§3. Applications . (a) Death of Alexander and Election of Athanasius . That the latter took place on June 8, 328, is established by the agreement of our sources, together with the numbering of the Festal Letters. Theodoret ( H. E. i. 26) and others, misled by some words of Athanasius ( Apol. c. Ar . 59), handed down to later ages the statement that Alexander died five months after the Council of Nicæa. It had long been seen that this must be a mistake (Tillemont, vi. 736, Montfaucon, Monit. in Vit. S. Athan .) and various suggestions 98 were made as to the terminus a quo for the ‘five months’ mentioned by Athanasius; that of Montfaucon remains the most probable (see ch. ii. §3 (1), p. xxi.). But the field was left absolutely clear for the precise and concordant statement of our chroniclers, which, therefore, takes undisputed possession. (Further details, supr . p. xx. sq .; Introd. to Letters , pp. 495, 303). (b) The first exile of Athanasius . The duration is fixed by the Hist. Aceph . (see Introd. p. 495, sq .) as two years, four months, and eleven days, and this exactly coincides with the dates given by the Index for his departure for Tyre, July 11, 335, and his return from exile Nov. 23, 337 (not 338; for the Diocletian year began at the end of August). Although, therefore, the Hist. Aceph . is not available for the date, the constructive agreement between it and the Index is complete. But it has been contended that the year of the return from this exile must still be placed in 338, in spite of the new evidence to the contrary. The reasons alleged are very weak. (1) The letter of Constantine II., dated Treveri, June 17, so far from making against the year 337, clinches the argument in its favour. Constantine is still only ‘Cæsar’ when he writes it (pp. 146, 272); he was proclaimed Augustus on Sep. 9, 337 (Montf. in ann . 338 tries in vain to parry this decisive objection to the later date. He appeals to Maximin in Eus. H. E . ix. 10, but overlooks the word σεβαστός there. Is it conceivable that a disappointed eldest son, as sensitive about his claims as Constantine was, would within so short a time of becoming ‘Augustus’ be content to call himself merely ‘Cæsar’?) The objection as to the distance of Treveri from Nicomedia has no weight, as we show elsewhere (p. xli., note 4); Constantine might have heard of his father’s death a fortnight before the date of this letter. (2) The law ( Cod. Th . X. x. 4) dated Viminacium, June 12, 338, if correctly ascribed to Constantius, would certainly lend plausibility to the view that it was at that time that Athanasius met Constantius at Viminacium (p. 240). But the names are so often con fused in mss. , and the text of the Theodosian Code requires such frequent correction, that there is no solid objection to set against the extremely cogent proofs (Gwatkin, p. 138) that the law belongs to Constantine, who in that case cannot have been at Trier on June 17, 338. As to Constantius, there is no reason against his having been in Pannonia at some time in the summer of 337. (3) The statement of Theodoret ( H. E . ii. 1) that Ath. ‘stayed at Treveri two years and four months’ seems to reproduce that of the Hist. Aceph . as to the length of the exile, and is only verbally inexact in applying it to the period actually spent in Trier. (4) The language of Letter 10, the Festal letter for 338, is not absolutely decisive, but §§3, 11 certainly imply that when it was written, whether at Alexandria or elsewhere, the durance of Athanasius was at an end. There can, we submit, be no reasonable doubt that the first exile of Athanasius began with his departure from Alexandria on July 11, 335, and ended with his return thither on Nov. 23, 337. (c) Commencement of the second exile . Here again the agreement of our chronicles is constructive only, owing to the loss of the earlier part of the Hist. Aceph .; but it is none the less certain. The exile ended, as everyone now admits and as both chronicles tell us, on Paoph. 24 (Oct. 21), 346: it lasted, according to the H. A ., seven years, six months, and three days. This carries us back to Phar. 21 (April 16), 339. Now we learn from the Index that he left the Church of Theonas on the night of Mar. 18–19, and from the Encyclical , 4, 5, that he took refuge first in another church, then in some secret place till over Easter Sunday (Apr. 15). This fits exactly with Apr. 16 as the date of his flight to Rome. To this there is only one serious objection, viz., that Ath. was summoned (p. 239) to Milan by Constantius after the end of three years from his leaving Alexandria. It has been assumed (without any proof) that this took place ‘just before’ the council of Sardica. As a matter of fact, Constans left Athan. in Milan, and (apparently after his summer campaign) ordered him to follow him to Trier, in order to travel thence to the Council. Athanasius does not state either how long he remained at Milan, or when he was ordered to Trier; for a chronological inference, in opposition to explicit evidence, he furnishes no basis whatever. I agree with Mr. Gwatkin (whom his Reviewer quite misunderstands) in placing the Milan interview about May, 342, and the journey from Trier to Sardica after Easter (probably later still) in 343 (Constans was in Britain in the spring of 343, and had returned to Trier before June 30, Cod. Th . XII. i. 36, see also supr . p. xlv.). A more reasonable objection to the statement of the Index is that of Dr. Bright (p. xv. note 5), who sets against its information that Athan. fled from ‘Theonas’ four days before Gregory’s arrival, the statement of the Encyclical that he left a certain church after Gregory’s outrages at Eastertide. But clearly Athan. first escaped from the church of Theonas, afterwards (between Good Friday and Easter) from some other church ( ἄλλη ἐκκλησία ), not named by him (‘Quirinus,’ cf. p. 95, note 1), and finally from the City itself. (Dr. Bright’s arguments in favour of 340 are vitiated in part by his placing Easter on April 9, i.e. on a Wednesday, instead of the proper day, Sunday, Mar. 30). The date, April 16, 339, is, therefore, well established as the beginning of the second exile, and there is no tangible evidence against it. It is, moreover, supported by the subscription to the letter to Serapion, which stands in the stead of the Easter letter for 340, and which states that the letter was written from Rome. (d) Council of Sardica and death of Gregory . The confusion into which the whole chronology of the surrounding events was thrown by the supposition (which was naturally taken without question upon the authority of Socrates and Sozomen) that the Sardican council met in 347, is reflected in the careful digest of opinions made by Newman ( Arians , Appendix, or better, Introduction to Hist. Treatises of S. Ath . p. xxvi.; cf. also Hefele, Eng. Tra., vol. 2, p. 188, sq ., notes), and especially in the difficulties caused by the necessity of placing the Council of Milan in 345 before Sardica, and the mission of Euphrates of Cologne to Antioch as late as 348. Now the Hist. Aceph ., by giving October, 346, as the date of the return of Athanasius from his second exile, at once challenged the received date for Sardica, and J. D. Mansi, the learned editor of the ‘Collectio Amplissima’ of the Councils, used this fact as the key to unlock the chronological tangle of the period. He argued that the Council of Sardica must be put back at least as early as 344; but the natural conservatism of learning resisted his conclusions until the year 1852, when the Festal Letters , discovered ten years earlier, were made available for the theological public of Europe. The date 347 was then finally condemned. Not only did Letter 18, written at Easter, 345, refer to the Council’s decision about Easter, and Letter 19 refer to his restoration as an accomplished fact; the Index most positively dated the synod in the year 343, which year has now taken its place as the accepted date, although the month and duration of the assembly are still open to doubt ( Supr . p. xlv., note 6). In any case it is certain that the Easter at which the deputies from Constans and the Council reached Antioch was Easter, 344. This brings us to the question of the date of Gregory’s death. Mr. Gwatkin rightly connects the Council which deposed Stephen for his behaviour to the Western deputies, and elected Leontius, with the issue of the ‘Macrostich’ creed ‘three years’ ( de Syn. 26) after the Council of the Dedication, i.e., in the summer of 344. This is our only notice of time for the Council in question, and it is not very precise; but the Council may fairly be placed in the early summer, which would allow time for the necessary preliminaries after Easter, and for the meeting of the fathers at reasonable notice. (Perhaps Stephen was promptly and informally deposed (Thdt.) after Easter, but a regular council would be required to ratify this act and to elect his successor.) After the Council (we are again not told how long after) Constantius writes a public letter to Alexandria forbidding further persecution of the orthodox (277, note 3). This may well have been in the later summer of 344. Then ‘about ten months later’ ( ib .) Gregory dies. This would bring us ‘about’ to the early summer of 345; and this rough calculation 99 is curiously confirmed by the precise statement of the Index xviii., that Gregory died on June 26 (345, although the Index , in accordance with its principle of arrangement, which will be explained in the proper place, puts the notice under the following year). Of course the date of the letter of Constantius, which Athanasius gives as the terminus a quo of the ‘ten months,’ cannot be fixed except by conjecture, and the date given by the Index is (1) the only precise statement we have, (2) is likely enough in itself, and (3) agrees perfectly with the datum of de Synod 26. That is to say, as far as our evidence goes it appears to be correct. (e) Return of Athanasius in 346. Here the precise statements of the Index and Hist . Aceph . agree, and are confirmed by Letter 19, which was written after his return. The date therefore requires no discussion. But it is important as a signal example of the high value to be assigned to the united witness of our two chronicles. For this is the pivot date which, in the face of all previously accepted calculations, has taken its place as unassailably correct, and has been the centre from which the recovery of the true chronology of the period has proceeded. The difficulty in dating the interview with Constantius at Antioch is briefly discussed p. xlvii. note 10. (f) Irruption of Syrianus and Intrusion of George . The former event is dated without any room for doubt on the night of Thursday, Feb. 8 (Mechir 13), 356 (see p. 301, also Index and Hist. Aceph. ). Here again the accuracy of our chronicles on points where they agree comes out strongly. It should be noted that an ill-informed writer could hardly have avoided a blunder here; for 356 was a leap-year: and in consequence of this (1) all the months from Thoth to Phamenoth, inclusive, began a day later, owing to the additional Epagomenon before the first day of Thoth: the 13th Mechir would, therefore, in these years correspond to Feb. 8, not as usual to Feb. 7. (2) Owing to the Roman calendar inserting its intercalary day at the end of February, Feb. 8 would fall on the Thursday, not on the Friday (reckoning back from Easter on Apr. 7: see Tables C, D., pp. 501 sq.). This date, then, may rank as one of the absolutely fixed points of our chronology. After the above examples of the value of the concordant testimony of the two chronicles, we must demand positive and circumstantial proof to the contrary before rejecting their united testimony that George made his entry into Alexandria in the Lent of 357, not 356. As a matter of fact all the positive evidence (supr., p. lii., note 11) is the other way, and when weighed against it, the feather-weight of an inference from a priori probability, and from the assumed silence of Athanasius ( Ap. Fug . 6), kicks the beam. (g) Athanasius in 362. The difficulty here is that Athanasius clearly returned after the murder of George, which, according to Amm. Marc. XXII. xi., took place upon the receipt at Alexandria of the news of the execution of Artemius at Antioch, which latter event must be placed in July. Therefore Athanasius would not have returned till August, 362. On the other hand the Hist. Aceph . makes George arrested four days after his return to Alexandria, and immediately upon the proclamation of the new Emperor, Nov. 30, 361. On Dec. 24 George is murdered, on Feb. 9 the edict for the return of the exiles is promulged, and on Feb. 21 Athanasius returns, to take flight again ‘eight months’ later, on Oct. 24. The difficulty is so admirably sifted by Mr. Gwatkin (pp. 220, 221) that I refer to his discussion instead of giving one here. His conclusion is clearly right, viz., that Ammianus here, as occasionally elsewhere, has missed the right order of events, and that George was really murdered at the time stated in Hist. Aceph . The only addition to be made to Mr. Gwatkin’s decisive argument is that Ammianus is inconsistent with himself, and in agreement with the Hist. Aceph ., in dating the arrest of George shortly after his return from court . As George would not have been at Julian’s court, this notice implies that the arrest took place only shortly after the death of Constantius. Moreover, George, who even under Constantius was not over-ready to visit his see, and who knew well enough the state of heathen feeling against him, would not be likely to return to Alexandria after Julian had been six months on the throne. We have then not so much to balance Ammianus against the Hist. Aceph ., as to balance one of his statements, not otherwise confirmed, against another which is supported by the Hist. Aceph ., and by other authorities as well, especially Epiph. Hær . 76. 1. (The Festal Index gives no precise date here, except Oct. 24, for the flight of Athanasius, which so far as it goes confirms the Hist. Aceph .) Moreover, “on the side of Ammianus there is at worst an oversight; whereas the Hist. Aceph . would need to be re-written.” The murder of George, Dec. 24, 361, return of Athanasius, Feb. and his flight, Oct. 24, 362, may therefore be taken as firmly-established dates. (h) Supposed Council at Alexandria in 363. This Synod assumed by Baronius, Montfaucon ( Vit. in Ann . 363. 3) and others, after Theodoret ( H. E . iv. 2) must be pronounced fictitious (so already Vales. in Thdt. l.c .). (1) The letter of Ammon (extract printed in this volume, p. 487) tells us on the authority of Athanasius that when Pammon and Theodore miraculously announced the death of Julian, they informed Athan. that the new Emperor was to be a Christian, but that his reign would be short; that Athanasius must go at once and secretly to the Emperor, whom he would meet on his journey before the army reached Antioch, that he would be favourably received by him, and that he would obtain an order for his restoration. Now (apart from the possibility of a grain of truth in the φήμη of the death of Julian) all these details bear the unmistakeable character of a vaticinium post eventum , in other words, we have the story as it was current when Ammon drew up the document in question at the request of Archbishop Theophilus (see also p. 567, note 1). At that time, then, the received account was that Athan. hastened secretly to meet Jovian as soon as he knew of his accession, and that he met him between Antioch and Nisibis. Now this native Egyptian account is transmitted independently by two other channels. (2) The Hist. Aceph . viii. tells us that the bishop entered Alexandria secretly ‘adventu eius non pluribus cognito,’ went by ship to Jovian, and returned with letters from him. (3) The Festal Index tells us that eight months (i.e., Oct. 24–June 26) after the flight of Ath. Julian died. On his death being published , Athan. returned secretly by night to Alexandria. Then on Sept. 6 he crossed the Euphrates (this seems to be the meaning of ‘embarked at the Eastern Hierapolis,’ the celebrated city, perhaps the ancient Karkhemish, which commanded the passage of the river, though some miles from its W. bank) and met the Emperor Jovian, by whom he was eventually dismissed with honour, returning to Alexandria Feb. 20, 364. Jovian was at Edessa Sept. 27, at Antioch Oct. 23. The agreement of the three documents is most striking, and the more so since the chronicles are clearly independent both of one another and especially of the letter of Ammon, as is clear from the fact that neither mentions the φήμη , while the Festal Index implicitly contradicts it. This appears to be a crucial case in many ways. Firstly , the three narratives are all consistent in excluding the possibility of any such council as is supposed to have been summoned (see above, p. lx.). Against this there is nothing but the hasty inference of Thdt. (corrected by Valois, see above, ib .); the valueless testimony of the Libellus Synodicus (9th cent.); the marvellous tale of Sozom. v. 7 (referred to this time by Tillem. viii. 219, but by Soz. to the death of George: probably an amplification of Hist. Aceph. ‘visus est’) that Athanasius suddenly to the delight of his people was found enthroned in his Church; and the more vague statement of Socr. (iii. 24) that he regained his church ‘at once after Julian’s death.’ As the three fifth-century writers are implicitly contradicted by three writers of Alexandria at the end of the previous century, the latter must be believed against the former. Secondly , the Index , the later as it appears, of the two chronicles, would seem to represent a form of the story less marvellous and therefore earlier than that of the Narratio . Now the latter certainly belongs to the Episcopate of Theophilus. The Index therefore can scarcely be placed later, and the Hist. Aceph . would fall, as Sievers, Einl . 2, had independently placed it at the beginning of the Episcopate of Theophilus. Thirdly , we have here an excellent example not only of the value of the combined evidence of the two chronicles, but also of their character as representing in many important respects the Alexandrian tradition of the last third of the fourth century. Before leaving this question it will be well to consider the dates a little more closely. Hierapolis was counted eight days’ journey from Antioch. From Alexandria to Antioch by sea was about 500 miles, i.e. with a fair wind scarcely more than four days’ sail (it might be less, cf. Conybeare and Howson, St. Paul , vol. 2, p. 376, sq . ed. 1877). This allows about twelve days for Athan. to reach the Euphrates from Alexandria, remembering that southerly winds prevail in the Eastern Mediterranean at this season (Sievers, Einl . 28). Now Athan. reached Hierapolis on Sept. 6 (Thoth 8, Egyptian leap-year). But according to the Index , he reached Alex. after Julian’s death was published, and this according to Hist. Aceph . was on Mesori 26, i.e. Aug. 19. From that day to Sept. 6 are eighteen days, leaving about a week’s margin for Ath. to hear the news, reach Alexandria, and perhaps for delay in finding a vessel, &c. But a far wider margin is really available, for the official announcement must have been preceded by many rumours, and was probably not despatched till more than a fortnight after Julian’s death (as is observed by Mr. Gwatkin, p. 221). If we remember that Athanasius, according to the Letter of Ammon, was making all possible haste ( supra , §9) we shall again realise the subtle cohesion of these three sources, and the impossibility of the ‘large Synod’ imagined by some historians for the year 363. (k) Exile under Valens . The date of this is discussed by Tillem. ( note 96) and Montf. Vit . who, on the unstable basis of a computation of Theophanes (about 800 a.d. ) and of the vague and loose sequences of events in Socr. and Sozom., tentatively refer the exile to the year 367. The only show of solid support for this date was that Tatianus (of later and unfortunate celebrity), whom the Photian Life and that by the Metaphrast connected with the expulsion, was known from Cod. Theod . to have been Prefect of Egypt in 367. But this airy fabric now gives place to the precise and accurate data of the Theophilan chronicles. Both Index and Hist. Aceph . place the occurrence not under Tatian but under Flavian, governor of Egypt 364–366. Both fix the year 365. The Hist. Aceph . (used by Soz. vi. 12, who however makes no use of the dates) gives May 5, 365, for the Imperial order against bishops restored by Julian, June 8 for the reference to the Emperor ( supra , ch. ii. §9), Oct. 5 for the retreat of Athan. and search for him by Flavian and Duke Victorinus , Feb. 1 for the return of Athanasius. This detailed chronology is corroborated in two ways; first by a letter of Libanius ( Ep . 569) to Flavian, thanking him for a present of [Egyptian] doves, and congratulating him on his ‘victory’ (a play on the name Victorinus is added), but with a satirical hint that if only Victorinus had any prisoners to shew for his pains (a clear allusion to the escape of Ath.) he (Libanius) would think him a finer fellow even than Cleon (Siev. Einl . 31). Secondly, the restoration of Ath. by Valens becomes historically intelligible, in view of the danger from Procopius, as pointed out supr . p. lxi., fin . We cannot then doubt that the chronicles are here once more the channels of the genuine chronological tradition. (1) Death of Athanasius . It is superfluous to discuss this date at the present day, but it may be worth while to point out for the last time how admirably the combined testimony of our chronicles confirms the judgment of the best critics (Montfaucon, Tillemont, &c.) antecedent to their discovery, and how clearly the secondary value to be assigned to the chronological statements of Socrates and Sozomen once more comes out (Socr. iv. 21 puts the date at 371, and was followed by Papebroke, Petavius and others (fuller details and discussion of the question on its ancient footing in Newman’s preface to Hist. Tracts of St. Athan ., pp. xx., sqq .). But no one any longer questions the date of May 2–3, 373. The fact that the Hist. Aceph . gives May 3 and the Index May 2 (the date observed in the later calendars) vouches for the independence of the two documents and for the very early date of the former: probably, as Sievers and others suggest, the true date is the night between May 2 and May 3. 98     E.g. that he died five months after his return home from the council (Tillem.), or after the reconciliation of Meletius (Montf.). As neither event is dated, both hypotheses render the ‘five months’ useless for chronology. 99     The above resumé of the details of the evidence makes it clear that Mr. Gwatkin’s alleged oversights are in reality those of his critic. The proposal of the latter to correct ‘Epiph.’ in Fest. Ind. to ‘Pharmuthi’ is especially gratuitous.