Unfortunately for this conjecture, both Marsden in his notes on Marco Polo, and after him Klaproth, have shown in the clearest manner, that the Ardandam of that noted traveller makes no part of Tibet, but is altogether correspondent to the S.W. part of the province of Yunnan in China, of which Yongchang is the capital, called Unchian (not Nokian, as in Malte Brun,) by Marco Polo. The name of this district, besides, is not Ardandam, or Arcladam, as Malte Brun has written it after Muller's edition of Marco Polo; nor is it Kardandam, as Marsden has it; but Zar-dandan, a Persian appellation, signifying 'teeth of gold;' because in that district gold is so abundant, that every man wore a small plate of that metal as a cover to his teeth, according to Marco Polo, and which was exchanged for an equal quantity of silver, brought to them by the inhabitants of Mien, (Burmah,) as it was not to be found at all in the country of Caridi. Malte Brun has also identified Caridi, as he writes it with Ardandam or Zardandam, whereas the proper reading is Carazan, and not Caridi; and it is a corruption of Karadjang, or the black country,' the N.W. part of Yunnan, from the colour of the people, called Ouman, or black barbarians,' by the Chinese. The capital of this district is the modern Ta-li-foo, called Dai-leiou by the Chinese, and the capital of an independent state till destroyed by the Mongols in 1255, subsequent to the conquest of Tibet. According to Marco himself, he took five days journey on horseback W. from Ta-li-foo to the province of Zardandan. The Caraiam of Marco Polo was not Assam, as Malte Brun imagined, and which, in his fancy, had some relation to the Garrow mountains, but lay to the W. and N. of Karadjang, and was called Karayan by the Mohammedan writers, and Thsuonman by the Chinese, who call the people Carains. The large lake of Eul-hai, or Tali, separates Karayan from Karazan or Karadjang. We dissent totally from Klaproth in placing Caraiam to the S.E. of Karazan, and making it the south part of Yunnan, which is inhabited by the Mons. Caraiam extended N. and W. as far as the Caindhu of Marco Polo, on the Bruis, or Kincha Keeang, or 'river of golden sand,'-not the Brahmaputra, as Malte Brun imagined, -nor the Sanpo, as Klaproth dreams, or his Irrawaddy,—but the genaine Kincha Keaung of Tibet, and the north-west part of Yunnan.From what we have been able to gather from these confused and imperfect accounts of the divisions of Tibet are the following: 1st, Nagaree, towards the chain that divides the sources of the Indus and Sutluj from Upper Tibet, and which contains the sources of the Yarou Sanpo, the Keaung-koo, and other large rivers; 2d, Kam, probably the Kahang of Giorgi, called by him a principality; and kam, in Tibetian, means the kingdom. It lies S.E. of Nagaree; 3d, Karra Tabet, on the Yarkia-Sanpoo; 4th, Hor, on the N. side of Tibet; 5th, Congbo, on the left of the Sanpoo; 6th, Tac-po, on the right of that river, and both S.E. of Lassa; 7th, Bregiong, to the W. of Congbo; 8th and 9th, Qui and Chang, both comprehended, according to Father Hyacinth, in the province of Chlassa, and called Oochang; 10th, Jiamdo; 11th, Zanba, or Teeshooloomboo; and 12th, Amdoa, on the borders of Sechwen. As for the districts of Keeaung and Daum, we cannot fix their relative positions.

Physical Features.] Tibet presents, on an inspection of the Laina's map, the most confused assemblage of mountains, valleys, lakes, and rivers, that can meet the eye, resembling a piece of net-work; and to attempt a verbal description of such an apparent chaos would be to make confusion

worse confounded; and as no triangulation of this vast region was ever made, as in the case of China, we cannot depend on the accuracy of the relative positions, nor the directions of the interior chains. The declination of the slope, as indicated by the courses of the rivers, seems to be chiefly to the S.E. and E. From China to Tibet is a very sensible ascent the whole way; and the mountains, which are very numerous, are far higher above the horizon on the side towards China than on that towards Tibet ; but when once they are passed, there is a descent, and the climate is much more mild, and the country less wild and savage, than it is on the borders of Sechwin and Yunnan. According to the account of a Chinese officer, who served in the war of 1792 against Nepaul, the army took 72 days to march from the frontiers of China to the vicinity of Lassa, 12 days more to Upper Tibet, and 15 days more to the western range which separates Tibet from Nepaul-total 99 days to the seat of war. The passes, in his account, are spoken of with horror, some of them requiring a whole day to cross; and, when once attempted, must be passed before night, as there is no halting place, nor any possibility of travelling in the dark. The Chinese generals were compelled to dismount and walk over these tremendous mountains, instead of riding across them, as they are totally impervious to cavalry. But on the side of Hindostan the descent is much more rapid than on that of China, as being much shorter. When once the pass of Pharee is passed, the contrast between Bootan and Tibet is striking. It appears to the eye as one of the least favoured countries under heaven, and seems in a great measure incapable of cultivation. It exhibits only low rocky hills, without any visible vegetation, or extensive arid plains, both of the most stern and stubborn aspect, promising fully as little as they produce. But the very gradual descent of 70 British miles, from this pass to the Sanpo, was all that Turner saw of Tibet.

Mountains.] Those of the S.W. and S. towards Hindostan are the only ranges known to us, as geographically described by such of our travellers and surveyors as have reached them,-Bogle, Tarner, Webb, Colebrooke, Crawford, and others; and the elevations of some of the passes and peaks towards India have been already given. The mountains, on the side of Kemaoon, Nepaul, and Bootan, correspond to the Emodus of the ancients, the Sanscrit Heemadree. There are, properly speaking, three distinct ranges, the Northern, the Central, and the Southern Emodus; the first of which is invisible from the plains of Hindostan, and bounds the valley of the Sanpo to the S., and is not perforated by rivers. The central range is that immediately to the N. of Nepaul, and which is perforated by several rivers, as the Gogra, the Ghandauk, the Arun, and the Tusta-all branches of the Ganges; but amidst such tremendous precipices, and by such narrow gaps, as to render their openings generally totally impracticable. It was through the pass of the Arun, by Lungercote, that the Nepaulese army retreated from Teeshooloomboo; but another division, which took the pass of Kootee, (the Cuthi of D'Orville and Grueber) to the west, lost 2000 men amidst the snow. The pass of Mount Langur, or Lungercote, is represented by the above fathers as the most tremendous and precipitous they had encountered all the way from Sining in Shensee. central range is upwards of 40 miles in horizontal depth, and at the source of the Tusta it is denominated Khawa Karpola, or the mountain white with snow.' The southern range is that called by Kirkpatrick the Kuchar Alps, immediately N. of Catmandoo; and there its

This

summits are not, as he imagined, covered only with patches of snow, but with perennial snow to a very great extent. The distance from Catmandoo to Lassa is 536 British miles, or 462 geographical miles, by the road, according to Father Giorgi, but not above 300 geographical miles in direct distance, supposing Lassa to be in 29° 35′ N. lat., as in the Lama's map; so that more than one-half the space traversed is taken up with windings, from the very mountainous nature of the road. It must be remarked, however, that Rennel has placed Catmandoo 33' too far N., and Lassa a full degree too far N.; but he had no other way of fixing them but by routes in his map. According to father Giorgi, Mount Langur is 50 road miles beyond Mount Rimola, (the southern Himmaleh,) and abounds in suffocating exhalations, which increase as it is ascended by the pass; and 25 miles beyond this is the beautiful alpine valley of Tingri,-an earthly paradise in every respect but the sharpness of the air.

Topography.] At the southern foot of Mount Langur is the fortress and town of Tankya, the first place recognised in the Lama's map; and 90 miles beyond this is Zuenga, on the Bontsu. From hence two roads lead to Lassa,-the most northern by Sgigatche, (the Jiekse of the Lamas,) and Ringboo, and the other by Kiangtse. Here wild horses, variously spotted, are seen sporting in great numbers on the banks of the Bontsu. Kiangtse is a fine city and fortress, with a convent of Ghyllongs, so very extensive and magnificent as to seem another city of itself. Fifty miles beyond this, to the N.E., and within three days' journey of Lassa, is the famous lake of Pallee, or Jamdro, or Jangso, said to be so large by the natives as to be 18 days' journey in circumference; but in the Lama's map the circumference is only 150 miles. In the middle, according to Giorgi, is a continued range of hillocks and islands; or, according to the map above-mentioned, and the only one we yet possess, one large island, encircled by a lake from three to eight miles broad. On the western shore of this island is a monastery, and the seat of a Lama priestess, called Lamissa Turcepamo, a female incarnation of Boodha. The road from Kiangse to Lassa lies alongst the north side of this singular lake, a day and a half's journey. Between this lake and the Sanpo is a very lofty range, called Mount Kambala, which travellers must cross in the way to Lassa. From the summit of this elevated range is seen, towards the N., a still higher range, covered with during snow. Seven miles beyond the base of Mount Kambala is the Sanpoo, which 12 miles farther on is crossed by a bridge of iron chains, stretched from side to side, with planks or logs laid across. This chain bridge is composed of 500 links, each a foot long, laid across the narrowest part of the stream, which here is very deep and rapid. N.E. of this crossing place of the Sanpoo is the famed city of Lassa, 24 miles distant by the road. Thus is Giorgi's itinerary of the route from Catmandoo to Lassa correct in its details, but, unfortunately, no bearings of the route are given; and the same is the case with Mr Bogle's route from Coosbeyhar, Tassisudon, and Paridrong, to Chanmanning, the then residence of the Dalai Lama; so that we are in the dark respecting the particular direction of both routes; and these are all we have to rest on for the relative positions of place, in the total absence of observations of latitude and longitude. Turner, in his journey to Teeshooloomboo, entered Tibet at the pass of Pharee, called Paridrong and Paridsong in the Lama's map, which was ascertained to be a full degree too far S. in the Lama's

map, as Rennel justly suspected, being in 27° 58′ N. lat., and 89° l' E. of Greenwich. The pass ascends steep for 12 miles up the southern face of the dreary Soomoonang till its summit be scaled. On this is a long row of small inscribed flags, fixed in rude cairns of stones, and fluttering in the wind. These mark the boundary between Tibet and Bootan. To the N.E. of this a few miles, is the noted peak of Chumularee, visible at a direct distance of 232 British miles. At the foot of the pass is the sterile vale and fortress of Pharee. The road goes to Tushooloomboo almost due N. along the banks of the Painomchoo, by Tueena, Sumdta, Chaloo, Nainee, and Jhansee Jeung, the distance from Pharee about 70 British miles direct. E. of this we hear of no other passes through the Himallah to Tibet. E. of the Langtang mountains, which separate Assam from the valley of the Seree Serhit, all the way to the frontiers of Yunnan and the course of the Lookiaung, a distance of 100 British miles at least, in the parallel of 27° 30', the range continues an unbroken course, of tremendous elevation, and skirted at its base by a large tract of rugged subalpine country, impracticable even for the hardy mountaineers themselves, and all communication is apparently barred to the N. The range seen from the summit of Mount Cambala is that of Koiran, esteemed the highest in Tibet, and runs to the S. of the lake of Terkiri. The only reason assigned for its superiority of elevation to the other chains is merely the circumstance of its central position, but it is no conclusive proof. The range to the W. is perhaps equally elevated, and communicates with the Moos Taugler to the N. and the Himmalah to the S. It is called Kentaisse and Kanteshan. The names Malaya and Kelasch, signifying 'snowy mountains,' are corruptions of the Sanscrit Himalaya, and Kailasa, or Khaillas. The Sanscrit Rimola, applied to the dividing range between Tibet and India, has been corrupted by transcribers into Moriul, as appears in the geographical maps of the 17th century. Respecting the other interior ranges, we can say nothing as to their elevation, or whether they are groups or ranges, as the maps present us with nothing but a confused assemblage of mountains, lakes, and rivers, which it is impossible to describe in

words.

Lakes.] This very mountainous region contains a host of lakes, a great number of which have no outlet, especially in the northern part. The Terkiri is 70 miles long and 25 broad, and contains a superficies of 2,300 square miles. Malte Brun observes, that if one line be drawn from the Terkiri lake 220 miles N., and another W. of the same 470 miles, we shall find 23 other lakes, which have no outlet, or which flow the one into the other. These masses of stagnant water are the result of the configuration of the Tibetian plateau, which is itself a collection as it were of smaller plateaus, encircled by mountains, or so many concave basins, whose waters, having no outlets, descend to the bottoms of their respective hollows, where they either form lakes, or find them already prepared for the reception of their waters. Fifteen days' journey from Tushooloomboo is a lake 20 miles in circumference, that produces tincal, or crude borax, which is formed or deposited in its bed, near the bank; from the deeper parts rock-salt is procured; and during one-half of the year this expanse of saline fluid is covered with a smooth sheet of ice.

Rivers.] Though many of the streams which arise on this most elevated of all terrestrial convexities are lost in the numerous lakes which here and there occupy the hollows of the circular depressions of the mountainous surface, yet several of the longest and largest rivers of the Eastern he

misphere originate in this region. The edict of the emperor Kaunghee, published in 1721, and given by Klaproth in his memoir on the sources of the Brahmapootra and Irrawaddy, published at Paris in 1828, affords a very curious and particular description of the Tibetian rivers, according to accurate data furnished by the superior Lamas. The origin of the Whangho is, in this document, said to be without the frontier of Sining. The real name of the incipient stream is Altun-kol, or Golden river,' as it is denominated by the Mongols. It is about three feet deep, and rises two degrees W. of the Tsing-soo-hai lakes. It has much gold mixed with its sands. The source of this small stream, the commencement of the Whang-ho, is in 35o N. lat. and 21° W. of Pekin. The tract watered by its upper course is called Moma in Tibetian, and Thokan in Chinese. The Whang-ho enters Shinsee, near Hochew, 10 days' journey from its source in direct distance, by a very narrow pass between two vast steep rocks. The next river of Tibet mentioned by Kaunghee is the Min-keeanng; but it is not a Tibetian but a Chinese river, its course being almost wholly within Sechwen.-The Yalong-keeaung forms, for 400 miles in a S.E. course, the S.W. frontier of the Seefaun, and, for 140 more, the western frontier of Sechwen, dividing it from Tibet; and after a course of 160 more, through the S.W. angle of Sechwen, it finally joins the Kincha-keeaung, coming N.E. from Yunnan; thus accomplishing a comparative course of 700 British miles, receiving in its broad and deep channel the waters of the Seefaun at the west of Sechwen. From its source to its entrance into Sechwen, in 29° 54′ N., it bears the appellation of the Sachoo-Tsitsirhana, and then of the Yalong. The Keeaung-koo, according to the emperor's memoir, rises in the N.E. of the states of the Dalai lama, and runs S.E. into Tibet, and then enters Sechwen. Farther on, the Keeang passes Kwei-chow-foo, enters Hookwang, waters King-choo-foo, and joins the Han-keeaung before Woo-chang-foo.-The Han-keeaung comes from the Po-chung-shan mountain—a mountain of Shinsee in the district of Nin-keang-chow, and bears at its source the name of Shih-yang-shweey; it runs to the eastward, and at Nan-ching-hun enters Hookwang, and joins the Great Keeaung near Han-yang-hun: the joint streams are called Hankenow. In the map of the Lamas, the Kincha-keeaung rises in Ngari, or Upper Tibet, (not in the Seefaun,) in 35° N. lat. and 90° 27′ E. long. from a lake called Pahaton Kol. Its name at the source is not given in their map. It runs E. to a place called Hurha, or the Custom-house,' where it is joined by a stream from the S. Two degrees E. of its source it receives the Pitee Muran, a large stream from the S., and a little beyond, a still larger, called the Aktam. In 4 degrees E. of its source, where it is separated by Mount Koolkoon from the parent stream of the Whang-ho, it turns to the S. and S.E., passing by Cocosay, a customhouse, and Tsitsirkhana, a ruined city, and is called in this part of its course the Porsic-ho. In 31° N. lat. it enters the country of Laton, and runs almost due S. till at Tachinquan, in 27° 32′ N. and 16° 40′ W. of Pekin, it enters Yunnan, after a comparative course of 1000 British miles through Tibet; and its course through China is at least double that distance. For volume of water, it is the largest in the Eastern hemisphere, being calculated at 464.400 cubical feet of water, per second of time, discharged into the sea. Including sinuosities, its length, of course, is to that of the Thames as 214 to 1, or as 3,780 to 180 British miles; and its basin to that of the Thames as 138 to 1, or 760,000 British square miles, as it comprehends all the central part of China and the eastern part of

V.