[6] p. 3.—“The Desert near the basaltic mountains of Harudsh.”

Near the Egyptian Natron Lakes, (which in the time of Strabo had not yet been divided into six reservoirs), there is a range of hills which rises steeply on the northern side, and runs from east to west past Fezzan, where it finally appears to join the chain of the Atlas. It divides in north-eastern Africa, as the Atlas does in north-western Africa, the inhabited maritime Lybia of Herodotus from the land of the Berbers, or Biledulgerid, abounding in wild animals. From the limits of Middle Egypt the whole region south of the 30th degree of North latitude is a sea of sand, in which are dispersed islands, or Oases, containing springs of water and a flourishing vegetation. The number of these Oases, of which the ancients only reckoned three, and which Strabo compared to the spots on a panther’s skin, has been considerably augmented by the discoveries of modern travellers. The third Oasis of the ancients, now called Siwah, was the Nomos of Ammon; a residence of priests, a resting place for caravans, and the site of the temple of the horned Ammon and the supposed periodically cool fountain of the Sun. The ruins of Ummibida, (Omm-Beydah), belong incontestibly to the fortified caravanserai at the temple of Ammon, and therefore to the most ancient monuments which have come down to us from the early dawn of civilization. (Caillaud, Voyage à Syouah, p. 14; Ideler in den Fundgruben des Orients, Bd. iv. S. 399–411).

The word Oasis is Egyptian, and synonymous with Auasis and Hyasis (Strabo, lib. ii. p. 130, lib. xvii. p. 813, Cas.; Herod. lib. iii. cap. 26, p. 207, Wessel). Abulfeda calls the Oases, el-Wah. In the later times of the Cæsars, malefactors were sent to the Oases; being banished to these islands in the sea of sand, as the Spaniards and the English have sent criminals to the Falklands or to New Holland. Escape by the ocean is almost easier than through the desert. The fertility of the Oases is subject to diminution by the invasion of sand.

The small mountain-range of Harudsh is said to consist of basaltic hills of grotesque form (Ritter’s Afrika, 1822, S. 885, 988, 993, and 1003). It is the Mons Ater of Pliny; and its western extremity or continuation, called the Soudah mountains, has been explored by my unfortunate friend, the adventurous traveller Ritchie. This eruption of basalt in tertiary limestone, rows of hills rising abruptly from dike-like fissures, appears to be analogous to the outbreak of basalt in the Vicentine territory. Nature often repeats the same phenomena in the most distant parts of the earth. In the limestone formations of the “white Harudsh” (Harudje el-Abiad), which perhaps belong to the old chalk, Hornemann found an immense number of fossil heads of fish. Ritchie and Lyon remarked that the basalt of the Soudah mountains, like that of the Monte Berico, was in many places intimately mixed with carbonate of lime,—a phenomenon probably connected with eruption through limestone strata. Lyon’s map even mentions dolomite in the neighbourhood. Modern mineralogists have found syenite and greenstone in Egypt, but not basalt. Possibly the material of some of the ancient Egyptian vases, which are occasionally found of true basalt, may have been taken from these western mountains. May “Obsidius lapis” also have been found there? or are basalt and obsidian to be sought for near the Red Sea? The strip of volcanic or eruptive formations of the Harudsh, on the margin of the African desert, reminds the geologist of the augitic vesicular amygdaloid, phonolite, and greenstone porphyry, which are only found at the northern and western boundaries of the Steppes of Venezuela and of the plains of the Arkansas, as it were on the hills of the ancient coast line. (Humboldt, Relation Historique, tom. ii. p. 142; Long’s Expedition to the Rocky Mountains, vol. ii. pp. 91 and 405.)

[7] p. 3.—“When suddenly deserted by the east wind of the tropics in a sea covered with weed.”

It is a remarkable phenomenon, well known among sailors, that in the vicinity of the African coast (between the Canaries and the Cape de Verde Islands, and particularly between Cape Bojador and the mouth of the Senegal), a west wind often takes the place of the general east or trade-wind of the tropics. It is the wide expanse of the desert of Sahara which causes this westerly wind. The air over the heated sandy plain becomes rarefied and ascends, the air from the sea rushes in to supply the void so formed, and thus there sometimes arises a west wind, adverse to ships bound to the American coast, which are made in this manner to feel the vicinity of the heat-radiating desert without even seeing the continent to which it belongs. The changes of land and sea breezes, which blow alternately at certain hours of the day or night on all coasts, are due to the same causes.

The accumulation of sea-weed in the neighbourhood of the African coast has been often spoken of by ancient writers. The locality of this accumulation is a problem which is intimately connected with our conjectures respecting the extent of Phœnician navigation. The Periplus, which has been ascribed to Scylax of Caryanda, and which, according to the researches of Niebuhr and Letronne, was very probably compiled in the time of Philip of Macedon, describes beyond Cerne a quantity of fucus forming a weed-covered sea—a kind of “Mar de Sargasso;” but the locality indicated appears to me to differ very much from that assigned in the work entitled “De Mirabilibus Auscultationibus,” which long bore, unduly, the great name of Aristotle. (Compare Scyl. Caryand. Peripl. in Hudson, vol. ii. p. 53, with Aristot. de Mirab. Auscult. in opp. omnia ex. rec. Bekkeri, p. 844, § 136.) The pseudo-Aristotle says, “Phœnician mariners, driven by the east wind, came in four days’ sail from Gades to a part where they found the sea covered with reeds and sea-weed (θρύον καὶ φῦκος.) The sea-weed is uncovered at ebb and covered at flood tide.” Is he not here speaking of a shallow place between the 34° and 36° of latitude? Has a shoal disappeared in consequence of volcanic eruption? Vobonne speaks of rocks north of Madeira. (Compare also Edrisi, Geog. Nub., 1619, p. 157.) In Scylax it is said, “The sea beyond Cerne is unnavigable on account of its great shallowness, its muddiness, and the great quantity of sea grasses. The sea grass lies a span thick, and is full of points at the top, so that it pricks.” The sea-weed found between Cerne,—(the Phœnician station for laden vessels, Gaulea, or, according to Gosselin, the small island of Fedallah, on the north-western coast of Mauritania),—and Cape de Verde, does not now by any means form a great sea meadow, or connected tract of fucus, a “mare herbidum,” such as exists beyond the Azores. In the poetic description of the coast by Festus Avienus, (Ora Maritima, v. 109, 122, 388, and 408), in the composition of which, as Avienus himself says, (v. 412) he availed himself of the journals of Phœnician ships, the obstacle presented by the sea-weed is referred to in a very circumstantial manner; but its site is placed much farther north, towards Ierne, the “Sacred Island.”

In remarking that the fucus and the mud or mire, (πηλός), the shallowness of the sea, and the perpetual calms, are always spoken of by the ancients as characteristics of the western ocean beyond the Pillars of Hercules, one is disposed, more particularly on account of the mention of the calms, to ascribe something to Punic artifice,—to the desire of a great trading people to deter others, by the apprehension of dangers and difficulties, from entering into competition with them in western navigation and commerce. But even in the genuine writings of Aristotle (Meteorol. ii. p. 1, 14,) he maintains this same opinion of the absence of wind in those regions, and seeks the explanation of what he erroneously supposes to be a fact of observation, but which is more properly a fabulous mariner’s tale, in an hypothesis concerning the depth of the sea. In reality, the stormy sea between Gades and the islands of the Blest or Fortunate Islands, (between Cadix and the Canaries), is very unlike the sea farther to the south between the tropics, where the gentle trade winds blow, and which is called very characteristically by the Spaniards, el Golfo de las Damas, the Ladies’ Gulf. (Acosta Historia natural y moral de las Indias, lib. iii. cap. 4.)

From very careful researches by myself, and from the comparison of the logs or journals of many English and French vessels, I infer that the old and indefinite expression, Mar de Sargasso, includes two banks of fucus, of which the greater and easternmost one, of a lengthened shape, is situated between the parallels of 19° and 34° N. lat., in a meridian of 7 degrees to the west of the Island of Corvo, one of the Azores; while the lesser and westernmost bank, of a roundish form, is situated between the Bermudas and the Bahamas, (lat. 25°-31°, long. 66°-74°.) The longer axis of the small bank which is crossed by ships going from Baxo de Plata (Caye d’Argent, Silver Cay) on the north of St. Domingo, to the Bermudas, appears to have a N. 60° E. direction. A transverse band of Fucus natans, running in an East and West direction between the parallels of 25° and 30°, connects the greater and lesser banks. I have had the gratification of seeing these inferences approved by my honoured friend Major Rennell, and adopted by him in his great work on Currents, where he has further supported and confirmed them by many new and additional observations. (Compare Humboldt, Relation Historique, tom. i. p. 202, and Examen critique, tom. iii. p. 68–99, with Rennell’s Investigation of the Currents of the Atlantic Ocean, 1832, p. 184.) The two groups of sea-weed, included together with the transverse connecting band under the old general name of the Sargasso Sea, occupy altogether a space exceeding six or seven times the area of Germany.

Thus it is the vegetation of the ocean which offers the most remarkable example of an assemblage of “social plants” of a single species. On terra firma, the savannahs or prairies, or grassy plains of America, the heaths (ericeta), and the forests of the North of Europe and Asia, consisting of coniferous trees, birches, and willows, offer a less degree of uniformity than do those thalassophytes. Our heaths show, in the north, in addition to the prevailing Calluna vulgaris, Erica tetralix, E. ciliaris, and E. cinerea; and in the south, Erica arborea, E. scoparia, and E. mediterranea. The uniformity of the aspect offered by the Fucus natans is greater than that of any other assemblage or association of plants. Oviedo calls the fucus banks “meadows,” praderias de yerva. Considering that the island of Flores was discovered in 1452, by Pedro Velasco, a native of the Spanish port of Palos, by following the flight of certain birds from the island of Fayal, it seems almost impossible, seeing the proximity of the great fucus bank of Corvo and Flores, that a part of these oceanic meadows should not have been seen before Columbus, by Portuguese ships driven by storms to the westward. Yet the astonishment of the companions of Columbus in 1492, when surrounded by sea-weed uninterruptedly from the 16th of September to the 8th of October, shews that the magnitude of the phenomenon at least was previously unknown to the sailors. The anxieties excited by the accumulation of sea-weed, and the murmurs of his companions in reference thereto, are not indeed mentioned by Columbus in the extracts from the ship’s journal given by Las Casas. He merely speaks of the complaints and murmurs respecting the danger to be feared from the weak but constant East winds. It is only the son, Fernando Colon, who, in writing his father’s life, endeavoured to depict the fears of the sailors in a dramatic manner.

According to my researches, Columbus crossed the great fucus bank in 1492, in lat. 28½°, and in 1493, in lat. 37°, both times in the long. of from 38° to 41° W. This is deducible with tolerable certainty from Columbus’s recorded estimation of the ship’s rate, and the “distance daily sailed over;” derived indeed, not from casting the log, but from data afforded by the running out of half-hour sand-glasses (ampolletas). The first certain and definite mention of a log (catena della poppa) which I have been able to discover, is in the year 1521, in Pigafetta’s journal of Magellan’s Voyage round the World. (Cosmos, vol. ii. p. 259, and Note 405, English ed.) The determination of the ship’s place, while Columbus was engaged in traversing the great meadows of sea-weed, is the more important, because we learn from it that for three centuries and a half the situation of this great accumulation of thalassophytes, whether resulting from the local character of the bottom of the sea, or from the direction of the Gulf stream, has remained the same. Such evidences of the permanency of great natural phenomena arrest the attention of the physical inquirer with double force, when they present themselves in the ever-moving oceanic element. Although the limits of the fucus banks oscillate considerably, in correspondence with the variations of the strength and direction of the prevailing winds, yet we may still in the middle of the 19th century take the meridian of 41° W. from Paris (38° 38′ W. from Greenwich) as the principal axis of the “great bank.” In the vivid imagination of Columbus, the idea of the position of this bank was intimately connected with the great physical line of demarcation, which, according to him, divided the globe into two parts, with the changes of magnetic variation, and with climatic relations. Columbus, when uncertain respecting his longitude, (February 1493), directed himself by the appearance of the first floating streamers of weed (de la primera yerva) on the eastern margin of the great Corvo bank. The physical line of demarcation was, by the powerful influence of the Admiral, converted on the 4th of May, 1493, into a political line, being made the celebrated “line of demarcation” between the Spanish and Portuguese rights of possession. (Compare my Examen Critique, tom. iii. p. 64–99, and Cosmos, English ed. vol. ii. p. 279–280.)

[8] p. 3.—“The Nomadic Tibbos and Tuaricks.”

These two nations inhabit the deserts between Bornou, Fezzan, and Lower Egypt. They were first made known to us with some exactness by Hornemann’s and Lyon’s travels. The Tibbos or Tibbous roam through the eastern, and the Tuaticks (Tueregs) through the western, parts of the great desert. The first are called by the other tribes, from being in continual movement, “birds.” The Tuaricks are distinguished into those of Aghadez and those of Tagazi. They are often engaged as conductors of caravans, and in trade. Their language is the same as that of the Berbers; and they belong unquestionably to the number of the primitive Lybian nations. The Tuaricks present a remarkable physiological phenomenon. Different tribes among them are, according to the climate, white, yellowish, and even almost black; but all are without woolly hair or Negro features. (Exploration scientifique de l’Algérie, T. ii. p. 343.)

[9] p. 4.—“The Ship of the Desert.”

In oriental poems, the camel is called the land-ship, or the ship of the Desert (Sefynet-el-badyet); Chardin, Voyages, nouv. éd. par Langlès, 1811, T. iii. p. 376.

But the camel is not merely the carrier of the desert, and the link which, rendering communication between different countries possible, connects them with each other: he is also, as Carl Ritter has shewn in his excellent memoir on the sphere of diffusion of these animals, the principal and essential condition of the nomadic life of nations in the patriarchal stage of national development, in the hot parts of our planet where rain is either altogether wanting or very infrequent. No animal’s life is so closely associated by natural bonds with a particular stage of the developement of the life of man,—a connection historically established for several thousand years,—as the life of the camel among the Bedouin tribes (Asien, Bd. viii. Abth. i. 1847, S. 610 und 758). “The camel was entirely unknown to the cultivated Carthaginian nation through all the centuries of their flourishing existence, until the destruction of their city. The Marusians first brought it into military use, in the train of armies, in Western Lybia, in the times of the Cæsars; perhaps in consequence of its employment in commercial operations in the valley of the Nile by the Ptolemies. The Guanches, the inhabitants of the Canary Islands and probably related to the Berber race, were not acquainted with the camel before the 15th century, when it was introduced by Norman conquerors and settlers. In the probably very limited communication of the Guanches with the Coast of Africa, the small size of the boats would prevent the transport of large animals. The true Berber race, diffused throughout the interior of Northern Africa, and to which the Tibbos and Tuaricks, as already mentioned, belong, owes doubtless to the use of the camel throughout the Lybian desert and its Oases, not only the advantages of intercommunication, but also the preservation of its national existence to the present day. On the other hand, the negro races never, of their own accord, made any use of the camel; it was only in company with the conquering expeditions and proselyting missions of the Bedouins, carrying their prophet’s doctrines over the whole of Northern Africa, that the useful animal of the Nedjid, of the Nabatheans, and of all the countries inhabited by Aramean races, spread to the westward and was introduced among the black population. The Goths took camels as early as the fourth century to the Lower Istros (the Danube), and the Ghaznevides conveyed them in much larger numbers as far as India and the banks of the Ganges.” We must distinguish two epochs in the diffusion of the camel throughout the northern part of the African continent; one under the Ptolemies, operating through Cyrene on the whole of the north-west of Africa; and the Mohammedan epoch of the conquering Arabs.

It has long been a question, whether those domestic animals which have been the earliest companions of mankind—oxen, sheep, dogs, and camels—are still to be met with in a state of original wildness. The Hiongnu, in Eastern Asia, belong to the nations who earliest tamed and trained wild camels as domestic animals. The compiler of the great Chinese work, Si-yu-wen-kien-lo, (Historia Regionum occidentalium, quæ Si-yu vocantur, visu et auditu cognitarum,) affirms that in the middle of the 18th century wild camels, as well as wild horses and wild asses, still wandered in East Turkestan. Hadji Chalfa, in his Turkish Geography, written in the 17th century, speaks of the frequent chase of the wild camel in the high plains of Kashgar, Turfan, and Khotan. Schott translates, from a Chinese author, Ma-dschi, that wild camels are to be found in the countries to the north of China and west of the Hoang-ho, in Ho-si or Tangut. Cuvier alone (Règne Animal, T. i. p. 257), doubts the present existence of wild camels in the interior of Asia. He believes they have merely “become wild;” because Calmucks, and others having Buddhistic religious affinities with them, set camels and other animals at liberty, in order “to acquire to themselves merit for the other world.” According to Greek witnesses of the times of Artemidorus and Agatharchides of Cnidus, the Ailanitic Gulf of the Nabatheans was the home of the wild Arabian camel. (Ritter’s Asien, Bd. viii. s. 670, 672, and 746.) The discovery of fossil camel bones of the ancient world by Captain Cautley and Doctor Falconer, in 1834, in the sub-Himalaya range of the Sewalik hills, is peculiarly deserving of notice. These bones were found with other ancient bones of mastodons, of true elephants, of giraffes, and of a gigantic land tortoise (Colossochelys), twelve feet in length and six feet in height. (Humboldt, Cosmos, Engl. ed. vol. i. p. 268.) This camel of the Ancient World has received the name of Camelus sivalensis, but does not show any considerable difference from the still living Egyptian and Bactrian camels with one and two humps. Forty camels have very recently been introduced into Java, having been brought there from Teneriffe. (Singapore Journal of the Indian Archipelago, 1847, p. 206.) The first experiment has been made in Samarang. In like manner, reindeer have only been introduced into Iceland from Norway in the course of the last century. They were not found there when the island was settled, notwithstanding the proximity to East Greenland, and the existence of floating masses of ice. (Sartorius von Waltershausen physisch-geographische Skizze von Island, 1847, S. 41.)

[10] p. 4.—“Between the Altai and the Kuen-lün.”

The great highland, or, as it is commonly called, the mountain plateau of Asia, which includes the lesser Bucharia, Songarei, Thibet, Tangut, and the Mogul country of the Chalcas and Olotes, is situated between the 36th and 48th degrees of latitude, and the meridians of 81° and 118° E. long. It is an erroneous view to represent this part of the interior of Asia as a single undivided mountainous gibbosity, continuous like the elevated plains of Quito and Mexico, and elevated from seven to nine thousand feet above the level of the sea. That there is not in this sense any undivided mountain plateau in the interior of Asia, has already been shewn by me in my “Researches respecting the Mountains of Northern India.” (Humboldt, Premier Mémoire sur les Montagnes de l’Inde, in the Annales de Chimie et de Physique, T. iii. 1816, p. 303; Second Mémoire, T. xiv. 1820, p. 5–55.)

My views concerning the geographical range of plants, and the mean degree of temperature requisite for certain kinds of cultivation, had early led me to entertain considerable doubts as to the continuity of a great Tartarian plateau between the Himalaya and the Altai. Writers continued to characterise this plateau as it had been described by Hippocrates (De Ære et Aquis, § xcvi. p. 74), as “the high and naked plains of Scythia, which, without being crowned with mountains, rise and extend to beneath the constellation of the Bear.” Klaproth has the undeniable merit of having been the first to make us acquainted with the true position, extent, and direction of two great and entirely distinct chains of mountains—the Kuen-lün and the Thian-schan, in a part of Asia which is better entitled to the name of “central” than Kashmeer, Baltistan, and the Sacred Lakes of Thibet, (the Manasa and the Ravanahrada). The importance of the Celestial Mountains, the Thian-schan, had indeed been already surmised by Pallas, without his being aware of their volcanic nature; but this highly-gifted investigator of nature, hampered by the then prevailing hypothesis of a dogmatic and fantastic geology, firmly believing in “chains of mountains radiating from a centre,” saw in the Bogdo Oola (the Mons Augustus, or culminating point of the Thian-schan) such a “central node, from whence all the Asiatic mountain chains diverge in rays, and which dominates over all the rest of the continent!”

The erroneous idea of a single vast elevated plain occupying the whole of central Asia, the “Plateau de la Tartarie,” took its rise in France, in the latter half of the 18th century. It was the result of historical combinations, and of a not sufficiently attentive study of the writings of the celebrated Venetian traveller, as well as of the naïve relations of those diplomatic monks who, in the 13th and 14th centuries, (thanks to the unity and extent of the Mogul empire at that time) were able to traverse almost the whole of the interior of the continent, from the ports of Syria and of the Caspian Sea to the shores of the Pacific on the east coast of China. If a more exact acquaintance with the language and ancient literature of India had dated farther back among us than half a century, the hypothesis of this central plateau, occupying the wide space between the Himalaya and the south of Siberia, would no doubt have had adduced in its support an ancient and venerable authority from that source. The poem of the Mahabharata appears, in the geographical fragment Bhischmakanda, to describe “Meru” not so much as a mountain as an enormous elevation of the land, which supplies with water at once the sources of the Ganges, those of the Bhadrasoma (Irtysh), and those of the forked Oxus. These physico-geographical views were intermingled in Europe with ideas of other kinds, and with mythical reveries relating to the origin of mankind. It was said that the elevated regions from which the waters first retreated, (geologists in general were long averse to the theory of elevation), must also have received the first germs of civilisation. Hebraizing systems of geology, and views connected with the Deluge and supported by local traditions, favoured these assumptions. The intimate connection between time and space, between the beginnings of social order and the plastic character of the surface of the earth, lent to the supposed “uninterrupted Plateau of Tartary” a peculiar importance, and an almost moral interest. Acquisitions of positive knowledge, the late matured fruit of scientific travels and direct measurements, as well as of a fundamental study of Asiatic languages and literature especially those of China, have gradually demonstrated the inaccuracies and exaggerations of those wild hypotheses. The mountain plains (ὁροπέδια) of Central Asia are no longer regarded as the cradle of civilization and the primitive seat of all arts and sciences. The ancient nation of Bailly’s Atlantis, happily described by d’Alembert as “having taught us everything but their own name and existence,” has vanished. The supposed inhabitants of the Oceanic Atlantis had already been treated, in the time of Posidonius, in a no less derisive manner. (Strabo, lib. ii. p. 102; and lib. xiii. p. 598, Casaub.)

A plateau of considerable but very unequal elevation, having the names of Gobi, Scha-mo (sand desert), Scha-ho (sand river), and Hanhai, runs in a SSW.-NNE. direction, with little interruption, from Eastern Thibet towards the mountain knot of Kentei south of Lake Baikal. This swelling of the ground is probably anterior to the elevation of the mountain chains by which it is intersected; it is situated, as already remarked, between 79° and 116° long. from Paris, (81° and 118° E. from Greenwich). Measured at right angles to its longitudinal axis, its breadth is, in the south between Ladak, Gertop, and H’lassa, (the seat of the great Lama,) 720 geographical miles; between Hami in the Celestial Mountains, and the great bend of the Hoang-ho near the In-schan chain, hardly 480; and in the north, between the Khanggai, where the great city of Karakhorum once stood, and the chain of Khin-gan-Petscha, which runs north and south (in the part of the Gobi traversed in travelling from Kiachta by Urga to Pekin) 760 geographical miles. The whole extent of this swelling ground, which must be carefully distinguished from the far more elevated mountain range to the east, may be approximately estimated, taking its inflections into account, at about three times the area of France. The map of the mountain ranges and volcanoes of Central Asia (Carte der Bergketten und Vulkane von Central-Asien), constructed by me in 1839, but not published until 1843, shows in the clearest manner the hypsometric relations between the mountain ranges and the Gobi plateau. It was founded on the critical employment of all the astronomical determinations accessible to me, and on a vast amount of orographic description, in which Chinese literature is beyond measure rich, examined at my request by Klaproth and Stanislas Julien. My map marks the mean direction and the height of the mountain chains, and represents the leading features of the interior of the continent of Asia, from 30° to 60° degrees of north latitude, and between the meridians of Kherson and Pekin. It differs materially from any previously published map.

The Chinese have enjoyed a threefold advantage towards the collection of so great an amount of orographic data in the highlands of Asia, and more especially in the regions (hitherto so little known in the west), north and south of the Celestial mountains, between the In-schan, the mountain lake Khuku-noor, and the banks of the Ili and the Tarim. The three advantages I allude to are,—the military expeditions towards the west, (under the dynasties of Han and Thang 122 years before our era, and again in the ninth century when conquerors advanced as far as Ferghana and to the borders of the Caspian), together with the more peaceful conquests of Buddhistic pilgrims;—the religious interest attaching to certain lofty mountain summits on account of sacrifices to be periodically offered there;—and the early and general use of the compass in giving the directions of mountains and of rivers. The knowledge and use of the “South pointing” of the magnetic needle twelve centuries before our era, has given to the orographic and hydrographic descriptions of countries by the Chinese, a great superiority over the descriptions of the same kind which Greek or Roman writers have bequeathed to us, and which are besides extremely few. The acute and sagacious Strabo, was alike imperfectly acquainted with the direction of the Pyrenees, and with those of the Alps and of the Apennines. (Compare Strabo, lib. ii. p. 71 and 128; lib. iii, p. 137; lib. iv. p. 199 and 202; lib. v. p. 211, Casaub.)

To the lowlands belong almost the whole of Northern Asia to the north-west of the volcanic chain of the Thian-schan;—the Steppes to the north of the Altai and of the Sayan chain;—the countries which extend from the mountains of Bolor, or Bulyt-Tagh, (“cloud mountains” in the Uigurian dialect) which follow a north and south direction, and from the upper Oxus, (whose sources were found by the Buddhistic pilgrims Hiuen-thsang and Song-yun in 518 and 629, by Marco Polo in 1277, and by Lieutenant Wood in 1838, in the Pamer Lake, Sir-i-kol, Lake Victoria), towards the Caspian; and from Tenghir or the Balkhash Lake through the Kirghis Steppe, towards the sea of Aral and the southern extremity of the Ural mountains. As compared with high plains of 6,000 to 10,000 feet above the level of the sea, it may well be permitted to use the expression of “lowlands” for flats of little more than 200 to 1200 feet of elevation. The lowest of the last two numbers corresponds nearly to the altitude of the town of Mannheim, and the highest to that of Geneva and Tubingen. If the word plateau, so often misemployed in modern works on geography, is to have its use extended to elevations which hardly present any sensible difference in climate and vegetation, the indefiniteness of the expressions “highlands and lowlands,” which are only relative terms, will deprive physical geography of the means of expressing the idea of the connection between elevation and climate, between the profile or relief of the ground and the decrease of temperature. When I found myself in Chinese Dzungarei, between the boundary of Siberia and Lake Dsaisang, at an equal distance from the Icy Sea and from the mouth of the Ganges, I might well consider myself in Central Asia. The barometer, however, soon taught me that the plains through which the Upper Irtysh flows, between Ustkamenogorsk and the Chinese Dzungarian Post, Chonimailachu, (sheep-bleating,) are scarcely raised 850, or at the most 1170, feet above the level of the sea. Pansner’s older barometric measurements (which, however, were not published until after my expedition), are confirmed by mine. Both refute the hypothesis of Chappe, relative to the supposed high elevation of the banks of the Irtysh, in Southern Siberia; an hypothesis based on estimations of river declivities. Even further to the East, Lake Baikal is only 222 toises, or 1420 English feet, above the level of the sea.

In order to connect the idea of the relation of the terms lowlands and highlands and of the various gradations in the height of elevated plains or undulating grounds, with actual examples ascertained by measurement, I have subjoined a table, forming an ascending scale of such districts in different parts of the Globe. What I have said above respecting the mean height of those Asiatic plains, which I have termed lowlands, may be compared with the following numbers:—

No portion of the so-called Desert of Gobi (parts of which contain fine pastures) has been so thoroughly explored in respect to the differences of elevation as the zone, of nearly 600 geographical miles in breadth, between the sources of the Selenga and the great Wall of China. A very exact series of barometric levellings was executed under the auspices of the Academy of St. Petersburgh by two distinguished Savans, the astronomer George Fuss, and the botanist Bunge. In the year 1832 they accompanied the mission of Greek monks to Pekin, to establish there one of the magnetic stations recommended by me. The mean height of this part of Gobi does not amount, as had been too hastily inferred from the measurement of neighbouring summits by the Jesuits Gerbillon and Verbiest to from 7500 to 8000 French (8000 to 8500 English) feet, but only to little more than half that height, or barely 4000 French or 4264 English feet. Between Erghi, Durma, and Scharaburguna, the ground is only 2400 French, or 2558 English, feet above the level of the sea, or hardly 300 French (320 English) feet higher than the plateau of Madrid. Erghi is situated midway, in lat. 45° 31′, long. 111° 26′ E. from Greenwich. There is here a depression of more than 240 miles in breadth, in a SW. and NE. direction. An ancient Mogul tradition marks it as the bottom of a former inland sea. There are found in it reeds and saline plants, mostly of the same kinds as those on the low shores of the Caspian. In this central part of the desert there are small salt lakes, from which salt is carried to China. According to a singular opinion very prevalent among the Moguls, the ocean will one day return and establish its empire anew in Gobi. One is reminded of the Chinese tradition of the bitter lake, in the interior of Siberia, mentioned by me in another work. (Humboldt, Asie Centrale, tom. ii. p. 141; Klaproth, Asia Polyglotta, p. 232.) The valley or basin of Kashmeer, so enthusiastically extolled by Bernier, and but too moderately praised by Victor Jacquemont, has also given occasion to great hypsometric exaggerations. By a careful barometrical measurement, Jacquemont found the height of the Wulur Lake in the valley of Kashmeer, not far from the chief city Sirinagur, 836 toises, or 5346 English feet. Uncertain determinations by the boiling point of water gave Baron Carl von Hügel a result of 910, and Lieutenant Cunningham only 790 toises. (Compare my Asie Centrale, tom. iii. p. 310, with the Journal of the Asiatic Society of Bengal, vol. x. 1841, p. 114.) Kashmeer,—respecting which, in Germany particularly, so much interest has been felt, but the delightfulness of whose climate is considerably impaired by four months of winter snow in the streets of Sirinagur (Carl von Hügel, Kaschmir, Bd. ii. S. 196),—is not situated, as is often supposed, upon the ridge of the Himalaya, but is a true cauldron-shaped valley (Kesselthal, Caldera,) on the southern declivity of those mountains. On the south-west, where the rampart-like elevation of the Pir Panjal separates it from the Punjaub, the snow-covered summits are crowned, according to Vigne, with formations of basalt and amygdaloid. The latter formation has received from the natives the characteristic name of “schischak deyu,” marked by the devil’s small-pox. (Vigne, Travels in Kashmeer, 1842, vol. i. p. 237–293.) The beauty of its vegetation has from the earliest times been very differently described, according as the visitor came from the rich and luxuriant vegetation of India, or from the northern regions of Turkestan, Samarcand, and Ferghana.

It is also only very recently that clearer views have been obtained respecting the elevation of Thibet; the level of the plateau having long been most uncritically confounded with the summits which rise from it. Thibet occupies the interval between the two great chains of the Himalaya and the Kuen-lün, forming the raised ground of the valley between them. It is divided from east to west, both by the natives and by Chinese geographers, into three portions. Upper Thibet, with its capital city H’lassa, probably 1500 toises (9590 English feet) above the level of the sea;—Middle Thibet, with the town of Leh or Ladak (1563 toises, or 9995 English feet);—and Little Thibet, or Baltistan, called the Thibet of Apricots, (Sari Boutan), in which are situated Iskardo (985 toises, or 6300 English feet), Gilgit, and south of Iskardo but on the left bank of the Indus, the plateau of Deotsuh, measured by Vigne, and found to be 1873 toises, or 11,977 English feet. On examining all the notices that we possess respecting the three Thibets, (and which will have received in the present year a rich augmentation by the boundary expedition under the auspices of the governor-general, Lord Dalhousie), we soon become convinced that the region between the Himalaya and the Kuen-lün is no unbroken plain or table land, but that it is intersected by mountain groups, undoubtedly belonging to wholly distinct systems of elevation. There are, properly speaking, very few plains; the most considerable are those between Gertop, Daba, Schang-thung (Shepherd’s Plain) the native country of the Shawl-goat, and Schipke (1634 toises, 10,450 English feet);—those round Ladak, which have an elevation of 2100 toises, or 13430 English feet, and must not be confounded with the depression in which the town is situated;—and lastly, the plateau of the Sacred Lakes Manasa and Ravanahrada (probably 2345 toises), which was visited so early as 1625 by Pater Antonio de Andrada. Other parts are entirely filled with crowded mountainous elevations, “rising,” as a recent traveller expresses it, “like the waves of a vast ocean.” Along the rivers, the Indus, the Sutlej, and the Yaru-dzangbo-tschu which was formerly regarded as identical with the Brahma-putra, points have been measured which are only between 1050 and 1400 toises (6714 and 8952 English feet) above the level of the sea; so also with respect to the Thibetian villages of Pangi, Kunawur, Kelu, and Murung. (Humboldt, Asie Centrale, T. iii. p. 281–325.) From many carefully collected measurements of elevation I think I may conclude that the plateau of Thibet, between 73° and 85° E. long., does not reach a mean height of 1800 toises (11510 English feet); this is hardly equal to the height of the fertile plain of Caxamarca in Peru, and is 211 and 337 toises (1350 and 2154 English feet) less than the height of the plateau of Titicaca, and the street pavement of the Upper Town of Potosi (2137 toises, 13,665 English feet).

That outside of the Thibetian highlands and of the Gobi, the boundaries of which have been defined above, there are in Asia, between the parallels of 37° and 48°, considerable depressions and even true lowlands, where one boundless uninterrupted plateau was formerly imagined to exist, is shewn by the cultivation of plants which cannot thrive without a certain degree of heat. An attentive study of the travels of Marco Polo, in which the cultivation of the vine and the production of cotton in northern latitudes are spoken of, had long called the attention of the acute Klaproth to this point. In a Chinese work, entitled “Information respecting the recently-subdued Barbarians (Sin-kiang-wai-tan-ki-lio),” it is said, “the country of Aksu, somewhat to the south of the Celestial Mountains (the Thian-schan), near the rivers which form the great Tarim-gol, produces grapes, pomegranates, and numberless other excellent fruits; also cotton (Gossypium religiosum), which covers the fields like yellow clouds. In the summer the heat is exceedingly great, and in winter there is here, as at Turfan, neither severe cold nor heavy snow.” The district round Khotan, Kashgar, and Yarkand, still pays its tribute in home-grown cotton as it did in the time of Marco Polo. (Il Milione di Marco Polo, pubbl. dal Conte Baldelli, T. i. p. 32 and 87.) In the Oasis of Hami (Khamil), above 200 miles east of Aksu, orange trees, pomegranates, and vines whose fruit is of a superior quality, grow and flourish.

The products of cultivation which are thus noticed imply the existence of only a small degree of elevation, and that over extensive districts. At so great a distance from any coast, and in those easterly meridians where the cold of winter is known to exceed that of corresponding latitudes nearer our own part of the world, a plateau which should be as high as Madrid or Munich might indeed have very hot summers, but would hardly have, in 43° and 44° latitude, extremely mild winters with scarcely any snow. Near the Caspian, 83 English feet below the level of the Black Sea, at Astrachan in 46° 21′ lat., I saw the cultivation of the vine greatly favoured by a high degree of summer heat; but the winter cold is there from -20° to -25° Cent. (-4° to -13° Fahr.) It is therefore necessary to protect the vines after November, by sinking them deep in the earth. Plants which live, as we may say, only in the summer, as the vine, the cotton bush, rice, and melons, may indeed be cultivated with success between the latitudes of 40° and 44° on plains of more than 500 toises (3197 English feet) elevation, being favoured by the powerful radiant heat; but how could the pomegranate trees of Aksu, and the orange trees of Hami, whose fruit Père Grosier extolled as distinguished for its goodness, bear the cold of the long and severe winter which would be the necessary consequence of a considerable elevation of the land? (Asie Centrale, T. ii. p. 48–52, and 429.) Carl Zimmerman (in the learned Analysis of his “Karte von Inner Asien,” 1841, S. 99) has made it appear extremely probable that the Tarim depression, i. e. the desert between the mountain chains of the Thian-schan and the Kuen-lün, where the Steppe river Tarim-gol empties itself into the Lake of Lop, which used to be described as an alpine lake, is hardly 1200 (1279 English) feet above the level of the sea, or only twice the height of Prague. Sir Alexander Burnes also assigns to that of Bokhara only an elevation of 1190 English feet. It is earnestly to be desired, that all doubt respecting the elevation of the plateaux of middle Asia, south of 45° of latitude, should finally be set at rest by direct barometric measurements, or by determinations of the boiling point of water made with more care than is usually given to them. All our calculations respecting the difference between the limits of perpetual snow, and the maximum elevation of vine cultivation in different climates, rest at present on too complex and uncertain elements.

In order to rectify in the smallest space that which was said in the last edition of the present work, relatively to the great mountain systems which intersect the interior of Asia, I subjoin the following general review. We begin with the four parallel chains, which follow with tolerable regularity an east and west direction, and are connected with each other at a few detached points by transverse elevations. Differences of direction indicate, as in the Alps of western Europe, a difference in the epoch of elevation. After the four parallel chains (the Altai, the Thian-schan, the Kuen-lün, and the Himalaya), we have to notice chains following the direction of meridians, viz. the Ural, the Bolor, the Khingan, and the Chinese chains, which, with the great bend of the Thibetian and Assamo-Bermese Dzangbo-tschu, run north and south. The Ural divides a part of Europe but little elevated above the level of the sea from a part of Asia similarly circumstanced. The latter was called by Herodotus, (ed. Schweighaüser, T. v. p. 204) and even as early as Pherecydes of Syros, a Scythian or Siberian Europe, including all the countries to the north of the Caspian and of the Iaxartes; in this view it would be a continuation of Europe “prolonged to the north of Asia.”

1. The great mountain system of the Altai, (the “gold mountains” of Menander of Byzantium, an historical writer who lived as early as the 7th century, the Altaï-alin of the Moguls, and the Kin-schan of the Chinese), forms the southern boundary of the great Siberian lowlands; and running between 50° and 52½° of north latitude, extends from the rich silver mines of the Snake Mountains, and the confluence of the Uba and the Irtysh, to the meridian of Lake Baikal. The divisions and names of the “Great” and the “Little Altai,” taken from an obscure passage of Abulghasì, are to be altogether avoided. (Asie Centrale, T. i. p. 247.) The mountain system of the Altai comprehends (a) the Altai proper, or Kolywanski Altai, the whole of which is under the Russian sceptre; it is west of the transverse opening of the Telezki Lake, which follows the direction of the meridian; and in ante-historic times probably formed the eastern shore of the great arm of the sea, by which, in the direction of the still existing groups of lakes, Aksakal-Barbi and Sary-Kupa (Asie Centrale, T. ii. p. 138), the Aralo-Caspian basin was connected with the Icy sea:—(b) East of the Telezki chain which follows the direction of the meridian, the Sayani, Tangnu, and Ulangom or Malakha chains, all running tolerably parallel with each other and in an east and west direction. The Tangnu, which sinks down and terminates in the basin of the Selenga, has from very ancient times formed a boundary between the Turkish race to the south and the Kirghis (Hakas, identical with Σάκαι) in the north. (Jacob Grimm, Gesch. der deutschen Sprache, 1848, Th. i. S. 227.) It is the original seat of the Samoieds or Soyotes, who wandered as far as the Icy Sea, and who were long regarded in Europe as a nation belonging exclusively to the coasts of the Polar Sea. The highest snow-clad summits of the Altai of Kolywan are the Bielucha and the Katunia-Pillars. The height of the latter is about that of Etna. The Daurian highland, to which the mountain knot of Kemtei belongs, and on the eastern side of which is the Jablonoi Chrebet, divides the depressions of the Baikal and the Amur.

2. The mountain system of the Thian-schan, or Celestial Mountains, the Tengri-tagh of the Turks (Tukiu) and of the kindred race of the Hiongnu, is eight times as long, in an east and west direction, as the Pyrenees. Beyond,—i. e. west of its intersection with the transverse or north and south chain of the Bolor and Kosuyrt, the Thian-schan bears the names of Asferah and Aktagh, is rich in metals, and has open fissures, which emit hot vapours, luminous at night, and which are used for obtaining sal-ammoniac. (Asie Centrale, T. ii. p. 18–20.) East of the transverse Bolor and Kosyurt chain, there follow successively in the Thian-schan,—the Kashgar Pass (Kaschgar-dawan); the Glacier Pass of Djeparle, which leads to Kutch and Aksu in the Tarim basin; the volcano of Pe-schan, which sent forth fire and streams of lava at least as late as the middle of the seventh century; the great snow-covered massive elevation Bogdo-Oola; the Sol-fatara of Urumsti, which furnishes sulphur and sal-ammoniac (nao-scha), and is situated in a coal district; the still active volcano of Turfan (or volcano of Ho-tscheu or Bischbalik), almost midway between the meridians of Turfan (Kune-Turpan), and of Pidjan. The volcanic eruptions of the Thian-schan chain, recorded by Chinese historians, reach as far back as the year 89, A.D., when the Hiongnu of the sources of the Irtysh were pursued by the Chinese army as far as Kutch and Kharaschar (Klaproth, Tableau hist. de l’Asie, p. 108). The Chinese General, Teu-hian, surmounted the Thian-schan, and saw “the Fire Mountains which send out masses of molten rock that flow for many Li.”

The great distance from the sea of the volcanoes of the interior of Asia is a remarkable and solitary phenomenon. Abel Rémusat, in a letter to Cordier (Annales des Mines, T. v. 1820, p. 137), first directed the attention of geologists to this fact. The distance, for example, in the case of the volcano of Pe-schan, to the north, or to the Icy Sea at the mouth of the Obi, is 1528 geographical miles; to the south, or to the mouths of the Indus and the Ganges, 1512 geographical miles; to the west, 1360 geographical miles to the Caspian in the Gulf of Karaboghaz; and to the east, 1020 geographical miles to the shores of the sea of Aral. The active volcanoes of the New World were previously supposed to offer the most remarkable instances of such phenomena at a great distance from the sea; their distance, however, is only 132 geographical miles in the case of the volcano of Popocatepetl in Mexico, and only 92, 104, and 156 geographical miles in those of the South American volcanoes Sangai, Tolima, and de la Fragua, respectively. I exclude from these statements all extinct volcanoes, and all trachytic mountains which have no permanent connection with the interior of the earth. (Asie Centrale, T. ii. p. 16–55, 69–77, and 341–356.) East of the volcano of Turfan, and of the fertile Oasis of Hami rich in fine fruit, the chain of the Thian-schan gives place to the great elevated tract of Gobi which follows a S.W. and N.E. direction. This interruption of the mountain chain, caused by the transverse intersection of the Gobi, continues for more than 9½ degrees of longitude; but beyond it the mountains recommence in the somewhat more southerly chain of the In-schan, or the Silver Mountains, running (north of the Pe-tscheli) from west to east almost to the shores of the Pacific near Pekin, and forming a continuation of the Thian-schan. As I have viewed the In-schan as an easterly prolongation (beyond the interruption of the Gobi) of the cleft above which the Thian-schan stands, so one might possibly view the Caucasus as a westerly prolongation of the same, beyond the great basin of the Aral and Caspian Seas or the depression of Turan. The mean parallel of latitude or axis of elevation of the Thian-schan oscillates between 40⅔° and 43° N. lat.; that of the Caucasus, according to the map of the Russian Etat-Major (running rather E.S.E. and W.N.W.), is between 41° and 44° N. lat. (Baron von Meyendorff, in the Bulletin de la Societé géologique de France, T. ix. 1837–1838, p. 230.) Of the four parallel chains which traverse Asia from east to west, the Thian-schan is the only one in which no summits have yet had their elevation above the sea determined by measurement.

3. The mountain system of the Kuen-lün (Kurkun or Kulkun), if we include in it the Hindu-Coosh and its western prolongation in the Persian Elbourz and Demavend, is, next to the American Cordillera of the Andes, the longest line of elevation on the surface of our planet. Where the north-and-south chain of Bolor intersects the Kuen-lün at right angles, the latter takes the name of the Thsung-ling (Onion Mountains), which is also given to a part of the Bolor at the eastern angle of intersection. The Kuen-lün, forming the northern boundary of Thibet, runs very regularly in an east and west direction, in the latitude of 36°. In the meridian of H’lassa an interruption takes place from the great mountain knot which surrounds the alpine lake of Khuku-noor, the Sing-so-hai, or Starry Sea, so celebrated in the mythical geography of the Chinese. The somewhat more northerly chains of Nan-schan and Kilian-schan may almost be regarded as an easterly prolongation of the Thian-schan. They extend to the Chinese wall near Liang-tscheu. West of the intersection of the Bolor and Kuen-lün (the Thsung-ling) I think I have been the first to shew (Asie Centrale, T. i. p. 23, and 118–159; T. ii. p. 431–434 and 465) that the corresponding direction of the axes of the Kuen-lün and the Hindu-Coosh (both being east and west, whereas the Himalaya is south-east and north-west) makes it reasonable to regard the Hindu-Coosh as a continuation, not of the Himalaya, but of the Kuen-lün. From the Taurus in Lycia to Kafiristan, through an extent of 45 degrees of longitude, this chain follows the parallel of Rhodes, or the diaphragm of Dicearchus. The grand geognostical view of Eratosthenes (Strabo, Lib. ii. p. 68; Lib. xi. p. 490 and 511; and Lib. xv. p. 689), which is farther developed by Marinus of Tyre, and Ptolemy, and according to which “the continuation of the Taurus in Lycia extends across the whole of Asia to India, in one and the same direction,” appears to have been partly founded on statements which reached the Persians and Indians from the Punjaub. “The Brahmins affirm,” says Cosmas Indicopleustes, in his Christian Topography, (Mountfauçon, Collectio nova Patrum, T. ii. p. 137) “that a line drawn from Tzinitza (Thinæ) across Persia and Romania, exactly cuts the middle of the inhabited earth.” It is deserving of notice that Eratosthenes had so early remarked that this longest axis of elevation in the Old Continent, in the parallels of 35½° and 36°, points directly through the basin (or depression) of the Mediterranean to the Pillars of Hercules. (Compare Asie Centrale, T. i. p. 23 and 122–138; T. ii. p. 430–434, with Kosmos, Bd. ii. S. 222 and 438, p. 188, and note 292, Engl. ed.) The easternmost part of the Hindu-Coosh is the Paropanisus of the ancients, the Indian Caucasus of the companions of Alexander. The now generally used term of Hindu-Coosh, belongs, as may be seen from the Travels of the Arab Ibn Batuta (English version, p. 97), to a single mountain pass on which many Indian slaves often perished from cold. The Kuen-lün, like the Thian-schan, shews igneous outbreaks or eruptions at many hundred miles from the sea. Flames, visible at a great distance, issue from a cavity in the Schin-khieu Mountain. (Asie Centrale, T. ii. p. 427 and 483, where I have followed the text of Yuen-thong-ki, translated by my friend Stanislas Julien.) The highest summit measured in the Hindu-Coosh, north-west of Jellalabad, is 3164 toises above the sea (20132 English feet); to the west, towards Herat, the chain sinks to 400 toises (2558 English feet), until, north of Teheran, it rises again to a height of 2295 toises (14675 English feet) in the volcano of Demavend.

4. The mountain system of the Himalaya. The normal direction of this system is east and west when followed from 81° to 97° E. long. from Greenwich, or through more than fifteen degrees of longitude from the colossal Dhawalagiri (4390 toises, 28071 English feet) to the breaking through of the long-problematical Dzangbo-tschu river (the Irawaddy, according to Dalrymple and Klaproth), and to the chains running north and south which cover the whole of Western China, and in the provinces of Sse-tschuan, Hu-kuang, and Kuang-si form the great mountain group of the sources of the Kiang. The next highest culminating point to the Dhawalagiri, of this east and west part of the Himalaya, is not, as has been hitherto supposed, the eastern peak of the Schamalari, but the Kinchinjinga. This mountain is situated in the meridian of Sikhim, between Bootan and Nepaul, and between the Schamalari (3750? toises, 23980 English feet) and the Dhawalagiri: its height is 4406 toises, or 26438 Parisian, or 28174 English feet. It was first measured accurately by trigonometrical operations in the present year, and as the account of this measurement received by me from India says decidedly, “that a new determination of the Dhawalagiri leaves to the latter the first rank among all the snow-capped mountains of the Himalaya,” the height of the Dhawalagiri must necessarily be greater than that of 4390 toises, or 26340 Parisian, 28071 English feet, hitherto ascribed to it. (Letter of the accomplished botanist of Sir James Ross’s Antarctic Expedition, Dr. Joseph Hooker, written from Dorjiling, July 25, 1848.) The turning point in the direction of the axis of the Himalaya range is not far from the Dhawalagiri, in 79° E. long. from Paris (81° 22′ Greenwich). From thence to the westward the Himalaya no longer runs east and west, but from SE. to NW., connecting itself, as a great cross vein, between Mozuffer-abad and Gilgit south of Kafiristan, with a part of the Hindu-Coosh. Such a bend or change in the direction or strike of the axis of elevation of the Himalaya (from E-W. to SE-NW.), doubtless points, as in the western part of our European Alps, to a difference in the age or epoch of elevation. The course of the Upper Indus, from the sacred lakes Manasa and Ravana-hrada (at an elevation of 2345 toises, 14995 English feet) in the vicinity of which the great river rises, to Iskardo and to the plateau of Deo-tsuh, (at an elevation of 2032 toises, 12993 English feet) measured by Vigne, follows in the Thibetian highlands the same north-westerly direction as the Himalaya. Here is the summit of the Djawahir, long since well measured and known to be 4027 toises (25750 English feet) in elevation, and the valley of Kashmeer, where at an elevation of only 836 toises, (5346 English feet), the Wulur Lake freezes every winter, and, from the perpetual calm, no wave ever curls its surface.

Having thus described the four great mountain systems of Asia, which in their normal geognostic character are chains coinciding with parallels of latitude, I have next to speak of the series of elevations coinciding nearly with meridians, (or more precisely, having a SSE.-NNW. direction), which, from Cape Comorin opposite to the Island of Ceylon to the Icy Sea, alternate between the meridians of 66° and 77° E. long. from Greenwich. To this system, of which the alternations remind us of faults in veins, belong the Ghauts, the Soliman chain, the Paralasa, the Bolor, and the Ural. The interruptions of the series of elevations are so arranged that, beside their alternate position in respect to longitude, each new chain begins in a degree of latitude to which the preceding chain had not quite reached. The importance which the Greeks (although probably not before the second century) attached to these chains induced Agathodemon and Ptolemy (Tab. vii. and viii.) to represent to themselves the Bolor, under the name of Imaus, as an axis of elevation extending as far as 62° N. lat. into the low basin of the Lower Irtisch and the Obi. (Asie Centrale, T. i. p. 138, 154, and 198; T. ii. p. 367.)

As the perpendicular elevation of mountain summits above the level of the sea (unimportant as in the eyes of the geologist the circumstance of the greater or lesser corrugation of the crust of the earth may be), is still, like all that is difficult of attainment, an object of popular curiosity, the following historical notice of the gradual progress of hypsometric knowledge may here find a suitable place. When I returned to Europe in 1804 after a four years’ absence, not a single Asiatic snowy summit either in the Himalaya, the Hindu-Coosh, or the Caucasus, had been measured with any exactness; and I could not therefore compare my determinations of the height of perpetual snow in the Cordilleras of Quito, or the mountains of Mexico, with any corresponding determinations in the East. The important journey of Turner, Davis, and Saunders, to the highlands of Thibet, does indeed belong to the year 1783, but Colebrooke justly remarks, that the elevation given by Turner to the Schamalari (lat. 28° 57′, long. 89° 30′, a little to the north of Tassisudan) rests on foundations as slight as those of the so-called measurements of the heights seen from Patna and the Kafiristan by Colonel Crawford and Lieutenant Macartney. (Compare Turner, in the Asiatic Researches, vol. xii. p. 234, with Elphinstone’s Account of the Kingdom of Caubul, 1815, p. 95, and Francis Hamilton, Account of Nepal, 1819, p. 92.) The excellent observations and writings of Webb, Hodgson, Herbert, and the brothers Gerard, have thrown great and certain light on the elevation of the colossal summits of the Himalaya; yet, in 1808, the hypsometric knowledge of this great Indian chain was still so uncertain that Webb wrote to Colebrooke: “The height of the Himalaya still remains a problem. I find, indeed, that the summits visible from the high plain of Rohilcund are 21000 English feet above that plain, but we do not know the absolute height above the sea.”

It was not until the beginning of the year 1820 that it began to be reported in Europe, that not only were there in the Himalaya, summits much higher than those of the Cordilleras, but also that Webb had seen in the Pass of Niti, and Moorcroft in the Thibetian plateau of Daba and the Sacred Lakes, fine pastures and flourishing fields of corn, at altitudes far exceeding the height of Mont Blanc. These accounts were received in England with much incredulity, and were met by doubts respecting the influence of refraction. I have shown the groundlessness of these doubts in two memoirs (Sur les Montagnes de l’Inde), printed in the Annales de Chimie et de Physique. The Tyrolese jesuit, P. Tiefenthaler, who in 1766 penetrated into the provinces of Kemaun and Nepal, had already divined the importance of the Dhawalagiri. We read on his map, “Montes Albi, qui Indis Dolaghir, nive obsiti.” Captain Webb always uses the same name. Until the measurements of the Djawahir (lat. 30° 22′, long. 79° 58′, altitude 4027 toises, or 25750 English feet) and of the Dhawalagiri (lat. 28° 40′, long. 83° 21′, altitude 4390? toises, 28072 English feet) were made known in Europe, the Chimborazo (3350 toises, or 21421 English feet), according to my trigonometric measurement, (Recueil d’Observations astronomiques, T. i. p. 73) was still everywhere regarded as the highest summit on the surface of the earth. The Himalaya now appeared, according as the comparison was made with the Djawahir or the Dhawalagiri, 676 toises (4323 English feet), or 1040 toises (6650 English feet), higher than the Chimborazo. Pentland’s South American travels, in the years 1827 and 1838, fixed attention (Annuaire du Bureau des Longitudes, 1830, p. 320 and 323) on two snowy summits of Upper Peru, east of the Lake of Titicaca, which were supposed to surpass the height of the Chimborazo respectively by 598 and 403 toises, (3824 and 2577 English feet.) I have remarked above, pp. 53–54, that the latest calculation of the measurements of the Sorata and Illimani shews this view to be incorrect. The Dhawalagiri (on the declivity of which, in the valley of the Ghandaki, the Salagrana Ammonites, so celebrated among the Brahmins as symbols of one of the incarnations of Vishnu, are collected) therefore still shews a difference between the culminating points of the Old and the New Continents of more than 6200 Parisian, or 6608 English feet.

The question has been raised, whether there may not exist behind the southernmost more or less perfectly measured chain, other still greater elevations. Colonel George Lloyd, who in 1840 edited the important observations of Captain Alexander Gerard and his brother, entertains an opinion that in the part of the Himalaya which he calls somewhat vaguely “the Tartaric chain,” (meaning therefore in north Thibet towards the Kuen-lün, and perhaps in Kailasa of the sacred lakes, or beyond Leh) there are summits of from 29000 to 30000 English feet,—one or two thousand feet higher therefore than the Dhawalagiri. (Lloyd and Gerard, Tour in the Himalaya, 1840, vol. i. p. 143 and 312; Asie Centrale, T. iii. p. 324.) So long as actual measurements are wanting, one cannot decide respecting such possibilities; as the indication, from which the natives of Quito, long before the arrival of Bouguer and La Condamine, recognised the superior altitude of the Chimborazo (namely, from the portion of its height above the region of perpetual snow being greater than in any of the other mountains), might prove very deceptive in the temperate zone of Thibet, where radiation is so active in the table-land, and where the lower limit of perpetual snow does not form a regular line at an equal elevation, as it does in the tropics. The greatest elevation above the level of the sea ever attained by human beings on the declivity of the Himalaya, is 3035 toises, or 18210 Parisian, or 19409 English feet, reached by Captain Gerard, with seven barometers, on the mountain of Tarhigang, a little to the north-west of Schipke. (Colebrooke, in the Transactions of the Geological Society, vol. vi. p. 411.) This happens to be exactly the same height as that reached by myself on the 23rd of June, 1802, and thirty years later by my friend Boussingault, on the 16th of December, 1831, on the declivity of the Chimborazo. The unattained summit of the Tarhigang is, however, 197 toises, or 1260 English feet, higher than that of the Chimborazo.

The passes across the Himalaya, leading from Hindostan into Chinese Tartary, or rather into Western Thibet, more particularly between the rivers of Buspa and Schipke or Langzing Khampa, are from 2400 to 2900 toises, or 15346 to 18544 English feet. In the chain of the Andes I found the pass of Assuay, between Quito and Cuenca on the Ladera de Cadlud, having a similar elevation, being 2428 toises, or 15526 English feet, high. A great part of the mountain plains of the interior of Asia would be buried throughout the year in perpetual snow and ice, if it were not, that by the great radiation of heat from the Thibetian plateau, by the constant serenity of the sky, by the rarity of the formation of snow in the dry atmosphere, and by the powerful solar heat peculiar to the eastern continental climate, the limit of perpetual snow is wonderfully raised on the northern slope of the Himalaya,—perhaps to 2600 toises, or 16625 English feet above the level of the sea. Fields of barley (Hordeum hexastichon) are seen in Kunawur up to 2300 toises, or 14707 English feet; and another variety of barley called Ooa, and allied to Hordeum cœleste, even much higher. Wheat succeeds extremely well in the Thibetian highlands up to 1880 toises, or 12022 English feet. On the northern declivity of the Himalaya, Captain Gerard found the upper limit of the higher birch woods ascend to 2200 toises, 14068 English feet; and small bushes which serve the inhabitants for fuel to warm their huts, attain, in the latitude of 30¾° and 31° of north latitude, a height of 2650 toises (16945 English feet), or almost 200 toises (1279 English feet) higher than the limit of perpetual snow under the equator. From the data hitherto collected it would follow, that we may take the lower limit of perpetual snow on the northern side of the Himalaya, on the average, and in round numbers, at 2600 toises, or about 16600 English feet; whilst on the southern declivity of the Himalaya the snow-line sinks to 2030 toises, or about 13000 English feet.

But for this remarkable distribution of temperature in the upper strata of the atmosphere, the mountain plain of Western Thibet would be uninhabitable to the millions who dwell there. (Compare my Examination of the Limit of Perpetual Snow on the two declivities of the Himalaya, in the Asie Centrale, T. ii. p. 435–437; T. iii. p. 281–326, and in Kosmos, Engl. ed. vol. i. note 403; S. 483 of the original.)

A letter which I have just received from India from Dr. Joseph Hooker, who is engaged in meteorological and geological researches, as well as those connected with the geography of plants, says: “Mr. Hodgson, who we regard here as the geographer best acquainted with the hypsometric relations of the snow ranges, completely recognises the correctness of your statement in the third part of the Asie Centrale, respecting the reason of the inequality in the height of the limit of perpetual snow on the northern and southern declivities of the Himalaya. In the ‘trans Sutlej region’ in 36° lat. we often saw the snow limit only commence at an altitude of 20000 English feet, while in the passes south of the Brahmaputra, between Assam and Burman, in 27° lat., where the most southern Asiatic snowy mountains are situated, the limit of perpetual snow sinks to 15000 English feet.” I believe we ought to distinguish between the extreme and the mean heights, but in both we see manifested in the clearest manner the formerly contested differences between the Thibetian and the Indian declivities.

The local differences vary still more, as may be seen from the list of extremes given in my Asie Centrale, T. iii. p. 295. Alexander Gerard saw the snow limit ascend, on the Thibetian declivity of the Himalaya, to 19200 Parisian feet (20465 English); and on the southern Indian declivity, Jacquemont once saw it, north of Cursali on the Jumnotri, even as low as 10800 Parisian (11,510 English) feet.