ated in that quarter of the globe; and when the expected island was not found, it was easier to believe it swallowed up by the sea than to admit that it never had any but a fabled existence.

In the accounts which the writings of the Hindoos give of the Indian seas, the same tendency to dispose every thing according to the symmetry of a religious system prevails over the simplicity of truth. In this quarter, also, there is a tra-nate, or group of three islands, composed respectively of gold, silver, and iron. Co-existing with these, there is a mysterious assemblage of seven islands; and when Jambolo stated that the Indian islands were seven in number, he only repeated the language of the natives; for the expression Yail Lancas, or the seven lancas, is still in use at the present day. It is obvious that these legends were all understood literally by the Greek geographers, and particularly by Ptolemy, who derived an unusually large share of information from India, and was careful to turn every atom of it to account. Accordingly, we find in his map a Heptanesia nesos, or Septuple island, which it is impossible to assign to any known position; and also a Tricadiba, or Trinity island, with a Tricanesia nesos, or translation of the former name, after his custom. His islands of satyrs, of monkeys, &c. all stand ranged in triple order. Then the countries of gold, of silver, and of brass, (this last being substituted for iron, according to the Greek notions of precedence among metals,) all belong evidently to the domains of fable, although posterity, always credulous when gold is in question, has never ventured to dispute the reality of their exist

ence.

But Ptolemy committed a graver error when he adopted the opinion of Hindoo geography, which unites the eastern peninsula of India with the African continent. Arrian, the author of the Periplus, although he believed it possible to sail round Africa into the Atlantic, was yet evidently impressed with the Hindoo notions, for he says, "it is believed that Taprobane approaches

very close to the coast of Africa." The south-eastern coast of Africa is named in Hindoo writings Sanc'hadwipa (Zanguebar), that is, the island of shells, an expression equivalent to the country of the Troglodytes; and as this Troglodytic region is supposed to extend round to the south-eastern extremity of India, many errors arose from the use of this equivocal language, which have found their way into the writings of the Greek and Latin geographers. With Anga-dwipa,

or China, the ancient writings of the Hindoos show but little acquaintance. "No great men, famous and learned among Bipeds, ever visited the island called Bhadrásvá, where there is a wondrous Canadab tree, with flowers like great water-pots." Yet silk, the peculiar production of China, found its way into India at a very early period. Silken garments worn by the queens at a time of festal display, are mentioned in the Ramâyuna, of Valmeeki, a poem written one thousand years at least before the commencement of our era.

There still remains a story transmitted to us by a Greek writer, which, from the indulgence shown to it by some eminent modern geographers, deserves to be compared with the legends of the Hindoos. Evemerus, a Macedonian, is said to have discovered, to the south of Arabia, a group of islands, three in number.* The largest of them, called Panchæa, was inhabited by four nations, one of which was ruled by three kings, who were nevertheless controlled by the college of priests. A magnificent temple there was covered with Egyptian hieroglyphs and inscriptions. Three cities adorned this terrestrial paradise. One of the lesser islands produced frankincense enough to supply the altars of all the gods in the world. In fine, Panchaa was the country of the Phoenix, and the island of the Triphyllian Jove. It is obvious that Evemerus derived this relation from the Egyptian priests, who debased the mythic tri-cuta of the Hindoos by attempting to unite it to facts. Just as Hecateus placed the Hyperboreans in Britain (an island

* Diodorus.

opposite to Gaul) amidst memorials of the Greeks, so the Egyptians appear to have fixed the wandering islands of the East, and to have added to the general principle of a triple Elysium the hieroglyphs and other particularities derived from their peculiar modes of thinking. There seems but little reason, therefore, to hope, as some of the learned have done*, that these blissful islands may be again discovered on the coasts of Africa or Arabia.

The persuasion that the dwelling-place of happiness is in the West, may have exercised an important influence on the early migrations of mankind. The existence and the wide diffusion of such an opinion are clearly established; nor is there any reason to believe that it was grounded in positive tradition. But then it will be asked, why was Paradise supposed to be in the West? An answer to this question may be found in the constitution of the human being, who is always more disposed to receive profound impressions at the hour when the natural day is coming to a close, and contemplates with the finest sensibilities that most glorious of celestial phenomena, the setting sun. † The Hindoos retain to the present day their old belief. The chalk with which the Brahmins mark their foreheads is from the West: they even pretend that it is brought over land from Britain. Yogees, followed by their trains of pilgrims, have attempted in modern times to reach the Hyperborean regions across Europe, and have even advanced as far as Russia; but the importunate curiosity by which they were assailed effectually subdued in every instance their piety and courage.

In the West the primitive tradition is still remembered. The lakes and seas of Scotland and of Ireland have all their floating and holy islands. The Inis Wen, or white island of the Gaels, and the Ynys y Cedeirn, or island of the mighty ones of the Welsh, are still objects of hope and veneration. The most westerly group of the

*Malte-Brun, Geogr. vol. i.

Gosselin,

The first hints on the natural principles of mythology are to be found in Heyne's Opuscula Academica.

Hebrides, the Flannan islands, which are devoutly believed to be seven in number, and are even laid down as seven in our maps, though only six are visible to the eyes of the sceptical*, are said to have the virtue of disposing to prayer and religious meditation all who land upon them. The Arran islands, on the west of Ireland, are entitled the isles of the living, that is to say, of those who have returned to life; but the language of this general superstition was carried far beyond the shores of Europe. It is found among the Indians of North America, who fervently believe in the existence of a land of happiness in the West beyond the ocean; but whether this tradition belongs to them originally, or was introduced among them by the Scandinavian adventurers of the eleventh and twelfth centuries, it is impossible to determine.

* Macculloch's Western Isles,

+ Martin's account of the Hebrides.

BOOK II.

GEOGRAPHY OF THE MIDDLE AGES.

CHAP. I.

THE ARABIANS.

ITINERARIES OF THE ROMANS. PEUTINGERIAN TABLE.-COSMAS INDICOPLEUSTES. -THE WORDS SINT AND HINDOO. THE TSINITZE. --THE ARABIANS. -THEIR CONQUESTS. -AND COMMERCE. EARLY TRAVELLERS INTO CHINA.-EDUCATION OF THE CHINESE. REGULATIONS OF THEIR PORTS. - FIRST MENTION MADE OF TEA. CHINESE EAT HUMAN FLESH. STRANGERS IN CHINA. CANFU. USES OF THE COCOA-NUT TREE. KINGS OF INDIA. — THE UNICORN. SOGDIANA.-THE ALANS. KHAZARS. THE FOSSIL IVORY OF BULGAR.COMMERCE OF THE ARABIANS WITH THE NORTH OF EUROPE. THE INTERIOR OF AFRICA COLONISED BY ARABIANS. - KINGDOMS OF GHANA AND TOCRUR.-LAMLAM.— THE ZINGES AND WACWAC. THE PERPETUAL ISLANDS. OTHER ISLANDS IN THE SEA OF DARKNESS. VOYAGE OF THE ALMEGRURIM.

THE geography of the ancients may be considered as having attained its greatest accuracy and comprehensiveness in the age of Ptolemy. The disorders which soon after followed in the Roman empire, the frequent migrations of the northern nations, and the invasions of the Goths and Huns, completely changed the geography of Europe, and rendered it difficult to procure any positive details amidst the disorders of so many revolutions.

Yet geography was still cultivated in the Roman world, and several treatises and itineraries remain, interesting to the critic, but of little importance in a general view. Of the itineraries the Romans had two kinds, the Picta and Annotata, or the drawn and written: the latter contained the names of the stations and chief places, with their distances from one another, without entering into any detail. In the Itineraria Picta, which were much more complete, all the great roads were drawn, the name and extent of the different provinces added, with