the middle of the 3rd century AD[24], then this is conceivably the earliest use of meaning-rich spolia from the Graeco-Roman world. It is not the only sensitive translation of spolia for an identical use: at Bodrum, that the Knights did re-use for the original purpose when they could is suggested by the complete architrave beam from the Mausoleum built into the castle[25]. At Aphrodisias, the smallish hill to the south of site (and called the acropolis) was defended, using spolia from nearby monuments, including the theatre and the adjacent theatre baths, and arranging them decoratively. At Miletus, where the only highish ground has the theatre built against it, it was necessarily the theatre which was converted into a fortress by the Byzantines, towers of which are still to be seen[26]. At Seljuk, 23 coins of Constans have been found, which provides some dating evidence.[27]


Where high ground was not available, convenient monuments would be employed to help make a smaller set of defences, and reduce the amount of building work required. At the flat peninsular site of Side, for example, the extent of the Roman defences was effectively halved about the middle of the fourth century by building a second wall at a narrower point along the finger of land, and employing the theatre walls in the process. Here a high arched gateway flanking the theatre was roughly filled in with stones, to make an aperture small enough to close with a gate. Although obviously make-and-mend in view of the treatment meted out to the now-external monuments, it features column shafts and blocks, which are used in the new wall in great quantities. The date of such work is unclear but, according to Foss, who discards a dated inscription because he believes it is itself a spoil[28], the repairs to the Hellenistic walls are a response to the Isaurian threat, and the inner gate, with spolia, of the 7th century - just as the entrance to the citadel at Sardis boasts a third-century inscription, although it too is 7th century.


A concern for monumentality attained by the employment of spolia is frequently in evidence, not only in defensive walls, but in civic structures inside cities. At Side, this takes the form not merely of the careful decoration of the new walls with the spolia, but also of the construction of a new monumental gateway which, like the previous (outer) one, has a monumental fountain adjacent to it (although the fountain antedates the gate). At Aphrodisias, we find restoration of the Agora Gate and its adaptation as a fountain[29]. At Ephesus, the Library of Celsus was apparently ruinous at about the same period, and its façade was converted into a fountain embellished with spolia reliefs (Amazono- Giganto- and Centauromachies) brought from elsewhere in the city. Thus possible influence from one site to the other, and a suggested date of 435/6 – which accords with an increase in building activities at both sites in the mid- and later 5th century. However, at Side such refinement is not extended to the later gate, by the side of the fountain, which blocks up the large arch to make a much smaller opening, and column shafts are incorporated in the infill in regular courses. The effect for the outer face is arguably decorative but, although the courses are well-laid, using large blocks and slimmer entablature blocks, and presumably the granite column shafts from the adjacent agora, it is not easy to see what final effect was intended because the opening remains like gaping teeth in an open mouth. And on the inside, the builders have not even bothered to saw level those shafts which are too long, although perhaps they were rubbled or plastered over. At Xanthus, parts of several monuments were retrieved by Fellows from the defensive wall built on top of the theatre; at Bodrum, the Mausoleum reliefs were decoratively displayed in the Knights' castle; at Pergamum, sculptures from the Altar of Zeus were also retrieved from a Byzantine defensive wall which, however, did not display them decoratively - although Winnefeld wonders[30] whether Cyriaco’s reference to colossaeque de marmore deorum herounve simulachra might be sightings of pieces of the Gigantomachy; given later accounts, this is quite possible. This, and indeed the whole upper city, survived much better than the buildings in the valley. Ibn Battuta, for example, when he visited Pergamum in 1331, found the city in ruins, but with a formidable fortress; which we know displayed the usual column shafts, neatly laid, together with other marble spolia although not, apparently, reliefs.


In North Africa[31] there are handsome forts which are clearly Byzantine (because they are comparable with similar ones in Asia Minor), as well as fortifications which were clearly thrown up in a great hurry (and, as it now appears, in unsuitable places). In this latter category are the fort at Haidra, which is surely the most spectacular of all such works; and the fort of similarly large dimensions at Ain Tounga, overlooking the main road through the pass. But the Triumphal Arch at Haidra is also fortified by being encased in additional (spolia) masonry and, to modern eyes, looks a mess. Likewise the Capitol temple at Dougga was once presumably completely fortified. But we must remember local conditions, including lack of manpower, equipment, transport or quarries, all of which bedevilled the French in Algeria (see below).


One antique practice which, curiously, the Middle Ages did not try to imitate (except perhaps at Arles) was to relate the famous dead to the august antiquity of the city by associating standing funerary monuments with the walls and approach-roads (although on one occasion at Antioch tombs were dismantled for fort-building[32]). The practice was common in Turkey, so there was no shortage of possible models, such as the streets of tombs at Hierapolis, or great tomb terraces at Elaiussa Sebaste, surely designed to be seen from out to sea. Again, at Kyaneae, the north-east wall of the city is beautified (in much the same way as the southern approaches to Termessus) by siting sarcophagi on top[33]. In places, column shafts are also employed in a much less suave fashion. Since the main approach to the city was indeed from the north-east, is this then, simply imitation of the “conspicuous display”, of the earlier builders, with the walls and the crowds of free-standing Lycean sarcophagi emphasising the venerable antiquity of Kyaneae? Probably, because the city also boasted an elaborately “triumphal” entrance gate, with pilasters and capitals, if no reliefs, which must have impressed her later inhabitants[34]. We have seen how, for Leo Africanus, the incorporation of Roman inscriptions in enceintes is a type of triumphalism: might it also be argued that placing specifically funerary inscriptions in walls is itself intended as a proof of venerable antiquity?




[1] B. C. P. TSANGADAS, The fortifications and defense of Constantinople, New York 1980, p. 17. F. KRISCHEN, Die Landmauer von Konstantinopel, I, Berlin 1938, plates 41-44; and vol II, Berlin 1943, pp. 39-60 for a description and analysis;

[2] C. MANGO & R. SCOTT, The Chronicle of Theophanes Confessor, Oxford 1997, p. 572;

[3] A. FAILLER, editor, Georges Pachymères, Relations Historiques, 2 vols, Paris 1984. Born Nicaea 1242, and in Constantinople from 1261. Cf. II.31;

[4] SHAT. MR1616, p. 55;

[5] S.VRYONIS, Byzantine Constantinople and Ottoman Istanbul: evolution in a millenial imperial iconography, in I.A. BIERMAN et al editors, The Ottoman City and its Parts: urban Structure and Social Order, New Rochelle NY 1991, pp. 13-52; cf. pp. 33-4;

[6] MANGO & SCOTT, Theophanes, cit., p. 572;

[7] MANGO, Antique statuary, cit., p. 75;

[8] WILLIAM OF TYRE, Hist. Rerum Transmarin., J-P MIGNE, PL CCI, Paris 1903. I.10 (col 226) on an earlier undated recapture of Jerusalem by "the Turk", when subvertebant calices et vasa divinis obsequiis mancipata pedibus conculcantes, confringebant marmora - which could mean either that they defaced the marbles ( = "affect") or carried them away ( = "exhaust").

[9] TSANGADAS, The fortifications, cit.,  p. 17;

[10] A. VAN MILLINGEN, Byzantine Constantinople: the walls of the city and adjoining historical sites, London 1899, pp. 59-73;

220 R. KER PORTER, Travels in Georgia, Persia, Armenia, ancient Babylonia, etc etc, during the years 1817, 1818, 1819 and 1820, 2 vols, London 1821 & 1822. pp. 740-1;

[12] cf. R. JANIN, Constantinople Byzantine, (Archives de l’Orient Chretien 4), Paris 1950. cf. pp. 252ff;

[13] FOSS, Ephesus, cit, pp. 96-9 T.C.F. BLAGG, The reuse of monumental masonry in late Roman defensive walls, in J. MALONEY & B. HOBLEY eds, Roman urban defences in the West (CBA Research Report 51), London 1983, pp. 130-5;

[14] S. JOHNSON, Late Roman fortifications, London 1983, passim, for a survey;

[15] D. FIORANI, Tecniche costruttive murarie medievali: Il Lazio meridionale, Rome 1996. Pp. 87ff for reworking of spolia, and their use in defensive structures and churches;

[16] D. PRINGLE, The defence of Byzantine Africa from Justinian to the Arab Conquest, 2 vols, Oxford 1981, p. 133;

[17] RAVEGNANI, Castelli, cit., p. 55;

[18] J.M. SPIESER, Thessalonique et ses monuments du 4e au 6e siècle. Contribution à l’étude d’une ville paléochrétienne, Paris 1984 ( Bibliothèque de l’Ecole Française d’Athènes et de Rome 254),, p. 64; and see also J. M. SPIESER, Note sur la chronologie des remparts de Thessalonique, in Bulletin de Correspondance Hellenique, 98 (1974), pp. 707-19;

[19] F. M. ULUHOGIAN, Demolizione delle mura ed espansione urbana, in C. DE SETA & J. LE GOFF, eds., La citta e le mura, Rome & Bari 1989, pp. 371-86; and L. PRINCIPI, Uccidere le mura. Materiali per una storia delle demolizioni in Italia, in ibid, pp. 387-417, for a mournful and very long catalogue of walls demolished since 1862;

[20] C. FOSS & D. WINFIELD, Byzantine fortifications: an introduction, Pretoria 1986, p. 133;

[21] C. WILLIAMS, Hellenistic and Roman buildings in the mediaeval walls of Mytilene, in Phoenix Toronto, 38 (1984), pp. 31-70. See p. 34;

[22] The Beule Gate illustrated in J. TRAVLOS, Pictorial Dictionary of Ancient Athens, London 1971, figs 462-3, s.v. Nikias Monument. And cf. T. GREGORY, The fortified cities of Byzantine Greece, in Archaeology, 35 (19820,  pp. 14-21; see especially p. 16;

[23] cf. FRANTZ, The Athenian Agora, cit., p. 134: Tower W6 incorporated blocks from a monument to Attalos, by his Stoa (destroyed in the Herulian attack), which stood in front of the position the tower took; and the same Tower W6 incorporated 97 column drums from the same Stoa in its side wall, together with metopes, triglyphs, capitals, etc;

[24] W.B. DINSMOOR, The Choragic Monument of Nicias, AJA 14, 1910, 459-484, believed (p. 484) that it must date from late Roman or even Byzantine times;

[25] G.B. WAYWELL, The free-standing sculptures of the Mausoleum at Halicarnassus in the British Museum: a catalogue, London 1978, pl. 45.1;

[26] F. KAUSS, Das Theater von Milet, (Milet: Ergebnisse der Ausgrabungen und Untersuchungen seit dem Jahre 1899, IV.1), Berlin 1973, plates 126-132; and cf. the remarks on the late city walls in A. VON GERKAN, Milet: Die Stadtmauern, (Milet: Ergebnisse der Ausgrabungen und Untersuchungen seit dem Jahre 1899, II.3) Berlin & Leipzig 1935, pp. 114-117; W. MULLER-WIENER, Mittelalterliche Befestigungen im suedlichen Ionien, in Ist. Mitt., 11 (1961), pp. 5-122. Pl.XIV.2 for a row of columns en boutisse in the east castle wall at Priene;

[27] C. FOSS, Ephesus, cit. 111ff for picture of the dark-age city, with a path across the abandoned agora, covered even then with debris a metre thick, leading to a north entrance made completely of spoils;

[28] C. FOSS, Attius Philippus and the walls of Side, in Zeitschrift fuer Papyrologie und Epigrafik, 26 (1977), pp. 172-80;

[29] C. ROUECHE, Aphrodisias in Late Antiquity, London 1989, p. 69; For an illustration of the Nymphaeum, cf. F. A. BAUER, Stadt, Platz und Denkmal in der Spaetantike. Untersuchungen zur Ausstattung des oeffentlichen Raums in den spaetantiken Staedten Rom, Konstantinopel und Ephessos, Mainz 1996, plate 30.3;

[30] cf. H. WINNEFELD, Alterthuemer von Pergamon, III.2: Die Friese des groszen Altars, Berlin 1910, pp. 1-6 for excavation notes; nevertheless A. CONZE et al, Alterthuemer von Pergamon, I..1: Stadt und Landschaft, Berlin 1912, p. 3; cf. FELLOWS, Journal, p. 35: at Pergamum, the walls of the Turkish houses are full of relics of marble, with ornaments of the richest Grecian art;

[31] PRINGLE, Byzantine Africa, cit., passim;

[32] ACADEMIE DES INSCRIPTIONS, Croisades, cit., vol. IV, Baldrici Episcopi Dolensis Historia Jerosolimitana, IV, 51F, anno 1098: they build at Antioch a castellum et de lapidibus ipsis quos de sepulturis dehumatorum abstraxerant munire deliberaverunt - although not clear whether these are modern graves (i.e. to insult the Turks), or antique ones reused for the stone;

[33] cf. F. KOLB, editor, Lykische Studien 2: Forschungen auf dem Gebiet der Polis Kyaneai Kampagne 1991, Bonn 1995, Tafel X & p. 69;

[34] F. KOLB, editor, Lykische Studien 3: Die Siedlungskammer von Kyaneai Sommer 1992, Bonn 1996, pp. 21ff: Ein Dreitoriger Ehrenbogen in Kyaneai;