1. An oral informant of the Batchili lineage, in Arnaud, Robert, “L'Islam et la politique musulmane française en Afrique Occidentale Francaise, suivi de la singulière légende des Soninké,” tirage à part du Comité de l'Afrique francaise (Paris, 1911), 151.Google Scholar

2. Early scepticism was expressed by Monteil, Charles, “Les empires du Mali,” Bulletin du comité d'études historiques et scientifiques de l'Afrique occidentale française, 12 (1929), 291–447, 364.Google Scholar See also Innes, Gordon, Sunjata: Three Mandinka Versions (London, 1974), 26.Google Scholar

3. See for example, Levtzion, Nehemia, Ancient Ghana and Mali (London, 1973), 16–20, 55–60.Google Scholar

4. Bathily, Abdoulaye, “A Discussion of the Traditions of Wagadu with some references to ancient Ghana,” Bulletin de l'Institut Fondamental d'Afrique Noire, 37B (1975), 8.Google Scholar

5. al-Mukhtar, Ibn (wrote ca. 1665) Tarikh el-Fettach, Arabic text and French translation by Houdas, O. and Delafosse, M. (Paris, 1913), 75–76Google Scholar; es-Saʿdi, Abderrahman (wrote ca. 1655), Tarikh es-Soudan, Arabic text and French translation by Houdas, O. (Paris, 1911), 18Google Scholar; Bathily, , “Discussion,” 9.Google Scholar

6. Arnaud, , “Islam,” 155.Google Scholar

7. Bathily, , “Discussion,” 13.Google Scholar

8. Al-Bakri, , Kitab al-masalik wa-'l-mamalik, ed. de Slane, Baron MacGuckin, in Hopkins, J.F.P. and Levtzion, N., eds., Corpus of Early Arabia Sources for West African History (Cambridge, 1981), 79.Google Scholar

9. Bathily, , “Discussion,” 13.Google Scholar

10. For more on this subject see Conrad, David C. and Fisher, Humphrey J., “The Conquest That Never Was: Ghana and the Almoravids, 1076. II. The Local Oral Sources,” HA, 10 (1983), 53–54.Google Scholar

11. Bathily, , “Discussion,” 22.Google Scholar

12. Adam, M.G., “Légendes historiques du pays de Nioro (Sahel),” Révue coloniale (1903), 83–95, 354–62Google Scholar; Delafosse, Maurice, “Traditions historiques et legendaires du Soudan occidentale,” Bulletin du comité de l'Afrique française et du comité du Maroc (1913), 293–301Google Scholar; Arnaud, , “Islam,” 145–58, 166–72.Google Scholar

13. Lanrezac, H.C., “Au Soudan: la légende historique,” La revue indigène (1907), 292–97, 380–86.Google Scholar

14. Mali National Archives, series 1D, “Monographies-Etudes-Coutumiers,” 1D-1 through 1D-305. Titles of individual studies are listed in Archives Nationales du Mali: Répertoire, 1855–1954 (Bamako, 1974), 8–27.Google Scholar

15. Beginning with the Coppolani mission to the Sahel in 1898/99, Arnaud had several assignments to the Western Sudan during which he could have visited Nioro and recorded traditions, but the most likely time appears to be when he was sent out on a study mission in 1906. For dates on Adam, Arnaud, and Chartier I am grateful to Mme. M. Pouliquen of the Section Outre-Mer, National Archives of France. For the date on Lanrezac I want to thank General Delmas, Chief of Services Historiques de l'Armée, Vincennes.

16. Monteil, Charles, “La légende du Ouagadou et l'origine des Soninké,” Mélanges Ethnologiques (Paris, 1953), 362–408.Google Scholar

17. Arnaud, , “Islam,” 159.Google Scholar

18. Adam, , “Légendes,” 81.Google Scholar

19. Lanrezac, , “Au Soudan,” 380.Google Scholar

20. Delafosse, , “Traditions,” 293.Google Scholar Delafosse stated that the Bambara version was dictated by two “notables,” Hadi Bà and Mamadou Sallama, but it is highly unlikely that two informants would dictate a single oral account and there is no evidence that Sallama was himself a narrator of oral testimony. It is much more likely that Bà was the narrator and Sallama the recorder. Though there may have been other local historians whose work has not survived, there also may have been fewer than it first appears. For reasons he did not give, Delafosse claimed that the recorder of his own Arabic version, Mamadi Aissa, had previously provided Adam with the account he published, and that Lanrezac had received his version from Aissa's nephew, Mamadou Sallama. There is nothing to contradict this in Adam's description of his informant as a young Soninke marabout, for this could have applied to a man who had become qādī by the time Delafosse heard about him several years later. Similarly, Lanrezac's Arabic scrivener could have been Mamadou Sallama. If the qādī and his nephew did indeed record the earlier versions and pass them on to Adam and Lanrezac, this only adds to the local scholars' importance as collectors and preservers of traditional history texts.

21. Adam, , “Légendes,” 85Google Scholar (he spelled it “Dowichs”); Arnaud, , “Islam,” 148Google Scholar (“Douaich”); Delafosse, “Traditions,” 295Google Scholar (“Douaissé”); Lanrezac, , “Au Soudan,” 382 (this version is muddled, identifying the “Douassi” as a son who went into the Sahel and founded a Moorish clan)Google Scholar; Monteil, , “Legende,” 373Google Scholar (Doysi). According to G.P. Murdock, the sahelian Duaish are culturally indistinguishable from the Zenaga, except that some of them still speak a Berber language Africa: Its Peoples and Their Culture History (New York, 1959), 112.Google Scholar

22. The snake also appears in other Wagadu descent lists, but as the offspring of different wives of Dinga, whose various mates produced the beginnings of numerous lineages.

23. Niane, D.T., Sundiata: An Epic of Old Mali (London, 1965), 41, 47.Google Scholar

24. Innes, , Sunjata, 26.Google Scholar

25. Characteristically, the times values are unclear, and many references to Sumanguru are anachronistic because ancient families tend to be identified in the traditions by their most distinguished member, while the rest are forgotten. Thus it is conceivable that the Kante-Bamagana lineage in general could be remembered through many generations simply as Sumanguru, and the one hero could be credited with deeds, not only of others in the lineage, but of anyone with whom they were associated in the minds of the oral traditionists. Also to be considered is the fact that “Sumanguru” is, in modern times at least, also a surname rendered “Soumarouo,” and this may have been so in ancient times as well. Arcin, André, Histoire de la Guinée franciase (Paris, 1911), 50Google Scholar, gives as alternative spellings, “Souma N'Kourou” and “Souma-Horo” and adds that Hor or Har means “noble” in Berber, and that it is the name of one group of the “Souma” people. We cannot attach much weight to differences in spellings, because these often resulted simply from the various ways they were written down by recorders of oral texts. Nevertheless, Arcin's comment is interesting in view of Sumanguru's alleged servile descent, suggesting the possibility that his forebears were of Berber origin. However, the Manding term for noble is also horo.

26. The definitive discussion of the origin and meaning of this term is by Bathily, Abdoulaye, “Discussion,” 9–11.Google Scholar Among the linguistic possibilities Bathily suggests for restoring the original Soninke pronunciation, he favors the idea that “Kaya Magha” is the mistranscription of “Kuya Manga,” a sovereign's flattering nickname meaning “generous ruler.”

27. Bathily, , “Discussion,” 36.Google Scholar He translates komo xooro (sing. komo xoore) as “great slaves.”

28. Monteil, Charles, Les Bambara du Ségou et du Kaarta (Paris, 1923), 193.Google Scholar

29. N'diaye, Bokar, “Les structures politico-sociales de l'ancienne société Mandingue,” Conference on Manding Studies, SOAS, University of London, 1972, 15.Google Scholar

30. Pacques, V., Les Bambara (Paris, 1954), 78–79.Google Scholar

31. Monteil, , Bambara, 191.Google Scholar

32. Monteil, , “Legende,” 403.Google Scholar In Bambara the term jòn kòròw (Monteil spells it jon gorow) designates people who have been slaves or whose ancestors were slaves. Arnaud's informant seems confused with respect to the term Komon Gallo, which he equates with the Kusa. He says the Komon Gallo were royal slaves at Kumbi from the clans Toun'Kara, Kouloubali, and Badiaka, while at the same time agreeing with all other sources that the Tunkara were Kusa (“Islam,” 167).

33. Arnaud, , “Islam,” 167.Google Scholar

34. Monteil, , “Légende,” 404.Google Scholar

35. Ibid., 403–405; Arnaud, , “Islam,” 167.Google Scholar

36. Meillassoux, Claude, “Les origines de Gumbu (Mali),” BIFAN, 34 B (1972), 270–71.Google Scholar

37. Arnaud, , “Islam,” 167.Google Scholar

38. Monteil, , “Légende,” 404.Google Scholar

39. Arnaud, , “Islam,” 167.Google Scholar As originally told to Arnaud in the Soninke language, the term “ancien captif” could have been meant to signify that Sumanguru was a descendant of those who had served the Kayamaga, rather than that Sumanguru himself had been a slave.

40. The informants generally agree that the ruler under whom Sumanguru or his antecedents had served as slaves was the last Kayamaga of Wagadu, though there is no agreement on who the last one was. Batchili claimed it was Kayamagha Tanné (Arnaud, , “Islam,” 167Google Scholar), and Adam's informant said it was Manga Koumas (“Légendes,” 358), a name not included in the descent list he gave, though it can be seen in Arnaud's (Koumma) and possibly Lanrezac's (Kounia).

41. Monteil, , “Empires,” 354–55Google Scholar and “Fin de siècle à Médine (1898–1899),” BIFAN, 28 B (1966), 166.Google Scholar In “Empires” (354), Monteil says the chief of slaves was called dyon-sandigui (jòn santigi), a term that makes no sense in Bambara. Santigi (or Satigi) is a proper name (translated roughly as “owner of the year”) often given to a surviving child of parents who had several previous non-survivors; san = to buy, jòn sanna = slave buyer, and by extension, the chief of slave buyers or handlers might have been called jòn son kuntigiw, a shortened form of jòn sanna kuntigiw.

42. Hopkins, /Levtzion, , Corpus, 333.Google Scholar

43. Barth, Heinrich, Travels and Disooveries in North and Central Africa 1849–1855 (3 vols.: London, 1857), 3:660.Google Scholar

44. Delafosse, Maurice, Haut-Sénégal-Niger (3 vols.: Paris, 1912), 2:165.Google Scholar

45. Monteill's informant in “Empires” (355) claimed that the Bamagana were a branch of the Kante lineage. Innes' informant Dembo Kanute (Sunjata, 311) gives Bamagana as the original surname, with Kante being acquired at the time of the conflict with Sunjata. This is consistent with Bemba Suso's statement (Innes, , Sunjata, 81Google Scholar) that presently familiar surnames originated in the time of Sunjata. There are many popular etymologies accounting for the name Kante, one of which relates it to Sumanguru's other name, Bamagana (Monteil, , “Légende,” 370n5Google Scholar, and “Fin de siècle,” 166).

46. Cornevin, Robert, Histoire de l'Afrique, Tome I: des origines au XVI^e siècle (Paris, 1962), 347–48Google Scholar; his Delafosse reference is Haut-Sénégal-Niger 1:256–57.Google Scholar

47. Pageard, Robert, “Note sur les Kagoro et la chefferie de Soro,” Journal de la Societé des Africanistes, 29 (1959), 264.Google Scholar

48. Levtzion, , Ancient Ghana and Mali, 51.Google Scholar

49. Monteil, , “Empires,” 354.Google Scholar

50. Delafosse, , Haut-Sénégal-Niger 2:163–64.Google Scholar

51. Lanrezac, , “Au Soudan,” 385Google Scholar; Adam, , “Legendes,” 92Google Scholar; Arnaud, , “Islam,” 151Google Scholar; Delafosse, , “Traditions,” 297.Google Scholar

52. Delafosse, , Haut-Sénégal-Niger 2:162.Google Scholar The Adam version (“Legendes,” 89) gives Soumane Fade's clan as the Diareni.

53. Delafosse, , Haut-Sénégal-Niger 2:163Google Scholar, where he attributed the kinglist to “local tradition.”

54. Elsewhere he mentions Tautain as being one of several who had reported on a Soninke legend of the founding of Wagadu, , Haut-Sénégal-Niger 1:256, and 256n2).Google Scholar

55. Tautain, L., “Légende et traditions des Soninké relatives à l'empire de Ghanata,” Bulletin de géographie historique et descriptive, 9/10 (1894/1895), 472.Google Scholar

56. Ibid., 475.

57. Adam, , “Légendes”, 89Google Scholar; Lanrezac, , “Au Soudan,” 385Google Scholar; Arnaud, , “Islam,” 147–48Google Scholar; Delafosse, , “Traditions,” 296.Google Scholar Notably absent from these sources and from Tautain's list are Al-Bakri's Tunka Manin (1063–1068) and his predecessor Basi (Hopkins, /Levtzion, , Corpus, 79)Google Scholar, as well as Kanissai, who is mentioned in the Tarikh el-Fettaah (76).

58. Tautain, , “Légende,” 475–76.Google Scholar He added that according to some, events in the story of the snake (and the destruction of Wagadu) occurred during Biramu's reign though, he noted, others say it was later.

59. Delafosse, , Haut-Sénégal-Niger 2:163–64.Google Scholar References to Kambine Jariso are rare, but a fragment in the Mali archives supports his importance as the primary hero of the Jariso. It describes how he was seriously wounded in a fierce battle with Manga Khonne. Though the manuscript has been partly destroyed by termites, it is possible to make out that the battle continued on foot after the horses were killed, and that Manga Khonne threw Kambine to the ground, giving him a bloody wound that stained his blue shirt. We are told that the Jariso later claimed that the blood issued from Manga Khonne instead of Kambine, and that thereafter the Jariso adopted the striped shirt in place of the blue, with their chiefs taking the name of Manga, apparently as an honorary title. “S/du Commandement chez les Diawara ‘Historique’ Nara - 1918,” Archives Natlonales du Mali, 1D-78.

60. Lanrezac has Diacouraga Traore for Diamera Sokhona (“Légende,” 385). Similarities in these lists stand out when compared to others, e.g. one collected by Penel near Bafoulabe ca. 1895: Igoorédinga's sons were Diamé-Sissé, Tiroboungari-Sissé, Kiné-Sourouna, Slramanganda-Touré, and Manga-Sissé. Penel, Julien, “Coutumes soudanaises (Malinké, Sarakolé, Khassonké)” (2 vols.: Paris, 1895) 1:8–9.Google Scholar

61. Whatever the truth of these relationships, the Wagadu tradition generally has Dinga occupying the mythical role of a more or less divinely creative figure, father of Wagadu and all its inhabitants, rather than a more historically-rooted figure, designated patriarch of a single lineage. The magha or kayamaga Diabe Cisse and his descendants of the same title seem to be the only rulers who are remembered, because tradition has them in power both at the founding of Wagadu and during the killing of the snake Bida which heralded the state's destruction. However, other lineages, such as those repeatedly mentioned as holding provincial governorships, may also have ruled at some period, though the possibility of this ever being confirmed is more complicated than usual because the provinces they are said to have governed also appear in the collective griot memory as Wagadu itself, and so do the later Soninke successor states. See especially Frobenius, Leo, “Gassires Laute” in Spielmannsgeschichten der Sahel, Atlantis VI (Jena, 1921), 53–60.Google Scholar

62. In an accompanying note, Tautain mentioned that a group of Soninke from the Jariso family moved from Kaarta to settle at Damfa, but this was not until 1753. These settlers claimed that their people were originally from Jimbala in Wagadu, from whence they had emigrated to Kaarta (“Légende,” 477n1). In the Tarikh el-Fettach (68n3) Houdas and Delafosse noted that Damfa, as well as Soso, was part of Kanlaga.

63. Tautain, , “Légende,” 477.Google Scholar He notes that “Suleyman” is the first Muslim name in the list, attaching a date of 1087 to Suleyman's reign. Among the other names succeeding Kambine's in this list, “Wagadu Makha II” is difficult to accept as a proper name rather than a title indicating Wagadu's second ruler. “Gane” is a rare occurrence in the oral sources of a name similar to the “Ghana” of the Arab writers who give this as the name of the ruler and his city or country: Ibn Hawqal (A.D. 988) in Hopkins/Levtzion, , Corpus, 49Google Scholar; Al-Bakri, (1068), Corpus, 79Google Scholar; Kitab al-Istibsar (1192), Corpus, 196. “Musa” is of course also a Muslim name, and “Biramu” stands out because it is close to “Birama,” the name of the famous Tunkara ancestor who ruled at Mema. On the whole, this appears to be a catch-all list of random names from the past.

64. Delafosse, , Haut-Sénégal-Niger, 2:163–64.Google Scholar

65. Tautain, , “Legende,” 476.Google Scholar A tradition preserved in the Tarikh el-Fettach (70–71) places the Kusa among the people of Kaniaga subsequent to the collapse of the Kayamaga's kingdom.

66. Delafosse, , Haut-Sénégal-Niger 2:164.Google Scholar

67. Tautain, , “Légende,” 476.Google Scholar

68. Delafosse, , Haut-Sénégal-Niger, 2:164–65.Google Scholar

69. At the beginning of the section he describes the post-Almoravid emigration of some Soninke clans to Kaniaga, citing a legend in Delafosse, , Haut-Sénégal-Niger, 1:256.Google Scholar

70. Cornevin says there were seven members of the dynasty, which matches Delafosse's number, but he names only six, passing directly from Kambine to Banna-Boubou, leaving out Suleyman. He also varies slightly from Delafosse's dates, giving 1078 instead of 1076 as the beginning of Kambine's reign.

71. Cornevin, , Histoire, 347–48.Google Scholar

72. Niane, , Sundiata, 92n47.Google Scholar In Niane's text (38) the informant identifies the Jariso as a “line of smiths” and claims that Sumanguru was descended from them.

73. See also Conrad, David C., “Maurice Delafosse and the pre-Sunjata trône du Mandé,” Bulletin of the School of Oriental and African Studies, (1983), 335–37.Google Scholar

74. Arnaud, , “Islam,” 167Google Scholar; Monteil, , “Empires,” 355.Google Scholar

75. Westermann, D. and Bryan, M.A., Handbook of African Languages II, Languages of West Africa (London, 1952), 32.Google Scholar

76. Arnaud, , “Islam,” 167Google Scholar; Monteil, , “Légende,” 403–04Google Scholar; Meillassoux, , “Origines,” 270–71.Google Scholar

77. Arnaud, , “Islam,” 167Google Scholar; Monteil, , “Légende,” 405.Google Scholar

78. Humblot, P., “Du nom propre et des appellations chez les Malinké des vallées du Niandan et du Milo (Guinée francaise),” Bulletin du Comite d'Etudes Historiques et Scientifiques de l'Afrique Occidentale Francaise, (1918), 523.Google Scholar

79. “L'organisation sociale et politique du cercle de Kita 1944: les origines du peuplement,” Archives Nationales du Mali, 1D-43-7. According to Westermann and Bryan (Handbook, 34), the Kagoro around Kolokani and Nioro are a mixture of Bambara and Fulani.

80. Monteil, , “Légende,” 369–70, 370n5.Google Scholar Another legend of Sumanguru's birth gives him two mothers and a birth more contemporary with Sunjata in Frobenius, Leo, Dichten un Denken im Sudan, Atlantis V (Jena, 1925), 321.Google Scholar

81. Monteil, , “Légende,” 369–70Google Scholar; Arnaud, , “Islam,” 167.Google Scholar

82. Although in testimony addressed to later times there is less emphasis on former Tunkara servility, and some informants make no reference to it at all, e.g. Archives Nationales du Mali, 1D-78 (1918) and 1D-43-7 (1944).

83. Monteil believed that there was a direct association between the name Tunkara and the Soninke term tunka, meaning “chief” or “king” (“Legende,” 404), but see Bathily, , “Discussion,” 13–14.Google Scholar

84. Arnaud, , “Islam,” 167.Google Scholar

85. Meillassoux, , “Origines,” 270.Google Scholar

86. Monteil, , “Légende,” 404.Google Scholar