TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 Seven Byzantine Revolutions and the Chronology of Theophanes Warren Treadgold FROM 685 TO 717, perhaps the most obscure period of Byzantine history, the Empire suffered unprecedented instability. A combination of repeated revolutions and poor evidence makes the course of events extraordinarily difficult to determine. For hardly any other time could one seriously argue, as I shall here, that the standard accounts of Byzantine history have systematically misdated almost every event during thirty years. For this period our most important source is the Chrono­ graphia of Theophanes Confessor, completed between 813 and 815.1 According to Ostrogorsky, Theophanes dates all his entries between 609 and 715 and 725 and 773 a year too early. Ostrogorsky's redating and the account that he based upon it in his History have won wide acceptance. They have influenced the relevant parts of nearly every study or reference work pro­ duced over the past sixty years, though a few individual dates have been challenged.2 No one, however, has reexamined the evidence as a whole, which seems to indicate that Theophanes is right about the great majority of his dates between 685 and 715, and that even when he is wrong his mistakes were of a different character than Ostrogorsky supposed. 1 On Theophanes see H. Hunger, Die hochsprachliche profane Literatur der Byzantiner I (Munich 1978) 334-39; I. S. Cicurov, Vizantijskie Istoriceskie SoCinenija (Moscow 1980). The most recent edition of Theophanes is that of c. de Boor, Theophanis Chronographia (Leipzig 1883). 2 See G. Ostrogorsky, History of the Byzantine State 2 (New Brunswick 1969) 87ff (esp. 88 n.1), 129-46, 152-56. The most detailed recent treatment of the years 685-711 appears in A. N. Stratos, Byzantium in the Seventh Century V (Amsterdam 1980). J. B. Bury, A History of the Later Roman Empire from Arcadius to Irene (London 1889) 320-30, 352-86, still provides in many respects the best account because, unlike Ostrogorsky and Stratos, Bury accepts Theophanes' chronology and avoids interrupting the course of events at 711 simply because the Heraclian dynasty then died out. 203 TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 204 BYZANTINE REVOLUTIONS AND THEOPHANES Ostrogorsky had treated Theophanes' chronology at length some ten years before the first editon of his History appeared in 1940. 3 He concluded that Theophanes' dates are a year early for two long periods, from the Alexandrian annus mundi (A.M.) 6102 (609/10) to A.M. 6207 (714/5), and again from A.M. 6218 (725/6) to A.M. 6265 (772/3). Although other scholars had already noticed problems with Theophanes' chronology in parts of the seventh and eighth centuries, Ostrogorsky was the first to offer a clear, precise, and comprehensive solution. Most of his argu­ mentation is sound, and for the second period and the part of the first to A.M. 6176 (684/5) his conclusions seem justified. But Theophanes' systematic error of a year need not imply that his revised dates are correct. Here and elsewhere he made various mistakes, repeating some from his sources and intro­ ducing others as he reworked his material into annual entries. Similarly, one or two correct dates in Theophanes would not necessarily show that he was not systematically misdating events, because he might have made miscalculations that can­ celed each other out. But the Alexandrian dating he used was so obsolete that he is not likely to have found it used in his sources for this period; indeed if he had done so, he would presumably have detected his original miscalculation. 4 Confirming a series of dates in the Chronographia would therefore suffice to dis­ prove Ostrogorsky's thesis for that part of the chronicle, even if an occasional date is wrong and others cannot be proved or disproved. Although Theophanes' uniqueness as the only evidence for many events of the seventh and eighth centuries and their dates presents difficulties, other sources do permit some control, especially the Breviarium of Patriarch Nicephorus, compiled probably in the 780·s. Nicephorus shared a source with The­ ophanes, who copied it more extensively, but Nicephorus sometimes includes facts omitted by Theophanes. Nicephorus also composed a briefer Chronographia, which clarifies the J Ostrogorsky, "Die Chronologie des Theophanes im 7. und 8. Jahrhun­ dert," ByzNeugrJb 17 (1930) 1-56; cf the review by F. Dolger, BZ 31 (1931) 351-55; see also the recent treatment in A. S. Proudfoot, "The Sources of Theophanes for the Heraclian Dynasty," Byzantion 44 (1974) 367-439 . .. On Theophanes' use of the Alexandrian era see C. Mango, "The Tradition of Byzantine Chronography," Harvard Ukrainian Studies 12-13 (1988-89) 367-70. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 205 chronology by recording the duration of emperors' reigns. s Several Arabic, Syriac, and Latin sources also provide occasional dates for comparison with those of Theophanes. When Ostrogorsky wrote, no one recognized the importance of the so-called Necrologium, recording the day and month of the death or deposition of each emperor and the length of each reign; from these notices the exact dates of reigns can usually be determined. Although the complete Necrologium survives only in an awkward, corrupt Latin translation in the thirteenth­ century Chronicon Altinate, Philip Grierson has demonstrated that the Greek original of the section of the Necrologium for the years 306 to 963 must have been of high accuracy. In a note appended to Grierson's article, Cyril Mango and Ihor Sevcenko identified the original text as a previously lost chapter from the tenth-century De cerimoniis of Constantine VII Porphyrogenitus. They also showed that part of this chapter survives in an almost illegible palimpsest. 6 Consequent­ ly the Necrologium is an official source, presumably based on the imperial archives. The imperial chancery had a special need for accurate records of the first and last days of reigns because state documents were dated by regnal years. The original dates in the Necrologium should therefore have been quite reliable. 1. Theophanes' Error and its Rectification Insofar as Theophanes' dates to 685 can be checked, they support Ostrogorsky's thesis. The last such date, Theophanes' entry for A.M. 6176, should correspond to the year beginning 1 September 683 and ending 31 August 684. Under this year Theophanes (p.361 de Boor) records a peace treaty between Constantine IV (668-685) and the Caliph 'Abd aI-Malik (685-705). The rulers' dates indicate that this treaty could only 5 On the Breviarium see text, translation, and commentary by C. Mango, Nikephoros, Patriarch of Constantinople: Short History (Washington 1990), esp. 2-19 for analysis of Nicephorus' historical works. For the text of the Chronographia see C. de Boor, Nicephori Archiepiscopi Constantinopolitani Opuscula Historica (Leipzig 1880) 79-135; also see Hunger (supra n.1) 344-47. 6 P. Grierson, "The Tombs and Obits of Byzantine Emperors (337-1042)," with an additional note (including readings from the palimpsest) by C. Mango and I. Sevcenko, DO P 16 (1962) 1-63; for the Latin text of the Necrologium see R. Cessi, Origo Civitatum Italiae seu Venetiarum (Chronicon Altinate et Chronicon Gradense) (Rome 1933) 104-14. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 206 BYZANTINE REVOLUTIONS AND THEOPHANES have been concluded in 685~ and the Syriac chronicler Elijah of Nisibis dates it precisely to 7 Tammuz A. H. 65 (7 July 685). Because Elijah was dependent on eastern sources~ this was doubtless the date of the final ratification of the treaty by the Caliph. 7 Since this date (and presumably the somewhat earlier date of the emperor's ratification) fell during A.M. 6177 (1 September 684-31 August 685), Theophanes placed the treaty one year too early. For A.M. 6177 Theophanes (p.361) gives only one event, the death of Constantine IV and the succession of his son Justinian II. Ostrogorsky dated Constantine's death to the beginning of September 685 on the weak authority of the Liber Pontificalis, the only source he knew. 8 If this were the actual date of the death rather than the date at which the death became known at Rome, Theophanes would have placed it a year too early. The Necrologium (108 Cessi) states~ however, that Constantine IV died on 10 July after a reign of seventeen years-therefore 10 July 685, because he succeeded to the throne on 15 July 668. 9 Theophanes (pp.353, 361) and Nicephorus' Chronographia (99 de Boor) agree that seventeen years was the length of Constan­ tine's reign. Theophanes has therefore corrected his error~ although dividing the events of A.M. 6177 between two entries. His entry for A.M. 6177 immediately suggests how the chronicler corrected himself. After two lines on Constantine's death, Theophanes (pp.36lf) includes a chronological excursus of a page and a half~ mainly concerned with the effective date of the decrees of the Sixth Ecumenical Council. The Council convened on 7 November 680 and adjourned on 16 September 681.10 Theophanes dates it correctly to the twelfth year of Constantine IV (15 July 680-14 July 681) but incorrectly to A.M. 6172 (1 September 679-31 August 680)~ thus making his old mistake of counting an annus mundi too early. He then lists the number of years in several imperial reigns: of Constantine IV 7 See the Latin translation of E. W. Brooks, Eliae Metropolitae Nisibeni Opus Chronologicum I (=CSCO 63, Syr. 23 [Louvain 1910]) 71f. B Ostrogorsky (supra n.3) 32; cf L. Duchesne, ed., Liber Pontificalis I (Paris 1886) 366. 9 Grierson (supra n.6) 49f; for an explanation of why the N ecrologium seems to date the death of Constans II to 5 November, see infra n.20. 10 See the dating formulas given by the acts of the council in G. Mansi, Sacrorum Conciliorum Nova et Amplissima Collectio XI (Florence 1765) 208E-209A (7 November, indiction 9), 624D-E (16 September, indiction 10). TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 207 after the Council, Justinian II for the first time, Leontius, Tiberius III Apsimar, and Justinian II for the second time. Subsequently, Theophanes cites an edict from Justinian's reign relating to 15 January of the fourth indiction (706). The­ ophanes misidentifies this indiction as A.M. 6199 (1 September 706-31 August 707), thus counting an annus mundi too late, the reverse of his usual error. Calculating from his year too early for the Council to his year too late for the edict, he states that the period between them was 27 years rather than the actual 25 (681-706). Finally he lists the years in office of the yatriarchs of Constantinople from the Sixth Ecumenical Counci to his own time, ending with the Patriarch Tarasius (784-806). The editor of Theophanes, Carl de Boor, bracketed this entire chronological excursus as an interpolation because in its Greek version it continues the sequence of patriarchs down to John VII the Grammarian (838-843), after Theophanes' death. As I have noted elsewhere, however, the interpolation consists only of the list of patriarchs after Tarasius, which assigns erroneous terms to their patriarchates and is absent from the ninth­ century Latin translation of Theophanes by Anastasius Bibliothecarius.ll The rest of the excursus is by Theophanes, as appears from his characteristic miscalculations of the Alex­ andrian annus mundi. In the course of this discussion, The­ ophanes rectifies his chronology without correcting his earlier error. He probably dated the edict of January 706 too early because, realizing that he had to add a year to correct his calculations, he added a year to a date that was already right. Although this excursus inspires little confidence in Theopha­ nes' arithmetic, his dates for events from 685 cease to be a year ahead of dates that can be checked. After the death of Constan­ tine IV, dated by the Necrologium, the next clearly verifiable date is a campaign against the Slavs during which Justinian II advanced to Thessalonica. Theophanes records it under A.M. 6180 (1 September 687-31 August 688). In an inscription found in 1885 at Thessalonica, Justinian mentions his visi t to the city and gives the Church of St Demetrius possession of a salt 11 Treadgold, "The Chronological Accuracy of the Chronicle of Symeon the Logothete for the Years 813-845," DOP 33 (1979) 178f with n.87; cf. Treadgold, The Byzantine Revival: 780-842 (Stanford 1988) 441 nA06. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 208 BYZANTINE REVOLUTIONS AND THEOPHANES works beginning with September of the second indiction, which in his reign could only be September 688,12 This inscription agrees perfectly with Theophanes, indicating that Justinian ordered the grant to take effect soon after his campaign, which evidently occurred in the summer of 688. But if Ostrogorsky's thesis held good for this entry, the expedition would have begun later than the inscription recording it. Finally and decisively, Theophanes (p.367) mentions a solar eclipse on Sunday 5 October of A.M. 6186 (1 September 693-31 August 694), and an eclipse did occur on Sunday 5 October 693. Ostrogorsky's chronology would put this event on 5 October 694, a Monday, and months away from any eclipse. 13 Thus three verifiable dates from Justinian II's first reign support Theopha­ nes against Ostrogorsky, while none supports Ostrogorsky against Theophanes.14 II. The Revolution of 695 By 695 Justinian II had imprisoned some aristocrats in the Praetorium for as long as eight years. He had kept Leontius, formerly his principal general, jailed for three of those years, probably as punishment for the Arab victory at Sebastopolis in Anatolia (now to be dated 692 rather than 693). Theophanes absurdly reports that by 695 the emperor planned to massacre the whole population of Constantinople beginning with the Patriarch-presumably an invention of those who overthrew Justinian. is The emperor's worst error was to release Leontius and appoint him Strategus of Hellas in 695, after thoroughly arousing the resentment of the aristocracy. 12 J.-M. Spieser, -Inventaires en vue d'un recueil des inscriptions historiques de Byzance I: Les inscriptions de Thessalonique," Tra'llMem 5 (1973) 156-59. 13 Cf V. Grumel, La Chronologie (Paris 1958) 462,316. 14 The simplest and probably the right solution to the apparent confusion in Theoph. p.363 (A.M. 6178=685/86) is to assume that all the events in this entry did take place between 1 September 685 and 31 August 686, but that the treaty mentioned at the beginning of the entry was made after the campaign mentoned at the end of the entry. A good case can in fact be made for transferring lines 26-32 to the beginning of the entry. 15 Theoph. 368f records the prisoners' years in the Praetorium with the story of Justinian's plans for the massacre; Theoph. 365f records the defeat at Sebastopolis under A.M. 6184 (1 September 691-31 August 692). TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGO LD 209 An exact date of Justinian's deposition by Leontius is not recorded: the Necrologium (108) mentions only the date of his death at the time of his second deposition; Theophanes (361, 363) and Nicephorus' Chronographia (99) put Justinian's first reign at ten years, a round number yielding an approximate date of July 695. Current opinion, relying (in the absence of evidence) on Ostrogorsky's theory, dates the deposition after 1 September 695. Theophanes (368f) describes it under A.M. 6187 (1 September 694-31 August 695), just after a raid by MouuJl£O on Armenia IV. This raid can be identified with that of Mu­ ~ammad ibn Marwan on the region of Melitene, just across the Euphrates from Armenia IV, recorded by the Arab chronicler Ibn al-AthTr under A.H. 76 (21 April 695-9 April 696). 16 Since the Arabs usually raided in summer, this should have been the summer of 695. So if Theophanes is right, Justinian was deposed about August 695. The detailed account given by Theophanes, somewhat differ­ ently excerpted from their common source by Nicephorus (Brev. 40), may even reveal the exact date of the revolution. The night before the Constantinopolitans were to be massacred and Leontius was to embark for his new command in Greece, he and his fellow conspirators seized the Praetorium and released the prisoners who had been held there for so long. These then dispersed throughout Constantinople shouting, "All Christians to St Sophia!" Early in the morning a crowd gathered before the church, where the Patriarch Callinicus greeted them with the words, "This is the day which the Lord hath made" (Ps. 118.24). They cursed Justinian and proclaimed Leontius emperor. Justinian's nose was slit to disqualify him from ruling, and he was exiled to Cherson in the Crimea. This episode resembles the later revolution of 713 when, on the morning after the Emperor Philippicus was blinded, a crowd gathered at St Sophia and proclaimed Anastasius II the new emperor. In 713, however, according to both Theophanes (383) and Nicephorus (Brev. 48) the day of the gathering of the crowd was Pentecost. Another revolution just before a great feast of the Church occurred a century later in 820, when conspirators murdered Leo V the Armenian early on Christ­ mas morning, and his successor Michael II the Amorian was 16 See E. W. Brooks, "The Arabs in Asia Minor (641-750), from Arab Sources," ]HS 18 (1898) 190 with n.2. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 210 BYZANTINE REVOLUTIONS AND THEOPHANES hailed and crowned in St Sophia on Christmas day.17 Apparently in these two cases the conspirators planned to take advantage of the crowd that would naturally come to St Sophia for the morning liturgy. The people could then acclaim the new ruler and witness the coronation there. In 695 the meeting at St Sophia and especially the Patriarch's choice of words seem also to indicate that the day was a major feast of the Church. The one important summer feast celebrated at this time was the Dormition of the Virgin on 15 August. Neither Theophanes nor Nicephorus is careful to note when events fell on church feasts; neither mentions that the Arab siege of Constantinople of 717-718 ended on the Feast of the Dormition, although both date the Arab withdrawal to 15 August and contemporaries regarded the coincidence as a sign of the Virgin's intervention to save the city.18 Like the scheduled massacre, Leontius' scheduled departure on that day is probably an elaboration. Thus Leontius may well have de­ posed Justinian on 15 August 695, and in any case did so in late summer, before 1 September. III. The Revolution of 698 Although Leontius was a capable and reasonably popular ruler, his status as an outright usurper made him highly vul­ nerable to rebellions. Probably realizing this, he avoided warfare as much as possible. When the Arabs took Carthage, however, and were about to end more than 150 years of Byzantine rule in Africa, Leontius had to act to avoid a devastating blow to his prestige. The large expedition he sent, though led by a fully trustworthy commander, soon became the instrument of his deposition. Again, an exact date for this revolution is unknown. The N ecrologium records only the restored Justinian II's later execution of Leontius. Here, as in omitting the date of Jus­ tinian's deposition, the Necrologium shows its dependence on the archives, for after his restoration Justinian would have in- 17 See Treadgold (1988: supra n.ll) 223ff with n.307. 18 Theoph. 399; Niceph. Bre'll. 56. Cf a sermon on the Feast of the Dormi­ tion given ten years later by the Patriarch Germanus: V. Grumel, MHomelie de Saint Germain sur Ie delivrance de Constantinople," REByz 16 (1958) 183-205. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 211 structed the chancery to ignore the revolutions of 695 and 698 as events without legal standing. At least the Necrologium (108), Theophanes (361, 369), and Nicephorus' Chronographia (99) agree in assigning Leontius a reign of three years, while Ni­ cephorus (Brev. 41) notes that Leontius was still in his third regnal year when he fell. Theophanes (370f) records Leontius' deposition under A.M. 6190 (1 September 697-31 August 698), giving an account of the year's events paralleling that of Ni­ cephorus. According to this account, Leontius, hearing that the Arabs had seized Carthage, sent a great naval expedition under John the Patrician that retook the city and wintered there. This evidently happened during the winter of 697-98. But the Arabs soon counterattacked, recaptured Carthage, and sent John and his armada fleeing to Crete. There the fleet rebelled and acclaimed as emperor one of its officers, Apsimar, whose apparently German name it replaced with the imperial­ sounding Tiberius III. While bubonic plague broke out in Con­ stantinople, to last for four months, Apsimar's fleet sailed to the capital and put it under siege. The city's population supported Leontius and held out for some time (£1tt Xp6vov bE nvu in both Theophanes and Nicephorus); but a few officers betrayed part of the wall. Apsimar led his men into the city, slit Leontius' nose, relegated him to a monastery, and ruled as emperor. All this must have happened between the autumn of 697 and mid-August of 698, since Leontius failed to complete his third year as emperor. The four months of plague seem to cor­ respond roughly to the time between Apsimar's proclamation on Crete and his capture of Constantinople. John must have sailed to Carthage in the fall of 697 and held it through the winter. He would have fled Carthage about March of 698 and arrived on Crete about April, when Apsimar was proclaimed. Then Apsimar besieged the capital approximately from May to July, when he took the city and became undisputed ruler. Again we have no reason to doubt Theophanes, or to follow Ostrogorsky in putting the accession of Apsimar after 1 Sep­ tember 698. IV. The Revolution of 705 Apsimar had even less justification for seizing power than Leontius, and after the loss of Africa the new usurper's success TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 212 BYZANTINE REVOLUTIONS AND THEOPHANES against the Arabs was mixed. Justinian II, despite his dis­ figurement, began plotting at Cherson to regain the throne. He escaped to the Khazar Khanate, where he married a Khazar princess and renamed her Theodora after the wife of Justinian I. Then, forced to flee to the Bulgar Khanate, he obtained an army from the Bulgar Khan Tervel. When Justinian led this force to Constantinople, he found the people unwilling to receive him; but after three days he made his way into the city through a broken aqueduct and resumed his interrupted reign. Apsimar was eventually captured at Apollonia in Thrace, and Leontius was taken from his monastery. Justinian celebrated his restora­ tion at games in the Hippodrome, where he sat in the imperial box with one foot on Apsimar and one on Leontius before having the two usurpers beheaded. The Necrologium (108) dates the deposition of Apsimar to 21 August and the executions of Leontius and Apsimar to 15 February, while agreeing with Theophanes (361, 371) and Nicephorus' Chronographia (99) on a seven-year reign for Apsimar. Theophanes (374f) dates Justinian's entry into Con­ stantinople to A.M. 6197 (1 September 704-31 August 705) and his capture of Apsimar to A.M. 6198 (1 September 705-31 August 706). Thus Justinian evidently returned to Constantinople on 21 August 705, but beheaded Apsimar and Leontius only on 15 February 706. Grierson (supra n.6: 51) has proposed that Justinian waited to execute them until after displaying them at the regular consular games in January 706. But since Justinian could easily have held special games if he had wished, this would still imply that Apsimar had been captured not long before, perhaps in December. Grierson at first accepted the dates in the N ecrologium for the death of Constantine IV and the deposition of Apsimar, but later concluded that one of them had to be wrong. He observed that if both dates are right Justinian, having succeeded his father on 10 July 685, would have been restored on 21 August 705, in a year that the emperor would have counted, ignoring his time in exile, as the twenty-first of his reign. But some bronze folIes of Justinian are dated with a regnal year that Grierson reads as xx. He concluded that "Since the evidence of the coins cannot be gainsaid, one has the alternatives of ignoring either 21 August as the date of the deposition or 10 July as that of Justinian's ac­ cession." As the right date for the death of Constantine IV he TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGO LD 213 suggested 4 November, which the Necrologium wrongly gives as the death date of Constantine's predecessor Constans II.19 Almost certainly, however,S November is the date of deposi­ tion of Constans' predecessor Heraclonas, which appears to be Constans' death date only because most of Heraclonas' entry has dropped from our text. 20 Furthermore, some of the coins in question show a crowned bust of Justinian's infant son Tiberius. According to both Theophanes (375) and Nicephorus (Brev. 42), Tiberius was brought from the Khazar Khanate to Con­ stantinople and crowned only after the executions of Leontius and Apsimar, presumably when the sailing season began in the spring of 706. That Tiberius' coronation could have occurred within twenty years of Justinian's accession is irreconcilable not only with the dates of the Necrologium but with any plausible reading of the other sources. As Grierson notes, the folIes he mentions have a mark after the XX that earlier numismatists read either as I or V and no one has explained otherwise. 21 Since either XXI (10 July 705-9 July 706) or XXV (10 July 709-9 July 710) would be perfectly compatible with the sources, the ambiguous numis­ matic data should be interpreted in light of the literary evidence. The restoration of Justinian in 705 raises an apparent contra­ diction between Theophanes and the N ecrologium that has no bearing on the chronology but needs to be resolved. Theopha­ nes (427f) mentions that Bishop Theodosius of Ephesus who presided at the iconoclast Council of Hieria in 754 was the son of Apsimar. The Necrologium (109) records that the emperor Theodosius III, who was deposed in 717, subsequently became a celebrated bishop of Ephesus: M ense iulii, die xxiiii, proiectus est Theodosius, qui et Adra­ mitinus, ex imperio a Leoncio [read: Leo n e] Isauro et effectus est clericus cum filio suo, etiam et episcopus Ephesi, et mortuus est ac sepultus in templo Sancti Phylippi [sic], in antiq ua urbe, iuxta portum, faciens mirabilia in sepultura. 19 P. Grierson, Catalogue of the Byzantine Coins in the Dumbarton Oaks Collection and in the Whittemore Collection 11.2 (Washington 1968) 624 with nA. 20 See Treadgold, "A Note on Byzantium's Year of the Four Emperors (641)," BZ 84 (1991), forthcoming. 21 Grierson (supra n.19) 654 no. 11,655 no. 11 bis, 656 no. 12a.5. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 214 BYZANTINE REVOLUTIONS AND THEOPHANES Two different men of imperial blood, both prominent bishops of Ephesus in the early to middle eighth century, arouse suspi­ cion. Graham Sumner supposed that the Emperor Theodosius III was the son of Apsimar, but both Theophanes (385) and Nicephorus (Brev. 50) say that before Theodosius became emperor he was a tax collector-in Mango's translation of Ni­ cephorus, "an ordinary man uninvolved in politics" (u1tpa:y­ )lOVa nva Kat tOtro'tTlV).22 The chroniclers' early source would hardly have referred to an emperor's son in this way, nor could Justinian II have allowed his executed predecessor's son to live without at least making him a cleric. In this connection the limitations of the N ecrologium should be recognized. Although its dates are highly reliable, its additional reports are less so. For example, the N ecrologium (106) repeats the myths that the Emperor Zeno (474-91) was buried alive by his wife and the Emperor Anastasius (491-518) was struck dead by lightning. The tenth-century compilers of the De cerimoniis seem to have found only the dates of reigns in the state archives, so that they had to compose the rest of the Necrologium from the inscriptions on imperial tombs, most of which they could see in the Church of the Holy Apostles, and whatever incidental information they found. In the case of Theodosius III, who was not buried as an emperor, they probably learned that a tomb existed at Ephesus of a Bishop Theodosius of imperial blood who had died in the eighth century. They apparently knew little else about the tomb, since they failed to realize that its supposed miracle­ working occupant had died an iconoclast heretic. Misled by the coincidence that Apsimar's son and Theodosius III had the same surname, the compilers identified the bishop with the emperor. Even in the Latin translation of the N ecrologium the words beginning with etiam et episcopus Ephesi look like a later addition, and they cannot have had anything to do with the original record of Theodosius Ill's deposition, made some forty years before the bishop's death. Since Theophanes should therefore be right, Justinian allowed Apsimar's young son to enter the church unharmed-except perhaps by castration. This 22 G. Sumner, • Philippicus, Anastasius II and Theodosius III," G R BS 17 (1976) 29lf; Mango (supra n.5) 119. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 215 act of clemency is a sign that Justinian's behavior in his second reign was not so recklessly vindictive as his enemies alleged and most modern historians have believed. V. The Revolution of 711 Overthrowing rulers had become a habit, and the memory of the revolutions of 695, 698, and 705 could not be erased. Jus­ tinian still had enemies-some not eliminated in the purges following his restoration. Turmoil in the Empire stimulated the Arabs to begin bolder raids on Anatolia, and these in turn helped to discredit Justinian. Nonetheless, his final downfall resulted from trouble in remote Cherson that he had done little or nothing to cause. Theophanes (377-81), in an account paralleled by Nicephorus (Brev. 45), records at length the events leading to Justinian's death, all dated to A.M. 6203 (1 September 710-31 August 711). First, Justinian sent an expedition to Cherson under Maurus the Patrician, who easily occupied the city. Although Theophanes says that Justinian ordered a general massacre of the Cher­ sonites because they had betrayed him during his exile, this is obviously another fabrication of Justinian's enemies. Since The­ ophanes notes that when Maurus arrived Cherson was under a Khazar governor, Maurus' orders were probably to retake the city after a Khazar invasion and then to punish collaborators. Theophanes reports that to Justinian's delight all adult Cher­ sonites were killed and Maurus' entire expedition of 73,000 was lost on its voyage back to Constantinople in October. The exaggeration is transparent, since the Chersonites and Maurus reappear later in the story.23 But presumably Maurus did exe­ cute some Chersonites (Theophanes describes 27 executions) and lose some men to shipwreck as he returned in October 710. According to Theophanes, the Chersonites invited the Khazars back, and Justinian sent a second, smaller expedition to Cherson under George the Patrician, whom the Chersonites killed. They then proclaimed Philippic us Bardanes emperor, whom Justinian had previously exiled to Cherson. After this 23 Nicephorus seems to have noticed one of these inconsistencies. because he omits Maurus' name from his account of the first expedition. mentioning only Maurus' co-commander Stephen Asmictus. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 216 BYZANTINE REVOLUTIONS AND THEOPHANES the emperor sent another expedition to Cherson, again under Maurus, who after failing to subdue the Chersonites joined them in proclaiming Philippicus. Meanwhile Justinian led an ex­ pedition to Sinope, which Theophanes connects with the revolt at Cherson but was more probably directed against the Arabs raiding Anatolia. 24 Justinian turned back when he learned that Maurus and Philippicus were sailing on Constantinople; but Philippicus arrived at the capital first. He killed Justinian's little son and heir Tiberius and sent troops to Justinian's camp in Bithynia. Justinian was beheaded when his army deserted him. Although the Latin translation of the Necrologium (108) dates the death of Justinian to 24 November, the Greek text (this time legible in the palimpsest) reads 4 November.25 Which is corrupt? The Greek, as the original language, might seem to have the stronger claim. Yet in both Latin and Greek notation the number 24 (xxiiii, 1()') is much more likely to be misread as 4 (iiii, ()' ) than the reverse; and there is no certainty that the text of the palimpsest was always better than the Greek text from which the Latin text is descended. In fact, as we shall see in considering the revolution of 713, the length of Philippicus' reign recorded by the Syriac chronicler Michael Syrus indicates that Philippic us became emperor be­ tween 18 November and 18 December 711. Michael, who as a Monophysite had a special interest in the Monothelete Philippicus, correctly dates Philippicus' rebellion to the year 1022 of the Seleucid era (1 October 710-30 September 711) and Philippicus' accession to A.S. 1023 (1 October 711-30 October 712).26 Michael's evidence therefore allows us to choose 24 rather than 4 November as the real date of Justinian's death and the correct reading in the text of the N ecrologium. Since Theophanes (361, 374) and Nicephorus' Chronographia (100) both state that Justinian's second reign lasted six years, the date of his death should have been 24 November 711. On the other hand, if we subtract the length of Philippicus' reign in months as given by Theophanes from the date of the Philip­ picus' deposition we arrive at a date between 18 August and 18 24 Cf Brooks (supra n.16) 193. 25 Grierson (supra n.6) 62 with n.3, 50£. 26 See J.-B. Chabot, Chronique de Michel Ie Syrien, Patriarch Jacobite d'Antioche (1166-1199) II.3 (Paris 1904) 479; cf n.31 infra. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 217 September 711. Sumner has persuasively argued that this was the date of Philippicus' proclamation in Cherson by Maurus' expedition.27 Theophanes must therefore be wrong to record Justinian's death under A.M. 6203, but he is not wrong in the way that Ostrogorsky's redating would require. As Theophanes shows by his October date for Maurus' shipwreck, which must be October 710, his entry for A.M. 6203 does begin with the events following 1 September 710. The second exredition, under George, should then be dated to the spring 0 711, because it was overly dangerous for a fleet to sail the Black Sea in winter: Maurus had already found it dangerous in October. The third expedition, again under Maurus, should belong to the summer of 711. Thus Theophanes, instead of confusing A.M. 6203 with A.M. 6204, has continued to recount some events that occurred after 1 September 711 as if they had occurred before that date. Evidently Theophanes believed that Justinian died before 1 September. But why? We have already seen that he began reckoning Philippicus' reign from a date between 18 August and 18 September 711, actually the date of Philippicus' procla­ mation by Maurus. But Theophanes had no reason to take the view that that Philippicus' reign began when he held only Cherson. The most likely solution is that Theophanes found in his source a date between 18 August and 18 September for Philip­ picus' accession, which he mistook either for the date when PhiIippicus captured Constantinople or for the date soon afterward when Justinian died. If so, the exact day, which Theophanes did not include in his text, must have been in late August, because it persuaded him that Justinian's death had taken place before 1 September. We can reasonably conclude that the date of Philippicus' proclamation was in late August, and that Theophanes, rather than miscalculating all his dates consistently, misunderstood a single date in his source. 27 Sumner (supra n.22) 287f. But note that by confusing Maurus' first and second expeditons, Sumner misdates Maurus' shipwreck to October 711 rather than October 710. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 218 BYZANTINE REVOLUTIONS AND THEOPHANES VI. The Revolution of 713 Philippicus Bardanes should have known that his hold on power was shaky after so many revolutions, particularly because the Arabs were making ever more alarming inroads into Anatolia. Yet Philippicus was a Monothelete, like many of his fellow Armenians, and he could not restrain himself from hastily repudiating the Sixth Ecumenical Council, which had condemned Monotheletism. By this time few Byzantines were Monotheletes and most resented Philippicus' attempt to revive the recently resolved controversy. After 712, when the Arabs and Bulgars both raided far inside Byzantine territory without effective opposition, Philippicus was ripe for deposition after a reign shorter than any earlier usurper's. Theophanes (383), echoed by Nicephorus (Brev. 48), records Philippicus' fall under A.M. 6205 (1 September 712-31 August 713). The Count of the Opsician Theme George Buraphus, then stationed in Thrace, sent to Constantinople a band of soldiers who abducted and blinded Philippicus on 3 June 713, the day before Pentecost. Doubtless George expected to be proclaimed emperor by the crowd that gathered the next day in St Sophia, but instead they proclaimed the Protoasecretis Artemius, who took the imperial name of Anastasius II. To punish usurpation, Anastasius soon blinded and exiled both George and the officer who had blinded Philippicus. Nicephorus, compatibly with Theophanes' implied date, re­ ports that Philippicus' blinding occurred in his second regnal year. For some reason the Necrologium (108) records the date not of Philippicus' blinding but of his later death on 20 January, perhaps of 714; but the text mentions the earlier blinding and notes that Philippicus reigned a year and a half-as he had done by 3 June 713. Theophanes' entry headings (382, 383) and Ni­ cephorus' Chronographia (100) both assign Philippicus a reign of two years, apparently rounding their figures upward. But later in his account Theophanes (386) unexpectedly assigns Philippic us a reign of 2 years and 9 months. Michael Syrus gives yet another length for the reign, 2 years and 6 months. 28 At first, Ostrogorsky relied on Theophanes' second reckon­ ing of the length of Philippicus' reign and dated the blinding to 28 Chabot (supra n.26) II.3 479. TREADGOLD, WARREN, Seven Byzantine Revolutions and the Chronology of Theophanes , Greek, Roman and Byzantine Studies, 31:2 (1990:Summer) p.203 WARREN TREADGOLD 219 the day before Pentecost of 714, assuming that Theophanes' annus mundi was again a year early.29 But Ostrogorsky changed his mind when he realized that a contemporary, the Char­ tophylax of St Sophia Agatho, dates Philippicus' blinding to the day before Pentecost of the eleventh indiction (1 September 712-31 August 713), thus of 713.30 The date of 3 June 713 is consequently secure. As Sumner suggested, the figures of The­ ophanes and Michael Syrus probably resulted from chroniclers' or copyists' altering correct figures to agree with the round number of two years that some sources gave for Philippicus' reign. 31 The original figures in the sources of Theophanes and Michael Syrus (and perhaps in their own texts before later copyists changed them) should thus have been 1 year and 9 months and 1 year and 6 months. Counting back from 3 June 713, these figures yield 3 September 711 and 3 December 711, with a margin of half a month either way. As already shown, the former result implies that Philippicus was proclaimed at Cherson in late August of 711, while the latter indicates that Justinian was killed on 24 rather than 4 November 711. VII. The Revolution of 715 By the time of Anastasius II, probably the ablest ruler in years, all order in the succession had been lost and the Arabs were preparing a full-scale attempt to take Constantinople. Ele­ ments of the Opsician Theme, the empire's largest military unit, appear also to have resented Anastasius' blinding of their Count George Buraphus, whom they must have expected to become emperor. Although Anastasius made considerable progress in fortifying his capital and putting in supplies for a siege, the Opsician Theme nonetheless brough t him down. 29 Ostrogorsky (supra n.3) 33f, giving the date as 27 May 714. 30 For Agatho's letter see Mansi (supra n.l0) XII 193 A-B. Cf Ostrogorsky's review of E. J. Martin, History of the Iconoclast Contr