^{page 639 note 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 (ace.) > Emine, with an intermediate form Ἔμμωνα, Emona, found in documents of the early 14th century; see Jireček, C., Das Fürstenthum Bulgarien, Vienna, 1891, pp. 4 and 527, n. 1.Google Scholar

^{page 640 note 1} Kowalski, T., Les Tures et la lange turque, de la Bulgarie du Nord-Est.Google Scholar Polska Akademja Umiejętności, Mémoires 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'Académie Polonaise des Sciences et des Lettres, Cracow, 1938, pp. 7–12Google Scholar, and ‘Les éléments ethniques turcs de la Dobroudja’, in: Rocznik Orjentalistyczny, xiv, 1938, pp. 66–80.Google Scholar

^{page 640 note 2} I have dealt briefly with this account in my article ‘La descendance chrétienne de la dynastie Seldjouk en Macédoine’, in: Échos d'Orient, XXX, 1934, pp. 409–12Google Scholar, and more fully in my study ‘Les Gagaouzes = les gens de Kaykāūs’, 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 I still hope it will one day appear in print, since it is by no means superseded by this present article; on the contrary, both are complementary one to the other.

^{page 640 note 3} The opuscule is entitled Ijmāl-i aḥvāl-i āl-i seljūq ber mūjib-i naql-i O ghuz-nāme. On Loqmān see Babinger, F., Die Geschichtsschreiber der Osmanen und ihre Werke, Leipzig, 1927, pp. 164 seqq.Google Scholar, where, however, p. 167, n. 1, Loqmān's indisputable authorship of the Ijmāl is, without any justification, denied. To the information given there can be added the firman in Aḥmed Refīq, Istānbūl ḥayāti, i, Istanbul, 1333h. = 1917–18, p. 52, No. 5 (German translation by Jacob, G. in Der Islam, ix, 1919, p. 251)Google Scholar, and the accompanying note, both of which give interesting information. For Loqmān's famous Hünernāme (by no means missing from the Topkapi Sarayi!) see also Karabacek, J., Zur orientalischen Altertumskunde IV = SB. Akad. d. W. Wien, cxxii, 1 (1913).Google Scholar Of Loqman's Qiyāfet ül-insānīye a new MS. has in the meantime been brought to notice and described in Fehmi Edhem and Ivan Stchoukine, Les manuscrits orientaux illustrés de la Bibliothèque de L'Université de Stamboul, Paris, 1933 (Mém. de l'Institut Français d'Archéologie de Stamboul I), p. 1 and pl. I.

^{page 640 note 4} In his Geschichte des osmanischen Reiches, i, p. 122Google Scholar; ii, p. 143; iii, p. 202; viii, p. 354, and also in his Geschichte der Goldenen Horde, p. 174.Google Scholar Hammer found Loqmān's opuscule in a MS. of his own between two works of Luṭfī Pasha wherefore he ascribed it to the latter, quoting it as ‘Lutfi's Oghusname oder Geschichte der Seldschuken’. Hammer's copy, apparently still the only known one, is now in the Nationalbibliothek of Vienna, H.O. 17b = Flügel, No. 1001, 2.

^{page 640 note 5} Lagus, I. J., Seid Locmani ex libro turcico qui Oghuzname inscribitur excerpta. Helsingfors, 1854.Google Scholar

^{page 641 note 1} Of the 15 paragraphs which constitute my résumé of the whole account (see below, pp. 648–51) only the first nine are found in Loqmān. On the other hand he has made some additions—above all the date which he gives for the immigration into the Dobruja, right at the beginning, in the form of a distich:—

This date 662h. = 1263–4 is clearly the result of calculation, nevertheless, it cannot be far from the truth.

Another addition is his statement (ed. Lagus, , Turk, text, p. 7)Google Scholar that the story of Sultan 'Izzeddīn's flight to the Byzantines is to be found in the ‘Destān of the blessed Ṣari Ṣaltiq’; in his source the chapter where Ṣari Ṣaltiq appears is in fact headed ‘Flight of Sultan 'Izzeddīn to the Byzantines’; clearly Loqmān's interest was so exclusively concentrated on Sari Saltlq that he felt the rest ofcthe account to be merely a subordinate framework.

^{page 641 note 2} As Th. Menzel did in his article ‘Gagauz’ in the Encyclopædia of Islām. The account was mercilessly analysed and rejected by the Bulgarian historian Mutafčiev, P., Die angebliche Einwanderung von Seldschuk-Türken in die Dobrudscha im XIII. Jahrhundert, Sofia, 1943, 129 p.Google Scholar; (Вългарска Академия на Наукит и Изкуствата, lxvi, 1). He found the support of an Orientalist: Duda, H. W., Zeitgenössische islamische Quellen und das Oguznāme des Jazyğyoġlu 'Alī zur angeblichen türkischen Besiedlung der Dobrudscha im 13. Jhd. n. Chr., ib. (Bulg. Ac., lxvi, 2), pp. 131–145.Google Scholar Other scholars who had doubts about the value of the information found in Loqmān are listed in Mutafčiev, p. 13, n. 1 (for those who accepted it see ib., p. 12, n. 2).

The late Mutafchiev was obviously inspired by his patriotic zeal to show the Dobruja Turks as Bulgarians who had adopted the Turkish language. In defence of his hopeless case he displays an enormous erudition and his book is certainly the most comprehensive study of the subject ever made: he reviews at length the various theories of his predecessors and at the same time he brings together a rich information from the sources, subjecting it to a thorough though strongly biased commentary. I wish to state that I am much indebted to Mutafchiev's work.

In the main the book is a bitter attack on a pamphlet by the Bulgarian scholar G. D. Balaschev: Γεώργιος Δ. Μπαλ⋯στζεφ Ὁ αὐτοκρ⋯τωρ Μιχα⋯λ Η´ ⋯ Παλαι⋯λογος κα⋯ τ⋯ ἱδρυθ⋯ν τ συνδρομ αὐτο κρ⋯τος τν Ὀγουζν παρ⋯ τ⋯ν δυτικ⋯ν ἄκτην το Εὐξε⋯νου, Sofia, 1930, 26 p., where Loqmān, reproduced in Greek translation, is accepted almost without criticism and used to reach rather rash and exaggerated conclusions. Nevertheless, Balaschev, too, has his merits, above all, his explanation of the name of ‘Gagauz’ (see below, p. 668) is a brilliant and important discovery. For Mutafchiev, of course, Balaschev has only the negative merit of discrediting, by carrying it ad absurdum, the identification of the Gagauz with 'Izzeddīn's Turks, a theory which had been advanced by Bruun, Ченоморье, ii, Odessa, 1880, p. 333, and accepted by Smirnov, Крымское ханство подъ верховенство Отоман. Порты, St. Petersburg, 1887, p. 17 (see Mutafčiev, , p. 82Google Scholar, n. 1, and p. 84, n. 2).

^{page 641 note 3} The first to state Loqmān's dependence on Yazijioghlu was M. Th. Houtsma in his Recueil des textes relatifs à l'histoire des Seldjoucides, iii, p. x; he shows him, however, not as dependent on Yazijioghlu himself but on an abridged version of his work—indeed, an abridgement, MS Paris, Bibl. Nat., Suppl. turc 1182 (see Blochet, E., Catalogue des manuscrits turcs, Paris, 1932–1933, ii, p. 190)Google Scholar was all that Houtsma had for comparison.

Mutafchiev, (p. 16, n. 1)Google Scholar knew from my remarks in Der Islam, xx, 1932, p. 202Google Scholar seq., that Loqmān comes from a much older source but took no account of it. The intervention of the orientalist Prof. Duda reassured him, indeed, that Loqmān's information, though taken from Yazijioghlu, did not occur in the contemporary Seljuq sources—and therefore was worthless. Both these scholars ignored my article in Échos d'Orient (see above, p. 640, n. 2) which would have shown them that Loqmān has omitted not only the larger but, indeed, the essential part of Yazijioghlu's account.

^{page 642 note 1} Complete manuscripts: two in Istanbul, Topkapi Sarayi, Revan Köşkü, 1390 and 1391, one in Berlin, Staatsbibliothek, Orient Quart 1823 (from these three MSS. my notes are taken), and one in Paris, Bibliothèque Nationale, Supplément Turc 737 (Blochet, , Cat., ii, p. 47).Google Scholar

Incomplete manuscripts: Leyden, Warner 419, Paris, Bibl. Nat., Ancien fonds turc 62 (Blochet, , Cat., i, p. 24)Google Scholar and Suppl. Ture 1185 (Blochet, , ii, p. 191).Google Scholar

Edition: Houtsma, M. Th., Recueil III: Histoire des Seldjoucides d'Asie Mineure d'après Ibn Bībī, Texte turc, Leyden, 1902Google Scholar, based on the two incomplete MSS. Leyden and Paris, A.f. 62, reproduces only that portion of the text which is in the main a translation from Ibn Bībī and this only as far as the then available MSS. permitted, i.e. less than the half. The end of Rec. III corresponds to Rec. IV (Ibn Bībī), p. 159, ult.Google Scholar

An edition of the complete work is, as I was glad to hear, being prepared by the Turkish Historical Society (Türk Tarih Kurumu).

^{page 642 note 2} Duda, H. W. in ZDMG., N.F. 14 (89), 1935, p. *19* seq.Google Scholar

^{page 642 note 3} Istanbul, Library of the Aya Sofya, No. 2985: Al-awāmir al-'alānīya fi'l-umūr al-'alā'īya, by Ḥusain b. Muḥammad b. 'Alī al-Ja'farī ar-Rugẖadī () al-mutahir bi-Ibn Bībī al-munajjima (MS., p. 10, 1. 7), 744 pages; cf. Tauer, F., ‘Les manuscrits persans historiques des bibliothèques de Stamboul, iv,’ in: Archiv Orientálni, iv, 1932, p. 92.Google Scholar

The Turkish Historical Society have announced their intention to publish a facsimile of the manuscript.

^{page 642 note 4} Notice on the donation, accompanied by the sultan's seal, on the frontispiece.

^{page 642 note 5} Edited by Houtsma, M. Th., Recueil IV: Histoire des Seldjoucides d'Asie Mineure d'après l'abrégé du Seldjouknameh d'Ibn Bībī, Texte persan, Leyden, 1902.Google Scholar

Prof. Duda has claimed (loc. cit.) that the abridgement was made in the lifetime of Ibn Bībī, assuming that the Mālik-i (or: amīr-i) dīwān aṭ-ṭughrā Amīr Nāṣir ad-dīn Yaḥyā al-ma'rūf bi-Ibn al-Bībī (Rec., iv, p. 2Google Scholar, 1. 3, and p. 196,1. 2) is the author of the original—which he is most certainly not since from the original we know (see above, n. ?.) that its author was named Ḥusain (al-mutahir bi-Ibn al-Bībī); there is no mention in the original of an Emīr Yaḥyā nor of the office which according to the abridgement he held. The chapter heading in Rec., iv, p. 196, appears in the original MS., p. 442, simply as . Prof. M. Fuad Köprülü (in Belleten, vii, 1943, p. 388 seq.) is certainly correct in regarding the Emīr Yaḥyā as the brother of Ḥusain but errs in believing that he is the author of the abridgement which explicitly attributes to him the authorship of the original . I see no solution of this puzzle other than the assumption that the author of the abridgement, who remains anonymous, has attributed to Yaḥyā Ibn Bībī what belonged to Ḥusain Ibn Bībī, probably after the latter's death and in order to ingratiate himself with his new superior in the divan. Though we do not know what office Ḥusain had held, it is probable that he was his brother's predecessor as mālik-i dīwān aṭ-ṭugẖrā. It is the very office which their father Majd ad-dīn Muḥammad Tarjumān had held until his death in 1272. He is described as chief of the inā office, which is probably identical with the dīwān aṭ-ṭugẖrā (see Björkman, W., Beitrāge zur Geschichte der Staatskanzlei im islamischen Ägypten, Hamburg, 1928, p. 44, n. 4).Google Scholar (According to the original he was appointed to this post after having been head of the ‘tent and carpet depot’

firā - āne-i āṣṣ [MS., p. 443,1. ii: .)

At least it is certain that the abridgement was, like the original, written in the reign of Mas'ūd; Rec., iv, p. 334, 1. 20 seq.: belongs to the epitomisi since in the original, MS., p. 736,1. 2, the corresponding passage reads:

^{page 643 note 1} Muḥammad ar-Rāwandī, Raḥat aṣ-ṣudūr (composed in 1202–3, retouched between 1207 and 1210), ed. Muḥ, , Iqbāl, , London, 1921 (Gibb Mem. Ser., N.S. ii).Google Scholar

^{page 643 note 2} I am not dissuaded from transcribing Qutlumush by Prof. Fuad Köprülü's argument in Tarih Dergisi, i, 1950, pp. 227–230Google Scholar; the passage from 'Ayn¯'s 'Iqd al-Jumān quoted in his postscript (ib., p. 236) even gives welcome support to my reading.

^{page 643 note 3} The last passage taken from Rāwandī (p. 128, 1. 5): reads in Yazijioghlu as follows:

(with this heading Rec. III begins).

In the ‘fore-runner’ (between the asterisks) Yazijioghlu shows himself influenced by Ḥamdullāh Mustaufī's Ta'rī -i Guzīde (Gibb Mem. Ser., xiv, 1), which reports immediately before the date 471h. Sulaimān's dispatch by Malikāh, not against Rūm, but against Antioch; ib., p. 441, the name of the Byzantine emperor is given as , to be read Urumānōs, i.e. ‘Romanos’; cf. p. 481, where it is corrupted into ( On the other hand, Yazijioghlu's , i.e. fasilyevs, βασιλε⋯ς is a first indication of his turning to Ibn Bībī where this is the usual term for the Byzantine emperor (Rec., iv, p. 14Google Scholar, 11. 2 and 19, p. 15, 11. 9 and 13, etc.); indeed, Sulaimān's dispatch to Rūm by Malikāh is mentioned in Ibn Bībī, in Qilij Arslan's speech to his son (MS., p. 18,11. 14–16; not in Rec., iv, p. 3)Google Scholar: (= Rec. iii, p. 9Google Scholar, 11. 6–9).

^{page 644 note 1} Houtsma states (Rec. iii, p. ix)Google Scholar that in the abridged version of the Ogẖuznāme (see above, p. 641, n. 3) the ‘continuation’ refers here and there to sources, e.g. to the Ta'rī -i Guzīde. As far as my reading of the original work goes I have come across no such reference.

^{page 644 note 2} I have reproduced, translated, and analysed the beginning of this closing chapter in my Das Fürstentum Mentesche, Istanbul, 1934, p. 32 seqq.Google Scholar

^{page 644 note 3} The MSS. used by Berezin for his edition of Rashīdeddīn's Jāmi' et-tevārīkh (Труды ВосточΗ. Отдјі. Имп. Арχеол. Ощества, vii, St. Petersburg, 1861, pp. 32–8), contain the tamgẖas though probably in a form far inferior to the fine tables found in the MSS. of Yazijioghlu's Ogẖuznāme. In Vambery, H., Das Türkenvolk, Leipzig, 1885, pp. 4–6Google Scholar, and (incomparably better) in Mayer, L. A., Saracenic Heraldry, Oxford, 1933Google Scholar, pls. 1 and li, the tamgẖas (in the last named work the entire tables) are reproduced from the Leyden MS.—they are still more finely executed in the Berlin MS.

Occasionally Yazijioghlu makes additions to Raīdeddīn's text, some of which are not without interest, e.g. when he expands Rashīdeddīn's (ed. Berezin, p. 5) in the following way (MS. Berlin, f. 2b):

This addition (from the asterisk onward) can be regarded as an observation of his own despite the fact that he refers to rāvīer. Some lines later he mentions an Ogẖuznāme in Uighur characters (); we read further on (f. 3a) that information on the Ogẖuz is to be found ‘in the Jāmi' et-tevārī and in the Ogẖuznāme’. It is possible that here Yazijioghlu has in mind the famous Uighur Ogẖuznāme made known by Radloff and subsequently studied by Biza Nour, Pelliot, Bang, and Arat—in his own time uigurica were, indeed, in fashion at the Ottoman court—but as far as my notes go there is no passage which could be traced back to that text; what I have said on this in Der Islam, xxx, p. 202Google Scholar, has therefore to be corrected.

^{page 645 note 1} Numerous examples in Rec. III, e.g. p. 74Google Scholar, 1. 21 (Ogẖuz boylari); p. 80, 1. 20 (beys); p. 99, 1. 11 (deeds); p. 204, 1. 3, p. 205, 1. 17 (feast, verses); p. 217, 1. 12 (Sultan 'Alāeddīn, the Ogẖuznāme and the türe). It goes without saying that all these passages are missing not only in the abridgement but also in the original of Ibn Bībī.

^{page 645 note 2} The passage on ‘Osman's election’ has been reproduced, in a very much shortened form, by Münejjimbai, Jāmi' ad-duval, iii, p. 278Google Scholar, and before him, at greater length, by Luṭfī Paa, Ta'rī , p. 21Google Scholar, and by Rūḥī Edrenevī (see Mordtmann, J. H. in Mitteilungen zur osmanischen Geschichte, ii, p. 136).Google Scholar The influence of Yazijioghlu's work appears indeed very early in Ottoman historical writing.

^{page 645 note 3} Though most of the elements of which these interpolations are made up come clearly from Raīdeddīn, some of them as well as the vivid way in which they are presented may well be due to Yazijioghlu's personal knowledge of ogẖuz traditions; cf. Abdülkadir, in Türk Hukuk ve Iktisat Tarihi Mecmuast, i, 1931, p. 123seq.Google Scholar

^{page 645 note 4} It would be unthinkable, of course, that there should be no āzī theme in this work. It appears frequently, often linked with the ogẖuzian theme, and is introduced right at the beginning of the work (in the MS. Berlin, 1. 6 of the first page): (The subject of vérdi is, of course, God.)

^{page 645 note 5} From the numerous examples found in Rec. III, I quote p. 10, 1. 15, where the yazijilar are added to Ibn Bībī's text in order to make them appear among the grandees, and above all the long poem, pp. 253–7.

^{page 646 note 1} Rec., iii, pp. 87, 372, and 382.Google Scholar

^{page 646 note 2} Duda, H. W., Zeitgenössische islamische Quellen, etc. (see above, p. 641, n. 2)Google Scholar, p. 138, n. 4 and 6, quotes from MS. Berlin and (I add the first hemistich): as well as the distich which reveals, or rather hides, the date:

Prof. Duda has succeeded in wresting from this chronogram an acceptable date, understanding that the reader is invited to add to = 540 the number value of = 287, which gives as date 827h./beg. 5 December 1423.

^{page 646 note 3} See in the preceding note Prof. Duda's interpretation, which can, however, just as well yield as date 833/beg. 30 September 1429, if we take into account the tadīd of and reckon the wāw twice.

^{page 646 note 4} Rec., iii, p. 87, 11. 15–19.Google Scholar

^{page 646 note 5} See below, p. 652.

^{page 646 note 6} The ‘Testament’, fully vocalized at the beginning to give it the aspect of a sacred text, and headed (MS. Berlin, f. 18a) , follows immediately the tables containing the tamgẖas; though it replaces Raīdeddīn's account (ed. Berezin, , vii, p. 38 seq.), it is obviously very much influenced by it.Google Scholar

^{page 647 note 1} The süῃük ‘bone’ is a certain part of a certain animal which at the festivities is reserved for a certain tribe. See Houtsma, M. Th., ‘Die Ghusenstämme,’ in WZKM., ii, 1888, p. 229.Google Scholar

^{page 647 note 2} MS. Berlin, f. 19a:

^{page 647 note 3} Following immediately the preceding quotation:

This passage has been placed at the beginning of the Kitāb Dede Qorqud (ed. Rif'at, Killisli Mu'allim, Istanbul, 1332, p. 3, 11. 1–6)Google Scholar; it is alien to the rest of the book and visibly a later addition.

^{page 647 note 4} Foreshadowed by the ‘fore-runner’ Rec., iii, p. 218, 1. 1.Google Scholar

^{page 647 note 5} See the remarks on the introduction of the ān title in my ‘Notes sur la tughra ottomane [II]’ in Byzantion, xx, 1950, pp. 279–282.Google Scholar

^{page 647 note 6} In Der Islam, xxx, p. 203Google Scholar, I have presented Yazijioghlu 'Alī and the famous mystic writer Yazijioghlu Meḥmed (and implicitly also the latter's brother Yazijioghlu Aḥmed Bījān) as ‘brothers’, in the belief (1) that Yazijioghlu 'Alī wrote in the later years of Murād II's reign, and (2) that Yazijioghlu Meḥmed was, as Evliyā, Seyāḥatnāme, iii, p. 366Google Scholar, says, the author of a risāle on Ṣari Ṣaitiq, which seemed to me to explain the interest his ‘brother’ 'Alī took in this holy man. Both assumptions have proved to be wrong. It is obvious that Evliyā has mistakenly attributed our 'Alī's account, which he like Loqmān (see above, p. 641, n. 1) calls a ‘Saltiq-nāme’, to the much better known Yazijioghlu, i.e. Meḥmed. It remains nevertheless true that all three have in common a family name of great distinction which was probably reserved for the members of one family, that all were literary people though in very different fields, and that they were contemporaries, for Meḥmed and Aḥmed were well advanced in age when they began writing about 1450, almost a generation after 'Alī had composed his O ghuznāme. Still another Yazijioghlu is known: Nerī (Ǧihānnümā, ed. Taeschner, F., i, p. 65Google Scholar; edd. Unat, F. R. and Köymen, M. A., i, p. 239Google Scholar) mentions a Yazijioghlu as ambassador to Egypt in the later years of Murād I's reign—a mission often entrusted to a high dīvān official. This ambassador could well be our 'Alī himself or his father, just as the two mystic writers could be his brothers or his sons (that Kātib Ṣalāḥeddīn has been regarded as their father [cf. Gibb, , History of Ottoman Poetry, i, p. 390Google Scholar seq.] is obviously nothing but an inference from Katib). As matters stand one can at best speak of a certain probability that the four Yazijioghlus belong all to one and the same family.

^{page 648 note 1} Text (MS. Berlin) and German translation in Duda, H. W., Zeitgenōssische islamische Quellen, etc., p. 143 seq.Google Scholar

^{page 649 note 1} Text (MS. Berlin) and German translation in Duda, , op. cit., p. 144.Google Scholar

^{page 649 note 2} MS. Revan K. 1391, f. 411b: (sic!)

^{page 649 note 3} Cf. the last chapter in Rec. IV.

^{page 649 note 4} ; this name was known to Yazijioghlu from Ibn Bībī who, however, writes it ; missing in Rec., iv, p. 334, 1. 15Google Scholar, where the original, MS., p. 735, 1. 6, reads as follows: … Could the change into (belā ‘calamity’) be meant as a pun?

^{page 649 note 5} MS. Revan K. 1391, f. 415a:

^{page 650 note 1} For this comparison see my ‘Islam und Kalifat’ in: Archiv für Sozialwissenschaften und Sozialpolitik, liii, 1925, p. 412, especially n. 82.Google Scholar The anonymous history of the Rūm Seljuqs there referred to (MS. Paris, suppl. pers. 1553) has ; see Uzluk, Feridun Nâfiz, Anadolu Selçuklari tarihi, Ankara, 1952, facsimile, p. 42,1. 6.Google Scholar

^{page 650 note 2} MS. Revan K. 1391, f. 415a:

^{page 650 note 3} MS. Revan K. 1391, f. 415b:

^{page 651 note 1} MS. Revan K. 1391, f. 444a: (The Passage tas been quoted in transcription by Aurei Decei in his article ‘Dobruca’ in Islâm Ansiklopedisi, iii, p. 632b.)Google Scholar

^{page 652 note 1} See above, p. 646.

^{page 653 note 1} Dobruja-éli is one of the numerous designations of countries formed by a name + él-i, the possessive suffix indicating that the country is regarded as belonging (or having belonged) to the person or people named in the first element; thus it means ‘Land of Dobruja (Dobrotitsa)’, as Chalkokondylas, ii, p. 98, 1. 15 Darkó, says: (⋯ το Εὐξε⋯νου παραλ⋯α,) Δοβροτ⋯κεω (το Μυσο) χώρα.

That ‘Dobruja’ = ‘Dobrotitsa’ becomes quite clear from Neri (ed. Taeschner, p. 68,1. 13; edd. Unat-Köymen, p. 249, 1. 15: ) where Dobrotitsa's son and successor Ivanko appears under the name of ‘Dobruja oghlu’.

^{page 653 note 2} Rec., iv, p. 296, 1. 1–p. 297, 1. 12.Google Scholar

^{page 653 note 3} Pachymeres, Georgius, i, p. 130, 1. 17–p. 132, 1. 16, and ii, p. 609, 1. 12–p. 611, 1. 15, Bonn. Nicephorus Gregoras, i, p. 82, Bonn.Google Scholar

^{page 654 note 1} Rec., iv, p. 298, 1. 11Google Scholar, the epitomist speaks of the episode as an efsāne but this expression is not to be found in the original.

^{page 654 note 2} Duda, H. W., Zeitgenössische islamische Quellen, etc. (see above, p. 641, n. 2), p. 145.Google Scholar

^{page 654 note 3} See Laurent, V., ‘La domination byzantine aux Bouches du Danube sous Michel VIII Paléologue,’ in Revue du Sud-Est européen, xxii, 1945, pp. 184–198.Google Scholar

^{page 655 note 1} Pachymeres, , i, p. 133, 11. 3–15Google Scholar, is, however, explicit at least about the fact that at the time of ‘Izzeddīn's stay in the empire Turkish nomads (σκην⋯ται) appeared on the Eastern frontier: hating all discipline, loath to submit to the Tatars, and anxious to be left alone, they infiltrated into the Byzantine defences, proclaiming themselves to be allies of the emperor but none the less plundering under the cover of darkness, though the frontier people were able to hold them in check. Finally the emperor bound them closely to his service, i.e. used them for military purposes. This measure, which meant the settlement of these nomads in the Eastern frontier zone, was designed to strengthen that frontier against the Tatars who, in spite of the prevailing good relations, had nevertheless to be deterred from all aggressive intentions. The dangers inherent in the measure may soon have become apparent, since there was no guarantee that at a given moment these Turks might not make common cause with nomads beyond the frontier. Besides the European campaigns drained off from Anatolia all available reserves for service in the Balkans. The soldiers levied among the newly arrived nomads may, sooner or later, have also been sent there—and their kinsfolk with them. It is not impossible, therefore, that Pachymeres’ nomads and those of Yazijiolu's account are one and the same.

^{page 655 note 2} Of. Rec., iv, pp. 297, 1. 12–298, 1. 2. The original, MS. p. 639, 1. 2, is still more explicit:Google Scholar

^{page 655 note 3} Pachymeres, , I, 131, 1.2Google Scholar: [ἅμα] κα⋯ γραιậ μητρ⋯, χριστιαν εἰς τ⋯ μ⋯λιστα οὔσῃ. See also EI., S.v. ‘Kaikā'ūs II’, where on the testimony of Frater Simon in Vincentius Bellovacensis, lib. xxxi, cap. 26, she is said to have been the daughter of a Greek priest.

^{page 656 note 1} Tafel, and Thomas, , Urkunden zur älteren Handels- und Staatsgeschichte, der Republik Venedig, vol. ii, p. 208Google Scholar, in the Tributa Lampsacenorum, of 1219: de CXX plinthis de vineis quas receperunt pro anacapsi pp. (= perpera, hyperpers) VIII annuatim. The term anacapsi of this Latin text can be regarded as a perfect rendering of the vernacular form for ⋯ν⋯καμψις, the annual payment due by a special category of tenants, the ⋯νακαμπτικς ἔχοντες. It was good luck, indeed, to come across Prof. D. A. Zakythinos' brilliant study on ‘La Société dans le Despotat de Morée’ in L'Hellénisme Contemporain, 2nd ser., vola, iv and v (Athens, 1950 and 1951), and to find there, at the very last moment, just what was needed to solve the anaqaplsi problem with which I had been struggling in vain for years.

^{page 656 note 2} Cf. Rec., iv, p. 298, 11. 2–6.Google Scholar

^{page 656 note 3} The story of ‘Izzeddīn's projects, imprisonment and liberation in Pachymeres, i, pp. 229, 1. 3–240, 1. 22, and in Gregoras, i, pp. 82,1. 10–83,1. 2; 99,1. 21–101, 1. 19, also in Aqsarāyī, ed. Turan, O., pp. 75Google Scholar, 1. 5–76, 1. 12. His liberation, in Maqrīzī, , Sulūk, i, Cairo, 1934, p. 522Google Scholar (transl. Quatremère, , Histoire des Sultans Mamlouks, I 2, p. 57 seq.Google Scholar); other Arabic sources in Tiesenhausen, W. de, Recueil de matériaux relatifs à l'histoire de la Horde d'Or, i, St. Petersburg, 1884, pp. 81Google Scholar (Rukneddīn Baibars), 133 (Nuwairī), 179 (Mufaḍḍal), 200 (ahabī), 482 ('Ainī).

^{page 656 note 4} Cf. Rec., iv, p. 298, II. 7–9.Google Scholar

^{page 656 note 5} Cf. Rec., iv, p. 298, II. 9–11.Google Scholar

^{page 657 note 1} Gregoras, , i, p. 101, 11. 16–19Google Scholar: ⋯ δ⋯ περ⋯ ⋯κενον [τ⋯ν σονλτ⋯ν Ἀζατ⋯νην] ⋯χλος, ἄνδρες δ᾽οτοι μ⋯λα τοι πλεστοι κα⋯ κρ⋯τιστοι τ⋯ πολ⋯μια, τ Χριστιανν ⋯ναγεννηθ⋯ντες βαπτ⋯σματι, τ `Ρωμα⋯ων συγκατελ⋯γοντο στρατιậ. Also p. 229, 11. 11–17, and p. 248, 11. 6–10.

^{page 657 note 2} Pachymeres, , ii, p. 574, 11. 5–7: τος ⋯ ὑπογ⋯ου Χριστιανος Τουρκοπο⋯λοις, ο δ⋯ κα⋯ οὐ πολλ πρ⋯τερον Χρ⋯νῳ ⋯κ τν βορε⋯ων βασιλε προσεøο⋯τησαν …Google Scholar

^{page 657 note 3} Ib., p. 550, 11.15–19: εἰσ⋯ δ᾽οἵ λ⋯γονσιν ὅτι κα⋯ ⋯ν⋯πυστον γεγον⋯ς ⋯κενοις [τος Τουρκοπο⋯λοις] τ⋯ ⋯π⋯ το Τουκτ⋯ πρ⋯ς βασιλ⋯α μ⋯νυμα περ⋯ τς αὐτν ⋯νθυποστροøς (δλον γ⋯ρ ν ὡς ⋯ Τουκτ⋯ς ⋯κε⋯νους ὡς ἰδ⋯ους κα⋯ τς τν Τοχ⋯ρων ⋯ρχθεν Χειρ⋯ς ⋯πτει παρ⋯ τ⋯ν ⋯ν⋯κτα πρεσβευ⋯μενος) …

^{page 657 note 4} See above, n. 1.

^{page 657 note 5} Gregoras, , i, p. 248Google Scholar, 1. 9 seq.: τας τν διαδοχας αὐηθ⋯ντας.

^{page 658 note 1} The türbe of the saint still exists at Aqehir (see Sarre, F., Reise in Kleinasien, Berlin, 1896, p. 22Google Scholar). Of its two inscriptions, one, bearing the date 1224, comes from a mosque and is here re-employed as an ornament (Huart, Cl., Epigraphie d'Asie Mineure, Paris, 1895, no. 15Google Scholar); the other, above the door, mentions the restoration of the türbe by the saint's great-grandson Seyyidī Muḥyī ed-dīn in 812h. = 1409–1410. I give here its full reading to replace the incomplete copy in

Huart, no. 16:

The three wooden coffins, wonderfully carved (Sarre, F., Seldschukische Kleinkunst, Leipzig, 1909, ii, pl. 14Google Scholar, shows them still in situ), are now in the Museum of Turkish and Islamic Antiquities (the former Evqāf Müzesi); in the guide book of the museum (Türk ve Islam eserleri rehberi, Istanbul, 1939, p. 13Google Scholar, nos. 193, 191, and 194) are fairly correctly reproduced the inscriptions of the three sarcophagi, i.e. of our saint Maḥmūd b. Mas'ūd, died in 667h. = 1268–9, of his brother Aḥmed b. Mas'ūd, died 649h. = 1251–2, and of his grandson ‘Alī b. Meḥmed (i.e. Muḥyī ed-dīn) b. Maḥmūd er-Rufā'ī (the father of the restorer of the türbe), without date.

Maḥmūd al-Ḥayrānī appears as a contemporary of Jelāleddīn Rūmī (died in 1273) in Eflākī; see Huart, Cl., Les Saints des Derviches Tourneurs, Paris, 1918–1922, ii, p. 108.Google ScholarGross, E., Das Vilājet-nāme des Ḥā ī Bektasch, Leipzig, 1927Google Scholar (Türkische Bibliothek xxv), p. 80 seq., shows our saint as claimed by the Bektaī, just as it is the case with Ṣari Ṣaltiq (ib., p. 73).

^{page 659 note 1} See Fu'ād, Köprülüzāde M. in Darülfünūn Edebiyāt Fakültesi Mejmū'asi, ii, 1922, p. 292 seq.Google Scholar, and more fully in his Influence du Chamanisme Turco-Mongol sur les Ordres Mystiques Musulmans, Istanbul, 1929, pp. 14–17Google Scholar; cf. also the same in Belleten, vi, 1943, p. 431, n. 1Google Scholar, where Prof. F. Köprülü promises a monograph on Baraq and also a study on Sari Ṣalti q which will make use of a newly discovered Ṣaltiq-nāme of the late 15th century.

^{page 659 note 2} Moravcsik, G., Byzantinoturcica, Budapest, 1942–1943, iiGoogle Scholar, records S.v. Παρ⋯κ two baptized Tatar women who died in 1280 and 1308 respectively, and S.v. Βαρ⋯κος an Ottoman army chief of the early 15th century as well as τ⋯ν Β⋯ραγκον τ⋯ν Κώναταν in an Athos document of 1292. For Mongol and Turkish rulers of the name of Baraq see El. S.v., and alīl Edhem, Düvel-i islāmīye (index) or de Zambaur, E., Manuel de généalogieGoogle Scholar (index, s.v. Borâq). For the central asiatic Baraq ān, who in the years 1422–7 dominated the events in the realm of the Golden Horde, see Spuler, B., Die Goldene Horde, Leipzig, 1943, p. 156 seq.Google Scholar

^{page 659 note 3} For this is doubtless its meaning. In classical antiquity the saliva was considered as a means of conferring spiritual power; see e.g. Davreux, J., La légende de la prophétesse Cassandre, Paris, 1943, p. 69Google Scholar, and ProfGoossens, R.' remarks thereon in L'Antiquité Classique, xiii, 1944, p. 178 seq.Google Scholar

^{page 659 note 4} Okiç, M. Tayyib, ‘Sari Saltuk'a ait bir fetva’ in: Ankara Üniversitesi Ilâhiyat Fakültesi Dergisi, i, 1952, pp. 3–13Google Scholar, giving (p. 10) the text of the fetva: the question submitted by Sultan Sulaimān I: ‘Is the person known by the name of Ṣari Ṣaltiq a saint? ‘and Abū's-Su'ūd's reply:

^{page 659 note 5} See above, p. 655, n. 3.

^{page 660 note 1} Pachymeres, , i, p. 258Google Scholar,1. 7: ὡς κα⋯ τος υἱ⋯σιν ⋯κε⋯νου [το σουλτ⋯ν] τ ἰδ⋯ῳ μεταδιδ⋯ναι τν ⋯γιασμ⋯των κελε⋯σειεν.

^{page 660 note 2} Ib., p. 259,11. 11–13: τ⋯ δ⋯ τ σουλτ⋯ν κ⋯ τος νἱ⋯σιν ὡς Χριστιανος προςø⋯ρεσθαι ⋯ρχιερ⋯ως μ⋯ν μαρτυροντος το Πισσιδ⋯ας ⋯νε⋯θυνον ομαι.

^{page 660 note 3} Gregoras, , i, p. 94, 11. 10–19.Google Scholar

^{page 661 note 1} Cantacuzenus, , i, p. 269, 1. 21, Bonn: ρχε δ⋯ αὐτς [τς ⋯κροπ⋯λεως] Λυζικ⋯ς Γεώργιος ⋯κ Βεῤῥο⋯ας, and pp. 271, 1. 12–272, 1. 22.Google Scholar

^{page 661 note 2} Ib., iii, pp. 161, 1. 7–163, 1. 3. His appointment as commandant of Edessa, , ib., pp. 129Google Scholar ult.–130, 2: καταλιπν… ⋯γεμ⋯να τ π⋯λει [Ἐδ⋯σσῃ] Λυζικ⋯ν Γεώργιον, κρ⋯τιστον δοκοντα τ⋯ πολ⋯μια κα⋯ συνετ⋯ν Probably Lyzikos was one of those notables of Berrhoia who had fled before the Servian occupation to the emperor, with whose army they returned to the reconquest of their town; he was certainly among the many men of Berrhoia who then went with the emperor to recapture also the neighbouring Edessa from the Servians (ib., p. 123, 11. 1–5, and p. 127,1. 2).

^{page 661 note 3} Sp. Lampros, Βραχ⋯α Χρονικ⋯, ed. Amantos, K. I., Athens, 1932, no. 29, 1. 7Google Scholar; no. 42, 1. 29; no. 49, 1. 46: 8 May 6895 A.M. = 1387 (Ἀχουμ⋯της is obviously a corruption of Ἀμουρ⋯της).

^{page 662 note 1} I owe this information to my friend and colleague Prof. Halil Inalcik, of Ankara University, who is engaged in preparing the edition of this defter; for the present see the brief information he gives on this matter in Belleten, XV, 1951, p. 650.Google Scholar

^{page 662 note 2} It seems to me very probable that in the important movement connected with the sheikh's name the ‘people of Kaikāūs’ played a foremost rôle, politically as well as ideologically, and that Bedreddīn was involved in it because of his descent from Kaikāūs. (Since Prof. Babinger, F.'s inspiring monograph in Der Islam, xi, 1921, pp. 1–106CrossRefGoogle Scholar, the Bedreddīn movement has been the subject of many studies of which the most recent is Kissling, H. J., ‘Das Menāqybnāme Scheich Bedr ed-Dīn's, des Sohnes des Richters von Samāvnā,’ in ZDMG., e, 1950, pp. 112–176.)Google Scholar

^{page 662 note 3} Gregoras, , i, pp. 227, 1. 4–233, 1. 13; pp. 244, 1. 16–249, 1. 2Google Scholar; pp. 254, 1. 2–258, 1. 14; pp. 262, 1. 20–269, 1. 23, Bonn.

^{page 662 note 4} Ib., p. 228, 11. 22–4: κα⋯ δ⋯ διαπρεσβε⋯ονται περ⋯ συμμαχ⋯ας πρ⋯ς ⋯ντιπ⋯ρας οἰκοντας τν Το⋯ρκων κα⋯ λαμβ⋯νουαι τ⋯τε μ⋯ν πεντακοσ⋯ους ⋯πλ⋯τας τρισχ⋯λιοι ⋯ντες αὐτο⋯.

^{page 663 note 1} Ib., p. 229, 11. 11–17; cf. above, p. 657, n. 1.

^{page 663 note 2} Ib., p. 232, 11. 11–13: πρ⋯ς Κατελ⋯νους αὐτομολοσιν οἱ ῥηθ⋯ντες Τουκ⋯πουλοι κα⋯ ⋯σμ⋯νως δεχ⋯ντες τος το Χαλ⋯λ ὡς ⋯μοφ⋯λοις συγκατελ⋯γησαν Το⋯ρκοις. Χαλ⋯λ δ᾽ ⋯ τν Το⋯ρκων ⋯ρχηγ⋯ς ὠνομ⋯ζετο.

^{page 663 note 3} Ib., p. 248, 11. 5–15.

^{page 663 note 4} Ib., p. 248, 11. 18–20: συστ⋯ντες ον οἱ τν Τουρκικν ⋯γεμονε⋯οντες, ὅ τε Μελ⋯κ κα⋯ ⋯ Χαλ⋯λ, τ τν Κατελ⋯νων ⋯ρχηγ λ⋯γους περ⋯ τς ἅμα εἰρ⋯νῃ διαλ⋯σεως κεκιν⋯κασιν.

^{page 663 note 5} Gregoras, , i, pp. 254, 1. 2–255, 1. 14Google Scholar: Οἱ δ⋯ Τορκοι μετ⋯ τ⋯ διαστναι τν Κατελ⋯νων εἰς δ⋯ο σχ⋯ζονται μο⋯ρας κα⋯ οἱ μ⋯ν τ Χαλ⋯λ, οἱ δ⋯ τ Μελ⋯κ ἕπονται. ⋯ μ⋯ν γ⋯ρ Μελ⋯κ … ⋯ντεθεν εἵλετο μλλον προσι⋯ναι καλοντι τ κρ⋯λῃ Σερβ⋯ας ἢ ῾Ρωμα⋯οις εἰς ⋯ψιν ⋯λθεν … ⋯ δ⋯ Χαλ⋯λ … ⋯ζ⋯τει σπε⋯σασθαι ῾Ρωμα⋯οις ⋯π⋯ δυον, ἵνα τε τ⋯ περ⋯ τ⋯ν Χριστο⋯πολιν ⋯φεθεεν στεν⋯ διελθεν, κα⋯ ἵνα ῾Ρωμαïκας ναυσ⋯ διαπεραιωσ⋯μενοι τ⋯ν ῾Ελλησπ⋯ντιον πορθμ⋯ν ⋯π⋯λθωσιν ⋯ς τ⋯ οἴκοι μετ⋯ τς λε⋯ας ς ⋯πεφ⋯ροντο π⋯σης. … ⋯ βασιλεὺς … διεβ⋯βαζεν αὐτοὺς ⋯κ Μακεδον⋯ας εἰς Θρᾴκην ἄχρι το ῾Ελλησπ⋯ντου. ⋯νταυθο δ⋯ … οὔτε νας ⋯δ⋯δουν, αἵ πρ⋯ς ᾽Ασ⋯αν αὐτοὺς διαβιβ⋯ζειν ἔμελλον, κα⋯ νυκτ⋯ς ⋯πιθ⋯σθαι σφ⋯σιν ⋯σκ⋯ψαντο.

^{page 664 note 1} Ib., pp. 255, 1. 14–256, 1.3: τοτο τοὺς Το⋯ρκους οὐκ ⋯λελ⋯θει … αἱροσιν ἓν τν παρακειμ⋯νων φρουρ⋯ων, κ⋯κεθεν ὡς ⋯ ⋯ρμηρ⋯ου πρ⋯ς μ⋯χας κα⋯ πολ⋯μους ὡπλ⋯ζοντο … π⋯μψαντες ⋯ς ᾽Ασ⋯αν πλε⋯στην ⋯κ τν ⋯μοφ⋯λων βαρβ⋯ρων ⋯ν βραχε διεπεραιώσαντο συμμαχ⋯αν. ὅθεν ⋯ε⋯ παρει⋯ντες … ⋯δουν τ⋯ν χώραν. Ib., pp. 257, 1. 11–258, 1.8: Michael IX's defeat. Here we read (p. 257, 11. 14–17) that the Turks have their women with them. Ib., p. 258, 11. 11–14: alīl's triumph (… ⋯ βασιλικ⋯ καλ⋯πτρα … τ ⋯αντο κεφαλ τ⋯ν Χαλ⋯λ ⋯πιθ⋯ντα φασ⋯ σκωπτικο⋯ς τε κα⋯ εἴρωνας λ⋯γους ⋯φι⋯ναι κατ⋯ το βασιλ⋯ως).

^{page 664 note 2} Ib., p. 265, 11. 15–18: Χιλ⋯ους πεζοὺς κα⋯ διακοσ⋯ους ἱππ⋯ας ⋯ Χαλ⋯λ ⋯πολε⋯μενος πρ⋯τριτα π⋯πομφε, π⋯σας τ⋯ς περ⋯ τ⋯ν Βιζ⋯ην καταστρεψομ⋯νους Χώρας….

^{page 664 note 3} Ib., pp. 267,1. 23–268, 1. 14. The action starts with a measure to prevent alīl from getting reinforcements from Anatolia: π⋯μπει δ⋯ κα⋯ π⋯ντε τρι⋯ρεις ⋯ βασιλεὺς εὐθὺς, ἴνα παραπλ⋯ουσαι τ⋯ν ῾Ελλ⋯σποντον παραφυλ⋯ττωσι, μ⋯ λ⋯θῃ παρ⋯ τν ⋯ντιπ⋯ρας βαρβ⋯ρων ⋯ληλυθυα δ⋯ναμις τ Χαλ⋯λ.

^{page 664 note 4} Ib., pp. 268, 1. 15–269, 1. 23: Only those are spared who fall into the hands of the Genoese of Galata who, commanded by their podestà, are present with their ships as allies of the emperor: (p. 269, 1. 21) ἔπειτα τοὺς μ⋯ν ⋯ αὐτν [τν Το⋯ρκων] ⋯κ⋯μισαν [οἱ Λατνοι] τ βασιλε, τοὺς δ⋯ διενε⋯μαντο.

^{page 664 note 5} Pachymeres, , ii, pp. 523,1. 18–524, 1. 2Google Scholar: τ⋯ ⋯κ παλαιο Περσικ⋯ν, οὔς κα⋯ Τουρκοπο⋯λους ὠ⋯μαζον, … περ⋯ … τοὺς οἰκε⋯ους κοσμ⋯τορας.

^{page 664 note 6} Ib., p. 574, 1. 5; see above, p. 657, n. 2.

^{page 664 note 7} Ib., p. 590, 1. 10 seq.: Τουρκ⋯πουλοι ⋯π⋯κλινον πρ⋯ς Καλλιο⋯πολιν ἄμα γυναι⋯ κα⋯ παισ⋯.

^{page 664 note 8} And their Turks, whose presence in the Catalan camp Pachymeres had mentioned before the battle of Aproi, , p. 550, 1. 2Google Scholar: προσεκ⋯λεσαν γ⋯ρ κα⋯ Περσικ⋯ οἱ Κατελ⋯νοι.

^{page 664 note 9} Ib., p. 590,11. 11–14.

^{page 664 note 10} Ib., p. 591,1. 17.

^{page 664 note 11} Ib., p. 591, 11. 1–7: Ἐν τοσο⋯τῳ δ⋯ κα⋯ τις τν διαπεραιωθ⋯ντων Περσν, Ἰσα⋯κ Μελ⋯κ, σατρ⋯πης ὢν κα⋯ πολλοὺς ἄγων, π⋯μπων πρ⋯ς κρυøηδ⋯ν παρ⋯ πσαν τν Κατελ⋯νων αἴσθησιν…

^{page 664 note 12} Ib., pp. 591, 1. 12; 608, 1. 18; 612, 1. 16.

^{page 664 note 13} Ib., p. 612, 11. 11–13.

^{page 665 note 1} Ib., p. 612, 11. 13–16.

^{page 665 note 2} Ib., pp. 612, 1. 18–613, 1. 7.

^{page 665 note 3} Ib., pp. 631,1. 13–632, 1. 17.

^{page 665 note 4} Ib., p. 632, 1. 11: τ⋯ν σΦν ⋯ξηγο⋯μενον Ταχαγτζι⋯ριν, ὃς ἰδ⋯ως γε τοὺς Τουρκοпο⋯λους. For the variants of the name Ταγχατζι⋯ρις (probably the most correct form), Ταχατζι⋯ρις, etc., see Moravcsik, G., Byzantinoturcica, iiGoogle Scholar, s.v. Ταγχατζι⋯ρις. in Aqsarayī, ed. O. Turan, passim (see index), is obviously the same name though not denoting the same person.

^{page 665 note 5} Ib., p. 633,11. 2–16.

^{page 665 note 6} Jireček, C., Staat und Gesellschaft im Mittelalterlichen Serbien, i, Vienna, 1912 (Denkschriften Ak. d. W. Wien, lvi), p. 78 seq.Google Scholar

^{page 665 note 7} Ib., p. 631, 1. 13: κατ⋯ Пηγ⋯ς, п⋯λιν пαραθαλ⋯σσιον, νυμΦαγωγ⋯α ηὺτρ⋯пιοτο, shows that the bride was already in Pegai.

^{page 666 note 1} Ib., pp. 601, 1. 11–603, 1. 11: ν χ⋯ριν ὑпεραλγοντες Τουρκ⋯пουλοι (οἱ γ⋯ρ αὐτν οἰκεοι пαρ' ⋯κε⋯νοις [τος 'Αλανος] ⋯ντες συνεξηλα⋯νοντο) and: пλ⋯ν γ⋯ρ τινων ⋯λ⋯γων (who had been sent to the emperor) пσαι αἱ γυνακες σὺν τ⋯κνοις. τν Τουρκοпο⋯λω пαρ' ' Αλανν κατεἰχοντο.

^{page 666 note 2} Ostrogorsky, G., Geschichte des byzantinischen Staates, München, 1940, p. 354.Google Scholar

^{page 666 note 3} He appears in the sources occasionally as Ya'qūb Eje Bey.

^{page 668 note 1} Γ. Δ. Μπαλ⋯στζεφ, ⋯ αὐτοκρ⋯τωρ Μιχα⋯λ Η΄, etc. (see above, p. 641, n. 2), p. 19, n. **. Balaschev seems to have been afraid of his derivation since he chose to hide it in a note. In support of it he quotes toponyms of the Dobruja where an original k has changed into g, e.g. Καλλι⋯κρα, which appears in a document of 1320 as Γαλι⋯γρα (for this he quotes, p. 15, Miklosich-Müller, , Acta Patr., i, p. 95)Google Scholar. Indeed, it is the ‘Gelara’ (Evliya, , ii, p. 133)Google Scholar and even ‘Gülgrad’ (Ḥājjī alfa; Hammer, Rumeli und Bosna, p. 27)Google Scholar of the Ottomans and the ‘Gelare’ (Mutafčiev, , op. cit., p. 37, n.)Google Scholar of to-day. As to the transition ay > ā (> a) instead of ay > ī, which would be the normal (Kowalski, , Les Turcs et la langue turque de la Bulgarie du Nord-Est, p. 19)Google Scholar it is sufficiently explained by the influence of the back vowels which follow.

^{page 668 note 2} E.g. I had to abstain from dealing with the Gadjal of the Deli-Orman, the neighbours of the Gagauz to whom linguistically they are so closely related that they must be of the same origin; they are distinct from the Gagauz only in that they are Muslims—though Muslims of a very unorthodox kind.