salt mines of Aroan, ten journies to the NNW of Tombuctoo, and in the road to Morocco, were meant ; and from whence Tombuctoo is at present supplied. It is not easy to guess how an inland salt mine should have been inistaken for an island in the ocean; but it is certain, that both Edrisi and Abulfeda supposed the Niger to discharge itself into the sea near the meridian of Tombuctoo. Ibn Al Wardi* speaks of Oulili as the principal city of Soudan (or Nigritia) situated on the sea coast, and having extensive salt works, from which salt was carried to the other states of Nigritia. Mr. Park mentions the city of Walet, capital of Beero, which may, perhaps, be the Oulili intended by Ibn Al Wardi,† but it has no salt pits, for the inhabitants fetch salt from Shingarin, six journies to the northward of it ;‡ and Walet is more than 24 journies from Sala, instead of 16, as stated by Edrisi.

Cadamosta and Leo, in the third and fourth centuries after Edrisi. say that the people of Tombuctoo had their salt from Tegazza, 40 journies to the westward of that city; and that the salt was carried so far to the east as Melli, which is opposite to Kassina. By Tegazza, Tisheet,¶ the salt mine of Jarra seems to have been meant, but is far short of forty journies from Tombuctoo. Now, if in the twelfth century, salt was procurable so near to Tombuctoo as Aroan, or Shingarin (the salt pits of Walet) why should they have. fetched it from a place 30 or 40 days distant in the fifteenth and sixteenth? This requires explanation; for Edrisi states very particularly, that salt was carried from Ulil in boats along the Niger, and distributed amongst the nations on its banks from Sala to Kauga.††

Mr. Beaufoy, quoted as above, says, that there is a salt lake, or lakes in Bornou; from whence Agadez, Kassina, and certain states on the south of the Niger are supplied.This, at least, implies that there are no salt mines in the Desert, in the quarter east of Tombuctoo.

*Hartman's Edrisi, p. 29.

Oulili, Oualet.

Mr. Park's MSS.

Edrisi wrote in the twelfth century; Cadamosta in the fifteenth, and Leo in the sixteenth.

Mr. Park's MSS.

tt Edrisi, p. 7.

## Af. Assoc. 1790; Q. p. 157, 167; O. p. 236, 251.

Ff

CHAPTER VI.

The subject continued. Course of the river Niger at large; has no communication with the Nile. Ptolemy's description of it consistent.

THE course of the Niger (or Joliba) as we have seen, is established by ocular demonstration, as far as Silla, and may, I conceive, be admitted as far as Houssa, about 400 miles farther to the east, on the foundation of the information collected by Mr. Park; since it agrees with the ideas communicated to Mr. Beaufoy, by an intelligent Moorish merchant, who had navigated the river; and as it agrees no less with the report of Mr. Magrah, obtained from Moorish merchants at Tunis; and of Maj. Houghton from Bambouk. Thus the first 700 G. miles of its course are from west to east; or rather from WSW to ENE. There remains then, a space of more than double that distance, between Houssa and the nearest part of the Egyptian Nile, near Dongola; and yet more, to the known parts of the White river, or Abiad, the SW branch of the Nile.

I shall divide the matter, respecting the course of this river, into three heads. 1. Respecting the continuity of its waters, from Houssa on the west, to Wangara on the east ; without regard to the direction of the stream. 2. Respecting the positive direction of the stream. And 3. Concern

ing its termination.

1. Respecting the continuity of its Waters.

Edrisi gives the most positive information concerning the course of the Niger, or Nile of the Negroes, from east to west; deriving it from the same lake through which the Egyptian Nile passes; and describing it to terminate at 16 journics west of Sala (that is a little to the west of the position occupied by Tombuctoo) and near the supposed island of Ulil before mentioned.* He thus cuts off about 1000 miles of the breadth of Africa This was an error common to all the ancient geographers, as well as to those of Arabia; for Ptolemy places the mouth of the Senegal river

* Page 7 of Edrisi.

only two degrees more to the west, than Edrisi does that of the Niger.

Abulfeda believed, with Edrisi, that the Niger had a common source with the Nile, and ran westward *

It is certain, that these opinions furnish no proofs of continuity of course; but it may be supposed that there was some foundation for them; especially as Edrisi says, that salt was carried upon the Niger in boats from the island of Ulil, and distributed to the people on its banks, from Sala to Wangara, and Kauga.t

Mr. Matra was told‡ that from Kabra, the port of Tombuctoo," people sometimes travelled along the river the space of 40 days, to Ginny (Ghana) a large city," &c.

The Moorish merchant, with whom Mr. Beaufoy conversed, and whom he speaks of as a clear and intelligent man, says, "That the country of Guinea, or Ginny, is on the same river with Houssa." (Mr. Beaufoy's MSS.)

Edrisi, besides, mentioning the cities of Sala, Tokrur, Berissa, Ghana, and Ghanara, all of which he says are situated on the Niger, remarks that the country of Wangara, to which Ghanara belongs, is surrounded by that river, as it would appear by means of a subdivision of its waters: for Gatterer says, that Ghanara, one of its cities, stands on the western arm of the Guin, by which name he mentions the Niger; of which more presently. Now, as Wangara extends, according to Edrisi, 300 Arabic miles along the river,** this extent, together with the distance of Wangara from Ghana, eight journies,tt or 152 miles, make up 496 G. miles of the course of this river, eastward from Ghana ; which being itself 500 miles east of Houssa; there will be 969 miles in direct distance, traced eastward of Housa; or, on the whole, as Houssa is 700 miles below the source, about 1670 G miles of water-course from the head of the Niger, above Manding to the eastern extremity of Wangara!

In addition to these authorities, I may state from Leo, that the people of Tombuctoo convey their merchandize in boats (or rather canoes) to Ginea, by the Niger; and that

*Article Soudan.

† Edrisi, p. 7.

§ Edrisi, p. 11 and 12. Edrisi, p. 11.

Mr. Beufoy's MSS.

¶ Hartmann's Edrisi, p. 48, notes,
tib. p. 11.

at Kabra they embark for Melli also. But it is proper to be noticed, that he says (p 249) that this communication with Ginea takes place in the rainy season only (July, August, September) which would imply a deficiency of water for navigation at other seasons.* Leo, however, certainly never saw the Niger, although he seems to report himself an eye-witness of many particulars relating to itHis intelligence is, therefore, often to be suspected; though it has probably happened, that being regarded as an original author, instead of a compiler, he has given weight to the systems of Edrisi and Abulfeda, respecting the course of the Niger.

Gatterer, as I have hinted before, calls the Niger, Guin, as well at Tokrur and Ghana, as at Wangara.† Now we learn from Mr. Park, that the northern branch of the Niger, above Tombuctoo, passes by the town of Jinbala, and collect also from Labat, that it is named the river of Guin; and here we have the same name extended even to Wangar; a presumptive proof of the prolongation of the same river!

Edrisi speaks of the same Niger, or Nile of the Negroes,t also, at Kauga, 10 journies to the east of Wangara; from which we collect that he must have supposed, that this emanation of the Egyptian Nile, as he supposed it to be, first ran to the north and then turned to the west, through Nigritia. And if any consequence can be deduced from his account of the conveyance of salt along the Niger to Kau ga, where the catalogue of places supplied ends, we should conclude that he supposed the navigable part of the river ended at Kauga.

Although there can be no question that a river named Nile (or rather Neele) passes through the quarter of Kauga, Angimi, &c. since Edrisi, Abulfeda and Leo, speak of it, yet it would be advancing too far within the region of conjecture, in this place, to attempt to decide whether it has any communication with the western waters. I shall therefore reserve this discussion till the last, that it may not be allowed to have any weight in the decision of the

If this report of Leo has any particular meaning, and as the river in question carries a great body of water at all seasons, one must suppose that there are falls or rapids in the river when in its low state. Time may discover. Edrisi, p. 7 and 13.

Hartman, p. 32, 48, 51.

great question concerning the continuity and direction of the Niger. Having, therefore, as I conceive, established the fact of a continuation of the waters from Manding to Wangara, I shall next proceed to inquire into the authori ties for the direction of the stream.

2. The direction of the course of the Niger.

Ocular demonstration has shewn, that its course is to the eastward as far as Silla; and no reasonable doubt can be entertained that it continues the same course to Houssa, 400 miles farther to the eastward, even if the information communicated to Mr. Park could be doubted. For the Moorish merchant before quoted, told Mr. Beaufoy that he had himself descended the Joliba, from Kabra to Houssa, although he had forgot the exact number of days employed in the navigation; and whether it was 8 or 10 days; Mr Park was told 11. But one circumstance dwelt on his mind; which was, that "by the favour of a brisk wind, they returned to Kabra against the stream, in as short an interval as they went down." (This is no new fact to those who are accus tomed to inland navigations, even of the natural kind)

The same Moor added, "that from Houssa, going still with the stream, boats went to Jinnee* and Ghinea; near the latter of which was the sea, into which the Neel, or Niger, discharged itself." That this Ghinea lies to the castward of Houssa and Tombuctoo, has been already shewn and that at the distance of 40 land journies.

Edrisi says that the navigation from Ghana to Tirka, which latter is in the way to Wangara, admitted by the same authority to lie to the east of Ghanat, is with the stream of the Niger, and if this be true, it ought unquestionably to have the same direction all the way from Housa.

To these notices, of which the most full and positive is that of an intelligent person who had visited the spot, are to be opposed, the reports of Edrisi and Abulfeda, who wrote at a distance, and from the information of others. As to Leo, although his declaration is in favour of the two Ara

* It is certain that one city of Jinne, or Jinnee, stands above Tom buctoo and Houssa.

Edrisi, p. 9, 11, and 12.

Sionita, p. 12, translates the passage thus: "Via cursum Nili comitante." And Hartmann, p. 51, "Nilum sequere D'Herbolt understood the same thing; article Vankara.