¹ Bovill, E. W., Golden Trade of the Moors (London, 1958), 191.Google Scholar

² Levtzion is virtually alone in indicating his source (Mauny) for the possible identification of Wangara as Bambuk; see Levtzion, N., Ancient Ghana and Mali (London, 1973), 155.Google Scholar

³ Delafosse, M., Haut-Sénégal-Niger (Paris, 1912), i, 55Google Scholar, translation from the French by SKM. Trimingham, J. S., A History of Islam in West Africa (Oxford, 1962), 54Google Scholar mentions that the eastern part of the region between the Faleme and Bakhoy rivers is still called Gangaran, but cites no authority. The same is true for de la Roncière, C., La Découverte de l'Afrique au Moyen Age; Mémoire of the Royal Society of Geographers of Egypt. No. 5 (Cairo, 1924), i, 98Google Scholar; Bovill, E., Golden Trade, 94Google Scholar; and Lewicki, T., Etudes Maghrebines et Soudanaises (Warsaw, 1976), 41–2.Google Scholar

⁴ Delafosse, , Haut-Sénégal-Niger, ii, 45Google Scholar, translated from the French by SKM. Other authors who make this same point include Crone, G. R., The Voyages of Cadamosto (London, 1937), xvGoogle Scholar; Monteil, C., ‘Les Gana des géographes arabes et des Européns’, Hespèris, xxxviii (1951), 451Google Scholar; and Bovill, E., Golden Trade, 94.Google Scholar In addition, Lombard, M., Les Métaux dans l'Ancien Monde du V^e au XI^e Siècle (The Hague, 1974), 210Google Scholar, and Cuoq, J. M., Recueil des Sources Arabes concernant l'Afrique Occidentale du VIII^e au XVI^e Siècle (Paris, 1975), 135Google Scholar, note 2, state that Wangara refers to the regions between the Faleme and Bafing and the Faleme and upper Niger, respectively, but neither author offers any arguments as evidence nor cites an authority for this identification.

⁵ As a consequence, Delafosse and other French authors have tended to ignore the possible role of the Lobi mines of Upper Volta in early gold traffic. French commercial exploitation of Lobi gold was severely hindered by the inaccessibility of the deposits and the refractory nature of the indigenes. Meniaud, J., Haut-Sénégal-Niger (Paris, 1912), ii, 195–9.Google Scholar

⁶ Mauny, R., Tableau GÉOgraphique De L'Ouest Africain D'AprÈS Les Sources Ecrites, La Tradition, Et L'ArchÉOlogie. MÉMoire De L'Institut Fondamental D'Afrique Noire. No. 61 (Dakar, 1961), 300.Google Scholar

⁷ Mauny, , Tableau Géographique, 302.Google Scholar

⁸ Marquart, J., Die Benin-Sammlung des Reichmuseums für Völkerkunde in Leiden (Leiden, 1913)Google Scholar, separate map; Miller, K., Mappa Arabicae: Arabische Welt und Länder-karten des 9–13 Jahrhunderts (Stuttgart, 1926); ii, 164Google Scholar; Taylor, E. G. R., ‘Pactolus: River of Gold’, Scottish Geographical Magazine, xliv (1928), 131Google Scholar; Kimble, G. H. F., Geography in the Middle Ages (London, 1938), 59, 195.Google Scholar

⁹ Bovill, for example, begins his discussion of Wangara with the statement: ‘our search for Wangara can be confined to the four gold bearing districts…Bambuk, Bure, Lobi, and Ashanti’, Bovill, , Golden Trade, 192.Google Scholar

¹⁰ The earliest historical mention of trade by boat in the Inland Delta and of the entrepot of Jenne is contained in a letter written in 1447 by Antonio Malfantes, a Genoan merchant. Historians have maintained that the emergence of Jenne as a major trading centre and of the Inland Niger Delta as an important transport axis in the fourteenth century was dependent on several factors: the replacement of Walata in the fourteenth century by Timbuktu as the southern terminus for trans-Saharan caravans; the opening of the Akan gold mines, for whose trade Jenne was well-situated; and the consolidation of the Niger waterways under the protection of the Empire of Levtzion, Mali., Ancient Ghana and Mali, 156–7Google Scholar; Mauny, , Tableau Géographique, 500Google Scholar; Monteil, C., Une Cité Soudanaise: Djénné (Paris, 1932), 39Google Scholar; Tymowski, M., ‘Le Niger, voie de communication des grands états du soudan occidentale jusqu'au fin du XVI^e siècle’, Africana Bulletin, xii (1967), 81–2Google Scholar; Devisse, J., ‘Routes de commerce et échanges en Afrique occidentale en relation avec la Méditerranée’, Revue d'Histoire Economique et Sociale, L (1972), 366Google Scholar; Herbert, E., ‘Timbuktu: a case study of the role of legend in history’, in Swartz, B. K. and Dumett, R. (eds.), West African Culture Dynamics (The Hague, 1980), 433.Google Scholar

¹¹ al-Fazari, , in Cuoq, Recueil, 42.Google Scholar

¹² Lyon, G. F., A Narrative of Travels in Northern Africa in the Years 1818, 19 and 20 (London, 1821), 146.Google Scholar

¹³ In Cuoq, Recueil, 52.

¹⁴ Herbert, , ‘Timbuktu’, 443–4Google Scholar

¹⁵ These results were discussed in McIntosh, R. J. and McIntosh, S. K., ‘The Inland Niger Delta before the Empire of Mali: evidence from Jenne-jeno’, J. Afr. Hist, xxi, i (1980).Google Scholar Full details of the excavations and archaeological survey are presented in McIntosh, S. K. and McIntosh, R. J., Prehistoric Investigations in the Region of Jenne, Mali. Cambridge Monographs in African Archaeology, No. 2 (Oxford, 1980).Google Scholar

¹⁶ In Cuoq, Recueil, 62. Cuoq comments in a footnote to this passage that the Niger inundates a large area of floodplain annually, at which time the highest land remains above water. Cuoq must be referring to the Inland Niger Delta here, since at no point upstream of Segou is the gradient small enough to create an extensive floodplain.

¹⁷ In Cuoq, Recueil, 134–5. Translated from the French by SKM.

¹⁸ The floodwaters of the Niger and Bani rivers, rising in August, annually inundate an area of 20,000 km² in the Inland Niger Delta.

¹⁹ Meniaud, , Haut-Sénégal-Niger, 169, 187–8Google Scholar; Michel, P., Les Bassins des Fleuves Sénégal et Gambie: Etude Géomorphologique, Mémoire de l'ORSTOM, No. 63 (Paris, 1973), 131–2, 656.Google Scholar

²⁰ Crone, Voyages, xv, Bovill, , Golden Trade, 194.Google Scholar

²¹ Brasseur, G., Les Etablissements Humains au Mali, Mémoire de l'Institut Fondamental d'Afrique Noire, No. 83 (Dakar, 1968), 431Google Scholar; Gallais, J., Le Delta Intérieur du Niger, Mémoire de L'Institut Fondamental d'Afrique Noire, No. 79 (Dakar, 1967) ii, 479Google Scholar; Monteil, C., Monographie de Djénné (Tulle, 1903), 14–15. 26–7, 240.Google Scholar

²² Fernandes, V., Description de la Côote d'Afrique de Ceuta au Sénégal, 1506–7, (translated de Cenival, P. and Monod, T.) (Paris, 1938), 85–7Google Scholar; Crone, , Voyages, 22Google Scholar; al-Sa'di, , Ta'rikh es-Sudan (translated Houdas, O.) (Paris, 1900), 22.Google Scholar

²³ Barth, H.(Travels and Discoveries in North and Central Africa (New York, 1857), iii, 366Google Scholar) confirms that caravans from the north usually arrive at Timbuktu in November and December, allowing traders to complete their transactions and be en route north again by February.

²⁴ As La Roncière (Découverte, I, 19) has rightly pointed out, no greater homage to Arab science was possible than that the Christian prince, Roger ii, should choose al-Idrisi to direct research on the geography of the world.

²⁵ Ibid., i, 24.

²⁶ From al-Bakri's account in 1068, it is clear that Tirka (Tirekka) is situated on the Niger Bend, Cuoq, Recueil, 105.

²⁷ Ibid., 135.

²⁸ Ibid., 136–7.

²⁹ La Roncière, , Recueil, i, 23.Google Scholar

³⁰ Cuoq, Recueil, 205–7. Translated from the French by SKM.

³¹ Map reproduced Ibid., 2.

³² Ibid., 206. It should be noted that the reference to Wangara as an ethnic group is clear in both Ibn Sa'id and al-Idrisi.

³³ Taylor, , ‘Pactolus’, 134Google Scholar; A funduq comprised residences and warehouses of foreign traders within a walled enclosure. Fisher, H. J., ‘The eastern Maghreb and the central Sudan’, Cambridge History of Africa (Cambridge, 1977), iii, 254.Google Scholar

³⁴ La Roncière, , Découverte, i, 13.Google Scholar

³⁵ Carignano's planisphere is reproduced in Kamal, Youssouf, Monumenta Cartogra-phica Africae et Aegypti (Cairo and Leyden, 1926–1951), IV, i, 1139.Google Scholar

³⁶ La Roncière, , Découverte, i, 23–58.Google Scholar

³⁷ Beazley, C. R., The Dawn of Modern Geography (Oxford, 1906), iii, 527.Google Scholar

³⁸ La Roncière, , Découverte, i, 128.Google Scholar

³⁹ Atlas reproduced in Kamal, , Monumenta, iv, iii, 1303.Google Scholar

⁴⁰ La Roncière, , Découverte, i, 118Google Scholar; Taylor, , ‘Pactolus’, 142.Google Scholar

⁴¹ Markham, C. (ed.), Book of Knowledge (Hakluyt Soc, 1912), 34.Google Scholar

⁴² ‘Pallole’ in Catalan means ‘gold dust’.

⁴³ Viladestes's map is reproduced in colour as the frontispiece to La Roncière, , Découverte, iGoogle Scholar; Fra Mauro is reproduced Ibid., I, 128–9.

⁴⁴ Perinbaum, B. M., ‘ Notes on Dyula Origins and nomenclature’, Bulletin de l'Institut Fondamental d'Afrique Noire (B), xxxvi (1974), 676–89.Google ScholarCurtin, P. D., Economic Change in Precolonial Africa (Wisconsin, 1975), 67–9.Google Scholar

⁴⁵ Perinbaum, , ‘Notes’, 688.Google Scholar

⁴⁶ Delafosse, , Haut-Sénégal-Niger, i, 124.Google Scholar

⁴⁷ Hunwick, J. O., ‘The mid-fourteenth century capital of Mali’, J. Afr. Hist., xiv (1973). 195–208.CrossRefGoogle Scholar

⁴⁸ Fernandes, , Description, 85–7Google Scholar. Translation from the French by SKM.

⁴⁹ See Levtzion, , Ancient, 164Google Scholar, and Monteil, C. ‘Les empires du Mali’, Bulletin du Comité d'Etudes Historiques et Scientifiques de l'Afrique Occidentale Française xii (1929)Google Scholar (reprinted by Maisonneuve and Larose in 1968), 44.

⁵⁰ Delafosse, , Haut-Sénégal-Niger, 1, 255–63Google Scholar; Monteil, , Cité, 31.Google Scholar

⁵¹ There is considerable evidence from the Late Stone Age sites at Tichitt to indicate that increasing desiccation was a factor in the abandonment of agricultural villages in the region by c. 400 B.C. It has also been thought that an additional pressure was the appearance of horse-riding Libyco-Berbers from North Africa, see Munson, P. J., ‘The origins of cultivation in the southwestern Sahara’, in Harlan, J. R, de Wet, J. M. J., and Stemler, A. B. L. (eds.), The Origins of African Plant Domestication (The Hague, 1976), 195Google Scholar; an analogous spread south of Saharan groups down the Tilemsi Valley is described in Smith, A. B., ‘ Biogeographical considerations of the colonization of the lower Tilemsi valley in the second millennium B.C.’ Journal of Arid Environments, 2 (1979), 358–60.Google Scholar

⁵² The cultural material from Jenne-jeno is described in detail in McIntosh and McIntosh, Prehistoric Investigations.

⁵³ Monteil, , Cité, 32.Google Scholar

⁵⁴ Delafosse, , Haut-Sénégal-Niger, i, 252–6Google Scholar; Levtzion, N., ‘The Western Maghrib and Sudan’, Cambridge History of Africa (Cambridge, 1977), iii, 275.Google Scholar

⁵⁵ Cuoq, Recueil, 105.

⁵⁶ Ibid., 106. Safanku may be situated in the vicinity of modern Lere, the westernmost point north of Lake Debo that is joined to the Niger Bend by a major canal. We might look for Bughrat in the area of modern Goundam. Barth outlined a route from Sansanding to Timbuktu which passed by Bassikounou, Lere, and Goundam, , Travels, iii, 694.Google Scholar

⁵⁷ In a footnote to the relevant passage in al-Bakri, Cuoq comments that the desert Arabs of this region still refer to Soninke as Gangara, , Cuoq, Recueil, 86.Google Scholar

⁵⁸ Delafosse, , Haut-Sénégal-Niger, i, 260–2.Google Scholar

⁵⁹ Trimingham, , History, 45.Google Scholar

⁶⁰ Cuoq, Recueil, 102.

⁶¹ E.g. by way of Ghana, Tirekka, Gao, or possibly all three?

⁶² It is worth recalling Meniaud's assessment of the Lobi gold deposits: ‘ Le Lobi est peut-être la région du Haut-Sénégal-Niger où l'or est le plus abondant, le plus facile à exploiter, et, en tout cas, le plus beau’, Meniaud, , Haut-Sénégal-Niger, ii, 193.Google Scholar Surely detailed archaeological investigation is warranted at the ancient gold mining sites in Upper Volta, e.g. those reported by Savonnet, G., ‘Habitations souterraines bobo ou anciens puits de mines en pays wilé?’, Bulletin de l'Institut d'Afrique Noire (1974), xxxvi, 228–245.Google Scholar