he supports this guess by a statement in the chronicle of Kézai, that the old country of the Magyars was watered by the rivers Togata and Etul. The Etul is, of course, the Volga; but Vámbéry has shown that Togora, not the Togata, is the right reading, and that the Ural is meant. But it is highly probable that the second part moger is really the same as magyar, and we have it also in hetu-moger, seven magyars,' as the Anonymus calls the seven chiefs of the Hungarian tribes, and in the land of Magoria,one of the three parts in which Scythia is divided in Kézai. What then is the meaning of Magyar? Hunfalvy and Roesler explain it as ma-ger or mo-ger, men of the land,' from ma, which means 'land' in the Vogulic tongue (cf. Finn maa), and ger, man,' corresponding to the Vogulic kär, and preserved in a diminutive form in the Hungarian gyerek, gyermek, 'child.' Vámbéry rejects this Ugrian etymology and proposes a Turkish one. He considers that Magyar stands for madsar, and that the original form was majar, meaning 'hero' or 'chief.' This would explain Μονάγερις, occurring in Theophanes † as the name of a Hun noble.

But the derivation of proper names is almost always uncertain and arbitrary, and in an inquiry touching the origin of a people it will generally be wise to seek for other evidence first, and only admit the testimony of proper names as confirmatory of conclusions otherwise arrived at it. The proofs which Vámbéry brings forward of his Turkish theory may be divided into three chief heads; (1) the testimony of Greek and Arabic writers; (2) ethnological and sociological considerations; (3) the character of Magyar civilisation, so far as we know it. Under the third head come the names of persons, offices, dignities, etc.; military tactics and customs.

(1) Both Leo the Wise and his son, Constantine VII., as we have already seen, knew the Magyars by the name of Turks. It may seem strange that this generic name should have been specially associated with the Magyars, while the Khazars, Pat

* It is stated in the chronicle that the Togora flows into the mare Aquilonis, which is the Caspian, not as Hunfalvy thought the Arctic Sea. + P. 176, ed. de Boor.

zinaks, and other neighbouring nations were designated by their own names. Vámbéry's solution of this difficulty is that the Magyars were only newcomers. Having been, of all the Scythian Turks, the most remote from the borders of the Empire, they did not appear on the scene of European history until the ninth century, and then the Greeks unfamiliar with their special name called them generally Turks. There may be some truth in this; but would not the same thing apply to the Patzinaks, who came westward later than the Magyars? It must be also remembered that the name Ovyypo seems to have been known at Constantinople. But in any case the fact that the Greeks called the Magyars Turks, does not even raise a presumption, much less prove that the Magyars are a Turkish race. With Leo and Constantine and their contemporaries 'Turks' had no more precise significance than 'Scythians.' They had no idea of the distinction between Turks and Ugrians. And in any case the Magyars had come under Turkish influence, and had partly Turkized their language before they set up their abode in Atelkuzu, so that even a competent and discriminating observer might easily have taken them for Turks. And for the same reason it seems to me that the statement of Ibn Dasta that they were a Turkish race has no independent value. Thus Vámbéry's first argument falls to the ground, as the facts are explicable on the other hypothesis.

(2) Vámbéry asserts the principle that all the nomads who marched against Europe, from Attila to Timur, belonged to the Tartar, and not to the Finnic family. Therefore, he concludes, the Magyars must, like Huns and Avars, like Bulgarians and Kumans, be Tartars or Turks, and not Ugrians or Finns. This is a very dangerous sort of argument-plausible at first sight as a bold generalisation, but carrying no conviction when one comes to look into it. There is nothing in the nature of the case to prevent our supposing that the Magyars were an exception to the rule. Whether the Bulgarians were also an exception it is beyond my scope to consider here; Hunfalvy holds that they were. It is quite conceivable that the Magyars, as the most southerly of the Ugrian peoples, and in constant communication with Turkish races, might have

been infected with the nomadic habit of the Turk. Vámbéry's rule proves nothing until it be otherwise shown that the Magyars conform to it.

(3) Vámbéry tries to show, in support of his theory, that the Magyar proper names mentioned by Constantine Porphyrogennetos are Turkish and not Ugrian. The etymologies which he proposes are in some cases very ingenious, and I should be very sorry to say that some of them may not be right. Zaltas, for example, might well be the Arab sultan,' but an Arabic word might reach Ugrians, through Turkish mediation, as well as the Turks themselves. Bultzus again looks very much like the Magyar bölcs, wise,' which is a Turkish word. On the other hand, his derivation of Arpad and other names seem farfetched and improbable. But what I would insist on is that even if some of these guesses were demonstrated to be right, they would not prove his case. For proper names may have been imported from abroad. It would require a very strong array of very sure cases to found a serious and solid argument on proper names. As for the names of the two rulers, the gylas and the karchas, mentioned by Constantine, we may readily admit that they are Turkish, borrowed by the Magyars from their neighbours. The gylas of Constantine is clearly the same as the dsila of Ibn Dasta. He was probably the general of the army, while the karchas was the judge. Ibn Dasta states that the King of the Magyars was called the kundu. This is clearly the same word as kender, by which, according to Ibn Fozlan, the Khazars designated the dignitary in their realm who was next in rank to the chagan. As we learn from Constantine that the institution of monarchy was introduced among the Magyars by Khazar influence, it does not surprise us to find that the title of the new dignity was borrowed from the Khazar vocabulary. Nor is there any difficulty, as Vámbéry and others seem to imagine, in the circumstance that while the Magyar kundu was the sovereign, the Khazar kender was only the under-king. On the contrary, this is just what we should expect, for it is clear that at the time of the election of Arpad, the Magyars were in a dependent relation to the Khazars. It is clear that the chagan would never have pro

posed the institution of monarchy to the Magyars if he did not intend to be himself the overlord of their kings. And this relation of overlordship might easily have been expressed by giving to the new lord of the Magyars the same title as that borne by the under-king of the Khazars. A kundu or kender or kende, or whatever the exact form of the name was, stood, we may suppose, in the same relation to the chagan, as a pýš to the Barileús, or a King to the Emperor.

6

In his work on Tactics Leo VI. gives some account of the mode of warfare of the Magyars, or as he calls them the Turks.* Vámbéry points out that the details mentioned by the Emperor correspond to practices existing at the present day among Tartar people in the east. Turkomans, Kirgizes, Özbégs are accustomed to sleep, eat, and drink on horseback, just as the Magyars did according to Leo. The same writer mentions that felt was used by the Turks' for the clothing both of the warrior and of his steed, and that each soldier carried with him on military expeditions a piece of iron of crescent shape for shoeing his horse. Vámbéry can 'cap' these details by instances from the habits of the Oriental Turks whom he knows so well. But all this does not prove his case nor overthrow the position of Hunfalvy. These customs and others like them were doubtless among the many things which the false Turks' learned from the true Turks; and Vámbéry can certainly not show that such customs could not be borrowed.

Vámbéry does not omit to summon to his aid the anonymous scribe of King Béla, whose untrustworthy chronicle has already been mentioned. But untrustworthy though it be as history, it

* In cap. xviii. Salamon (F.) has written a valuable little work on Leo's Tactics, of which I have noted the title at the head of this article. He tries to show that the Strategic of the Emperor Maurice was not really Maurice's work, but dates from the reign of Basil I., Leo's father. In arguing this thesis he falls into a strange mistake when he says (p. 6) hogy a ix. század elött nem kelhetett oly görög könyv, mely turkokrúl beszél.' He has evidently never read Menander, who lived in the reign of Maurice. Vámbéry makes a much queerer mistake in speaking of Salamon's book. He writes of 'Leo the Wise, or his father Maurice' (Bölcs Leo vagy atyja Maurikios)-a curious jumble.

may yet have some value. For it rests on the basis of tradition and népmondák or folk lore, and may therefore preserve dim memories of a past otherwise forgotten. The Hungarian scholar pertinently asks those who denounce the anonymous scribe as absolutely worthless,-whence did he get his proper names? For he did not get them from Dares Phrygius or from Regino, whom he uses so liberally in other matters We may fairly suppose that the names of persons and places mentioned in the chronicle rest on popular tradition. Vámbéry accordingly proceeds to deal with them, as he dealt with the names recorded by Constantine Porphyrogennetos. He interprets, for example, Esküllö, 'the place of the oath,' by the Turkish ecskilik. But I have already spoken of the value of arguments from proper names.

The battle between Ugrians and Turks for the possession of the Magyars must be fought on the field of language. And may I venture to suggest that Turanian philology will have to establish itself on a far surer footing and set up for itself a higher standard of exactitude, before we can have much confidence in its results. At present it is in that stage in which vowels count for nothing and consonants for very little, a stage which Indogermanic philology has fortunately passed. When we remember how much ingenuity has been thrown away, even in recent years, on attempts to connect Aryan with Hebrew words, we naturally think how easy it might be to make a Tartar vocabulary appear to be Ugrian or a Ugrian vocabulary appear to be Tartar. For while the genius of Aryan and the genius of Semitic tongues are totally different, the genius of Vogul and Ostjak and that of Turkish languages are alike—both, as we say, 'Turanian.' But though one must feel diffident in approaching a subject, which is still, I fear, at the mercy of punsters and guessers, there is one capital and obvious fact which raises a very strong presumption in favour of Budenz and Hunfalvy. In the comparison of languages, almost the first thing that one looks at is the numbers. If the Magyars were a Turkish people they would surely have kept the Turkish words for one, two, three; if a Ugrian people, the Ugrian. One's own words for the simple numbers are the last part of the vocabulary one is likely to replace by the words of strangers. In the case of a higher number, borrowing is con