¹ British Library MS Sloane 3114, published in facsimile in Elçin, S. (ed.), Ufki, A., Mecmûa-i sâz ü süz (Kültür Bakanhiği türk musikisi eserler, 1, Istanbul: Milli Eğitim Basimevi, 1976)Google Scholar. The repertoire is that of the middle of the seventeenth century.

² Cantemir, Demetrius, The collection of notations, Part 1: text (SOAS Musicology Series, 1, London: School of Oriental and African Studies, 1992)Google Scholar. This collection contains the instrumental repertoire of the end of the seventeenth century. Both Cantemir and ‘Alī Ufḳī were Europeans, for whom notation was a familiar phenomenon (the latter, indeed, used Western notation in his collection): it is not until the nineteenth century, with the introduction of the Hamparsum system (to which one may add, in Central Asia, the invention of the Khwarezmian system) that indigenous forms of notation become widely used.

³ Risāla fī 'l-luḣūn wa-'l-nagham, ed. Yūsuf, Z. (Baghdad, 1965)Google Scholar, translation and transcription in Shiloah, A., ‘Un ancien traité sur le ‘ūd d'Abū Yūsuf al-Kindī’, Israel Oriental Studies, 4, 1974, 179–205Google Scholar.

⁴ British Library MS Or. 11190, fol. 201b (text published in Ibn Sīnā, Kitāb al-Shifā', al-riyāḍiyyāt 3:Google Scholarjawāmi‘ ‘ilm al-mūsīqī, ed. Yūsuf, Z., Cairo, 1956)Google Scholar; D'Erlanger, R., La musique arabe, 2 (Paris: Geuthner, 1935), 232Google Scholar.

⁵ Ṣafī al-Dīn al-Urmawī, Kitāb al-Adwār, ed. al-Rajab, H. M. (Baghdad, 1980), 159–63;Google Scholared. Khashaba, Gh. ‘A. and al-Ḣifnī, M. A. (Cairo, 1986), 311–32;Google Scholar transl in D'Erlanger, R., La musique arabe, 3 (Paris: Geuthner, 1938), 185–565, music examples at pp. 553–62Google Scholar. Transcriptions of music examples may be consulted in Kitāb al-Adwār, ed. al-Rajab, H. M., 164–8;Google ScholarBardakçi, M., Maragali Abdülkadir (Istanbul: Pan Yaymcilik, 1986), 131Google Scholar; D.'Erlanger, , La musique arabe, 3, 553–62Google Scholar, Farmer, H. G., ‘The music of Islam’, in Wellesz, E. (ed.), The new Oxford history of music, 1 (London: Oxford University Press, 1957), 454–5;Google ScholarWright, O., The modal system of Arab and Persian music, A.D. 1250–1300 (Oxford: Oxford University Press, 1978) ]henceforth referred to as Modal system[, 219–31Google Scholar.

⁶ See Modal system, 245–55Google Scholar.

⁷ British Library MS Add. 7694, fols. 241b–242a (text published in Quṭb al-Dīn al-Shīrāzī, Durrat al-tāj, ed. Mashkut, S. M. and Taqwā, N. A., Tehran, 1939–1946)Google Scholar. Notation transcribed in I. Rajabov, Maḳomlar masalasiga doir (Tashkent, 1963), 89–92 and Modal system, 233–44Google Scholar.

⁸ Nuruosmaniye MS 3647, fols. 58b–59a (text published in Ṣafī al-Dīn al-Urmawī, al-Risāla al-sharafiyya, ed. al-Rajab, H. M., Baghdad, 1982), transl. in R. D'Erlanger, op. cit., 3, 3–182, music examples at pp. 181–2)Google Scholar.

⁹ Even the grandest of all, the so-called Sharḣ mubārak shāh bar adwār, failes to extend the range of material (British Library MS Or. 2361, fols. 151b-153a. D'Erlanger, op. cit., 3, 558–63).

¹⁰ In addition to the treatise of al-Lādhiqī (al-Risāla al-fatḣiyya, British Library MS Or. 6629, transl. in D'Erlanger, R., op. cit., 4, Paris: Geuthner, 1939, 259–498)Google Scholarone may cite here those of al-Shirwānī, (Majallafi 'l-mūsīqī: first version Topkapi MS Ahmet HI 3449, facsimile in Publications of the Institute for the history of Arabic-Islamic science, series C, 29 (Frankfurt, 1986);Google Scholar second version British Library MS Or. 2361, fols. 168b–219b, transl. in R. D'Erlanger, op. cit., 4, 3–255) and Jāmī (Risāla-i mūslqī, facsimile and transl. by Boldirev, A. N., Tashkent, 1960)Google Scholar.

¹¹ A point insisted on, quite rightly, by Manik, L. (Das arabische Tonsystem im Mittelalter, Leiden:Brill, 1969, 54)Google Scholar.

¹² See Modal system, 31–2, 37–43.

¹³ Kitāb al-Adwār, ed. M.al-Rajab, H., 159–60Google Scholar(the transcription, p. 164 top, has been copied out back to front and should, therefore, be read from right to left); Modal system, 219–20, 245–6.

¹⁴ It is possible that the intonation of what is written as c in ḣijāzī tended, in practice, to be somewhat higher (c#), thus producing a tetrachord consisting of three quarter tone, tone and a quarter, and semitone (see the discussion in Modal system, 128–30)Google Scholar.

¹⁵ To the above the Sharḣ mubārak shāh bar adwār (D'Erlanger, op. cit, 3, 558) adds yet one further possible substitution, in rahāwī.

¹⁶ Jāmiެ al-alhān, ed. Bīnish, Taqī, Tehran: Mu'assasa-i muṭāla ‘āt wa-taḣqīqāt-i farhangī, 1366/1987, 238; Maqāsid al-alḣān, ed. Taqī, BīnishGoogle Scholar (majmūެa-i mutūn-i fārsī, 26, Tehran: Bungāhi tarjama wa-nashr-i kitāb, 2nd. ed. 2536/1977), 102.

¹⁷ This mode has a core pentachord notated as A B d e but, as with ḣijāzī, it is possible that in practice the intonation of C was high (c#), resulting in the interval below being approximately a tone and a quarterGoogle Scholar.

¹⁸ Three versions of mukhammas are described in the JA (pp. 225–6), one of 16 time units (kahīr) with an internal structure of 3 + 3 + 2 + 3 + 3 + 2; one of eight time units (awsaṭ), internally 3 + 3 + 2; and one of four (ṣaghīr). The MA (p. 95) gives only the second of theseGoogle Scholar.

¹⁹ This example is also transcribed by Bardakçi (op. cit., 133) who goes on to explore the ways in which a rhythmic articulation might be imposed by subdividing note values, only to conclude that there can be no single correct method.

²⁰ ެAbd al-Qādir al-Marāghī lists (JA, 219) a ramal kabīr of 24 time units, but ramal proper is of 12, with the inner structure 2 + 2 + 2 + 2 + 4, exactly the definition that had been given by Ṣafī al-Dīn al-UrmawīGoogle Scholar (Kitāb al-Adwār, ed. al-Rajab, H. M., 150)Google Scholar. The MA (p. 95), however, has the variant form 2 + 2 + 4 + 4.

²¹ In the JA (pp. 238–9) ެAbd al-Qādir al-Marāghī also_suggests an innovation in the technique of notation, namely to extend the writing of the text for a given pitch commensurate with the number of time units covered and to insert along the extended line dots to the number of time units. The idea may have been derived from Quṭb al-Dīn al-Shīrāzī's scheme of notation, in which duration is displayed visually; but it is essentially redundant, since duration is already indicated numerically (when, that is, the relevant numerals are actually entered), and ެAbd al-Qādir al-Marāghī does not, in fact, employ this technique.

²² Kiesewetter, R. G., Die Musik der Araber nach Originalquellen dargestellt (Leipzig: Breitkopf und Hôrtel, 1842, repr. Wiesbaden: Dr. Martin Sôndig, 1968)Google Scholar. For a general appreciation of this pioneering work see Bohlman, P., ‘R. G. Kiesewetter's Die Musik der Araber’, Asian Music, 18/1, 1986, 164–96Google Scholar.

²³ op. cit., 56: ‘obgleich ich gestehen muss, dass ich in diesem Exempel kaum noch die Prosodie, vielweniger etwas von den, in dem Texte des Autors beschriebenen Rhythmen zu erkennen vermag.’

²⁴ Fétis, F.-J., Histoire générate de la musique, 2 (Paris: Firmin Didiot Fréres, Fils, 1869), 68–70Google Scholar.

²⁵ op. cit., 1, 141. They are blamed upon too great a reliance on prosody as a determining factor.

²⁶ ibid. The argument about pitch is made, curiously, in relation to the numerals of the original notation, which define, rather, duration. With regard to measures 9–11, Fétis states, with a conviction based on instinct rather than evidence, that ‘il est évident que le chant est ici suspendu, et que les paroles sont presque parlées’.

²⁷ ‘Tonschriftversuche und Melodieproben aus dem muhammedanischen Mittelalter’, Vierteljahrsschrift fūr Musikwissenschaft, 2, 1886, 354Google Scholar.

²⁸ He suggests that either a time unit has been added to the final note of the second hemistich or a note with the duration of one time unit has been omitted from the area of the syllable /bul/.

²⁹ The two a re referred to (ibid.) as ‘zwei Melodien’, and the second is introduced (p. 355) as ‘das andere Beispiel’.

³⁰ ibid. ‘Die Silben dir dirun (so im Original vocalisirt) sind mir unverstăndlich: tanani u.s.w. sind Formeln des īqā' welche ich aber in dieser Fassung sonst nirgends nachweisen kann.‘ It may be observed that theorists of the period do, indeed, use the syllables ta and na in defining rhythmic cycles, but not ni.

³¹ Bey, Rauf Yekta, ‘La musique turque’, Encyclopédie de la musique (Lavignac), 1 (Paris: Librairie Delagrave, 1922), 2945–3064, at p. 2977Google Scholar: ‘mais sa traduction a pris une forme si désordonnée que Fétis, de son cúté, a éprouvé la nécessité de la corriger et d'en donner une traduction soi-disant correcte. Malheureusement, Fétis n'a pas réussi davantage; il a déformé complétement le rythme …’

³² Leiden MS no. 270–1, which dates from this year. ެAbd al-Qādir al-Marāghī is referred to (Fétis, op. cit., 1, 140) as an ‘auteur persan d'un traité de la musique arabe, écrit à Constantinople vers l'an 824'/1421. The claim that it was written in Constantinople is explained by its having been dedicated to Sultan Murād II. Other fifteenth-century copies are: Bodleian MS Ouseley 264, a holograph dated 821ü1418; Mashhad Razavi Library MS 539, also dated 821/1418 (the holograph on which the printed edition (see n. 16) is based); a further holograph, dated 826/1423, once in the possession of Rauf Yekta Bey, auctioned at Sotheby's in 1985 and now in private hands (Bardakçi, op. cit., 142); Tehran Malik Library MS 832/1, dated 837/1433; and Topkapi Saray MS R. 1727, dated 838/1434.

³³ But largely completed by 808/1405.

³⁴ Bodleian MS Marsh 282, fol. 93b; ed. , T. Bīnish, 237Google Scholar.

³⁵ ed. T. Bīnish, 101.

³⁶ ެAbd al-Qādir al-Marāghī places it in the preceding chapter (ed. Bīnish, T., 231–7). This is divided into two parts (fast), and the second, which moves on to deal with aspects of practice, could more logically have been held over to the nextGoogle Scholar.

³⁷ It is closely modelled on ch. 14 of the Kitāb al-Adwār (ed. al-Rajab, H. M., 157)Google Scholar, using the same terminology and even quoting some phrases verbatim. There is a natural extension from the shūdud modes (the only ones mentioned by Ṣafī al-Dīn al-Urmawī) to include the other important groups (āwāzes and shuެbas), although there is n o mention of individual modes belonging to these categories. The passage ends not, as in the Kitāb al-Adwār, with a verse unsuitable for a given mode but, rather, with three lines exemplifying what is deemed appropriate.

³⁸ ed. H. M. al-Rajab, 159.

³⁹ ed. T. Bīnish, 237–8: tā kasī ki mubāshir-i ެamal bāshad bidān sūrat … ެamal kunad … wa-chūn māhir shawad bar kayfiyyat-i ṣūrat-i ān aṣwāt wuqūf yābad tā az khūd tarkīb tawānad kard wa-har ānchi az taṣānif khwāhad bidān nawެ wadެ kardan chunānki dīgarī … tawānad istikhrāj kardan wa-dar ެamal āwardan. There is, however, nothing to suggest that the use of notation for transmission was more than a theoretical possibility.

⁴⁰ ’Part 2: Commentary’, to appear in BSOAS, LVIII, 1, 1995Google Scholar.

⁴¹ Although they are, ultimately, from different sources. Apart from being grammatically eccentric as vowelled by ‘Abd al-Qādir al-Marāghī, the first line is odd when compared with the other two because it is in a different metre (khajīf as against the following munsarih). It is interesting to note that when the anecdote does resurface in the MA (having been shifted to a section (ed. T. Bīnish, 138) beginning with the mythological inventors of instruments and other sages of antiquity with musical skills and proceeding to caliphs and other significant musical figures in Islam) only the second and third lines are given. However, evidence that the conjunction of all three was by no means new, going back at least to the early twelfth century, is provided by their inclusion in the Bawāriq al-ilmā‘ of Majd al-Dīn al-Ghazālī (d. 1126) (ed. J.Robson, Tracts on listening to music (Oriental Translation Fund 34, London: Royal Asiatic Society, 1938), 172).

⁴² The term ṭarīqa-i matla' does not appear at all in the MA.

⁴³ The description of the qawl form (JA, ed. Bīnish, T., 241) states categorically that it has both (qawl-rā dū tarīqa bāyad ki bāshad awwal tarīqa-i jadwal thdrīT tarīqa-i matla')Google Scholar.

⁴⁴ See Wright, O., Words without songs: a musicological study of an Ottoman anthology and its precursors (SOAS Musicology Series, 3, London: School of Oriental and African Studies, 1992) [henceforth Words without songs], 53Google Scholar.

⁴⁵ Marsh, Bodleian MS 282, fol. 95b, ed. Bīnish, T., 242: agar tarīqa-i jadwal-rā bar yak misrā' sākhta bāshand miyānkhāna bar yāk bayt bāyad zīrāki yak miṣrā' barāyi taghayyur-i ṣawt wa-yak miṣrā' barāyi i'āda-i tarīqaGoogle Scholar.

⁴⁶ This being precisely the position 'Abd al-Qādir al-Marāghī refers to, although his wording (ān nuqūsh ki miyān-i ṣawt u i'āda-i ṭarīqa ast ānrā nuqūsh-i mulsaqa kh^wānand) suggests that it might not be the only possible one.

⁴⁷ ibid.: agar matla'-rā bar bayīi sākhta bāshand ṣawt bar dū bayt bāyad bārdyi ānki āhang-I mutaghayyira-i miyānkhāna bar yak bayt bāyad wa-baytī-yi dīgar i'āda-i ṭarīqa.

⁴⁸ The meaning suggested for tashyī'a seems the least unlikely among those listed by the dictionaries. Perhaps wisely, no attempt is made to gloss it in Faruqi, L. I. al, An annotated glossary of Arabic musical terms (Westport, Connecticut: Greenwood Press), 1981Google Scholar. The meaning of bāzgasht, is, in contrast, not at all problematic. But it is, at first sight, a curious term nevertheless: incidentally odd is that it is not synonymous with its Arabic equivalent tashyī'a, but centrally peculiar is that it does not seem to indicate what it ought, since it stands at the head of a section the primary function of which is not to return to something that has gone before but to introduce something new. A possible explanation is suggested in part 2 (‘Conclusion: intiqāl’). Al Faruqi calls it a refrain section, and gives as source D'Erlanger's translation of al-Shirwānī (op. cit., 4, 235–6)Google Scholar, where tashyī'a is glossed as ‘ritournelle’. But there is nothing in the original (British Library MS Or. 2361, fol. 215v) to correspond to this gloss.

⁴⁹ I am throughout, but particularly in relation to the verse, grateful to Mr. A. Morton for kindly checking (and often correcting) the readings of the Persian texts.

⁵⁰ Marsh, Bodleian MS 282, fols. 94b–95a, ed. T. Bīnish, 240Google Scholar. Presumably to be identified with Mīr Sayyid Sharīf Jurjāni (d. 1413), although he is known primarily as a writer of prose works in Arabic.

⁵¹ Written in above, as a later addition, is tashyī'a. As before, it is a synonym of bāzgasht adding nothing to the meaning.

⁵³ Although they are defined by Quṭb al-Dīn al-Shīrāzī, whose achievement in this respect is both unprecedented and extraordinary.

⁵⁴ al-Urmawī, Safī al-Dīn, Kitāb al-Adwār, ed. al-Rajab, H. M., 98, 121Google Scholar; for al-Shīrāzi, Quṭb al-Dīn see Modal system, 68Google Scholar (where the A–e segment is explicitly indicated as prominent); ‘Abd al-Qādir al-Marāghī, JA, ed. Bīnish, T., 114, 120Google Scholar.

⁵⁵ On the basis of the Bodleian MS alone it is not absolutely certain that the second and third notes in the second line should be c rather than e , since they have an intermediate shape with a short final flourish , but reference to other versions (e.g. Nuruosmaniye MS 3644) confirms that the correct reading is c . In the nonsense-syllable section the vertical alignment of pitches and syllables is, from the fourth note on, precise, but for the first three is asymmetrical, a single pitch being written beneath the first two syllables and two pitch indications beneath the third. The same lay-out appears in the Nuruosmaniye holograph.

⁵⁶ In Nuruosmaniye MS 3644 the second note is aligned with the whole word ruqyatī rather than just the final syllable, but the third is again squeezed up to correspond only to the penultimate syllable yā.

⁵⁷ The first F is written with the letter shape of A , but with a dot, suggesting that is intended, and this is confirmed by Nuruosmaniye MS 3644.

⁵⁸ Apart from the varying interpretations of the size of the F–G versus A–B interval the only difference between this and ex. 10 is in the duration of the antepenultimate note.

⁵⁹ Again, interpretations of intonation apart, there is one difference of detail between this and ex. 11, concerning the identity of the penultimate note. B (equivalent to G in ex. 11) seems more likely, but it is possible that the Leiden MS has another reading. A more significant difference concerns duration: Land here opts for 1 = , having transcribed ex. 10 on the basis of 1 = )

⁶⁰ Unless the new material it contains is prefigured in a revised version of the MA, for Rauf Yekta Bey had had a copy (see n. 32) of this in his possession since 1893 (Bardakçi, op. cit., 142).

⁶¹ A confusion of the earlier and later JA versions occurs in Words without songs (p. 219), where the description of the melodic outline relates to one and the definition of the mode to the other.

⁶² The first bracket relates to a definition of duration (4) for which there is no corresponding pitch indication but, rather, a dash suggesting a link between the preceding and following As and, therefore, the likelihood that the same pitch is intended (together with the possibility that the duration might be a global one including that of the following note). A similar notation is used in the next line for fa–'indahū; the pitch is defined twice (for fa and 'indahū) with a n intervening dash under which is written the global value 6. The second bracket reflects what is probably another global value, although there is no accompanying dash: 8 is written under the second B and no duration is separately indicated for the two notes flanking it.

⁶³ As in the preceding example the bracket encloses three notes in relation to which there is only one indication of duration. In this case, however, there is a dash underlining the first two of the three, so that it is possible that the 8 time units specified apply only to these, leaving the final note without any indication of duration.

⁶⁴ ed. Bīnish, T., 102: ammā īnjā alfāẓ-i naqarāt-rā isti'māl kardīmGoogle Scholar.

⁶⁵ As is also the case in ‘Abd al-Qādir al-Marāghī's second ṭarīqa notation based on them (ex. 7), while the first (ex. 6) spans exactly two rhythmic cycles.

⁶⁶ In the light of this difficulty the (not very convincing) suggestion is made in Modal system (p. 222) that an underlying half cycle unit might be postulated, thus reducing the number required to two time units per hemistich in the first example.

⁶⁷ That is, if we adopt the 3 + 3 definition of durational values in H2 in the later holograph, Nuruosmaniye MS 3644.

⁶⁸ Discounting the one note without an assigned duration (in HI). To be assumed is a fractional short value to be subtracted from that of the preceding or following note.

⁶⁹ See n. 62 to ex. 22.

⁷⁰ See n. 63 to ex. 23.

⁷¹ By the Markaz-i nashr-i dānishgāhī, Tehran.

The writer identifies himself (p. 2) as ‘Alī b. Muḥammad al-Mi‘mār al-mushtahir bi’1-Binā’i (or: al-Bannā'ī). He may be identified with Kamāl al-Dīn Shīr ‘Alī b. Muḥammad Binā’ī Harawī, a poet, historian and musician sometime active at the Timurid court in Herat but for long periods forced to seek patronage elsewhere. The present work was written while he was attached to the Aq Qoyunlu Sultān Ya'qūb. Becka, J. (in J. Rypka, History of Iranian literature, Dordrecht: D. Reidel, 1968, 497–500)Google Scholar mentions his treatises on music as not having survived. On Binā'ī and Nawā'ī see Subtelny, M. E., ‘Scenes from the literary life of Tīmūrid Herāt’ in Savory, R. M. and D., Agius (ed.), Logos islamikos: studia islamica in honorem Georgii Michaelis Wickens (Papers in medieval studies, 6, Toronto: Pontifical Institute of Medieval Studies, 1984), 137–55Google Scholar (at pp. 147–8).

⁷² The manuscript is in the private possession of Dr. Yūsuf Nayyirī, and there is no mention of it in Dānishpazhūh, M. T., ‘Ṣad u sī and athar-i fārsi dar mūsīqī’, Hunar wa-Mardum, 94 (Aug. 1970), 23–44; 100 (Feb. 1971), 24–27Google Scholar.

⁷³ pp. 53, 67, 70. The argument is put forward in Modal system (pp. 27–9) that, in general, the theoretical division of the upper wholetone of an octave scale is misleading, and that the pitch which in other contexts might be transcribed as should here have the value . As far as ‘irāq is concerned this value is confirmed by Quṭb al-Dīn al-Shīrāzī (Modal system, 66), who gives ratios dividing the upper whole tone into two intervals of 119 and 85 cents respectively, and by Binā’ī himself (p. 67), for whom they are slightly closer together in size, the ratios cited being 16:15 (112 cents) and 20:19 (88 cents). These add up to marginally less than a wholetone (Binā'ī's ratios elsewhere contain some obvious blunders, so this need occasion no great surprise), but the message is clear: the upper whole tone in 'irāq was divided into two slightly unequal semitones, the larger being below the smaller.

⁷⁴ Different, therefore, from that recorded by ‘Abd al-Qādir al-Marāghī (see n. 18), which is repeated by Binā’īs contemporary al-Lādhiqī (D'Erlanger, op. cit., 4, 474). But this version is not peculiar to Binā'ī, for it is also given by another contemporary, ‘AlIshāh b. Būka Awbahī (Muqaddima-i usūl, MS İstanbul Üniversitesi Kütüphanesi F 1079, 73).

⁷⁵ This is stated (p. 125) to be a convention of naqsh pieces in the mukhammas rhythmic cycle setting a text in the ramal metre.

⁷⁶ The note in cycle 7 marked by an asterisk is given, in the original, the duration 2 rather than the 1 assigned to it here. It is difficult to see this as other than an obvious blunder. To say that it would destroy what appears to be a clearly intended symmetry in rhythmic articulation might be regarded as insufficient grounds for emendation, but there is also a more objective reason for considering it potentially wrong, namely, that the line in which it occurs infringes thespatial logic established by the first three lines, each of which covers exactly two cycles, with the third, indeed, being extended, in order to do so, beyond the frame into the margin. But the arithmetic argument does not of itself identify this particular note as the one that breaks the rules, for there are two others in the line of two time units duration. It will be seen, however, that the first of these is slightly unusual in the angle of its stem, and it may be suggested as the most likely explanation that it was first written as 1 and then later corrected to 2 but that, by oversight, the compensatory correction to the note emended was omitted.

In cycle 9 the note marked by an asterisk has been transcribed as d rather than c# because of its proximity to the intrusive e . The latter indicates a modulation in the context of which d is the more apprpriate equivalent.

⁷⁷ But in neither the JA nor the MA, where the lists and descriptions of forms contain no mention of it. His account comes, rather, in the Fawā'id-i'ashara, Nuruosmaniye MS 3651, fol.95a. Words without songs (p. 215) erroneously cites as the first theorist to describe this form al-Shirwānī, whose definition simply reproduces the slightly earlier one of ‘Abd al-Qādir al-Marāghī.

⁷⁸ p. 126 (Words without songs, 216).

⁷⁹ Nuruosmaniye MS 3652. See Words without songs, 107–8Google Scholar.

⁸⁰ JA, ed. Bīnish, T., 238Google Scholar.

⁸¹ In his mukhtaṣar dar ‘ilm al-mūsīqī, Nuruosmaniye MS 3649, fols. 44a–47a.

⁸² In Bodleian MS Ouseley 128, fols. 43b–48b, 65b–72a.

⁸³ Words without songs, 128.