extensive swamps, overgrown with rushes, which effectually put a stop to their further explorations. The coast of Eastern Africa was known as far south as Rhapta, an important emporium exporting ivory, rhinoceros horns, and tortoise-shell.

Unfortunately, with the exception of a " Periplus" of the Erythræan Sea, usually ascribed to Arrian, the reports of these travellers and explorers have only reached us in a fragmentary shape. Neither Strabo nor Pliny, to whom we are indebted for nearly all these fragments, have succeeded in reconciling the various authorities whom they were able to consult. The main facts, so far as they bear upon our subject, appear to be as follows:

The Nile, below Meroe, or rather below its confluence with the Atbara, was known as Siris. Our Atbara was represented by the Astaboras, said to have its source in a lake, and asserted, by Artemidoros, to bifurcate, and to send one of its arms into the Red Sea. This supposed eastern arm of the Astaboras is undoubtedly the modern Wadi Baraka, which has, however, no connection whatever with the Atbara. The Blue Nile was known anciently as Astasobas, or Astosapes, that is, the river of Soba, or Sape, a town identical with Ptolemy's Esar, the ruins of which still exist above Khartum. The Astapus, lastly, was the river which flowed along the western side of Meroe; and when Strabo tells us that this river was called by some Astasobas," he only gives expression to popular opinion, in accordance with which the "Blue Nile " was looked upon from the earliest times as the true head-stream of the Nile, the floods of which, produced by heavy rains in the mountains of Ethiopia, were the cause of Egypt's fertility. As to the White Nile, it must have been known, for we can scarcely doubt that it was this river which was explored by Nero's centurions; but its native name has not been placed on record, unless we suppose it to have been "Astapus." All these rivers are in a vague way supposed to have their sources in lakes, but the only lake mentioned by name is Strabo's Psebo, "above Meroe," which is usually identified with Lake Tsana, but may equally well be some other Abyssinian lake, such as Lake Haik.

Very little was known of the country just mentioned. In the time of the "Periplus" (100 A.D.) Greek merchants were in the habit of going from Adulis to a place called Koloe, an ivory mart, three days in the Interior, whence it was a five days' journey to the capital of the Auxumitae, our modern Axum. Koloe is represented by the modern village of Halai, at the edge of the plateau, and is referred to as Kole in ancient Ethiopian documents. Beyond Axum these adventurers crossed a river "Nile," in this instance clearly the upper Takazze, or Atbara, into a district called Cyeneum, which abounded in ivory. The mountains forming the eastern edge of Abyssinia are referred to by Agatharchides (116 B.C.) as the Pseboean mountains, a name recalling Strabo's Lake Psebo. The name of many tribes are mentioned. Of these I only need refer here to the Simbarri and Paluogges, because, owing to a similarity of name, they have been placed on the Upper Nile-quite erroneously, as I conceive as a careful perusal of Pliny's work shows distinctly that they were supposed to dwell between the Blue Nile and the Abyssinian

mountains. As to the "Pygmies," first mentioned by Aristotle, and placed by Pliny, that indiscriminate compiler, in the "marshes in which the Nile rises," they were not believed in by Strabo, and have found no place upon Ptolemy's map.

The coast outside the Red Sea, as far as Rhapta, is described with considerable detail in the anonymous Periplus of the Erythraean Sea, already referred to. Its author, unfortunately, says nothing about the interior of the country. It is clear, however, that the ivory-hunters established on the Afer and Somal coasts proceeded some distance inland, for Artemidoros of Ephesus (100 B.C.) mentions a salt sea and a freshwater lake inhabited by crocodiles and hippopotami, together with a river Isis, as existing in the Myrrh country, and tells us, moreover, that a river Nile rose farther to the east. It scarcely admits of doubt that the salt lake in question is identical with Lake Assal, at the back of the bay of Tajura; that the fresh-water lake still exists in Ausa; that the Isis is now known as Hawash, whilst the "Nile " is the upper Leopard River, which loses itself in a marsh, or lake, near the East Coast, not far from the mouth of the river Jub. Marinus of Tyre, the immediate predecessor of Ptolemy, confirms the existence of lakes near the East Coast of Africa, and one of these lakes he looked upon as the head of the Nile of Egypt. The maps and writings of Marinus have unfortunately been lost. All we know about them is derived from Ptolemy, who only refers to them for the purpose of subjecting them to a scathing criticism. The views of Marinus, however, appear to have influenced later writers, and more especially the Arabs.

PTOLEMY.-From what I have stated, it must be perfectly clear to you that the ancients, up to the days of Ptolemy, were absolutely ignorant of the fact that the main branch of the Nile rises in a vast lake lying under the Equator. The lakes vaguely referred to by them were evidently the lakes of Abyssinia or of the Maritime region of Eastern Africa. Ptolemy, however, with that uncompromising definitiveness almost inseparately connected with the mapping of geographical facts or hypotheses, derives the famous rivers from snow-clad mountains, and from two lakes which he places far to the south of the Equator.

When the existence of snow mountains and vast lakes in Equatorial East Africa first became known to us about half a century ago through inquiries instituted among the leaders of caravans who had visited the Interior, and when soon afterwards these mountains and lakes were actually found by European explorers to exist, geographers somewhat rashly saw in this a remarkable confirmation of Ptolemy's hypotheses. Dr. Beke, as early as 1846, looked for the "Mountains of the Moon "in Unyamwezi. Subsequently (1848) he assumed that they formed the eastern edge of the African plateau, and that the Nile had its sources on their inward slopes, a hypothesis on the strength of which he considered himself entitled to be looked upon as the "theoretical discoverer of the sources of the Nile." A few years afterwards (1861) he identified the Victoria Nyanza and the Tanganyika with Ptolemy's two Nile lakes. Sir R. Burton (1864) sought these Nile lakes in the Tanganyika, of which he himself was the discoverer, and in the Baringo, and he associated the

snow-clad Kilma-njaro with the Mountains of the Moon. After Baker's discovery of the Albert Nyanza, M. Berlioux (1874) identified that lake and the Victoria with Ptolemy's lakes, whilst Kenia and Kilima-njaro still represented the Mountains of the Moon. Still more recently, after the discovery of the snow-clad Ruwenzori, and the exploration of the Albert Edward, Mr. Stanley transferred Ptolemy's Lunar Mountains to that remote part of Africa, and this rather startling hypothesis has not been wanting supporters.

Almost the only modern critic who doubted Ptolemy's knowledge of Equatorial Africa has been Mr. W. D. Cooley (Ptolemy and the Nile, 1854). Mr. Cooley identifies the Astaboras with the Mareb, the Astapus with the Atbara, Strabo's Lake Psebo with Lake Tsana, and Ptolemy's Nile lakes with the marshes in which the Leopard or Haines River loses itself. Ptolemy, according to Mr. Cooley, knew nothing of the White Nile, whilst the sentence referring to the Mountains of the Moon "does not belong to the genuine text of his work."

After a careful study of the subject, I have arrived at conclusions identical, in several respects, with those put forward by Mr. Cooley. We cannot examine Ptolemy's map of the part in Africa here in question without perceiving that much information, which we know to have existed, and which must have been available at Alexandria, has not been utilised. Ptolemy does not appear to have consulted the anonymous Periplus of the Erythraean (witness his misplacements of Coloe and Menuthias); he knew nothing about the expedition which Nero sent to Meroe and the Upper Nile; nay, he cannot even have read such standard authors as Strabo and Pliny. On the other hand, his map embodies information not given by his predecessors, and most likely derived from the work of Marinus of Tyre; nor must we lose sight of the erroneous graduation adopted by the Alexandrian geographer, who assumed a degree of the Equator to measure 500 stades (or fifty geographical miles) instead of 600 stades. Wilberg and Cooley have sought to eliminate the errors arising from this mistake by overlaying Ptolemy's map with a corrected network of parallels and meridians. This proceeding, however, would have been admissible only if the whole of the map were based upon itinerary materials, not checked by occasional latitudes obtained by astronomical observations. Meroe, for instance, the latitude of which had been observed, occupies approximately its true position, whilst Adulis, which is dependent upon a dead-reckoning carried southward from Alexandria, is placed 3° 34' too far to the south. An amended graduation would give to the latter very nearly its true place upon the map, but it would shift Meroe more than two degrees to the northward, and correct in no wise the latitudinal difference between the two places, which would still amount to 4° 45', whilst in reality it does not reach 1° 45'. I have, therefore, discarded an amended graduation, but furnished Ptolemy's map with an amended scale instead, which enables us to measure the direct distances between places shown upon the map, quite irrespective of the degree-lines which cross it.

After these preliminary observations, I proceed to a more detailed examination of Ptolemy's map, and begin with Meroe, in order to identify,

as far as possible, the rivers and localities which have found a place upon it, and to ascertain whether there exists any connection between them and the Upper Nile as now known to us. The Astaboras, which joins the main Nile below Meroe, I do not hesitate to identify with our modern Atbara. The unnamed tributary of that river, which is shown on Ptolemy's map, but is not referred to in the text, may possibly be the Mareb, along the banks of which dwelt the Molibae. The Astapus is almost unanimously identified with the Blue Nile, and Lake Coloe, from which it issues, with Lake Tsana.1 Such an identification, however, presents very considerable difficulties. The island of Meroe, according to Ptolemy, included also the country of the Sebridae or Sembritae, whose capital was Eser or Sape (Bion ap. Pliny, c. 35). The position of Sape has been satisfactorily ascertained, and the country of the Sembritae extended undoubtedly along the Blue Nile. As Ptolemy's As Ptolemy's "Astapus enters the main Nile above Eser, we are bound to identify the latter, that is, the main Nile, with the modern Blue Nile, whilst the Astapus would be the Rahad or the Dender, which flow into that river from the Abyssinian highlands. Other features of Ptolemy's map favour the same hypothesis. The Ptomphanae, who, according to Pliny, had a dog for their king, must be looked for in Fazogl, where Brun-Rollet and Marno found customs surviving which explain the name given to this tribe by the ancients. The "Catadupi," or cataracts, which Ptolemy places on his main Nile, may possibly refer to the falls which the Abai forms after leaving Lake Tsana. As a matter of fact, all is confusion in this part of his map. Nor do we fare better if we approach Ptolemy's Astapus from the Red Sea. Three days from Adulis, so the Periplus tells us, we reach Coloe, a name found on Ptolemy's map in quite a different position. Five days beyond Coloe we reach Axum, and still farther west we cross the "Nile," that is, the Astapus of Ptolemy's map, or our modern Takazze. Mr. Cooley, it will thus be seen, was justified in identifying the Astapus with the Atbara, for the Takazze is the Upper Atbara, and I quite agree with him when he identifies Ptolemy's Pylaei Montes, which lie beyond that river, with the snow-clad mountains of Semen. But if the Astapus is not the Blue Nile, Lake Coloe cannot possibly represent Lake Tsana, from which that river rises. Fortunately Ptolemy himself enables us to give that lake a more correct position than that assigned to it upon his map. He places Lake Coloe to the south of Gabartus Mons, and in the Myrrh country. Gabartus is readily identified with the Jabarta of the Arab geographers, said to extend from Zeila to Shoa. If Gabartus Mons bet anywhere in Ifat, its position on Ptolemy's map is surprisingly correct, for he places it at a distance of 250 miles from Aualites (Zeila), the real distance being 225 miles. The lake is placed south of Gabartus Mons, and may possibly be the Zuway, which has, however, no connection whatever with the Nile. If we place the Coloe in the Myrrh region, which ncient authors unanimously describe as being situated around the head of the Gulf of Aden, it may be identical with one of the lakes at the back of Tajura Bay, in which case the river connected with it would be the

1 De Barros, Asia, 1552, first identified Lake Coloe with the Barcena or Lake Tsana.

Hawash! Our result then is as follows:-Ptolemy's Astapus is the Upper Astaboras (Atbara)—that is, the Takazze-and is supposed to send an arm into the main Nile south of the country of the Sembritae, whilst communicating at the same time with Lake Zuway, or perhaps with the river Hawash. The confusion could not be greater. Any notion that Ptolemy heard of the existence of his "Nile Lakes " from Meroe must be abandoned.

Such knowledge, if ever he possessed it, can only have been obtained on the East Coast, and that only from caravans which traded between Rhapta, the great emporium of Azania, and the Interior. Rhapta, which derives its name from the sewed boats employed all along the coast of Eastern Africa, as also in the Red Sea, has been identified by various authorities with quite a number of places. De Barros (1552) and Dr. Vincent placed it at Kilwa; D'Anville and Dr. Beke at the mouth of the Lufiji; Dr. Berlioux (1870) at that of the Pangani; and Mr. Cooley (1854) at that of the Jub. Most authorities, however, agree that it should be looked for somewhere in the bay facing Zanzibar Island.

This view certainly derives confirmation from the statements made in the Periplus, as also from an inspection of Ptolemy's very imperfect map. As an instance of these imperfections, I may mention that Ptolemy tells us (1. 17) that Rhapta was on a river not far from the sea, and close to a promontory bearing the same name; whilst, on his map, Rhaptum promontory is situated a hundred miles to the south-east of the river, and the town stands 75 miles above its mouth. Loose or contradictory statements, such as these, unfortunately disfigure much of what Ptolemy tells us about the East Coast of Africa. I believe, however, that Bagamoyo or the village at the mouth of the Kingani may fairly be assumed to occupy the site of ancient Rhapta.

In adjusting the information given in the Periplus and by Ptolemy to a modern map, it is of importance that the positions of Aromata promontory and Zingis peninsula or Opone should be ascertained. As to the former, it is beyond a doubt our modern Cape Guardafui, although Ptolemy places it in lat. 6° N. on his map, and in 4° 15', or 5° 41' N. in his text (Books I. and VIII.); whilst its correct lat. is 11° 50'. It is equally certain that Zingis peninsula is represented by our modern Ras Hafun, whilst the village of that name represents Opone. The distance between Opone (Ras Hafun) and Bagamoyo, following the sinuosities of the coast and touching at Pemba, amounts to 1255 geographical miles. A voyage from Opone to Rhapta (Bagamoyo) occupied, according to the Periplus, 25 days, or rather "courses," a day's and a night's sail being reckoned as two "courses." A day's sail would thus have amounted to about 500 stades or 50 miles. This is a reasonable distance. Ptolemy, when discussing the distances given by Marinus, admits that 400 or 500 stades daily would be a fair average; whilst Dr. Krapf, who travelled in 1843-4 from Cape Guardafui to Zanzibar in native craft, spent 24 "courses" upon his voyage, his daily progress having varied between 25 and 66 miles, and averaged 55 miles. Theophilus, one of the mariners cited by Ptolemy, performed the voyage from Rhapta to Aromata in 20 days; whilst Diogenes, sailing in the opposite direction, did so in 25.