three days’ journey from the capital. Toward the south it could not have
reached very far. The independent kingdom of Tekrúr was, at the utmost,
eighteen or twenty days distant in that direction. Still nearer was
Ṣínghánah, apparently an independent state, which carried on trade with
the Benú Goddálah. This trade was guided in its channel, as must always
be the case in the early stages of society, by natural circumstances. A
branch of the desert penetrates south-eastward to the very banks of the
Great River, in a tract of which we know not the exact width, but which
embraces the western shores of Lake Debú.[83] Such a road, laid open by
nature, could not fail to exert a great influence on the history of
Negroland. And indeed, the fact that Tekrúr, situate near that part of
the river, was the first converted of the Negro states; that the trade
of the Berbers occupying the sea shore at Aúlíl, and subsequently their
sway also, extended to the same quarter, might alone create a suspicion,
that the tribes of the desert found in that tract of country a nature
congenial to their habits. This suspicion is converted into certainty by
the narrative of Alexander Scott, who crossed the tract in question.[84]
The Berbers were actual possessors of territory south of Ghánah, where
the desert approached the Great River between that country and Tekrúr.
North of Ghánah, the dry desert of Tíser or Azawad was but eight or ten
days distant. In that direction, however, as well as towards the west,
the wilderness opposed no precise limits to the claims of empire, but
allowed pretensions of sovereignty to expatiate freely over territories
of two months’ journey in extent.
Thus we have seen that Ghánah was the frontier kingdom of the Blacks
contiguous to the advanced portion of the Great River at its north-west
angle; and extending in front of that portion of the desert, over which
lay the commerce with Sijilmésah,—a commerce guided by a principle,
which if not strictly unchangeable, at least fluctuates only within
narrow limits,—namely, that of choosing the shortest and safest route
across the desert. It comprised the country between Waláta and the Great
River, near the future site of Tomboktú, and enjoyed the identical
advantages of position which subsequently made the latter city so
prosperous.
But what were the revolutions, it may be asked, which caused Ghánah to
disappear? This question shall be fully answered hereafter; our inquiry
at present regards the place where Ghánah existed, and not the events
which led to its extinction. Yet it will not be alien from our purpose
to observe, that although the name, or rather title, of Ghánah became
politically extinct, and was erased from the list of sovereignties, yet
it still adhered obscurely, in the sixteenth century, to at least one
spot of the territory originally designated by it. For Marmol informs
us, that in his time Walata was also called Ganata; and that he did not
in this instance hazard an erudite conjecture, but spoke the plain
language of habit and experience, is evident as well from the
unostentatiousness of the remark, as from the frequency with which he
indifferently employs these two names one for the other.[85]
Let the reader now recal the account of Negroland, and of Ghánah in
particular, given in the preceding pages; let him fix his attention on
those features of the description pointed out for the purpose of showing
that Ghánah was near the site of Tomboktú; let him consider well that
those features have a magnitude incompatible with the supposition of
their being repeated, and a permanence derived from their dependence on
the physical constitution of the African continent. Let him, in fact,
figure to himself a great and navigable river, flowing from a town
called Silla north-eastwards for three weeks, through the country of the
Blacks who first embraced the Mohammedan faith, skirting the desert
eastwards for six days, and then turning southwards to a place called
Kaúkaú, or Kághó; let him place the emporium of Negroland near the
north-western angle of that river, at a distance of two months’ ordinary
travelling from the shores of the Atlantic, two months from Sijilmésah,
and fifty days from Tádmekkah, not far from the modern Aghades. He may
then trace the road from Sijilmésah to that emporium, dividing the whole
distance into its distinct portions, viz.—eleven days south-westward to
the border of the desert, then six days over the hills, about seventeen
more to the zone of drifting sand, passing near the salt mines of
Tagháza, and eight or ten over an utterly inhospitable tract near the
southern limit of the Ṣaḥrá. Along this road he may distribute the tents
of the wandering Masúfah; and, a little to the east of it, he may mark
the boundary line of the great Berber nation, the Zenágah. Let him then
write above Silla, on the left hand, this remark:—“Trade carried on with
the Berbers on the sea shore;” and below Kaúkaú, on the right,—“Obscure
and little known:”[86] and when, having finished this delineation, he
finds that, though drawn in conformity with the descriptions of Ghánah,
it is yet perfectly applicable to Tomboktú; and that it is equally true
and faithful, whichever of these names be given to the emporium of the
Blacks: then, even if he throw aside all other considerations, such as
the relations of Ghánah with the Morabites and with Mímah, the town
whence Tomboktú derived its Berber population, he certainly cannot
refuse to admit that the Ghánah of Arab writers was contiguous to that
part of the Great River where Tomboktú now stands.
In the sixteenth and seventeenth centuries, when numerous accounts of
Barbary were published in Europe, and when the trade of Morocco and
Táfílélt (the ancient Sijilmésah) with Negroland was highly rated, we
find that, besides Tomboktú, whither the routes from Táfílélt and Wád
Nún conducted, no places of any importance in Negroland were mentioned,
except Jenni near Silla, and Gago, which is the same as Kaúkaú. The
caravans across the desert directed their march to the nearest point of
Negroland, and the merchants, arriving there, never looked beyond the
marts with which they could thence maintain a direct intercourse. In the
same manner, and exactly within the same limits, was El Bekrí’s
information circumscribed. He says nothing of the Mandingoes,
Serakholies, Wolofs, Fellatah, or other black nations of the west. He is
equally silent respecting Houssa, toward the south-east. Of Kánem, which
was reckoned to be only forty days distant from Zawílah, he speaks in
terms indicative of the faintness of his knowledge:—“The people,” he
says, “are blacks and idolators; and the country is hardly ever visited
by travellers.”—Thus the analogous modes in which Ghánah and Tomboktú
present themselves respectively to the ancient and modern historians of
Negroland, corroborate the opinion that the site of the former of those
places was in the vicinity of the latter.[87]
In the attempt here made to demonstrate that Ghánah was near the site of
Tomboktú, it will be perceived that no aid has been sought from
etymologies or fancied resemblance of names; much less have such
delusive lights been allowed to guide our investigations. The arguments
urged in the foregoing pages rest wholly on necessary deductions from
the obvious sense of our Arab authorities, without any attempt to wrest
their meaning. The topics touched on in those arguments may be thus
briefly recapitulated:
The description of Ghánah in respect to its bearing and distance from
Sijilmésah—the details of the road to it, and characteristics of the
desert—the relations of Ghánah with the Zenágah, and particularly with
the Morabites—its intercourse with the tribes on the sea shore—the
position of Genéwah—the course of the River—the relations of Ghánah with
Mimah, Másín or Masena, Anbárah, &c.—the name Ganata remaining to
Walata—the laws and usages of Ghánah.
There still remains an argument of no common weight, the substance of
which, though belonging to another part of this inquiry, yet may,
without impropriety, be briefly stated here. The Blacks of the country
named Málí, who, it will be seen, were Mandingoes, issuing from the
south-west, conquered Ghánah. Their empire (to use the words of their
historian) extended from Ghánah in the east, to the ocean in the west.
Nearly a century after that conquest, they advanced eastward (or rather
south-eastward) to Kaúkaú, whence they marched into the desert and made
themselves masters of Tekaddá. They subsequently relinquished, however,
the possession of that place, and retired to Kaúkaú; so that the Great
River formed the boundary of their sway towards the east, with the
exception, perhaps, of a short space below Kaúkaú. Now there can be no
uncertainty respecting the ground over which these conquerors marched.
The celebrated traveller, Ibn Baṭúṭah, visited, in the fourteenth
century, the capital of Málí, on the Joliba above Sego; he thence
travelled to Tomboktú; thence descended the river to Kaúkaú, and from
that place he went to Tekaddá, which was on the road from Kaúkaú to Ghát
and Ghodémis. He thus appears to have followed exactly the course which
had been taken by the Mandingo conquerors; the only difference in his
described route being, that he found Tomboktú where they had found
Ghánah.
* * * * *
[Footnote 5: Chénier (Recherches sur les Maures, tom. III. pp. 16 and
79) was the first to pronounce, Walckenaer (Recherches sur l’Afrique
Septentrionale, 1822, p. 285) the first to prove, the identity of
Sijilmésah سجلماسه with Táfílélt تافيلالت. The arguments of the latter
amount to demonstration, and need no reinforcement. Yet we may be
permitted to add, that all the Arab writers, without exception, make
Sijilmésah contiguous to Darʿah درعه; and that the uncritical Marmol,
although he subjoins to his description of Sijilmésah (vol. III. fol. 8)
an account also of Táfílélt, denuded of every circumstance which could
help to determine its geographical position, yet in copying Leo’s list
of the provinces of Numidia, omits Sijilmésah, and substitutes for it
Táfílélt (Leo Africanus, pt. I. c. 5. in Ramusio, 1554, vol. I. fol. 1
V, Marmol, I. fol. 12). The name Táfílélt seems to have come into use
with the rise of the dynasty of the Fílélí sherífs. The tribe, ennobled
by events, gave its name to the country in which it was established. M.
Gråberg af Hemsö, therefore (Specchio di Marocco, 1834, p. 65), who
refuses his assent to M. Walckenaer’s conclusions, and separates
Táfílélt from Sijilmésah for reasons of comparatively little weight,
cannot certainly vindicate his mode of using the former of these names,
by showing that it occurs in the pages of any historian, native or
foreign, anterior to the sixteenth century. The Fílélí tribe or family
are however of ancient standing, for Ibn Baṭúṭah informs us that at
Kaúkaú, in Negroland, in A.D. 1353, he became acquainted with the faḳíh
or doctor, Mohammed the Fílélí الفيلالي.]
[Footnote 6: Ghánah غَانَة and غَانَةَ (MS. B.M.). The final _hé_ ﻪ of
the Arabs, when pointed ﺔ, is pronounced as _t_ before a vowel; غَانَة
and غَانَةَ therefore, when not immediately followed by consonants, are
read Ghánat̤ and Ghánat̤a. But the suppression of the _t_ in this
instance seems to be due altogether to the analogies of the Arabic
language, and probably was not always imitated by the Berbers, in which
language _t_ is a frequent termination. Hence it is not surprising that
in an extract from El Bekrí, sent from Tripoli by M. Gråberg af Hemsö,
to M. Jaubert (Recueil, &c. par la Soc. de Géogr. tom. II. 1825), we
should find Ghánat غَانَتْ constantly written instead of Ghánah غانه.
The importance of this remark will appear hereafter.]
[Footnote 7: The Arabic _wa_ و when it begins a word, is a consonant,
like our _w_. Hence, when the Arabs would write a name beginning with a
long _o_ or _u_, they are obliged to prefix an aleph ا to the wa و to
preserve to the latter its vocal function; thus اودغست, اوليل, اوكار,
would be written to express Odaghost or Udagost, Olíl or Ulíl, Okár or
Ukár. The áú او may be also intended for a diphthong. But it must be
observed that the prefixed aleph in the Berber language is a sign of
case, and may have other offices; there is some temerity therefore in
excluding the form Awadagost, and in reading Berber names according to
the analogies of a foreign language.]
[Footnote 8: Not. et Extr. p. 642. The reasons for concluding that Aúkár
lay to the south-west of Aúdaghost, will be shown hereafter. El Bekrí
states more than once that Ghánah was the king’s title. Not. et Extr.
pp. 630 and 642.]
[Footnote 9: Támedelt تَامدَلتْ (MS. B.M.) is read by M. Quatremère,
Tamdoult. Sús el Aḳṣa, i.e. the remote Sús, is the most southern
province of Morocco, on the confines of the desert.]
[Footnote 10: The routes to Ghánah and Aúdaghost went southward,
according to Abulfedá, but this expression need not be strictly
understood. (Abulfedá’s Geography, translated by Reiske in Büsching’s
Magazin, vol. IV. 212, V. 354). Tomboktú is here spelt as dictated by
Ibn Baṭúṭah. The river of Negroland, which, in the successive parts of
its course is named Joliba, Issa, Quorra, &c., and which theory and
false learning have styled the Niger, will be generally denominated,
throughout the following pages, _the Great River_.]
[Footnote 11: Caillié reckoned fifty-seven days of actual travelling
between Tomboktú and Táfílélt.]
[Footnote 12: Mabrúk is said by some (Itinerary of Ḥáj Ḳásim, in
Walckenaer, Rech. p. 426) to be eleven days, by others (Mohammed of
Tripoli, in the Quarterly Review, No. 45, p. 231) fifteen from
Tomboktú.]
[Footnote 13: Since the intercourse between Fez and Támedelt need not be
considered as merely commercial, it would be injudicious to measure the
routes between them by the journeys of a loaded caravan. El Bekrí (Not.
et Extr. p. 598) reckons eight days’ journey between Fez and Sijilmésah,
but one of them was a long journey (across the desert of Angad) of sixty
miles. Abulfedá makes the distance between these cities to be ten days’
journey. We reckon nine days, so as make the whole distance of Fez from
Támedelt twenty days’ journey, and allow twenty-two geographical miles
to each. The bearing of Támedelt from Iklí is said to have been in the
direction of the Ḳibla بقبلي (MS. B.M. fol. 105), or temple of Mekkah,
towards which the followers of Mohammed turn their faces when they pray.
The direction of the Ḳibla is understood in Morocco, according to Windus
(Journey to Morocco, p. 49) to mean east by south.]
[Footnote 14: The caravans from Sijilmésah to Ghánah, like those to
Tomboktú, assembled on the confines of Sús, Darʿah, and the desert.
Tatta, the gathering place of the merchants going to Tomboktú, as placed
by Major Rennell, on information derived from the British consul at
Mogadore (Proceedings of the African Assoc. 1810, vol. I. p. 254), is
not more than two days distant from the site of Támedelt; and El Harib,
the point at which Caillié, travelling northward, quitted the desert,
was but four or five days distant from these places.]
[Footnote 15: Bír el Jemmálín بير الجمالين (MS. B.M. fol. 101), the
Camel-keepers’ well. M. Quatremère (Not. et Extr. p. 612) reads Bír el
Ḥammálín, that is, the Porters’ well. But as this name occurs in the
route from Wádi Darʿah, it is fortunate that a pointed MS. enables us to
distinguish clearly between those two places, the proximity of which
would conduce to the embarrassment likely to result from their being
confounded together.]
[Footnote 16: Azawwar أزَوَّرْ MS. B.M. fol. 102; Azour in Not. et Extr.
p. 613. This name may, with much probability, be read Azawwad ازود, that
is, the dry or sterile country.]
[Footnote 17: Tendefas تندفس MS. B.M. fol. 102; Tendefak, Not. et Extr.
p. 613. Weínhílún وَيْنهيْلُون MS. B.M.; Wirhaloun, Not. et Extr.]
[Footnote 18: Tázḳa تازقَي MS. B.M.; Tarka, Not. et Extr. This word,
written _Taskha_ by Capt. Lyon (Travels in North Africa, p. 315), is
still retained in the dialect of the Tawárik.]
[Footnote 19: Weíṭúnán ويْطُونان MS. B.M.—Aúkázenta اوكازنْتَ MS. B.M.;
Oukarit, Not. et Extr.]
[Footnote 20: Sidi Hamed, leaving Wád Nún, went six days round the
mountains towards the south, that is, he cleared the hills on the sixth
day. He then travelled fifteen days over hard ground, on which the
camels left no trace; then three days on hard sand, and then entered the
hills of drifting sand. Riley’s Narrative of the Shipwreck of the brig
Commerce, p. 322.]
[Footnote 21: The Harib of Caillié, who was not fortunate in seizing the
sounds of the Arabic and Berber languages, ought probably to be Gharíb
غريب M. D’Avezac (Bulletin de la Soc. de Géog. 1834, tom. II. p. 169)
proposes reading ʿArib; but since the tribe of the Gharíb is mentioned
in M. Gråberg af Hemsö’s list of Berber names (Journal Roy. Geog. Soc.
vol. VII. p. 255) and Marmol (tom. III. fol. 9,) places a tribe named
Garib in the neighbourhood of Tatta, we cannot avoid concluding that the
French traveller means to speak of the same tribe as the last-named
author, and that its true name is El Gharíb. But in the map drawn by M.
Jomard to illustrate Caillié’s journey, the position of El Harib is even
more faulty than its orthography. It detracts little from Caillié’s
merit to say, that under all the circumstances of his journey his
observations of the compass were worth but little, and his estimates of
distance are not to be implicitly relied on. Nineteen miles a day for
the average of forty-three days between Tomboktú and El Harib, and
twenty-nine miles daily during the worst part of the journey, over deep
and burning sands, are rates of travelling much too high for a loaded
caravan. By the undue lengthening of the early part of the route, El
Harib has been carried about fifty miles too far north, so that Tatta,
instead of being north-west of it, according to the traveller’s text, is
made to lie to the south-west. All the other bearings described are in
like manner displaced. M. D’Avezac has corrected this error of latitude,
but has, at the same time, unfortunately introduced a new error of
longitude, and carried all his positions too far eastward.]
[Footnote 22: Quart. Rev. No. 75, 1828, p. 102; Edinburgh Phil. Journal,
vol. IV. p. 42.]
[Footnote 23: Wanzamín ونزمين MS. B.M.; Wabermin, Not. et Extr. p. 614.]
[Footnote 24: Wárán وارَان. The name of the inhabitants of this part of
the desert has been read by M. Quatremère, Benú Hareth, instead of Benú
Wareth وارث; but the restoration of this name is important; since we
learn from another passage in El Bekrí that the Benú Wareth were to the
east of the Lumtúnah; and are thus enabled to perceive that the road to
Aúdaghost lay eastwards from the road to Ghánah.—Agharef أَغَرَفْ MS.
B.M. 102 v.—Akríri اقرٮري Not. et Extr. p. 615. اقرٮٮدى MS. B.M.]
[Footnote 25: Azjúnán (pronounced Azgúnán) ازجونان MS. B.M.; Arkounat,
Not. et Extr.]
[Footnote 26: In Nubia showers of rain are of rare occurrence north of
the 18th parallel (Rüppell, Reisen in Nubien, &c. p. 75). Denham
(Travels, &c. I. p. 164) fixed the first appearance of fresh vegetation
and the limit of the tropical rains on his route to Bornú, near the 16th
parallel, which is probably too low for the limit of rain. Tomboktú, we
are told, has annually six weeks or more of rain (Proc. Afr. Assoc. I.
p. 285; Narrative of R. Adams, p. 42)—and this blessing seems to extend
some distance north-eastward of it (Riley’s Narrative, p. 346). Towards
the coast the gum forests which extend as far north as the 18th degree
of latitude probably do not fall far short of the limit of regular
rains.]
[Footnote 27: Tarḳá ترقا Not. et Extr. p. 623; Tárga تارجي MS. B.M. 105
r.—Tezámt تزامت MS. B.M. 105 v; Baramet, Not. et Extr. p. 624.—Bir el
Ḥammálín بير الحمالين—Máleki مالكي Not. et Extr. p. 264; Nálelli ناللِّي
MS. B.M. 105 v.]
[Footnote 28: In the Parisian MS. the expression is “the mountains, the
name of which, in Berber, signifies _the mountains of iron_.” Not. et
Extr. 624. But the MS. B.M. gives the Berber name Adarérén wazzél
ءَاَدَرَارَانْ وَزّالْ in which Adarérén is the plural of Adrar, a
mountain, and wazzél, iron, corresponds with the _ouzail_ of Shaw’s
vocabulary (Travels in Barbary, II. p. 382).]
[Footnote 29: Yentesír ينْتسِير, MS. B.M.; Belis, Not. et Extr.—Moddúken
مُدُّوكَن MS. B.M.; Merouken, Not. et Extr. From the village of
Moddúken, which belonged to the Zenágah, to the city of Ghánah مَدَينة
غَانة was a journey of only four days. But for Ghánah in this place M.
Quatremère proposes reading Akka, عاقه, being perplexed by the
faultiness of his MS. which seems to conduct beyond Ghánah to the
country of the Lumtúnah. The MS. B.M. p. 105 v, clears up the difficulty
by these words; “and from the wells before mentioned, (viz. the wells of
El Ḥammálín and Nálelli,) the water is carried a four days’ journey to
Mount Aízal or Izal ايزَلْ in the desert,” &c. Thus the road to the
desert (the Lumtúnah not being named in this passage) does not begin
from Ghánah, but from “the before-mentioned wells.” Jebel Aízal may be
suspected of being another form of Adarérén Wazzél.]
[Footnote 30: The name Tíser تيسر (Jaubert’s Idrísí. Recueil de Mém. &c.
tom. V. p. 106) is extremely doubtful. Some of the MS. copies of Idrísí
have Níser نيسر, others Nesír نسير. The epitome offers Bansar بنسر;
Abulfedá writes Yasr يسر.]
[Footnote 31: El Bekrí (Not. et Extr. 624) gives to the absolutely
waterless desert between Sijilmésah and Ghánah, an extent of eight days’
journey. Sheríshí, in his commentary on Harírí (Maḳámah 9) gives it a
width of ten days. Ibn el Wardi increases it to twelve days, and El
Idrísí (Jaubert’s Transl. p. 106) to fourteen. It was natural enough
that the first of these writers, who may be supposed to have derived
much of his information from the early Morabites themselves, should
receive a less exaggerated account of the inhospitable nature of the
western desert than those who followed him.]
[Footnote 32: The road to Aúdaghost passed through the territory of the
Benú Wáreth. But this tribe were to the east of the Lumtúnah, through
whose country was traced the road to Ghánah (see note 24). And Ghánah
was at least four days south of the desert of Tíser, while Aúdaghost was
east of that desert, according to Abulfedá (Büsching’s Magazin, vol. IV.
p. 212). El Idrísí also places Aúdaghost in the northern part of the
kingdom of Ghánah; and by stating its distance from Wergelán and Jermah,
he plainly intimates that it was likewise in the eastern part. All these
particulars combine to prove that Aúdaghost was to the north-east of
Ghánah.]
[Footnote 33: Ibn Baṭúṭah travelled at the slow rate of a heavily laden
caravan. Halts included, he was two months in reaching Aïwalátin, or
Walata, on the southern border of the desert. His accounts of the pools
of Táserahlá which shall be given hereafter, exactly correspond with El
Idrísí’s description of the wells of Tíser. In Jaubert’s ‘Idrísí’ (p.
11), is the following passage: “Il y existe cependant des mares d’eau de
pluie qu’on rencontre après deux, quatre, cinq ou douze journées de
marche, semblables à celle du désert situé sur la route de Sedjelmasa à
Ghana, et où l’on ne trouve de l’eau qu’au bout de quatorze jours de
marche.” Instead of the word _desert_ in this place, the Epitome of El
Idrísí has the name Bansar بنسر, evidently for Tíser تيسر.]
[Footnote 34: Leo Afr. pt. VI. c. 54. Marmol writes Azaoat. Beyond, or
eastward of the desert of the Zenágah, Leo places that of the Zuenziga,
“which extends from Segelmesse, Tebelbelt, and Benigorai, to the desert
of Ghir, in the south, which faces the kingdom of Gubar. On the west it
has Tegaza, and on the east the desert of Air, inhabited by the Targa
tribe (the Tawárik).” It is manifest that the several divisions of the
desert described by Leo (pt. VI. c. 54-8), all extend from north-west to
south-east, conformably to the boundary line which we have ascribed to
the country of the Zenágah on the east. It is plain also, that these
people were not in the vicinity of the Houssa country. The desert of
Ghir brings to mind the Káhir كاهر of Ibn Baṭúṭah; but we can have no
doubt that the Air of Leo is the country of Ahír (Ḥáj Ḳásim in
Walckenaer, Rech. p. 448) or Aáheer (ʿAhír?), which we learn from Sultan
Bello (Denham’s Travels, II. 447, where Aáheer is erroneously said to be
south of Bornú) to be the portion of the desert lying north of Houssa
and Bornú. But it appears that the name in question has extended further
southwards since Leo’s time, a proof that the Tawárik have been gaining
ground.]