1. Winks, Robin W., ed., The Historian as Detective: Essays on evidence (New York, 1968), xii, xiv.Google Scholar

2. For the most recent discussion of al-Bakri's geography see Hunwick, J.O., Meillassoux, Claude, and Triaud, Jean-Louis, “La géographie du Soudan d'après al-Bakri: trois lectures” in Le sol, la parole et l'écrit: mélanges en hommage à Raymond Mauny (Paris, 1981), 401–28Google Scholar; on Wangaran gold see Mcintosh, Susan Keech, “A Reconsideration of Wangara/Palolus, Island of Gold,” JAH, 2 (1981): 145–58.CrossRefGoogle Scholar

3. For example see McDougall, E. Ann, “Salts of the Western Sahara: Myths, Mysteries and Historical Significance,” IJAHS, 23 (1990): 231–57Google Scholar; Herbert, Eugenia W., Red Gold of Africa. Copper in Precolonial History and Culture (Madison, 1984) esp. chs. 1 and 5Google Scholar; and on ‘peoples and places’ more generally in the Sahara, H.T. Norris, Saharan Myth and Saga (London, 1972).Google Scholar

4. This is not to underestimate the contributions of other sources, especially archeological finds, that have been critical to the ongoing revisionist views of the Arabic accounts. Roderick and Susan Mcintosh have perhaps been the most outspoken about the problems raised for interpreting early African history as a consequence of so many years of allowing inquiry to be directed by their contents. Susan Keech Mcintosh and Roderick J. Mcintosh, “From Stone to Metal: New Perspectives on the Later Prehistory of West Africa,” (forthcoming), and “From Siècles Obscurs to Revolutionary Centuries on the Middle Niger,” World Archaeology, 20 (1988), 141–65.CrossRefGoogle Scholar The work of Fisher and Conrad on the now infamous “conquest that never was” was undertaken from a different critical perspective but their findings are a superb example of precisely this difficulty. (Fisher, Humphrey and Conrad, David C., “The Conquest That Never Was: Ghana and the Almoravids, 1076, I: The External Arabic Sources,” HA, 9 (1982), 21–59.Google Scholar Similarly, my own critical work on the “conflict theory” of pastoral-sedentary relations draws on anthropological analyses to derive a new perspective on medieval Arab predjudice. McDougall, E. Ann, “The Sahara Reconsidered: Pastoralism, Politics and Salt From the Ninth Through the Twelfth Centuries,” African Economic History, 12 (1983), 263–86.CrossRefGoogle Scholar Nevertheless, it is the continued dialogue among the Arabic accounts and other useful sources and resources, as orchestrated by historians, which will ultimately reveal whatever can be uncovered of historical reality prior to the fifteenth century.

5. This view was given its most invidious expression in the well-known statement by Trevor-Roper that prior to the arrival of the Europeans, Africa had no history! (“Perhaps in the future there will be some African history to teach. But at present there is none: there is only the history of Europeans in Africa. The rest is darkness…and darkness is not a subject of history…” The Rise of Christian Europe [New York, 1965].Google Scholar)

6. Stone, Thora G., “The Journey of Cornelius Hodges in Senegambia, 1689— 90,” English Historical Review, 39 (1924), 89.CrossRefGoogle Scholar The letter is drawn from the records of the Royal African Company, Letter Book 17 and is dated James Island (Gambia), 16 Sept. 1690. In fact, there are several instances of placenames which are simply described as “doubtful” or “unknown” in the notes.

7. Stone, , “Cornelius Hodges,” 93.Google Scholar Hodges placed himself at “Yafarra” in “Bamboo,” near the gold mines of “Nettico” (91). Philip Curtin has located Nettico on the edge of the goldfield surrounding modern Sanjola (Map 1) in the region of Bambuk (Bambuhu). Curtin, , Economic Change in Precolonial Africa: the Senegambia in the era of the Slave Trade (2 vols.: Madison, 1975), 51Google Scholar and maps pp. 52, 200-201.

8. Stone, , “Cornelius Hodges,” 93.Google Scholar

9. I have discussed this trade in “Salt, Slaves and the Saharan trade: Nineteenth-Century Developments,” Slavery and Abolition, (forthcoming).

10. Stone, , “Cornelius Hodges,” 93.Google Scholar There are several discrepancies in spelling here, e.g., verry-very; Emperour-Emperor; Countrey-countrey; I would suspect that the variations arise from Stone's rendering of the document rather than Hodges' original but here can only reproduce the published version, complete with flaws.

11. Ibid., 92.

12. Ibid., 93.

13. Ibid., 94.

14. Timbuktu was destined to experience a certain economic revival in the eighteenth and nineteenth centuries based principally on its ecozonal position. It seems never to have recovered the international fame and fortune of its days as a medieval caravan terminus.

15. This was in the context of asking after the fate of Major Houghton, who had perished in an attempt to travel through “Moorish country” to Tishit some years earlier. He was told that the “Moors” hired to convey him there had betrayed and robbed him some two days' journey out of Jara. He had attempted to return and got as far as the watering-place “Tarra” before perishing in unknown fashion. Park, MungoTravels in Africa, ed. and intro. Miller, Ronald (London, 1969), 78.Google Scholar Paul Marty has attributed the death of Houghton to the Mbarek, Awlad chief, Ali wuld Umar: Etudes sur l'Islam et les tribus du Soudan, III, Les tribus maures du Sahel et du Hodh (Paris, 1921), 385.Google Scholar

16. This same consideration tends to preclude an identification with the now-abandoned Taranni, discussed more fully in note 64 below. In fact, Mungo Park's ‘Tarra” is more likely to have referred to the generic Saharan expression tarha from the Arabic term meaning place where travelers halt (for water, for example); see note 35 below.

17. Curtin, Economic Change, 51n5. He acknowledged the possibility that “Tarra” might have been the watering-hole identified by Park, or even the much more modern “Nara” located further east. However, his analysis then proceeded on the basis of a Jara/“Tarra” identification.

18. Park, , Travels, 84.Google Scholar

19. Boyer, G., Un peuple de l'ouest soudanais: les Diawara, (Dakar, 1953), 34.Google Scholar

20. Ly-Tall, Madina, Contribution à l'histoire de l'empire du Mali (XIIIe-XVIe siècles) (Dakar, 1977), 74.Google Scholar

21. Boyer does not give an exact date for the siege, but it clearly followed the Moroccan ‘conquest’ of Songhay, ca. 1590/91, and may well date to sometime during the next decade when the new government faced continual rebellions from outlying provinces like Jawara (Boyer, , Diawara, 34–36Google Scholar). In this case Jara had given refuge to opponents of the Arma government. See also the most recent work on the Jawara state by Diawara, Mamadou, La graine de la parole (Wiesbaden, 1990), 28–30.Google Scholar He draws critically on Boyer's earlier work, using other secondary sources but most especially, recently-collected oral history, to enlarge our understanding of Jawara's past.

22. Each ‘side’ sought to augment its power by contracting new alliances among such strange bedfellows as Arma and Peul from Bakunu, Soninke of Wagadugu, Bambara from Kaarta, and Saharan “Moors” of the Awlad Mbarek.

23. Boyer, , Diawara, 36–38.Google Scholar

24. Curtin, , Economic Change, 51.Google Scholar

25. He does, on the other hand, note that a small village not far from Jara was walled (Travels, 86).

26. Diawara, , Graine, 29.Google Scholar

27. Stone, , “Cornelius Hodges,” 92, 93.Google Scholar

28. Park, , Travels, 72.Google Scholar

29. Ibid., 79.

30. Ibid., 72, 84. “Ludamar” was the name he gave to the Awlad Mbarek territory centered in what we know as the Mauritanian Hodh.

31. This might be considered a debatable point, however. Diawara, for example, accepts Louis Tauxier's dates (1672 to 1754) for the existence of Ludamar and its domination over the Kinigi region, after which time it spread its authority over Kaarta as well (Graine, 29). However, other sources looking at these developments from the Mauritanian perspective date the crucial movements to the eighteenth century. Marty, for example, wrote that it was during the reign of the son of the Awlad Mbarek chief Umar (1755-1762)—that is, Ali wuld Umar whose long reign lasted from 1762 to 1808—that a ‘royal house’ became recognized. Houghton and Park attached this family name “wuld Umar” to the Awlad Mbarek kingdom itself, rendering it “Ludamar” in transliteration (Tribus, 384). Most recently, Abdel Wedoud ould Cheikh, drawing largely on hitherto unused Mauritanian manuscripts, confirms this dating. The Awlad Mbarek fought alongside the Ida Aish against the Awlad Dayman of the Trarza in 1695. At that time they were still based in the Tagant, and some traditions hold that the Ida Aish were in a position of vassalage. During the course of the following century, the Awlad Mbarek moved in the direction of “the rich pastures of the Hodh the attraction of the agriculture of Bakunu and to Kingi, the crossroads of the salt-gold exchange.” This movement from their political center in the Tagant favoured the gradual emancipation of the Ida Aish; the eighteenth century saw the establishment of states (‘emirates’) in both the Tagant, and Hodh, Eléments d'histoire de la Mauritanie, (Nouakchott, 1988), 110.Google Scholar

32. On Jawara's origins, relations with Saharans and with the great ‘empires’ of the middle Niger, see Boyer, , Diawara, 21–37.Google Scholar

33. Ibid., 84.

34. It was Farias, Paulo de Moraes who first pointed this out in “The Almoravids: Some Questions Concerning the Character of the Movement During its Period of Closest Conduct With the Western Sudan,” BIFAN, 29 (1967), 829–30Google Scholar; see also my discussion in “Salts of the Western Sahara,” 241-42. Other examples include ‘Adrar,’ a proper name appearing in several places in the Sahara which means, in Berber, “mountain”; or “tagant,” which has become a recognized proper name, but originally referred only to “forest,” as is discussed below.

35. See Ahmed Lamine ech Chenguiti, el-Wasît, (Saint-Louis, 1953) 144, for the meaning and origin of tarha.

36. Witness the rendering of ‘wuld Umar’ as “Ludamar” (notes 30, 31, above). Presumably the logic underlying Curtin's assumption of the “Tarra”/Jara identification is that of sound similarity, as well as the throw away, but unpursued, suggestion that “Tarra” could “even have been Nara (further east);” Economic Change, 51n5.

37. See note 39 below.

38. H. J. Norris, “The History of Shinqît, According to Idaw Ali Tradition,” Bulletin de l'Institut Français de l'Afrique Noire, 24B (1962), 401. The same tradition was recounted to me by the chief of the Ida Ali in Shinqit, although he referred only to 3,000 camels (which would seem more realistic), all belonging to the “Shinqitois.” Interview, Ahmed Ould Abderrahaman (Ida Ali) 18 August 1984 Shinqit. Other renditions: El-Wasît, 15; Paul Marty, cited in Charles Monteil, “La Langue Azer,” from documents collected by Brosset, Diego, in Monod, TheodoreContribution à l'étude du Sahara occidental (2 vols.: Paris, 1938/1939), 2:215Google Scholar; Bonafos, V., “Les Ida ou Ali Chorfa Tidiania de Mauritanie,” Revue Monde Musulman, 31 (1915–1916), 230–31.Google Scholar Bonafos refers to the town of “Zar.” The exact era during which this was to have taken place varies according to account; while the Norris and Bonafos versions suggest the seventeenth century, and that would also be consistent with my oral source, the Monteil-Mauny account places it in the sixteenth century, and a reference to “caravans travelling with regularity from the [mine] Ijil towards Zara” by Raymond Mauny on the history of Shinqit is situated in a timeless past. Mauny, , “Notes d'histoire et d'archéologie sur Azougui, Chinguitte et Ouadane,” BIFAN, 17 (1955), 148.Google Scholar

39. See for example, sources cited in note 38.

40. Hughes, Thomas P., A Dictionary of Islam, (London, 1965), 713–15.Google Scholar The dictionary entry gives India and Central Asia as examples, but the term ziara (or zyara) is also used in Mauritania, as, for example, in reference to the tomb of Shaykh Mohammed Fadel in the Adrar.

41. Monteil, “Langue Azer,” 216n1; in ibid., 216n2. The “Zar/Dyara” identification is made in n2 with reference to Mauny's account of the Shinqit-Tishit caravan. The other names cited for Dia are “Zaber” or “Zaberma/Dyerma” in Songhay. However, the proximity of the Niger River to Dia makes it an unlikely candidate for Hodges’ “Tarra Forest.”

42. For a discussion of Azer in this context, and references to most of the existing literature on it, see McDougall, E. Ann, “The View from Awdagust: War, Trade and Social Change in the Southwestern Sahara, From the Eighth to the Fifteenth Century,” JAH, 26 (1985), 27–28.CrossRefGoogle Scholar See Monteil, especially Charles, “La langue azer d'après les documents recueillis par Th. Monod et D. Brosset” in Contribution, 2: 212–343, esp. 215–16Google Scholar; and in addition, Norris, H. T., “Future Prospects in Azayr Studies,” African Language Review, 9 (1970/1971), 99–109.Google Scholar

43. Tadeusz Lewicki provides an excellent example of the historian as detective in his application of linguistic analysis to the existing written and oral evidence regarding the mysterious medieval kingdom of Zafunu, in the course of which he refers to Azer and Monteil's findings. See Lewicki, , “Un état soudanais médiéval inconnu: le royaume de Zafun(u),” Cahiers d'Etudes Africaines, (1971), 501–25.Google Scholar

44. Charles Toupet identified Tagant as the feminine of the Berber word agan, meaning “forest of great trees.” He cautioned that this is often an exaggeration or, as Odette de Puigadeau put it during her travels through Mauritania in the 1940s, “the African forest: most of the time the voyager towers above it from the height of his camel.” In this case, however, earlier twentieth-century colonial reports confirm that the region was one of sizable trees, housing a varied animal life. See Toupet, , La sédentarisation des nomades en Mauritanie central sahelienne (Paris, 1975), 218.Google Scholar

45. Colonial reports confirm this. It seems agricultural ‘development’ policies had much to do with clearing inappropriate lands and altering soil conditions. Not only did much of the ‘forest’ disappear, but with it the animals and wild-fruit-producing plants that the region's occupants used to depend on to augment food supplies.

46. Lewicki, , “Royaume,” 521.Google Scholar The Triq-Lamtuna was the commercial route followed from the Wadi Draa region of Morocco to ancient Ghana by the Almoravids; see Paulo de Moraes Farias, “The Triq Lamtuni (XIth-XIIth c): A Trade Route?” Paper presented for the ‘Muslims and non-Muslims in Africa’ Seminar, S.O.A.S., December 1970.

47. By the 1680s and 1690s Tidjikja had achieved fame in this regard through Sidi Abdullah wuld al-Hajj Brahim. In the course of the seventeenth century, he traveled to Egypt and Mecca and wrote theological poetry prolifically. He was renowned by the latter part of the century and was attracting followers, telamidh, to Tidjikja by the early eighteenth century. Bonafas, “Les Ida Ali;” Toupet, Charles, Sédentarisation, 271–72Google Scholar; Abdellahi wuld Yubba wuld Kalifa, “Essaie de monographie historique de Tidjikja,” Centre Culturel Français, Nouakchott.

48. Bonafas, “Les Ida Ali;” Toupet, , Sédentarisation, 271–72Google Scholar; Abdellahi wuld Yubba wuld Kalifa, “Essaie de monographie historique de Tidjikja.”

49. The tomb is in fact located about a day–and-a–half's journey away. Lewicki, , “Royaume,” 521.Google Scholar

50. Abitol, Michel, Tombouctou et les Arma. De la conquête marocaine du Soudan nigérien en 1591 à l'hégémonie de l'empire Peulh du Macina en 1833, (Paris, 1979), 186.Google Scholar

51. Daveau, Suzanne, “Etude géographique de la region de Tegdaoust,” 39Google Scholar; and Charles Toupet, “Les conditions climatiques et végétales et la mise en valeur de la région (le R'Kis),” 66; both in Dévisse, Jean, Robert, R., and Robert, D., Tegdaoust I: recherches sur Aoudaghost 1 (Paris, 1970).Google Scholar

52. Also “Tolba” or “Tugba” in some sources, from tulba (pl. tulub), a reference to clans who have at least theoretically renounced the use of force in the settlement of disputes, in keeping with the Tajakant who left Tinigi for Togba in the sixteenth century. See the glossary in Stewart, Charles C., Islam and Social Order in Mauritania, (Oxford, 1973), xvii, xviii.Google Scholar

53. Toupet, , “Les conditions climatiques,” 62.Google Scholar Its similarity to Tidjikja in terms of topography and suitability as a caravan or travelers' resting spot, tarha, is also striking.

54. Toupet, , “Les conditions climatiques,” 62Google Scholar; Laforgue, Pierre, “Note sur Togba, ancienne capitale des Tadjakant du R'Kis (Mauritanie saharienne),” Bulletin de la Société géographique et archéologique d'Or an, 62 (1941)Google Scholar, passim; Robert-Chaleix, Denise and Richir, Claude, “Un site jusqu'à présent inconnu, dans le Rkiz: Moulay Arbad” in Tegdaoust III. Recherches sur Aoudaghost. Campagnes 1960/65, enquêtes générales (Paris, 1983), 349–50Google Scholar; ibid., Robert, D., “Fusaiöles décorées du site de Tegdaoust,” 510–12.Google Scholar

55. Robert, , “Fusaiöles,” 510–12.Google Scholar

56. Amilhat, , “Les Ida Aich,” 112.Google Scholar

57. al-Chennafi, “Sur les traces d'Awdagust,” 101; also note 42, above.

58. Claude Meillassoux, “A propos de deux groupes Azers: les Giriganko-Tegdawst et les maxanbinnu (Mali),” in Tegdaoust III, 526n15. Curiously, he says his informants placed the emigrants' new home in the Hodh between Timbedra and Alaso, but said it was located some 300kms northwest of Alaso. This distance and direction cannot refer to someplace between Timbedra and Alaso, but Togba is located 370kms to the northwest.

59. Laforgue, Pierre, “Note sur Togba,” 61.Google Scholar

60. Curtin, , Economic Change, 51.Google Scholar

61. Abitbol, , Tombouctou, 186Google Scholar, citing Marty. Marty's brief rendition of this history places the marriage at the beginning of the eighteenth century and more clearly links Sidi al-Mahjub to the northern Tajakant (Maures du Sahel et du Hodh, 97). However, another tradition recounted by Norris places the “saint” al-Mahjub al-Jakani en route to the Qsar (walled town) “Tugba” earlier in the century, at which time he prophesized the emergence of the jihad leader Nasr al-Din. (“Znaga Islam during the Seventeenth and Eighteenth Centuries,” Bulletin of the School of Oriental and African Studies, 32 [1969], 512–13.Google Scholar)

62. Mohammed el-Chennafi, “Sur les traces d'Awdagust: les Tagdawest et leur ancienne cité” in Tegdaoust 1, 106n2.

63. Curtin, Economie Change, 51n5.

64. It also encompasses Taranni, founded by the Ahel Taranni (Azer-speaking Masna emmigrants from Tishit), probably towards the end of the seventeenth century. (al-Chennafi, “Sur les traces d'Awdagust,” 101n3.) Nara, however, was developed only in the nineteenth century.

65. Stone, “Cornelius Hodges,” 90n4. Hodges corrected Jobson's mileage from James Island to Barracundy Falls on the Gambia from 900 to 558 miles; Stone in turn, corrected this to 257 miles.

66. Stone, , “Cornelius Hodges,” 93.Google Scholar

67. The only other possibility is that salt was being brought from Tawdeni. While this cannot be ruled out completely, it is highly unlikely, given what we know of salt-trading patterns in the late seventeenth century in the Tagant-Hodh region. See McDougall, “Salts of the Western Sahara.”

68. Stone, , “Cornelius Hodges,” 90.Google Scholar

69. Curtin, , Economic Change, 280.Google Scholar

70. Quotation from Emerson, Ralph Waldo, cited on the title page of HA, 1 (1974).Google Scholar Our “quest” epitomizes both Emerson's sentiment, and HA's concerns, for it can be said unequivocally to have discovered no truths and few conclusions!

71. See for example, Hair, P.E.H., “Barbot, Dapper, Davity: a Critique of Sources on Siera Leone and Capte Mount,” HA, 1 (1974), 25–54.Google Scholar

72. See Henige, David “On Method: An Apologia and a Plea,” HA, 1 (1974), 1–8.Google Scholar

73. See Map 1. We have by no means exhausted the possibilities here; on the other hand, if the archeological work from the Tegdaoust site is any indication, we have a lot of work to do on the ground (quite literally…) before we should begin thinking in terms of conclusive or comprehensive interpretations.

74. Curtin drew our attention to this issue some years ago, suggesting that a paucity of suitable source materials might prove an insurmountable barrier to fuller investigation. However, (and with reference to my comments in the preceding note), work in Mauritania itself, as well as the maturing of oral history as an accepted methodology, have opened new doors to historians. It is time to look again at some of these ideas. (Curtin, , Economic Change, 51–53.Google Scholar)

75. It is true that Saharan written materials are usually in Arabic, a language few historians of the “sub-Sahara” master.

76. An example of this problem, pertinent to the “Tarra” question can be seen in note 31.

77. The first draft of this paper was completed in October of 1990 when political conditions in the Persian Gulf region had already revealed something of this potential. The final version was completed as the Gulf War raged without end in sight, and as Mauritanian and Sudanese allegiances to Sadam Hussein's vision of a new Arab World were being strongly proclaimed. Presumably, the outcome of this conflict will reaffirm those allegiances to an Arab rather than an African world, and these politics will only exacerbate what is already a problem in the writing of West African history.