^{page 417 note 1} Some 75 km. to the N.E. of Solṭānābād, and some 50 km. to the S.W. of Sāva.

^{page 417 note 2} I have lost my notes of it, but presumably it was identical or similar, to the Khalajī of Kūshkak (on the road Tehran-Hamadān) described by Brugsch, , Reise der Preussischen Gesandschaft, Leipzig, 1862–1863, i, 337:Google Scholarkié “a room”, berr “a door”, azbé “a dog”, zāg “a boy”, agīr “fire”, jīr “below”, beshan agirem “I shall seize”, vergibukhma “I shall eat”, burana “come back!” Cf. Minorsky, Tāt in EI.

^{page 418 note 1} i.e. learning to read.

^{page 418 note 2} Abbreviations: Deny, Grammaire de la langue turque, 1921; Āz., Āzarbāyjān Turkish; K., Kondūrūd; Kh., Khoräkäbād; Ott., Ottoman Turkish; P., Pougerd; Q.A., Qanbar-'Ali.

^{page 419 note 1} Evidently an abridged form from the verb qairmaq “to do”: *qairay, as below: chixay.

^{page 421 note 1} Even in Persian translated as: sārī che? But the word sari seems to be Turkish: tañ sarï “eastwards”, cf. Deny, § 906. However, its use in Khalaji is uncommon.

^{page 423 note 1} Köntär- seems to stand both for “to send” and for “to make to burn” (from kön-).

^{page 424 note 2} See n. 1, p. 424.

^{page 424 note 1} Qanbar-‘AIī may have used these forms in -mish as the usual “pidgin” forms, as they are usually used in Russo-Turkish “soldiers’ slang”. It is possible, however, that the forms in -mish stand here instead of the gerund in -üp, -ip, as suggested in my translation.

^{page 425 note 1} Mélanges Asiatiqties, Petersburg, S., 1863, iv, 63–74Google Scholar. I regret not to have been able to consult the recent Turkman grammar by Potseluyevsky, Ashkhabad, 1929.

^{page 425 note 2} Cf. in Āz. ged-änmirä “he cannot go”; cf. anniyamirux in Foy, K., Azerbajganische Studien in MSOS., 1904, p. 238Google Scholar, paenult.

^{page 425 note 3} Perhaps in the sense of “I am of the children of P.”, etc., unless -chä be of a different origin here.

^{page 425 note 4} Unless -ch- represents the suffix chä (vide supra), namely *xalaj tāifäsi-chä-ik'(?): “we are of the tribe Kh.” (literally: in the tribe ?).

^{page 426 note 1} Q.A.'s predilection to ch- sound (qairchom, qalmich, -chik', -chiz) needs to be checked on the spot.

^{page 426 note 2} On the history of the Khalaj see Marquart, , Erānšahr, 251–Google Scholar, idem, Das Reich Zābul in Festschrift Sachau, p. 258, note 1; Barthold, Ocherlc istorii Turkm. naroda, 8; idem, 12 Vorlesungen, 1935, 100; Minorsky, Ḥudūd al-' Ālam, 347–8; Validi, A. Z., ZDMG., 90, 1936, 34Google Scholar.

^{page 426 note 3} Abū 'Abdillāh Muḥammad b. Aḥmad al-Khuwārizmī, Mafātīḥ al-'ulūm, 119.

^{page 426 note 4} However, the Khallukh too were in touch with the Hephthalites, see Ḥudūd al-'Ālam, 288.

^{page 426 note 5} Erānšahr, 253: “Ein Überrest der Hephthaliten.…” “Eine Abzweigung eines sehr alten türkischen Volkes.”

^{page 427 note 1} In fact, Zacharias Rhetor, supposed to be identical with the Bishop of Mitylene who took part in the council of A.D. 536 and was dead before 553, wrote in Greek. The Syriac author follows this text down to his ch. vi; from chap, vii on, he uses other sources carrying the history down to A.D. 568–9, see Wright, W., Hist, of Syriac Liter., pp. 107–8Google Scholar [Russian transl. with additions by Kokovtsov, , 1902, pp. 74–5Google Scholar], and the German translation of the Syriac original by Ahrens, and Leipzig, Krüger, 1899, Introduction. The geographical passage (cb. xii, 7Google Scholar), ibid., 251–3, accompanied by Gelzer's, commentary, pp. 381–3Google Scholar.

^{page 427 note 2} As preserved in Menander Protector's fragments, see Müller, , Fragmenta hist. Graec., iv, 227–230Google Scholar.

^{page 428 note 1} For more detail see my commentary on the Ḥudūd al-‘ Ālam, pp. 347–8.

^{page 428 note 2} The forms in Khurdadhbih, I., 31, and Mas'ūdī, Murūj, v, 302Google Scholar, must stand for the original form of the name Khallukh, namely Kharlukh (Qarluq). In Das Reich Zābul, p. 258, Marquart himself admitted this possibility.

^{page 428 note 3} The only possibility I can see is the pronunciation of the Turkish a as ả > zo under the influence of some Iranian language (Soghdian on the Jaxartes ?). In the present day Sart (“Uzbek”) dialects qaqmaq sounds qōqrnag.

^{page 428 note 4} Über d. Volkstum der Komanen, 73; Wehrot und Arang, 93, note 3.

^{page 428 note 5} chäbär, a Mongolian word!

^{page 429 note 1} See Pelliot, in T'onng-Pao, xxvii, 1930, p. 297Google Scholar. In the latest edition of the Oghuzname by Bang, and Rachmati, , in Sitzungsber. d. preus. Akad., 1932, 683–726: Tömürtü QayulGoogle Scholar.

^{page 429 note 2} Firdausī, Shāh-nāma, ed. Mohl, , v, 682, ed. Tehran, 1314, vii, 2,202,Google Scholar in describing Bahrāin Gōr's exploits in Transoxiana, says:-

Bar-āvard mīlī za sang-u za gaj

Ki kas rā ba-Irān za Turk-u Khalaj

Nabūdī guδar juz ba farmän-i shāh

Hamān nīz Jayhūn miyānjī ba rāh

Here the Khalaj seem to be distinguished from the Turks, but Firdausī's geography is often fantastic.

^{page 429 note 3} In this ease Kāshghari has in view the particle aj which strengthens the imperative, vide ibid., i, 38, line 11.

^{page 429 note 4} One might say that the story places them somewhere to the east of the Jaxartes (from which Alexander was advancing), but no great weight can be attached to the details of the legend.

^{page 430 note 1} The additional names of the Ghuz clans found in Rashīd al-dīn, i, 34, namely Yayïrlï, Qïrïq, and Qarqïn cannot refer to the Qalaj who are mentioned separately.

^{page 430 note 2} Ed. Bérézine, , Trudï V.O., vii, 7Google Scholar: “Uyghur, Qanqlï, Qïpchaq, Qarluq, Qalaj, Aghajeri.”

^{page 430 note 3} Cf. ibid., 24, a similar story about the Qarluq who received their name from Oghuz-khan while he was returning from his campaign in Ghūr and Gharchistān, i.e. from a region considerably to the east of Persia.

^{page 430 note 4} In the account of the province of Dāvar on the Hilmand.

^{page 430 note 6} After all, Mas'ūdī's vague passage may even not refer to the Khalaj but only to the Kharlukh and the Turkmans (often quoted alongside with the Khalaj).

^{page 431 note 1} Probably based on the history of Ibn al-Azhar al-Akhbārī. see Barthold, , Zur Geschichte der Ṣaffāriden in Oriental. Siudien Th. Nöldeke, 1906, pp. 173, 186Google Scholar.

^{page 431 note 2} Siyāsat-nāma (485/1092), eh. xxvii, p. 96.

^{page 431 note 3} Gardīzī, 56. The text has Turkān-i ṣulḥ but the editor has already suggested the reading *Khallukh. I admit the necessity of the emendation, but, in view of the circumstances, I prefer *Khalaj.

^{page 431 note 4} Aḥlās al-ẓuhūr.

^{page 432 note 1} The alternance of initial i- and yi is frequent; cf. Inal/Yinal.

^{page 432 note 2} But certainly not all the Khalaj.

^{page 433 note 1} i.e. Khalaji. In Indian pronunciation the middle short vowel of a trisyllabic word is regularly omitted (shafaqat > shafqat), while a monosyllabic word ending in two consonants becomes bisyllabic (fahm > faham).

^{page 433 note 2} His father had the Turkish title yughrush, see Köprülü, M. F., Zur Kentniss der alttürkischen Titulatur, in K“orösi Csoma Archivum, 1938, Ergänzungsband, p. 339Google Scholar, who quotes Tārīkh-i Firishta, i, 152, 155.

^{page 433 note 3} Or with a further reduction of the vowel: Ghilzae, in Persian Ghiljā'ī.

^{page 433 note 4} See Longworth Dames, Afghānistān and Ghilzai in El. The author seems not to have realized the weight of the earlier historical evidence and disbelieved the possibility of the transformation Khalaj > Ghilzai, fully admitted by other collaborators of the El. (Barthold, Sir W. Haig); cf. also Marquart, op. cit., 253. In fact there is absolutely nothing astonishing in a tribe of nomad habits changing its language. This happened with the Mongols settled among Turks and probably with some Turks living among Kurds. [Sir Haig, W. in the Cambridge History of India, III, 90Google Scholar, gives a pertinent reply to Raverty: “If the Ghilzay be not Khaljis it is difficult to say what has become of the latter.”]

^{page 433 note 5} Cf. Ṭabarī, iii, 1416: Ghamish < Turkish gamïsh “a reed”.

^{page 434 note 1} Bustān al-siyāḥa, Tehran, , 1315/1897, p. 273Google Scholar.

^{page 434 note 2} Cf. Abul-Ghāzī's additional passage quoted above, p. 430.

^{page 434 note 3} Particularly in the neighbourhood of Bust on the Hilmand. On the Afghan Ghilzay see El.

^{page 434 note 4} Some fragments of the tribe may have travelled very far afield. I am tempted to render the Khalaj responsible for the name of the Russian town Kalach, situated on the Don where the latter comes the nearest to the Volga, i.e. on the road undoubtedly followed by innumerable nomad migrations.

^{page 434 note 5} See Samoylovich, Khivinslciye marshrutï XIX veka po Qara-qumam, in the presentation volume to Oldenburg, S. F., 1934, p. 462Google Scholar.

^{page 434 note 6} Materialï po istorii Turkmen, 1938, ii, 422Google Scholar.

^{page 434 note 7} Tumansky, Ot Kaspiyskago morya, St. Petersburg, , 1896, p. 120Google Scholar.

^{page 435 note 1} Ẓafar-nāma, ii, 573. Kara-rūd (< Karaj-rūd) is the river of the ancient Karaj Abī-Dulaf, near the present day Solṭānābād. Charrā and Farāhān are districts of Solṭānābād.

^{page 435 note 2} Eastern Persian Irak, 1896, 129Google Scholar.

^{page 436 note 1} Inhabited by Khalach-e Hertī.

^{page 436 note 2} 1 farsakh from Jehrud, 1 f. from Āshtiyān, 3—4 farsakhs from Kondūrūd. In 1917 half of it was possessed by small owners (khorde-mālek), while the other half belonged to Sheikh Murtaḍā of Tehran.

^{page 437 note 1} See my article Shāhī-sevän in El. In the course of the same journey in December, 1906, I spent a night at the village of Sahlābād (near Sāva) inhabited by Baghdād Shāhsevans. We spoke in “Āzarbayjān Turkish”, but I noticed the following unusual forms in their speech: g'örmiyipik' “we have not seen”, g'örürän and görüriz (instead of g'örürsän and g'örürsiz), babayiñ gordüm “I saw your father”, aparayiz (instead of aparsaz) “if you take away”