liturgy ilgainst the Cirocks, by disciples of Methodius, who utilised facts which they had learned from him. The Lives were also intended to serve theological instruction ; to teach the Bulgarians methods of pologetic and controversy (against Jews, Saracens, and the Latin Church). We cannot regard is historical the disputations (in V'.(.) with John the ex-Patriarch or with the Mohammadans; and the arguments against the Jews and Khazars ille the work of the biographer. Brückner dwells on what ho calls schematism in the missions to the Mohammadans, the Khazar's, and the Moravians; in each cise Constantine is represented as being sent by the Emperor. The Mohammadan cpisode is unhistorical, the others are historical ; but the purt assigned to the Byzantine government is probably a misrepresentation of fact.

But incidental bits of information, not necessity to the writer's pragmatical purposes, ille trustworthy with some reservations. We may accept the statement about the parentage of the apostles, the patronage accorded to Constantine by the logotheto (Theoktistos), his appointment its librarian of the Patriarch. llis friendship with Photins is known from Anastasius.

If he was :ppointed librarian by Photins, the date could not be earlier than 859, and it would follow that, if the order of events in 1'.('. is correct, the visit to the Khazar's could har:lly have been carlier thiin 860. But we can hardly accept the statement that he wils ciucated with the son of Theophilus, for he was at least ten years older than Michael III.'

APPENDIX XII

THE MAGYARS

1. Date of the Second Magyar Migration (to Atelkuzu) WESTBERG has put forward a new view as to the date of the migration of the Hungarians to Atelkuzu (in K anal. ii. 49-51) which he places c. A.D. 825. His argument is based on a passage in Constantine, De alm. imp. 175, relating to the four sons and four grandsons of Arpad. The descent may conveniently be represented in a table.

Salmutzes (Almus)

Arpad

Tarkatzus

Telukla
Iutotras

Zuillas
1
1
1

1
Teles
Ezelekli
Phalitzis

Taxis
.

(Phales)

(Taser) Termntzuis

When Constantine was writing (A.D. 950-952), Phalitzis was the Hungarian king (Tùy vuvi úp xortu), Teleles was dead, and his son Tormatzi18 was adult and hnd recently visited Constantinoplo on an embilssy (é úpríws uvedbwv pídos mistranslated by W'estberg, its by most others). Westberg infers that Tebeles died not later than 945, and that the surviving grandsons of Arpad, Phalitzis and Taxis,” were advanced in years. Reckoning thirty years to it gonerition, ho goes on to place the death of Tarkatzus about 915, that of Arpul c. 88.5, that of Salmutzes c. 855. At the time of the elevation of Arpad, Salmutzes was alive and considered (by · Lebedias) capable of ruling the Magyar nation. Therefore the election of Arpad must belong to :he second quarter of the ninth century, not later than A.1), 850. But the migration to Atelkuza occurred not long before Arpad's election (1) wilm. imp. 169.4); so

Il lato pointed this out in B.l. xv. wlio, loe thinks, Wres the chilest son of 602.

Arpaul (15.%. vi. 687-688). But the grassage ? | assume that Taxis and Tuses are implies that Tieses Bias been already incora the sime. l'erz, however, lits conjectural tioncil, and the identification withi Taxis that Times was it son of Liuntis or Levente seems inevitabile,

#SO

11

w the

presence of the Magyars in Atelkuzni covers tho poriod from approximately 825 to 89.5."

This argumentation carries no conviction. We can readily accept 885 as the approximate date of Arpad's death, for c. 889 his son Levente (who is not mentioned in this passage) was king. But this does not necessitate the inference that Arpad was olected before 850, or even before 860. Suppose that he was sixty years old when he died; then he would have been born in 825. Suppose that Salmutzes, his father, was then twenty-five years old, he would have been sixty, a "bodrii starik,” in 860. This hypothesis, which might be varied (there is no reason to suppose that Arpad was old when he died; he may have been much younger than sixty), is sntlicient to show that Westberg's reasoning is arbitrary, and that the data admit of no such conclusion as he drilw's.

Our fixed dato ante quam for the first migration of the Magyars is A:D. 862, the year in which they invaded the empire of the Franks, for it is improbable that this invasion was undertaken before they had settled west of the Dnieper. Our fixed date post quem is the time of the visit of Constantine the Philosopher to Cherson and the Khazars, which we can only define approximately as before A.D. 863 (see above, p. 396). At that time, as we learn from the lila Constantini, the Magyars were still in the neighbourhood of the Crimea. Although there are many unhistorical details in this Vita, the cpisode of the Hungarians evidently preserves a genuine fact, for when the l'itu was written the Hungarians were filr a will, and no inventor of fiction would have dreamed of introducing them on the scene. Westberg (il. 51) admits the genuineness of the notice, but seems to think that the Hungarians in vided the Crime:t from Atelkızı. This is possible, but less probable; once they left their old seats, they were not likely to return cross the Dnieper and trespass on the hunting grounds of the Patzinaks, whom they drealed.

As the mission of Constantine wils probably about A.1), 860, we en deduce A.1). 860-861 as it probablu date for the first historicul migration of the Magyars. Their second migration, to their abiding home, occurred izbout 89.5, so that their period in Atelkuzi was about forty years. The election of Arpad may bo placcel roughly about A.1), 860.

The appearance of the Magyars west of the Dnieper c. A.1). 837 (see above, p. 371) proves only that, as we should expect, they made predatory expeditions into Atelkuzi long before they occupied it.

2. Dute of the First Magyar Migration (to Lebeilia) The question of the date of the migration of the Magyar's into their earlier home between the Don and Drioper is more difficult.

According to Constantino (op. cit. 168) they called this territory Lebediu, after tho namo of their most important tribal leader, Lebedias. I tako this to mean that in later times, when they were in Atelkuzu and Hungary, they described this territory, having no other name for it, as the country of Lebodias--the country which they ussociated with his leadership. According to the text of Constantine, ib., they occupied this country, on the borders of the land of the Khazars, for three years (évuautols Tpeis). This is certainly an error; and we can indeed refute it from Constantine himself, who goes on to say that during this period the Magyars fought for the Khazar's “in all their wars," a statement which naturally presupposes a much longer period. The probability is that there is a textual error in the number. Westberg (ib. 51) proposes to read Tpukurtu Tpeis or tpuskurtu. If we adopted the

τριάκοντα . former, which is the less violent, correction, we should obtain c. 822-826 as the date of the arrival of the Magyars in Lebedia.

It must be considered doubtful whether they had como to Lebellia from beyond the Caucasus, where there were Magvars known to the Armenians as the Sevordik. See above, p: 410. Constantine indeed says that they were still known by this name (Supptog udou) in Lebedia. It is true that the troubles which distracted Armenia and the adjacent regions in the reign of Mamun (sce the account of Yakubi, apud Marquart, Streifzüge, 457 $74.) might have forced a portion of the Suvordik to seek a new habitation under the protection of the Khazars.

We can say with certainty that the Magyars diil not arrive in Lebedia ilt it later period than in Mamun's reign, and there is perhaps a probability that if they had been there long before that period, some indication of their presence would have been preserved in our sources. The conjectural restoration of Constiintino's text (thirty-three years) cannot be relied on; but it mily be noted that the Bulgarian Warfare on the Dnieper in Omurtay's reign (seo above, p. 366), if it was provokud by the prosence of the Magyar's, would be chronologically compatible.

Constantine clocs not tell as the source of his information about the Magyars and their earlier history. We can, however, form i probable opinion. While he wis ongged in writing his treatise known as De suministru nulo imprio, or just before he had begun it, an Hungarian embassy arrived at Constantinople (referred to above, p. 489) consisting of Termat zus, it grandson of Arpad. and Bultzus, who hold the dignity of karchas (the third dignity in the realm, after the king and the gvliis). It seems very likely that Constantine derivad much of what he tells us about the. Mayvars from this friendly embassy. Compare my paper on “The Treatise De alm. imp." 1.2. xv. 562-5633.

3. The names Magyur, 1/ungarian, Turk While they were in Lebeilia, thu llungurians soem already to havo called theriselves Mayvaurs, for they were known by this name to in Arabic writer (efore A.1), 850), who reproduced it as Buzhghar (cp. Marquart, op. cit. 68).' In their own ancient chronicles the name appear's as Moyor. It is obviously identical with the name of one of their tribes, the Meyépn, mentioned by Constantine,

We may conjecture that this was the tribe of which Lebedias was chieftain, and that his pre-eminence was the cause of its becoming it lime for the nation.

To the Slavs and Litins, the Magyars were known by the more comprehensive nitne of the Ugrian race, to which they belonged : Ingri, whence Ilungari; and the Greek chronicle, which describes their appearance west of the Dnieper in the reign of Theophilus, likewise calls them Oiyypou (cholil

, Cicurge 818). But this designation in it Greek writer of the ninth and tenth centuries is exceprional, for the Cireeks regularly applied to them the term Toipkot, und even in this pilnsage they are also called Torpkou 3 and Ovrou. Why did the Greeks call them Turks? The simplest Lliswer is that the milme come into use after the union of the Miyvars with the Kabar's who were Torks.

Marynart has pait forward in ingenious but hardly convincing explanation of Torpaille .llo iilentities it with the 'liprint of Herodotus 1. ...., who seem to appear in Pliny, vi. 19, ils Tyrrur, and in Pompons Meln, i. $ 11.6, ils Turcu. Je supposes that Iurkui

. is the silme word ils lugra, Vyrian, with metalesis of 1; that the word afterwards acquired an initial / in Scythian dialects, and that the Cireeks borrowed it from the Alans us a designation of the Maglar's (op. cit. 51 377.) before their imion with the Kabar's, According to this theory, the Turks are fillso. "Turkw," and the Wag Vill's ille U'NO " Turkin," incrording to the original denotition of the name: in fact, the Ugrini namne, in its Saythian formi, calmo in the course of history 10 bo transferred from the Vyrian to the Turimi: mcr.