Constantine always give them, and also by the name Oûvvol, Huns, which was constantly given to them in the eleventh and twelfth centuries, the Greek historians, lovers of antiquity, wishing to identify them with the people of Attila. The pas

sage also enables us to determine approximately the date of the Hungarian occupation of Atelkuzu. Allowing Arpad a long reign, and supposing that his son had been in power a good while before the collision with Simeon in 895, we see that they cannot have been long in that land when they attacked the Macedonian exiles in 836. If we said that the migration from Lebedia must have taken place about 830, we might not be far astray.

We may now see what further light, by way of expansion or correction, may be obtained from Arabic writers. Ibn Dasta, a Persian, writing in Arabic according to the prevailing custom, composed a work before 913 A.D., in which he gives a valuable account of the Khazars, the Bulgarians, the Magyars, and other peoples of those regions.* It is to be observed that he knows the Magyars by that name-Mazgars,—unlike Constantine. Five of his statements have a special importance for us: (1) The first district of the Magyars lay between the Patzinaks and the Eszegel Bulgarians,—as the Arabs called the Bulgarians of the Volga. They were bounded by the 'Roman Sea' and two rivers. (2) They reign over Slaves, and the Khazars fear them. They bring Slavonic slaves for sale to the Roman town of Kerch. (3) They are Turks. (4) They are fireworshippers. (5) They have two rulers over them, a

kende and a dsila.

It is clear that Ibn Dasta is describing the Magyars in their second home, Atelkuzu, not in Lebedia. For they live on the Black Sea, and have relations with the Empire, and are in a Slavonic country, ruling over Slaves. The rivers are conjectured by Hunfalvy to be the Dnieper and the Bug. 'The tributary of the Magyars,' he writes (Magyarország ethnographiája, p. 203),

* All students of early Magyar and Bulgarian history are deeply indebted to the work of Chwolson entitled Izviestiya o Chozarach, Burtasach, Bolgarach, Madyarach, Slavanyach i Rusach, which contains a translation of the important evidence of Ibn Dasta.

'might be Slaves dwelling on the banks of the Bug and Dniester, or perhaps still further west, of the Pruth and Seret, or even in Siebenbürgen.' Vámbéry thinks that one river is the Dnieper or Don and the other the Danube.

The first statement of Ibn Dasta is most important for determining the whereabouts of Lebedia; in fact it is the only distinct evidence that we have. If the first, clearly the most north-easterly, district" of the Magyars of Atelkuzu marched on the land of the Patzinaks, who had succeeded the Magyars in Lebedia, it is clear that Lebedia was west of the Volga. This is confirmed by the circumstance that in the reign of Theophilus, as Greek historians record, the Khazars built the fort of Sarkel on the Don as an outpost against the Patzinaks. We may guess that Lebedia corresponded partly to the western part of the province of Kazan, and that its boundaries were on the north-east, Black Bulgaria, on the south-east, the Khazar territory.

In Constantine, however, we find something which seems inconsistent with this conclusion. We have seen that in his account of the Magyars he does not say where Lebedia is. But in his chapter on the Patzinaks he states that their old dwelling, before they came to Atelkuzu, was between the rivers Itil and Geêch. There is no doubt whatever that the Itil is the Volga, and that the Geêch is the Jaik or Ural. This statement, then, does not agree with our conclusion that Lebedia was west of the Volga. It seems likely that Constantine here † mixed up the land which the Patzinaks occupied before they entered Lebedia with Lebedia itself, the land which they held during most of the ninth century. This mistake was the easier as there is no doubt that a portion of the Patzinaks had remained behind in their eastern home. In fact, as Vámbéry has well shown, the split of the Patzinak race began before the end of the ninth century (p. 84). It is significant that Constantine does not mention Lebedia in his special

* Vámbéry (p. 105) says simply a Magyarok földje'; Hunfalvy more accurately 'a Magyarok első vidéke' (p. 201).

The expression he uses (cap. 37 init.), тò åπ' ȧpxês 'originally' suggest this.

chapter on the Patzinaks. This is characteristic of the whole treatise. He makes statements in one place which do not tally with statements in another, and the difficulty of reconciling them never seems to occur to him. The question arises, where did he get information about these barbarian peoples of Eastern Europe? It seems evident to me that he got some of it at least from Slavonic sources, perhaps from Slavonic interpreters. This comes out undesignedly, without the intention of the imperial writer. When he speaks of the voevodes of the Turks,' it is clear that he is using the words of an informant to whom it was as natural to use the Slavonic' voevode' as for a Greek to say 'archon.' For the Magyars themselves did not call their chieftains voevodes. Again, when he tells us that the Patzinaks took oaths according to their zakona, one must suspect some Slavonic mediation. But it is clear that he did not get all his knowledge from the same source. His accounts of the Patzinaks and his accounts of the Turks seem to be founded on different stories-not necessarily incompatible, but he at least did not combine them. We have had an instance of this in the matter of Lebedia, but there is another still more striking instance in the same connexion. In his chapter on the Hungarians, he knows them only by the name of Turks and the 'older' name Sabartoasphaloi; but in the chapter on the Patzinaks, he states that this people, in their old home, were neighbours of the Uzes and the 'Magyars,' without apparently having the slightest notion that Mášapoi, Madyars, was the true name of the people whom he calls Turks.* These Magyars were, of course, those who had migrated eastward after the Patzinak conquest of Lebedia, and became neighbours of the eastern Patzinaks.

The likelihood that Constantine used Slavonic sources will support an acute guess of Zeuss as to the strange name Sabartoasphaloi. He thought that the word is nothing else than Svarte Falen, that is schwarze Falen, 'swart Falians.'† The

* There can be no doubt, I think with Vámbéry, that the Mášapo are the Magyars of the East. Otherwise Roesler and Hunfalvy (who thinks it a mistake for Khazars).

+ Approved by Roesler and Hunfalvy, rejected by Vámbéry.

Germans in their Latin chronicles used the name Fali to designate the Kumans in later times, and it seems to be the same word as Polovtzi, by which they go in the Russian Chronicle. May not the Scandinavians have used that name in the tenth century to designate the Magyars? May they not have called the Magyars in the east the 'swarthy Falians,' as opposed to the Magyars of the west, who would be the White Falians, just as Black and White Bulgaria are distinguished? Sabartoasphaloi would thus be a Graecized form of a Teutonic appellation, and Constantine would have been mistaken in regarding it as the name which the Turks' had themselves applied to the whole people before they split. He could easily have heard the name from Slavonic subjects of the Scandinavian princes of Kiev, or perhaps from Varangian guards.

But it is time now to consider the question, whence the Magyars originally came and who they were. For though Lebedia is their first recorded dwelling-place, it is clear from the record itself that it was not their original home. In fact, Constantine's narrative seems almost to imply that they abode only three years in that country. However this may be, no written record has come down to us of an earlier habitation. There is, however, certain evidence which makes it highly probable that the original country of the Magyars was in a north-easterly direction from Lebedia, beyond the Volga Bulgarians, north of the Kama. This evidence is the Hungarian language, which, whatever theory we hold as to its origin, points to a time when the Hungarians lived in close communion with Ugro-Finnic peoples. For, if we hold with Budenz and Hunfalvy that the Hungarians are themselves an Ugrian people, and that the Ugro-Finnic element in their language is the original basis of it, we must seek their first home in the UgroFinnic, and more narrowly in the Ugrian, circle. If, on the other hand, we accept the view of Vámbéry, that the Hungarians are of Turkish race, and that the basis of their language is Turkish, we have to account for the Ugro-Finnic element which was afterwards blended in their tongue, and we can only explain it by supposing that they lived in the neighbourhood of the Ostjaks, or the Voguls, for a considerable period.

According to Vámbéry, the great plain spreading from the Lower Jaxartes along the north shores of the Aral and Caspian, over the source-lands of the Ural, and beyond the Don, Dnieper, and Dniester to the confines of Hungary, was inhabited by one and the same branch of the Turk stock, divided into different groups. Of these the most northerly was the Magyar division, which coming in contact with Ugrian peoples north of the Kama, absorbed a new element into their language. In the same way he explains the origin of the Csuvas tongue. The Bulgarian people split up into two parts towards the end of the fifth century, just as the Hungarians and the Patzinaks split in the ninth. That part which remained in the east was pushed northward by the Khazars to the Kama-Volga region about 650, and there came in contact with a Finnic or Ugrian people, perhaps the Cseremiszes. Through mixture with them the Bulgarians lost their physical peculiarities, and their tongue was transformed. Such is the origin of the Csuvas language, according to the ingenious guess of Vámbéry, who holds that the Bulgarians were a Turkish people, and not Ugrian, as Hunfalvy tries to make out.

But whether we go with Vámbéry or with Hunfalvy, in either case the original home of the Magyars must have been in the same regions, north of the Kama. Hunfalvy thinks that memories of this ancient home survived among the Magyars even after they had reached their abiding home on the Danube, and are to be found in their oldest chronicle, that of the anonymous notary of King Béla IV., which was pulled to pieces by the merciless pen of Roesler.* The Anonymous says that the Magyars originally lived in Scythia, a large country of which the eastern part is called Dentu-moger. Hunfalvy identifies moger with Magyar, and dentu with Tangat, which he tries to show might have been a name of the river Irtish; and

There is a special work on this chronicle by J. Vass, entitled Béla Király névtelen jegyzőjének kora (1865), but it is practically worthless. Jászay's work on early Hungarian history (a Magyar Nemzet napjai a legrégibb idötöl az arany bulláig) is quite uncritical, and in its earlier part a mere reproduction of the Anonymous. The first pages of Szalay's great Hungarian history, Magyarország történelme, have also very little value.