I am grateful to the Management Committee of the British School at Athens for allowing me to publish this set of drawings.

¹ Papageorgiou, P. N., ‘Μνημεῐα τῆς ἐν Θεσσαλονίκῃ λατρείας τοῦ μεγαλομάρτυρος’, Byzantinische Zeitschrift xvii (1908) 321–81Google Scholar (afterwards cited as Papageorgiou (1908)). An account of the building previous to 1907 is supplied by Texier, C. and Pullan, R. P., Byzantine Architecture (London, 1864) 123–30.Google Scholar

² The Turkish record in the form Hegira 898 was already known by Texier and Pullan; various transcriptions are on offer in the literature.

³ For plans of the church see Soteriou, G. and Soteriou, M., Ἡ Βασιλική τοῦ Ἁγίου Δημητρίου Θεσσαλονίκης (Athens, 1952)Google Scholar, afterwards cited as G. and M. Soteriou (1952).

⁴ The value of the activities of Strzygowski, Millet, van Millingen, and Mordtmann has long been recognized. C. Mango, in his editor's preface to Ainalov, D. V., The Hellenistic Origins of Byzantine Art (New Brunswick, N.J., 1961)Google Scholar, pays tribute to the fundamental contribution of Russian specialists in this field. The work sponsored by the Byzantine Research Fund of the British School at Athens, of which the material published in this paper is only one example, belongs to this period of research.

⁵ The expedition is recorded in the Report of the Institute's activities in 1908 in the xiv (1909) 169–71. Their visual record appears as plates 1–20 of this periodical. The results of their observations were published in two papers in the same tome: T. E. Uspenskij, 1–61, afterwards cited as Uspenskij (1909); and N. K. Kluge, 62–7, afterwards cited as Kluge (1909).

⁶ Papageorgiou (1908). His transcriptions were not free from errors (e.g. κώστας for κτίστας).

⁷ For an account of this restoration work and of problems connected with Sophia, S., see my Ph.D. thesis, Ninth Century Monumental Painting and Mosaic in Thessaloniki (University of London, 1968).Google Scholar

⁸ Diehl, C. and le Tourneau, M., ‘Les Mosaiques de Saint-Démétrius de Salonique’, Monuments et Mémoires (Fondation Piot) xviii (1910) 225–47Google Scholar, afterwards cited as Diehl (1910). This account formed the basis for the section in Diehl, C., le Tourneau, M., and Saladin, H.Les Monuments Chrétiens de Salonique (Paris, 1918)Google Scholar, and afterwards cited as Diehl (1918).

⁹ le Tourneau, M. and Diehl, C., ‘Les Mosaïques de Sainte-Sophie de Salonique’, Monuments et Mémoires (Fondation Piot) xvi (1908–1909) 39–60.Google Scholar Diehl is responsible for the notion of a chronological division between the peripheral figures and mandorla group in the cupola, rejected by all critical observers.

¹⁰ Tafrali, O., ‘Sur la date de l'église et des mosaïques de Saint-Démétrius de Salonique’, Revue Archéologique, 4th series, xiii (1909) 83–101Google Scholar, afterwards cited as Tafrali (1909); and O. Tafrali, ‘Sur les réparations faites au VII^e siècle à l'église de Saint-Démétrius de Salonique’, ibid., 4th series, xiv (1909) 380–6.

¹¹ Boissonnas, F., Salonique et ses Basiliques, avec un introduction de Baud-Bovy, D. (Geneva, 1919)Google Scholar, afterwards cited as Boissonnas (1919).

¹² The lack of records is lamented by Grabar, A., Martyrium (Paris, 1946) vol. 2, 87 n. 2.Google Scholar This study relies mainly on the Russian material (afterwards cited as Grabar (1946)). Grabar, A., L'Iconoclasme byzantin (Paris, 1957) 84 ff.Google Scholar submits the same material to a further scrutiny; afterwards cited as Grabar (1957). The Russian material was partly reprinted by Kondakov, N. P., Иконорафія Ъоомамерu. vol. 2 (Petrograd, 1915)Google Scholar; and used also by Hoddinott, R. F., Early Byzantine Churches in Macedonia and Southern Serbia (London, 1963)CrossRefGoogle Scholar; afterwards cited as Hoddinott (1963). Lazarev, V. N., Исморuя бuзанмuŭскоŭ Мuбоnuсu (Moscow, 1947)Google Scholar selected his illustrations from both the Russian and French collections.

¹³ Rice, D. Talbot, ‘The lost mosaics of St. Demetrios, Salonica’ Πεπραγμένα τοũ Θ' Διεθνοũς Βνзαντινολογικοũ Συνεδρίου, vol. 1 (Athens, 1955) 474.Google Scholar A detail of the set was photographed to assist the research of Theotoka, N., ‘Περί τῶν κιβωρίων τοũ ἁγίου Δημηρίου Θεσσαλονίκης καί Κωνσταντινοπόλεως’ Μακεδονικά ii (1953) 395–413.Google Scholar The Byzantine Research Fund collection also includes photographs taken in S. Demetrios by W. M. Harvey before 1917 (Plate 15).

¹⁴ Kitzinger, E., ‘Byzantine Art in the period between Justinian and Iconoclasm’, Berichte zum XI Internationalen Byzantinisten-Kongress (Munich, 1958) 20 and n. 74.Google Scholar Afterwards cited as Kitzinger (1958).

¹⁵ Dalton, O. M., Byzantine Art and Archaeology (Oxford, 1911) 378–82, and figs. 198 and 224.CrossRefGoogle Scholar Afterwards cited as Dalton (1911).

¹⁶ A plan was drawn up in 1933 for a joint English-Greek corpus on the churches of Thessaloniki, and work was begun by A. H. S. Megaw and others. G. and M. Soteriou (1952) incorporated some of the survey material into their monograph.

¹⁷ George, W. S., The Church of Saint Eirene at Constantinople (Oxford, 1912).Google Scholar For an appreciation of the accuracy of this survey, see Perkins, J. B. Ward in The Great Palace of the Byzantine Emperors, Second Report, edited by Rice, D. Talbot (Edinburgh, 1958) 73.Google Scholar

¹⁸ For example, the two panels on the left pier. Sheet 16 is a complete tracing of the inscription below the south face. It begins with an area of loss followed by broken letters (Plate 14b). The distance between the left end of the panel and the alpha of άνθρώποις was 6¾ in. (17 cm.). Uspenskij (1909) restored this as [Δῶρ] ον after considering [Δόμ]ον. G. and M. Soteriou (1952) 195 give [Ὁ ἐν]. Neither solution appears satisfactory. The inscription reads: […]ον ἀνθρόποις ἀπελπ[ιο]θεὶς παρἀ δὲ τῆς σῆ(ς) δυνάμεως ӡωοποιηθὶς εύχαριστῶν ἀνεθέμην The proper name Κλήμης given by G. and M. Soteriou (1952) 195, did not appear in 1909. The palaeographical evidence in this church still awaits treatment. For example, the similarities of this inscription with the forms exhibited on the founder's panel facing it on the right pier require assessment to test the hypothesis that stylistically this Virgin panel is homogeneous with phase two mosaics. Cf. Hoddinott (1963) 154–5; to whose arguments might be added the use of the lotus ornament of the architrave in this panel, which appears in the ornamental mosaics of the north inner aisles; and the comb-like hatching on the saint's arm in this panel, which recurs on the saint's arms on the east panel of this pier.

Concerning this east panel, George's record confirms that no inscription was preserved to identify the saint. G. and M. Soteriou (1952) 194–5 identified him as S. Demetrios, but Kitzinger (1958) suggested S. Bacchos, as a pendant to S. Sergios on the right pier. Underwood, P. A., ‘The Frescoes in the Kariye Çamii’, Dumbarton Oaks Papers xiii (1959) 189 n. 9Google Scholar, rejects both identifications and proposes that this type represents S. George. If this is correct, the three panels of the left pier may be the trio of warrior saints, George (west), Theodore (south), and Demetrios (east).

¹⁹ Sheet 9 published in black and white by Dalton (1911), fig. 198.

²⁰ Description by Nordhagen, P. J., ‘The Mosaics of John VII (705–707 A.D.), Acta ad Archaeologiam et artium historiam pertinentia ii (1965) 121–66, esp. 158–9.Google Scholar Afterwards cited as Nordhagen (1965). Nordhagen apparently underestimates the significant proportion of green tesserae in the beard. The predominant colour in the neutral flesh parts of the face is buff. The possibility of fading since 1909 must be reckoned with.

²¹ Sheet 5 published in black and white by Dalton (1911), fig. 224.

²² Krautheimer, R., Early Christian and Byzantine Architecture (London, 1965) pl. 32Google Scholar, reproduces Boissonnas (1919) pl. 27, of the north side of the church (taken in 1913).

²³ Texier, C. and Pullan, R. P., Byzantine Architecture (London, 1864) 123–30.Google Scholar

²⁴ Tafrali (1909) prints the additional text found in Paris, Bibliothèque Nationale, gr. 1517. The priest mentioned may be the figure portrayed on the east face of the right pier, cf. Hoddinott (1963) 152–3; the portrait reappears in the left of the three central medallions in the north inner aisle mosaics (Plate 4).

²⁵ G. and M. Soteriou (1952), esp. 143–6.

²⁶ Diehl (1910) 238 ff., Diehl (1918) 104 ff., and Diehl, C., ‘A propos de la mosaïque d'Hosios David à Salonique’, Byzantion vii (1932) 333–8.Google Scholar Cf. Bertelli, C., La Madonna di Santa Maria in Trastevere (Rome, 1961) III, n. 15Google Scholar, who applies this principle to the enthroned Virgin panel, and deduces a seventh-century date, preferably in the time of Justinian II in Thessaloniki. Grabar (1946) and (1957) assumes a licence of similar free dissociation in dating panels.

²⁷ The nature of mosaic sutures is discussed by Underwood, P. A., ‘The evidence of restorations in the sanctuary mosaics of the Church of the Dormition at Nicaea’, Dumbarton Oaks Papers xiii (1959) 235–42.CrossRefGoogle Scholar

²⁸ Dalton (1911) 378–82, quotes George as saying, ‘the dividing line between the work of the two periods is conspicuous’.

²⁹ This medallion head is the only part where the line of the suture was obscure. George (sheet 3) writes, ‘halo certainly belongs to the restoration, probably the whole panel also’. On this point Kluge appears the more careful observer. At another point (arch 4), George may be the more correct in seeing the donor below S. Matrona as homogeneous, cf. Kluge (1909) 65.

³⁰ n. 74.

³¹ Illustrated in sheets 1 and 10; Papageorgiou (1908) pl. 1, 1; Uspenskij and Kluge (1909) pl. 17; Boissonnas (1919) Pl. 31.

³² G. and M. Soteriou (1952) fig. 77.

³³ n. 92.

³⁴ When studied by Bayet, C., Mémoire sur un ambon conservé à Salonique, in Mémoire sur une mission au Mont Athos (Paris, 1876) 249–99Google Scholar, the ambo was in two pieces, one outside the Rotonda and the other outside S. Panteleimon. Mendel, G., Catalogue des sculptures (Constantinople, 1914) 2, 392–405Google Scholar, attributed it to the Rotonda, and dated it to the first half of the sixth century, presumably on the grounds that the mosaics of the church belonged to that time. Recently, Grabar, A., Sculptures byzantines de Constantinople, IV^e–X^e siècle (Paris, 1963) 81–4Google Scholar, has reaffirmed an early sixth century date by comparison with the S. Polyeuctos finds. But since the date of the church conversion of the Rotonda has been set by Torp, H., Mosaikkene i St. Georg-Rotunden i Thessaloniki (Oslo, 1963)Google Scholar, at around 400, a similar dating for the ambo requires consideration.

³⁵ Cf. Grabar (1957) 85 n. 5.

³⁶ Translation: ‘As a prayer for one whose name God knows.’

³⁷ G. and M. Soteriou (1952) pl. 71(a); and Diehl (1918) 57 and 95. I am throughout this paper using the name Basilica of the Virgin for the present church of S. Paraskeue; for its full Byzantine appellation see Pelekanides, S., Παλαιοχριστιανικà Μνημεῖα Θεσσαλονίκης. ἈχειροποίητοςΜονή λατόμον (Thessaloniki, 1949)Google Scholar, and Xyngopoulos, A., Αἱ περì τοῦ ναοῦ τῆς Ἀχειροποιήτου εἰδήσεις τοῦ Κωνσταντίνου Ἁρμενοπούλου Τόμος Κωνσταντίνου Ἁρμενοπούλου (Thessaloniki, 1952) 1–26.Google Scholar

³⁸ Hoddinott (1963) pl. 20(d).

³⁹ Illustrated in sheet 1; Papageorgiou (1908) pl. 1, 2 and pl. 2, 3; Uspenskij and Kluge (1909) pl. 12.

⁴⁰ Illustrated in colour by Hoddinott (1963) pl. 4, facing p. 142.

⁴¹ This type of banded ornament must have been widespread in the Greek East; it occurs frequently in the manuscript Paris, Bibliothèque Nationale, gr. 510 (e.g. folios 71 verso, 149, 239, 355), whose artists must have been familiar with pre-iconoclastic models.

⁴² Cf. the representation in the manuscript Vatican, gr. 1613 folio 61. Since Diehl speaks of a prophet, Hoddinott (1963) 145, takes him to mean Zechariah.

⁴³ Illustrated and discussed by Janashian, M., Armenian Miniature Paintings of the Monastic Library at San Lazzaro, i (Venice, 1966)Google Scholar; MS. no. 1144/86.

⁴⁴ Illustrated on sheets 2 and 18; Papageorgiou (1908) pls. 2, 3, and 4; Uspenskij and Kluge (1909) pls. 3 and 5 (in colour) and 10 and 11; Diehl (1910) pl. 16 (colour); and Diehl (1918) pls. 27 and 34, and fig. 45.

⁴⁵ Mango, C. and Hawkins, E. J. W., ‘The Apse Mosaics of St. Sophia at Istanbul’, Dumbarton Oaks Papers xix (1965) 115–51.Google Scholar

⁴⁶ Schazmann, P., ‘Des Fresques byzantines récemment découvertes par i'auteur dans les fouilles à Odalar Çamii, Istanbul’, Studi bizantini e neoellenici vi (1940) 372–86.Google Scholar The dating of this composition requires re-examination.

⁴⁷ Oakeshott, W., The Mosaics of Rome (London, 1967) 378–9.Google Scholar

⁴⁸ Bertelli, C., La Madonna di Santa Maria in Trastevere (Rome, 1961), esp. 88–9Google Scholar; he finds the contrast here sufficiently marked to postulate two painters.

⁴⁹ Schmidt, T., Die Koimesis-Kirche von Nikaia (Berlin and Leipzig, 1927).Google Scholar

⁵⁰ Illustrations of Hosios David in Hoddinott (1963) pl. vii, facing p. 174; and of the Cyprus apse in Megaw, A. H. S., ‘The mosaics in the church of Panayia Kanakaria in Cyprus’, Studi bizantini e neoellenici viii (1953) 199–200.Google Scholar

⁵¹ Cf. Underwood, P. A., ‘The frescoes of the Kariye Çamii’, Dumbarton Oaks Papers xiii (1959) 194 n. 20.Google Scholar

⁵² G. and M. Soteriou (1952) 233–40, fig. 93. The close association of the cult of the Virgin and S. Demetrios was a commonplace of local religious attitudes. Cf. Laurent, V., ‘Une homélie inédite de l'archevêque de Thessalonique, Léon le Philosophe sur l'Annonciation (25 mars 842)’, Mélanges Tisserant (Vatican, 1964) vol. ii, 281–302, esp. 293–4.Google Scholar

⁵³ Grabar (1946) 2, 98; and Grabar, A., ‘A propos d'une icone byzantine du XIV^e siècle au Musée de Sofia’, Cahiers archéologiques X (1959) 289–304, esp. 298.Google Scholar

⁵⁴ Migne, J. P., Patrologia Graeca, vol. 116, 1278.Google Scholar

⁵⁵ George, W. S., The Church of St. Eirene at Constantinople (Oxford, 1912) 51–2.Google Scholar

⁵⁶ p. 379 n. 1.

⁵⁷ Cf. Underwood, P. A. and Hawkins, E. J. W., ‘The Mosaics of Hagia Sophia at Istanbul; the Portrait of the Emperor Alexander’, Dumbarton Oaks Papers XV (1961) 189–217, esp. 194 n. 25.Google Scholar

⁵⁸ Weitzmann, K. and others, Icons from South Eastern Europe and Sinai (London, 1968) pls. 1–3.Google Scholar

⁵⁹ Cf. Papageorgiou (1908), 342. Inscription two translation: ‘And the Lady, the holy Mother of God …’ Inscription three translation: ‘And you, my Lord Saint Demetrios, aid us your servants and your servant Maria, whom you gave to us.’

⁶⁰ Cf. Diehl (1910). The notion of a ‘miraculous’ birth is perhaps too extreme.

⁶¹ Illustrated on sheet 2; Papageorgiou (1908) pl. 3, 5; Uspenskij and Kluge (1909) pl. 10.

⁶² Theotoka, N., ‘Περί τῶν κιβωρίων τοũ ἁγίου Δημηρίου Θεσσαλονίκης καί Κωνσταντινοπόλεως’ Μακεδονικά ii (1953) 395–413Google Scholar, reconstructs the ciborium. She assumes its form before and after a fire in 580 or 581 and its restoration before 586 to be unaltered. She accepts the hexagonal base found on the left side of the nave in 1917 to represent the position of the ciborium since the foundation of this church, and to indicate its shape throughout the Byzantine period. In her opinion, some church of S. Demetrios in Constantinople must have contained an octagonal ciborium, whose existence may be inferred from the reliquary copy in Moscow. Grabar, A., ‘Quelques reliquaires de Saint Démétrios et le martyrium du saint à Salonique’, Dumbarton Oaks Papers V (1950) 3–28Google Scholar, concluded that the ciborium in Thessaloniki was not altered in the eleventh century.

⁶³ On Christ types cf. Breckenridge, J. D., The Numismatic Iconography of Justinian II (New York, 1959), esp. 46–62.Google Scholar This bust would be closest to his type B, but a considerable variety of types is found in the pre-iconoclastic mosaics of Thessaloniki. Breckenridge argues that an unsupported book, implying the existence of a left hand, indicates a reduced version of a larger representation, and so cannot be strictly called an imago clipeata (which he defines as a medallion portrait complete within its frame). Chatzidakis, M., ‘An Encaustic Icon of Christ at Sinai’, Art Bulletin xlix (1967) 197–208CrossRefGoogle Scholar, discusses prototypes influential on early representations of Christ, and suggests the Pantocrator Christ on the Sinai icon, shown with a small beard, might derive from the Chalke icon, which he dates after 532. However, the destruction of this postulated model in the 532 riots is not certain; and the developments leading up to this representation would require definition.

⁶⁴ Hoddinott (1963) 148–9, gives his approval to a syncretist theory of a Great Goddess-Eutaxia-Virgin fusion; but the relative values of historical fact and psychological speculation in his account are not clarified.

⁶⁵ Cf. Weitzmann, K., ‘The Survival of Mythological Representations in Early Christian and Byzantine Art, and their impact on Christian Iconography’, Dumbarton Oaks Papers xiv (1960) 45–68, esp. 55.Google Scholar Grabar (1957) fig. 87, supplies a formal parallel, but the reference there is to the death of a child.

⁶⁶ e.g. the diptych in the State Museum, Berlin; illustrated by Volbach, W. F., Early Christian Art (London, 1961), pl. 224.Google Scholar Cf. also the Sinai mosaic medallion, illustrated by Weitzmann, K., ‘The Classical in Byzantine Art’, Byzantine Art, Lectures (Athens, 1966) fig. 132.Google Scholar

⁶⁷ Illustrated in sheet 3; Papageorgiou (1908) pls. 3, 5, and 6; Uspenskij and Kluge (1909) pls. 2 and 9; Diehl (1910), pl. 19, 2; Diehl (1918), pl. 30, 2 and fig. 46. Kluge's colour plate is a valuable supplement to George's copy.

⁶⁸ Illustrated by Hoddinott (1963) pl. v, facing p. 156.

⁶⁹ Illustrated in sheets 3 and 11; Papageorgiou (1908) pls. 3, 6 and 4, 7; Uspenskij and Kluge (1909) pl. 8.

⁷⁰ Cf. Kitzinger, E., ‘The Cult of Images in the age before Iconoclasm’, Dumbarton Oaks Papers viii (1954) 85–150.Google Scholar

⁷¹ Illustrated in sheets 4 and 12; Papageorgiou (1908) pl. 4, 7 and 8; Uspenskij and Kluge (1909) pls. 6 and 7; Diehl (1910) pl. 19, 3; Diehl (1918) pl. 30, 3; Boissonnas (1919) pl. 29.

⁷² Cf. Morey, C. R., Early Christian Art (Princeton, 1942) 139Google Scholar, for the ‘Italo-Gallic school’ attribution of this form.

⁷³ p. 146, following Diehl (1918).

⁷⁴ Lafontaine-Dosogne, J., Iconographie de l'enfance de la Vierge dans l'Empire byzantin et en Occident vol. i (Brussels, 1964) 49–50.Google Scholar

⁷⁵ P. 347.

⁷⁶ Illustrated in sheets 4, 8, 13, and 14; Papageorgiou (1908) pl. 4, 8; Uspenskij and Kluge (1909) pls. 1, 4, and 6; Diehl (1910) pl. 19, 1; Diehl (1918) pl. 30, 1. See G. and M. Soteriou (1952), fig. 77, and Procopiou, A., The Macedonian Question in Byzantine Painting (Athens, 1962) pl. 10Google Scholar, for illustrations of the surviving fragment. (A colour slide is available from the firm Lykides.)

⁷⁷ Cf. Demus, O., Byzantine Mosaic Decoration (London, 1948) 38.Google Scholar

⁷⁸ (1965), esp. 146–7.

⁷⁹ p. 343.

⁸⁰ p. 59.

⁸¹ p. 379 n. 1.

⁸² (1965), esp. 150 ff. His theory has difficulties; e.g. SS. Cosmas and Damian must be interpreted as local work, and S. Lorenzo fuori le Mura must be severely carved up into sections.

⁸³ Illustrated in sheets 1–4; Papageorgiou (1908), esp. pl. 3, 6; Uspenskij and Kluge (1909), esp. pls. 8 and 9; Diehl (1918) pls. 30, 32 and fig. 46.

⁸⁴ Translation: ‘Made young again in the times of Leo, you see the church of Demetrios, previously burnt.’ The theme is clearly restoration. The first vowel of Demetrios was written as η, not ι, as it was in the inscription three of phase one.

⁸⁵ Illustrated by Hoddinott (1963) pl. 33 (c). Probably it was this same wheel shape which appeared on the thigh of S. Onesiphoros in the Rotonda, obscured by his chlamys: illustrated by Volbach, W. F., Early Christian Art (London, 1961) pl. 126.Google Scholar

⁸⁶ The wall-painting is visible in Papageorgiou (1908) pl. 4, 8. The state of this section of the arcade is illustrated by Boissonnas (1919) pl. 27. G. and M. Soteriou (1952) date the extreme east end of the arcade to the seventh-century restoration.

⁸⁷ Papageorgiou (1908) pl. 11, 1; and Uspenskij (1909) 59–61, give their transcriptions of the tituli.

⁸⁸ p. 335.

⁸⁹ Nordhagen, P. J., ‘The Frescoes of John VII (A.D. 705–7) in S. Maria Antiqua in Rome’, Acta ad Archaeologiam et Artium Historiam pertinentia, iii (1968) 110Google Scholar, is ready to accept a mid-seventh-century date for at least one of the wall-paintings in S. Demetrios; cf. G. and M. Soteriou (1952) pl. 75, pp. 204 ff.

⁹⁰ See esp. pp. 246–7. Out of 108 fifth-century coins, only one belonged to the first half of the century.

⁹¹ Krautheimer, R., Early Christian and Byzantine Architecture (London, 1965) 96–7 and 328–9 n. 49.Google Scholar Certain of the chronological features adduced by him require proper documentation (e.g. ‘patterns of marble revetment, brickwork of piers, and rhythm of support and arcades’).

See pp. 90–101 for discussion of late fifth-century churches in the Balkans.

⁹² Kautsch, R., Kapitellstudien (Berlin and Leipzig, 1936), esp. 72–5.Google Scholar

⁹³ For a convenient bibliography on these churches see Krautheimer and Hoddinott (1963); the latter favours an early fifth-century date for the group, but with insufficient justification.

⁹⁴ Cf. Kitzinger (1958) n. 86, who accepts a date c. 450.

⁹⁵ On the secondary literature, see Kitzinger (1958) nn. 101–2.

⁹⁶ Cited in n. 10.

⁹⁷ Petit, L., ‘Les évêques de Thessalonique’, Échos d'Orient iv (1900–1901) 136–45, 212–21, and v (1901–2) 26–33, 90–7, 150–6, 212–19, has the documentation for the pre-iconoclastic archbishops.Google Scholar

⁹⁸ Bariŝić, F., Чуәа Дuмuмрuја Сопунско као uсморuскu uзборu (Belgrade, 1953), esp. pp. 81 ff.Google Scholar, proposed a lacuna of some considerable time between chapters 1–3 and 4–6, but this was not the opinion of two other scholars; Lemerle, P., ‘La composition et la chronologie des deux premiers livres des Miracula S. Demetrii’, Byzantinische Zeilschrift xlvi (1953) 349–61Google Scholar; Lemerle, P., ‘Invasions et migrations dans les Balkans depuis la fin de l'époque romaine jusqu'au VIII^e siècle’, Revue Historique ccxi (1954) 265–308Google Scholar; and A. Burmov, xlvii (1952) 167–213. Burmov dates Perbund to 645–7 and Kouver to the 660s; Barišić and Lemerle date the Perbund episode to 674–7. Barišić dates Kouver to 680–5, whereas Lemerle identifies him with the Bulgar Kouvrat or Kovratos, known to have been in revolt in 635, and thus dates chapter 5 to 635–. This latter dating reverses the chronological order of chapters 4 and 5, and would give a terminus ante quem for the fire of 635–.

⁹⁹ Cf. Breckenridge, J. D., ‘The “long siege” of Thessalonica; its date and iconography’, Byzantinische Zeitschrift xlviii (1955) 116–22.Google Scholar

¹⁰⁰ For a definition of ἒπαρχος or ὕπαρχος see Lemerle (1954), cited in n. 98, pp. 270 ff.

¹⁰¹ Traced in sheet 17: ✠ Κτίστας θεωρεῐς τοῦ πανε[ν]δόξου δόμου ἐκεῐθεν ἒνθεν μάρτυρος Δημητρίου/τοῦ βάρβαρον κλύδωνα βαρβάρων στόλω(ν) μετατρέποντος κ(αὶ) πόλιν λυτρουμένου ✠. As in inscription five, the first vowel of Demetrios was an η. The founders must be contemporary with the phase two restoration; one is probably the likeness of the living archbishop (later in date than John); the other of the prefect Leo, if Tafrali (1909) is right. If the priest on the east face of this pier is the one of chapter 3, then the restoration took place in his lifetime too.

¹⁰² Esp. p. 22.

¹⁰³ Esp. p. 26.

¹⁰⁴ Esp. p. 21. Among the number of those who deny or ignore the relevance of the fire as a terminus ante quem may be added Grabar (1957), esp. 85. No reassessment of the evidence has appeared since Kitzinger; cf. however comments by Hoddinott (1963), 141–55; Matthiae, G., ‘La Cultura figurativa di Salonicco nei secoli V e VI’, Rivista di archeologia cristiana 38 (1962) 163–213Google Scholar; and Kalokyris, K., ‘La basilique Saint Démétrius de Thessalonique: ses mosaïques’, Corso di cultura sull';arte ravennate e bizantina xi (1964) 224–36.Google Scholar

¹⁰⁵ Lazarev, V. N., Исморuя бuзанмŭuскоŭ мuбоnu (Moscow, 1947) 56 ff. and 288–9 n. 12.Google Scholar The second edition reaffirms his opinion, Storia della pittura bizantina (Turin, 1967) 35 and 56 n. 24.

¹⁰⁶ G. and M. Soteriou (1952) 103 ff.; 205, fig. 26; pl. 27.

¹⁰⁷ The only earlier possible occasion for major alterations known from the Miracula was the burning and reconstitution of the ciborium between 580 and 586 (Book 1, Chapters 6 and 12). However there is no mention of serious damage or repair work to the structure. This occasion is of course also a possible one for the mosaics of phase one.

¹⁰⁸ See n. 105.

¹⁰⁹ The theory of the middle and second half of the seventh century as a period of great artistic activity in Thessaloniki derives from Laurent, J., ‘Sur la date des églises St-Démétrius et Ste-Sophie à Thessalonique’, Byzantinische Zeitschrift iv (1895) 420–34Google Scholar, where it depends on erroneous attributions of the structures of both these churches to this period. A more realistic picture of conditions in Thessaloniki is given by Teall, J. L., ‘The Grain Supply of the Byzantine Empire, 330–1025’, Dumbarton Oaks Papers xiii (1959) 89–139Google Scholar and G. Ostrogorsky, ‘The Byzantine Background of the Moravian Mission’, ibid. xix (1965) 3–18. It must be admitted that the Miracula despair over money and workmen (cited n. 24) might be seen as an indication of outside assistance for the restoration.

¹¹⁰ Cf. Weitzmann, K., ‘The Mosaic in St. Catherine's Monastery on Mount Sinai’, Proceedings of the American Philosophical Society cx (1966) 392–405.Google Scholar

¹¹¹ His list on p. 23 may be increased: e.g. the ornamental use of oval and rectangular cabochons, cornice decoration on the aediculae, alternation of straight and twisted fluting on columns, forms of capitals, details of garments, etc. in the north inner aisle mosaics may derive from the Rotonda; so might the fir-cone form on the aedicula of the south aisle donor panel, which also appeared in the Onesiphoros panel.

¹¹² Lazarev (cited in n. 105) in 1947, p. 41, accepted an early sixth-century date for the Rotonda, but revised his opinion in 1967, p. 35 and p. 56 n. 24. When he accepted the attribution to c. 400, he made no corresponding adjustments to his dating of other mosaics in Thessaloniki.

¹¹³ The face of S. Demetrios on the west wall of the south aisle has, I suspect, suffered from modern restoration work. The saint stands in front of a panelled door, presumably of his ciborium.

¹¹⁴ Illustrated by Hoddinott (1963) pl. 32 (b).

¹¹⁵ Cf. M. Chatzidakis, cited in n. 63, who believes the icon was part of a set sent from Constantinople for the foundation.

¹¹⁶ Illustrated by Volbach, W. F., Early Christian Art (London, 1961) pl. 225.Google Scholar

¹¹⁷ The scheme is characterized as ‘Trabantenbild’ by Schramm, P. E., ‘Das Herrschbild in der Kunst des frühen Mittelalters’, Vorträge der Bibliothek Warburg ii (1922–1923) 1, 145–224, esp. 177 ff.Google Scholar

¹¹⁸ Cf. Stern, H., Le Calendrier de 354 (Paris, 1953).Google Scholar

¹¹⁹ See section 3. The episode is entitled: Περὶ τοῦ ἀπελπισθέντος ἐπάρχου. This verb appeared in the inscription below the south panel of the left pier, cf. Uspenskij (1909).

¹²⁰ Cf. the second edition (1967), cited in n. 105, p. 73, and p. 98 n. 17.

¹²¹ nn. 74 and 90. He does not discuss at length the ‘Apotheosis’ of S. Demetrios panel (illustrated by Hoddinott (1963) pl. 4, facing p. 142. The stylized clouds and angel emerging from the sky could be forerunners as easily as ‘echoes’ of S. Vitale, since both features occur in S. Maria Maggiore in Rome.

¹²² pp. 23–4.

¹²³ Cf. Kostof, S., The Orthodox Baptistery of Ravenna (New Haven and London, 1965), esp. 124–36.Google Scholar A case can be made for attributing the Orthodox Baptistery and Mausoleum to the same workshop, and for deriving its traditions from Rome. The analysis of mid-fifth-century art in Ravenna by Kostof does suggest the need for a revision of ideas on the growth of ‘Justinianic’ art.

¹²⁴ Cf. Stylianou, A. and Stylianou, J. A., The Painted Churches of Cyprus (Cyprus, 1964) pp. 23–7.Google Scholar

¹²⁵ Xyngopoulos, A., ‘Αἱ δύο ψηθιδωταì προσωπογραθίαι τῆς Νικοπόλεως’, ADelt xxii (1967) part 1, 15–20.Google Scholar

¹²⁶ Cf. Kitzinger, E., ‘Some Reflections on Portraiture in Byzantine Art’, Зборнuк Раәоба Вuзанмопошко Инсмuмума viii/i (1963) 185–93.Google Scholar

¹²⁷ Cf. Grabar (1957) 21–4.

¹²⁸ Laurent, V., ‘Une homélie inédite de l'archevêque de Thessalonique, Léon le Philosophe, sur l'Annonciation (25 mars 842), Mélanges Tisserant (Vatican, 1964) ii 281–302.Google Scholar The homily is attributed to an Archbishop Leo of Thessaloniki, whom Laurent identifies as the famous Iconoclast. The emphasis on images would seem inappropriate to this candidate, and so the twelfth century (c. 1133–45) or a pre-iconoclastic candidate must be considered.

¹²⁹ Diehl (1918) 57.

¹³⁰ The Nicopolis mosaic fragment may be an independent witness of such a general current of development in the Greek East; cf. Krautheimer, R., Early Christian and Byzantine Architecture (London, 1965) 98–9Google Scholar, who emphasizes the Italian affiliations of the structure of Basilica B.