Chapter III: Part 3

[Footnote 46: Not. et Extr. p. 631. The fact that the campaign of the Morabites in one year embraced both Aúdaghost and Sijilmésah, is enough to show that the former place was contiguous to the Western Ṣaḥrá.]

[Footnote 47: The language of El Idrísí (Rec. de Voy. V. p. 109), paints the decay of Aúdaghost: he describes it to be “a little town, deficient in water; with a scanty population and miserable trade, which consists in camels.” This is the town which modern geographers, induced by a supposed resemblance of names—though Rennell (Geogr. Illust. of Park’s Journey, in Proc. of Afr. Assoc. I. p. 501,) took the precaution to convert Aúdaghost into Agadost—have chosen to identify with Aghades, or Aghdes, which Leo Africanus (pt. VII. c. 9), writing in 1541, calls “a city built by the moderns;” while Marmol (III. fol. 24), more precise, says that it was founded 160 years before the time of his writing, or in 1438.]

[Footnote 48: Not. et Extr. p. 630. Bucklers made of the skins of the Dant or Lant (probably el-ant), which is supposed to be the _Antilope Leucorix_, were chiefly manufactured in Wád Nún. By the amber carried to Aúdaghost from the sea shore, we must understand ambergris, to which El Idrísí alludes when describing the western shores of Africa (Rec. de Voy. pp. 64 and 135).]

[Footnote 49: Not. et Extr. p. 615.]

[Footnote 50: Ghábah غابه MS. B.M. fol. 112 r; Alghábat الغابت Rec. de Voy. II. p. 2; Ghaïah غايه Not. et Extr. 643. The predominant idea in the meaning of the name Ghábah or Ghábat̤, which is undoubtedly the true reading, is _obscurity_: lowness of situation and overhanging gloom are both implied by it.]

[Footnote 51: Ṣínghánah صينغانه. Caillié (tom. II. p. 237) mentions a place called Sangouno, on the left bank of the Great River, three or four days from Jenni.—Tekrúr تكرور.]

[Footnote 52: Silla سلي, سيلي, and سلا. El Bekrí mentions cotton as one of the chief productions of this country; no house, he says, was without its cotton tree. Leo, in like manner, says (pt. VII. c. 3), that cotton was the staple merchandise of Jenni, which is but two days from Silla; and Caillié observed the general cultivation of that article in the country south-westward of Jenni (tom. II. pp. 156-167.) The people of Silla, being slave-dealers, made constant war on their pagan neighbours, of whom the nearest were the Kalembú قلنبو, a day’s journey distant. Now the district of Negroland at present characterized by the termination _bú_, is that contiguous to the modern Silla towards the west, and on the northern side of the river. There we find Modiboo, Doolinkeaboo, Fanimboo, &c., within a small compass. “From Tarankat̤i ترنْقَةِ (MS. B.M. fol. 111 r), near Silla, the inhabited country (says our author) extends to Záfḳú زافقو” which name M. Quatremère reads Afnou (Not. et Extr. p. 641). But if we suppose that a Nún ﻨ is here mistaken for the Maghrebí Kaf ڧ, the two readings will be reconciled in Zafnú, the Jafnoo of our maps, which is a very likely limit to our author’s exact information. To point out unequivocally the direction in which his narrative led him, he adds, that “the country continues populous to the ocean.”]

[Footnote 53: Aúghám اوغام MS. B.M. 1140; Audagam اودعام Not. et Extr. p. 651—Merásah مراسه—Tírḳa تيرْقَي MS. B.M.—Tádmekkah تادمكه—Seghmárah سغمارة—Kaúkaú كوكو.]

[Footnote 54: Not. et Extr. pp. 652, 653. The ten journeys allowed between Wérgelán وارجلان MS. B.M., وارقلان Not. et Extr., and Ghodémis غُدامِس, show the scale by which we are to measure this route, and allow us to stretch the forty journeys between the latter place and Tádmekkah farther than could be done without such an intimation.]

[Footnote 55: Macrízí, in Hamaker’s Specimen Catalogi Cod. Or. MSS. Academiæ Lugd. Bat. pp. 207, 9. In the passage in question M. Hamaker reads Taoumcah تاومكه, instead of which it is an obvious correction to restore Tádmekkah تادمكه. If we increase forty in the ratio of thirty to fifty, the numbers by which our authors respectively measure the distance between Ghánah and Tádmekkah, we shall have sixty-six days for the distance of the latter place from the Tajúah, according to El Bekrí’s scale. The Tajúah or Tajuwín of the Arabs, are the people whom Browne calls Dageou (Travels in Africa, p. 325), and who once ruled Darfur.]

[Footnote 56: The name Tádmekkah signified The Likeness of Mekkah, (Not. et Extr. p. 653.) But Ned Roma, as Leo informs us (pt. IV. c. 6), signified The Likeness of Rome. A single point in Arabic writing discriminates between these two prefixes. But since El Bekrí, who writes Tádmekkah, also writes Nádrúmah, we must be satisfied to ascribe the apparent discrepancy, in this case, to difference of dialect. But it may be here remarked, that the pages of Leo Africanus are not quite free from the inaccuracies which originate in an unpointed Arabic text. They offer, for instance, Perzegreg for Ber Zegzeg, Nefreoa for Nefzeoa, Amarig for Amazig, and frequently Ibn Racu for Ibn Rachic. In the early French translation of Leo (by Jean Temporal, 1556), we find also Cairaran frequently written for Cairaoan, Azarad for Azaoad, and Araran for Araoan. Whether these errors have been all copied from the version of Leo in the first edition of Ramusio’s first volume, we have not had the means of determining; but the second edition of Ramusio (1554) is free from the more glaring of them.]

[Footnote 57: Safnaḳú سفنقوا—Búghrát بوغرات MS. B.M. 115 v; Not. et Extr. 652.]

[Footnote 58: Silla is fourteen days from Tomboktú by land, and a month by water. From these extremes it is easy to derive the distance assigned above. Sidi Hamed (in Riley’s Narrative, p. 334) going from Tomboktú to Houssa, first travelled six days along the river, a little south of east till he came to hills, where the stream turned southwards.]

[Footnote 59: “Verso mezzogiorno, e quasi inchina alla parte di scilocco.” Pt. VII. c. 3.]

[Footnote 60: The statements of this valuable writer, as well as the journey of Ibn Baṭúṭah, will be given at length further on.]

[Footnote 61: Jaubert’s Idrísí, in the Rec. de Voy. V. p. 17; Ibn el Wardí.]

[Footnote 62: Alexander Scott, in his pilgrimage beyond the Great River, found that the desert continues to the very shores of Lake Debú, where there was a town or encampment of the Orghebets (Raghabát?)—Edinb. Phil. Journ. vol. IV. p. 43.]

[Footnote 63: Not. et Extr. p. 640.]

[Footnote 64: The people of Melli (Málí), according to Leo (pt. VII. c. 4), were the first to embrace the Mohammedan faith.]

[Footnote 65: This point will be more fully considered when we come to speak of Tekrúr.]

[Footnote 66: The Brebísh often encamp eastward of Tomboktú, in which quarter nevertheless the Tawárik seem to have gained ground on the Zenágah.]

[Footnote 67: It is a strong argument in favour of the construction here given to the route to Ghaïárú غيارُوا (MS. B.M. 1120), that it establishes a uniform and consistent method in El Bekrí’s narrative. That author begins his account of Negroland with the Ṣínghánah, who traded with the Benú Goddálah in the west. He then goes to the south- west to Silla and Tekrúr; then having mentioned Ghánah, he passes to the route to Ghaïárú, and finally turns due east, and describes the route to Tádmekkah. It is requisite for clearness and exact order, that the route to Ghaïárú should lie between the south-west and the east.]

[Footnote 68: Sámaḳanda سامَقَنْدَي MS. B.M. 113 r; Sámaghondi سَامَغُنْدِي Rec. de Voy. II. p. 4. The corruptions of this name, which is probably significant, are enumerated in the notes to Hartmann’s Idrísí, p. 42.—Ṭáḳah طاقة MS. B.M.; Ṭáḳat طاقت Rec. de Voy. p. 5; Tanah, Not. et Extr. p. 646.—Zúgú زوغُوا MS. B.M.; Zoghárá زُغَارَا Rec. de Voy.; Rougou, Not. et Extr.]

[Footnote 69: Gharnatil غرنتل MS. B.M.; Garbil, Not. et Extr. عُونْتِل Oʿuntil, Rec. de Voy. This latter reading brings to mind the place called by Mohammed Maséní (Clapperton’s Second Journey, p. 330), Oodel or Goodel (with the same doubtful initial letter), where the Great River is crossed between Sokkatú and Maséna. In the absence of better guidance, Ghúntil غونتل shall be here assumed to be the true reading, and the name of the identical place called Goodel by Bello’s servant.]

[Footnote 70: The MS. B.M. fol. 113, has Yersana يرسني, which seems too violently opposed to the other MSS. In the Rec. de Voy. and Not. et Extr. it is Bersa برسي.—This is the Berísa بريسي of Idrísí, the Berísá بريسا of Abulfedá. Bersana was the resort of certain negroes who brought gold from the interior, and were called Benú Nʿamrát بَنو نَعمْرَات (Rec. de Voy. p. 7), or Wangamranah ونعمراٮه (Not. et Extr. p. 647), or Benú Zammakhrátah بنو زمخْراتة (MS. B.M. 113 v). It is plain that the text translated in the Not. et Extr. has been curtailed of the first two letters of the name, but if these be supplied, together with the diacritic points, it agrees with the text of the Rec. de Voy. The name, therefore, will be Benú Nʿamrátah, or Namrát. But who can be the negroes bearing such a name? Sultan Bello informs us, (Appendix to Denham and Clapperton’s Travels, vol. II. p. 454) that the people of Yarba or Yariba “originated from the remnants of the children of Canaan, who were of the tribe of Nimrod.” The people of Yariba therefore seem to be the Benú Nʿamrát. But to this it may be objected that Nʿamrát is not the correct Arabic mode of writing Nimrod. Truly not; neither can the blacks of Yariba, we verily believe, prove their descent from the great hunter. But the name and the historical tradition in this case are both equally spurious; they were both probably suggested by a sound—we think indeed, by the same sound, or, in other words, we believe that the name which was shaped into Benú Nʿamrát, and afterwards into Benú Nemrúd or descendants of Nimrod, belonged to the people of Yariba.]

[Footnote 71: Daur دور Not. et Extr. p. 647; Daú دوْ MS. B.M. fol. 113 v.; Dawa دَوَ Rec. de Voy. II. p. 7.—Malelo مَللُ MS. B.M.; Malik مَلكْ Rec. de Voy. For the various readings of the names Ghaïárú (Ganarah of D’Herbelot) and Ghuntil, see Hartmann’s Edrísí.]

[Footnote 72: In the Kissour language, spoken, according to Caillié, in Tomboctú, Jenni, and in the intervening country, the word Ganda (Caillié III. p. 313), or Gunda (Clapperton’s First Journey, p. 182), signifies Land or Country, so that Samaḳanda or Samaghondi, explained by it, would mean Samah-land. Nor is this explanation less probable from the circumstance that the name Sami, and the termination Kanda or Konda, signifying town, is common among the Mandingoes, who overwhelmed Ghánah from the south, as shall be shown hereafter, and who now people the country south of Tomboktú (Caillié, tom. II. 252).]

[Footnote 73: Marmol (vol. III. fol. 17). Yça, that is, Issa, (Hissa in Caillié’s vocabulary) signifies _river_ in the language of Tomboktú. The Serakholies inhabit Galam. The Tucorones therefore interposed between them and the Wolofs, must be the Fúlah or Fellatah, who occupy both banks of the Senegal, in the neighbourhood of the Isle de Morfil. We find in a MS. vocabulary of their language, brought home by Clapperton, the word _mio_ signifying a lake, probably any large sheet of water. With respect to the name here applied to them, a respectable authority (Dard, Grammaire Wolofe, p. 148) informs us, that a division of the Fúlah nation bears the appellation of Teukirères. The name Tucorones seems related to the plural Tekayrne, used by Burckhardt (Trav. in Nubia, p. 365); while Teukirères rather resembles the Tekrírí of Ibn Baṭúṭah. It is manifest that the route pointed out by this series of names is that of the slave-dealers between Galam and Tomboktú. They cross the desert at a distance from the Great River where it is called Joliba, and first reach the stream where it turns eastward, north of Lake Debú. There, according to Bowdich’s informant (Mission to Ashantee, p. 193), Jinbala is on the left bank of the river. It is placed on the right by all other authorities except Marmol and his copiers, who give the name to the river.]

[Footnote 74: Caillié’s account of Jimbala, or, as he writes it, Ginbala, was confirmed by Abú Bekr, the intelligent native of Tomboktú who accompanied Mr. Davidson in his ill fated attempt to cross the desert from Wád Nún; and of whom an interesting account may be read in the Journal of the Royal Geographical Society, vol. VI. p. 100. The initial sound in Jimbala is one hard to be seized by a foreign ear. It is the same which Caillié (II. pp. 82, 160) sought to express by a triple form in Jaulas, Diaulas, and Iolas. Park wrote the same name Jules (First Journey), while Mr. Watt preferred Nyalas (Proc. Afr. Ass. I. 436). The natives themselves often express the sound in question by yʿe يع.]

[Footnote 75: Marmol elsewhere (III. fol. 27 r) distinctly places the _Baganos_ on the river at the point where the road from Galam to Tomboktú first reaches its banks. Livio Sanuto also (Geografia, 1588, fol. 83) says, “that Zimbala or _Bagano_ adjoins Tomboktú on the south, beyond the Sanaga,” that is, the Great River. Mohammed Maséní (Clapperton, p. 331) mentions a lake Búkma as being in the same tract as Lake Jeboo, that is, Dhiebú or Debú.]

[Footnote 76: Not. et Extr. p. 617. Másín ماسين MS. B.M. fol. 103. For Aúghám, see page 29. The Massina of our maps, and Maséna of the Translations appended to Clapperton’s Second Journey, is generally written Másín or Máshín by the natives (see the documents appended to Bowdich’s Mission, and to Dupuis’ Residence in Ashantee).]

[Footnote 77: Anbarah اَنْبارَة Rec. de Voy. p. 8. The king of this country was styled Tárim تَارِمْ. If for this we could read Farim فارم, we should have a true Mandingo title. The difference between Anbárah اَنْبارَهْ and Oonbori, probably اُنْبُرِي supposing the vowels not to have been supplied conjecturally, (for otherwise the latter might be read Anbara,) is no greater than may be expected where orthography is unsettled. The title of the king of Oonbori is Farma (Clapperton’s Second Journey, p. 331), which, as well as Farim, imports a governor or local chief. That by Kúghah كُوغَه (Rec. de Voy.) El Bekrí meant the Cochia of Cadamosto (Ramusio, I. fol. 108 v) and Gago of Leo, can hardly be doubted; but this point shall be examined hereafter. His statement, that Anbárah, nine days from Kúghah, west of Ghánah, does not admit of any plausible defence; but if Sámah سامه be read in this place for Ghánah غانه then not only does all difficulty vanish, but the author’s discourse acquires coherence and natural order.]

[Footnote 78: Kitábu-l-Jʿaráfíah (Book of Geography), &c. MS. in the collection of D. Pascual de Gayangos. Ráyawen رايون has a suspicious resemblance to Arawan. It is quite gratuitous to suppose that the Morabites, who were all of the Zenágah nation, and who rushed at once, as soon as they felt their strength, from their own deserts to the conquest of Barbary and Spain, ever went eastward as far as Houssa, or even to Kághó.]

[Footnote 79: “No one (in Aïwalátin, that is, Walata) is named after his father, but after his maternal uncle; and the sister’s son always succeeds to property in preference to the son: a custom I witnessed nowhere else except among the infidel Hindoos of Malabar.” (Lee’s Ibn Baṭúṭah, p. 234.)]

[Footnote 80: “Of all people the Blacks debase themselves most in presence of their king.... When the Sultan addresses one of them, he (who is addressed) will take the garment off his back and throw dust upon his head” (Lee’s Ibn Baṭúṭah, p. 240). The ceremonial of Tomboktú (Leo, pt. VII. c. 5), and that of Ghánah (Not. et Extr. p. 644), are described in nearly the same terms.]

[Footnote 81: Proc. of Afr. Assoc. I. p. 149. Though Clapperton says little of the laws or government of Houssa, yet his narrative discloses the subdivision of power in that country. The people there have never been trained up under a paramount tyranny.]

[Footnote 82: Makrízí (Quatremère, Mémoires sur la Nubie, tom. I. p. 28; Burckhardt’s Travels in Nubia, p. 456) relates of the court of Kánem, and Ibn Baṭúṭah of that of Bornú, the ceremony of audience, as it was witnessed in the latter place by Denham (I. p. 231).]

[Footnote 83: Caillié saw (Voy. II. p. 253) a line of hills of red sandstone without any vegetation, on the left bank of the river, about forty miles south of the lake; and, at an equal distance north of it, sand hillocks bordered the stream (p. 266). It is explicitly stated by Marmol (III. fol. 15 v), that Jenni had all the trade of the Zenágah, the Brebísh, the Ludayas, and the Arabs of Arguin. But he erred in supposing that the conflux of Arabs and Berbers from the shores of the Ṣaḥrá to that city was owing to its western position. It was rather due to the character of the intervening country, which may be called a fine desert.]

[Footnote 84: It is plain, from Scott’s narrative (Edinb. Phil. Jour. vol. IV. p. 45), that the level desert continues, without any change, save in the frequency of brackish rills, to the very shore of the lake; southwards from which the country seemed uninhabited; but a little to the north was the town of the Orghebets, in which the dwellings were constructed of canes and bamboos.]

[Footnote 85: “_Gualata_, que otros llaman _Ganata_.” Marmol, III. fol. 21 v. It is hardly necessary to observe, that, in the orthography of Southern Europe, Gualata represents our Walata: “_Gualata_ o _Ganata_,” (I. fol. 17.) “Vled Vodey andan en los desiertos que estan entre Iguaden y _Ganata_; son señores de Iguaden, y el Rey Negro de _Ganata_ les paga cierto tributo,” &c. (I. fol. 39.) “Alarabes llamados Udaya, y por otro nombre Vled Vodey, que moran el desierto de Lybia que está entre esta poblacion (Guaden) y _Gualata_ reyno de negros.” (III. fol. 3.) “En Gelofe, Geneúa, Tombuto, Meli, Gago y _Ganata_, hablan una lengua llamada Zungay.” (I. fol. 44.) This last sentence is taken from Leo (pt. I. c. 11), who, however, writes Gualata. Marmol, in his first volume, seems to prefer Ganata, but, in the third, he generally follows Leo with little deviation. When Ali Bey (Badia) speaks of caravans going “from Sús and Táfílélt to Ghánah and Tomboktú,” (Travels in Barbary, &c. I. 45,) does he inadvertently mix ancient with modern times—his reading with his recent intelligence?—or does he mean by Ghánah, Ghanata, that is, Walata? What was surmised in Note 6 respecting the predominance gained by the analogies of the Berber language over those of the Arabic, and the change of the contingent t̤ into the absolute t, seems confirmed by the MS. extract of El Bekrí published in the Rec. de Voy. II. That MS., however inferior in other respects, is yet good authority on the subject of the relation subsisting between ancient Arabic and Moorish orthography; and we find that it writes Ghánat, Sámat, and Ṭáḳat, for Ghánat̤ or Ghánah, Sámat̤, Ṭákat̤, &c.]

[Footnote 86: “It appears singular that the country immediately to the eastward of Timbuctoo as far as Kashna should be more imperfectly known to the Moorish traders than the rest of central Africa” (Quart. Rev. No. 45, May 1820, p. 234.) The reviewer, however, errs in ascribing the obscurity which involves that tract to the wars of the Fellatah. But the fact is, that between Tomboktú and Houssa passes the line of demarcation separating what may be called the two commercial provinces of Negroland, which depend on the two great roads (from Fezzán and Táfílélt), and have little communication with each other on their northern frontiers. Besides, the interposed desert supports a formidable population of Tawárik.]

[Footnote 87: The Arab geographers, unable to form an exact conception of the country westward of Ghánah, diminished the distance between that place and the ocean. El Idrísí makes the distance between Silla and Aúlíl to be sixteen days’ journey (Rec. de Voy. V. p. 11); but Abulfedá sets Ghánah only four degrees eastward of the ocean (Reiske’s Trans. in Büsching’s Mag. V. p. 354). In like manner Leo (pt. VII. c. 3) says that the kingdom of Jenni, extending 250 miles along the river, reaches the ocean; and he supposes Walata to be only a hundred miles from the sea shore. As he was copied in all his errors, our maps of the sixteenth and seventeenth centuries uniformly placed Tomboktú too far westward.]

EL IDRISI COMPARED WITH EL BEKRI.

MAGHRAWAH.

The account of Negroland contained in the foregoing pages is drawn altogether from El Bekrí. So much does his description of Ghánah and the adjoining countries excel in copiousness of detail as well as in clearness, that in all the Arab writers succeeding him, there is hardly a single new particular worth adding to it. Some of these writers, nevertheless, are much better known than El Bekrí; and one of them, El Idrísí, whose work, entitled ‘The Amusement of one desirous of knowing all the Countries of the World,’ was composed about the year 1153, has been long regarded as the first authority on questions relating to the geography of Central Africa. It will be worth while, therefore, to set these two authors side by side, and to compare them carefully, so that we may be able to choose between them when they disagree, and to determine whether the later of the two improved on his precursor.

The first peculiarity of El Idrísí that strikes the attentive reader, is his general reduction of distances in the desert. He begins to contract even in the vicinity of Atlas, and makes Sijilmésah to be only thirteen days distant from Wád Nún; whereas these places must be at least twenty good journeys asunder.[88] But to confine ourselves to the consideration of the principal dimensions of the Ṣaḥrá, we may recal the statement of El Bekrí that Aúlíl was a two months’ journey distant from Wád Nún, going along the shore. Now from Sijilmésah to Aúlíl, which was a greater distance, is reckoned by El Idrísí to be a journey of only forty days. This supposes (since Arguín is exactly 900 miles from Sijilmésah) a rate of 22½ geographical miles a day in a straight line—a rate far exceeding what is practicable on a journey of such a nature and extent.[89] Nor can El Idrísí have here the benefit of any objection to the hypothesis which places Aúlíl at Arguín, since he informs us that Aúlíl was one day’s sail from the mouth of the river called the Nile of the Blacks, by which he must be understood to mean the Senegal. But the mouth of this river is at least two days’ sail from Arguín; so that to make that author consistent with himself, it must be allowed not only that Aúlíl was in the Bay of Arguin, but also that his numerical expressions of distance are, in this instance, too low.[90] In like manner he reduces the distance between Sijilmésah and Silla to forty days; and that from Wád Nún to the latter place to thirty-two days,—viz. seven from Nún to Arkí, and thence twenty-five to Silla; his estimate in each case being less than two-thirds of the reality.[91]

But the same author’s reduction of the longitudinal dimensions of the Great Desert is still more remarkable. Instead of a journey of two months between Aúlíl and Silla, as may be inferred from El Bekrí, whose measures of distance accord strictly with the results of modern inquiry, El Idrísí separates those places by a distance of only sixteen days. Again, he makes Aúdaghost to be but a month from Aúlíl, and twenty-five days from Jermah in Fezzán; so that from the latter place to the shores of the Atlantic, the desert should be crossed in fifty-five days, or less than half of the time actually required for that journey.[92] Again, he states Kúghah to be a month from Dongolah, and six weeks from Ghánah (though the genuineness of the text is here liable to suspicion); and consequently he reduces the whole distance between Dongolah, or the Nile in the east, and the Atlantic Ocean in the west, to a journey of three months and a half, which is not more than a moiety of the true distance.[93]

The incorrectness of El Idrísí exemplified above, cannot be explained away by supposing that he made use of a large scale of measures, or a conventional day’s journey of great length. His contraction of space is not sufficiently uniform to sustain that plea: it operates chiefly on the uninhabited country. Like modern geographers, he seems to have had an invincible dislike to large blanks in a map; and among the expedients to which he had recourse for the purpose of filling them up, was the common one of dilating as much as possible the contiguous inhabited countries. A perfect illustration of this remark is afforded by his description of the course of the Great River in the vicinity of Ghánah.

It has been seen that El Bekrí places the capital of Ghánah not immediately on the Great River, but at a distance of perhaps three days’ journey from it. But Ṣínghánah, with which place he begins his description of Negroland, is described by him as standing on both banks of the river, and having Tekrúr on the south-west. Now in El Idrísí’s geography there is no mention made of Ṣínghánah, but its description is transferred to Ghánah, which is made to stand on both banks of the river, and to have Tekrúr on the south-west. The distance of twenty days, according to El Bekrí, between Silla and Ghánah, is increased by El Idrísí to twenty-four days. The former of these writers thus describes the route eastward and along the river from Ghánah. First, five days to Rás el má, then six days through the country of the Merásah to Tírḳa, where the river turned southwards, and three days further along its banks, to the country of the Seghmárah, which commenced, therefore, at the distance of fourteen days from Ghánah. The same course is thus described by El Idrísí:—From Ghánah, six or eight days to Tírḳa; six more to Merásah; and another six to Seghmárah: so that the fourteen days of his author are here increased to eighteen.[94]

In accordance with the construction given above to El Bekrí’s route from Ghánah to Ghaïárú, El Idrísí makes the portion of the river therein mentioned to form a part of its course below Seghmárah. But the manner in which the latter writer puts together his materials in this place demands attentive consideration. The following are his details of distance and bearing:—

From Samghadah (Sámaḳanda) to Seghmárah, 8 days.

From ditto to Kúghah, _eastwards_, 10

From ditto to Gharbíl (Ghúntil), 9

From Seghmárah to Gharbíl, _southwards_, 6

From Gharbíl (Ghúntil) to Ghanárah (Ghaïárú), _westwards_, 11

From Ghánah to Ghanárah, 11

Thus it appears that El Idrísí makes the river flow first north-eastward from Silla to Ghánah; then eastwards to Seghmárah; then southwards to Ghúntil, and finally westwards again to Ghaïárú.

The Sámaḳanda of El Bekrí, which was four days from Ghánah, and his Kúghah, fifteen days from the same capital, and nine from Anbárah, (which appears to have been near Sámaḳanda,) are evidently the Samghadah and Kúghah of El Idrísí. This writer agrees with his predecessor in making the river flow, first north-eastwards from Silla to Ghánah, then eastwards, and then southwards. So far he seized with some felicity the idea of the general winding of the river. But his turning it westward from Ghúntil can be explained only by supposing that he misunderstood his authority. El Bekrí says, that opposite to the Seghmárah, whose territory extended from the Great River to Tádmekkah, was Kaúkaú. And again, he remarks, that the road to the country of the Remrem went westwards along the river from Kaúkaú. Now this last passage furnishes the explanation of the westward course which El Idrísí has given to the river, if we suppose that he confounded for a moment Kúghah with Kaúkaú.

El Bekrí mentions Bersana after Ghaïárú; and accordingly El Idrísí sets this place, under the altered name of Berísa, due west of Ghaïárú, and on the Great River, half way between Silla and Ghánah. The series of names which the latter found in his author he thus arranged in a circle, under the influence of misconception, so as to make it terminate in itself. But the artificial division of climates, by severing Berísa from the group of names to which it originally belonged, fortunately obviated the ready exposure of so absurd a concatenation.[95] The Malilo and Daúr or Daú of El Bekrí are evidently the Malel and Daú of El Idrísí; and in consequence of the same mistake which placed Berísa on the river between Ghánah and Silla, they are brought into the neighbourhood of the latter place. Their character, too, is as much misrepresented as their position. Instead of being countries of some extent and importance, they become, in the pages of El Idrísí, only towns of Lemlem, the wretched inhabitants of which, possessing but few camels, wander over deserts destitute of water;—a picture of Negro poverty, more likely it must be confessed, to originate in the imagination of an Arab, than in the physical character of the country south of the Great River.

From El Idrísí’s delineation of the Great River we may return with advantage to consider the position assigned by him to Aúdaghost. That town was, according to him, thirty days from Aúlíl, thirty-one from Wergelán, and twenty-five from Jermah. The short distance of fifty-five days herein allowed between Aúlíl and Jermah, will not admit of being applied to a circuitous route. The distance from Wergelán, therefore, which is relatively long, and by reaching far southwards has the effect of elongating the preceding line, must be supposed to be circuitous. And this is a well-founded supposition, inasmuch as the road from Wergelán to Aúdaghost must have passed through Twát, and probably also by Wanzamín. Now the point which satisfies the conditions of distance specified above, and at the same time best eludes surrounding difficulties, will be found to be not far from the 20th parallel of north latitude, and the 1st meridian of east longitude, or about 120 miles east of the position assigned to Aúdaghost in our map. Thus it appears that El Idrísí’s statements respecting the position of Aúdaghost, do not, when taken together, lend the slightest countenance to the hypothesis which makes that place identical with Aghades. For that position, deduced in the strictest possible manner from the assigned conditions, still leads to the conclusion that Ghánah was situate on the northern bend of the river of Tomboktú. But since we likewise learn from the same writer, that it was situate on the western, and not the eastern portion of that northern bend, we have no alternative but to correct his distances with respect to the angle of the river, and to remove Aúdaghost further west, so as to place it exactly half way between Aúlíl and Jermah; the correction, in this case, amounting to only a twenty-second part of the whole distance between those places.

It would be running into needless digression to point out all the contradictions in which El Idrísí involves himself by reducing distances so as to fit them to the frame in which he combines his information, or by expanding details so as to distribute them more equally. It will be here sufficient to have shown that he learned the course of the Great River from El Bekrí, yet that he did not copy his author faithfully, but took liberties with him, which are rendered more conspicuous by the incongruities into which they lead him. He contracts the Desert, spreads out the River; makes Silla on one side only sixteen days distant from the Ocean, and Kúghah, near Seghmárah, on the other, only a month from Dongolah. He wholly misunderstands the account of the lower portion of the river, and by turning the stream westward, he falls into glaring inconsistencies. In conclusion, whatever is reasonable in El Idrísí’s account of Ghánah and its vicinity, is taken from El Bekrí, and nearly all of it which is not taken from El Bekrí is absurd. Nevertheless, his statements, when carefully analysed and freed from misconceptions, plainly indicate that Ghánah was situate near that part of the Great River where Tomboktú now stands.[96]

The only novelty worth notice in El Idrísí’s account of Western Negroland, is his statement respecting the river of Ghánah, which he informs us was navigable in large boats, and flowed into the ocean. Its mouth was one day’s sail from Aúlíl. The river flowing into the ocean near the Bay of Arguin, where it has been shown that the isle and salt mine of Aúlíl were situate, is obviously the Senegal. The short distance of one day’s sail, allowed by the Arab geographer in this case, is in just proportion with all his other measures affecting the area of his map.[97] He supposed the Nile of Ghánah, or Great River of the Interior, to unite with the Senegal, and to run westwards into the ocean. Nor is there any rashness in ascribing to him so great a misconception. Leo Africanus makes a precisely similar statement respecting the river of Tomboktú. Having navigated that river from Tomboktú to Jenni, the latter author affirms most positively that it flows westward to the ocean. The only excuse that can be offered for Leo’s mistake is, that the part of the river with which he was practically acquainted, has little current, and shows no diminution of magnitude as it is ascended; to a careless observer, therefore, it presents nothing capable of controlling speculation, or guiding to a correct inference respecting the course of the stream. Higher up, the hypothesis was less tenable, and so Melli was placed on a branch of the river. Being biassed by the early Arab writers, particularly El Idrísí, Leo zealously adopted their erroneous opinion, which being repeated by De Barros and other writers on African geography, continued in vogue till the middle of the last century, or nearly six centuries after it was first promulgated.[98] El Idrísí states that salt was carried from Aúlíl to the mouth of the Nile, one day distant, and then up that stream to Silla, Ghánah and Kúghah. Modern authorities, on the other hand, have reported the Senegal to be navigable up to Jenni (two days from Silla), Tomboktú and Gago, at which point their information always terminated. And herein is another point of resemblance between Ghánah and Tomboktú; inasmuch as they hold similar positions in the hypothetical system, connecting the Senegal with the Great River of the Interior.[99]

The Western Desert is represented by El Idrísí with the changed aspect consequent on the movement of the Morabites. The Lumtúnah had gone northwards to Morocco, and the Benú Goddálah, to whatever quarter they had migrated, were no longer predominant in the south-western portion of the Ṣaḥrá. The inhospitable tract extending between the desert of Tíser and the Ocean, is named by El Idrísí Kamnúdíyah, the chief town of which was half way between Silla and Arkí. South of Kamnúdíyah, he places a country, the name of which, vitiated by copyists, occurs under a great variety of forms, as, for example, Maghráwah, Meghrárah, Meghzárah, &c. Of these readings, the first alone admits of a satisfactory explanation, and shall therefore be here adopted.[100] Maghráwah lay to the west of Ghánah, and as it extended from Aúlíl, on the sea shore, to Silla and Tekrúr inclusively, it also embraced a territory lying to the south of that State.[101] The country named by El Idrísí Maghráwah is therefore obviously the same which a century earlier had been occupied by the Benú Goddálah, and the change of its name may be naturally ascribed to the revolution which carried away the latter people with the hordes of the Morabites. The Arab historians are silent with respect to those who took the place of the Benú Goddálah; but the want of information may be in this instance supplied by a very probable conjecture.

When the Morabites, having subjugated Sús, Darʿah, Sijilmésah, and the province wherein they afterwards founded Morocco, still continued to press northwards, they met with a vigorous resistance from the Maghráwah, who had long ruled over Fez and its dependencies, and who now united with the Miknésah and other Zenátah tribes to oppose the invaders.[102] The victory fell to the Morabites, who entered Fez in triumph in A.D. 1067. But, grown negligent through continual success, they were soon after taken by surprise, overpowered, and expelled. Their enthusiasm, however, was not to be subdued by slight reverses; they returned to the struggle, and again entered Fez in 1069, slaughtering, it is said, 20,000 of the Maghráwah, whose sway in the west thus terminated, after a continuance of just a century.[103]

Of the fortunes of the defeated tribes, there is nothing recorded; but the general tenor of the history of Barbary justifies the supposition that they betook themselves to the desert.[104] In 1084, Yúsef ben Táshifín, the Amír or chief of the Morabites, sent messengers into the Ṣaḥrá, to the Lumtúnah, Goddálah, and Masúfah, announcing to them that he possessed extensive territories, well watered, which he was ready to bestow on the first comers; “and in a few days,” says the historian, “the whole land of Maghreb [Western Barbary and Morocco] was filled with colonists from the Lumtúnah and the other tribes of the desert.”[105] It is manifest that the Maghráwah, and their adherents, must have deserted the fine country around Fez, before the half-wild tribes of the Ṣaḥrá were called in to occupy it; and it is probable that, in the course of revolution, they stepped into the place of the Benú Goddálah soon after the latter had accepted the invitation sent to them to fill the vacancy left by the expelled tribes. Thus we are led to conclude, that the territory of the Benú Goddálah passed into the possession of the Maghráwah at a period subsequent to, and probably not far removed from, the year 1084.

The natural and probable supposition, that the tribes expelled from Mauritania by the Morabites changed places with the latter, and fixed themselves in the desert at the same time that their conquerors rushed into the occupation of the cultivated country, explains at once the great difference between the accounts given of the Western Desert by two authors, one of whom wrote seventeen years before the migration referred to, and the other sixty-nine years after that event. Nevertheless, the desire inherent in the human mind to give importance to whatever is obscure, favoured by the corruption of the name Maghráwah, and the garbled accounts of the country so designated, may revolt against a conjecture which confines that name to the Desert, instead of extending it over a large tract of Negroland. Yet El Idrísí plainly states that Maghráwah was a desert; that it was bounded on the north by the middle tract of the Ṣaḥrá, named Kamnúdíyah; and that it extended from Aúlíl, which was its capital, to Silla and Tekrúr; so that it must have been on the northern side of the Great River, of which the Senegal, according to his system, was a part.[106] The same writer indeed includes Silla and Tekrúr in Maghráwah, whence it may be inferred that the exiles from Fez soon obtained the ascendancy due to superior civilization, and became the rulers of their black neighbours. But since the Arabs nowhere mention the Mandingoes, Serakholies, Fúlahs, Wolofs, or other black nations between Silla and the ocean, it must be presumed that they had no direct intercourse with that part of Negroland, and knew nothing of it; and, besides, it is unreasonable to suppose that they described the country south of the Senegal under the general name of Maghráwah, of which general name, in any shape, not the least trace now remains in the region to which it is supposed to have been applied. North of the river, on the other hand, the disappearance of a particular tribe, or of its name; or a loss on the part of any tribe of that predominance which determines the name of a territory, is much more explicable. And if it be admitted that the name Aúlíl or Aúlílí was derived from that of Walílí, the chief place of the Maghráwah and the capital of Western Barbary under their dominion, then the presumption will arise, that some of that nation were always mingled with the Goddálah, and carried the local name to which they were attached, from the shores of the Mediterranean to those of the Ṣaḥrá.[107]

[Footnote 88: El Idrísí places Nún, or as he writes it, Núl, at a distance of three days from the sea, and thirteen (erroneously reduced to three by Hartmann) from Sijilmésah (Rec. de Voy. V. p. 205). But the town of Wád Nún is one or two days (22 miles) from the sea (Davidson’s Notes), twelve from Tatta, and sixteen from the chief town of Darʿah (Proc. of Afr. Assoc. I. p. 224), which is six days from Sijilmésah (Jackson’s Shabeeny, p. 3). The last-named place must, therefore, be twenty-two days from Wád Nún, and twenty-three, at least, from the sea. In like manner El Idrísí reduces to three and eight days respectively the distances of Sijilmésah from Darʿah and Aghmát, which El Bekri, confirmed by modern itineraries, estimates at six and fourteen days.]

[Footnote 89: Major Rennell, in his ‘Memoir on the rate of Travelling as performed by Caravans’ (Phil. Trans. Vol. LXXXI. p. 144), concludes that in Africa fourteen geographical miles and five-sixths of horizontal distance, is the mean daily rate of loaded caravans. M. Walckenaer (Recherches, &c. p. 266,) adopts fifteen geographical miles as the ordinary rate.]

[Footnote 90: From Cape St. Anne in the Bay of Arguin, to the mouth of the Senegal, is a distance of 260 nautical miles, or about forty hours of moderate sailing.]

[Footnote 91: Rec. de Voy. pp. 12, 206. Arkí (see Note 38) appears under various forms in the copies of El Idrísí; as Arḳi ارقى Rec. de Voy. pp. 12, 107; Azḳi ازقى; Azki ازكى Ibid. p. 206, and in the Abridgment. Ibn el Wardi also writes Azki, which, he says, is the place where travellers (ascending the Desert from Wád Nún) begin to climb the rocks (Not. et Extr. II. p. 23). Ibn Baṭúṭah travelled forty-five days from Sijilmésah to Aïwalátin, or Walata, whence Silla cannot be less than twenty days’ distant. From the latter place to Wád Nún is now reckoned a journey of fifty-five days (Davidson’s Notes, &c.).]

[Footnote 92: From the capital of Fezzán to Tomboktú is reckoned a journey of three months (Lyon, Travels in N. Afr. p. 144); or to calculate more accurately, from Morzúk to Twát is a distance of thirty- nine or forty days, and thence to Tomboktú, is a journey of forty-five or forty-seven days (Walckenaer, Rech. p. 423; Quart. Rev. No. 45, p. 230). If to the sum of eighty-four days thus found, be added the journey of two months, or rather two months and a half, between Tomboktú and the sea (see Note 41), we shall have for the distance between Fezzán and the ocean nearly three times the space assigned by El Idrísí; and, reducing the route to a straight line, with all possible allowance, more than double.]

[Footnote 93: The text stating the distance of Kúghah from Ghánah to be a month and a half, occurs in the Abridgment of El Idrísí (Hartmann’s Idrísí, p. 42), but is wanting in the larger work. There is good reason for believing it to be an interpolation. From Ghánah to Seghmárah, according to El Idrísí, was eighteen days; thence to Sámaḳanda eight days; and thence to Kúghah ten days. So that if these places be all arranged in a straight line from west to east, Kúghah will be still only thirty-six days from Ghánah, instead of forty-five. But while El Idrísí expressly traces eastwards the route from Ghánah to Seghmárah, and sets Kúghah eastward of Sámaḳanda, he says nothing of the bearing of Sámaḳanda from Seghmárah; so that we are at liberty to set the former place west by south from the latter, under the guidance of El Bekrí, whose Sámaḳanda was only four days from Ghánah. The two authors will then be found to coincide in general design, and El Idrísí’s Kúghah will be not above twenty days from Ghánah. It would be easy, were it worth while, to explain why the epitomator, trying to supply an apparent deficiency in his author’s ill-connected details, should have separated Kúghah from Ghánah by a month and a half’s journey.]

[Footnote 94: From Silla to Berísa, according to El Idrísí, was twelve days; thence to Ghánah twelve days; or to Aúdaghost twelve days; and between the two last-named places twelve days. This is a handsome arrangement. Again, from Ghánah to Tírḳa six days; thence to Merásah six days; thence to Seghmárah six days; thence to Ghúntil six days. Then come distances of eight, nine, ten, and eleven days. These numbers alone are enough to excite suspicion.]

[Footnote 95: The Berísa of El Idrísí is the same place of which the name is written in the copies of El Bekrí, Bersa برسى Not. et Extr. p. 647; Yerma يرمى (rather Yersa يرسى), Rec. de Voy. II. p. 6; and Yersana يرسنى MS. B.M. El Bekrí represents Ghúntil as a great country; he does not state the distance of Ghaïárú from it, but places the latter twelve miles from the river, and Bersana, or Yersana, west of it on the river. But, in El Idrísí, the corresponding names all designate towns which stand eleven or twelve days’ journey asunder.]

[Footnote 96: The map of El Idrísí does not represent the conceptions explained above. It makes the Great River divide at Tírḳa into two branches, so as to form a great island, which he names Wangárah. On the southern branch he places Ghúntil and Ghaïárú, the latter place being 75 days, or 2½ months from Aúlíl, measuring along the river, while only 3½ months at the utmost are allowed for the whole breadth of the continent. El Idrísí writes sometimes in conformity with one of these systems, sometimes with the other. He appears, on examination, to be an unsound author, who, with good materials before him, often wrote without understanding them.]

[Footnote 97: The distance of Aúlíl (in the Bay of Arguin) from Sijilmésah, as stated by El Idrísí, supposes, as we have seen, a mean daily journey of 22½ geographical miles, instead of 15, which is the ordinary rate. Now, if the forty hours’ sail from Arguin to the mouth of the Senegal, be reduced in the ratio of 22½ to 15, or 3 to 2, we shall have 26⅔ for the number of hours, according to El Idrísí’s scale. But there is no need of such exactness. It is highly improbable that the navigation spoken of by the Arab author, and which formed an ornament of his theory, had any existence even so far as it was within the limits of possibility.]

[Footnote 98: Labat (Ethiop. Occid. 1728, tom. II. p. 125) describes the course of the Niger or Senegal from the lake of Bornú to the sea. Moore (Travels in the Inland Parts of Africa, 1738) also maintained that the Senegal (of which the Gambia was supposed to be a branch) is the Niger, although he at the same time published the Journal of Capt. Stibbs, who was adverse to that opinion. An earlier writer tells us that “the English were frustrated in their attempts to ascend the Niger to the gold countries of Gago, by the osiers among other things.” (Charant, Réponses à diverses questions, &c. appended to Frejus, Voyage dans la Mauritanie, 1666.)]

[Footnote 99: El Idrísí says (Rec. de Voy. p. 11) that the salt of Aúlíl was carried up the river to Silla, Tekrúr, Berísa, Ghánah, Kúghah, and the other countries of the blacks. Let it be observed, that while propounding the hypothesis of a navigable river extending across Africa from the Western Ocean to Bornú, the Arab author knew nothing of its navigated course except between Silla and Kúghah or Kághó; that is to say, the generally navigated part of the river of Ghánah, which was evidently identical with that of the river of Tomboktú. The information of our early travellers respecting the Great River of the interior always terminated at Gago. This form of the name was taken from Leo; but Cadamosto had written Cochia (Kúghah), which was probably borrowed from the Mandingoes.]

[Footnote 100: To the usual various readings, Meghzárah, Meghrárah, Meghwárah, Meḳzárah, &c., M. De Humboldt (Histoire de Géographie, I. p. 291) has added Mufráda. D’Herbelot (under the title Maczarat) seems to consider this to be the name of a fortress, and not of an extensive region. In the Rec. de Voy. V. this name, where it first occurs (p. 10), is written Maghráwah مغراوة (afterwards changed into Maghzárah مغزارة); and the copy of Ibn el Wardi, in the possession of D. P. de Gayangos, has Maghráwah throughout.]

[Footnote 101: Rec. de Voy. V. pp. 10, 13, 18. In the first of the passages here cited, Maghráwah is represented as extending from Aúlíl to the Great River, and including also those countries which the author, by misconstruction of El Bekrí’s statements, brought into the vicinity of that part of the river. It is to be lamented that the Translation of El Idrísí’s Geography, published in the Rec. de Voy. swarms with false readings, against which little care has been taken to guard the reader.]

[Footnote 102: The Maghráwah مغراوة rose into importance about A.D. 945 (Marmol, I. fol. 127). Their name is written, by Leo Africanus, Magraoa; by Marmol, Magaraoa, or Magaraúa; by Moura (Historia dos Soberanos Mohametanos, &c. Lisbon, 1828) Magraua; and by Conde (Historia de la Dominacion de los Arabes, &c. Madrid, 1820) Magaraba and Magarava. Their original seat, according to Ibn Khaldún, was on the western side of the province of Afrikíah. They are evidently the Machurebii (Μαχυρήβιοι) of Ptolemy, who places them on the right bank of the Chinalaph or Shelíf, near Julia Cæsarea or Shershel, where Dr. Shaw (Travels, I. p. 56) still found an encampment of them; and also on the northern side of the Daradus, the modern Wádi Darʿah. El Idrísí, in giving their name to a country, only took the same liberty with it as with those of the Merásah, Seghmárah, &c. which he has converted into the names of towns. The appellation Maghráwat̤ es-Súdán, or of the Blacks, clearly intimates that there was another Maghráwah not on the borders of Negroland. To the scanty account of Maghráwat̤ es-Súdán given by El Idrísí, nothing has been added by later Arab writers save their mistakes.]

[Footnote 103: Moura, Hist. dos Soberanos Moham. p. 121.]

[Footnote 104: The Miknésah, with whom the Maghráwah were associated in their misfortunes, had formerly inhabited the Ṣaḥrá, whither they returned in their adversity (Marmol, I. 95; Conde, Dom. de los Arab. I. 411).]

[Footnote 105: Conde, II. pp. 99, 100.]

[Footnote 106: After naming the towns of Maghráwah, he adds (Rec. de Voy. p. 11) that the rest of the country bordering on the river is a sandy desert; and again, he says (p. 107) that the country between Kamnúdíyah and the river, that is, Maghráwah, is all desert. Hence Ibn el Wardi (Not. et Extr. II. 35) describes Maghráwah as an unfrequented and uninhabitable region. The same writer also observes, in a passage not translated by M. De Guisnes, that Maghráwah is the same country as Maghrebu-l-aḳṣa, or the Extreme West, a name certainly not applied to Negroland.]

[Footnote 107: Tanjah or Tangiers was anciently called Walílí. Another place of the same name, and of much celebrity, was situate near Fez. Let it be observed, that the name read in the text Aúlílí, may be also read Awalílí; and that there is some reason to suspect that nouns of race or nation are formed by prefixing aleph.]

MALI.—The Extinction of Ghánah.

The catastrophe which caused the disappearance of Ghánah from the political horizon of Negroland, is not distinctly described by any of the Arab historians. Nevertheless, so much light is thrown on the circumstances attending the extinction of that kingdom, in Ibn Khaldún’s sketch of the history of Málí, as may enable us to trace the course of those early events with tolerable precision. The statements of that valuable author shall be here given in his own words[108]:—

“When the conquest of the West (by the Arabs) was completed, and merchants began to penetrate into the interior, they saw no nation of the Blacks so mighty as Ghánah, the dominions of which extended westward as far as the Ocean. The King’s court was kept in the city of Ghánah, which, according to the author of the Book of Roger (El Idrísí), and the author of the Book of Roads and Realms (El Bekrí), is divided into two parts, standing on both banks of the Nile, and ranks among the largest and most populous cities of the world.[109]