of the many migratory races who have since swept through the country. At the foot of the mountains again are tracts of surpassing fertility, rich well-watered plains, where the great mass of the population congregate in towns and villages and pursue the peaceful arts of life, the miscellaneous character of the inhabitants of these marts of commerce and industry being unequalled perhaps in any other part of the East.* And thirdly, beyond the cultivated plains stretches out in every direction the pathless desert, which has been tenanted by pastoral nomades ever since the earth was peopled. Here rapine and disorder seem to have their natural home, and here, at the present day, to the ordinary excesses of brigandage is superadded the detested occupation of man-stealing.

Those who have been accustomed to regard Central Asia solely under its present condition of political and social degradation, may find it difficult to realize the idea that it was ever the seat of arts and industry, or had made any great advance in civilization, yet such was undoubtedly the case. We are not able, it is true, as in the case of Egypt, or Babylonia, or Assyria, to appeal to contemporary monuments in support of a Central Asiatic development at a period of any remote antiquity, but the evidence to this effect, derived from a large field of induction, is not less significant and sure. In the first place, the belief in a very early empire in Central Asia, coeval with the institution of the Assyrian monarchy, was common among the Greeks long anterior to Alexander's expedition to the East, and could only have been derived from the traditions current at the court of the Achamenian kings. This belief again is connected through the names of Oxyartes and Zoroaster with the Iranian division of the Arian race, and receives confirmation from the earliest memorials of that people. Without seeking, indeed, to penetrate the myth

The following list, which is given in the anonymous Russo-German travels, of the component parts of the population of Cashgar-though the numbers, if referring to individuals rather than to families, are far too limited, and though the pretended Armenian element is probably an invention of the author-would seem to be otherwise relatively correct, and may be taken as a fair sample of the mixture of races in a Central Asiatic town:

6

of Iran-víj, the legendary birthplace of the so-called Persian race, there can be no reasonable doubt that the enumeration of the other fifteen localities successively created by Ormuzd, which is given in the opening chapters of the Vendidad,' indicates the progress of Iranian colonization during the earliest phases of the national existence; and it is thus of much ethnological importance to find that the empire commenced with Sogdiana, Merv, and Bactria; that in its subsequent development it included the modern provinces of Khorassan, Afghanistan, and Kharism,* and finally at its period of greatest extension stretched from Seistan on the south to the Jaxartes on the north, and from the Indus on the east, till it touched the extreme limit of the Median frontier to the west. It was formerly argued from this classification of the Iranian settlements that the antiquity of the legend must be enormous, since neither the Medes nor Persians, whose cognate nationality is unquestioned, and who are historically mentioned as early as 2000 B.C., were not included in the series; but modern criticism prefers to explain their omission from the list by supposing that there were in reality two distinct systems of civilization among the Iranian division of the Arian race, synchronous in their action, though geographically and politically divided. Of these the Eastern system described in the Vendidad' may have had its primæval seats upon the Oxus, and have been identified with the dualism of the Zend Avesta; while the other-or Western-system may have been more immediately connected with Magism, and have belonged to western Persia, being, perhaps, locally centralised in northern Media, but with ramifications extending into Armenia and Asia Minor. The Persians, indeed, when they are first met with in the Assyrian inscriptions of the ninth century B.C., are not settled in the south, but appear as a cognate race with the Medes in the modern province of Azerbijan; and that offshoots of this race must even at that early time have been pushed far on towards the west is proved by the names of Kustaspa, king of Comagene, and Aspabara of Armenia, who are mentioned amongst the adversaries of Tiglath Pileser and of Sargon. In this western, or Magian, division of the Iranians

It is singular that neither Mons. Khanikoff himself, nor any of the modern commentators on the Vendidad whom he enumerates-Spiegel, Bréal, Haug, and Justi-should have recognised Kharism among the sixteen localities created by Ormuzd. There can be no doubt, however, but that the eighth name, which immediately precedes Vehrcan, or Hyrcania (modern Gurgan), and which is read as Urvan, represents the well-known title of Urganj, the old capital of Kharism, the Zend being regularly replaced by the Persian g, and the terminal j being dialectic, as in the names of the Kharismian months, which, according to Abu Rihan, are optionally written with or without a final j.

were

were included no doubt both the Medes and Persians; their language was probably that with which we are acquainted from the tri-lingual inscriptions of Persia, and it may be conjectured from many circumstances connected with the history of Darius Hystaspes that it was in his reign, and in connection with the overthrow of the Magian usurpation that the dualism of Oromasdes and Arimanes was first introduced from the far East, in supercession of the old national faith.*

It is with the Eastern Iranians, however, that we are principally concerned, as the founders of Central Asian civilization. This people, on the authority of the Vendidad, may be supposed to have achieved their first stage of development in Sughd. Their language was probably Zend, as distinguished from the Achæmenian Persian, and somewhat more removed than that dialect from the mother-tongue of the Arians of the South. To them must be referred the old Greek traditions of the Bactrian Zoroaster, and the entire framework of Persian historical romance, culminating in the famous Epic of Firdousi. A more important evidence, however, of the very high state of power and civilization to which they attained is to be found in the information regarding them preserved by the celebrated Abu Rihan, himself a native of the country, and the only early Arab writer who investigated the antiquities of the East in a true spirit of historical criticism.† This writer supplies us with an extensive specimen of the old dialects of Sughd and Kharism. He gives us in those dialects the names of the twelve months, the names of the thirty days of the month, and of the five Epagominæ, together with the names of the signs of the zodiac, of the seven planets, and, lastly, of the mansions of the moon. A portion of this nomenclature is original, and offers a most curious subject for investigation; but the majority of the names can be compared, as was to be expected, with the Zend correspondents, and, indeed, are much nearer to the primitive forms than are the better known Parsee equivalents. According to Abu Rihan, again, the solar calendar of Kharism was the most perfect scheme for measuring time with which he was acquainted; and it was maintained by the astronomers of that country, that both the solar and the lunar zodiacs had originated with them, the divisions of the signs in their system being far more regular than

This subject is ably and exhaustively treated in Rawlinson's 'Herodotus,' vol. i. p. 426.

We quote from a most excellent work of Abu Rihan's on general chronology, which has not, we think, received the attention that it merits at the hands of European scholars, though there is a copy of the MS. at Paris, which was formerly much referred to by Quatremère, under the title of Athar-el-Bakíeh.'

those

those adopted by the Greeks or Arabs; and the very name, moreover, by which an astronomer was designated in the language of Kharism being taken from the asterism of the eighth mansion of the moon, All this information is exceedingly curious in its bearing upon the controversy which has so long raged in the scientific world, as to the superior antiquity of the lunar zodiac used respectively by the Indians and Chinese, leading as it does to a suspicion that neither the one nor the other of these systems may have been original, but that their similarity may be explained by their derivation from a common centre in Bactria, where astronomy was first cultivated by the Eastern Iranians. An argument of some weight, indeed, in favour of this derivation is furnished by another statement of Abu Rihan's, which asserts that the Kharismians dated originally from an epoch anterior by 980 years to the era of the Seleucidæ, a date which agrees pretty accurately with the period assigned by our best scholars to the invention of the Jyotisha or Indian calendar.† Abu Rihan further speaks of the Kharismian writing and records, which were carefully investigated by Koteibah Ibn Moslem when he conquered the country, and strengthens the authority of these native documents, by showing that a single family named the Shahiyeh, and supposed to be derived from Cyrus, had reigned in Kharism-with the exception of a Turkish or Scythian interregnum of 92 years-from the Achæmenian period down to the time of the Mahomedan invasion. ‡

We have specially alluded to this evidence of early Arian

*This term is Akhir vínak, the observer of Akhir,' which is the name of the 8th mansion.

The date of 980 years before the era of the Seleucidæ is equal to B.c. 1304. Now, the date derivable from the Jyotisha observation of the Colures has been variously calculated by different Sanscrit scholars; the earliest date being that determined by Davis and Colebrooke in the last century, namely B.C. 1391, and the most recent being that lately adopted by Archdeacon Pratt and approved by Professor Max Müller, namely B.C. 1181. Professor Whitney, it is true, does not agree with either of these results, and thinks, indeed, that the data for calculation are so faulty and uncertain that a margin of several centuries should be allowed for possible error; but Sir E. Colebrooke, on the other hand, in replying to his criticism, has shown that a mean calculation of the two Equinoctial stars Revatí (or & Piscium) and Chitrá (or Spica) will bring us to the end of the thirteenth century B.C.; which is almost identical with the Kharismian date of B.C. 1304. This date, too, is almost certainly an astronomical rather than a political era, and was connected with the institution of the lunar zodiac, which, like the original Indian zodiac, commenced with the asterism of the Pleiades.

As an example of the accuracy of the chronology of the Kharismians, Abu Rihan further quotes from their annals the date of the building of the famous castle of 'Ir, near the city of Kharism, in A.S. 616 (=A.D. 292)-a date which afterwards became a national era-and adds that this place continued to be the royal residence till it was destroyed by inundations of the Oxus in A.S. 1305 (A.D. 981). We have never seen in any other Arabic author an account of this castle, which is compared by Abu Rihan with the celebrated Ghamdán of Yemen; nor, indeed, do we think that the antiquities of Kharism are elsewhere at all noticed. Vol. 120.-No. 240. 2 L civilization

civilization in Central Asia, furnished by a writer of Abu Rihan's authority, as we believe it to be entirely new to Oriental students; but there are many other notices of a corroborative character, which have been often quoted. Justin's notice, for instance, of the thousand-citied Bactria which revolted under Theodotus, indicates a very high state of prosperity and power. Balkh and Kharism, again, furnish all the most favourite illustrations for the old Persian romance. The fire temple of Núbehar, at the former place, in which the Barmecide family, previous to their emigration to Baghdad, were servitors, was one of the most famous shrines of the Zoroastrian faith throughout the East; and the original fire of King Jem, the Eponym of the Iranian race, was supposed to have survived unextinguished on an altar in Kharism until the introduction of Islam. Of course fable was abundantly mingled with truth in these glimpses of old-world history. The pretended expedition of the Himyarite king, who founded Samarcand, from the south of Arabia, cannot, for instance, command a moment's attention; but there was certainly an ancient tablet, in an unknown tongue, over one of the gates of the city, which was supposed to commemorate this expedition; for Jeyhani, the Samanide vizier, distinctly says that he saw it in about A.D. 920, and that it was destroyed during a popular émeute whilst he was resident in the city. Whether this inscription was in Zend, or in Greek, or in Bactrian Pali, can be now , of course a mere matter of conjecture, but the mention of such a tablet may well excite our curiosity.

The Iranian people who were thus settled between the Oxus and the Jaxartes, as early as the time of the Judges of Israel, still hold their ground in the country, notwithstanding the continuous flood of foreign races which has ever since swept over the region, surging up from those prolific slopes of the Altai that have been called the 'officina gentium.' Under the names of Tat, Tajik, Sert, Galsha, and Parsiwan, a primitive and not impure Iranian population is to be found in almost every district from the Indus to the Jaxartes, subject to the dominant Afghans to the south, and Uzbegs to the north. The same nationality prevails throughout the valleys of the upper Oxus in a quasiindependent position; and these mountaineers, who, in their modern name of Vakhani, retain the old ethnic title which originated the "Oos of the Greeks, are perhaps the best representatives extant of the primæval race.* To the east of the Pamír

*This primitive title we suppose to have been Vakh, or Vakhsh, but its signi fication is unknown. It gave rise, however, not only to the "Ogos of the Greeks, but to the title of Vah-rúd, by which the Oxus is known in the Bundehesh, and among the old Zoroastrians generally, and also to the modern names of Vakhan, Vakhsh-ab, Vash-jird, &c.

the