CHAPTER L

THE SOURCE OF THE SUTLEJ

We had scarcely dressed in the morning before the storm raged again. Galsan and a gova from Parka overtook us here. The former brought provisions, the latter had strict orders from his chief, Parka Tasam, to tell me that if I did not at once betake myself to Parka, he would send off all my baggage to Langak-tso, and force me to move on to Purang. But the gova himself was a jovial old fellow, and he received my answer that if Parka Tasam ventured to meddle with my boxes, he should be immediately deposed. If he kept quiet a couple of days, I would come to Parka, and the rather that I found it impossible to navigate the lake at this season of the year.

Then we marched on westwards, in and out of the bays and round all the projections produced by a mountain elevation north of Gurla, which prolongs its ramifications to the lake. The constantly changing views, as we wind in and out and wander between land and water, are indescribably beautiful and charming. The two large islands lying far out in the lake we see wherever we may be. One is named Dopserma; other water-birds breed there, but no geese. In winter yaks and sheep are driven over the ice to the island, where there is good pasturage. When cattle disease rages in the country the animals on Dopserma do not suffer.

We passed round the sharp-pointed westernmost bay in a furious storm and blinding clouds of sand, and encamped on the shore again. The same agreeable weather continued also on the last day of August as we travelled north-eastwards and saw the Langak-tso in a new and beautiful aspect. The air was now clear, Kang-rinpoche and Gurla Mandatta were unclouded, and stood as sentinels above the lakes. We passed the point where Tundup Sonam, Ishe, and I had waited so long, and by the sand-dune where we had lain four hours.

At the north-western bay we cross the old bed of the Sutlej, consisting of treacherous, quaking bog or dry, hard clay; it is broad, has no terraces, and has been much degraded and smoothed down by deflation and driftsand in later times. Two springs rise up in the middle, and flow in the direction of the lake. Westwards the bed seems quite level, but actually it rises slowly and evenly to a flat culmination, on the other side of which it dips down towards the Indus.

Now it had become dark, and we rode hour after hour among low hills and dunes and over meadows and water channels. I thought we had lost our way, when the bells of grazing cattle were heard, a fire appeared, and Rabsang came to meet us with a lantern in order to lead us to the village Parka, where my tent was set up in a courtyard.

During the much-needed day of rest we allowed ourselves in Parka, I negotiated now and then with the govas of the neighbourhood. They asked me to set off definitely for the west next day, and I promised to do so, but on the condition that I might stay three days in Khaleb, half a day's journey to the west. They consented without inquiring into my further intentions. I wished, be it known, to pass round the holy mountain by the pilgrim road, but saw that the authorities would never grant their permission. It could be done only by stratagem.

Here I received a second very kind letter from Mr. Cassels, who happened to be in Gyanima on official business. Unfortunately the force of circumstances prevented us meeting. He gave me a pleasant surprise with three packets of tea, which were the more welcome as I had latterly had to put up with brick-tea.

Here also the truth of the report that had so long followed us, that six Chinese and Tibetan officials from Lhasa had been sent to bring me to reason, was at length made clear. The report was certainly true, but when the gentlemen on reaching Saka-dzong heard that I had marched on westwards some time before, they simply turned back again.

I obtained all kinds of information about the two lakes and their periodical outlets, from Tibetans who had long lived in the country. Four years before some water had flowed from Tso-mavang to Langak-tso, which confirms Ryder's statement. Twelve years ago the outflow had been so abundant that the channel could not be passed except by the bridge. The channel is sometimes called Ngangga, sometimes Ganga. The water of Langak-tso is said to drain off underground, and to appear again at a place in the old bed called Langchen-kamba, and this water is said to be the true source of the Sutlej, and to find its way to the large streams which form this river, called in Tibetan Langchen-kamba. Twelve years and forty-eight years ago the spring in the old bed is said to have emitted much more water than now. Sherring collected similar data in 1905.

Langak-tso is said to have been so poisonous in former times that any one who drank of its water died, but since the holy fish broke through the isthmus and passed into the lake, the water has been sweet. Langak-tso freezes in the beginning of December, half a month sooner than its eastern neighbour, and the freezing proceeds slowly and in patches, whereas Tso-mavang freezes over in an hour. Langak-tso also breaks up half a month before Tso-mavang. Both have ice 3 feet thick. In winter the surface of Tso-mavang falls 20 inches beneath the ice, which consequently is cracked and fissured, and dips from the shore; but Langak-tso sinks only one or two thirds of an inch. This shows that it receives water constantly from the eastern lake, but only parts with a trifling quantity in winter.

With regard to the goose island, I was told that three men are commissioned by the Devashung to settle on the island as soon as the wild-geese arrive, to protect them from wolves and foxes. They receive 8 rupees, a sheep, and a lump of butter as wages. At this time, in May, the ice is still two feet thick, but the egg-gatherers must take care that they are not cut off from the mainland by a storm. Some years ago it happened that two watchmen were isolated on the island in this way. They lived there eight months, subsisting on eggs and green food, and returned over the ice next winter as soon as the lake was frozen over. But one of them was so enfeebled that he died on reaching Parka.

After a lively feast held by the Ladakis in the evening, we rode on September 2 north-westwards, accompanied by an old grey-headed gova, who had become a particular friend of mine. The weather was fine, but we now felt biting cold in the morning, much as at home on the islets off the coast when the yellow leaves have fallen and a thin sheet of ice has spread over the inlets. All Parka was on foot to witness our departure. With us set out a high lama whom I had known in Leh. His retinue looked well in their yellow dresses against the grey and green ground. He had been in Shigatse, and had lately made the circuit of the holy mountain. During the march we waded through the rivers Dam-chu, Sung-chu, La-chu, and Khaleb, which together carried about 350 cubic feet of water per second to Langak-tso.

The nearer we came to the holy mountain, the less imposing it appeared; it was finest from Langak-tso. In form it resembles a tetrahedron set on a prism. From the middle of its white top a belt of ice falls precipitously down, and below it stands a stalagmite of ice, on to which a thick stream of water pours from above. The stream splits up into glittering drops of spray and thin sheets of water--a grand spectacle, which one could watch with pleasure for hours.

Our camp on the Khaleb moor had the advantage of being far from the haunts of men--a very necessary condition, for here I contrived to make three excursions without permission. The second of these took a whole day, September 6, and its aim was the old bed of the Sutlej. Where we reached it, the bed seemed to contain stagnant water both to the east and west, and the ground was quite level. At the place which seemed highest, we tested it with the boiling-point thermometer and found that it stood about 30 feet above the lake. Following the bed westwards we come first to a large pool of sweet water with large quantities of ducks and water-weeds, then to a series of freshwater swamps connected by channels, and at length to a brook, which flows slowly south-westwards. The brook pours into a large freshwater pool, No. 2, which has no visible outlet. But when we proceed farther west to the point where the bed is contracted between walls of solid rock, we come upon two springs forming a new brook, which flows through a clearly marked valley to the south-west. I am convinced that this water filtrates underground from Langak-tso. A year later I followed the old bed a day's march farther west, and found at Dölchu-gompa permanent springs of abundant water, which likewise well up on the bottom of the bed. From here and all along its course through the Himalayas the Tibetans call the Sutlej Langchen-kamba, the Elephant river; the hill on which the convent Dölchu-gompa is built is supposed to bear some resemblance to an elephant, and hence the name. The spring at Dölchu is called Langchen-kabab, or the mouth out of which the Elephant river comes, just as the Brahmaputra source is the Tamchok-kabab, or the mouth out of which the Horse river comes, and the Indus source is the Singi-kabab, or the mouth from which the Lion river comes. The fourth in the series is the Mapchu-kamba, the Peacock river or Karnali. The Tibetans assert that the source of the Sutlej is at the monastery Dölchu, not in the Himalayas or the Trans-Himalaya, from which, however, it receives very voluminous tributaries. They are also convinced that the source water of the Langchen-kamba originates from Langak-tso. And I would draw particular attention to the fact that the first of the two holy springs which pour their water into the Tage-tsangpo is also called Langchen-kamba (see p. 105), a proof that in old times the source was supposed to lie to the east of Tso-mavang.

Now I advise any one who takes no interest in the source of the Sutlej to skip the following quotation. During my stay in Kioto in December 1908, Mr. Ogawa, Professor of Geography in the University there, showed me a collection of Chinese books. One of them, _Shui-tao-ti-kang_, or _The Elements of Hydrography_, is a compilation of the author Chi Chao Nan in the 26th year of the Emperor Kien Lung, that is, the year 1762, and in this work, Book 22, is the following communication concerning the source of the Sutlej, which Professor Ogawa was kind enough to translate for me literally:[1]

The Kang-ka-kiang comes out from Kang-ti-ssu-shan, on the south-east of which there stands Lang-chuan-ka-pa-pu-shan ( = Langchen-kabab), magnificent like an elephant. The relief is gradually accentuated more and more towards the south-western frontiers, and culminates at Kang-ti-ssu-shan ( = Kailas). The mountain has a circumference of more than 140 li. On all sides the mountain forms precipitous walls more than 1000 feet high above the surrounding mountains, and accumulated snow seems as if hung on cliffs. Hundreds of springs pour down from the top, but flow under the ground on the foot of the mountain. It is situated in the extreme west of the Tsang region, 310 li north-east of Ta-ko-la-cheng in A-li, more than 5590 li south-west of Si-ning-fu in Shensi province. Its longitude is 36° 4' W., and its latitude 30° 5' N. In olden times the place was unknown, but can be doubtfully identified with A-nok-ta-shan in the annotation of Shui-ching. In the neighbourhood there are four high mountains, of which the southern is called Lang-chuan-ka-pa-pu-shan, lying 250 li south-by-east of Kang-ti-ssu-shan, and 270 li east of Ta-kola-cheng. The natives call it so, because the form of the mountain resembles an elephant. On the east of this mountain there stands Ta-mu-chu-ko-ka-pa-pu-shan (= Tamchok-kabab), which is the source of the Ya-lu-tsang-pu river (= Yere-tsangpo or Brahmaputra). Springs come out from the northern foot of the mountain, and accumulate into a lake (35° 5' W., and 29° 1' N.). The water flows north-westwards for 70 li, and receives a stream coming from the north-east. The stream lies in the mountains 80 li north-east of Lang-chuan-ka-pa-pu. Two streams flow westwards from the mountain and turn north-westwards after their junction. It now takes a sinuous course for 60 li, turns south-westwards, and joins the main river. This is a source.

The river flows further to the west-by-north for 40 li, then to the north-east, to be met by the water of lake Kung-sheng (= Gunchu-tso), which sinks under the ground of the lake basin, but which, after reappearing, and after receiving three northern affluents, runs south-westwards to the river.

The lake of Kung-sheng-o-mo has two sources--one coming from the north-east, from Ta-ko-la-kung-ma-shan, and flowing 150-160 li; the other from the east, from the western foot of Ma-erh-yo-mu-ling ( = Marium-la) in the western frontiers of Cho-shu-tê. This last-mentioned mountain forms the eastern boundary of A-li, and is the chief range going south-eastwards from Kang-ti-ssu. The water (of the lake Kung-sheng) flows westwards for more than 50 li, and forms another lake, 80 li wide, and without an outlet. However, more than 10 li farther to the west there is a third lake with a subterranean source and with a length of 30 li. A stream comes from the north to the lake. The river now flows south-westwards for 60 li, and receives a stream coming from the north-east. 40 li farther south-westwards it receives a stream coming from the northern mountains; farther south-westwards the river meets the water from Lang-chuan-ka-pa-pu-shan.

The water forms the lake Ma-piu-mu-ta-lai ( = Tso-mavang). From south to north it is 150 li long, from east to west 80 or 100 li wide, and has a circumference of more than 200 li. On the northern side of the lake there are two streams coming from the north. The lake is situated 120 li to the south of Kang-ti-ssu. The water flows out from the west of the lake into the lake Lang-ka (= Lan-gak-tso) at a distance of 60 li. The latter lake receives a stream coming from the north-east. Lake Lang-ka has a narrow rectangular shape, pointed and elongated, the length from south to north being 170 li, and the width from east to west 100 li. Its northern pointed corner has the stream coming from north-east. There are three sources on the southern foot at a distance of 70 li from a southern branch of Kang-ti-ssu; they flow southwards, unite into one stream, which takes a south-westerly course for 150-160 li before entering the lake. The lake is the same in circumference and area, but different in outline.

The water (of lake Lang-ka) flows out from the west, and after running westwards for more than 100 li it turns to the south-west, and is now called the Lang-chu-ho, and takes a sinuous course for more than 200 li. Then it receives the Chu-ka-la-ho coming from the north-east.

This description of the position of the source of the Sutlej is of such extraordinary interest that I do not like to reserve it for my scientific work, and the less so that it supports the theory I expressed when in India, that the Tage-tsangpo is nothing but the uppermost section of the course of the Sutlej, or, in other words, that the source of the Tage-tsangpo is also that of the Sutlej. Many quotations have been looked up during the discussion that has arisen on this problem, but they cannot compare in importance with the one just cited, which, moreover, is sixty years older than the oldest of the others, namely, Gerard's opinion that the Gunchu-tso is the source of the Sutlej.

The description in Chi Chao Nan's _Hydrography_ is distinguished by the same careful conformity with the truth and conscientiousness as all other Chinese geographical descriptions. Compare the description of Kailas (Kang-rinpoche) with what I have already said about it.

Lang-chuan-ka-pa-pu is the Chinese translation of the Tibetan Langchen-kabab, which literally means the "Source of the Sutlej." When the Chinese author informs us that east of Langchen-kabab lies Tamchok-kabab, which is the source of the river Yere-tsangpo (Brahmaputra), we must admit that his description is quite in accordance with the truth, as I, the first European to visit this country, have myself discovered; for on the Tamlung-la I stood on the pass which parts the water between the Brahmaputra and Sutlej, and immediately to the south of the pass I saw Gang-lung-gangri and the glacier from which the Tage-tsangpo takes its rise, and in which the source of the Sutlej lies.

It is further said that the lake Gunchu-tso has two source streams--one from the north-east, from the mountain Ta-ko-la-kung-ma, which is evidently identical with D'Anville's Tacra-concla; the other from the west side of the pass Marium-la: an account which agrees with Ryder's map in all particulars. At present the Gunchu-tso is completely cut off and is salt; it therefore is no longer connected with the Sutlej system. But 147 years ago it had an outlet which ran partly underground, and then, rising up again, joined the Langchen-kamba or Tage-tsangpo. And that the Tage-tsangpo was at one time considered by the Tibetans to be the headwater of the Sutlej is apparent from the fact that its name, Langchen-kamba, is still applied to the upper of the two sacred source streams in the valley of the Tage-tsangpo.

And, again, it is said: This water, that is, the water of the Langchen-kabab, or the headwater of the Sutlej, forms the lake Ma-piu-mu-ta-lai, the Tso-mavang or Tso-mavam, as the name is also pronounced; on D'Anville's map (Map 2) it is written Ma-pama Talai, and D'Anville explains that Talai signifies lake. He might have added that it is the same word as in Dalai Lama, the priest, whose wisdom is as unfathomable as the ocean; for the Chinese word Talai or Dalai means ocean. By the use of this word the Chinese author wished to imply that Tso-mavang is much larger than the other lakes mentioned in his text.

The surplus water, as there is every reason to assume, flowed in the year 1762 from Tso-mavang through the channel to Langak-tso. The length of the channel was 60 li, which corresponds to my 5½ miles. All the northern tributaries which flow into the two lakes from the valleys of the Trans-Himalaya are correctly noted. The lake Langak is called Lang-ka. On D'Anville's map, the material for which was supplied by the Jesuits who lived in Pekin in the time of the Emperor Kang Hi (at the beginning of the eighteenth century), the lake is named Lanken. On the same map the river flowing thence westwards is called Lanc-tchou (Sutlej), but it is suggested, absurdly enough, that it is the upper course of the Ganges. D'Anville names the mountains south of Tso-mavang Lantchia-Kepou, which is Langchen-kabab, and the mountains lying to the south-east of them Tam-tchou, that is, Tamchok, in which he quite correctly places the origin of the Yarou Tsanpou, the Brahmaputra. The material for the map of the whole Chinese Empire, which the Jesuits presented to the Emperor Kang Hi in the year 1718, was collected between the years 1708 and 1716, and the Emperor procured information about Tibet through natives, who were prepared for their work by the Jesuits, just as in later times English topographers have trained Indian pundits.

From D'Anville's map we learn that 200 years ago the Sutlej flowed out of Langak-tso through the bed I have already described. Professor Ogawa's translation of the Chinese text shows us that even in the year 1762, or perhaps some years before, the river still emerged from the Langak-tso. And it is expressly said that the river Chu-ka-la-ho (Chu-kar, which, however, is said to descend from the north-east instead of the south-east) is only a tributary.

In the year 1846 Henry Strachey found no visible outlet, but he says that there is one underground, and considers it probable that the channel also may carry water when the lake has risen after heavy rains.

On July 30, 1908, I heard from the chief lama of the monastery Dölchu-gompa, who was born in the neighbourhood and was then fifty-five years old, that when he was quite young, water occasionally flowed out of the lake. But when he was ten years old, that would be in the year 1863, this water had failed, and since then no more had been seen. On the other hand, the springs in the bed are constant both in winter and in summer, and are independent of the precipitation. The monks believe that the water comes from Langak-tso, but nevertheless they call it the Langchen-kabab, the river which flows out of the mouth of the elephant.

My investigations on the spot, as well as the Chinese quotation, prove that Colonel S. G. Burrard is quite right in his masterly description of the rivers of the Himalayas and Tibet (Calcutta, 1907), when he includes Tso-mavang and Langak-tso and all their affluents in the drainage basin of the Sutlej, and therefore I will here cite two sentences of Colonel Burrard:

The connection between the two lakes may be taken as established, but that between the western lake and the Sutlej basin is still open to question. If the water from Rakas-tal flows into the Sutlej once a century, and then only for such a short period as to be observed by no one, we shall still be justified in including the lakes in the catchment area of the river.

And in this connection I would point out that the water-level of Tibetan rivers and lakes is subject to periodical fluctuations, dependent on the precipitation, of the same kind as the Brückner periods. The level in the two lakes varies from year to year. At the present time they are very low, but there is nothing to prevent them rising gradually in a more or less distant future. Tso-mavang may rise so that its water may again flow through the channel to Langak-tso, and this lake at length may discharge its surplus water, as formerly, through the dry bed of the Sutlej. It is more probable, however, that Langak-tso is approaching a time when it will lose its subterranean outlet also, and be quite isolated, like Gunchu-tso and Panggong-tso, and consequently become salt in time. But after it has lost its outlet it may be a long time, as Professor Brückner informs me, before the lake becomes noticeably salt. The next step in the development will be that Tso-mavang will be cut off from Langak-tso and likewise become salt.

However, we need not plunge into speculations and prognostications of the future, which may have surprises in store about which we can form only more or less probable conjectures. It is our duty to rely solely on fact and observation.

And now that we are agreed that the two lakes belong to the drainage area of the Sutlej, the question is: Which of the rivers debouching into Tso-mavang is the headwater of the Sutlej? Naturally, the longest and the one which carries most water. The river which once flowed out of Gunchu-tso has no claim to this honour, and the Gunchu-tso must be rejected as the source of the Sutlej. The Tage-tsangpo discharged 388 cubic feet of water per second, while all the other streams entering Tso-mavang carried at most 100 cubic feet each. The source of the Tage-tsangpo in the front of the Gang-lung glacier is therefore the source of the Sutlej.

FOOTNOTE:

[1] I have only omitted a couple of sentences, which have no immediate connection with the problem.