KUEN-LUN, or KWEN-LUN, a term used to designate generally the mountain ranges which run along the northern edge of the great Tibetan plateau in Central Asia. In a wider application it means the succession of ranges which extend from the Pamirs on the W. to 113 E., until it strikes against or merges in the steep escarpments of the S.E. flank of the Mongolian plateau. In the narrower acceptation it applies only to those ranges which part the desert of Takla-makan on the N. from the Tibetan plateau on the S. between the Pamirs and the transverse glen of the Kara-muren, that is, nearly to the longitude of the town of Cherchen (about 85$ E.). Although the use of the name is thus restricted in geographical usage, the mountain system so designated does, as a fact, extend eastwards as far as the great depression of Tsaidam (say 95 E.), though it is uncertain whether its direct orographical continuation eastwards is to be identified with the Astin-tagh, or, as F. Grenard and K. Bogdanovich believe and with them Sven Hedin is inclined to agree with the parallel ranges of Kalta-alaghan and Arkatagh, which lie S. of the Astin-tagh. At any rate the Astintagh, whether it is the principal continuation of the Kuen-lun or only a subsidiary flanking system, is itself the westward continuation of the Nan-shan or Southern Mountains, which reach down far into China (to 1 13 E.).
Taken in its widest meaning, the Kuen-lun Mountains thus stretch in a wavy line for nearly 2500 m. from E. to W., and while in the W. their constituent ranges are folded and squeezed by lateral compression into a breadth of some 15^200 m., their summits being forced up to correspondingly higher altitudes, in the E. they spread out to a breadth of some 600 m., the ranges being in that quarter less folded, and consequently both flatter and lower. In the tectonic structure of Asia the Kuen-lun forms, as it were, the backbone of the continent. In point of age it is very much older than either the Himalayas to the S. or the Tian-shan to the N. But although the crests of its component ranges reach altitudes of 21,500 to 22,000 ft., they are not as a rule overtopped by individual peaks of commanding and towering elevation, as the Himalayas are, but run on the whole tolerably uniform and relatively at little greater altitude than the lofty valleys which separate them one from another. It is a strikingly marked characteristic of the northern edge of the Tibetan plateau that its outermost border-range (e.g. Western Kuen-lun and Astin-tagh) is throughout double; and this " twinning " of the mountain-ranges, as also of the intermont lake-basins among the Kuen-lun ranges, is a peculiar feature of the Tibetan plateau.
The supreme orographic importance of this great Central Asian mountain system was recognized in a fashion even by the geographers of ancient Greece. They used to suppose that an immense range of mountains crossed Asia from west to east on the parallel of the island of Rhodes, extending through Asia Minor, the Kurdish highlands, the N. of Persia, the N. of Eactria (Afghanistan), the Hindu- kush, and so on into China. This long range they supposed to separate the waters which flow N. to the Arctic from those which flow S. to the Indian Ocean. K. Ritter (Asien, ii.) was the first of modern geographers to recognize the true character of the Kuen-lun as a border range of the Tibetan plateau ; and Baron von Richthofen (China, i. 1876) still further defined and accentuated the conception of the system by representing it as a complex arrangement of several parallel ranges, running in wavy lines from the Pamirs (76 E.) eastwards to 118 E. But though von Richthofen's general conception of the Kuen-lun system was broadly sound and in accordance with facts, the details both of his description and of that of his pupil Wegener 1 require now very considerable revision, and need even to be in part recast, as a consequence of explorations and investigations made since they wrote by, amongst others, the Russian explorers N. M. Przhevalsky, M. V. Pyevtsov, V. I. Roborovsky, P. K. Kozlov, K. Bogdanovich, V. A. Obruchev, and (?) Skassi ; by the Englishmen A. D. Carey, A. Dalgleish, St G. R. Littledale, H. Bower, H. H. P. Deasy and M. S. Wellby; by the American W. W. Rockhill; the Frenchmen J. L. Dutreuil de Rhins, F. Grenard, P. G. Bonvalot and Prince Henri d ^Orleans; by the Hungarians L. von Loczy and Count Szechdnyi; and above all by the Swede Sven Hedin.
Western Kuen-lun. On the east the Pamir highlands are fenced off from the East Turkestan lowlands by the double border-ridge of Sarik-kol (the Sarik-kol range and the Muztagh or Kashgar range), which has its eastern foot down in the Tarim basin (4000-4500 ft.) and its western up on the Pamirs at 10,500 to 13,000 ft. above sealevel, while its own summits, e.g. the Muztagh-ata (25,780 ft.), shoot up far above the limits of perpetual snow. This double borderridge is continued east of the meridian of Yarkand or Yarkent (7jr E.) by a succession of twin ranges, all running, though under different names, from the W.N.W. to the E.S.E. According to the investigations of F. Stoliczka and K. Bogdanovich, the same fossils occur in both sets of border ranges, in the Sarik-kol and in their eastward continuations, e.g. corals ; Stromatophorae, Bryozoa, A try pa reticularis, A. latilinguis and A. aspera, Spirifer verneuili, etc., and these the latter geologist assigns to the Devonian epoch. These eastward continuations of the double border-range of the Pamirs are the constituent ranges of the Kuen-lun proper. The names given to them are the Kilian or Kiliang, the Khotan and the Keriya Mountains in the more northerly range and the Raskem or Raskan, the Sughet and the Ullugh-tagh Mountains in the more southerly range. Although they alldecrease in altitude from west to east, they nevertheless reach elevations of 19,000 ft., with individual peaks ascending some 2000-2500 ft. higher. From the East Turkestan lowlands on the north the ascent is very steep, and the passes across both sets of ranges lie at great altitudes; for example, the pass of Sanju-davan in the lower range is 16,325 ft. above sea-level, and the Kyzyl-davan, farther east, is 16,900 ft., while the Sughetdavan in the higher range is 1 7,825 ft. The latter range is separated from the Karakorum Mountains by the deeply trenched gorge of the Raskem or Yarkand-darya, while the deep glen of the Kara-kash or Khotan-darya intervenes between the upper (Sughet Mountains) and the lower (Kilian Mountains) border-ranges. Altogether this western extremity of the Kuen-lun system is a very rugged mountainous region, a consequence partly of the intricacy of the flanking ranges and spurs, partly of the powerful lateral compression to which they have been subjected, and partly of the great and abrupt differences in vertical elevation between the crests of the ranges and the bottoms of the deep, narrow, rugged glens between them. In the broad orographical disposition of the ranges there is considerable similarity between north Tibet and west Persia, in that in both cases the ranges are crowded together in the west, but spread out wider as they advance towards the east. To the two principal ranges in this part of the system F. Grenard, who accompanied J. L. Dutreuil de Rhins on his journey in 1890-1895, gives the names the Altyn-tagh and Ustun-tagh, though he names no less than six parallel ranges altogether. Now as Altyn-tagh* is an accepted, though in point of fact erroneous, name for Astin-tagh, it is clear that Grenard considers the main Kuen-lun ranges to be continued directly by the Astin-tagh.
From the transverse breach of the Keriya-darya (about 8it E.) to that of the Kara-muren in the longitude of Cherchen (about 85$ E.) the parallel border-ranges of the Tibetan plateau trend to the E.N.E., and here occur in the lower or outer range the passesof Dalai-kurghan-art (14,290 ft.), Choka-davan, i.e. Littledale'sChokur Pass (9530 ft.) and others at altitudes ranging from 8600 to 1 In " Orographie des Kwen-lun," in Zeitschrift der Gesellschaft fur Erdkunde zu Berlin (1891).
1 It is used, for instance, on the map of " Inner-Asien " (No. 62) of Stieler's Hand-atlas (ed. 1905) and in the Atlas of the Russian General Staff. Etymologically the correct form is Astin-tagh or Astun-tagh, meaning the Lower or Nearer Mountains. Ustun-tagh, which appears on Stieler's map as an alternative name for Altyn-tagh, means Higher or Farther Mountains, and though not used locally of any specific range, would be appropriately employed to designate the higher and more southerly of the twin border-ranges of the Tibetan plateau.
11,500 ft., while in the upper range are the At-to-davan (16,600 ft.), Yapkak-lik-davan (15,550 ft.), Sarshu-davan (15,680 ft.) and others not named at 16,590 and 17,300 ft.
Middle Kuen-lun. Between the upper transverse glensof the Karamuren (or Mitt River) and the Cherchen-darya stretches the short range of Tokuz-davan. From it, on the east side of the Cher chendary t, in about 86 E., the component ranges of the middle Kuen-lun begin to diverge and radiate outwards (i.e. to north and to south) like the fingers of the outspread human hand. And here at least four principal ranges or groups of ranges admit of being discriminated, namely the Astin-tagh, the Chimen-tagh, the Kalta-alaghan and the Arka-tagh, all belonging to the mountainous country which borders on the north the actual plateau region of Tibet. Although these several ranges, or systems of ranges, differ considerably in their orographical characteristics, the following description will apply generally to the entire region from the Astin-tagh southwards to the Arka-tagh. The broad features of the surface configuration are a series of nearly parallel mountain-ranges, running from W.S.W. E.N.E. to W.N.W. E.S.E., and separated by high interment valleys, which are choked with disintegrated material and divided into a chequered pattern of self-contained, shallow lacustrine basins. As a rule the crests of the ranges are worn down by aerial denudation and have the general appearance of rounded domes. Hard rock (mostly granite and crystalline schists, with red sandstone in places) appears only in the transverse glens, which are often choked with their de'bris in the form either of gravel-and-shingle or loose blocks of stone or both. The flanks of the mountains are so deeply buried in disintegrated material that the difference in vertical altitude between the floors of the valleys and the summits of the ranges is comparatively small. But as each successive range, proceeding south, represents a higher step in the terraced ascent from the desert of Gobi to the plateau of Tibet, the ranges when viewed from the north frequently appear like veritable upstanding mountain ranges, and this appearance is accentuated by the general steepness of the ascent; whereas, when viewed on the other hand from the south, these several ranges, owing to their long and gentle slope in that direction, have the appearance of comparatively gentle swellings of the earth's service rather than of well-defined mountain ranges. As a rule, the streams flow alternately east and west down the interment latitudinal valleys, until they break through some transverse glen in the range on the northern side of the valley. In the western parts of the system they mostly go to feed the Kara-muren or the Cherchen-darya, while farther east they flow down into some larger self-contained basin of internal drainage, such as the Achik-kol, the two lakes Kara-kol, or the Ghaz-kol, and even yet farther east make their way, some of them into the lakes of the Tsaidam depression or become lost in its sands or in those of the Kum-tagh desert on the north, or go to feed the headstreams of the great rivers, the Hwang-ho (Yellow River) and the Yangtsze-kiang (Blue River) in the south. It appears to be a rule that the rivers which eventually terminate in the deserts of Gobi and Takla-makan grow increasingly larger in magnitude from east to west. Another law appears to distinguish the hydrography of at any rate the great latitudinal valleys of the Arka-tagh and the Chimen valley (north of the Chimen-tagh) : the streams flow close under the foot of the range that shuts in each individual valley on the north. But in respect of precipitation there is a very marked difference between the valleys of the north and those of the south. Whereas both the mountains and valleys of the Astintagh and of the Akato-tagh (the next large range to the Astin-tagh on the south) are arid and desolate in the extreme, smitten as it were with the desiccating breath of the desert, those of the Arka-tagh and beyond are supersaturated with moisture, so that, at any rate in summer, the surface is in many parts little better than a quaking quagmire. Throughout vegetation is scanty and faunal life poor in species, though in some respects certain of the species, e.g. wild yaks, wild asses (kulans), antelopes (orongo and others), marmots, hares and partridges exist locally in large numbers. The wild camel approaches the north outliers of the Astin-tagh, but rarely, if ever, ventures to enter their fastnesses. Bears, wolves, foxes, goats (kokmet), wild sheep (arkharis), lizards, earth-rats, and a small rodent (teshikan), with ravens, eagles, wild ducks and wild geese are the other varieties principally encountered. The vegetation consists almost entirely of scrubby bushes of several varieties, including tamarisks and wild briers, of reeds (kamish), and of grass on the yaylaks (pasture-grounds) of the middle ranges. On the Arka-tagh even the moss, the last surviving representative of the flora, disappears entirely. In the eastern Astin-tagh a variety of wild tea (chay, mountain tea) is used by the Mongols. gold is obtained in very small quantities in a few places in the Astin-tagh and the Kalta-alaghan. The nomenclature of the numerous ranges in this part of the Kuen-lun is extremely confusing, owing to different travellers having applied the same name to different ranges and to different travellers have applied different names to what is probably often identically the same range. In this article the nomenclature adopted is that employed by the latest, and probably the most thorough, explorer of this part of Central Asia, namely, Sven Hedin. Nevertheless, owing to the fact that nearly all the longer and more important crossings of Tibet and its northern montane region have been made from north to south, or vice versa, that is, transversely across the ranges, and comparatively few from east to west along the Jntermont latitudinal valleys, the identifications between ranges in the east and ranges in the west are in more than one instance more or less doubtful.
The Astin-tagh, although it occupies a similar position to the twin ranges of the Western Kuen-lun, in that it forms the outermost escarpment or border-ridge on the north of the Tibetan plateau, would appear in the opinion 01 the most competent judges (e.g. Grenard, Bogdanovich, Sven Hedin, Przhevalsky), to be only a branch or subsidiary range of the main range of the Kuen-lun. It is not however a single, long, continuous chain, as it is shown, for example, on the map of the Russian general staff, but consists of two parallel main ranges, and in the east of three, and even to the N.E. of Tsaidam of four, parallel main ranges, flanked throughout by several subsidiary chains, spurs and offshoots. Beyond that it swells out into the vast massif of Anambaruin-ula, which is traversed by at least three minor parallel chains. But on the east of the Anambaruin-ula it once more contracts to two main ranges, the more southerly being that which Przhevalsky called the Humboldt Range (crossed by a pass at 13,200 ft.). This branch is probably Continued in the range which overhangs the Koko-nor on the south, namely, the south Kokonor Range. The northern branch merges eastwards into the Nanshan or Southern Mountains. 1 The passes in the Lower Astin-tagh range from altitudes of 10,150 to 10,700 ft., and in the Upper Astintagh at 1 1,770 to 15,680 ft. (Tash-davan), though one pass beside the Charkhlik-su is only 9660 ft. high. And as the relative altitudes of crest and pass remain approximately the same as in the Western Kuen-lun, it is evident how greatly the general elevation of the twin border ridge decreases towards the east. But there exists a striking difference between the crests of the Astin-tagh and those of the ranges which give rise to the gigantic ridge and furrow arrangement on the Tibetan plateau. " Here in the Astin-tagh the mountains, like those in the Kuruk-tagh, 2 are indeed severely weathered, but they always consist, from base to summit, of hard rock, bare and barren, most frequently piled up in eccentric, rugged masses, denticulated, pinnacled crests and peaks. On the Tibetan plateau, on the other hand, most of the ranges are distinguished by their rounded outlines and soft consistency, and their striking poverty in hard rock, which in the best cases only crops out near the summits. There too disintegration has been to a remarkable extent operative. This gives rise to the great morphological difference, that in the former regions, the Astin-tagh and the Kuruk-tagh, the products of disintegration are almost always carried away by the wind, and so disappear; no matter how powerful or how active the disintegration may be, none of the loosened material ever succeeds either in gathering amongst the mountains or in accumulating at their foot. The climate is so arid, and precipitation so extremely rare, that the fine powdery material falls a helpless prey to the winds. On the other hand, the precipitation on the Tibetan plateau is so copious, and so uniformly distributed, that it is able to retain the loosened material in situ, and causes it to heap itself up in rounded masses on the flanks of the mountains that are its primitive source of origin, these projecting in great part like skeletons from the midst of their own ruins." The twin ranges of the Astin-tagh are fairly equivalent in point of magnitude and regularity ; but while the Lower Range, on the north, sensibly decreases in altitude towards the east, the Upper Range, on the south, maintains its general altitude in a remarkable way, and is gapped by steep, wild, deeply incised transverse glens directed towards the north, and generally fenced in by dark precipitous walls of rock. The great valley between the two is ' cut up into a series of self-contained basins, each serving as the gathering ground of the brooks that run down off the adjacent mountains. Outside the lower end of each large transverse glen there is a scree of sedimentary matter. These screes are however very flat and their lower edges generally reach all the way down to the central part of the basin, which is occupied by an expanse of yellow clay, perfectly flat and fairly hard, as well as dry and barren, often cracked into polygonal cakes and drawn out in the direction of the long axis of the valley. . . . But though the great morphological features of this latitudinal valley forcibly recall the latitudinal valleys of Tibet, the climatic differences give rise to differences between the basins corresponding to the differences between the mou n- tain-ranges themselves. For while the self-contained basins of Tibet generally possess a salt lake in the middle, into which brooks and streams of greater or less magnitude gather, often from very considerable distances, these self-contained basins of the Astintagh are very small in area, and it is extremely seldom that their central parts receive any water at all, only in fact after copious ram. These terminal lakes, or more accurately sedimentary plains are therefore almost always dry." 4 The next parallel range on the south, the Akato-tagh, and the valley which separates it from the Astin-tagh, are equally arid and waterless. The valley, known by the general name of Kakir, meaning a hard, dry, sterile expanse of clay," is chequered with shallow selfcontained basins of the usual type and has remarkably gentle slopes
1 The Northern Mountains are the Pe-shan in the desert of Gobi (see GOBI).
2 On the opposite or north side of the desert of Lop (desert of Gobi)
3 Sven Hedin, Scientific Results, iii. 308.
4 Ibid. 310-311.
up to the mountains on both north and south. Its surface slopes from altitudes of 10,100 to 10,600 ft. in the west, where is the lake of Uzunshor (9650 ft.) to 9400 ft. in the east, in which direction it continues as far as the Anambaruin-ula (see below) and the plain or flat basin of Sartang, a north extension of Tsaidam. This range of Akato-tagh, the Altun Range of Carey, is the same as that which on the map of the Russian general staff bears the name Chimen-tagh. Like the Astin-tagh it stretches towards the E.N.E., and, like it, appears to be built up of granite and schists, but its crest is greatly denuded, so that it is a mere crumbling skeleton protruding above the deep mantle of disintegrated material which masks its flanks. The slopes on both north and south are extremely gentle, but that .on the south is eight to ten times as long as that on the north. In the east the range is mostly narrow, and dies away on the edge of the Tsaidam depression ; but in the west it swells out into the lofty and imposing mass of the Ilve-chimen or Shia-manglay, which is capped with perpetual snow. This part of the range is crossed by the pass of Chopur-alik at an altitude of 16,160 ft., but farther east the passes lie at altitudes f '3.380 to 10,520 ft. The latitudinal valley that intervenes between the Akato-tagh and the next great range on the south, the Chimen-tagh, slopes for the most part eastwards, from 12,500 ft. down to the shallow salt lake of Ghaz-kol or Chimen-koli (9305 ft.). In the western part of this valley occurs the very important transverse water-divide of Gulcha-davan (14,150 ft.), which separates the basin of the Cherchen-darya that goes down into the Tarim basin from the area that drains down to the Ghaz-kol, Which belongs to the Tsaidam depression. This, the Chimen valley, contains in places a good deal of drift-sand, which however is stationary in the mass and heaped up along the northern foot of the Chimen-tagh. Nevertheless the Akato-tagh is only of secondary importance in the general Kuen-lun system, being nothing more than a central ridge running along the broad Kakir valley that separates the Astin-tagh from the Chimentagh.
The latter range, the Chimen-tagh, is identical in its western parts with the Piazlik-tagh and in the east must be equated with the Tsaidam chain of Przhevalsky ; and it is probably continued westwards by the range which the Russian explorers call the Moscow Range or the Achik-tagh, running north of the Achik-kol and, according to Przhevalsky, connecting on the west with the Tokuz-davan. The Chimen-tagh rises into imposing summits, some rounded, some pyramidal in outline, which are capped with snow, though the snow melts in summer. This range acts as a " breakwater " to the clouds, arresting and condensing the moisture which is carried northwards by the south winds. Hence its slopes are not so arid as those of the Akato-tagh and the Astin-tagh. Snow falls all the year round on the Chimen-tagh, even in July, and water is abundant everywhere. The southern slope of the range is gentle but short, the northern slope long and steep. Grass is able to grow, and animal life is more abundant. The range is crossed by passes at 13,970, 13,230 and 13,760 ft., and the Piazlik-tagh by a pass at an altitude of 13,640 ft.
The next important range, still going south, is the Kalta-alaghan, Carey's Chimen-tagh Range, Przhevalsky's Columbus Range and the range which is variously designated (e.g. by Pyevtsov) as the Ambal-ashkan, Kalga-lagan and Ara-tagh. This last is, however, properly the name of a short secondary range which rises along the middle (ara = middle) of the valley between the Chimen-tagh and the Kalta-alaghan. Not only is it of lower elevation than them both, but it dies away towards the west, the valleys on each side of it meeting round its extremity to form one broad, open valley, with an altitude of 11,790 to 13,725 ft. The Ara-tagh is crossed by a pass at an altitude of 14,345 ft. In the Kalta-alaghan, which is the culminating range of this part of the Kuen-lun, and is overtopped by towering, snow-clad peaks, the passes climb to considerably higher altitudes, namely, 14,560, 14,470, 14,430 and 14,190 ft., while the pass of Avraz-davan ascends to 15,700 ft. This range appears to be linked on to the Tokuz-davan by the Muzluk-tagh, in which there are passes at 16,870 and 15,450 ft. It is possible however that the Muzluk-tagh belongs more intimately to the Chimen-tagh system, that is, to the Moscow or Achik-kol ranges. Indeed Bogdanovich considers that the Tokuz-davan, the Muzluktagh, the Moscow Range and the Chimen-tagh form one single closely connected chain, in which he also places Przhevalsky's isolated peak of Mount Kreml (15,055 ft.). Sven Hedin, whilst agreeing that this may possibly be the true conception, inclines to the view that the Achik-kol Range dies away towards the E., and that the Chimen-tagh and the Kalta-alaghan merge westwards into the border-ranges that lie north of the Muzluk-tagh and the Tokuzdavan. Unlike most of the other parallel ranges of N. Tibet, the Kalta-alaghan does not decrease, but it increases in elevation towards the east, where, like the Chimen-tagh, it abuts upon and merges in the ranges that border Tsaidam on the south.
Immediately south of the Kalta-alaghan comes a relatively deep depression, the Kum-kol valley, forming a very well-marked feature in the physical conformation of this region. It is crossed transversely by a water-divide which separates the basin of the twin-lakes of Kum-kol (12,700 ft.) from the basin of Tsaidam, some 3500 ft. lower. The floor of the valley consequently slopes away in both directions, like the Chimen valley between the Akato-tagh and the Chimen-tagh ; and in so far as it slopes westwards towards the Kum- kol lakes it differs from nearly all the other great latitudinal valleys that run parallel with it, because they slope generally towards the east. Not far from the Kum-kol lakes there is a drift-sand area, though the dunes are stationary. The upper lake of Kum-kol (Chon- kumkol) (12,730 ft.), which contains fresh water, is of small area (8 sq. m.) and in depth nowhere exceeds 13 ft. ; but the lower lake (Ayak- kumkol) (12,685 ft.), which is salt, is much bigger (283 sq. m.) and goes down to depths of 64 and 79 ft. Farther west, lying between the Muzluk-tagh and the Arka-tagh, is the lake of Achik-kol (13,940 ft.), l6J m. broad and 50 m. in circuit.
The next great parallel range is the lofty and imposing Arka-tagh, the Przhevafsky Range of the Russian geographers, which has its eastward continuations in the Marco Polo Range (general altitude 15,750-16,250 ft.) and Gurbu-naiji Mountains of Przhevalsky. The Arka-tagh 1 is the true backbone of the Kuen-lun system, and in Central Asia is exceeded in elevation only by the Tang-la, a long way farther south, this last being probably an eastern wing of the Karakorum Mountains of the Pamirs region. At the same time the Arkatagh is the actual border-range of the Tibetan plateau properly socalled ; to the south of it none of the long succession of lofty parallel ranges which ridge the Tibetan highlands seems to have any connexion with the Kuen-lun system. Of great length, the Arka-tagh, which is a mountain-system rather than a range, varies greatly in configuration in different parts, sometimes exhibiting a sharply denned main crest, with several lower flanking ranges, and sometimes consisting of numerous parallel crests of nearly uniform altitude. Amongst these it is possible to distinguish in the middle of the system four predominant ranges, of whicn the second from the north is probably the principal range, though the fourth is the highest. The passes across the first range (north) lie at altitudes of 15,675, 16,420, 17,320 and 18,300 ft.; across the second at 16,830, 17,020, 17,070 and 17,220 ft.; across the third at 16,800, 16,660, 17,065, 17,830 and 17,880 ft.; and across the fourth at 16,540, 16,765, 16,780, 18,100 and 18,110 ft. The crests of the ranges lie comparatively little higher than the valleys which separate them, the altitudes in the latter running at 14,940 to 16,700 ft., if not higher, and being only 500 to 1000 Ft. lower than the crests of the accompanying ranges. The Arka-tagh ranges do not culminate in lofty jagged, pinnacled peaks, but in broad rounded, flattened domes, a characteristic feature of the system throughout. These Arka-tagh mountains are built up, at all events superficially, of sand and powdery, finely sifted disintegrated material. Where the hard rock does crop out on the surface, it is so excessively weathered as to be with difficulty recognized as rock at all. The culminating summits of the ranges generally present the appearance of a flat, rounded swelling, and when they are crowned with glaciers, as many of them are, these shape themselves into what may be described as a mantle, a breastplate, or a flat cap, from which lappets and fringes project at intervals; nowhere do there exist any of the long, narrow, winding glacier tongues which are so characteristic of the Alps of Europe. But not the slightest indication has been discovered that these mountains were ever panoplied with ice. The process of disintegration and levelling down has reached such an advanced stage that, if ever there did exist evidences of former glaciation, they have now become entirely obliterated, even to the complete pulverization of the erratic blocks, supposing there were any. The view that meets the eye southwards from the heights of the Kalta-alaghan is the picture of a chaos of mountain chains, ridges, crests, peaks, spurs, detached masses, in fact, montane conformations of every possible description and in every possible arrangement. Immediately north of the Arkatagh the country is studded with three or four exceptionally conspicuous and imposing detached mountain masses, all capped with snow and some of them carrying small glaciers. Amongst them are Shapka Monomakha or the Monk's Cap; the Chulak-akkan, which may however be only Shapka Monomakha seen from a different point of view; Tomiirlik-tagh * (i.e. the Iron Mountain) ; and farther west, Ullugh-muz-tagh, which, according to Grenard, reaches an altitude of 24, 140 ft. But the relations in which these detached mountain-masses stand to one another and to the Arka-tagh behind them have not yet been elucidated. In the vicinity of the Ullughmuz-tagh there exist numerous indications of former volcanic activity, the eminences and summits frequently being capped with tuff, and smaller fragments of tuff are scattered over other parts of the Arka-tagh ranges.
The next succeeding parallel range, the Koko-shili, which is continued eastwards by the Bayan-khara-ula, between the upper headstreams of the Hwang-ho or Yellow River and the Yangtszekiang, belongs orographically to the plateau of Tibet.
The succession of ranges which follow one another from the deserts of Takla-makan and Gobi up to the plateau proper of Tibet rise in steps or terraces, each range being higher than the range to the north of it and lower than the range to the south of it. The difference in altitude between the lowest, most northerly range, the Lower Astin-tagh, and the most southerly of the Arka-tagh ranges amounts to nearly 7500 ft. With one exception, namely the climb out of the Kum-kol valley to the Arka-tagh, the first three steps are This is the correct form, Arka-tagh meaning the Farther or Remoter Mountains. The form Akka-tagh is incorrect. 1 The form Tumenlik-tagh is erroneous.
individually the biggest ; whereas the Upper Astin-tagh exceeds the Lower Astin-tagh by an altitude of some 1350 ft., it is itself exceeded by the Akato-tagh to the extent of 1760 ft. There is also a considerable rise of 880 ft. from the Akato-tagh to the Chimen-tagh. But between the Chimen-tagh, the Ara-tagh and the Kalta-alaghan there is comparatively little difference in point of elevation, namely, 730 ft. in all. The biggest ascent is that from the Kalta-alaghan to the Arka-tagh, namely, nearly 1850 ft. The ranges of the Arkatagh, again, run at pretty nearly the same absolute general altitudes, namely, 16,470 to 17,260 ft. When the altitudes of the intermont latitudinal valleys are compared, the significance orographically of the Chimen valley and of the Kum-kol valley is strikingly emphasized. Both are much more deeply excavated than all the other latitudinal valleys that run parallel to them, the Chimen valley being 875 ft. above the valley to the north of it, but no less than 2235 ft. below the valley to the south of it. The case of the Kum-kol valley is altogether exceptional, for it lies not higher, but 680 ft. lower, than the valley to the north of it, and consequently the climb up out of it to the first (on north) of the Arka-tagh valleys amounts to no less than 2900 ft. Hence these ten parallel ranges of the middle Kuen-lun system may be grouped in three divisions (i) the more strictly border ranges of the Upper and Lower Astin-tagh and the Akatotagh; (2) the three ranges of Chimen-tagh, Ara-tagh and Kaltaalaghan, which may be considered as forming a transitional system between the foregoing and the third division; (3) the Arka-tagh, which constitute the elevated rampart of the Tibetan plateau proper. (J. T. BE.)
The Nan-shan Highlands overlook Tsaidam on the N.E. They embrace a region 380 m. long and 260 m. wide, entirely occupied with parallel mountain ranges all running from the N.W. to the S.E. Broad, flat, longitudinal valleys, at altitudes of 12,000 to 14,000 ft. (9000 to 10,000 at the south-western border) and dotted with lakes (Koko-nor, 9970 ft. ; Khara-nor, 13,285 ft.), fill up the space between these mountain ranges. In the S.E. the Nan-shan highlands abut upon the highlands of the Chinese province of Kansuh, and near the great northward bend of the Hwang-ho they meet the escarpments by which the Great Khingan and the In-shan ranges are continued, and by which the Mongolian plateau steps down to the lowlands of China. On the N.E. the Nan-shan highlands have their foot on the Mongolian plateau (average altitude, 4000 ft.), i.e. in the Ala-shan. On the N.W. they are fringed by a border range, the Da-sue-shan, a continuation of the Astin-tagh, which rises to 12,200-13,000 ft. in its passes, and is pierced by several rivers flowing west to Lake Khala-chi or Khara-nor. This border-range, which continues on to the 97th meridian, separates the Nan-shan range from the Pe-shan range.
On the S.W. the Nan-shan mountains consist of short irregular chains, separated by broad plains, dotted with lakes, which differ but slightly in altitude from Tsaidam (8800-9000 ft.). Next a succession of narrow ranges intervene between this lower border terrace and the higher terrace (12,000-13,500 ft.). The first mountain range on this higher terrace is Ritter's range, covered in part with extensive snow-fields. The passes at both ends of this snow-clad massif lie at altitudes of 15,990 ft. and 14,680 ft. The next range is Humboldt or Ama-surgu range, which runs N.W. to S.E. from the Astin-tagh to about 38 N., and is perhaps continued by the southern Kuku (Koko)-nor range, which strikes the Hwangho with an elevation of 7440 ft. It includes, in fact, several other parallel ranges e.g. the Mushketov, Semenov, Suess, Alexander III., Bain-sarlyk the mutual relations of which are, however, not yet definitely settled.
Small lateral chains of mountains, rising some 2000 ft. above the general level of that plateau, connect the central Nan-shan with the next parallel ranges, namely, those of the eastern Nan-shan. The mutual relations of the latter, as well as the names of the several constituent chains, are equally unsettled. Thus, one of them is named indiscriminately Nan-shan, Richthofen Range and Momoshan. In fact, the region is dominated by three ranges of nearly equal altitude, all lifting many of their peaks above the snow-line. Finally, there is a range of mountains, about 10,000 ft. high, named Lung-shan by Obruchev, which borders the Kan-chow and Lianchow valley on the N.E., and belongs to the Nan-shan system. But the string of oases in Kan-suh province, which stretches between the towns named, lies on the lower level of the Mongolian plateau (4000 to 5000 ft.), so that the Lung-shan ought possibly to be regarded as a continuation of the Pe-shan mountains of the Gobi.
Generally speaking, the Nan-shan highlands are a region raised 12,000 to 14,000 ft. above the sea, and intersected by wild, stony and partly snow-clad mountains, towering another 4000 to 7000 ft. above its surface, and arranged in narrow parallel chains all running N.W. to S.E. The chains of mountains are severally from 8 to 17 m. wide, seldom as much as 35, while the broad, flat valleys between them attain widths of 20 to 27 m. As a rule the passes are at an altitude of 12,000 to 14,000 ft., and the peaks reach 18,000 to 20,000 ft. in the western portion of the highlands, while in the eastern portion they may be about 2000 ft. lower. The glaciers also attain a greater development in the western portion of the Nan-shan, but the valleys are dry, and the slopes of both the mountains and the valleys, furrowed by deep ravines, are devoid of vegetation. Good pasture grounds are only found near the streams. The soil is dry gravel and clay, upon which bushes of Epkedra, Nitraria and Salsolaceae grow sparsely. In the northeastern Nan-shan, on the contrary, a stream runs through each gorge, and both the mountain slopes and the bottoms of the valleys are covered with vegetation. Forests of conifers (Picea obovata) and deciduous trees Przhevalsky's poplar, birch, mountain ash, etc., and a variety of bushes are common everywhere. Higher up, in the picturesque gorges, grow rhododendrons, willows, Polentilla fruticosa, Spriaeae, Lonicereae, etc., and the rains must evidently be more copious and better distributed. In the central Nan-shan it is only the north-eastern slopes that bear forests. In the south, where the Nan-shan enters Kan-suh province, extensive accumulations of loess make their appearance, and it is only the northern slopes of the hills that are clothed with trees. (P. A. K.)
AUTHORITIES. An enumeration of the works published before 1890, and a map of itineraries, will be found in Wegencr's Versuch einer Orographie des Kuen-lun (Marburg, 1891), but his map is only approximately correct. Of the books published since 1890 the most important are Svcn Hedin's Scientific Results of a Journey in Central Asia, /5pp-/pO2 (Stockholm, 1905-1907, 6 vols.), with an elaborate atlas and a general map of Tibet on the scale of 1 : 1 ,000,000 ; H. H. P. Deasy's In Tibet and Chinese Turkestan (London, 1901), with a good map; F. Grenard's vol. (iii.) of J. L. Dutreuil de Rhins's Mission scientifique dans la haute Asie, iSpo-iSpS (n.p., 1897), also with a very useful map; W. W. Rockhill's Diary of a Journey through Mongolia and Tibet in l8<}i and 1892 (Washington, 1894); M. S. Wellby's Through Unknown Tibet (London, 1898); P. G. Bonvalot's De Paris au Tonkin a travers le Tibet inconnu (Paris, 1892) ; St G. R. Littledale's " A Journey across Tibet," in Geog. Journal (May 1896) ; H. Bower's Diary of a Journey across Tibet (London, 1894); the Izvestia of the Russian Geog. Soc. and Geog. Journal, both passim.
Note - this article incorporates content from Encyclopaedia Britannica, Eleventh Edition, (1910-1911)