^{page 90 note 1} L'Ecole grecque dans l'architecture byzantine (Paris, 1916). My debt to this invaluable study will be apparent throughout the following pages. I should also like to record my gratitude to Mme. Soteriou and her husband the Director of the Byzantine Museum at Athens for advice and assistance in collecting my material, to Mr. Schultz Weir for permission to reproduce figs. 1 and 2 on pl. 31 from his unpublished drawings and to Mr.H. M. Casson for his photograph of Hagia Mone (pl. 28, 4).

^{page 91 note 1} In the Εὑρετήριον τῶν Μεσαιωνικῶν Μνημείων, Ι,Ἀθηνῶν (Athens, 1927 and 1929), the Greek Archaeological Service has published an admirable survey of the Athenian churches which includes full bibliographical notes and a complete series of plans. For elevations and architectural details generally it is unfortunately dependent on previous publications which are incomplete and often inaccurate; consequently, some of the points which I have observed on the churches themselves and which are here recorded for the first time must await confirmation in the first complete publication of the buildings.

^{page 91 note 2} The two Synaxaria were published by Vasilievski, Pravoslavni Palestinski Sbornik, VI. fasc. 2. Cf. K. Konstantopoulou, Ἡ Μονὴ Ὁσίου Μελετίου (Δελτ. Χριστ. Ἀρχ Ἑτ.), 49 ff. and Millet, Le monastère de Daphni, 18 ff.

^{page 92 note 1} Plan in Orlandos, Μοναστηριακὴ Ἀρχιτεκτονική (Athens, 1927), fig. 7; photograph from South-East published by Konstantopoulou, op. cit. 58, fig. 2.

^{page 92 note 2} Discussed by Lampros, Νέος Ἑλληνομνήμων, VI, 382 ff.; cf. Orlandos, Ἡ Μονὴ Βαρνάκοβας(Athens, 1922), 7 ff. It comprises a series of notes in Greek and Italian relating to the early history of the monastery. The bilingual entries do not exactly correspond and it is evident that both Greek and Italian versions are based on a third, much older.

^{page 92 note 3} C.I.G. 8730; published with corrections by Orlandos, op. cit. 7.

^{page 93 note 1} Diehl, , L'Église et les mosaiques du couvent de Saint-Luc (Paris, 1887), 6–7.Google Scholar

^{page 93 note 2} Ibid. 8–9.

^{page 93 note 3} Schultz, and Barnsley, , The Monastery of St. Luke of Stiris in Phocis (London, 1901; henceforth S. and B.), 22.Google Scholar Their interpretation, inso far as it affects the dating of the churches, has since been generally accepted. Cf. Wulff, , Das Katholikon von Hosios Lukas (Die Baukunst, ser. II, 11, 1903), 3–4Google Scholar; id., Altchristliche und Byzantinische Kunst (Berlin-Neubabelsberg, 1914), II, 461; Diehl, , Manuel d'art byzantin ¹ (Paris, 1910), 435Google Scholar; ibid.² (1925–6), 463.

^{page 93 note 4} Demus, and Diez, , Byzantine Mosaics in Greece (Harvard, 1931), 108.Google Scholar Meletios granted remission of sins with the permission and authority of the Patriarch Nicholas III (1081– I I I I ) ; v. supra p. 91².

^{page 93 note 5} The two sarcophagi in the crypt which tradition associates with the Emperor Romanus II and his wife do not help to date the church. The Emperor in question is known to have been buried at Constantinople in the church of the Holy Apostles. Again the sarcophagi seem to be of different dates in view of their different decoration and they need not necessarily be connected with the founders. Cf. Demus and Diez, 107.

^{page 94 note 1} Millet, Le Monastère de Daphni, 17.

^{page 94 note 2} Ibid. 18 ff

^{page 94 note 3} Demus and Diez, 110.

^{page 94 note 4} Metropolitan of Athens 1182–1204. Subsequently exiled in Keos (d. 1220).

^{page 94 note 5} Lampros, , Μιχαήλ Ἀκομινάτομ τά σωƷόμενα (Athens, 1880), II, 247Google Scholar and 628 ff.

^{page 94 note 6} Ibid. II, 311. Both letters were written from Keos, that to Ioannes Kynegos in 1207, the other two years later.

^{page 94 note 7} Struck, , Vier Byzantinische Kirchen der Argolis (A.M. 1909, 230)Google Scholar.

^{page 94 note 8} Miklosich, u. Müller, , Acta. Dipl., V, 178.Google Scholar The relevant passages are discussed at length by Struck (op. cit. 230 ff.).

^{page 95 note 1} Op. cit. 233.

^{page 95 note 2} Ibid. 236.

^{page 95 note 3} Miklosich u. Müller, V, 253.

^{page 95 note 4} Νέος Ἑλληνομνήμων, XIII (1916), 363.

^{page 96 note 1} Ἐφημ. 1853, 937, no. 1589; C.I.G. IV, 9336; Antonin, , O Drevnikh Kristianskikh Nadpis'ach u. Afinakh (St. Petersburg, 1874), 4, no. 4Google Scholar, facsimile pl. iii, no. 4; cf. Millet, L'École grecque., 7¹.

^{page 96 note 2} Ἐφημ. 1854, 1214, no. 2448; Εὑρετήριον, I, fig. 66.

^{page 96 note 3} Op. cit. 7².

^{page 96 note 4} C.I.G. 8803; Εὑρετήριον, I, fig. 65.

^{page 96 note 5} Εὑρετήριον, I, 68, 73.

^{page 96 note 6} Couchaud, , Choix d'é;glises byz. en Grèce (Paris, 1842), pl. 9. I.Google Scholar

^{page 96 note 7} Ibid. pl. 10, 2; Castellazzi, , Ricordi di Architetlura Orientale (Venice, 1871), pl. 5.Google Scholar

^{page 97 note 1} Some opinions : Mme. Soteriou : eleventh century ('Εφημ 1931, 137); Xyngopoulos: probably twelfth century (Εὑρετήριον, I, 74).

^{page 97 note 2} The relevant passage in the inscription is :

τὸν πρὶν παλαι(ὸν ὄ)τα σου ναόν, μάρτ(υς,καὶ μικρ)ὸν καὶ πήλνον καὶ σαθρὸν λίαν

^{page 97 note 3} For additional note v. infra, p. 130.

^{page 97 note 4} Struck, op. cit. 229.

^{page 98 note 1} C.I.G. 8752; Kampouroglou, , Ἡ Στήλη τοῦ Νεοφύτου in Μελέται καί Ἔπευναα -τὰ Ἀττικά (Athens, 1923); 130 ff.Google Scholar; Koukoules, , Τὸ Κιόνιον τοῦ Νεοφύτου, Ἐπετηρὶς ‘Ετ. ΒυƷ. Σπουδῶν Η’ (1931), 148 ff.Google Scholar

^{page 98 note 2} Soteriou, , Guide du musée byzantin d'Athènes (Athens, 1932), 67, fig. 40Google Scholar; Koukoules, op. cit. 150.

^{page 98 note 3} Ἡ Μονὴ τοῦ Κυνηγοῦτῶν φιλοσόφων (Δελτ. Ἱστ. Ἐθν. Ἑτ., III, 1889), 121. The three inscriptions have attracted the attention of scholars to the church, which has thus attained a position of importance it does not deserve, for it is small in scale, illconstructed and of meagre architectural interest.

^{page 98 note 4} Ibid. 122.

^{page 98 note 5} Published by Soteriou, Ἐφημ. 1924, 22, fig. 38.

^{page 98 note 6} For additional note v. infra, p. 130.

^{page 99 note 1} Op. cit. 122¹. This is Lampardopoulos' foundation at Demitsane. A sigillion of the Patriarch Polyeuctos dated 964 (Ἐφημ. 1854, 1216; Miklosich u. Müller, V. 250) mentions the monastery as recently built; it is now deserted and in ruins. It is unlikely that the small church which Zachos has published (Δελτ. VIII (1923), 59 ff.) belongs to the original foundation. Its dome has features which are not found elsewhere before the twelfth century and it possibly belongs to an even later date.

^{page 99 note 2} Strzygowski à propos of Kaisariane ('Εφημ. 1902, 62) claimed that difference of style is not interpretable as the result of continually developing technique, but rather as the product of different conditions; the resources of the builder are one controlling factor, the individual taste of the architect another. This view has been disregarded both by later historians of byzantine architecture, and by Strzygowski himself in other fields of research.

^{page 99 note 3} See especially Choisy's chapter ‘L'Art byzantin et les classes ouvrières au bas empire,’ op. cit. 169 ff.

^{page 100 note 1} Frazer, , Pausamos, X, 35, 8Google Scholar; S. and B., 23.

^{page 100 note 2} Dating on structural grounds is on this account impossible. The latest of the marble fragments built into the church are of twelfth-century style; a detailed study of these, the most satisfactory evidence for the dating of the building, is not within the scope of the present article and the church is therefore not included in the chronology.

^{page 101 note 1} S. and B., 69. A third example is the Vlacherna church near Mezappo in Mani.

^{page 101 note 2} Struck, op. cit. pl. X: Millet, op. cit. fig. 129.

^{page 101 note 3} E.g. Panagia near Vatheia, Euboea: ibid. fig. 22 ; Lampakis, , Mémoire sur les antiquités chrétiennes de la Grèce (Athens, 1902), fig. 85.Google Scholar The church has many features of the architecture which developed under the despots of Epirus, and dates from the second half of the thirteenth century at the earliest.

^{page 101 note 4} E.g. The basilican church and Martyrium at Corinth; cf. A.J.A., 1929, 348.

^{page 101 note 5} S. and B., pl. 6.

^{page 101 note 6} Choisy, op. cit. 12. Cf. the interior and exterior views of Christianou: Millet, op. cit. figs. 57–58.

^{page 101 note 7} Cf. Kaisariane, where the cloisonné system is used only on the east end and the north and south gables (Strzygowski, loc. cit. 59, fig. 4 b); Ioannes Kynegos, rubble through out save in the apse and dome (pl. 28, 2) ; Omorphe Ekklesia, a plinth of rubble all round the church, elsewhere cloisonné (Orlandos, , Ἡ Ὄμορφη Ἐκκλησια, Athens, 1921, figs. 6, 7)Google Scholar.

^{page 102 note 1} This is visible on the south façade, shown though not clearly on two published photographs: Millet, op. cit. fig. 75; Struck, , Griechenland I, Athen, u. Attika (Vienna-Leipzig, 1911), fig. 73Google Scholar; cf. (Schultz, Weir) The Athenian Churches (The Builder 57 (1889), 379 ff.), fig. 4.Google Scholar

^{page 102 note 2} North façade: Millet, Daphni, pl. VI, 1; south façade before restoration: Schlum-berger, , L'Epopée byzantine, III, 549.Google Scholar This feature has been incorrectly rendered by the draughtsman of Millet's plates IV and V.

^{page 102 note 3} Millet, op. cit. fig. 69; Struck, Vier Byz. Kirchen der Argolis, pl. XI, 4.

^{page 102 note 4} Ibid. pl. IX, 4 and 5.

^{page 102 note 5} Lampakis, Mémoire, fig. 42; Struck, op. cit. fig. 8.

^{page 102 note 6} H. Georgios, Alaï Bey, near Skala in Laconia. Two photographs have been published: by Millet, L'École grecque, fig. 128 (from south); by Soteriou, Mme., Λακωνικά I (1932), 45Google Scholar, fig. 8 (from north-east).

^{page 102 note 7} Cf. Millet's hypothesis, op. cit. 228.

^{page 103 note 1} Op. cit. fig. 15.

^{page 103 note 2} L' École grecque, 254.

^{page 103 note 3} Χριστιανικὴ Ἀρχαιολογία τῆς Μονῆς Δαφνίου, Athens, 1889 (henceforth: Lampakis, Daphni), 84 ff.; Mémoire, 21 ff.

^{page 103 note 4} Comptes Rendus du Congrès International d'Archéologie (Athens, 1905), 312. Strzygowski elaborated his thesis in Amida, 372 ff.

^{page 104 note 1} The term is used to cover all figures which derive ultimately from the early Arabic alphabet, whether they are interpretable as specific Arabic characters or not.

^{page 104 note 2} Notably a series of Cufic inscriptions, cut in Pentelic and Hymettan marbles which epigraphists date within the limits of the ninth and eleventh centuries. The most important of these, from the Asclepieion, records the dedication of a mosque. See Soteriou, , Ἀραβικά Λείφανα ἐν (Πρακτ. Ἀκαδ. Ἀθ., 1929), 266 ff.Google Scholar

^{page 104 note 3} Cf. A(ntonin), Khristianskiã Dreṿnosti Gretsii, i, O Drevnikh Tserkvakh Goroda Afin (Zhurnal Ministerstva Marodnago Prosv'eshtcheniā, LXXXI (1854), ii, 31 ff.), fig. 13, line 2.

^{page 105 note 1} S. and B., pl. II.

^{page 105 note 2} H. Apostoloi: pl. 30, no. 19; Theotokos: S. and B., pl. 11, line 6, No. 7; Panagia Lykodemou: Lampakis, Daphni, 84, no. 6; Mémoire, fig. 69. It will be seen that the motif is formed by the opposition of two identical characters; this symmetrical reduplication is typical of the Cufic inscripition which served as models for this style of ornament.

^{page 105 note 3} S. and B., pl. 11, line 1.

^{page 105 note 4} Ibid. line 3.

^{page 105 note 5} Ibid. line 2.

^{page 106 note 1} Lampakis, Daphni, 85, nos. 9 and 10; Mémoire, figs. 72 and 73.

^{page 106 note 2} Lampakis, , Daphni, 86, nos. 17 and 18Google Scholar; Mémoire, figs. 74 and 75. Lampakis' figures are inaccurate in some details; the same champlevé frieze has been reproduced, but no more correctly, in Castellazzi, op. cit. pl. 41, 2.

^{page 106 note 3} Soteriou, loc. cit. fig. 41; Εὑρετήριον, I, fig. 64. The Cufie panels have lately been deciphered; the readings have not yet been published but, as might be expected, their message is Mohammedan rather than Christian in spirit. Here is another convincing proof of the presence of Arab craftsmen in Athens.

^{page 106 note 4} Castellazzi, op. cit. pl. 42, 3; Millet, op. cit. fig. 105; Lampakis, Mémoire, fig. 88.

^{page 106 note 5} Op. cit. pl. II, 1.

^{page 106 note 6} Cf. the marble string-courses at Daphni: Millet, Daphni, fig. 36.

^{page 107 note 1} In the restoration of this church during the last century the western part was refaced without patterns. The original masonry is visible only at the east and in the lower courses of the lateral façades. A(ntonin) evidently saw the church before it was restored, for in illustrating the north Narthex door he shows a pattern in practically every joint of the surrounding masonry where now there is none (op. cit. fig. 15).

^{page 107 note 2} Near the Monument of Lysikrates. The plan is of thetetrastyle type (Εὑρετήριον, I, fig. 108). The east end has been plastered and a modern aisle encloses the other three façades so that only in the gables is the masonry exposed.

^{page 107 note 3} The masonry of the church ends abruptly a little to the east of the South Porch; that of the Exo-narthex does not bond with it.

^{page 108 note 1} The ill-adjustment of its masonry to that of the Exo-narthex is plainly visible on the building itself and distinguishable on the published photographs: Struck, Athen u. Attika, fig. 173; Millet, L'École Grecque, fig. 75: Εὑρετήριον, I, fig. 55.

^{page 108 note 2} Millet, op. cit. 125; Xyngopoulos, Εὑρετήριον, I, 71.

^{page 108 note 3} At the east end of the village towards the hamlet Koroni. The plan has been published by Villard, Monneret de (Inedita byzantina (Monitore Tecnico, XVIII 1912), 431Google Scholar, fig. 4, whence Millet, op. cit. fig. 139) with the title Koroni. Millet by discussing and indexing the church under two names—Ligourio and Koroni—gives the impression that there are two churches where in fact there is only one.

^{page 108 note 4} Millet, op. cit. fig. 115, c.

^{page 109 note 1} West façade, actually the inner wall of the Narthex which is a later addition and whose north wall partly obscures the frieze.

^{page 109 note 2} Millet, op. cit. fig. 115, a.

^{page 109 note 3} Published by Traquair, , Frankish Architecture in Greece (Journal Royal Inst. of British Architects, XXXI, 1923–1924), 80 ff.Google Scholar For the brickwork fig. 31; cf. Millet, op. cit. fig. 114, c.

^{page 110 note 1} École grecque, 257.

^{page 110 note 2} Lampakis, Mémoire, fig. 80.

^{page 110 note 3} S. and B., pl. 11, line 6, no. 2. It seems to be a combination of κ and ω and differs from its neighbours in the simple un-cut units of which it is composed.

^{page 110 note 4} Millet, op. cit. 253¹.

^{page 111 note 1} The church has features which connect it very closely with the twelfth-century Argive group (cf. infra, p. 127). Millet suggests that these are later additions (op. cit. 210⁴), but a careful inspection has revealed to me nothing to support this hypothesis. The whole of the church proper is without question of one period. The Exo-narthex alone is later, as has been indicated on both the published plans: Monneret de Villard, op. cit. 432, fig. 7, whence Millet, op. cit. fig. 141; Traquair, op. cit. fig. 29.

^{page 111 note 2} Lampakis, Daphni, 87, no. 20; Mèmoire, fig. 76.

^{page 111 note 3} Daphni, 87, no. 21; Mémoire, fig. 77.

^{page 111 note 4} Struck, Vier Byz. Kirchen der Argolis, 207, fig. 2, d.

^{page 111 note 5} Daphni, 85, nos. 11 and 12.

^{page 111 note 6} I, 16 and XXII, 16.

^{page 111 note 7} Millet, op. cit. fig. 118.

^{page 112 note 1} A comparison of the patterns from the Katholikon of Hosios Loukas and the Amphissa church is sufficient to illustrate the progress in this last respect.

^{page 112 note 2} On the north façade between the arch which marks the transept and the east end, out of a total of 59 vertical joints I counted 31 made without tiles.

^{page 112 note 3} Millet, op. cit. 228.

^{page 113 note 1} Choisy, op. cit. 12.

^{page 113 note 2} Orlandos, Αἱ Βλαχέρναι τῆς Ἠλείας, Ἐφημ. 1923, 22, fig. 36.

^{page 113 note 3} Millet, op. cit. fig. 119, e.

^{page 113 note 4} Traquair, loc. cit. fig. 26.

^{page 113 note 5} Three tiles all of the same pattern; two of these are used in horizontal courses.

^{page 113 note 6} H. Petros; cf. Orlandos, Ναοὶ τῶν καλυβίων Κουβαρᾶ (Ἀθηνά, XXXV), 183, fig. 13.

^{page 114 note 1} Millet, op. cit. fig. 115, a; cf. pl. 30, 56.

^{page 114 note 2} A(ntonin) figures a similar tile among the Lykodemou patterns (op. cit. fig. 13, line 2, 2) and in position by the north Narthex door (fig. 15).

^{page 114 note 3} Millet, op. cit. fig. 115, c; incorrectly illustrated by Struck, loc. cit. pl. IX, 7.

^{page 114 note 4} The Omorphe examples are actually at the south-east corner of the Parekklesion (Orlandos, Ἡ Ὄμορφη Ἐκκλησία, fig. 5) and do not necessarily prove a twelfth-century date for the whole building. Orlandos dates the church proper in the eleventhcentury and maintains that the Parekklesion was a later addition. Apart from the evidenceof the cut brick he points out (p. 41) that the south gable window is in part obscured by the roof of the chapel, and secondly that the window in the south wall of the S.W. angle compartment now opens, not to the outer air, but into the Parekklesion (for plan, fig 4; cf. fig. 14). To my mind neither of these arguments is valid. In the first case the gable window is not obscured by the original chapel roof but by that which, as Orlandos admits (p. 42), is a repair dating from the Frankish occupation. With regard to the other window it need only be said that in other churches of this size and type, unencumbered with Parekklesia, windows in this position are rare and it should be noted that at the corresponding point on the north side of the church, the wall is not pierced. The window has plainly been introduced to provide additional communication between chapel and church. Further,the masonry of the east end is uniform throughout and there is no break in its continuity at the point of junction (fig. 7). The Parekklesion window is arched in stone and identical to that of the Prothesis, and like the Bema window it has a bowl built in above it. These considerations combine to prove that the church proper is contemporary with the chapel and therefore dates, on the evidence of the cut tiles, from the second half of the twelfth century.

^{page 115 note 1} Orlandos, , Μεσαιωνικὰ μνημεἴα Ὀρωποῦ καὶ Συκαμίνου (Δελτ. χριστ. ἈρχἙτ. IV, 1927), 25 ff.Google Scholar, fig. 16.

^{page 115 note 2} Ibid. 43, fig. 17.

^{page 116 note 1} Orlandos, , Ἐφημ 1923, 22Google Scholar, fig. 36.

^{page 116 note 2} Op. cit. 268 ff.

^{page 116 note 3} S. and B., fig. 14.

^{page 116 note 4} Ibid. pis. 9 and 8, whence Wulff, Hosios Lukas, fig. 4; cf. Diehl, op. cit. 15.

^{page 116 note 5} Five in all; Rivoira, , Lombardic Architecture (London, 1910), I, fig. 262Google Scholar; Struck, Athen u. Attika, fig. 164; Strzygowski, Die Baukunst der Armenier, II, fig. 721; Εὑρετήριον, I, fig. 73.

^{page 116 note 6} S. and B., pl. 9.

^{page 116 note 7} For illustrations v. infra, p. 117¹.

^{page 116 note 8} Millet, Daphni,pl. V.

^{page 117 note 1} Cf. various views of the east end: Couchaud, op.cit. pl. 4; Castellazzi, op. cit. pl. 4; Rivoira, op. cit. I, fig. 36, Le Origini dell' Architettura Lombarda² (Milan, 1908), fig. 27; Diehl, , Manuel ¹ (1910), fig. 212Google Scholar; ibid.² (1925), fig. 220; Struck, op. cit. fig. 168; Εὑρετήριον I, fig. 61.

^{page 117 note 2} Millet, Daphni, pl. V.

^{page 117 note 3} S. and B., pl. 9.

^{page 117 note 4} Millet, op. cit. fig. 27.

^{page 118 note 1} Millet, op. cit. fig. 118.

^{page 118 note 2} Lampakis phot. 1846. This fine church was demolished and replaced by a wooden-roofed chapel in 1914. The large and equally spaced windows recall Daphni; the masonry, though like Daphni in having no brick patterns, is less regular, which suggests a somewhat earlier date. Lampakis' photograph has been published by Mme. Sotiriou (Ἐφημ. 1931, 138, fig. 15), who proposes another limit for the dating of the church: ‘not earlier than the beginning of the eleventh century.’

^{page 118 note 3} Struck, Vier Byz. Kirchen der Argolis, pl. IX, 7; Millet, op. cit. fig. 115, c

^{page 119 note 1} The mortar is of inferior quality and has so disintegrated that it is impossible to discover if the two rows were connected by smaller fragments or not.

^{page 119 note 2} Struck, op. cit. pls. VII, I, IX, 6.

^{page 119 note 3} Ibid. pl. XI, 5; cf. Diehl, op. cit¹.(1910), fig. 210; ibid.² (1925), fig. 218.

^{page 119 note 4} Orlandos, Ἐφημ. 1923, 22 fig. 34; Millet, op. cit. fig. 119, c.

^{page 119 note 5} Op. cit. pl. 3, 4.

^{page 120 note 1} Millet, op. cit. fig. 118.

^{page 120 note 2} E.g. Panagia Acheiropoietos (Eski Djouma), Thessalonika; see Diehl, Le Tourneau et Saladin, Monuments chrétiens de Thessalonique, pl. IV.

^{page 120 note 3} See my drawing in Ἐφημ. 1931, 124, fig. 5.

^{page 120 note 4} S. and B., pl. 9.

^{page 120 note 5} Three-light window in the Bema, two lights elsewhere; for illustrations v.supra, p. 116⁵.

^{page 120 note 6} S. and B., pls. 9, 10.

^{page 120 note 7} Couchaud, op. cit. pl. II, I, whence Εὑρετήριον I, fig. 7g.

^{page 120 note 8} North façade (five windows, four of arcade type): Couchaud, op. cit. pl. 12, I, whence Wulff, Hosios Lukas, fig. 10, and Εὑρετήριον I, fig. 82. South façade (six windows, five of arcade type): Rivoira, , Lombardic Architecture ¹ (1910), IGoogle Scholar, fig. 279.

^{page 121 note 1} Kaisariane south gable (Strzygowski, Ἐφημ. 1902, figs, I, 42, whence Wulff, Altchr.u. Byz. Kunst II, fig. 358), Ligourio (apse window) and some minor Athenian churches. In all these the windows have two lights, never more. Note too that in certain church annexes, such as the Kapnikarea Exo-narthex, the desire for a more open treatment has resulted in a return to the broad double arcade-window (Couchaud, op. cit. pl. 14; Castellazzi, op. cit. pl. 71; for other illustrations v. supra, p. 108¹), or, as in the case of the west porch at Hosios Meletios, to the true arcade opening to the ground. But such annexes lie really outsidethe main current of church-building here examined; they are few in number and presented problems of design not met with in the churches them selves. They have thus an individual quality which contrasts sharply with the peculiar homogeneity of the church series as a whole.

^{page 121 note 2} Op. cit. 206.

^{page 122 note 1} East windows: Couchaud, op. cit. pl. II, 2; cf. my pl. 31, 4. Lateral façades: v. supra, p. 120⁸.

^{page 122 note 2} The architectonic affinity of the two churches also suggests imitation (Millet, Daphni, 53), while the use in some of the windows of the later church of a true column instead of a shaft, a feature rare elsewhere outside the Katholikon, points to the same conclusion.

^{page 122 note 3} West gable window: Castellazzi, op. cit. pl. 71; Εὑρετήριον I, fig. 55. South gable window: v. supra, p. 108¹.

^{page 122 note 4} For illustrations v. supra, p. 117¹.

^{page 122 note 5} Millet, Daphni, fig. 27 and pl. V.

^{page 122 note 6} Gastoune; both the Bema window (pl. 29, 3) and that in the south gable (pi. 29, 2) follow the earlier type with three equal lights.

^{page 122 note 7} The grouped window with the raised centre light survived the Frankish occupation and was still in use during the sixteenth century to judge by the examples in the Katholikon Exo-narthex at Hosios Loukas, erected in 1582 but removed during the last century: Wulff, Hosios Lukas, pl. II, 2; S. and B., fig. 11.

^{page 123 note 1} Op. cit. pl. 3, I; whence, Εὑρετήριον I, fig. 9.3.

^{page 123 note 2} S. and B., fig. 45, A.

^{page 123 note 3} Ibid. 69¹.

^{page 123 note 4} Millet, op. cit. fig. 118.

^{page 123 note 5} ‘Corniche de dents’; cf. Millet, op. cit. 266 ff.

^{page 123 note 6} Ibid. fig. 106.

^{page 123 note 7} Struck, op. cit. 192, fig. I

^{page 123 note 8} Op. cit. 267

^{page 123 note 9} Struck, op. cit. pl. VII, I.

^{page 123 note 10} Apse and dome only; cf. pl. 28, 2.

^{page 124 note 1} E.g. a window on the north facade: S. and B., fisr. 17.

^{page 124 note 2} Ibid. pl. 7.

^{page 124 note 3} Ibid. pl. 9.

^{page 124 note 4} The east windows of this church have been walled-up and plastered.

^{page 125 note 1} This feature, clearly visible on Lampakis' negative, is barely distinguishable in the reproduction: Ἐφημ. 1931, 138. fig.15.

^{page 125 note 2} (Weir Schultz). ob. cit. 381. fis. 7.

^{page 125 note 3} Millet, op. cit. fig. 96.

^{page 125 note 4} Daphni, 84. nos. 2, 3; Mémoire, figs. 78, 79.

^{page 125 note 5} S. and B., fig. 47, a.

^{page 125 note 6} Millet, Daphni, fig. 27.

^{page 125 note 7} S. and B., fig. 47, b.

^{page 125 note 8} Strzygowski, op. cit. fig. 7b; Lampakis, Mémoire, fig. 49; Millet, op. cit. fig. 106

^{page 125 note 9} Parabemata windows: Struck, op. cit. pl. VII. I.

^{page 125 note 10} ibid. pl. VIII.

^{page 125 note 11} Orlandos, Ἡ Ὄμορφη Ἐκκλησία, fig. 11.

^{page 126 note 1} Lampakis, Daphni, 87, no. 20; Mémoire, fig. 77.

^{page 126 note 2} S. and B., fig. 47, b.

^{page 126 note 3} West gable: pl. 31, 3. South gable: Couchaud, op. cit. pl. 10, 2; Castellazzi, op. cit. pl. 5. Central apse: v. supra, p. 117¹, and add (Weir Schultz), op. cit. 381, fig. 7.

^{page 126 note 4} Merbaka, Parabemata windows (Struck, op. cit. pl. VI, 2); Omorphe Ekklesia (Orlandos, op. cit. fig. 7, II).

^{page 126 note 5} Millet, op. cit. fig. 118.

^{page 126 note 6} Cf. Talbot Rice, Byzantine Glazed Pottery, 16.

^{page 126 note 7} op.cit. 207 ff.

^{page 127 note 1} Op. cit. fig. 104; S. and B., pl. 8, whenceWulff, op. cit. fig. 4.

^{page 127 note 2} S. and B., pl. 58.

^{page 127 note 3} Ἐφηµ 1931, 138, fig. 15; cf. supra, p. 118².

^{page 127 note 4} South gable: Struck, op. cit. 206, fig. 3, pi. X, 4; Lampakis, Mémoire, fig. 91. North gable: Struck, op. cit. pl. X, 5; cf. Diehl, Manuel ¹ (1910), fig. 211, ibid.², (1925), fig. 219.