142. Jeli Baba Sissoko, from a tape in the collection of the Institut des Sciences Humaines, Bamako, Republic of Mali, recorded 24 July 1975 with the permission of the director, Mamadou Sarr. Subsequent references to the Sissoko version are from the same source. Wagadu can be roughly translated as meaning “land of great herds,” implying fine pasturage resulting from an ample water supply, which might account for the griot's abstruse reference to bottomless wells. However, according to Charles Monteil's sources, the Soninke aristocracy were known as wago (sing. wage), and he believed “Wagadu” simply meant “land of the Wago” (“La légende du Ouagadou et l'origine des Soninké,” Mélanges Ethnologiques (Paris, 1953), 397)Google Scholar; Monteil's principal informant when he collected his version of the Wagadu tradition in 1898 was a Soninke qesere (griot) of Gumbu named Tudo Yaressi (ibid., 365).

143. Conrad, David C. and Fisher, Humphrey J., “The Conquest that Never Was: Ghana and the Almoravids, 1076, I, the External Arabic Sources,” HA, 9(1982), 21–59.Google Scholar

144. Tautain, L., “Légende et traditions des Soninké relatives à l'empire de Ghanata… 1887,” Bulletin de géographie historique et descriptive, 9/10 (1894/1895), 473Google Scholar; Meillassoux, C., Doucouré, L., and Simagha, D., Légende de la dispersion des Kusa (épopée Soninké) (Dakar, 1967), 8.Google Scholar Members of the proprietary classes of ancient Ghana were known as wago. Monteil, , “La légende,” 397Google Scholar and below, note 182.

145. See for example, Monteil, Vincent, L'Islam Noir (Paris, 1971), 65–70Google Scholar; Levtzion, Nehemia, Ancient Ghana and Mali (London, 1973), 20–22.Google Scholar

146. Though usually implied, an overt expression of the time-lessness of Wagadu appears in Frobenius, Leo, Spielmannsgeschichten der Sahet, (Jena, 1921), 60.Google Scholar The material cited from Frobenius was collected in 1909 from a Jerma in the Songai dialect of northern Togo.

147. Although Jeli Baba Sissoko mentions this date in connection with the founding of a village after the ruin of Wagadu but in the very year of the hijra, he is sufficiently suspicious of it to remark that “this is what some people have estimated. To say that the people of Kumbi dispersed in a certain year is very difficult, but this is how it was told to me.” See also the hijra reference from the Tārīkh al-Fettāsh quoted below.

148. Comparing concepts of the “remote past” and the “recent past,” G.I. Jones remarks that “the one changes abruptly to the other, or is linked by a series of disconnected episodes whose position is uncertain and whose sequence is in doubt.” Jones, , “Time and Oral Tradition with Special Reference to Eastern Nigeria,” JAH, 6 (1965), 153.CrossRefGoogle Scholar

149. There is a useful map in Levtzion, , Ancient Ghana, 2.Google Scholar See also Mauny, Raymond “The Question of Ghana,” Africa, 24 (1954), 200–213.CrossRefGoogle Scholar

150. Delafosse, Maurice, La langue Mandingue et ses dialectes, (2 vols.: Paris, 1929–1955), 2:23.Google Scholar

151. Abderrahman es-Saʿdi (wrote c. 1655), Tarikh es-Soudan, Arabic text and French translation by Houdas, O. (Paris, 1911), 9/18.Google Scholar

152. According to Westermann, D. and Bryan, M.A., Handbook of African Languages Part II, Languages of West Africa, (London, 1970), 31Google Scholar, Soninké is in the Mandé Tan group of Mandé (Manding) languages, which includes the Malinké–Bambara-Dyula dialect cluster, as well as the Khasonké and Vai.

153. It does appear in the Frobenius version, Spielmannsgeschichten, 53, where “Ganna” is described as one of four epochs in the history of Wagadu, but there are no useful spatial or temporal references for the epochs, and no other versions of the episode exist for comparison.

154. Delafosse, , Langue Mandingue, 2: 241.Google Scholar Moreover, Monteil's informant claimed that the Soninké ancestor Dinga carried the title Taga-du-n-kana, or “chief of the blacksmiths,” “Légende,” 372n2.

155. See Arnaud, R., “L'Islam et la politique musulmane française en Afrique Occidentale Française, suivi de la singulière légende des Soninké,” tirage à part du Comité de l'Afrique française (Paris, 1911), 155.Google Scholar See note 252 below.

156. Frobenius, , Spielmannsgeschiahten, 49–52.Google Scholar Many of Frobenius' conclusions are highly questionable (Monteil lists some of them in “Légende,” 367–68), and the possibility of the existence of the dausi and of the related genre of shorter material called pui as perceived by him, should be studied further before being accepted as having been part of west African tradition. See, for example, Knappert, Jan, “The Epic in Africa,” Journal of the Folklore Institute, 4 (1967), 178–79.CrossRefGoogle Scholar

157. Perhaps the best example of this is the Manding epic of Sunjata, which consists of a series of episodes, the number and contents of which depend on the knowledge and volition of the griot (bard) presenting them. For accurately reproduced versions see Innes, Gordon, Sunjata: Three Mandinka Versions (London, 1974).Google Scholar

158. For a description of the techniques of the Manding bards known as griots see ibid., 2–30, and the companion volume, Kaabu and Fuladu: Historical Narratives of the Gambian Mandinka (London, 1976), 27–30.Google Scholar

159. Both the patriarch Dinga and his descendant Diabé Cissé appear in this role, depending on the version.

160. See Curtin, Philip D., “The Uses of Oral Tradition in Senegambia: Maalik Sii and the Foundation of Bundu,” Cahiers d'Études Africaines, 15(1975), 190.CrossRefGoogle Scholar

161. Meillassoux, Claude, “Histoire et institutions du kfo de Bamako d'après la tradition des Niaré,” Cahiers d'Études Africaines, 4(1964), 189n2.Google Scholar

162. This would have been a relatively late addition to the legend of Wagadu. The story of Jacob and Esau is included in many collections of stories dating from the early days of Islam--e.g., al-Kīsa'ī, Muhammad ibn ʿAbd Allāh (fl. eleventh century), Vita Prophetarum (Qisas al-anbiyā), ed. Eisenberg, Isaac (Leiden, 1923), I: 153–54Google Scholar, and al-Tha ʿlabī, Ahmad ibn Muhammad (d. 1035), Qisas al-anbiyā (Cairo, 1954), 101Google Scholar--and these were probably carried to the western Sudan centuries ago by pilgrims returning from Mecca, who introduced them to generations of griots. These stories can still be heard weekly on Radio Mali.

163. Of Dinga's overall place in the tradition, it can be said that he occupies the mythical role of a semi-divine creative force, father of Wagadu and all its inhabitants, as opposed to being a more historically-rooted figure, who is thought of as the patriarch of a single lineage.

164. Arnaud, , “Islam,” 150–51Google Scholar; Monteil, , “Légende,” 376Google Scholar; Frobenius, , Spielmannsgeschichten, 62Google Scholar; Adam, M.G., “Légendes historiques du pays de Nioro,” Revue Coloniale (1903) 87Google Scholar, informant a young Soninké marabout; Lanrezac, , “Au Soudan: La légende historique,” La Revue Indigène (1907), 383Google Scholar, informant an Arabic scrivener at Nioro; Delafosse, M., “Traditions historiques et légendaires du Soudan occidentale” in “Renseignements coloniaux et documents,” Bulletin du Comité de l'Afrique Française et du Comité du Maroc (1913) 295Google Scholar, translated from an Arabic manuscript, the original of which is thought to have been produced by a Muslim judge at Nioro, ibid, 293.

165. Levtzion, , Ancient Ghana, 47.Google Scholar

166. Farias, P.F. de Moraes, “Great States Revisited,” JAH, 15 (1974), 484.CrossRefGoogle Scholar

167. Jan Vansina refers to such historical residue as “memories of features that no longer exist,” Oral Tradition (Chicago, 1965), 158.Google Scholar

168. See for example, Dieterlen, G., “Mythe et organisation sociale au Soudan français,” Journal de la Société des Africanistes, 25(1955), 39–76CrossRefGoogle Scholar; idem, “Mythe et organisation sociale en Afrique occidentale,” Journal de la Société des Africanistes, 29(1959), 119–38; idem, “The Mande Creation Myth,” Africa, 27(1957), 127–34. For the Bambara, specifically see her Essai sur la religion Bambara (Paris, 1951).Google Scholar For the Dogon see Griaule, M., Conversations with Ogotemmêli: An Introduction to Logon Religious Ideas (London, 1965).Google Scholar

169. It is, for example, generally believed that Sunjata Keita was a moving force in the early thirteenth-century emergence of the Mali Empire.

170. Against which Farias, , “Great States,” 484Google Scholar, warns.

171. Writing on Saharan tradition H.T. Norris says that the serpent is a “typically African” element, Saharan Myth and Saga (Oxford, 1972), 8–9Google Scholar; but compare the snake in Ishaq's, IbnSirat Rasul Allah, tr. Guillaume, A. as The Life of Muhammad (London, 1955), 84.Google Scholar

172. Jeli Baba Sissoko, a distinguished griot of Mali, is himself a Muslim.

173. Tarikh es-Soudan, 9/18.

174. al-Mukhtār, Ibn (wrote ca. 1665) Tarikh el-Fettaah, Arabic text and French translation by Houdas, O. and Delafosse, M. (Paris, 1913) 42/78.Google Scholar

175. Levtzion, N., “Patterns of Islamization in West Africa,” in McCall, D.F. and Bennett, N.R., eds., Aspects of West African Islam, (Boston, 1971), 34.Google Scholar

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

177. Ibid, 157. The loosely-employed Manding (Bambara, Maninka) term for “Arab” or “Moor” is “Suraka,” the Soninké variance of which is “Suraké,” as is noted by Arnaud (ibid, 157n2) whose “Maure” is the usual French translation. See Bazin, H., Dictionnaire Bambara-Française (Paris, 1906), 569 and above n.20.Google Scholar

178. Al-Bakrī, , Kitāb al masālik wa al-mamālik, ed. de Slane, M.G. (Paris, 1911), 175/328.Google Scholar See also a translation of sixteen chapters by Monteil, V., “Al-Bakri (Cordue, 1068), routier de l'Afrique blanche et noire du Nord-Ouest,” BIFAN, 30B(1968), 70.Google Scholar

179. Meillassoux, , “Histoire,” 188.Google Scholar

180. Monteil, , “Légende,” 379Google Scholar; Sissoko, interview.

181. Adam, , “Légendes,” 92–93Google Scholar; Sissoko, interview.

182. Monteil, , “Légende,” 397.Google Scholar He was told that there were six categories of wago, including those occupied by Dinga's descendants, the clans of the four fado or provincial chiefs, and certain other functionaries and leading citizens.

183. Arnaud, , “Islam,” 156Google Scholar; Frobenius, , Spielmannsgeschichten, 67Google Scholar; Monteil, , “Légende,” 379.Google Scholar

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

185. Certain servile clans are said to have been state property or “captifs de la couronne” who were known by the generic term Kusa, and possibly as Kagoro. Arnaud, , “Islam,” 167Google Scholar; Monteil, , “Légende,” 403–05.Google Scholar The Soninké term for enfranchised slaves was komon gallo, Monteil, , “Légende,” 403Google Scholar, the Bambara equivalent of which is jon gorow, Monteil, C., Les Bambara du Ségou et du Kaarta (Paris, 1924), 193.Google Scholar Monteil concluded that during Wagadu's declining years the power of certain slave chiefs increased to such a degree that the head of state was obliged to maintain extreme vigilance against these “usurpateurs présomptifs,” “Légende,” 404.

186. Tarikh el-Fettach, 75–76.

187. Ibid. Delafosse was aware that in the Manding language variations on the term maga meant “master” (La langue Mandingue 2:493Google Scholar), and he acknowledged that the Tarikh el-Fettach (75) gave “Kayamaga” as “king” or “master” of gold. Nevertheless, on the strength of the term's popular usage in recent times as a proper name, and because of the way it appears in the Wagadu descent lists, e.g., “Maghan Diabé Cissé” (see Adam, , “Légendes,” 85Google Scholar; Arnaud, , “Islam,” 147Google Scholar; Delafosse, , “Traditions,” 295Google Scholar; Lanrezac, , “Au Soudan,” 382Google Scholar; Monteil, , “Légende,” 373Google Scholar), he concluded that “Maga, Maghan or Makka” was the name rather than the title of the rulers in the Cissé dynasty (Tarikh el-Fettach, 75n1). On the contrary, “Maghan” (Maga, Manga) appears much more likely to have been the title held by the Wagadu rulers, the Cisse in particular, and may even be the term construed by Arab writers into the familiar “Ghana.”

188. Saint-Père, J.H., Les Sarakollé du Guidimakha (Paris, 1925), 12.Google Scholar In this fragment Diabé Cissé himself is identified as the snake killer, perhaps because the informant could not recall names of other characters in the legend. See also Frobenius, , Spielmannsgeschichten, 65–67.Google Scholar

189. Monteil, V., “Al-Bakri,” 70.Google Scholar See also Levtzion, N., “Was Royal Succession in Ancient Ghana Matrilineal?” UARS, 5 (1972), 91–93.Google Scholar

190. He flees to Kingui (Tellier, G., Autour de Kita, [Paris, 1898], 208–09Google Scholar); Sama (Frobenius, , Spielmannsgeschichten, 69Google Scholar; Monteil, , “Légende,” 380Google Scholar; Arnaud, , “Islam,” 159Google Scholar); and to his uncle at Nyamina (Sissoko).

191. Cuoq, J.M. ed. and tr., Recueil des sources arabes concernant l'Afrique occidentale du VIIIe au XVIe siècle (bilad al-sudan), (Paris, 1975), 133.Google Scholar Al-Harrānī (ca. 1332) also mentioned that the town of Ghana was on a river (Cuoq, , Recueil, 249–50).Google Scholar

192. Frobenius, , Spielmannsgeschichten, 53.Google Scholar

193. Adam, , “Légendes,” 85Google Scholar; Arnaud, , “Islam,” 147Google Scholar; Delafosse, , “Traditions,” 295Google Scholar; Monteil, , “Légende,” 373.Google Scholar

194. Lanrezac, “Au Soudan,” 385n.1. The town is not named, though it could have been Diara, which lay just north of the present Nioro. The legend of Gassiré's lute mentions “Dierra” as the name of one of the four Wagadus, Frobenius, , Spielmannsgeschichten, 53.Google Scholar

195. Tarikh el-Fettach, 42/77.

196. Frobenius, , Spielmannsgeschichten, 53, 60.Google Scholar

197. Delafosse, , “Traditions,” 293–94Google Scholar; Monteil, , “Légende,” 369, 371.Google Scholar See also a new translation of the Monteil version by Bathily, A., “La légende du Wagadu (texte soninké de Malamine Tandyan),” BIFAN, 29B(1967), 135Google Scholar, as well as Arnaud, , “Islam,” 145Google Scholar; Monteil, , “Légende,” 369Google Scholar; Adam, , “Légendes,” 83.Google Scholar

198. Arnaud, , “Islam,” 145.Google Scholar

199. Depending on the version, the band of followers is composed of several hundred slaves, sorcerers, indigenous priests, marabouts, or warriors.

200. Two versions (Arnaud, , “Islam,” 146Google Scholar; Monteil, , “Légende,” 371Google Scholar) feature a slave of the jinn, who issues the initial challenge at the well. The slave's name is Tori Gabé Séné Ouali (Arnaud) or Teri-gabe Senewali (Monteil), which translates as “Longhaired Senewali.” Monteil laboriously transforms “Senewali” into “Sanhaja” via “Senegani” and “Senaga,” suggesting that it shows that the Sanhaja “sont antérieurs aux Soninké.” He ignores the simpler alternative that “Senewali” is very close to “Sarawule,” one of the many names by which the Soninké are known (see Westermann, and Bryan, , Languages of West Africa II, 32Google Scholar). Hence the “slave” could, in the original event, have been one or several of the local longhaired Soninké people defending their water supply against shaven-headed Muslims.

201. Arnaud, , “Islam,” 157–58.Google Scholar

202. The encounter at a well between newcomers and local people or spirits appears to be a standard motif in sudanic oral tradition, and beyond the simple statement that someone arrived from elsewhere, can have no historical value.

203. In contrast, other episodes portray Dinga as the standard of pre-Islamic Soninke culture. Implying that those who boast direct descent from Dinga are resistant to Islam, Tudo Yaresi says they were honored when greeted by the name “Sukhuna” or “Sorona” and that they claim to have inherited their ancestor's rainmaking powers. But those who have been converted to Islam prefer to be called “Cissé” after the later ancestor Diabe, who is said to have been of the earliest possible western sudanic Muslim stock (Monteil, , “Légende,” 398Google Scholar). Cf. Wâra Dyâbe ruler of the Takrur animists (1040) who converted to Islam and introduced Muslim law to his people (Monteil, , “Al-Bakri,” 68Google Scholar).

204. Arnaud, , “Islam,” 157.Google Scholar The term diammou (jamu) can be variously defined as “identity,” “family name,” or “patronymic group.”

205. Monteil, , “Al-Bakri,” 70.Google Scholar

206. Monteil, , “Légende,” 369, 398.Google Scholar

207. Arnaud, , “Islam,” 150–51Google Scholar; Frobenius, , Spielmannsgeschichten, 62Google Scholar; Delafosse, , “Traditions,” 295Google Scholar; Adam, , “Légendes,” 87Google Scholar; Lanrezac, , “Au Soudan,” 383Google Scholar; Monteil, , “Légende,” 376.Google Scholar

208. Adam, , “Légendes,” 85Google Scholar; Arnaud, , “Islam,” 147Google Scholar; Delafosse, , “Traditions,” 295Google Scholar; Lanrezac, , “Au Soudan,” 382Google Scholar; Monteil, , “Légende,” 373.Google Scholar When specifically identified, the snake is a python, but according to Delafosse, (La langue Mandingue II, 54)Google Scholar, “bida” means “big black snake”; in Monteil's version (374) it has cobra-like characteristics. Sissoko identifies bida as a lethal black snake known in Bambara as “n 'koroko.“

209. Lanrezac, , “Au Soudan,” 384Google Scholar; Adam, , “Légendes,” 89Google Scholar; Delafosse, , “Traditions,” 296.Google Scholar The Arnaud (147) and Monteil (374) sources agree that two “Bidas” were born: “Bida-sere-kalla” (man-killing blacksnake) and “Bida-sere-toka” (harmless blacksnake), and the Arnaud version is alone in identifying the two as men, translating the names as “Blacksnake killer” and “Blacksnake respecter.” The motif of the benevolent snake with human parents appears elsewhere in western sudanic tradition, e.g., Tautain, , “Légende et traditions,” 475Google Scholar; de Zeltner, Frantz, Contes du Sénégal et du niger (Paris, 1913), 135–36Google Scholar; Traoré, D., “Folklore Soudanais,” Outre-Mer (1932), 113–15Google Scholar; Mamary Kouyate (a Bambara griot), “Chanamba,” recorded September, 1975, at Kolokani, Mali.

210. Tautain, , “Légende et traditions,” 473Google Scholar; Tautain claimed he heard this tradition some fifty times in many places from informants of various Soninké families, though he did not name them. For diluted or unusually distorted versions of the Wagadu snake legend collected far from the center of Soninke influence see Bérenger-Féraud, L.J.B., Les peuplades de la Sénégambie, (Paris, 1879), 169–72Google Scholar, and Hervé, H., “Recueil de contes du pays Malinké,” L'Education Africaine, 93/98(1936/1937), 50–62.Google Scholar

211. Tellier, , Autour de Kita, 206–09.Google Scholar In a fragment collected by Penel, J. in Coûtumes soudanaises: Malinké, Sarakolé, Kkassonké, (2 vols.: Paris, 1895)Google Scholar, Bibliothèque Nationale MS°4 Li³³. 27, the snake is “le génie protecteur” of Wagadu.

212. Adam, , “Légendes,” 94.Google Scholar

213. Monteil, , “Légende,” 376–79.Google Scholar

214. Frobenius, , Spielmannsgeschichten, 68.Google Scholar

215. Delafosse, , “Traditions,” 297.Google Scholar This translation of the original Arabic text includes the statement that “les hommes se mangèrent entre eux.”

216. Arnaud, , “Islam,” 156–57.Google Scholar

217. Pâques, V., “L'estrade royale des Niaré,” BIFAN, 15 (1953), 1644Google Scholar; Levtzion, , Ancient Ghana, 18Google Scholar, notes the prominence of this belief.

218. See also Meillassoux, , Légende de Za dispersion, 8.Google Scholar According to Sissoko this is the origin of the name “Diara” (Jara), which is one of the most distinguished Bambara patronymics. This is a popular etymology, the kind of explanation for something unknown that griots enjoy devising, often through the use of homonyms. A completely different etymology for the same name using the Bambara word jara (“lion”) was included in another tradition collected in Segou, 1976. For more on this subject see Vansina, , Oval Tradition, 44Google Scholar, and Zemp, H., “La légende des griots Malinké,” Cahiers d'Études Africaines, 6 (1966), 630.CrossRefGoogle Scholar

219. There is a variation in the 1964 version collected by Meillassoux, , “Histoire,” 189–90Google Scholar, in which the consequence of the python's death is sterility throughout the land.

220. Minton, S.A. Jr. and Minton, M.R., Giant Reptiles, (New York, 1973), 197.Google Scholar

221. Tautain, , “Légende et traditions,” 474–75.Google Scholar

222. Tauxier, Louis, La Religion Bambara, (Paris, 1927), 157n1Google Scholar; the tradition itself is in Tauxier, , Le Noir du Yatenga, (Paris, 1917), 496–97Google Scholar, a legend of the Nioniossé, which is what the Mossi call the Kurumba.

223. Pitman, C.R.S., A Guide to the Snakes of Uganda, (London, 1974), 69.Google Scholar Remarking on a possible interchange of mythical elements between the populations of ancient Ghana and their Saharan neighbors, H.T. Norris has pointed out that in both regions, jinn as well as serpents are considered to be denizens of water-points, caves, and the like, and that they are often confused with one another as protective entities (Saharan Myth, 8–10).

224. Tautain, , “Légende et traditions,” 475.Google Scholar

225. Brun, P.J., “Le totémisme chez quelques peuples du Soudan Occidentale,” Anthropos 5(1910), 857–58.Google Scholar The same motif occurs in Zeltner, , Contes, 135–36Google Scholar, and in Mamary Kouyaté's “Chanamba.”

226. Cansdale, G.S., West African Snakes (London, 1961), 20.Google Scholar

227. Tautain, , Religion Bambara, 156–57Google Scholar

228. Tautain, , “Légende et traditions,” 474.Google Scholar

229. Dieterlen, , Essai, 145.Google ScholarHenry, F. Joseph, L'âme d'un peuple africain: les Bambara (Münster, 1910), 76Google Scholar, declared that any idea of the snake being a Bambara divinity seemed to him a piece of stupidity.

230. Arcin, A., La Guinée Française (Paris, 1907), 402Google Scholar; Dieterlen, , “The Mande Creation Myth,” 135, 135n.2.Google Scholar

231. Monteil, , “Légende,” 398Google Scholar; Saint-Père, , Sarakollé, 12.Google Scholar

232. The motif of fate determined by magical hairs from an animal also appears in a Songai tradition in Dupuis-Yakouba, M.A., Les Gow ou Chasseurs du Niger, Légendes Songai de la Région de Tombouctou (Paris, 1911), 75–87.Google Scholar Hair as a symbol of kingship occurs elsewhere in Manding tradition, notably in the legend of Sunjata, where in one version the hero is born covered with hair from head to foot (Diabété, M.M., Kala Jata [Bamako, 1970], 30Google Scholar). Compare the Gambian Mandinka expression mansaya tio le be a bata (“kingship hair is on his body”), meaning the person has the look of a king, or of someone who will one day be a king (Innes, , Sunjata, 105Google Scholar).

233. Arnaud, , “Islam,” 156.Google ScholarAl-Bakri, (Monteil, V., “Al-Bakri,” 72)Google Scholar gives more details about pre-Islamic Soninké religion. Compare Dinga's possession of “idols” (probably from the Manding boli, a kind of indigenous altar), in Monteil, , “Légende,” 369.Google Scholar

234. Adam, , “Légendes,” 92.Google Scholar

235. Cuoq, , Recueil, 137Google Scholar; Levtzion, , Ancient Ghana, 17.Google Scholar

236. See Palmer, H.R., Sudanese Memoirs (3 vols.: London, 1928), 3: 133–34.Google Scholar

237. Lanrezac, , “Au Soudan,” 384–85Google Scholar; though he does not specifically identify the snake-slaying episode, Lanrezac recounts all the usual events leading up to it, before abruptly terminating the narrative with the comment that the “légende historique” had got mixed up with a much older one about the end of human sacrifice.

238. This aspect of the ritual could be derived from historical events. Dieterlen states that in former times the Bambara performed human sacrifice during periods of extreme crisis (Essai, 94), and according to a 1909 colonial report Soninke and Bozo who were constructing the “new” town of Jenne in the first half of the thirteenth century and who desired for it the protection of the jinn of the river, sacrificed a young Bozo girl named Tifama Diénapo by immuring her in the wall (“Resumé historique de Djenné” (1909), Archives Nationales du Mali, Koulouba, ID–38–2). C. Monteil heard the same tradition, identifying the girl as Tapama, daughter of a Bozo named Kayantao, Modi (Une Cité Soudanaise: Djénné [Paris, 1932], 36).Google Scholar The two sources agree that the sacrificial site was near the Kanafa gate, and in Monteil's time citizens still guarded that section of wall against mishap. The Frobenius version implies that in earlier times, not one but ten maidens were sacrificed to the snake and that the ritual had subsequently been altered to a less severe form (Spielmannsgeschichten, 64–65).

239. Pâques, , “L'estrade royale,” 1644.Google Scholar In ancient times their patronymic was Niakaté, but it was changed to Niaré at some stage of their ancestors' long migration from Kumbl to Bamako.

240. Ibid, 1645.

241. Ibid, 1643–44.

242. Levtzion, , Ancient Ghana, 115–16.Google Scholar

243. Delafosse, , “Traditions,” 296Google Scholar; Adam, , “Légendes, 91Google Scholar; Monteil, , “Légende,” 378Google Scholar; Frobenius, , Spielmannsgeschichten, 64Google Scholar; Arnaud, , “Islam,” 155Google Scholar; Adam, , Légendes,” 94Google Scholar; Tellier, , Autour de Kita, 208–09Google Scholar; Delafosse, , “Traditions,” 297Google Scholar; Monteil, , “Légende,” 380.Google Scholar Levtzion describes this as an echo of the fact that the shift of the source of gold from Bambuk to Buré in the eleventh or twelfth century coincided with the decline of Ghana (Ancient Ghana, 155).

244. Delafosse, , “Traditions,” 296.Google Scholar For further discussion of the gold in Wagadu tradition see Farias, , “Great states,” 485–86.Google Scholar

245. Frobenius, , Spielmannsgeschichten, 65.Google Scholar

246. Delafosse, , “Traditions,” 296.Google Scholar

247. Monteil, , “Légende,” 378.Google Scholar

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

249. Al-Idrisi, , Kitāb nuzhat al-mushtāq fī ikhtiraq al-āfāq, eds. Dozy, R. and de Goeje, M.J., (Leiden, 1866), 9.Google Scholar

250. Tarikh es-Soudan, 18–19; Tarikh el-Fettach, 75.

251. Fūdī, ʿUthm¯n ibn, Bayān wujūb al-hijra ʿala'l–ʿibād, ed. and tr. El Masri, F.H., (Khartoum, 1978), 51.Google Scholar

252. This note is added to clear up a confusion which has crept into notes 20 and 155 above. In 1911, the Comité de l'Afrique française published a separate volume, a tirage à part, by R. Arnaud, entitled “L'Islam et la politique musulmane française en Afrique Occidentale Française, suivi de la singulière légende des Soninké,” as cited in note 155, and appearing thereafter as “Arnaud, “Islam.” In 1912, the Bulletin du Comité de l'Afrique française reprinted the first part of Arnaud's work, “L'Islam et la politique musulmane française en A.O.F.,” in its supplement, Renseignements coloniaux, but not the “Singulière légende.” In note 20, we wrongly locate “Singulière légende” in the Bulletin. The tirage à part is, alas, very rare. Even the British Museum does not possess a copy, and only one of the authors of this paper has ever seen it. Arnaud's informant for the legend was a Muslim of the Batchili lineage.