2. The literature is quite extensive. In addition to what is cited below, one should note the following works: Honigmann, Ernst, Die Ostgrenze des byzantinischen Reiches von 363 bis 1071 (Brussels 1935)Google Scholar; Poidebard, A., La trace de Rome dans le désert de Syrie, 2 vols. (Paris 1934)Google Scholar; Kennedy, Hugh, ‘The Last Century of Byzantine Syria: A Reinterpretation’, BF9–10 (1985), 141–183 Google Scholar; Gray, E.W., ‘The Roman Eastern Limes from Constantine to Justinian — Perspectives and Problems’, Proceedings of the African Classical Association 12 (1973), 24–40 Google Scholar; Whitby, M., ‘Procopius and the Development of the Roman Defences in Upper Mesopotamia’, The Defence of the Roman and Byzantine East, edd. Kennedy, H. and Freeman, P. (BAR Int. Series 297, 1986), 717–735 Google Scholar; idem, , ‘Arzanene in the Late Sixth Century’, Armies and Frontiers in Roman and Byzantine Anatolia, ed. Mitchell, S. (BAR Int. Series 156, 1983), 205–217 Google Scholar; Liebeschuetz, J.H., ‘The Defences of Syria in the Sixth Century’, Journal of the 10th International Congress of Roman Frontier Studies (Cologne 1974), 487–499.Google Scholar

3. For summaries, see Bury, John B., A History of the Later Roman Empire from the Death of the Theodosius the Great to the Death of Justinian, 2 (London 1923), 10–15, 89–100 Google Scholar; Downey, G., ‘The Persian Campaign in Syria in 540’, Speculum 28 (1953) 340–348 CrossRefGoogle Scholar. On the Persian War of Anastasius, an unvarnished reading of the chronicle of Joshua still has its merits, : The Chronicle of Joshua the Stylite, ed. tr.Wright, W. (Cambridge 1882).Google ScholarA new translation with extensive commentary is currently in progress by Trombley, F. and Watt, J.W.. On the Persian War of 572–591 Google Scholar, see Whitby, M., The Emperor Maurice and His Historian (Oxford 1988), 195–308 Google Scholar (with bibliography on previous scholarship). The justification for the present article is the need for a closer look at the epigraphy.

4. Millar, Fergus, The Roman Near East 31 BC-AD 337 (Cambridge, Mass. 1993), 159–173, 319–336.Google Scholar

5. Ammianus Marcellinus, Res Gesta 19.9.

6. Bury, LRE 2, 1–10. Cf. Rubin, Z., ‘Diplomacy and War in the Relations between Byzantium and the Saracens in the Fifth Century AD’, Defence of the Roman East, 677–695 Google Scholar; Greatrex, G., ‘Two Fifth-Century Wars between Rome and Persia’, Florilegium 12 (1993), 1–14.Google Scholar

7. The principal source for this war, the chronicle of Joshua the Stylite, contains exceptional information on the countryside of Osrhoene. Much of it seems to go back to an archive containing documents from the officium of the provincial governor of Osrhoene. See above, n. 3.

8. Procopius, Wars 2.7.1–37. The events of the siege deserve careful study as an example of the behavioural side of warfare. This includes the poor state of the fortifications of the lower city, the military mounts’ drinking dry the spring or cisterns in the akropolis, the bishop Megas’ purchase of security for the city by collecting some 288,000 solidi, and the defection of large numbers of regular soldiers to the Persians on the ground that their pay was in arrears. A perusal of the indices of Dewing’s translation of Procopius in the Loeb edition will convince the reader that there were large numbers of Persarmenians, Huns, Iranians, and Turks among the Roman even before the demographic crisis spawned by the plague of 542. Cf. Whitby, M., ‘Recruitment in Roman Armies from Justinian to Heraclius (ca.565–615)’, The Byzantine and Early Islamic Near East III: States, Resources and Armies, ed.Cameron, Averil (Princeton 1995), 92–99.Google Scholar

9. Morony, Michael, ‘Trade and Exchange: The Sasanian World to Islam’, Late Antiquity and Early Islam, Fifth Workshop: Trade and Exchange in the Late Antique and Early Islamic Near East (14–16 November 1996), prepublication manuscript, p.7.Google Scholar

10. Procopius, Wars 2.8–10. The interior of the Limestone Massif was accessible from several directions. In 540 the most obvious route for Khusau was the paved Roman military road (2nd c. A.D.) from Beroia to Antioch that ran through the Dana plain (the geographical centre of the Limestone Massif) by way of Tchalenko, Chalkis.G., Villages antiques de la Syrie du Nord 1 (Paris 1953), 44 (note 1), 82f. and Plate CLXX, 3–4; 2 (Paris 1953)Google Scholar, Plate XXXIX;cf. Plates XXXVII and XXXVIII.

11. Noted in Trombley, F., ‘Religious Transition in Sixth-Century Syria’, BF 20 (1994), 192 Google Scholar. The localities in question are conveniently represented in Meer, F. van der and Mohrmann, C., Atlas of the Early Christian World, tr. Hedlund, M.F. and Rowley, H.H. (London 1958)Google Scholar, Maps 15a-b.For a detailed topographical analysis along with the epigraphic and archaeological evidence, see: Mécérian, J., ‘Les inscriptions du Mont Admirable’, Mélanges de l’Université de Saint-Joseph, Beyrouth (=MUSJB) 38 (1962), 295–330 Google Scholar; idem, , ‘Expédition archéologique dans l’Antiochène occidentale’, MUSJB 40 (1964), 1–144.Google Scholar

12. La vie ancienne de Syméon Stylite le Jeune, ed. tr. den Ven, Paul van, 1 (Brussels 1962), 50–55.Google Scholar

13. The clearest statement on this comes from Galatia in Asia Minor in the life of St. Theodore of Sykeon, which requires a brief digression in view of the importance of this endemic phenomenon that lasted straight through the wars of the sixth century. See infra, n. 104 and Appendix.

14. In a vision Christ reputedly tells Symeon: The place was uninhabited and undeveloped for agriculture, but all this changed in the following decades. On this see Trombley, ‘Religious Transition in Sixth-Century Syria’, 182–194. Cf. Mécérian, ‘Mont Admirable’, 299 and Plate I.

15. Infra, Appendix.

16. Infra, n. 52.

17. For the site, see: Butler, H.C., Publications of the Princeton University Archaeological Expeditions to Syria in 1904–5 and 1909, Division II: Architecture, Section B: Northern Syria (Leiden 1920), 83–87 Google Scholar. This series is hereinafter abbreviated as PAES, with the relevant section and division.

18. Ibid., map facing p. 1.

19. Ibid., Illustrations 94 and 95.

20. Ibid., 84f.

21. Butler reports none. Ibid., 83f.

22. PAES, Division III, Greek and Latin Inscriptions, Section B: Northern Syria, ed. Prentice, W.K. (Leiden 1922), no. 1023 Google Scholar. Hereinafter cited as Prentice, PAES III B and inscription number.

23. The precise dispositions of the limitanti in this district c. 460 are unknown. Allowing that we cannot say what the ancient toponym of Idjaz was, its position does not seem to correspond to any of the billets under the command of the duces Syriae et Eufratensis listed in Notitia Dignitatum Or. XXXIII composed about half a century earlier. Formations were certainly reassigned. See Jones, A.H.M., The Later Roman Empire: A Social, Economic and Administrative Survey (Oxford 1964), 661f Google Scholar., who cites the example of the legio IV Parthica which was transferred from Circesium to Beroia in Syria I sometime prior to the later sixth century. Drill and other regulations for the limitami are set forth in the Twenty-fourth Novel of Theodosius II (443).

24. Prentice, PAES III B, 1020=IGLS 1600.

25. Jones, LRE, 17A. Bury, LRE 1, 317 and n., 319, 395.

26. Cf. Jones, LRE, 651. Novel 24 of Theodosius II contains much information about the economic life of the limitanei and their settlements. Given at Constantinople on 12 September 443, it provides for the regular drilling of these troops by the dux of the province or the principes of the camps and for keeping them at full establishment. A subsistence allowance was granted to the limitanei, apparently from the annona of the province, as also to the Arab allies who presumably collected these from the depots under the control of the dux. This seems to explain the visit of the sheikh Nu’man to Damascus, headquarters of the dux of Phoenice Libanensis, between c. 442–460. Simeon bar Apollon and Bar Chatar, ‘Lobrede auf dem Herrn Simeon das Haupt der Eremiten’, tr. Hilgenfeld, H., Das Leben des heiligen Symeon Stylites, ed. Lietzmann, H. (Leipzig 1908), 146f Google Scholar. Limitanei certainly owned small estates, farming them personally or letting them on leaseholds. These properties enjoyed exemption from corvées (sordida munera or . The rosters of each formation were in theory updated by officials of the Praetorian Prefect (and not the dux) in January and sent to Constantinople. A combination of privileges, subsidies, and exemptions, along with regular drilling, promoted cadre identity in the zones behind the frontier.

27. Procopius, Historia Arcana 24. 12–14. Jones, LRE, 661f. It is quite probable that regular drilling and the other features of the system came to an end, and effectively led to the disbanding of the limitanei. There was a clear decline in the garrisoning of the Provincia Arabia from the time of Justinian onward. Parker, S. Thomas, Romans and Saracens: A History of the Arabian Frontier (ASOR Dissertation Series 6, Winona Lake, Indiana 1986), 149–155.Google Scholar

28. Butler, PAESIIB, 71ff. See p.74, bottom.

29. Various local efforts to restore fortifications are mentioned in the chronicle of Joshua the Stylite, as for example, at Birta-kastra (present-day Birejik) in 505/6: “The excellent Sergius, bishop of Birta-kastra, which is situated beside the Euphrates, began likewise to build a wall to his town; and the emperor gave him no small sum of money for his expenses’. Joshua Styl., Chron. 91 (Wright 71). The place seems to have had no fortifications before 505/6; the monies received came in the form of reimbursements. It is conceivable that other towns and villages, like those at Taroutia, had their expenses covered in this way. The Persian War and building programme are briefly discussed in Capizzi, C., L’imperatore Anastasio I (491–518) (Orientaria Christiana Analecta 184, Rome 1969), 180–185, 216–221.Google Scholar

30. Butler, PAES II B, 15f.

31. Prentice, PAES III B, 993=IGLS 1631.

32. There is also a presbyter John named in one of the inscriptions of Taroutia. Prentice, PAES III B, 988=IGLS 1619.

33. Prentice, PAES III B, 992=IGLS 1630.

34. Cf. the foundation of the Qartmin monastery at Tur ‘Abdin by the monk Samuel c. 397. Palmer, A., Monk and Mason on the Tigris Frontier: The Early History of Tur ‘Abdin (Cambridge 1990), 12, 33–40 Google Scholar. For Anatolia, Trombley, F., ‘Monastic Foundations in Anatolia and their Role in the Social and Economic Life of the Countryside’, GOTR 30 (1985), 53f Google Scholar. Repr. with corrections in Byzantine Saints and Monasteries, ed. Vaporis, N.M. (Brookline, Mass. 1985), 53f Google Scholar. Thomas archbishop of Amida organised a survey of the site of Dara in 506 and his clergy supervised the construction. Rhetor, Zachariah, Die sogenannte Kirchengeschichte, tr. Kruegei, G. (Leipzig 1899), 115–119 Google Scholar = HE7.6.

35. Cf. supra, note 29.

36. Butler, PAES IIB, 75, 111. 82. Prentice, PAES III B, 998=IGLS 1615.

37. IGLS 1598.

38. Prentice, PAES III B, 998=IGLS 1607. The stonecutter used the omicron-upsilon ligature. This puts the inscription in the later 480s or after, but most probably in the sixth century.

39. IGLS 1610. Prentice’s transcription in PAES III B, 1058 is erroneous. He came up with The building was in fact a metatum or military hostel of the saints named.

40. This process is strongly implied in the fragment of John of Epiphaneia, Historiei Graeci Minores, 1, ed. Dindorf, L. (Leipzig 1870), 378, lines 21–28 Google Scholar. The mobilisation of the two of the two Phoenicias for dispatch’ to Antioch in 540 outside the provincial boundaries is another case in point. Procopius, Wars 2.8.2.

41. Cf. the parading of the (‘made without human hands’) icon of Christ, the ‘theandric image’, in front of Philippicus’ army at the battle of the Arzamon river in 586. Theophylact Simokatta, Historiae 2.3.4 (De Boor 73f.). On this see: Kitzinger, E., ‘The Cult of Images in the Age before Iconoclasm’, DOP 8 (1954), 111f Google Scholar. See also: Trombley, F., ‘The Decline of the Seventh Century Town: The Exception of Euchaita’, Byzantina kai Metabyzantina 4 (1985), 83, n. 23.Google Scholar

42. Capizzi, , L’imperatore Anastasio (as in note 29 supra). Rhetor, Zachariah, HE 8.5 (Ahrens-Krueger 157)Google Scholar. A monk named Dada vouched for a story that al-Mundhir, pagan phylarch of the Lakhmids, sacrificed 400 nuns taken from the convent of the apostle Thomas outside Emesa to the Arab divinity al-’Uzza.

43. Joshua the Stylite, Chron. 90 (Wright 70). Procopius, Buildings 2.1.11–2.3.26. Cf. Whitby, M., ‘Procopius’ Description of Daraxs (Buildings II 1–3)’, Defence of the Roman East, 737–783.Google Scholar

44. The Sabir Huns also plundered the territorium of Dara while working their way round Tur ‘Abdin, where two famous monks were killed. Rhetor, Zachariah, HE 8.5 (Ahrens-Krueger 160f.)Google Scholar.

45. Ibid., 367f.

46. IGLS 571 and 545. For the chronology of the inscriptions, see IGLS 4, p.377f. A considerable village clergy was in place again at Zerzita in Djebel Sim’an by 538. IGLS 456 and 462. For the Soghane martyrion see Jarry, J., ‘Inscriptions arabes, syriaques et grecques du Bélus en Syrie du Nord’, Annales islamologiques 9 (1970), no. 56.Google Scholar

47. IGLS 456 and 462.

48. IGLS 562.

49. Grumel, Chronologie, 174. Procopius notes that Khusrau I crossed the frontier just as the 13th regnal year of Justinian was expiring (ante 4 April 540). Procopius, Wars 2.5.1. The Persians took Antioch in June 540. Malaias, John, Chronographia, ed. Dindorf, L. (Bonn 1831), 479f.Google Scholar

50. On the Lakhmids see: Rothstein, G., Die Dynastie der Lahmiden in al-Hira (Berlin 1899)Google Scholar and Trimingham, J.S., Christianity among the Arabs in Pre-Islamic Times (London 1979), 188–202, esp. 191ff Google Scholar. The Lakhmids’ sixth-century raiding activities are variously reported in Shahid, Irfan, Byzantium and the Arabs in the Sixth Century I/I (Washington, D.C. 1995), 339–364, etc.Google Scholar

51. Procopius, Aedificia 2.9.3–4. For the smaller forts around Theodosiopolis and Amida, ibid., 2.4.14–24.

52. John of Ephesus, ‘Lives of the Eastern Saints’, ed. tr. Brooks, E.W., PO 17 (1923), 79–82.Google Scholar

53. The Qartmin monastery in Tur ‘Abdin had an enclosure wall, but little survives except foundations. Palmer, Monk and Mason, 43–45. The Persians sacked the place in a raid of 581. Ibid., 191.1 have rejected the thesis that the Sabir Huns successfully besieged Euchaita in the Helenopontus in 515. Trombley, ‘Euchaita’, 83f., n. 28.

54. In 504/5 the metropolitan of Edessa received 20 pounds in gold (1,440 solidi) to subsidise the repair of the city’s fortifications after the Persian siege of the previous year. Joshua Styl., Chron. 87 (Wright 69). Construction work at Darà began the following year. Joshua Styl., Chron. 90 (Wright 70). Some monies were nevertheless expended on smaller places. For example, Eulogius, governor of Osrhoene, repaired the wall of Batnan-qastra, a site in the territorium of Sarug. The magister militum per Orientem handled the erection of a new wall at the village of Europos or Oropos in the territorium of Hierapolis-Mabbug in Euphratesia. Joshua Styl., Chron. 89–90 (Wright 70f.).For later Justinianic building activity in Mesopotamia, see Procopius, Aedificia 2.1.4–2.3.26. (Dara), 2.3.27 (Amida), and 2.7.1–16 (Edessa).

55. In late 506, after the start of the new indiction, Anastasius wrote to Celer the magister officiorum and generalissimo in Oriens and Calliopus the Praetorian Prefect asking about the advisability of tax remissions. In reply, the officials recommended the remission of all the taxes of Mesopotamia and half those of Osrhoene for the indiction of 505/6. Joshua Styl., Chron. 99 (Wright 75). This plan proved to be too optimistic. Not long afterward, in November 506, Celer had to distribute 200 solidi in largess at the news of peace being concluded. This act was •accompanied by the dubious circumstance that the army was to winter in eastern Osrhoene. Joshua Styl., Chron. 100 (Wright 75f.). Tax remissions of this sort were nothing new. This had been done for certain provinces of the Western Empire between 423–445 for the same reasons. Jones, LRE, 452f.

56. The figures for the depopulation of Mesopotamia and Osrhoene between 502–506 are far from complete. Estimates made at the time included over 80,000 dead at Amida. Joshua Styl., Chron. 53 (Wright 43). This figure is based on the number of bodies that the Persians carried out by the North Gate. The majority must have been villagers who had taken refuge there during the siege. The figure may be broken down as the population of c. 30,000 men, women, and children from Amida and its environs plus those of some 50 villages (with 1000 as a very rough calculation of the average population of each). The latter would have come mainly from the smaller towns and rural sites ahead of the Sassanid line of march from Armenia.For some estimates of village sizes in sixth-century Anatolia, see Trombley, F., ‘Paganism in the Greek World at the End of Antiquity’, HTR 78 (1985), 331,Google Scholar n. 25. The Lakhmid Arabs took 18,500 captives in Osrhoene in a single raid at the outset of hostilities. Joshua Styl., Chron. 52 (Wright 40f.). The figure represents the population of about 20 villages. Most of the Jews of Tella-Constantina were wiped out by the citizens during the Persian siege. Joshua Styl., Chron. 58 (Wright 48). At times the locals could take refuge in small forts, like the castle of Ashparin, but the Lakhmid cavalry must have caught many others outside and led them into captivity. Joshua Styl., Chron. 57 (Wright 46). On balance, the demographic loss in Mesopotamia must have exceeded 100,000 persons. The rural society that arose afterward is described in some detail in John of Ephesus’ ‘Lives of the Eastern Saints’. For examples, see Trombley, ‘Religious Transition in Sixth-Century Syria’, 153–167.

57. Bury, LRE 11, 79. The work at is-Seqe’ah, completed in the fifth indiction, may belong either to 511/2 and be synchronous with the works at Taroutia, or to 526/7, in which case they would be contemporaneous with the tower completed at il-Burdj.

58. Bury, LRE II,23,81.

59. Butler, PAES IIB, 47–63 with map.

60. Evident from Mouterde, R. and Poidebard, A., Le Limes de Chaléis: Organisation de la steppe en haute Syrie romaine, 2 vols. (Paris 1945)Google Scholar, Carte I, and Butler, PAES II B, map of the district facing p. 1.

61. The Princeton Expeditions did not excavate these dwellings seriously nor did they take note of pottery samples, concentrating instead on the main public buildings. The old enceinte encloses a wide expanse full of ancient houses (Plate CXI). No modern structures existed at the time of Butler’s survey. The mid-sixth century kastron was built with alternating bands of oblong blocks in triple courses, interspersed by brick bands five courses thick. Despite its southerly geographical position, Adrona may have belonged to the territorium of Chalkis in the Antiochene rather than that of Apamea. SEG 39 (1989), no. 1613.

62. Butler, PAES IIB, 58–61. Even if the niches between the wall buttresses were used as tombs, as Butler shows, the specific defensive purpose of the enclosure cannot be doubted. The wall buttresses in fact suggest firing platforms behind the curtain wall. The villagers probably buried their dead inside the to protect the Christian dead from being plundered by Arab raiding bands, most of which were still pagan in the mid-sixth century.The Lakhmid phylarch al-Mundhir is said to have been pagan (after 556). Vie de Symeon (as in n. 12), 164f. Cf. Shahid, , Byzantium and the Arabs in the Sixth Century 1/2 (Washington, D.C. 1995), 722–726 Google Scholar. Butler’s preoccupation with the architecture prevented him from considering the possibility that the complex served the dual purposes of .

63. Cf. the equites promoti indigenae, Not. Dign. Or. 33.

64. Mouterde-Poidebard, , Limes de Chalcis, Carte I, and vol. 1, pp. 61ff.Google Scholar Not. Dign. Or. 33.16.

65. In Butler’s day the site was practically buried in rubble. Butler, PAES II B, 50–52. For the size of the late sixth-century , see Maurice, , Strategikon, ed. Dennis, G., tr. Gamillscheg, E. (Vienna 1981), 88f.Google Scholar

66. IGLS 1682.

67. Cf. infra for a near parallel example at Anasartha in the Chalkidike.

68. Expressions like and the like are attributed to the gods in Hellenistic inscriptions, but kings often qualified for these titles as well. Cf. Welles, C. Bradford, Royal Correspondence in the Hellenistic Period (New Haven 1934), no. 5, lines 5, 34, 36 Google Scholar; and no. 69, line 6.

69. Cf. the recurrence of this expression with its variants during the emperor Herakleios’ campaigns against Persia in 622–628. The melting down of church plate partly paid the expenses of these operations. Theophanes the Confessor, Chronographia, 1, ed. Boor, C. de (Leipzig 1883), 315, line 18;319, line 22 Google Scholar; etc. Cf. Herakleios’, dispatch from Persia in Chronicon Paschale, ed. Dindorf, L. (Bonn 1832), 732, line 6; 734, line 13 Google Scholar; etc. Paucis verbis it was an expression used in time of invasion and military necessity.

70. The chronicle of Joshua the Stylite details the complaints of the citizens of Edessa against the soldiery billeted there. These included drunkenness, theft of cattle, appropriation of food and clothing, abuse of women, and enforced quartering in civilian houses. Joshua Styl., Chron. 86, 94 (Wright 68, 73). The church suffered as well:’They were billeted even upon the priests and deacons, though these had a rescript from the emperor exempting them therefrom.’ Ibid., 86 (Wright 68).

71. IGLS 2501. The inscription’s date is lost, but its round uncials suggest a fifth-century date. Rights of asylum were granted to many local churches and monasteries in Anatolia and Syria. No complete list can be given here. For Syria cf. Kaufmann, CM., Handbuch der altchristlichen Epigraphik (Freiburg 1917), 407–410 Google Scholar; Phoenice:Dain, A. and Rouillard, G., ‘Une inscription relative au droit d’asile, conservée au Louvre’, Byzantion 5 (1930), 315–326 Google Scholar; Cyprus:Mitford, T.B., ‘New Inscriptions from Early Christian Cyprus’, Byzantion 20 (1950), 162–165 Google Scholar. The list could be extended considerably.

72. One hardly gets a hint of this in Procopius. A bilingual inscription in Latin and Greek from the reign of the coemperors Justin I and Justinian I identifies similar problems in Anatolia. Found at a site in Pisidia and dated 1 June 527, it proscribes the abuse of agricultural personnel on the estates belonging to a chapel of St. Diehl, John. C. (ed. pr.), ‘Réscrit des empereurs Justin et Justinien en date du 1er juin 527’, BCH 17 (1893), 501–520. Recueil des inscriptions grecques chrétiennes d’Asie Mineure, ed. Grégoire, y (Paris 1922), no. 314.Google Scholar

73. IGLS 1685.

74. IGLS 1684.

75. IGLS 1809.

76. IGLS 2524. This terminology carried on well into the Byzantine period in Anatolia, as we learn from an inscription on the fortifications of Attaleia in Pamphylia dating from 911/12: ‘(Cross) With his sweet son Constantine, the all-serene pious emperor Leo, who has always acted with fatherly foresight on behalf of his subjects as though they were his own children and who has [---] to accomplish all things sympathetically, as though he pondered the safety of all men, wisely fortified this Christ-beloved city with a second (i.e. outer) wall in order to save it thereby showing it to be stronger than itself (i.e. with reference to the earlier, inner wall) and taller than every siege engine of the enemy. The imperial hand is the provider (lit. ‘patron’) of the work, as also its master and defrayer of costs. Euphemios son of Krates the confidential clerk was the noble superintendent who performed the work zealously’. Bean, G.E., ‘Inscriptions in the Antalya Museum’, Türk Tarih Kürümü. Belleten 22 (1958), no. 41.Google Scholar

77. Cf. Kirsten, Ernst, ‘Die byzantinische Stadt’, Berichte zum XI. Internationalen Byzantinisten-Kongress 5/3 (Munich 1958), 19–34.Google Scholar

78. Bury, LRE II, 113–123.

79. Cod. lust. 1.4.26 (540 A.D.). Cf. Trombley, F., ‘“Dark Age” Cities in Comparative Context’, To Hellenikon: Studies in Honor of Spews Vryonis, Jr. I: Hellenic Antiquity and Byzantium, edd. Langdon, J., Reinert, S. et alii (New Rochelle, New York 1993), 443f.Google Scholar, n. 35.

80. Butler, PAES II B,73f.

81. Ibid.,74.

82. This interpretation is consistent with Whitby, Emperor Maurice, 250–304. On the background to the war, see also: Lee, A.D., ‘Evagrius, Paul of Nisibis and the Problem of Loyalties in the Mid-Sixth Century’, Journal of Ecclesiastical History 44 (1993), 569–585.CrossRefGoogle Scholar

83. Theophylact Simokatta, Historiae 5.3.4–7. It is harder to trace Late Roman institutions in Dara thereafter. On the episcopate of Dara and Tur ‘Abdin in the seventh century, see Palmer, Monk and Mason, 149–155.

84. Ibid., 256–8.

85. A perusal of IGLS indicates the complete absence of dated inscriptions in the Limestone Massif between the termini 540–550. The last of these before the Persian march mentions the completion of a gallery of the second storey of a house at Zerzita near Djebel Sim’an and is dated 539 A.D. IGLS 456. The editors have erroneously put IGLS 562 into the index at 540 A.D. The correct date for the latter is 480 A.D. It is remarkable indeed that plague theories have been used to explain the decline in building activity in the Limestone Massif during the 540s, particularly as the terminus comes in 539 and not 542.1n contrast, the southern Apamene does show signs of building activity between 540–544, including a small fortress at Tamak (after September 1, 540) which must have been put up in direct response to the Sasanid invasion that took Apamea earlier that year. IGLS 1957. At Kome Olbanon, also in the far south, a tower was erected in 543. IGLS 1889.The other two buildings constructed between 540–544 were civilian houses. IGLS 1873, 1818.The Persians did not reach this part of the Apamene in 540, but the region may have been troubled by the threat of Lakhmid Arab attacks.

86. For the siege of Apamea, see John, of Ephesus, The Third Part of the Ecclesiastical History of John Bishop of Ephesus, tr. Smith, R. Payne (Oxford 1860), 385–387 Google Scholar = HE 6.6.

87. Evagrius, HE 5.10 (Bidez-Parmentier 206).

88. Millar, Fergus, The Roman Near East 31 BC-AD 337 (Cambridge, Mass. 1993), 250.Google Scholar

89. John of Ephesus, HE 6.6 (Payne Smith 385f.).

90. The sponsor of the mosaic was the civil official Apellion. IGLS 1344. The main ornamentation of the basilica was executed by archbishop Paul between 533–540. SEG 26 (1976–77), 1627–1630.

91. The inscriptions are respectively IGLS 1709, 1598, 1609, 1682, 1505.

92. For the southern parts of Syria II, see IGLS 1714–1997.The Apamene proper: IGLS 1311–1713.

93. IGLS 1555, 1673, 1678, 1392 respectively. For a chronology of the inscriptions see IGLS 4, p. 376.

94. On the agricultural regime of the Limestone Massif and some aspects of sixth-century demographic decline, see: Tchalenko, G., Villages antiques de la Syrie du Nord 1 (Paris 1953)Google Scholar. See now Gatier, R-L., ‘Villages du Proche-Orient protobyzantin (4ème-7ème s)’, The Byzantine and Early Islamic Middle East II: Land Use and Settlement Patterns, edd. King, G.R.D. and Cameron, Averil (Princeton 1994), 17–48 Google Scholar. Semi-arid farming techniques as adapted to the Negev desert in Palaestina III are discussed in Mayerson, P.. ‘The Ancient Agricultural Regime of Nessana and the Central Negeb’, Excavations at Nessana (Auja Hafir), Palestine I, ed. Colt, H.D., 1 (London 1962), 211–257.Google Scholar

95. The series in IGLS runs 355–698 and excludes the steppe east of Beroia, the so-called limes of Chalkis.

96. The inscriptions in IGLS are 493, 541, 603, 572, 620, 635, 362, 546, 642, 403, 478, 357. I have omitted IGLS 391 from the series. Its date is disputed because of a damaged letter (chi versus upsilon). As a ‘One God’ inscription it belongs more plausibly in 363 than 563. The inscriptions of Tawam and Kefr Qalbloze are noted in Jarry, J., ‘Inscriptions arabes, syriaques et grecques du Massif du Bélus en Syrie du Nord’, Annales islamologiques 7 (1967), 139–220, with 74 plates, nos. 93 and 126 Google Scholar. For the inscription at Kfellousin: idem, ‘Inscriptions arabes, syriaque et grecques du Massif du Bélus en Syrie du Nord (suite)’, AI 9 (1970), no. 37.

97. An inscription mentioning a fortress at Me’ez (Djebel Barisha) seems to belong to the period of Byzantine reconquest in the 10th-11th c. Jarry, ‘Inscriptions du Massif du Bélus’, no. 115=IGLS 586. The depopulation of the Limestone Massif thus had little to do with the plague of 542, as the chronology is off by 2–3 years in the epigraphy. If anything, the plague would have offered incentives to recolonise the villages there, as the casualties from the plague were popularly perceived to be lower in the countryside. Arguments have been made for the plague being a factor in the ruralisation of the sixth-century economy in Trombley,’Monastic Foundations in Anatolia’, 57–59 (Lycia) and ‘Religious Transition in Sixth-Century Syria’, 192 (western Antiochene). On this see also: Conrad, L., ‘Epidemic Disease in Central Syria in the Late Sixth Century: Some New Insights from the Verse of Hassan ibn Thabit’, BMGS 18 (1994), 21f Google Scholar. See also the observations of Georges Tate, who leaves the question open (infra, note 186).

98. IGLS 502, 533,403,530, 500,401, 563. The date of no. 403 is disputed, being either 571/2 or 586/7. Sarfud: Jarry, J., ‘Inscriptions de Syrie du Nord revelées en 1969’, Annales islamologiques 9 (1970), no. 13.Google Scholar

99. IGLS 530.

100. IGLS 528, 589.

101. IGLS 563.

102. IGLS 401.

103. Sources: Theophanes, Chronographia 1, 290–299. Nikephoros of Constantinople, Historia Syntomos, ed. Boor, C. de (Leipzig 1880), 3ff.=Short History, ed. tr. Mango, C. (Washington, D.C. 1990), 34–37 Google Scholar. Older literature includes Stratos, A.N., Byzantium in the Seventh Century 1/1, tr. Ogilvie-Grant, M. (Amsterdam 1968), 69ff.Google Scholar, and Ostrogorsky, G., A History of the Byzantine State, tr. Hussey, Joan (Brunswick, New Jersey 1969),76ff Google Scholar. More recently: Olster, D., The Politics of Usurpation in the Seventh Century (Amsterdam 1993), 81–97 Google Scholar, the most thorough survey of the subject to date.

104. George the Monk, Vie de Théodore de Sykeon, ed. tr. Festugière, A.-J. 1 (Subsidia Hagiographica 48/1, Brussels 1970), 96–119 Google Scholar. Kaegi, W., ‘New Evidence on the Early Reign of Herclius’, BZ 66 (1973), 308–330.Google Scholar This material has not been studied in detail as evidence for rural society during the Persian War of 604–628.

105. Theophanes, Chronographia 1, 292, line 27f.

106. Ibid., 293, lines 23–26.

107. Ibid., 295, lines 14–16.

108. See generally Whitby, ‘Procopius’ Description of Dara’, (as in note 43).

109. Whitby plays down the significance of Dara in the overall scheme of frontier defence until the war of 572–591, a useful caution against monocausal explanations for the military catastrophes that followed: Maurice and His Historian, 209–213.

110. On the chronology, Olster, Politics of Usurpation, 95–97.

111. On this, see infra.

112. IGLS 348.Cf. 349.For the military gate of Chalkis, see: Foudrin, J.-P., ‘Une porte urbaine construite à Chaléis de Syrie par Isidore de Milet le Jeune (550/551)’, TM 12 (1994), 299–307 Google Scholar, including a new edition of the inscriptions by Feissel, D.. Cf. Mouterde-Poidebard, , Limes de Chalcis, Plates II–IV.Google Scholar

113. Bosworth, CE., ‘Iran and the Arabs before Islam’, Cambridge History of Iran 3/1: The Seleucid, Parthian and Iranian Periods, ed. Yarshater, E. (Cambridge 1983), 607 Google Scholar. The battle was fought near the spring of ‘Ouadaye, an otherwise unknown site: Shahid, Byzantium and the Arabs in the Sixth Century 1/1, 240–251.

114. Theophylact Simokatta, Historiae 5.11.5–7. See Whitby’s comment, Maurice and His Historian, 147.

115. Mouterde-Poidebard, Limes de Chalcis, 53.

116. The war of attrition between 575–591 saw few pitched battles, but much raiding and many sieges. For the course of operations, see: Whitby, , Maurice and His Historian, 262–292 Google Scholar. It is most probable that Herakleios’ incursions into Persia between 622–627 were dictated by the historical experience of the previous war.

117. Mouterde-Poidebard, Limes de Chalcis, nos. 1, 28, 62, and Syriac no. 7.

118. Prentice, PAES III B, 947 = IGLS 1673. Butler, PAES IIB, 63f. and Plate IX.

119. Mouterde-Poidebard, Limes de Chalcis, 167f.

120. Ibid., 167f.

121. Ibid., 170f. and Plates XCII.2 and XCIV.

122. IGLS 270. Mouterde-Poidebard, Limes ofChalcis, Plates XXXIII-XXXV.

123. IGLS 316. Mouterde-Poidebard, Limes ofChalcis, Plates XCVII-CI.

124. Zebed seems not to have had urban status. Nor is it clear that a regular cavalry formation was ever stationed there. Dussaud, R., Topographie historique de la Syrie antique et mediévale (Paris 1927), 203f.CrossRefGoogle Scholar

125. IGLS 310–313.

126. The dated inscription is IGLS 307. Mouterde-Poidebard, Limes of Chalcis, 162–167 and Plan VI.

127. The various dating possibilities are discussed in the different editions of the inscriptions. The editors of IGLS 288 came down in favour of 594/5 as the date. Previously edited in Publications of an American Archaeological Expedition to Syria in 1899, Part III: Greek and Latin Inscriptions, ed. Prentice, W.K. (New York 1908), no. 318 Google Scholar. Hereinafter cited as AAES III. Prentice suggests that portraits or effigies of the town’s benefactors completed the programme of the inscription.For photos of the site, see: Mouterde-Poidebard, Limes de Chalcis, Plates XXXI-XXXII.

128. Noted in Sebastian Brock, ‘Syriac Sources for Seventh-Century History’, BMGS 2 (1976), 29.

129. Khusrau II Parviz’s interest in the intercessory powers of St. Sergius at the shrine in Resapha can only have been viewed with scepticism by Domitianus, archbishop of Melitene (Armenia II), who as Maurice’s cousin or nephew conducted the negotiations that led to the recognition of Khusrau’s claim to the Sasanid throne. Theophylact Simokatta, Historiae 5..2.1–6 and 5.12–15. On the Avar war, see Whitby, Maurice and His Historian, 138–183.

130. Theophanes, Chronographia 1, 291 Sebeos, f., Histoire d’Héraclius, tr.Macler, F. (Paris 1904), 56f Google Scholar. Olster rejects the argument that Narses allied himself with Khusrau II. Politics of Usurpation, 92–94.

131. A.H.M. Jones’ views on this subject are not strictly applicable to the later sixth century. Even if the actual construction of frontier fortresses was supervised by the duces of the provinces, the Anasartha inscription of 594/5 makes it clear that the funds came from the praetorian prefecture. This will have included the salaries of the military engineers, funds for hiring stonecutters and carpenters, and so forth. This was in fact an important feature of the ‘public works’ function of the prefecture. Jones, LRE 461f. and 1198, n. 125. Most of the epigraphic evidence cited by Jones is of western provenance and of fourth-century date. The chronicle of Joshua the Stylite and Codex Iustinianus are in full agreement with the epigraphy that the bishops of the eastern provinces had begun to take considerable responsibility for this by the early sixth century. The division of tasks and source of funds was complex and locally differentiated. Cf. Trombley, ‘“Dark Age” Cities in Comparative Context’, 443f., n. 35.

132. IGLS 92=Prentice, AAES 111, 325. On the Arabic derivation of the family name , Shahid, Byzantium and the Arabs in the Sixth Century 1/1, 626–631.

133. IGLS 281.Cf. no. 292.

134. Cf. the discussion of the seventh-century sections of the life of St. Theodore of Sykeon, infra, Appendix.

135. IGLS 291.

136. Inscriptiones Latinae Selectae, ed. Dessau, H. 1 (Berlin 1954), no. 837 Google Scholar. Ostrogorsky has misunderstood the point of the inscription, which Smaragdus (and not the pope) erected as governor of Italy (praepositus palatii ac patricias et exarchus ltaliae). The act had nothing to do with pope Gregory I’s demand that the patriarch of Constantinople abjure the title of ‘oecumenical patriarch’. Byzantine State, 83f. The column was in any case erected during the pontificate of Boniface DI, whose biography provides no evidence whatever of a pope with his hands in secular affairs. Liber Pontificalis, Pars Prior, ed. Mommsen, Th. (Berlin 1898), 164 Google Scholar. Boniface reigned for nine months.

137. A later example of damnatio memoriae may be found in the deposition of the brothers of Constantine IV in 680/1.

138. See IGLS 298 for the question of provenance. For the months of the Macedonian calendar, see Grumei, V., Traité d’études byzantines I: La chronologie (Paris 1958), 168.Google Scholar

139. The documentation on this common phenomenon has yet to be brought together. To cite one important example, brick crosses were embedded in the Late Roman fortifications of Thessalonike in Macedonia Tafrali, I.O, Topographie de Thessalonique (Paris 1912)Google Scholar, Plate IV. 1. There is also the inscription on the Hormisdas Tower (late 4th a): ‘Hormisdas completed the city with unbreakable walls…’ Ibid., 32f. The church had taken over this task of maintaining parts of the fortifications by the late sixth century, as we learn from an inscription on the sea wall: ‘It was completed at the command of the most holy bishop Eusebius’ (c. 600 A.D.); ibid., 40f. Cf. Spieser, J.-M., ‘Inventaire en vue d’un receuil des inscriptions historiques de Byzance’, TM 5 (1973), 154 Google Scholar; idem, ‘Note sur le rempart maritime de Thessalonique’, TM 8 (1981), 477–485; Vickers, M., ‘The Byzantine Sea Walls of Thessaloniki’, Balkan Studies 11 (1970), 261–280 Google Scholar. The continuity of the idea of fortifications as the demarcator of ‘sacred space’ in the Late Roman town is aptly expressed by Digest 1.8 and again as recapitulated in the Appendix to the Ekloga c. 800 A.D.: ‘The walls are sanctified, that is to say holy. It is forbidden for anyone to damage them. Buildings devoted to public use, whether in town or countryside, are sacred’ . Appendix 3.8 in Appendix Eclogae, ed. Burgmann, L. and Troianos, Sp. in Forschungen zur byzantinische Rechtsgeschichte 4, ed. Simon, D. (Fontes Minores III, Frankfurt am Main 1979), 119.Google Scholar

140. Mouterde-Poidebard, Limes de Chalcis, Plates LIV and LV.

141. Ibid., no. 39. Only the first two letters of the name of the donor are visible. The identification with Gregorios is perfectly plausible. The site has a small building or fort with cisterns. Its ancient name was perhaps Ammatha, the fourth-century encampment of the Cohors Prima Victorum.Ibid., 85f.

142. The various agreements are noted in Whitby, Maurice and His Historian, 297–304.

143. Al-Balādhurī, Kitāb futūh al-buldān in Liber de expugnationis regionum, ed. Goeje, M. de, 2nd ed. (Leiden 1968), 144–152 Google Scholar. This work mentions the existence of both an akropolis and lower city fortifications at Chalkis at the time of the Arab conquest.

144. For the configuration of the fort, See: Mouterde-Poidebard, Limes of Chalcis, Plate XL.

145. IGLS 271 = Prentice, AAES III, 332. Ephrem Syrus’ Hymn 1.8, an invocatory incantation against the Persians in the siege of 350 against Late Roman Nisibis, also alludes to the Matthaean ‘rock’. Bonian, S., ‘St. Ephrem on War, Christian Suffering, and the Eucharist’, Parole de l’Orient 11 (1983), 158.Google Scholar

146. IGLS 273.

147. These deductions are based on Mouterde-Poidebard, Limes of Chalcis, Plate XLI.

148. IGLS 272. The date is based on the omicron-upsilon ligature.

149. IGLS 276.

150. IGLS 274. St. Dometius was executed by Julian the Apostate in 363. The inscription could be of fifth- or sixth-century date. Cf. Gheyn, Van den, ‘Acta Graeca S. Dometii Martyris’, Analecta Bollandiana 19 (1900), 285–317.Google Scholar

151. IGLS 272 note. I am unable to find the passage in the chronicle of Michael the Syrian cited by the editors of IGLS. It is perhaps a misprint for p.547 of the Syriac text, mentioning the burial of Mar Dionysius of Tell Mahre, the Jacobite patriarch of Antioch, at the ‘monastery of Qenneshrin’ on 22 August 845 A.D. John of Ephesus mentions a convent at Chalkis-Qenneshrin, so it seems going too far to link this reference to the site at Mu’allaq. Cf. Michael, the Syrian, Chronique III, ed. Chabot, J.-B. (Paris 1905), 116 Google Scholar. A monastery of Mar Sargis is later known at Mu’allaq, but the oblivion of the earlier name of the site — that of St. Barapsabba — is a sure sign of discontinuity.Cf. Honigmann, Ernst, Le couvent de Barsauma et le patriarcat jacobite d’Antioche et de Syrie (CSCO 146, Subsidia 7, Louvain 1954), 117 Google Scholar. Honigmann seems to have been unaware of the Mu’allaq inscriptions.

152. Theophanes, Chronograpkia 1, 296, lines 6–10.

153. Thus Theophylact Simokatta, Historiae 5.1.7 and 5.13.

154. These dates are consistent with the evidence of the Syriac Chronicle of 724: Saint Anastase le Perse et l’histoire de la Palestine au début du Vile siècle, ed. tr. Flusin, B., II (Paris 1992), 71–74 Google Scholar. It indicates that in 608/9 the Persians operated mainly in northern Mesopotamia, taking Mardin and Rheshaina. In 609/10 came the turn of Osrhoene and Euphratesia, with the loss of Edessa, Harran, Callinicum, Circesium and all remaining territory on the east bank of the Euphrates.

155. Theophanes, Chronographia 1, 299, lines 14–17.

156. Theophanes indicates that the Persians merely reached the environs of Antioch in 611. With an evident desire to absolve Herakleios and blame Phokas for the catastrophe, Michael the Syrian has moved the date backwards to 609/10, associating the fall of Chalkis and Beroia with that of Antioch. Chronique 2, 378. All three cities probably fell in the same year, 613. This is consistent with Sebeos’ account, who mentions a great battle outside the walls of Antioch shortly before Herakleios’ first son’s elevation to the rank of Augustus on 13 January 613. Histoire d’Héraclius, 67. For the date, see: Grumei, Chronologie, 356. Unfortunately, it is not clear from Sebeos’ account who controlled the city at this point. If, as Theophanes suggests, Antioch was still in Roman hands at the beginning of 613, we may safely put the Sasanid occupation of it not earlier than the spring of 613. Kaegi, following Norman Baynes, concurs. ‘Early Reign of Heraclius’, 328f. The great Arab raid against Syria mentioned by Theophanes for 613 would have put the seal on the depopulation of northern Syria and the Apamene. The failure to read Theophanes’ Greek closely may have caused scholars to conflate two sieges, an unsuccessful one in 611 and its unpleasant sequel in 613. See Michael the Syrian, Chronique 2, 400. The excavators prefer 613 for the capture of Apamea.

157. The series of dated coins from the reign of Herakleios runs: one from 615–24 (no. 29), one from 624–29 (no. 45), one from 626–29 (?)(no. 32), two from 627/8 (nos. 30 and 60), one from 629/30 with countermarks (no. 70), and one from the time of the Byzantine reconquest (difficult to decipher) (no. 44). Fouilles d’Apamée de Syrie I/I: L’église à l’atrium de la grande colonnade (Brussels 1969), 89–92 and 91, note 5.

158. Of the coins found in the ‘Mound’ of Hama-Epiphaneia (a section of the city occupied in the Late Roman period) all but one came from the mint of Constantinople. The sequence runs: 2 folles (612/3), follis (614/5), follis (615/6), follis (624/5), follis (mint of Cyprus, 626/7), follis (628/9), etc. Hama: Fouilles et recherches 1931–1938111/1, edd. Christensen, A.P., Thomsen, R., and Ploug, G. (Copenhagen 1986), 62f Google Scholar. The number of coins for the reign of Herakleios per annum is considerably greater than that for his predecessors, 19, compared with 9 tor Justinian I, 5 for Justin II, 2 for Tiberius Constantine, 4 for Maurice, and 4 for Phokas.

159. Jarry, ‘Inscriptions du Massif du Bélus’, nos. 5, 11, 15, 16, 74, 101, 118 and 154. IGLS 564 (Ba’ouda). The editors of the latter suggest the possibility that the date of the inscription is given in the Seleucid era rather than that of Antioch, putting the construction of the church in 392/3. The argument should be rejected on the ground of probability. There is not even one example of the use of the Seleucid era in the Limestone Massif, even in the 7th c. inscriptions listed above, and Greek was clearly the dominant epigraphic language at that time. The era of Antioch, rather than Seleucid, is always to be preferred in reading inscriptions of ostensibly fourth-century date, when the meddling of the council in the affairs of the city’s territorium was certainly at its height. Cf. Butler, PAES II B, 163 and 111. 174. For the inscriptions of Deir Tell-’Adeh, see note 184 infra.

160. Gerasa: A City of the Decapolis, ed. Kraeling, C. (New Haven 1938), 486f.Google Scholar, no. 335.

161. Ibid.,72.

162. The procedure for the donor system was established by emperor Zeno c. 470–477. Cod. lust. 1.2.15.

163. The numismatic data published for Gerasa in 1938 put the ‘sharp break’ in the circulation of coinage during the reign of Justin II (565–578).Kraeling, Gerasa, 503. We might suppose from this that tax burdens were shifted to neighbouring provinces to compensate for the revenues lost after the depopulation of Syria I and II in 573.

164. Procopius, Buildings 5.8.1–9. Forsyth, George H. et alii, The Monastery of Saint Catherine at Mount Sinai: The Church and the Fortress of Justinian (Ann Arbor 1973).Google Scholar

165. Cyril of Scythopolis, Vita Sabae, S. 73 in Kyrillos von Skythopolis, ed. Schwartz, E. (Texte und Untersuchungen 49/2, Leipzig 1939), 178.Google Scholar

166. On this see: Morony, M., ‘Syria under the Persians 610–629’, Proceedings of the Second Symposium on the History of Bilad al-Sham during the Early Islamic Period (English and French Papers), ed. Bakhit, M. A. (Amman 1987), 87–95 Google Scholar. Schick, R., The Christian Communities of Palestine from Byzantine to Islamic Rule: A Historical and Archaeological Study (Princeton 1995), 20–48.Google Scholar

167. A useful exception is Kaegi, W., Byzantium and the Early Islamic Conquests (Cambridge 1992), 26–46.CrossRefGoogle Scholar

168. A large number of building and mosaic inscriptions dated by the era of the Provincia Arabia between the years 614–628 are found in: Meimaris, Y.E. et alii, Chronological Systems in Roman-Byzantine Palestine and Arabia: The Evidence of the Dated Greek Inscriptions (Athens and Paris 1992),279–292 Google Scholar, nos. 445, 450, 453 (=SEG 7. 1197), 455, 456, 461, 464, 474, 480, 481 (=Waddington 2412m), 483 and 491 (=Waddington 1997). For Waddington, see infra, note 170.

169. Ibid.

170. Recueil des inscriptions grecques et latines de la Syrie, ed. Waddington, W.H. (Paris 1870), no. 2497.1 Google Scholar wish to thank Johannes Koder who supplied me with a copy of M. Restle’s photograph of this inscription and the latter’s corrected transcription that puts the arrival of the plague and bishop Wa’ir’s death between 22 March 542 and 21 March 543. Meimaris gives the uncorrected date (512/3) in Chronological Systems, 226, no. 243.

171. Meimaris, Chronological Systems, passim.

172. Cf. the brief observations in Trombley, F., ‘The Greek Communities of Umayyad Palestine (661–749 A.D.)’, Proceedings of the First International Congress on the Hellenic Diaspora from Antiquity to Modern Times 1: From Antiquity to 1453, ed. Fossey, J.M. et alii (Amsterdam 1991), 261f.Google Scholar

173. Meimaris, Chronological Systems, 293–304, nos. 494–531. The last two of these, dated to 756 and 785, belong to the early Abbasid period.

174. The bronze coin series shows a sharp break between 614/5 to 624/5. The six coins before the break date from 603/4, 606/7(two examples), 607/8, 602–10, and 614/5, representing the mints of Cyzicus, Nicomedia and Constantinople. Those immediately after the break are at least two, of 624/5 and 629/30, both from the Constantinople mint. They were probably deposited after the Sasanid evacuation. Four coins from the Alexandria mint belong generally to 610–641 (the latter city being under Sasanid control between 619-c. 628). Excavations at Nessana (Auja Hafir, Palestine, ed. Colt, H.D. et alii, 1 (London 1962), 73f Google Scholar. The epigraphic record is in agreement, with texts from September and October 605, 614/5, 624, 628, and 630. All but one of these inscriptions are either funerary or without clear function, the exception being a building inscription from an unknown location of 18 September 605. Ibid., 194, and nos. 12, 13, 14, 72, 81, and 152. The lack of funerary inscriptions between 614/5 to 624, coupled with an identical gap in the coinage, could reflect the temporary abandonment of the site in the face of the Sasanid invasion of 614. The destination of the migrants may have been Alexandria. The papyri consist of two archives, one covering 602–608, the other 674–690. Excavations at Nessana III: Non-Literary Papyri, ed. Kraemer, C. (Princeton 1958), 28–35 Google Scholar. But the argument is vitiated by the fact that they derive from discrete archives rather than chance finds: other archives from the period between the termini 608–674 may have been lost or destroyed. Schick mentions this material, but addresses a different set of problems, those of the continuity of Christian demography and institutions under the Umayyad and early Abbasid caliphates. Christian Communities of Palestine, 420f.

175. Excavations at Nessana l, no. 106.

176. This point is repeatedly stressed by the articles in the volume edited by Averil Cameron and G.D.R. King, supra, note 94.

177. PAES III, Section A: Syria, Southern, edd. Littmann, Enno, Magie, D. and Stuart, D.R. (Leiden 1921), no. 686.Google Scholar

178. There is an error in the date. The Acta Anastasii give the twentieth regnal year of Herakleios (5 October 629–4 October 630) and the fifth indiction (1 September 631–31 August 632). The earlier date is preferred. Acta S. Anastasii Persae, ed. Usener, H. (Bonn 1894), 12f Google Scholar. Flusin, Saint Anastase le Perse I, 76f., 101, note 11.

179. Theophanes, Chronographia 1, 321f. The existence of Christian institutions (a monastery and perhaps a village church) in the territorium of Palmyra around this time is illustrated by a Syriac inscription of 574/5. R. Mouterde, ‘Inscription syriaque duGebel Bil ‘ās’, MUSJB 25 (1942–3), 81–86.

180. Winnett, F.V., Safaitic Inscriptions from Jordan (Toronto 1957), no. 88 Google Scholar. There are arguments for and against this date. Winnett initially preferred a seventh-century date for the inscription, reckoning it contemporaneous to the Sasanid evacuation of Palestine and Arabia c. 628. He later reversed himself and put the text in the first century B.C. He was put off by the square letters of its Safaitic script which are typical of the earliest epigraphie examples around this time. G. Lankester Harding took the view that the late, even archaizing use of square letters cannot be excluded. Ibid., nò. 78. The argument has nothing against it. Contacts between the settled populations of Phoenice Libanensis and Arabia and the Arabs of the Safa’ would have enabled them to observe the cultural fashion for square scripts in Greek, Syriac, and even Arabic inscriptions. PAES 111 A and B, passim. Cf. Répertoire chronologique d’epigraphie arabe, ed. Sauvaget, J. et alii, 1 (Cairo 1931), no. 3 Google Scholar (Harran, Arabia, 568 A.D.). The fourth-century inscription of Imru’ al-Qays at Namara on the eastern edge of the Provincia Arabia has rhany rectilinear features. Shahid, Man, Byzantium and the Arabs in the Fourth Century (Washington, D.C. 1984)Google Scholar, Frontispiece (photo and transcription). Cf. Cantineau, J., Le Nabatéen II: Choix de textes — lexique (Paris 1932), 49–51.Google Scholar

181. The sections of al-Baladhuri on the Chalkidike are cited supra, note 143.

182. Cf. the many examples collected in Meimaris, Chronological Systems, passim.

183. Tate, Georges, Les Campagnes de la Syrie du Nord du le au Vile siècle. Tome I (Paris 1992)CrossRefGoogle Scholar. Cf. Sodini, Jean-Pierre, ‘La contribution de l’archéologie à la connaissance du monde byzantin (IVe-VIIe siècles)’, DOP 47 (1993), 139–184 Google Scholar. This is now brilliantly illustrated in: Grèce, Michel de et alii, (Athens 1993).Google Scholar

184. Cf. supra note 159. Early medieval Syriac inscriptions often refer to monasteries, as for example that at Deir Tell’Adeh in the Sermada plain of the Limestone Massif. PAES, Division IV: Semitic Inscriptions, Section B: Syriac Inscriptions ed. Littmann, E. (Leiden 1940), nos. 16 and 17 Google Scholar (907/8 and 941/2 A.D.). A Christian householder erected an inscription on his door in 772/3 (no. 51).The monastery at Burdj is-Sab’ belongs to 858/9 (no. 19). Syriac became the dominant Christian epigraphic language on buildings in parts of the Beq’a valley in Phoenice Libanensis. Cf. Mouterde, R., ‘Inscriptions en syriaque dialectal à Kamed (Beq’a)’, Mélanges de l’Université Saint Joseph, Beyrouth (=MUSJB) 22/4 (1939), 71–106 Google Scholar. Four are dated by year 96 of the hira or 714/5 A.D. (nos. 10, 20, 21 and 28). ‘Trente ans après les inscriptions de Kamed (Complément)’, MUSJB 44 (1968), 21–29. More striking is a lone Greek inscription also from Kamed dated like the four above to 714/5 A.D. with the Muslim era. It is a Christian dedication by a Graeco-Syrian intendant or commissary (rxKTOuápioç) named George from Edessa on behalf of himself and the ‘brother of Gabriel’. It reflects plausible literary Greek but with some difficulty in transliterating vowels. IGLS 2988. The mosaic of the bishop Michael (of an unknown see) on the church at Niha was commemorated in 739 A.D. Mouterde, R., ‘Un ermitage melkite en Emesène au Ville siècle’, MUSJB 18 (1934), 101–106 Google Scholar. The continuity of an agricultural and probably Christian Syriac-speaking society in the Limestone Massif down to the 9th or 10th c. is strongly suggested by the archaeology of Dehes (Djebel Barisha), the only site in the district to have been thoroughly excavated. Sodini, J.-P., Tate, Georges, et ai, ‘Déhès (Syrie du Nord). Campagnes I-III (1976-1978)Google Scholar. Recherches sur l’habitat rural’, Syria 57 (1980), 1–304. Unfortunately, its inscriptions are insignificant. Iarry, ‘Inscriptions du Massif du Bélus’, nos. 19 and 106. Cf. P.L. Gatier, ‘Les inscriptions grecques d’époque islamique (Vile-Ville siècles) en Syrie du Sud’, La Syrie de Byzance à l’Islam, ed. Canivet, P. and Rey-Coquais, J.-P. (Damascus 1992), 145–155.Google Scholar

185. The earliest Arabie inscription mentioning the construction of a mosque in the Apamene seems to be the one at Selemiyeh (Salamias) and dates from no earlier than 767 A.D. Four other Islamic inscriptions belong to the 9th-11 th c. Publications of an American Archaeological Expedition to Syria in 1899, Part IV: Semitic Inscriptions, ed. Littmann, E. (New York 1904), nos. 1–5 Google Scholar. An Arabic inscription commemorating the birth of a child at Zorava-Ezra’ in the Ledja in Provincia Arabia dates from 837 A.D. It could be either Christian or Muslim. Ibid., no. 8. A mosque was founded at Hass in Djebel Zawiya in May 944. J. and D. Sourdel, ‘Notes épigraphique et de topographic sur la Syrie du Nord’, Annales archéologiques de Syrie 3 (1953), 81–83. In general: Sourdel-Thomine, J., ‘Inscriptions et graffiti arabes d’époque umayyade’, Revue des études islamiques 32 (1964), 115–120 Google Scholar. There is, of course, much ‘Umayyad epigraphy elsewhere, as for example in the palaces on the Syrian desert fringe. Cf. Abu ‘1-Farağ al-Us, ‘Inscriptions arabes inédites à Djabal ‘Usayr’, Annales archéologiques de Syrie 13 (1963), 225–238; Ory, S., ‘Les graffiti umayyades de ‘Ayn al-Ğarr’, Bulletin du Musée de Beyrouth 20 (1967), 97–148 Google Scholar; and Schlumberger, D. et al., Qasr el-Heir el-Gharbi (Paris 1986), 28.Google Scholar

186. Georges Tate cites any number of possible causes for the demographic decline of the Limestone Massif but withholds any final judgement. Compagnes de la Syrie du Nord, 343–350. On the plague, see notes 97 and 170 supra. Clive Foss subscribes to a military explanation while rejecting plague theories and other hypothetical constructs such as soil erosion, deforestation or climatic change (on which see n. 187 infra). As usual, however, he prefers the period after 610 in ‘the great Persian War of 602–630 (sic)’ for the disintegration of the Syro-Greek civilisation of the Limestone Massif, a periodisation not fully consistent with the discussion of epigraphy given in the present article: ‘The Near Eastern countryside in late antiquity: a review article’, The Roman and Byzantine Near East: Some Recent Archaeological Research (Journal of Roman Archaeology Supplementary Series Number 14 [Ann Arbor 1995]), 216f.

187. Cf. Butzer, K.W., ‘Der Umweltfaktor in der grossen arabischen Expansion’, Saeculum 8 (1957), 359–371.CrossRefGoogle Scholar

188. The barest hint of this may be found in the tax assessment c. 636–641 reported in the chronicle of Theodore Scutariotes. It could have been a fiscal measure to assist tax collections from unregistered migrants who had been settling in eastern Anatolia since the early 600s. On the military side of the question, see Kaegi, Byzantium and the Islamic Conquests, 256f.