¹ See e.g. Villes et peuplement dans I'Illyricum protobyzantin. Actes du colloque organisé par l'Ecole Française de Rome (Rome, 12-14 mai 1982) (Rome 1984)Google Scholar. We await D.J. Geagan's article on Athens from 267 to Justinian, in ANRW iii 4; and A. Frantz, The Athenian agora. Late antiquity (267-700).

² Bintliff, J. L. and Snodgrass, A. M., ‘The Cambridge/Bradford Boeotian Expedition: the first four years’, JFA xii (1985) 123–61Google Scholar; Bintliff, J. L., ‘The development of settlement in south-west Boeotia’, in La Béotie antique: Lyon—Saint-Etienne 16-20 mai 1983 (Paris 1985) 49–70Google Scholar; Bintliff, J. L. and Snodgrass, A. M., ‘Mediterranean survey and the city’, Antiquity lxii (1988) 57–71CrossRefGoogle Scholar.

³ Bradford, J.'s report, ‘Fieldwork on aerial discoveries in Attica and Rhodes. Part II. Ancient field systems on Mt. Hymettos, near Athens’, Ant J xxxvi (1956) 172–80CrossRefGoogle Scholar, on the Glifadha area, divided into building plots in 1955, exemplifies what has been lost. The meagreness and ambiguity of what remains is exemplifield by Langdon, M. K., ‘Hymettiana I’, Hesperia liv (1985) 257–70.CrossRefGoogle Scholar

⁴ E.g. Hohlfelder, R. L. (ed.), City, town and countryside in the early Byzantine era (New York 1982)Google Scholar; Leveau, P., Caesarea de Maurétanie, une ville romaine et ses campagnes (Rome 1984)Google Scholar; Osborne, R., Classical landscape with figures. The ancient Greek city and its countryside (London 1987)Google Scholar; Wickham, C. J., The mountains and the city. The Tuscan Appennines in the early Middle Ages (Oxford 1988)Google Scholar.

⁵ Paus. i 32.1-2.

⁶ Apul., Met. iv 9, 13; Luc, Demon. 1. For Albanian brigands on Parnes in the seventeenth century, see Les voyages du Sieur du Loir, contenu en plusieurs lettres écrites du Levant (Paris 1654) 305–7Google Scholar.

⁷ Ober, J., Fortress Attica. Defense of the Athenian land frontier 404-322 B.C. (Leiden 1985), esp. 87–100, 208-22Google Scholar.

⁸ Townsend, R. F., ‘The Roman rebuilding of Philon's Porch and the Telesterion at Eleusis’, Boreas x (1987) 97–106Google Scholar.

⁹ Zos. i 29.3 (perhaps here dependent on Dexippus [see below]: Armstrong, D., ‘Gallienus in Athens, 264’, ZPE lxx [1987] 240Google Scholar); Geagan, D. J., ‘Roman Athens: some aspects of life and culture I. 86 B.C.-A.D. 267’, ANRW ii 7.1 (Berlin 1979) 410, 436Google Scholar. Eleusis: Ziro, D. G., H κύρια είσοδος του ιερού της Eλευσίνας (diss. Athens Polytechnic 1986) 273–8Google Scholar.

¹⁰ Armstrong (n. 9) 235-58.

¹¹ Thompson, H. A. and Wycherley, R. E., The agora of Athens: the history, shape and uses of an ancient city center (Princeton, N.J. 1972) 208–10Google Scholar; T. E. Gregory, ‘Fortification and urban design in early Byzantine Greece’, in Hohlfelder (n. 4) 44-51. Numismatic evidence: Kroll, J. H., ‘The Eleusis hoard of Athenian imperial coins and some deposits from the Athenian agora’, Hesperia xlii (1973) 312–33CrossRefGoogle Scholar; Walker, A. S., A chronological study of the Greek imperial coinage of Athens based on the collection of Agora Excavations at Athens (diss. Pennsylvania 1980) 125–33, 138-9, 146-8Google Scholar; and below, n. 25.

¹² Dexipp. (FGrH 100) fr. 28 (tr. Millar, with adjustments).

¹³ Several historical sources allude to the attack on Athens, and there is a significant body of epigraphical material on Dexippus and his family, gathered by Millar, F., ‘P. Herennius Dexippus: the Greek world and the third-century invasions’, JRS lix (1969) 20–9Google Scholar, and including IG ii—iii².3669, which numbers Dexippus among ‘famous men, mighty in courage, in word and in counsel’. Yet only H.A., V. Gall, xiii 8 actually says that the by then aged Dexippus led the resistance. Bergman, J., De P. Herennio Dexippo et Gothorum illa in Atticam incursione quid scriptores et inscriptiones doceant (Stockholm 1897) 17–22Google Scholar, argued that I) these are poor grounds for assuming that the anonymous speech recorded by Dexippus was actually delivered by Dexippus, rather than somebody else; 2) H.A., V. Gall. xiii 8 should read, not ‘duce Dexippo’, but ‘indice/docente Dexippo’. These reasonable arguments (the second of which accords with T. D. Barnes's conclusion that the V. Gall's ‘main source’ is Dexippus himself: The sources of the Historia Augusta [Brussels 1978] 109–11Google Scholar, and cf. Armstrong [n. 9] 240-6) have left no trace in subsequent scholarship, though they are in part repeated, unwittingly, by de Ste. Croix, G. E. M., The class struggle in the ancient Greek world (London 1981) 654–5Google Scholar, who reaches the radical conclusion that the resistance itself is a fiction, incredible because ‘a man of letters… in his sixties’ is by definition incapable of such exertion, and no other city is known to have mounted such fierce resistance to barbarian invaders. But if the leader was somebody other than Dexippus, the former argument (not anyway conclusive) falls; and de Ste. Croix forgets that the expulsion of the barbarians was brought about only once the imperial fleet had arrived: see below. He also fails to explain whose the fiction was, or who was supposed to be fooled, at so short a remove in time. (Bergman further suggested that the leader of the Athenian resistance may in fact have been Cleodamus of Byzantium, who according to H.A., V. Gall, xiii 6 and Zon. xii 26, p. 605, successfully discharged Gallienus's commission to organize resistance to the invaders. But Zonaras probably associates Cleodamus with the defence of Athens only because of a confusion induced by the name of his colleague and compatriot Athenaeus, whose identity with the Panathenios who restored a tower at Athens [IG ii-iii². 5201] is more doubtful than is implied by [e.g.] Groag, E., Die römischen Reichsbeamten von Achaia bis auf Diokletian [Vienna 1939] 96–7Google Scholar.)

¹⁴ Geo. Sync. 717 implies that the Heruli reached Attica by sea. Certainly, if they threatened Aegina, and caused the building of the late Roman acropolis wall (Walter, H. [ed.], Alt-Agina i 2 [Mainz 1975] 12Google Scholar), they must have disposed of boats. But they had already been worsted in at least one naval battle with the Romans (H.A., V. Gall. xiii 7), and the Athenians will have made the reasonable assumption that, if defeated, the Heruli would retreat by land.

¹⁵ Katsimidhi:Ober (n. 7) 115, 142-4, 184; Ober, J., ‘Pottery and miscellaneous artifacts from fortified sites in northern and western Attica’, Hesperia lvi (1987) 203 (plan), pl. 26Google Scholar (photographs); M. H. Munn, letter 25.5.88, reporting a fourth- to sixth-century cookingpot and combed amphora. Phyle: Ober (n. 7) 116-17, 145-7, 185-6; Adam, J.-P., L'architecture militaire grecque (Paris 1982) 206–7 (plan)Google Scholar; Schoder, R. V., Ancient Greece from the air (London 1974) 74 (photograph)Google Scholar.

¹⁶ Phot. Bibl. 82.64a; Stein, F. J., Dexippus et Herodianus rerum scriptores quatenus Thucydidem secuti sint (diss. Bonn 1957), esp. 59–60, 70-1Google Scholar.

¹⁷ X. HG ii 4.15—16.

¹⁸ Our tendency to overestimate the significance of fortifications, and to underestimate the role of natural defences, especially forests, which are only now beginning to become the object of historical research, is well illustrated by Cousinéry, E. M., Voyage dans la Macédoine (Paris 1831) ii 142–3Google Scholar.

¹⁹ E.g. Lampros, S. P., ‘‘lστορικά μελετήματα (Athens 1884) 92Google Scholar; Loucas, I. K., Φλύα, ∑υμβολή στην μελέτη της ιστορίας του αρχαίου χαλανδρίου (Chalandri 1986) 29.Google Scholar On Decelea, see Ober (n. 7) 141-2.

²⁰ The identification is at least as old as Kahrstedt, U., Das wirtschaftliche Gesicht Griechenlands in der Kaiserzeit (Bern 1954) 50 and n. 2Google Scholar. Note also Finlay, G., Greece under the Romans² (Edinburgh 1857) 111Google Scholar: ‘a strong position in the Olive Grove’. On the road, see Pritchett, W. K., Studies in ancient Greek topography, Part III (Roads) (Berkeley, Ca. 1980) 197–237Google Scholar.

²¹ Nezis, N., Tα βουνά της Aττικής (Athens 1983) 11–12Google Scholar; Kirsten, E., ‘Aigaleos’, Kl. Pauly i 156–7Google Scholar; Ober (n. 7) 148-9 (watchtowers; M. H. Munn, letter 25.5.88, now reports classical and Hellenistic rooftiles and coarse amphora sherds, and a few late Roman lamp fragments, from the ‘Karydallos Tower’), 188 (Sacred Way).

²² Letter, 2.88.

²³ Zon. xii 26, p. 605; Geo. Sync. 717 ('Aθηναῖοι κατά τινας στενς δυσχωρίας ένεδρεύσαντες αύτούς πλείστους άνείλον); H.A., V. Gall. xiii 8.

²⁴ O. W. Reinmuth, ‘Ephebia’, Kl. Pauly ii 290-1, surveys the number of ephebes mentioned in the later Roman lists. Spawforth, A. J. and Walker, S., ‘The world of the Panhellenion I. Athens and Eleusis’, JRS lxxv (1985) 98Google Scholar, speculate that an ephebic militia may, until 267, have been housed in the Library of Hadrian. On disputed evidence for an ephebic celebration of epineikia after the repulse of the Heruli, see Follet, S., Athènes au II^e et au III^e siècle. Etudes chronologiques et prosopographiques (Paris 1976) 325Google Scholar. Kapetanopoulos, E., ‘Some remarks on Athens of about 270’, AAA xvi (1983) 51–7Google Scholar, argues for the continuation of the ephebia after 267/8; but cf. S. Follet, Bull. Epigr. (1987) 590.

²⁵ Note also the coin-hoard found in an unknown cave ‘in the mountainous area between Porto Raphti and Brauron’, and probably deposited during the Herulian invasion: A. Walker, ‘A hoard of Athenian imperial bronzes of the third century A.D. from eastern Attica’, Coin hoards iii (1977) 40-9. The group of about 50 Athenian imperial bronzes found at the Cave of Pan on Parnes (K. Romaios, ‘Eύρήματα άνασκαφῆς τοῦ ὲπί τῆς Πάρνηθος ἄντρου’ AE [1906] 98) may also have been a hoard left in 267.

²⁶ Cf. Bowersock, G. W., Roman Arabia (Cambridge, Mass. 1983) 103–4Google Scholar.

²⁷ An inscription published by Raubitschek, A. E., ‘Iamblichos at Athens’, Hesperia xxxiii (1964) 63–8CrossRefGoogle Scholar; cf. Bull. Epigr. (1965) 155. Zos. iv 18 denies Attica was damaged by the earthquakes of 375; but see below, on his prejudice.

²⁸ Zos. v 5-6, with Paschoud's notes. But the stratigraphical evidence is controversial: Ziro (n. 9) 285; Archaeological reports for 1985-6, 12-13, 1986-7 7-8.

²⁹ Notes 27-8 above.

³⁰ Thompson and Wycherley (n. 11) 210-14.

³¹ Fowden, G.,‘The pagan holy man in late antique society’, JHS cii (1982) 43–5, 52, 53Google Scholar.

³² Harrison, E. B., The Athenian agora 1: Portrait sculpture (Princeton, N.J. 1953) 92Google Scholar; Koch, G. and Sichtermann, H., Römische Sarkophage (Munich 1982) 457–60Google Scholar; below, n. 55.

³³ Langdon, M. K., ‘Hymettiana II: An ancient quarry on Mt. Hymettos’, AJA xcii (1988) 83Google Scholar.

³⁴ Syn., ep. 136.

³⁵ Above, n. 2. Similar findings are now reported from the Skourta Plain, between Attica and Boeotia (Munn, M. H. and Munn, M. L. Zimmerman, ‘The Stanford Skourta Plain Project—1985 season’, Teiresias xvi (1986), appendix 9–10Google Scholar); from Keos, the closest of the Cyclades to the mainland (Mantzourani, E., Cherry, J. F., Davis, J. L., ‘Aρχαιολογική έρευνα επιφάνειας στη νήσο Kέα’ Παρουσία, 4 [1986] 198–9Google Scholar, corroborated by the results of the Athens University Keos Surface Survey, Communicated to me by Lina Mendoni); and from the S. Argolid (Andel, T. H. van, Runnels, C. N., Pope, K. O., ‘Five thousand years of land use and abuse in the Southern Argolid, Greece’, Hesperia lv [1986] 103–28, esp. 120-2, 125CrossRefGoogle Scholar; Andel, T. H. van, Runnels, C. N., Beyond the Acropolis. A rural Greek past [Stanford, Ca. 1987] 113–17, 171Google Scholar, and ‘The evolution of settlement in the Southern Argolid, Greece: an economic explanation’, Hesperia lvi [1987] 303–34, esp. 317-20Google Scholar). Perhaps the accumulation of survey evidence will eventually vindicate Whittacker, C. R.'s thoroughgoing scepticism about ‘Agri deserti’, in Finley, M. I. (ed.), Studies in Roman property (Cambridge 1976) 137–65CrossRefGoogle Scholar. In the meantime Osborne, R., ‘Buildings and residence on the land in classical and Hellenistic Greece: the contribution of epigraphy’, ABSA lxxx (1985) 119–28Google Scholar, and (n. 4) 69-70, while accepting the correlation between presence of rural structures and intensity of land use, warns against the automatic assumption that such structures were inhabited or provide an index of population-levels. In the case of Boeotia, though, the absence of rural structures in the Hellenistic and earlier Roman periods is matched by the shrinkage or even desertion of the urban sites surveyed (Askra, Haliartos, Thespiae): letter from A. M. Snodgrass, 27.10.87; Bintliffand Snodgrass, Antiquity lxii (1988) 60–8Google Scholar.

³⁶ Eliot, C. W. J., Coastal demes of Attika (Toronto 1962) 59Google Scholar, and cf. 20, 21, 44, 58-9, 67, 114. The comparison is (presumably) with the post-classical situation.

³⁷ H. Lohmann, ‘Atcne ('Aτήνη), eine attische Landgemeinde klassischer Zeit’, Hetlenika Jb. (1983) 99 and ‘Landleben im klassischen Attika’, Ruhr-Univ. Bochum, Jb. (1985) 71-96; Lauter, H., Der Kultplatz auf dem Turkovuni (Berlin 1985) 148–9Google Scholar.

³⁸ Letter from J. Ober, 2.88: ‘There appears to be pretty intensive occupation in the fourth/fifth AD, some reoccupied sites, some newly built’. Cf Muller, A., ‘Megarika’, BCH cvi (1982) 394–403Google Scholar.

³⁹ Wickens, J. M., The archaeology and history of cave use in Attica, Greece from prehistoric through late Roman times (diss. Indiana 1986), esp. i 204–27 and table 4Google Scholar, discusses 28 caves definitely or probably used in the late Roman period, and connects this observation with the results of the Boeotia survey.

⁴⁰ Lauter (n. 37) 149; Lauter, H. and Lauter-Bufe, H., ‘Ein attisches Höhenheiligtum bei Varkiza’, in Festschrift zum 60. Geburtstag von Werner Böser (Karlsruhe 1986) 304–5Google Scholar.

⁴¹ Langdon, M. K., A sanctuary of Zeus on Mount Hymettos (Princeton, N.J. 1976), esp. 7-8, 73-4, 76, 94–5Google Scholar. It may be that the presumed site of the altar of Zeus the Rain-God on Parnes had a similar history: ibid. 100-1.

⁴² Lauter (n. 37) 156.

⁴³ Paul. Sil., Soph. 679-80 (a purely literary touch?); Day, J., An economic history of Athens under Roman domination (New York 1942) 268Google Scholar; Butcher, S. A., ‘Late Roman lamps from a mine gallery at Thorikos’, in Spitaels, P. (ed.), Studies in South Attica i (Gent 1982) 137–43Google Scholar; cf. v. Andel, Runnels, Beyond the Acropolis (n. 35) 116-17, connecting the economic upswing in the S. Argolid with loss of supplies from N. Africa, Spain, etc. The late Roman re-occupation of hill/mountain sites and caves in the West too (Johnson, S., Late Roman fortifications [London 1983] 226–44, 249Google Scholar; Conges, E. el al., ‘Un dépotoir de la fin de l'antiquité dans la grotte de La Fourbine, Saint-Martin-de-Crau (B.-du-Rh.)’, RAN xvi [1983] 359Google Scholar; Premiers temps chrétiens en Gaule méridionale. Antiquité tardive et haul moyen age III^eme-VIII^eme siècles [Lyon 1986] 109–12, 154-63Google Scholar—these two references courtesy of Caroline Nicholson; Ciglenečki, S., Höhenbefestigungen aus der Zeit vom 3. bis 6. Jh. im Ostalpenraum [Ljubljana 1987]Google Scholar) is a further hint that the phenomena documented in Greece may have a wider context.

⁴⁴ Ober (n. 15) 201 (Hymettus), 204 (Beletsi), 223 (Kantili; on a similar site further along the same route see van de Maele, S., ‘La route antique de Megare à Thèbes par le défilé du Kandili’, BCH cxi [1987] 201, 203), 224—5 (Kerata)Google Scholar; Munn, M. H., Studies on the territorial defenses of fourth-century Athens (diss. Pennsylvania 1983) 236, 250-1Google Scholar (Dema tower), and letter of 25.5.88 (Katsimidhi, Korydallos tower, Plakoto-Palaiokastra).

⁴⁵ Ober (n. 7) 162-3, (n. 15) 219-20.

⁴⁶ Ober (n. 7) 142-3.

⁴⁷ See below, n. 75.

⁴⁸ Wrede, W., ‘Phyle’, MDAI(A) xlix (1924) 153–224, esp. 200-2, 220, 223-4Google Scholar; Ober (n. 15) 206, nos 6.11-12. Säflund, G.'s assertion, ‘The dating of ancient fortifications in southern Italy and Greece’, Opuscula archaeologica i (1935) 108–10Google Scholar, that the earth fill is Hellenistic is rejected by Ober (n. 7) 146 n. 43, (n. 15) 207.

⁴⁹ See G. Dagron, ‘Les villes dans l'lllyricum protobyzantin’, in Villes et peuplement (n. 1) 7-9, esp. n. 44; also Lawrence, A. W., ‘A skeletal history of Byzantine fortification’, ABSA lxxviii (1983) 191–2Google Scholar.

⁵⁰ Wrede (n. 48) 211.

⁵¹ On the significance of the route, and of Phyle as a halt, in the Turkish period, see A. N. Skias, ‘'Aνασκαφαί παρὰ τὴν Φυλὴν” PAAH (1900) 42-3. (Work in Ottoman archives is gradually producing a picture of Athens and Attica from the fifteenth century onwards, which could be of considerable comparative value for the ancient historian. Karydis, D. N., ‘'Aθήνα–’Aττικὴ στὸν πρῶτο αℓὠνα 'Oθωμανικῆς κατοχῆς. 'H σχέση πόλης–ύπαίθρου’, in Πρακτικὰ τοῦ Διεθνοῦς ∑υμποσίου ‘ℓστορίας Nεοελληνικὴ Πόλη: 'Oθωμανικές κληρονμιές καί έλληνικό κράτος (Athens 1985) i 49–58Google Scholar, discusses unexpected evidence for demographic and economic growth in the fifteenth and sixteenth centurarchaeologica ies, and mentions the organization of the Parnes dervenochoria, designed to protect traffic on rural roads.)

⁵² Lauter and Lauter-Bufe (n. 40) 285-309.

⁵³ Letter from H. Lauter, 22.10.87.

⁵⁴ Weller, C. H. et al., ‘The cave at Vari’, AJA vii (1903) 263–349CrossRefGoogle Scholar, esp. 284-5, 335—7 (coins), 338-49 (lamps); cf. Wickens (n. 39) ii 90-121.

⁵⁵ Butcher (n. 43), especially the note by J. Binder. Three very similar Pentelic marble altars from Phlya (Chalandri), discussed by Svoronos, I. N., τό έν 'Aθηναις 'Eθνικόν Mουσεῖον i (Athens 1903) 473–83Google Scholar, and E. and Loucas, I., ‘Un autel de Rhéa-Cybèle et la Grande Déesse de Phlya’, Latomus xlv (1986) 392–404Google Scholar, imply co-ordinated pagan artistic production for cultic purposes at least as late as the 380s.

⁵⁶ Leglay, M., Saturne africain: histoire (Paris 1966) 102Google Scholar.

⁵⁷ Skias, A. N., ‘Tὸ παρὰ τὴν Φυλὴν ἄντρον τοῦ Πανός’ AE (1918) 1–28Google Scholar; Wickens (n. 39) 2.245-69, including discussion of the inscriptions, to which add L. Robert, in Gagniers, J. deset al., Laodicée du Lycos. Le nymphée: campagnes 1961-1963 (Quebec 1969) 344–5, 349-50Google Scholar.

⁵⁸ Renfrew, C., The archaeology of cult: the sanctuary Phylakopi (London 1985) 398, 407Google Scholar, on tenth and ninth century BC Crete. On cave cults in late antique Crete see Sanders, I. F., Roman Crete (Warminster 1982), esp. 40Google Scholar; Chaniotis, A., ‘Plutarchos, Praeses Insularum’, ZPE lxviii (1987) 227–31Google Scholar; and the annual reports on G. Sakellarakis's excavations at the Idaean Cave, in Tὸ 'Eργον τῆς 'Aρχαιολογικῆς 'Eταιρείας.

⁵⁹ See e.g. Porph., Antr. 5-6, 20.

⁶⁰ IG ii-iii².4831; Fowden, G., ‘Nicagoras of Athens and the Lateran obelisk’, JHS cvii (1987) 51–7, esp. n. 10CrossRefGoogle Scholar.

⁶¹ Procl., In Ti. 37ef.

⁶² Marin., Procl. 33, 36; and cf. 15, on travels in Asia.

⁶³ Dam., Isid. 94, 131 (Ep. Phot.).

⁶⁴ Riginos, A. S., Platonica. The anecdotes concerning the life and writings of Plato (Leiden 1976) 17–21Google Scholar. Oliver Nicholson reminds me that this is a topos in ancient biography.

⁶⁵ Cf. Vanderpool, E., ‘Pan in Paiania: a note on lines 407-409 of Menander's Dyskolos’, AJA lxxi (1967) 309–11, esp. n. 9CrossRefGoogle Scholar.

⁶⁶ Note that Synesius, ep. 136, writes 'Aναγυρουντόθεν, i.e. from Vari. Hymettus and the late Platonists are also associated by local antiquarian speculation that the toponym Kaisariani recalls a school supposedly established there by Proclus's teacher Syrianus: I. Gennadios, 'H Kαισαριανή (Athens 1930) 30, 51.

⁶⁷ Above, n. 7.

⁶⁸ Gregory, T., ‘The fortified cities of Byzantine Greece’, Archaeology xxxv (1982) 14–21Google Scholar; Clement, P. A., ‘Isthmian notes’, in Φίλια 'Eπη είς Γεὠργιον E. Mυλωνάν ii (Athens 1987) 381–3Google Scholar. Cherf, W. J., ‘Procopius, lime mortar C¹⁴ dating and the late Roman fortifications of Thermopylai’, AJA lxxxviii (1984) 594–8CrossRefGoogle Scholar, suggests Thermopylae should be added to the list.

⁶⁹ Millet, G., Le monastère de Daphni (Paris 1899) 3–16Google Scholar; Orlandos, A., in Kourouniotis, K. and Sotiriou, G. A., Eύρετήριον τῶν μνημείων τῆς 'Eλλάδος ι Eύρετήριον τῶν μεσαιωνικῶν μνημείων iii (Athens 1933) 217–18Google Scholar.

⁷⁰ Molisani, G., ‘Un miliare di Arcadio e Onorio nel Museo Epigrafico di Atene’, SCO xxvi (1977) 307–12Google Scholar. The erection of milestones does not automatically Harmondsimply road-repair (note the remarks of Gounaropoulou, L. and Hatzopoulos, M. B., Les milliaires de la Voie Egnatienne entre Héraclee des Lyncestes et Thessalonique [Athens 1985] 72–3Google Scholar); but in this particular case the connection seems plausible enough.

⁷¹ The military function of the original Daphni monastery is underlined by its plan. The first church at Daphni, unlike the second, had the same alignment as the walls. Forsyth, G. H., ‘The monastery of St. Catherine at Mount Sinai: the church and fortress of Justinian’, DOP xxii (1968) 6 n. 5Google Scholar, discussing the nonalignment of church and walls at Sinai, remarks that ‘in dealing with large complexes of important buildings the Greek tradition is more flexible and organic than the Roman, more concerned with the changing viewpoints and the vitality of diagonal planning (Delphi, Athenian Acropolis, etc.) than with the forensic confrontation of symmetrical balance which appealed to the military and legal mind of the Roman. In the situation at Mt. Sinai a Roman architect might have been inclined to impose a plan like that of a Roman camp by rotating the fortress rectangle counter-clockwise so as to align its main gate with the axis of the church.’ But Daphni has never been discussed from this point of view; see Krautheimer, R., Early Christian and Byzantine architecture⁴ (Harmondworth 1986) 260Google Scholar: ‘Fortified posts and garrisoned sanctuaries of the fifth and the sixth centuries apparently provided the model for major monasteries even where no military protection was needed. At Daphni in Greece… The Daphni pass could of course be turned via the Aigaleos-Parnes gap; but the exiguous late Roman finds at the Dema tower (above, n. 44) do not encourage the supposition that the Dema wall was refortified in late antiquity.

⁷² Eun., VS vii 3.1-5 (end of Eleusinian Mysteries); Paus. i 37. 6-7 (Daphni temple). In obscure language, Eunapius connects the demise of Eleusis with the spread of monasticism and the betrayal by monks of a pass—at Thermopylae.

⁷³ Ober (n. 7) 221 n. 31, (n. 15) 226.

⁷⁴ Procop., Aed. iv 2.23-6.

⁷⁵ Ibid., iv 2.2-22 (Thermopylae), 27-8 (Peloponnese). Archaeology (Travlos, I. N., Πολεοδομικη εξέλιξις τῶν 'Aθηνῶν [Athens 1960] 144–5Google Scholar) supports Procopius's vague implication that Athens was refortified (Aed. iv 2.24-5) rather than his assertion (Arc. xxvi 33-4) that no public works were undertaken there. On Plataea, see Lawrence (n. 49) 194, 195. The late masonry (including aligned joints) at Plataea (Lawrence [n. 49] pl. 13b) and Katsimidhi (Ober [n. 15] pl. 26b) is very similar; but this need not mean Katsimidhi was rebuilt under Justinian. The problems faced by incompetent masons attempting to build with old stones cut for a different purpose did not change much with the passage of the centuries.

⁷⁶ Cameron, Av., Procopius and the sixth century (Berkeley 1985) 106–10.Google Scholar