² These were: Ibn Amīr Hādjib, the host of Mūsā Mansā during his stay in Cairo; al-Mihmandār, the official charged with looking after Mansā Mūsā during his visit to Cairo; Muhannār b. ‘Abd al-Bāqī al-‘Adjramī, Mūnsā Mūsā's guide on his way from Cairo to Mecca. (Al-‘Umarī, Masālik al-Absār fī Mamālik al-Amsār (L'Afrique moins L'Égypte), traduit et annoté par M. Gaudefroy-Demombynes (Paris, 1927), 76, 78.)Google Scholar Another informant of al-‘Umarī was ash-shaykh Abū ar-Ruh ‘Isā az-zawāwī (from Zawāwa in Qabiliya, Alger), who met Mansā Musā outside the Sudan (ibid. 80–1).

³ Ibid. 58. He was from Dukkāla in Morocco.

⁴ Ibn-Battūta, , Tuhfat an-Nuzār fī Gharaā'ib al-Amsār wa-‘Adjā'ib al-Asfār, Texte Arabe et Traduction Française par Defrémery, C. et Sanguinetti, B. K. (Paris, 1922), Tome IV, 377 (3–6), 447 (8).Google Scholar

⁵ Ibn-Khaldūn, Kitāb Ta'rīkh ad-Duwal al-Islāmīya bil-Maghrib min Kitāb al-'Ibar, Teste Arabe, pub. par De Slane, M. G. (Alger, 1847), 1, 263. id. Histoire des Berbères et des Dynasties Musulmanes de l'Afrique Septentrionale, trad. par M. G. De Slane, 2^eme édition, 1925–1956, 11, 109–10.Google Scholar

⁶ Ibid. 261(4)/trans. 105.

⁷ Ibn-Khaldūn, 263(12–14)/trans. 110. See also 264 (2, 14)/trans. 110–11, where, concerning the exact pronunciation of the names of the kings, he twice says ‘In that way it was dotted by the shaykh ‘Uthmān’. The title faqīh ahl Ghāna may be attributed not to the state of Ghana (which probably did not exist any more) but to the area where this state once flourished.Google Scholar (Cf.Trirningham, J. S., A History of Islam in West Africa (Oxford University Press, 1962), 60.)Google Scholar

⁸ Ibn-Khaldūn, 264 (19–20), 265(15–17)/trans. 111–13.Google Scholar

⁹ Ibid. 266(17)–267(9)/trans. 114–15.

¹⁰ Ibid. 265(4–18)/trans. 112–13.

¹¹ Ibid. 11, 73(7–13)/trans. 111, 287–8 on the confusion of Takkadā and Tādmakka, see: Lhote, H., ‘Contribution à l'étude des Touaregs Soudanais’, Bulletin de l'I.F.A.N., 1955, 359–67.Google Scholar

¹² Ibid. 1, 557 (4–12)/trans. 111, 37–8. See also the article ‘Ibn-Khaldūn’ in the Encyclopaedia of Islam.

¹³ Ibid. II, 459(8–14)/trans. IV, 343.

¹⁴ Delafosse, M., Haut-Sénégal-Niger (Paris, 1912).Google ScholarMonteil, Charles, ‘Les Empires du Mali’, Bulletin du Comité des Etudes Historiques et Scientifiques de l'A.O.F. (1929), 295–447.Google Scholar

¹⁵ Ibn-Khaldūn, 264(1–3)/trans. II, 110–11.Google Scholar

¹⁶ Al-Bakri, Al-Masālik wal-Mamālik, K. al-Mughrib fi dhikr bilād Ifrīqiya wal-Maghrib, pub. par Slane, M. G. Dc (Alger, 1911), 178(3–19), and Description de l'Afrique Septentrionale, trad. par M. G. De Slane (Alger, 1953), cf. Trimingham, op.cit. 62; This identification was made by Delafosse, op. cit. 11, 174–5.Google Scholar

¹⁷ Ibn-Khaldūn, 264(3–6); trans. 111.Google Scholar

¹⁸ Ibn-Khaldun (ibid.) interprets the name Mārī-Djāta as follows: ‘Mārī means in their language Amir, a prince, who is an offspring of a Sultan; Djāta means a lion.’ Niane, Dj. T., ‘Recherches sur l'empire du Mali au Moyen Age’, Recherches Africaines (1959), 45, and Delafosse (op. cit. 177–8), who finds Ibn-Khaldūn's interpretation satisfactory, suggest interpretations of the name Sun-Djāta, and both admit that Djāta is the proper name.Google Scholar

¹⁹ Except for al-Qalqashandī, who is copying Ibn-Khaldūn. (Al-Qaiqashandī, , Subh al-A'shā (Cairo, 1915), V, 293.) Ibn-Battūta, 419(10)–420(2), relates that the grandfather of one Mudrik ibn Faqqūs had converted to Islam Sāriq-Djāta, the grandfather of Mansā Mūsā. This Sāriq-Djāta should be identified with Māri-Djātta and Sun-Djata, although he was not a grandfather of Mansā Mūsā but a brother of his grandfather, as we shall show below. Ibn-Battūta had not studied the genealogy of the kings of Mali.Google Scholar

²⁰ Ibn-Khaldūn, 264(6–9)/trans. 111.Google Scholar

²¹ Delafosse, op.cit. 184.Google Scholar

²² Ibn-Khaldūn, 264(9–10)/trans. 111.Google Scholar

²³ Ibid. 264(10–13)/trans. 111. Ibn Battūta (388) relates that among the Masūfa, the Berber inhabitants of Walata, the son is attached by genealogy not to his father but to his maternal uncle. A man's heirs are not his sons but the sons of his sister. The same traveller found that the sons of the sister of the Sultan of Takadda, a Berber, lived with him and would succeed him to the throne. Al-Bakri (175, trans. 328) describes the same custom in Ghana, where the king was succeeded by his sister's son. The ruling dynasty in Ghana, being in contact with the Berbers could, therefore, be Berber-influenced and have adopted a matrilineal succession to the throne (cf. Baumann, H. & Westermann, D., Les peuples et les civilisations de l'Afrique (Paris, 1948), 392–8). Al-Bakrī's account relates only to the succession to the throne and not to inheritance among the common people. Al-Bakrī himself (179, trans. 334) gives evidence on patrilineal inheritance among the Bakama, the inhabitants of Sāma, a province of Ghana, where the eldest son inherits all the property of his father.Google ScholarMauny, R. (Tableau Géographique de l'Ouest Africain au Moyen Age (IFAN, Dakar, 1961), 126) identified the Bakama with the Bambara, thus bringing us nearer to the people of Mali, where the patrilineal pattern probably prevailed.Google Scholar

²⁴ Niane, op. cit. (1959), 39.Google Scholar

²⁵ About the identification of this place see: Beckingham, C. F., ‘The Pilgrimage and Death of Sākūra king of Mali’, Bull. S.O.A.S. (1953), 391–2.Google Scholar

²⁶ Ibn-Khaldūn, 264(13–20)/trans. 111.Google Scholar

²⁷ Al-Malik an-Nāsir B. Qalā'ūn, reigned three times over Egypt. (i) 693–4 A.H. (A.D. 1293–94); (ii) 698–708 A.H. (A.D. 1298–1308); (iii) 709–41 A.H. (A.D. 1309–40). His first period of reign was ephemeral and could hardly be thought of as ‘the days of al-Malik an-Nāsir’. During his third period of reign, al-Malik an-NSsir was visited by Mansā Musā. Thus we have to assume that Sākūra's pilgrimage occurred in this second reign.Google Scholar

²⁸ Ibn-Khaldūn, 264(20–21)/trans. 112.Google Scholar

²⁹ Ibid. 268(2–4)/trans. 116.

³⁰ Delafosse (op. cit. it, 186) suggests that Mansā Qū was very old at his accession and died in the first year of his reign. There is nothing to support this supposition, which might however explain the gap in time between Qū and his brothers.Google Scholar

³⁰ According to the traditions of Kangaba-Niani, Sundjata had only one son—Yerelinkon, identified with the Mansā Ulī of Ibn-Khaldūn. The traditions of Dioma give a list of four sons: Yerelinkon (Mansā Ulī), Ko Mamadi (Mansā Qū), Bata Mande Bory (Abū. Bakr), and Niani Mamadu (Muhammad). (Niane, op. cit. (1960), 22.) Wātī and Khalīfa, mentioned by Ibn-Khaldūn, are omitted from the list, probably because of their insignificance. According to our reconstruction of the genealogy based on Ibn-Khaldūn: Qū was a son of Mansā Ulī; Abū-Bakr was a son of Māri-Djāta's daughter; Muhammad was a son of Qū. Thus we find that out of the four sons of Sundjata mentioned by the traditionalists of Dioma, only Yerelinkon (Ulī) was a son of Sundjata, while the others were his grandsons. This Yerelinkon is the only son of Sundjata recorded by the traditions of Kangaba-Niani, and the only powerful king from among the sons of Mārt-Djāta according to Ibn-Khaldun.Google Scholar

³² Ibn-Khaldūn, 264(23–3)//trans. 112.Google Scholar

³³ Delafosse, op. cit. II, 186; Monteil, op. cit. 393–2; Trimingham, op. cit. 67. See also the genealogical tables given by these authors.Google Scholar

³⁴ It seems as if the Arabic text which was translated by De Slane had: liā waladi ukhtihi. But De Slane himself has edited this text, and he gives no note on any variant. It is obvious that even in the very old MS. used by a1-Qalqashandī it was written: liā wuidi akhīhi (al-Qaiqashandī, V, 294(9–11)).Google Scholar

³⁵ Niane, Dj. T., Soundjata ou l'épopée mandingue (Paris, 1960), 17, 56, 147; Monteil, op. cit. 356, 362. Niane (‘Recherches’ etc. 39–40) is confronted with a difficulty, as he is convinced that according to the traditionalists, Abū-Bakr was a brother of Sundjata, while according to the French translation of Ibn-Khaldūn he was a son of Sundjata's Sister. This brings Niane and others to underestimate Ibn-Khaldūn as a reliable source. In fact, the fault in this case is not with the medieval historian, but with the modern translator.Google Scholar

³⁶ al-Umarī, 74; The former ruler had left the country for an expedition to discover the limits of the ocean. He appointed Mūsā as his deputy during his absence. The king did not come back, and Mūsā remained the ruler.Google Scholar

³⁷ It is possible to understand the Arabic word ‘ibn’ as denoting not only a son but also a grandson. Thus, ‘Mūsa ibn Abū-Bakr’ could mean also: Mūsā the grandson of Abū-Bakr. Similarly ‘Mansā Qū ibn Māri-Djāta’ could also mean Mansa Qū grandson of Mārī-Djāta (cf. supra.).Google Scholar

³⁸ Niane, op. cit. (1959), 40.Google Scholar

³⁹ Ibn-Khaldūn, 266(7–8)/trans. 114.Google Scholar

⁴⁰ al-‘Umarī, 73.Google Scholar

⁴¹ Ibid. 74, see above, n. 36.

⁴² Ibn-Khaldūn, 266(9)/trans. 114.Google Scholar

⁴³ Following Monteil, op. cit. 414–16.Google Scholar

⁴⁴ al-‘Umarī, 53.Google Scholar

⁴⁵ Ibn-Bauttūṭa, IV, 299(11), 400(9)–401(6), 419(6–8).Google Scholar

⁴⁶ This is also the meaning of the word in modern Mandinke (Mauny, op. cit. 449, quoting Delafosse).Google Scholar

⁴⁷ Ibid. 417–19.

⁴⁸ According to other MSS., his name was Qanbā or Fanbā, and according to al-Qaiqashandī (v, 297), Qantabā.Google Scholar

⁴⁹ Ibn-Khaldūn, 266(10–11)/trans. 114Google Scholar; ibid. 11, 459(7–8), trans. IV, 343.

⁵⁰ According to Ibn-Khaldūn, he was a grandson of Sulaymān's brother, while Ibn-Battūta referred to him as his paternal cousin. The Arabic word ‘Ibn ‘amm’, that means ‘paternal cousin’, may also denote any close relative in the partrilinear lineage. See above, n. 19, another fallacy of Ibn-Battūta concerning the genealogy of the ruling dynasty.Google Scholar

⁵¹ Was there endogamy in the ruling clan of Mali ? It is possible that Ibn-Batt¯ta was influenced by customs of the wider Muslim world. The situation might be still more complicated if one assumes that Qasa, the queen, was the mother of Qāsā, the son and successor of Sulaymān. (See Monteil, op. cit. 429.) This assumption is based on the common custom of placing the name of the mother before the son's name. Hence, Mansā Mūsā is called Kanku Mūsā after his mother, Kanku (see Mahmud Kati, Ta'rikh al-Fattāsh, Texte Arabe et Traduction Française par Houdas, O. & Delafosse, M. (Paris, 1913), texte 32–3; trad. 56). On the name Sundjata see: Niane, ‘Recherches’ (1959), 45. There are at least three arguments against this assumption: First, Qāsā was the title of the queen and not her proper name. Second, if Qāsā was the name of the mother, then the proper name of the son was omitted, while generally it remains together with the mother's name. Third, there are in the MSS. other versions of the name Qāsā (see above, n. 48). These versions, which are more close, rule out this assumption.Google Scholar

⁵² Ibn-Khaldūn, 11, 459(4–12)/trans. IV, 343.Google Scholar

⁵³ On the lengths of reign see ibid. 1, 266(7–11)/trans. 114.

⁵⁴ A slight mistake might occur in this calculation, since Ibn-Khaldūn is using Muslim years, which are 10–11 days shorter than Christian years.Google Scholar

⁵⁵ al-'Umari, 73.Google Scholar

⁵⁶ Ta'rikh Ibn-Khaldūn (Beirut, 1958–1959), V, 932. According to other sources the debts were paid off: see Ibn-Batūta, 431(4)–432(6), al-'Umari, 75; al-Qalqashandī, v, 296(9–11).Google Scholar

⁵⁷ Ibn-Khaldūn, 11, 394(7–10)/trans. IV, 243.Google Scholar

⁵⁸ Ibid. 11, 161(8–10)/trans. III, 411.

⁵⁹ Ibid. 395(1) ‘He sent them to the King of Mali, Mansā Sulaymān the son of Mansā Musā, as his father had died before the return of his deputation.’ Mansā Sulaymān is here called the son of Mansā Mūsā probably by mistake, as Mūsā was first succeeded by his son, while the deputation arrived in Mali already in the reign of Mansā Sulaymān.

⁶⁰ Cf. ‘Ibn-Khaldūn’, Encyclopaedia of Islam, 11.Google Scholar

⁶¹ This was first suggested by Monteil, op. cit. 321–2. Delafosse, M. (op. cit. 190) gives 1332 as the date of Mansā Mūsā's death, but he wrongly Connects it with the deputation to Fes, as he did not consider the date of the conquest of Tlemcen.Google Scholar

⁶² See above, 2–3.Google Scholar

⁶³ Ibn-Khaldūn interprets this name saying that ‘Marl’ means ‘wazir’. Early in the same chapter he interpreted the title ‘Marl’ as denoting ‘prince’ (see above, n. 18). This shows that interpretations of Sudanese names and titles by Ibn-Khaldūn are very flexible, although some of them are satisfactory. It is possible, however, that this Mārī-Djāta was a member of the ruling dynasty, as this name was probably common in the family.Google Scholar

⁶⁴ Mauny (op. cit. 450), quoting Delafosse, explains that ‘santigui’ is the official who is in charge of the royal expenditure (tigui=chief, sa=purchasing). Monteil (op. cit. 432) suggests that it was the ‘dyon-sandigui’, the chief of the slaves. This interpretation confirms his analysis of the government of Mali, where he assigns an important role to slaves.Google Scholar

⁶⁵ Ibn-Khaldūn, 1, 266(11)–268(4)/trans. II, 115–16.Google Scholar