¹ The most comprehensive list of MSS is that contained in Shiloah, A., The Theory of music in Arabic writings (c. 900–1900) (Répertoire international des sources musicales: B X), Munich: Henle, 1979, 309–12Google Scholar. The number of copies detailed is 21 (2 from Vienna; 1 Dublin; 1 Paris; 2 London; 4 Oxford; 10 Istanbul; 1 Yale). Others are to be found in Baghdad, Cairo, Leningrad, Madras, Mashhad, Rampur and Tehran (see for some of these Yūsuf, Z., Makhṭūṭāt al-mūsīqa al-'arabiyya fī 'l-'ālam) 3 vols. Baghdad: Matba'at shafī', 1966–1967Google Scholar; see also n. 2). Commentaries include Sharh Mawlānā Mubārak Shāh, British Library MS Or. 2361 ff. 68b–153 (translation—in which is embedded KA itself—in D'Erlanger, , La musique arabe, 3, Paris: Geuthner, 1938, 185–565)Google Scholar; Shark al-adwār, British Library Add. 7371, fols. 41b–91b and Or. 2361, fols. 33b–68a; and that of 'Abd al-Qādir al-Marāghī, Nuruosmaniye 3651. Given the number of marginal glosses they contain, a number of MSS of KA can be regarded as text plus commentary. KA was also translated into Persian and Turkish, although for later Persian writers Safī al-Dīn al-Urmawī's ideas were generally mediated through the treatises of Quṭb al-Dīn al-Shīrāzī and, especially, 'Abd al-Qādir al-Marāghī. For basic biographical information see e.g. EI², s.v. Safī al-Dīn, al-Bakrī, 'A., Sofī al-Dīn al-Urmawī, mujaddid al-mūsīqā al-'abbāsiyya (Silsilat al-a'lām wa-'l-mashhūrīn, 4, Baghdad: Dār al-hurriyya li-1-ṭibā'a, 1978)Google Scholar and the introductions to the two editions of KA cited in n. 2. A number of specific points are discussed below.

² There have appeared so far two MS in facsimile: kitāb al-Adwār fī ma'rifat al-nagh(a)m/nigham (see on these forms n. 7 below) wa-'l-adwār, Husayn, akhrajah d.Mahfūz, 'Alī, Baghdad, 1961Google Scholar (scribe: Abū Ishāq al-Kirmānī; copied 870/1465; presented to Sulṭān Ibrāhīm b. Jahāngīr in 882/1477; kitāb al-Adwā, with introduction by Neubauer, E., Publications of the Institute for the History of Arabic-Islamic Science, series C, 6, Frankfurt, 1984Google Scholar (Nuruosmaniye 3653, dated 633/1235–6)) and two editions: kitāb al-Adwār, sharḥ wa-taḥqīq Hāshim Muḥammad al-Rajab (al-turāth, Silsilat kutub, 192, Baghdad: Manshūrāt wazārat al-thaqāfa wa-'l-i'lām, 1980)Google Scholar (al-Rajab lists 14 copies but gives variants from only three: British Library Or. 2361 (dated 1073/1663); the copy published by Husayn 'Alī Maḥfūẓ; and Nuruosmaniye 3653, on which the text is principally based (wa-qad ja'alt minhā aṣlan fī 'l-taḥqīq wa-'l-sharh, 18)Google Scholar, for all that chronology alone is an insufficient justification for considering this the most authoritative text); and kitāb al-Adwār fī 'l-mūsīqā, tahqīq wa-sharḥ Ghaṭṭās 'Abd al-Malik Khashaba, murāja'a wa-taṣdīr d. Maḥmūd Aḥmad al-Hifnī (Cairo: al-Hay'a al-miṣriyya al-'āmma li-1-kitāb, 1986) (utilizes MSS Dār al-kutub 428 funūn jamīla (scribe: 'Abd al-Karīm al-Suhrawardī), copied 727/1327; Dār al-kutub 507 funūn jamīla, a photocopy of Topkapi A. 2130, dated 726/1326; and Bodleian Marsh 521 (scribe: Yūsuf b. Nu'mān al-Kātib al-Mārīdīnī), dated 734/1334). Page references for KA will be to the edition by al-Rajab, identified as R.

³ Geschichte der arabischen Litteratur, Supplementband 1, 1937, 907.Google Scholar

⁴ Farmer, H. G., The sources of Arabian music (Bearsden, 1940)Google Scholar. (Mention should also be made of his earlier ‘Arabic musical manuscripts in the Bodleian Library’, Journal of the Royal Asiatic Society, 1925, 639–54Google Scholar, some of the MSS in question being copies of KA.)

⁵ 2nd ed. Leiden: Brill, 1965, 49. But no details are given.

⁶ loc. cit. The entry reads: ‘Fatih 3662, 27 ff. 239 x 164 (170 x 105) mm. 131. Ch. I (3a); II (3b); III (5a); IV (la); V (8b); VI (9b); VII (13a); VIII (14a); IX (18b); X (17a); XI (18b); XII (20b); XIII (22a); XIV (23b); ch. XV missing.’

IX should read (16a).

⁷ The standard dictionaries give, corresponding to the singular naghma, the plurals naghm and nagham (or regard these as verbal nouns of which naghma is a singulative), but both F and Nuruosmaniye 3652, which are (almost) fully vowelled, generally agree on nigham (which is also noted as a plural in L. I. al Faruqi, An annotated glossary of Arabic musical terms, Westport: Greenwood Press, 1981), although in the heading for chapter 1 F has nagham as against the nigham of Nuruosmaniye 3652. As plurals two of the three are rare forms, while the third would not normally relate to a singular naghma. The most common plural form in F and KA is in fact the more predictable naghamāt. There is one instance in F (chapter 11, fol. 18b) of nagham, but in the phrase bi-ziyādat al-nagham, where it seems to be, rather, a verbal noun form meaning ‘resonance’. In Nuruosmaniye 3652 (p. 2) nigham has masculine concord, suggesting that it might be viewed as a collective; but concord in this MS is somewhat erratic.

⁸ In general, ‘scale’ will be used in relation to the representation of the modes as abstract pitch sequences: but no attempt is made to discriminate rigorously between ‘mode’ and ‘scale’.

⁹ The Sharḥ Mawlānā Mubārak Shāh reads, rather, iṣṭikhab (‘clamour’) which may be considered, semantically, a lectio difficilior, and adds in relation to the first appearance of this word (in the introduction preceding the list of chapter headings) a specific comment, buttressed by a verse quotation, relating to its meaning (on the authority of which this reading is adopted in Wright, O., The modal system of Arab and Persian music, a.d. 1250–1300, London Oriental Series, 28, Oxford: Oxford University Press, 1978, 23)Google Scholar. But both F and Nuruosmaniye 3652 always have iṣṭiḥāb (‘association’), which must be the preferred reading. (The author of the Sharḥ Mawlḥnḥ Mubḥrak Shḥh refers elsewhere to the poor quality of the MS of KA on which he had to rely, and it may be conjectured that it was at least partially unpointed.)

¹⁰ The heading both in the initial chapter list and of the chapter proper has the rather unhelpful formulation fī 'stikhrḥj al-shudūd min al-shudūd, but in both cases uṣūl and adwār are added in the margin.

¹¹ For confirmation of this last point I am indebted to Dr A. Contadini.

¹² Facsimile of MS Topkapi Ahmed III 3640 (published in the same volume as KA), Publications of the Institute for the history of Arabic-Islamic science, series C, 6, Frankfurt, 1984, 5.

¹³ kitāb al-Mūīsūīqūī al-kabūīr, ed. Khashaba, Gh. 'A. (Cairo, 1967), 214Google Scholar (D'Erlanger, , op. cit., 1, 1930 81).Google Scholar

¹⁴ Identified as al-shaykh al-ra'īs D'Erlanger, , op. cit., 2, 1935, 114.Google Scholar

¹⁵ Majalla fī 'l-mūsīqī, (first version) MS Topkapi Ahmet III 3449, facsimile in Publications of the Institute for the history of Arabic-Islamic science, series C, 29, Frankfurt, 1986, 51Google Scholar. His mentor, 'Abd as-Qādir al-Marāghī, incorporates the whole passage from al-Risāla al-sharafiyya (Jāmi' al-alḥān, ed. Bīnish, Taqī, Tehran: Mu'assasa-i muṭāla'āt wa-tahqīqāt-i farhangī, 1366/1987, 15–16)Google Scholar. Later in the century, however, we find al-Lādhiqī repeating the definition by al-Fārābī, and claiming that this was the one that satisfied Safī al-Dīn al-Urmawī (D'Erlanger, , op. cit., 4, 269)Google Scholar. But even if misguided, this again shows that the source text is al-Risāla al-sharafiyya.

¹⁶ The theoretical sizes for these are an apotome (180 cents) and a limma (90 cents) respectively. See e.g. Manik, L., Das arabische Tonsystem im Mittelalter (Leiden: Brill, 1969), 67–9Google Scholar (for the argument that the former value distorts the norms of practice see Wright, , op. cit., 31–2, 37–43).Google Scholar

¹⁷ Khashaba, ed. (see n. 2), 87Google Scholar. The difference between F and KA had previously been noted by Sayyid, F. (Fihris al-makhṭūṭāt al-muṣawwara, 4: al-ma'ārif al-'amma wa-l-funūn al-mutanawwi'a (Jāmi'at al-duwal al-'arabiyya: Ma'had al-makhṭūṭāt al-'arabiyya), Cairo: Dār al-ma'ārif, 1384/1964, 45)Google Scholar, but without futher comment save to suggest for F a possible eighth(/fourteenth) century date.

¹⁸ Sources, 1st ed., 48.Google Scholar

¹⁹ Revue de I'Académie arabe de Damas, 3, 365Google Scholar. Mention is made of Ṣafī al-Dīn al-Urmawī's works (albeit on p. 366), but no dates are given.

²⁰ Geschichte der arabischen Litteratur, 1, 1898, 496.Google Scholar

²¹ EI¹ supp. s.v. Safī al-Dīn.

²² Farmer, H. G., A history of Arabian music to the xiiith century (London: Luzac, 1929, repr. 1973), 229Google Scholar. Whereas in the El¹ article Brockelmann is referred to as claiming that the text written ‘about the year 1252’ is al-Risāla al-sharafiyya.

²³ Geschichte der arabischen Litteratur, Supplementband 1, 1937, 907Google Scholar (citing Farmer, , History, 227 ff.)Google Scholar. Brockelmann also refers to Revue de I'Académie arabe de Damas, 3, 365Google Scholar, and to Hājjī Khalīfa (ed. Flügel, , 3, 423Google Scholar, for which read, with Farmer, 413). But neither the latter, nor the Habīb al-siyar of Kh^wāndmīr (lithograph, Bombay: Ahmadee Press, 1857, 3/1, 61) on which he draws, provide a date.

²⁴ As witness e.g. the incipient embarrassment of Manik, , op. cit., 53Google Scholar and the rather more obvious discomfort of Wright, , op. cit., 1Google Scholar. A more recent citation of 1252 as the date of composition may be found in Dzhumajev, A., ‘From parda to maqām: a problem of the origin of the regional systems’, in Elsner, J. and Jähnichen, G. (ed.), Regionale maqām-Traditionen in Geschichte und Gegenwart (Materialien der 2. Arbeitstagung der Study Group “maqam” des International Council for Traditional Music vom 23. bis 28. März 1992 in Gosen bei Berlin), 2 (Berlin, 1992 [pub. 1994]), 151Google Scholar. But the most bizarre adventure to befall 1252 occurs in Shiloah's Introduction (op. cit., 7), where it achieves its apotheosis as the date of his death (although this is evidently a momentary aberration, the correct date appearing in the main body of the text). Mercifully unaffected by these problems, al-Rajab settles down, after a vague fī ākhir al-'ahd al-'abbāsī (op. cit., 12), to 1235(-6) (pp. 17, 39) as the date of KA. This earlier date is also given by Neubauer, E. (EI²Google Scholar, s.v. Safī al-Dīn al-Urmawī), which makes all the more curious his following comment that it ‘was written while he still worked in the library of al-Musta'ṣim’—which sounds almost as if it were a last faint echo of 1252.

²⁵ Ḥusayn 'Alī Maḥfūẓ heads his biographical note with Safī al-Dīn al-Urmawī's name and date of death only (op. cit., j), but says (on the authority of a library catalogue) that he was born c. 613(/1216). This date is repeated in the introduction by al-Ḥifnī to the Khashaba edition (p. 3), without any reference, and likewise by al-Bakrī, (p. 31)Google Scholar. Farmer, (History, 227)Google Scholar says that he was probably born in Baghdad ‘in the early years of the thirteenth century’. There are no references. Al-Rajab, (op. cit., 7)Google Scholar also gives Baghdad as his place of birth, and the date as 613/1216, and cites as authority the introduction of the Husayn 'Alī Maḥfūẓ edition … More recently, the same date is given by Neubauer, E. (EI², s.v. Safī al-Dīn al-Urmawī)Google Scholar, who does, however, provide a textual reference (al-Fuwaṭī, Ibn, al-Ḥawādith al-jāmi'a (Baghdad: al-Maṭba'a al-'arabiyya, 1351/1932), 480)Google Scholar. According to this authority Safī al-Dīn al-Urmawī was approximately 80 when he died in 693/1294. If it is accepted that ‘approximately’ (nahw) allows him to be a few years older the conjectured range 1210–1215 appears not unreasonable.

²⁶ According to Ṣafī al-Dīn al-Urmawī himself, as reported by al-'Izz Ḥasan al-Irbilī, who met him is Tabriz in 689/1290 (the source text is Muḥammad b. Shākir al-Kutubī (d. 764/1362), Fawāt_ al-wafayāt). This and other biographical sources are quoted in al-'Azzāwī, 'A., al-Mūsīqā al-'irāqiyya fī'ahd al-mughūl wa-'l-turkwnān (Baghdad: Sharikat al-tijāra wa-'l-ṭibā'a al-maḥdūda, 1951), 23–31Google Scholar. The general trajectory of his career prior to the Luḥāẓ episode as portrayed there is that he did not set out as a musician, but only took up the lute seriously after a conventional scholarly training and having already established a reputation as a calligrapher, on the strength of which he was appointed royal copyist in al-Musta'ṣim's library.

²⁷ See EI² s.v. Mustanṣir (bi'llāh).

²⁸ This is the account reported by al-'Izz Ḥasan al-Irbilī (al-'Azzāwī, , op. cit., 24)Google Scholar. According to the Fawāt al-wafayāt he was a teacher at the Mustanāiriyya (the edition of 'Abd al-Ḥamīd, M. M. (2, Cairo: Maṭba'at al-sa'ada, 1951, 39)Google Scholar has ataytu faqīhan, that of Iḥsān 'Abbās (2, Beyrut: Dār ṣādir, 1974, 412)Google Scholaruthbittu faqīhan), although it should be noted that al-Bakri, (op. cit., 33)Google Scholar understands faqīh to mean, rather, ‘student’, an interpretation that may be related to his assumption (p. 37), which the sources do not compel us to accept, that Ṣafī al-Dīn al-Urmawī enrolled at the Mustanṣiriyya immediately after his arrival in Baghdad.

²⁹ Masālik al-absār fī mamālik al-amṣār, book 10 (Publications of the Institute for the History of Arabic-Islamic Science, series C, 46, 10, Frankfurt, 1988), 309Google Scholar. The authority is al-Shaykh Abū 'l-Khayr Sa'īd al-Dihlī (Khashaba (KA edition) reads al-Dhuhlī). Although it is included by al-'Azzāwī neither al-Rajab nor al-Bakrī refers to this account.

³⁰ He is said to have stayed initially at a ribāt and to have effected contact with the caliph through the presentation to him of an elegant calligraphic (bi-khaṭṭ mansūb) Qur'ān.

³⁰ He is said to have stayed initially at a ribāt and to have effected contact with the caliph through the presentation to him of an elegant calligraphic (bi-khaṭṭ mansūb) Qur'ān.

³¹ The claim is made by Farmer (History, 229)Google Scholar and repeated by al-Bakrī, (op. cit., 87)Google Scholar but disproved by Neubauer (EI² (s.v. Safī al-Dīn al-Urmawī)).

³² al-Bakrī, , op. cit., 36Google Scholar (although the one line given as being by him is decidedly pedestrian).

³³ The possibility that N might be a holograph is suggested, albeit tentatively, by Neubauer in his introduction to the facsimile (see n. 2), and repeated in EI² (s.v. Safī al-Dīn al-Urmawī).

³⁴ The Masālik al-abṣār fl mamālik al-amṣār names both Yāqūt al-Musta'ṣimī and Ibn al-Suhrawardū as his pupils (p. 310).

³⁵ In any case, establishing which of them might be genuine is highly problematic. See James, D., The master scribes: Qur'ans of the 10th to 14th centuries AD (The Nasser D. Khalili Collection of Islamic Art, 2, London and Oxford: Nour Foundation, 1992), especially pp. 58–9Google Scholar, ‘The problem of Yaqut al-Musta'simi’, and cf. the examples of naskh on p. 69.Google Scholar

³⁶ p. 39.

³⁷ p. 11.

³⁸ Risaāla fī 'ilm al-mūsīqī, ed. Yūsuf, Z. (Cairo, 1964).Google Scholar

³⁹ Bodleian MS Marsh 521, dated 734/1334 (Khashaba, pp. 86–7)Google Scholar and collated with one written by a pupil of Ṣafī al-Dīn al-Urmawī, Shams al-Dīn al-Suhrawardī.

⁴⁰

⁴¹ Avoiding e.g. the rather clumsy nisbat muṭlaq naghmat watar.

⁴² But, as Manik, (op. cit., p. 54)Google Scholar rightly insists, the abjad sequence is at the same time a numerical notation, and that it is viewed as such by Ṣafī al-Dīn al-Urmawī is shown by the fact that the continuation after y (10) is not with k (20) but with yā (11).

⁴³ It could be argued that in terms of the notation employed it is F that is here the more logical, for the abjad letter sequence begins to repeat at the perfect fifth (yā), so that just as ā-yā (D–A) is a perfect fifth so is d-yd (E–B) and z–yz (F#-c#). See Manik, , op. cit., 66–7Google Scholar, Wright, , op. cit., 26Google Scholar, Husmann, H., Grundlagen der antiken und orientalischen Musikkultur (Berlin, 1961), 107Google Scholar and n. 45, Sharḥ Mawlānā Mubārak Shāh, D'Erlanger, , op. cit., 3, 228.Google Scholar

⁴⁴ Wright, , op. cit., 27–9.Google Scholar

⁴⁵ Although it should be observed that in immediately restating the set F gives f, b and d. However, subsequent references are all to ṭ.

⁴⁶ cf. Wright, , op. cit., 28–9, 66.Google Scholar

⁴⁷ Mention is made of combinations with the rāst tetrachord both above and below. The latter possibility, ignored in KA, is of considerable interest for the history of this mode. Cf. Wright, , op. cit., 61, 75–6.Google Scholar

⁴⁸ The name is omitted from the circle, so that the only identification is the abbreviation h in the text. The original form could have been rāhawī as, usually, in KA.

⁴⁹ The name is omitted from the circle, so that the only identification is the abbreviation bz in the text. F defines buzurg first as zangūla with c# added, but then says that practitioners (ahl hādhih al-ṣinā'a) normally reverse (ya'kisūn) this, i.e. switch the D–G tetrachord and G–d pentachord around. (The notes defined around the circle for buzurg are wrong.)

⁵⁰ The name is omitted from the circle, so that the only identification is the abbreviation z in the text. F defines zīrafgand as a different reversal of buzurg, but the text then becomes confused in giving the details, defining the lower tetrachord first as nawā and then crossing this out and adding above the abbreviation for ḥijāzī. But this should also be rejected in favour of the following verbal definition (also crossed out) of this tetrachord as the fifth division, i.e., nawrūz, even though zīrafgand thus defined would still not coincide with the version in KA. Nothing is said about the remainder of the scale (possibly because the sequence of intervals occurring in it has not been defined). There is, however, a definition in terms of lute frettings in chapter 11 (fol. 19a), and this does coincide with the version in KA. The notes given on the accompanying circle relate to the second definition of buzurg. The relationship between this and zīrafgand may most readily be expressed by following Ṣafī al-Dīn al-Urmawī in juxtaposing their respective intervals to show the (extensive) area of identity. When transposing, Ṣafī al-Dīn al-Urmawī takes no account of the difference in size between steps (so that e.g. ‘ushshāq, which from D begins with a single step of a limma (D–E, the next note above), is portrayed from E^-c as beginning with a single step of a comma (E^-c–E, the next note above), and comparisons may most readily be effected by expressing intervals in terms of the number of steps (1, 2 or 3) they involve, irrespective of possible differences of size. We thus have:

buzurg: 2 3 2 2 1 3 2 2

zīrafgand: 2 2 3 2 2 1 3 2

⁵¹ Excluding from the original set of seven Iṣfahān, which is not mentioned as occurring in any other scale.

⁵² 52 Wright, , op. cit., 48–9.Google Scholar

⁵³ Or rather the more likely contender, for as 3 2 2 and 2 2 2 already exist (as rāst and rahāwī respectively) only 3 2 and 2 2 remain, and it is the latter that better relates to the structure of the mode as denned by Quṭb al-Dīn al-Shīrāzī (see Wright, , op. cit., 180, 197).Google Scholar

⁵⁴ The descriptive text begins D c A, but the following list converts this to D c B.

⁵⁵ Typical of the formulation is the very beginning:

⁵⁶ The problem would not be avoided by opting for the other sequence, that of rising minor sevenths/falling wholetones, in which x + 5 is the second step above, for e.g. ‘Irāq from both D and E^-c contains G; nor is it avoided by assuming that x + 5 is simply the fifth scale degree above x, for ‘irāq from both D and F#^-c again contains G.

⁵⁷ As given by al-Rajab the table is not wholly correct: yd should read yz in 1.4, and yw should read yz in 1.5.

⁵⁸ Risāla fī ‘l-luhūn wa-'l-nagham, ed Yūsuf, Z. (Baghdad, 1965)Google Scholar, tr. Shiloah, A. in ‘Un ancien traité sur le ‘ūd d'Abū Yūsuf al-Kindī’, Israel Oriental Studies, 4, 1974, 179–205.Google Scholar

⁵⁹ The text is defective towards the end, the definitions of abūsatīk being truncated virtually to the point of disappearance.

⁶⁰ Properly speaking a very close approximation to the Just Intonation interval consisting of a (9:8) wholetone plus two limmas. The corresponding value in practice was probably a (flexible) neutral third (see Wright, , op. cit, 37–43).Google Scholar

⁶¹ The definition is incomplete, three notes being omitted.

⁶² With one note missing.

⁶³ The chapter heading now adds ajma' ‘all’.

⁶⁴ Here, as also in buzurg and zīrafgand, there is an extra note, so that the octave note of rāst has to be lowered (and the total compass will no longer be two octaves).

⁶⁵ Which in its smaller manifestation is only one comma larger than the usual limma value.

⁶⁶ Although aqsām also occurs in F (chapter 6, fol. 9b).

⁶⁷ There is also an instance (chapter 11, fol. 18b) of shadd being used in its prior sense of ‘tuning’.

⁶⁸ cf. Wright, O., ‘Abd al-Qādir al-Marāghī and ‘Alī b. Muḥammad Binā'T: two fifteenth examples of notation. Part 2: commentary’, BSOAS, LVIII, 1, 1995, 31–3.Google Scholar

⁶⁹ For later instances see al Faruqi, op. cit.

⁷⁰ In chapter 10 (fol. 18a), for example, the set of transpositions through all 17 pitch positions is called jumlat taqāsīm al-hunūk (‘the complete set of hunūk divisions’). This plural form would relate more naturally to a reduced singular form, hank.

⁷¹ It should be recalled that the chapter heading refers to ‘unusual positions’ (al-amkina al-ghayr al-mu'tāda)—which probably outnumbered the ‘usual’ ones.

⁷² See Wright, , Modal System, 262–3, 270–81.Google Scholar

⁷³ KA chapter 9 (R 96), where they are dismissed as dissonant. Cf. Wright, , op. cit., 100.Google Scholar

⁷⁴ This form of classification was probably dispensed with because, on Ṣafī al-Dīn al-Urmawī's own admission, there was not a mursal version for each cycle.

⁷⁵ cf. Wright, , op. cit., 81–7.Google Scholar

⁷⁶ The text of the former is dated to 729/1329, that of the latter to 727/1327.

⁷⁷ This material, together with that relating to Ibn al-Khaṭīb al-Irbilī and al-Hasan b. ‘Abdallāh al-Ṣafadī, is taken from the comprehensive survey in Neubauer, E., ‘Arabische Anleitungen zur Musiktherapie’, Zeitschrift für Geschichte der arabisch-islamischen Wissenschaften, 6, 1990, 227–72.Google Scholar

⁷⁸ al-Ṭanjī, M. b. T, ‘al-ṭarā'iq wa-'l-alḥān al-mūsīqiyya fī ifrīqiya wa-'l-andalus', al-Abḥāth: Quarterly Journal of the American University of Beirut, 21/2, 3, 4, 1968, 93–116.Google Scholar

⁷⁹ No attempt is made here to reproduce the individual variations in the orthography of the mode names.