^{page 367 note 1} Tehran, 1950.

^{page 367 note 2} See Rosen, V., Les manuscrits persans de l'Institut des Langues Orientates, Petrograd, 1886, No. 26, pp. 146–59.Google Scholar

^{page 367 note 3} Ankara, 1951. A detailed account of the reign of Sanjar will be found in Büyük Selçuklu imperatorluğu tarihi, II, Ankara, 1954,Google Scholar by the same author.

^{page 367 note 4} Lubábu'l-albāb of Muḥammad ‘Awfí, ed. Browne, E.G. and Muḥammad Qazwīnī, Mīrzā, London, Leyden, 1906, I, 78–80.Google Scholar

^{page 368 note 1} For these and other details see the Introduction to the Persian edition of the ‘Atabat alkataba.

^{page 368 note 2} Two at least of these letters belong to the period after the death of Sanjar and must have been added after the original compilation (see Introduction to the Persian text).

^{page 368 note 3} ‘AK, 52.

^{page 368 note 4} ‘AK, 83.

^{page 368 note 5} ‘AK, 53.

^{page 368 note 6} cf. ‘AK, 79.

^{page 368 note 7} ‘AK, 61.

^{page 368 note 8} ‘AK, 45–6.

^{page 369 note 1} ‘AK, 43.

^{page 369 note 2} ‘AK, 48.

^{page 369 note 3} ‘AK, 67.

^{page 369 note 4} ‘AK, 46, 66.

^{page 369 note 5} ‘AK, 29.

^{page 369 note 6} ‘AK, 20.

^{page 369 note 7} ‘AK, 64.

^{page 370 note 1} ‘AK, 78.

^{page 370 note 2} ‘AK, 44.

^{page 370 note 3} ‘AK, 23, 27.

^{page 370 note 4} ‘AK, 45.

^{page 370 note 5} ‘AK, 50.

^{page 370 note 6} ‘AK, 56.

^{page 370 note 7} ‘AK, 46.

^{page 370 note 8} ‘AK, 47.

^{page 370 note 9} ‘AK, 72, 73.

^{page 370 note 10} ‘AK, 52.

^{page 370 note 11} ‘AK, 87.

^{page 370 note 12} ‘AK, 35.

^{page 371 note 1} ‘AK, 76.

^{page 371 note 2} ‘AK, 17.

^{page 371 note 3} ‘AK, 40.

^{page 371 note 4} ‘AK, 60.

^{page 371 note 5} ‘AK, 21.

^{page 371 note 6} cf. ‘AK, 10, 11, C5, and also my article ‘Quis custodiet oustodes’, Studia Islamica, V and VI, 1956.Google Scholar

^{page 371 note 7} ‘AK, 13–14.

^{page 371 note 8} ‘AK, 78.

^{page 371 note 9} For a further discussion of this see the text of my inaugural lecture, Islamic society in Persia, London, School of Oriental and African Studies, 1954.Google Scholar

^{page 372 note 1} Ibn ul-Athīr, (ed. Tornberg), X, 180.

^{page 372 note 2} ‘AK, 75.

^{page 372 note 3} cf. ‘AK, 30, 35, 39, 82. The deed of investiture for Burhān ud-Dīn as qāḍī of Nīshāpūr refers to him also as brother (‘AK, 57).

^{page 373 note 1} Professor Claude Cahen in his valuable article, ‘Évolution de l'iqṭa‘’, has pointed out that wilāya, iqṭā‘, and mulk were carefully distinguished in the Eastern Persian documents belonging to this period (Annales, Économie-Sociétés-Civilisations, 1953, 43).

^{page 373 note 2} cf. C. Cahen, ‘Évolution de l'iqṭa‘’.

^{page 373 note 3} ‘AK, 84.

^{page 373 note 4} ibid.

^{page 373 note 5} ibid.

^{page 373 note 6} ‘AK, 39–40.

^{page 374 note 1} In spite of the fact that Bundārī states that Qumāj and his son were killed in the engagement with the Ghuzz in which they captured Sanjar in A.H. 548 ( Zubdat an-nuṡrat wa nukhbat al-‘uṡrat, ed. Houtsma, M.T., Leyden, 1889, p. 283 Google Scholar), it is clear from the documents in the ‘Atabat al-kataba that Qumāj and his son, Abū Bakr, were killed in a previous engagement. Bundārī does in fact mention that one of Qumāj's sons was killed by the Ghuzz prior to 548 (p. 282). Ibn ul-Athīr states that Qumāj fled from the Ghuzz to Marv prior to 548 (XI, 116), and that he and his son Abū Bakr were killed in the battle of 548 (XI, 118); he also states that Muḥammad b. Abī Bakr [b.] Qumāj was with Sanjar during this engagement (XI, 118). Rāwandī, however, states that Qumāj and his son were killed by the Ghuzz prior to the engagement of 548 ( Rāḥat-us-ṡudūr wa āyat-tis-surūr, ed. Iqbāl, Muḥammad (GMS), London, Leyden, 1921, 117–18Google Scholar). See also my Landlord and peasant in Persia, O.U.P., 1953, 58–9.

^{page 374 note 2} ‘AK, 76.

^{page 374 note 3} ‘AK, 84.

^{page 374 note 4} ‘AK, 76–7. Amlāk are private estates and hence transmitted by inheritance. Qumāj, as stated above, was a mamlūk and consequently on his death his estates in theory, if not in practice, would have escheated to the sultan, and Qumāj's descendants could not legally have taken possession of them unless they had received a formal grant from the sultan.

^{page 374 note 5} ‘AK, 39–40.

^{page 374 note 6} It appears that this ‘journey to Transoxania’ coincided with some disorganization in the affairs of the empire, and consequently it seems more likely that it refers to the expedition of A.H. 535–6 in which Sanjar was defeated by the Qarā Khiṭāy, rather than to the shorter and more successful expedition of A.H. 524.

^{page 375 note 1} ‘AK, 21.

^{page 375 note 2} ‘AK, 69.

^{page 375 note 3} ‘AK, 68. This document, if I am right in dating it after Sanjar's return from his defeat by the Qarā, Khiṭāy, was issued at a time when conditions were far from normal. The khutba had been read in Nīshāpūr in the name of the Khwārazmshāh Atsiz in A.H. 536, and there had been disorder and bloodshed in parts of Khurāsān and the neighbourhood.

^{page 375 note 4} ‘AK, 77.

^{page 375 note 5} ‘AK, 73.

^{page 375 note 6} ‘AK, 20. Conditions when this grant was made also appear to have been somewhat unsettled, because the document instructs Mas‘ūd to remember that it is not permissible to punish or reproach anyone in the absence of some clear and evident fault, ‘so that those who were in service should not be broken at heart and those who had fled from service should turn back encouraged’. Mas‘ūd was in Gurgān on a number of occasions, and it is not clear on which this document was issued. He appears to have been sent twice by Sanjar to Gurgān between the years A.H. 513 and 521 (see Mar‘ashī, ຒahīr ud-Dūn, Tārīlch-i Ṭabarestān wa Rūyān ua Māzandaān, ed. ‘Shāyān, Abbās, Tehran, 1955,Google Scholar 160 ff.). He went to Khurāsān with Sanjar c. 522 (Bundārī, 154) and left Khurāsān, where he had been with Sanjar, in 524. He then met Maḥmūd b. Muḥammad in Kirmānshāhān and the latter assigned to him Ganja (Ibn ul-Athīr, X, 469). In the following year, however, Mas'ūd appears to have been in Gurgān again, because Ibn ul-Athīr states that on Maḥmūd's death he set out from Gurgān for Tabrīz (X, 474).

^{page 376 note 1} There is, for example, mention of an iqṭā‘ of the mustawfī‘ al-mamālik in a document dated A.H. 563 (Munsha'āt-i ‘ahd-i Saljūqī wa Khwārazmshāhīan wa avā'il-i ‘ahd-i Mughul, f. 22b).

^{page 376 note 2} Thus Ḥamdullah Mustawfī states that Khwārazm was ‘allocated’ to (dar wajh-i) the ṭashtkhāna, and since Anūshtegīn was ṭashtdār (ewer-bearer) the office of shiḥna of Khwārazm belonged to him ( The Ta'rīkh-i-guzīda, ed. Browne, E.G. (GMS), Leyden, London, 1910, 486).Google Scholar

^{page 376 note 3} ‘AK, 85.

^{page 376 note 4} cf. also a diploma for the office of dādbeg issued by the dīwān of Abū'l Muຓaffar Ṭughril Shāh b. Muḥammad, the Seljūq ruler of Kirmān (551–63/1156–67), which mentions the iqṭā‘ and nānpāra of the grantee and also the dues (rusūm) of the office of dādbeg (Munsha'āt-i ‘ahd-i Saljūqī wa Khwārazrnshāhīān wa avā'il-i ‘ahd-i Mughul, f. 30a).

^{page 376 note 5} ‘AK, 19.

^{page 376 note 6} ‘AK, 18.

^{page 377 note 1} ‘AK, 19.

^{page 377 note 2} ‘AK, 19.

^{page 377 note 3} ibid.

^{page 377 note 4} ‘AK, 20.

^{page 377 note 5} cf. ul-Mulk, Niຓām, Siyāsat-nāma, ed. Schefer, , Paris, 1891, 18.Google Scholar

^{page 377 note 6} ‘AK, 20.

^{page 377 note 7} Ra‘āyā, lit. subjects. The term seems to be used in the documents often to designate the people in general as opposed to the military and official classes, and I have translated it where it appears to be used in this way by people (rather than subjects). Ra‘iyyat seems often to be used to designate the ‘civil’ population as opposed to the military.

^{page 377 note 8} ‘AK, 21.

^{page 378 note 1} ‘AK, 20.

^{page 378 note 2} ‘AK, 18.

^{page 378 note 3} e.g. in the diploma for ‘Imād ud-Dīn Abū'l Fatḥ b. Abī Bakr b. Qumāj as governor and shiḥna of Balkh (pp. 78–9). cf. also the abstract of a diploma written on behalf of Ṭughtegīn for the sons of Muḥammad b. as-Ṣūfī for the office of ra'īs of Damascus ( al-Qalānisī, Ibn, History of Damascus, ed. Amcdroz, H.F., Beyrut, 1908, 144–5Google Scholar).

^{page 378 note 4} ‘AK, 15.

^{page 378 note 5} See above, p. 374, n. 6.

^{page 378 note 6} ‘AK, 31.

^{page 378 note 7} ‘AK, 31–2.

^{page 379 note 1} It is not clear who issued this document. The grantor was presumably a Seljūq malik, since he claims that he had inherited the province of Ray from Muhammad b. Malikshāh. The document also states that Sanjar had confirmed his title to Ray and that Mas'ūd had re-granted him the province. Ray enjoyed a special status during the reign of Sanjar, who, having defeated Maḥmūd b. Muḥammad near Ray in A.H. 513, kept Ray in his own hands as a precaution lest Maḥmūd should rebel again ( ul-Athīr, Ibn, al-Kāmil, X, 389).Google Scholar

^{page 379 note 2} ‘AK, 44.

^{page 379 note 3} ‘AK, 43.

^{page 379 note 4} It is a notable fact that merchants are seldom mentioned in the documents in the ‘Atabat al-kataba.

^{page 379 note 5} ‘AK, 44.

^{page 379 note 6} ‘AK, 75.

^{page 379 note 7} See above, n. 1, for the special position of Ray under Sanjar.

^{page 379 note 8} ‘AK, 71.

^{page 380 note 1} Ḍarā'ib-i shahr. The precise meaning of this term is not clear from the text. It is conceivable that it is here used to mean mustaghillāt, i.e. rentable property in towns such as shops baths, and caravansarais. Rāwandī states that Sanjar after he had defeated Maḥmūd in A.H. 513 took for the royal dīwān (bā khāṡṡ girift) estates (ḍiyā’) and ḍarība in every town in ‘Irāq and the main regions. (Rāhat-us-ṡudūr wa āyat-us-surūr, 171.)

^{page 380 note 2} Or perhaps casual revenue, windfalls. For a discussion of the meaning of the term ṭayyārāt see Minovi, M. and Minorsky, V., ‘Naṡīr al-Dīn Ṭūsī on finance’, BSOAS, X, 3, 1941, 774 Google Scholar). See also Balkhī, Ibn, Fārs-nāma, ed. Strange, G. le and Nicholson, R.A. (GMS), London, 1921, 132.Google Scholar

^{page 380 note 3} ‘AK, 72.

^{page 380 note 4} ‘AK, 72–3.

^{page 380 note 5} ‘AK, 73.

^{page 380 note 6} ‘AK, 73–4.

^{page 380 note 7} ‘AK, 73.

^{page 380 note 8} ‘AK, 72.

^{page 380 note 9} The term dādbeg may possibly also be used in the sense of shiḥna towards the end of the Great Seljūq period. The document for the office of dādbeg mentioned above, p. 376, n. 4, however, states that this office ranked second to that of atābeg.

^{page 380 note 10} Al-Muntaຓam, Haidarabad, A.H. 1357–1359, X, 102.Google Scholar

^{page 380 note 11} Tārīkh-i Rቫyān (ed. ‘Khalīlī, Abbās, Tehran, 1934–1935), 89.Google Scholar

^{page 380 note 12} Zubdat an-nuṡrat wa nukhbat al-‘uṡrat, 191.

^{page 381 note 1} Al-Kāmil, XI, 116.

^{page 381 note 2} Rāḥat-us-ṡudūr wa āyat-us-surūr, 177.

^{page 381 note 3} ‘AK, 77.

^{page 381 note 4} Rāḥat-us-ṡudūr wa āyat-us-surūr, 177.

^{page 381 note 5} ‘AK, 75.

^{page 381 note 6} Thus Muhammad Ibrāhīm appears to use the word shiḥna both as the ‘representative’ of the sultan and as a governor ( Tārīkh-i Saljūqīān-i Kirmān, ed. Houtsma, M.T., Leyden, 1886, 10,Google Scholar 26, 33, 147, 153, and 169).

^{page 381 note 7} ‘AK, 77.

^{page 381 note 8} ‘AK, 79.

^{page 381 note 9} ‘AK, 79.

^{page 381 note 10} ‘AK, 78.

^{page 381 note 11} ‘AK, 85.

^{page 382 note 1} ‘AK, 61. According to the diploma the office of shiḥna of Juwayn had been handed over to the dīwān of Sanjar's sister, Tāj ud-Dawla wa'd-Dīn Nūr Balkā, and Sayf ud-Dīn Yaranqush had been appointed by her. This diploma confirms his appointment.

^{page 382 note 2} ‘AK, 61.

^{page 382 note 3} ibid.

^{page 382 note 4} See below, pp. 386–7.

^{page 382 note 5} ‘AK, 80.

^{page 382 note 6} ‘AK, 82.

^{page 382 note 7} cf. Siyāsat-nāma, 94.

^{page 382 note 8} ‘AK, 81.

^{page 382 note 9} ‘AK, 82.

^{page 382 note 10} ‘AK, 85.

^{page 383 note 1} ‘AK, 23.

^{page 383 note 2} ‘AK, 28.

^{page 383 note 3} This term presumably means here government officials and military officers passing through the area rather than ordinary travellers.

^{page 383 note 4} ‘AK, 41.

^{page 383 note 5} ‘AK, 56.

^{page 383 note 6} ‘AK, 55.

^{page 384 note 1} ‘AK, 23.

^{page 384 note 2} ‘AK, 41.

^{page 384 note 3} ‘AK, 25.

^{page 384 note 4} ‘AK, 26.

^{page 384 note 5} ‘AK, 23.

^{page 384 note 6} ibid.

^{page 384 note 7} ‘AK, 27.

^{page 385 note 1} ‘AK, 56.

^{page 385 note 2} ‘AK, 29.

^{page 385 note 3} ibid.

^{page 385 note 4} ‘AK, 23.

^{page 385 note 5} ‘AK, 27.

^{page 385 note 6} ‘AK, 41.

^{page 386 note 1} ‘AK, 79.

^{page 386 note 2} ‘AK, 56.

^{page 386 note 3} ‘AK, 24. The mandate also instructs him to exercise care concerning the property of orphans (p. 29).

^{page 386 note 4} ‘AK, 24.

^{page 386 note 5} ‘AK, 24, 28–9.

^{page 386 note 6} ‘AK, 24–5, 29.

^{page 386 note 7} ‘AK, 25, 28. The ra'īs appears in some cases to have had powers of apprehension. On one occasion while Ghazāī was mudarris of the Niຓāmiyya madrasa in Nīshāpūr, to which post he was appointed in A.H. 499, the ra'īs of Khurāsān imprisoned a man who had secretly altered the text of two of Ghazālī's works and then attempted to get Ghazālī to sign the manuscripts question, so that charges of unorthodoxy could be preferred against him ( Makātīb-i fārsī-i Ghazālī bi-nām-i Faḍā'il al-anām, ed. ‘, Abbās Iqbāl, Tehran, 1954–1955, 11–12 Google Scholar).

^{page 387 note 1} ‘AK, 25. See also p. 29.

^{page 387 note 2} ‘AK, 26, 29.

^{page 387 note 3} ‘AK, 41.

^{page 387 note 4} ‘AK, 25, 28.

^{page 387 note 5} ‘AK, 28.

^{page 387 note 6} ‘AK, 29–30.

^{page 387 note 7} ‘AK, 23–4, 29.

^{page 387 note 8} ‘AK, 30.

^{page 388 note 1} ‘AK, 30.