entrench themselves on the peninsula. Reinforced by Turks whom they summon from Asia Minor, they repulse and defeat their aggressors . Once more for two years they devastate the neighbouring districts of Thrace  until the Byzantines are compelled to make a supreme effort . This is the end of Khalīl’s Turks: in a desperate effort to break out and to find ships for the crossing to Qaras, they perish (1311) .
Gregoras’ account is supplemented by that of Pachymeres, whose history, however, stops at 1308. According to him the Tourkopouloi served under their own officers,  were Christians of recent date, and had come only lately to the emperor — from the ‘Northern regions’ . After the battle of Aproi they move to Gallipoli  to join the Catalans . Their wives and children follow them but fall into the hands of the Alan mercenaries who send a small number of them to the emperor . He uses them to draw the Tourkopouloi once more to his side  but has no success. In these negotiations a certain Isḥāq (Ἰσαάκ) had offered his good services. He is a mighty Turkish leader (probably from Qaras); though in the camp of the Catalans, he is ready to abandon them . He must have nourished high ambitions since he gives himself the title of ‘Melik’ and asks in marriage the daughter of Sultan Mas’ūd, ‘Izzeddīn’s grand-daughter . The princess had been left in Constantinople when Mas’ūd, after his return to Rūm, was compelled to take refuge momentarily with the Byzantines, and there grew up generously provided for by the emperor . She has also an uncle there,
Melik Konstantinos (Μελὴκ Κονσταντῖνος), the second of ‘Izzeddīn’s two sons, who, unlike his brother Mas’ūd, had not left Constantinople, had readily accepted baptism and become a perfect Rhomaios . Ishaq, now returned to Anatolia, asked that he should be sent to him and proposed nothing less than to make him sultan of Rūm. Though Andronikos II found this latter project inopportune, he nevertheless sent Melik Konstantinos with the princess, his niece, to Pegai (Bigha), in the Troas, still a Byzantine possession, to be governor there and at the same time arrange with Isḥāq the matter of the marriage . The dubious game which Isḥāq played with the emperor, the Catalans, and the Tourkopouloi, ended in his death at the hand of the Catalans . The leader of the Tourkopouloi, Taghachar , barely escaped the same fate but was eventually spared and left in command of his troops — only to desert at the first opportunity with a number of his men to the Byzantines .
Pachymeres does not tell us what became of Melik Konstantinos but it is clear that he is the Melik whom Gregoras shows, two years later, as chief of the Tourkopouloi when they separate from the Catalans and who leads them afterwards into Serbia (where he eventually perishes ). Residing at Pegai so near to the events  and being in contact with Ishaq over the marriage of his niece, he must have become involved in Isḥāq’s intrigues to find himself one day, whether he wished it or not, drawn into the Catalan camp, where after Taghachar’s flight to the Byzantines he was the obvious man to take the command of the Tourkopouloi.
Combining the two accounts we learn that the Tourkopouloi, who had deserted at, or after, Aproi and then were most intimately connected with Khalīl’s Qaras Turks, spent the two years 1307-8 in Gallipoli; there was ample time for their kinsfolk to be sent over to Qaras. If their women and children had, after Aproi, fallen into the hands of the Alan mercenary corps — though probably not all of them — there was still a possibility of recovering them. We do not know if those few (probably the wives and children of officers) whom the Alans handed over to the emperor and whom he subsequently used as a means to separate the Tourkopouloi from the Catalans, were eventually restored to them. Those who remained in the hands of the Alans may have been recaptured when the Tourkopouloi attacked and annihilated the Alans in the Balkan passes while the latter were crossing into Bulgaria for service with the
tsar — indeed, the whole action was undertaken with this aim in view . The far-reaching raids the Tourkopouloi undertook throughout Thrace offered them other opportunities to recover dispersed groups of their families. At the end of Khalīl’s adventure there were again two years, 1310-11, spent in the peninsula of Gallipoli so that, if some families had followed their men to Thessaly and back, there was time and opportunity to pass them over to Qaras. Passing twice close to Karaferia, Khalīl could easily have been joined also by Turks of that region. Dobruja Turks would have joined Khalīl rather during his first stay in Gallipoli, in 1307-8. Yazjoghlu’s account establishes a connexion between their exodus and the advance of the Bulgarians — it was, indeed, just in 1307 that Andronikos II had to make peace with the Bulgarian tsar Svetoslav, leaving him in possession of all his conquests including the two important Black Sea ports of Anchialos and Mesembria . This meant that the Dobruja Turks were now completely cut off from the empire and from their kinsfolk there. It is only too understandable that some of them should have tried to join the Tourkopouloi at Gallipoli. Finally, what Gregoras describes as the complete annihilation of all the Turks of Gallipoli in 1311, may well have been nothing more than the last phase in a prolonged evacuation to Qaras. Above all, Khalīl himself seems to have escaped to safety, for Gregoras would not fail to mention his death if he had felt able to do so.
As we see, these closing paragraphs of the account, like the earlier ones, prove to contain information which is founded in fact or is at least historically possible, indeed, even probable. The story of the Dobruja Turks therefore appears to represent a genuine tradition. Where should it live on if not in Qaras — among people descended from the Dobruja Turks (the Tourkopouloi and their kinsfolk) and in the tales commemorating the great adventure of Khalīl Eje. When the Ottomans repeated, a generation later, the Gallipoli adventure, destined this time to initiate an era of great conquests, they certainly made good use of the help and advice which the experienced Turks of Qaras were able to offer — hence the appearance in the Ottoman historical legend of Eje Bey. This Eje Bey, an old and tried warrior, who shows Orkhān’s son Sulaimān Pasha how to cross over to Gallipoli, is obviously a man of Qaras, perhaps conceived, though hardly as Khalīl Eje himself , yet at least as representing the veterans of 1307-1311. There is one figure, unmentioned before in Ottoman history, who with the first Ottoman conquests in Rumeli suddenly appears on the scene as a brilliant leader; he may well be one of those Qaras Turks who only then entered into the Ottoman community (and therefore into their history) and were predestined to be in the forefront of the Rumeli adventure. This figure bears the name of Evrenos — a name which seems to
point to an origin from Varna, the chief town of the Dobruja (cf. the Avren Dagh, to the south of the town). These are, of course, and may never be more than, conjectures.
We have endeavoured to show that some of the Rumeli Turks could and probably did return with Khalīl Eje to Anatolia ; we have to stress, however, that their numbers must have been small. The bulk of the Tourkopouloi had gone with Melik Konstantinos to Serbia; another, smaller group, led by Taghachar, had rejoined the Byzantines; other Tourkopouloi, individuals or isolated groups, may never have deserted the emperor. As to those of their kinsfolk who had managed to reach Gallipoli in 1307-8, they may have been evacuated to Qaras before the march into Thessaly began — but their numbers were certainly limited. In the Qaras tradition, of course, only those counted who actually reached Anatolia, and their arrival was easily magnified into a mass immigration. Since as we know for certain the Tourkopouloi were Christians and this must also be true for their nomad kinsfolk with whom they maintained the closest contact, the newcomers arrived in Qaras not as Muslims but as Christians. They had, however, to turn Muslim at once — if only in appearance, and their Christian past fell into oblivion or was at least not openly spoken of. When Yazjoghlu presents them as Muslims who had left Rumeli because of their distaste for life among the Infidels, he is no doubt to a large extent following the Qaras tradition. In two other points, however, it is rather his reluctance to mention apostasy from Islam, which must have led him into misrepresentation of the facts: the Turks who remained in Rumeli after the events of 1307-11 were by far the majority and they were by then no longer Muslims, having accepted baptism perhaps a generation before. These christianized Turks in the Dobruja and their splinter-groups in Karaferia and Zikhna, explicitly attested by our account (§13), are beyond all possible doubt identical with the Gagauz, those Christian Turks who speak a Turkish of Anatolian character and until modern times had their main abode in the Dobruja, with small isolated colonies at Karaferia and Zikhna.
To conclude: the account as a whole shows remarkable consistency and chronological soundness, but these merits are no doubt to a great extent the result of Yazjoghlu’s skill in blending the four stories into a ‘Destan of ‘Izzeddīn’s and his people’s exile in the Dar ul-Harb’. By the sultan’s flight to the Byzantines he himself and all those who had followed him — his family, his warriors, and the nomads — became involved ever more deeply in a situation most perilous for Muslims. The destan carries the adventure of each of the dramatis personæ to a point where for a Muslim it has to end — in death, in return to the Dār ul-Islām, or in apostasy. Throughout, Yazjoghlu has endeavoured to spare the feelings of the Muslim reader: ‘Izzeddīn dies in exile but on Muslim soil and one of his sons returns to the throne of Rūm; another son ends as a Muslim saint; the third one dies as a faithful Muslim among the Infidels — it is true, alas, his descendants become Christians, nevertheless they are valiant soldiers fighting for the Ottoman sultan, the champion of Islam;
the warriors and the nomads eventually return to Anatolia, to the Dār ul-Islām — alas, some of them remain in Rumeli and turn Christian. Even in this form the account must have shocked pious Muslims, for Loqmān thought it wise to suppress the later part of the account and thus to avoid all mention of apostasy; for him the story ends with the emigration to the Desht-i Qpchaq, i.e. to the safety of the Dār ul-Islām.
Already Ibn Bībī’s chapter on ‘Izzeddīn’s flight is in itself a destan but strictly limited to the sultan’s own adventure. To expand this destan into a much richer one comprising also the adventures of ‘Izzeddīn’s people, was certainly suggested to Yazjoghlu by the dominant role ‘Izzeddīn played in a number of other stories which had come to his knowledge. Indeed, the Dobruja Turks not only enter the empire at ‘Izzeddīn’s summons, they also follow him into the Desht-i Qpchaq; a number of them join his son at Karaferia while his other son, the future Baraq, naturally finds shelter with his father’s people in the Dobruja. The intimate connexion in which these stories show the sultan and his sons on the one hand and his people on the other, must be the reflection of an underlying historical reality, all the more since in the Byzantine accounts of events which took place 40 years after ‘Izzeddīn’s disappearance from the scene, the sultan’s name is still very much alive and his son appears as the natural leader of the Tourkopouloi. It would certainly not be surprising if a Turkish community were to take their name from the man who had shaped their destiny. In this light Balaschev’s ingenious derivation of ‘Gagauz’ from ‘Kaikāūs’ becomes self-evident . The Gagauz are, indeed, the people of Kaikāūs.
It lies beyond the scope of this article to deal with all the problems involved  and to attempt a reconstruction of the historical reality. I shall be satisfied if my study restores the reputation of a source whose value and importance the Russian scholars Bruun and Smirnov had rightly perceived. To their illustrious compatriot Vladimir Minorsky, whose work has thrown so much light on the history of the Turkish peoples, I offer these pages as an expression of gratitude and homage.
1. The easternmost chain of the Balkan range which here, near to the coast where Greek survived the longest, retained its classical name of Haemus : Αἵμον (acc.) > Emine, with an intermediate form *Ἔμμωνα, Emona, found in documents of the early 14th century; see C. Jirecek, Das Füerstenthum Bulgarien, Vienna, 1891, pp. 4 and 527, n. 1.
2. T. Kowalski, Les Turcs et la lange turque de la Bulgarie du Nord-Est. Polska Akademja Umiejetnosci, Memoires de la Commission Orientaliste No. 16, Cracow, 1933, 28 p. In the introductory pages of this study is to be found an excellent survey of the rich literature on the subject. Kowalski has supplemented this paper by two shorter ones: ‘Compte-rendu de l’excursion dialectologique en Dobroudja, faite du 10 septembre au 1 octobre 1937,’ in : Bulletin de l’Academie Polonaise des Sciences et des Lettres, Cracow, 1938, pp. 7-12, and ‘Les elements ethniques turcs de la Dobroudja’, in : Rocznik Orjentalistyczny, xiv, 1938, pp. 66-80.
3. I have dealt briefly with this account in my article ‘La descendance chretienne de la dynastie Seldjouk en Macedoine’, in: Echos d’Orient, xxx, 1934, pp. 409-12, and more fully in my study ‘Les Gagaouzes = les gens de Kaykaus’, written for the Tadeusz Kowalski Memorial volume in 1948. I still hope it will one day appear in print, since it is by no means superseded by this present article; on the contrary, both are complementary one to the other. 4. The opuscule is entitled Ijmal-i ahval-i al-i seljuq ber mujib-i naql-i Oghuz-name. On Loqman see F. Babinger, Die Geschichtsschreiber der Osmanen und ihre Werke, Leipzig, 1927, pp. 164 seqq., where, however, p. 167, n. 1, Loqman’s indisputable authorship of the Ijmal is, without any justification, denied. To the information given there can be added the firman in Ahmed Refiq, Istanbul hayati, i, Istanbul, 1333h. = 1917-18, p. 52, No. 5 (German translation by G. Jacob in Der Islam, ix, 1919, p. 251), and the accompanying note, both of which give interesting information. For Loqman’s famous Hünername (by no means missing from the Topkapi Sarayi!) see also J. Karabacek, Zur orientalischen Altertumskunde IV = SB. Akad. d. W. Wien, cxxii, 1 (1913). Of Loqman’s Qiyafet ül-insaniye a new MS. has in the meantime been brought to notice and described in Fehmi Edhem and Ivan Stchoukine, Les manuscrits orientaux illustres de la Bibliotheque de l’Universite de Stamboul, Paris, 1933 (Mem. de l’Institut Francais d’Archeologie de Stamboul I), p. 1 and pl. I.