arrangement is the beginning of an administrative connexion between Thrace and the East; they would now be governed by the same Prætorian Prefect.

Praxagoras (F. H. G. iv. p. 3) : Ελλαδος τε και Μακεδονιάς και τῆς κατω (ita leg. pro Kara) Arias were acquired by Constantine. Anon. Val. 18; Licinius: orientem, Asiam, Thraciam, Moesiam, minorem Scythiam.

V. A.D. 335. [The arrangement of this year was not a division of the Empire, but partly a confirmation of the assignment of administrative spheres, already made to his sons, and partly a new assignment of administrations to his nephews. Constantine did not directly sacrifice the unity of the Empire, which was still realized in his own sovereignty, though he adopted a policy which might at any moment endanger it. "Von einer Erbtheilung ist dabei nicht die Rede, sondern nur von einem Antheil an der Verwaltung" (Ranke, Weltgeschichte, iv. 2, 270).]

(1) Constantine had Gaul, Britain and Spain ( the later "Prefecture of Gaul"); (2) Constantius, Asia and Egypt; (3) Constans, Italy, Africa, and Illyricum (including Thrace). For Delmatius the ripa Gothica was cut off from the portion of Constans; Hannibalian had (at the expense of Constantius) a "kingdom" composed of principalities in the regions of Pontus and Armenia.

The question is, what were the limits of the province of Delmatius? Is ripa Gothica [I have not seen noticed a parallel expression in De Mortibus, 17, where Galerius reaches Nicomedia, per circuitum ripa striga, where the emendation Istrica is doubtless right] to be interpreted as Eastern Illyricum (=dioceses of Dacia, Macedonia, and Thrace)? So Schiller (ii. 235), Ranke, Burckhardt and others. But the Epitome of Victor (41, 20) includes in the share of Constans "Dalmatia, Thrace, Macedonia and Achaia". Ranke supposes that Dalmatiam here is a scribe's mistake for Dalmatius, and that we should interpret the ripa Gothica of the Anonymous by the words thus amended. If we adopted this view, it would be better to read: Dalmacius Daciam Thraciam Macedoniam Achaiamque.

But a view that necessitates tampering with a text which in itself gives perfect sense cannot be accepted as satisfactory. There is a further objection here. The text of the Epitome agrees remarkably with the statement of Zonaras, xiii. 5, which assigns to Constans Italy, Africa, Sicily and the islands, Illyricum, Macedonia, "Achaia, with the Peloponnesus". The Epitome was not a direct source of Zonaras; but the agreement is explained by the fact they both (the author of the Epitome directly, Zonaras indirectly) drew from a common source (probably Ammianus: cp. L. Jeep, Quellenunt. zu den gr. Kirchenhistorikern, p. 67). Thus the assumption of a textual error in the Epitome means the assumption of an error in the text of an earlier authority; and therefore becomes decidedly hazardous and unconvincing. Add to this that the interpretation of ripa Gothica to include or to imply Macedonia and Greece is extremely forced. The natural meaning of the expression is: the provinces of Dacia, Moesia I. and II. and Scythia,' and perhaps Pannonia and Noricum. The actual testimonies of the two best authorities, that are explicit, concur in showing that the main division of A.D 335 was tripartite-between the Emperor's three sons-and that only subsidiary (though highly responsible) posts in frontier regions were given to the two nephews. This view is also more in accordance with Zosimus, ii. 39, who distinctly marks a triple division. Nor is it contradicted by Eusebius, Panegyr. ch. iii., which only proves that Delmatius (unlike Hannibalian) was a Cæsar, and thus co-ordinate in dignity with his cousins.

VI. A.D. 337-8. (1) Constantius: as before, along with the kingdom of Hannibalian, and the four provinces of D. Thrace, south of Haemus; 3 (2) Constans: as

1 Chron. Pasch., p. 532, ed. B. gives Mesopotamia to Delmatius (Godefroy accepted the statement). I conjecture that Μεσοποταμίαν may have arisen from Μυσιαν παραποταμιαν Moesiam ripensem.

2 He pretends to mark it as it existed at the death of Constantine (before the destruction of Delmatius); though he seems really to give the subsequent division.

3 The dates in the early edicts of the C. Th. are not certain enough to permit us to draw an inference from xi. 1, 4 (professedly issued by Constantius at Thessalonica in November 337). 36

VOL. II.

before, along with ripa Gothica, including Moesia II. and Scythia; and without (?) Raetia or part of Africa; (3) Constantine: as before, along with some part of Africa or of the Diocese of Italy (?).

We have not data for determining the details of this partition. The problem was to divide the provinces held by the two nephews into three parts. To secure geographical continuity Constans would naturally take the ripa Gothica, and hand over some part of his western dominions to Constantine; he likewise resigned Thrace south of Haemus (not Moesia and Scythia, I infer from Zos. ii. 39, who gives to Constans and Constantine τὰ περὶ τον Εύξεινον πόντον) to Constantius The war which broke out between Constans and Constantine was probably connected with the question of the territorial compensation to be received by the latter; seeing that Zos. ii. 41, ascribes it to a dispute about Africa and Italy.

Gibbon (with Tillemont) has accepted from the Chron. Alex. of Eutychius a curious notice (under Ol. 279) that Constantine_the younger reigned for a year at Constantinople. The only possible support I can see for this statement must be derived from the passage of Zosimus. He groups together the lands of Constantine and Constans, as if they ruled jointly over an undivided realm, in which he includes "the regions of the Euxine". A defender of Eutychius might urge that for some months at least Constans did not assert his independence, that his elder brother may have governed for him, and that the transference of Thrace to Constantius may have been subsequent. But without further evidence it is better to leave the Eutychian notice aside; and I may call attention to Ranke's remark that there is a tendency in the account of Zosimus, who desiring to justify Magnentius is hostile to Constans and anxious to throw on him the blame for the war with Constantine.

The division of 338 A.D. is given as follows in the Life of St. Artemius (Acta Sanct., Oct. 20)—a document which merits more criticism than it has received:

(1) Constantine: αἱ ἄνω Γαλλίαι καὶ τὰ ἐπέκεινα ̓Αλπεων (an expression often used to include Spain), ai re BpеTтavikai vñσol (Britain and the Orcades, etc.? cp. Eutropius 7, 13, and the interpolation in the Laterculus of Polemius Silvius, see above, App. 11), καὶ ἕως τοῦ ἑσπερίου ὠκεανοῦ. (2) Constans: αἱ κάτω Γαλλίαι ήγουν ai Irakia (Italy with its adjuncts, Sicily, Africa, etc.), ka avτn ʼn Póμn. (3) Constantius : τὸ τῆς ἀνατολῆς μέρος, Βυζάντιον, τὰ ἀπὸ τοῦ ̓Ιλλυρικοῦ (implying that Illyricum went to Constans) μέχρι τῆς Προποντίδος όποσα ὑπήκοα τοῖς Ρωμαίοις την το Συρίαν και Παλαιστίνην και Μεσοποταμίαν και Αἴγυπτον και τας νήσους ἁπάσας.

The Vita Artemii (the Greek text was first published by A. Mai in Spicilegium Romanum, vol. iv.) was composed by "John the Monk," and professes to be compiled from the Ecclesiastical History of Philostorgius and some other writers. Eusebius, Socrates and Theodoret are also referred to. There is evidence that Philostorgius was largely used, and consequently the Life of Artemius becomes an important mine of material for the restoration of the history of that Arian writer. The story of Gallus is, I presume, derived from him, and I conjecture that the statement of the partition of the Empire among the sons of Constantine comes from the same source. If so, both passages ultimately depend on Eunapius, who was doubtless the source of Philostorgius.

From the same source is certainly derived the statement of the partition in Constantine Porphyrogennetos, de Them., ii. 9 (ed. Bonn, p. 57). The portion of Constantine is described in exactly the same words as in the Vita Artemii (τὰς ἄνω Γαλλίας και τα ἐπέκεινα Αλπέων ἕως τοῦ ἑσπερίου Ωκεανοῦ,, except that instead of "the British Isles" the imperial geographer says as far as the city of Canterbury itself" (Κανταβριν). The expression αἱ κατω Γαλλίαι is also used, but, in expanding the concise expressions of his source, Constantine falls into error and assigns Illyricum and Greece to Constantius.

66

VII. A.D. 364. (1) Valentinian i.: Prefectures of Gaul, and of Italy and Illyricum; (2) Valens: Prefecture of the East, including D. of Thrace.

VIII. A.D. 378. (1) Gratian and Valentinian ii. Prefectures of Gaul and of Italy, including Western Illyricum: (2) Theodosius: Prefecture of the East, along with Dioceses of Dacia and Macedonia (Soz. vii. 4).

This partition, which drew a new line of division between East and West,

probably established definitely the system of four prefectures which Zosimus attributed to the express enactment of Constantine. Up to this time three pr. prefects seem to have been the rule, four an exception. But now, instead of adding Eastern Illyricum to the large Prefecture of the East, Theodosius instituted a new Prefecture.

16. THE SARMATIANS-(P. 216)

It is often asserted that "Sarmatian" was a generic name for Slavonic peoples. It is certain that a great many Slavonic tribes must have been often described under the name, but it is extremely doubtful whether any of the chief Sarmatian peoples-the Bastarnae, the Roxolani (? Rox-alani) or Jazyges-were Slavonic. I believe that Šafarik, in taking up a negative position on this question, was right (Slawische Alterthümer, ed. Wuttke, i. 333 sqq.). But I cannot think that he has quite made out the Slavonic race of the Carpi (ib. 213-4), though this is accepted by Jireček (Gesch. der Bulgaren, p. 77); he has a more plausible case, perhaps, for the Kostoboks. On the other hand it is extremely likely, though it cannot be absolutely proved, that in the great settlements of non-German peoples, made in the third and fourth centuries in the Illyrian peninsula by the Roman Emperors, some Slavonic tribes were included. This is an idea which was developed by Drinov in his rare book on the Slavic colonization of the Balkan lands, and has been accepted by Jireček. There is much probability in the view that Slavonic settlers were among the 300,000 Sarmatae, to whom Constantine assigned abodes in 334 A.D. It is an hypothesis such as, in some form, is needed to account for the appearance of Slavonic names before the beginning of the sixth century in the Illyrian provinces.

Safarik tried to show that the Alani, Roxolani, Bastarnae, Jazyges, &c., were of Iranian race, allied to the Persians and Medes,-like the Scythians of Herodotus.

17. BATTLE OF SINGARA—(P. 227)

I have shown in the Byzantinische Zeitschrift (vol. 5) that we should accept Julian's notice as to the date of this battle (and place it in A.D. 344), instead of following Jerome's date (adopted by Idatius), A.D. 348. One might be tempted to guess that there were two battles at Singara, and that the nocturna pugna was placed in the wrong year by an inadvertence of Jerome; this might be considered in connexion with Forster's reconstruction of the corrupt passage of Festus, Brev. ch. 27: Verum pugnis Sisaruena, Singarena, et iterum Singarena praesente Constantio ac Sicgarena, &c. The VUKTopaxía is described below as: nocturna Elliensi prope Singaram pugna. Elliensi is mysterious.

The events of the Persian wars of Constantius and Julian are briefly narrated by General F. R. Chesney in his Expedition for the Survey of the Rivers Euphrates and Tigris, vol. 2, p. 430 sqq. (quarto ed.).

18. SOURCES AND CHRONOLOGY OF ARMENIAN HISTORY
UNDER TRDAT AND HIS SUCCESSORS-(C. XIX.)

Some works bearing on Armenia have been mentioned in connexion with general oriental history in vol. i. Appendix 13. In addition to these must now be mentioned (besides St. Martin's Mémoires sur l'Arménie and the notes to his edition of Lebeau's Bas-Empire): Ter Mikelian, Die armenische Kirche in ihren Beziehungen zur byzantinischen (saec. 4-13), 1892; Chalatianz, Zenob of Glak (in modern Armenian; known to me through Stackelberg's summary in Byz. Zeitschrift, 4, 368-70), 1893; and above all Gelzer's highly important essay, Die Anfange der armenischen Kirche (in the Ber. der kön. sächs. Gesellschaft der Wiss.), 1895.

1. Sources. (a) Faustus. For Armenian history in the fourth century after death of Trdat (Tiridates), a.d. 317, our only trustworthy source is Faustus, who

wrote his History of Armenia in Greek (before the Armenian alphabet was introduced; the Greek original is quoted by Procopius, Pers. i. 5), probably in first years of King Vram Sapuh, who reigned from 395 to 416 (Gelzer, p. 116). The work is marked by enthusiasm for the clergy, and a certain prejudice against the policy of those who were loyal to the kings, also by chronological errors. Faustus is completely a national Armenian; therein lies his strength and his weakness" (ib. 117). He consulted official documents in the royal archives (ib.) and made use of old songs. It is announced that H. Gelzer and L. Babajan will issue a translation of Faustus, and Gelzer's name is a guarantee that it will be trustworthy. (b) Agathangelos, who lived about half a century later, contains a work which is our only good source for the reign of Trdat. His work (preserved both in Armenian and in a Greek translation, which mutually check each other) has been dissected by A. von Gutschmid (Kleine Schriften, 3, 395, sqq.). It contains an earlier Life of St. Gregory (perhaps originally composed in Syriac, Gelzer, p. 114) and an Apocalypse of Gregory written between 452 and 456 by a priest of Valaršapat. The latter is valuable as throwing indirect light on the church history of the fifth century, but worthless for the history of Trdat. (c) The conclusion of Carrière (mentioned in vol. i. App. 13) that the date of Moses of Chorene is very late (beginning of eighth century) is accepted by Chalatianz and Gelzer, and seems to be established. (d) The worthlessness of the History of Taron by Zenob of Glak has been shown by the investigation of Chalatianz (op. cit.). Hitherto supposed to have been written in Syriac in the fourth century and translated into Armenian in the seventh, it is now shown to be an apocryphal work of an impostor of the eighth or ninth century. There is a French translation by Langlois, F. H. G. vol. v.

2. Chronology. The student who consults the translation of Langlois (Agathangelos and Faustus; op. cit.) must be warned that the chronological indications in the notes are set down at random and contradict one another. And, if he has read the note in Smith's edition of the Decline and Fall, vol. ii. p. 369, which is taken from St. Martin's edition of Lebeau, and compares it with the chronological list of kings in the same scholar's Mémoires, he will find that the two accounts diverge. (In the Mémoires, p. 412-3, the dates are: death of Trdat, 314; interregnum; accession of Chosroes II., 316; Tiran II., 325; Arsaces, 341; Pap, 370. According to the old view, which appears, though not consistently, in Langlois' collection, and seems to be assumed in Ter Mikelian's op. cit., Trdat reigned from 286 to 342.) The following reconstruction seems most probable :Death of Chosrov I., accession of Trdat, Accession of Chosrov II.,

261 A.D.

317

"

There are not sufficient data for determining the dates of the Catholici; the statements of Moses will not bear criticism, see Gelzer, p. 121 sqq. The only certainties we have are that Aristakēs, son and successor of Gregory, attended the Council of Nicæa, 325; and that Nerses was poisoned by King Pap before 374.

3. Trdat and Constantine (Gelzer, 165 sqq.). Officially the Armenian kings adopted the style "Arsaces" (just as the Severian Emperors adopted Antoninus), and he appears in Cod. Theod. xi. i. 1 (Constantine and Licinius A. D. 315) as Arsacis regis Armenia. In the previous year, he and Gregory visited Constantine in Illyricum ("the land of the Dalmatians" in the Armenian Agathangelos) in "the royal city of the Romans," probably Serdica. There the alliance mentioned by Faustus (iii. 21; Langlois, p. 232) was concluded, which endured till 363. The authenticity of the account of Agathangelos (doubted by Gutschmid) has been successfully vindicated by Gelzer.

On Trdat's death the Romans intervened to put Chosrov on the throne, and Tiran likewise owed his elevation to Constantine. In 337 he was betrayed to the Persians by his chamberlain, seized by the governor of Atropatene, and blinded. The armed intervention of Constantine and Constantius led to the elevation of

Aršak, the son of Tiran, who declined to resume the sovereignty. Aršak first married Olympias, a Greek lady connected with the Constantinian house; and afterwards a daughter of the Persian king. His policy was to hold the balance between Rome and Persia throughout the wars of Constantius and Julian.

4. In Eusebius, H. E. vi. 46, 2, we find this notice: kai roîs karà 'Apμevíav ὡσαύτως περὶ μετανοίας ἐπιστέλλει ὧν ἐπεσκόπευε Μερουζάνης, Gelzer (p. 171 sqq.) points out that this bishopric of Meruzanes cannot have been in the Roman provinces called Armenia, and therefore was in Great Armenia; and he seeks to show that it may have been in the south-eastern corner, the district of Vaspurakan. The words in Eusebius are from a letter of Dionysios of Alexandria (248265), and the inference seems to be that Christianity was introduced into an outlying district of Armenia in the fifties of the third century. But the formal conversion of Armenia began about 280 under the auspices of King Trdat, through the labours of Gregory the Illuminator. The destruction of the temples of the gods, in spite of strong opposition from the priests, was one of the first parts of the change, and preceded Gregory's journey to Cæsarea (between 285 and 290 according to Gelzer) to be consecrated by Leontius. The Armenian Church was dependent on the see of Cæsarea, and under Greek influence for nearly a century. After the death of the Patriarch Nersēs, it was severed and made autocephalous by King Pap (circa A.D. 373-4. Cp. Ter Mikelian, p. 31). During the fourth century the seat of the Catholicus and the spiritual centre of Armenia, was Aštišat in the southern district of Taron, as has been well brought out by Gelzer. It was afterwards removed to Valaršapat, when no longer dependent on Cæsarea, and then the priests of Valaršapat invented stories to prove the antiquity of their seat and the original independence of the Armenian Church. In the fourth century, the chief feature of the domestic history of Armenia is the struggle between the monarch and the Catholicus, between the spirit of nationality and the subjection to foreign influences. It culminated in the reign of Pap, who solved the question by poison.

În regard to the conversion of Armenia, its progress was partly determined by the feudal condition of the country (Gelzer, 132). The nobles were easily won over by the personal influence of the king; the priests were naturally the most obstinate opponents. The new faith seems to have been slow in taking root among the people, and it is noteworthy that women, even in high rank, clung tenaciously to the old religion (like the wife of Chosrov, Faustus, iii. 3, and the mother of Pap, ib. 44).

I have read with interest the remarkable study of N. Marr, O nachalnoi istorii Armenii Anonima, in Viz. Vremennik, i. 263 sqq. (1894). He discusses the character of the brief History of Armenia, which is prefixed to Sebeos' History of the Emperor Heraclius (Russ. tr. by Patkanian, 1862); and its relation to Moses of Chorene. This document (which appears in the collection of Langlois under the title Pseudo-Agathange) he regards as the earliest extant Armenian history of early Armenia; it was worked up by a later (also anonymous) writer, of whose composition a large extract has been preserved in Moses of Chorene, bk. i. c. 8 (in Langlois, under the title, Mar Apas Catina). Moses also used the original work. Marr points out a number of resemblances between Faustus and the first Anonymous, and hazards the conjecture (295 sqq.) that this history of Armenia may be part of the first two books of Faustus, whose work, as we have it, begins with book iii.

19. CONSTANTINE AND CHRISTIANITY—(C. XX.)

The attitude of Constantine to the Christian religion has been the theme of many discussions, and historians are still far from having reached a general

1 My friend Mr. F. C. Conybeare is inclined to believe that Gregory the Illuminator used an Armenian version of New Testament Scriptures made from a pre-Peshito Syriac text, long before the time of Mesrop. This version may have been due to the Church in Vaspurakan. Apparently the non-existence of Mesrop's alphabet did not prevent literary composition in Armenian.