the Chola range, for the most part, consists of gneiss. This, again, has been carefully ascertained by Mr. Blanford. Of course Kanchanjanga, in the distance, is for the most part composed of granite; certainly all the upper part is so composed. The upper part of the Chola range is absolutely sterile; not a tree, not a shrub, grows in the locality-nothing but a few kinds of Alpine grasses and scanty herbage, which are dried up in most seasons of the year.

Our next picture (3) represents the Chokham Lake. Whilst the Chola Lake lies exactly on the boundary between Sikkim and Chumbi, and therefore between the Indian and the Chinese empires, the Chokham Lake is situated a little above it and just within the Chinese or Tibetan border. I say Chinese or Tibetan," for Tibet is virtually Chinese. The precise relations between the Grand Lama of Tibet and the Emperor of China I need hardly at this moment undertake to explain, but that part of Tibet is thoroughly dominated by China. Tibet has a local government of its own, no doubt, but its affairs are controlled by a Chinese resident supported by Chinese troops, and the Chinese are particularly careful to put boundary marks along the border. They do this in the most jealous and particular manner possible. Wherever I went, my footsteps were dogged by Chinese and Tibetan officials. Even if I stopped to take a sketch, these gentlemen were always considering whether my foot was upon one side of the border or the other. On one side of the wooden boundary-pillars was an inscription written in the Hindi character of British India, and on the other side an inscription in Chinese. When I went up to sketch the Chokham Lake, the Chinese local officers with their Tibetan attendants were very careful to warn me that I had transgressed the limits of my jurisdiction. However, we secured the sketch, and there it is, or at least the copy of it. This is decidedly the finest of all the lakes. I have been reminded by the high authority of our President that I ought not to call these sheets of water "lakes," and that they are more strictly "tarns,"

CHAP. IV. BEAUTY OF THE CHOKHAM LAKE.

79

because they have no outlets. In some respects they present geological problems; however, you will allow me to use "tarn" or "lake" as interchangeable. If the Chokham Lake is a tarn, it attains to the proportions of a lake, and is the loftiest and largest of the series. I do not know its dimensions, but it comprises an area of several square miles. The colour of the water is superb. The altitude is fully a thousand feet higher than that of the Chola Lake, and therefore must be nearly

FIG. 3.-CHOKHAM LAKE ABOVE CHOLA PASS. MOUNT CHANGU KANG IN THE DISTANCE.

16,000 feet above the level of the sea. We ourselves soon began to find that we were approaching a great altitude, because as we climbed the rocks we became short of breath, and our heads began to ache. However, when we surmounted the rocks we were more than rewarded by the splendour of the spectacle. Emerald, azure, turquoise-all these phrases combined will fail to give you an impression of the indescribable beauty of the water. In the background of our view will be

seen a tract of low forest which creeps along the base of the hills overhanging the Chumbi Valley, and beyond the valley rise a series of purple, pinkish mountains. I need not say that, to a spectator on the hills themselves, their colour would be of the dullest and most opaque yellow ochre; but the effect of distance in this clear atmosphere is to throw a sort of etherealised pinkpurple over the mountains, which has a lovely effect. Beyond the range of hills rises the snowy mountain of Changu Kang, a sort of pyramid. This Changu Kang is not found on the maps, but is very well known in that locality, and it separates Bhutan from Tibet.

On the southern spur of the Chola range lies the subject of our next picture (4), the Bhewsa Lake, which is much lower than those which I have been describing. It may be considered as having an altitude of 12,000 feet, being perhaps 4000 feet below the Chokham Lake. In the foreground near that lake, especially on the left-hand side of the picture, you will see indications of

CHAP. IV. ASPECT OF MOUNT KANCHANJANGA.

81

vegetation. These represent vast expanses of scrub rhododendron, a sort of rhododendron that grows very low and very thick, and spreads its branches in a tangled mass over the ground. It is a dreadfully difficult scrub for a pedestrian to push through; but it is very valuable to the traveller, because it supplies him with fuel for his fire, which he needs in the cold, and cannot otherwise get, and in the event of extreme want of water it will supply him even with moisture, because its leaves have upon them the hoar frost which can be wiped off, and be made to afford relief to thirsty men, as I and my staff sometimes found out. This scrub is the only vegetation met with in this zone, lying above the limits of trees. The Bhewsa Lake has a purplish-violet colour. I am unable to explain the reason of this variety of colour in the lakes, but the Chola Lake was principally blue, the Chokham Lake a mixture of emerald, azure, and turquoise, and Bhewsa Lake somewhat violet and purple. Behind the Bhewsa Lake, in the middle background, rise the gneiss rocks overhanging the Chomnaga Valley. In our illustration of the Chola Lake these rocks are seen in shadow: here, in this illustration, they are in the full blaze of the setting sun, and beyond them Mount Kanchanjanga overhanging the lake is seen with the same evening effect. Instead of the glittering white in the blaze of the morning sun, it is now tinged with the pink and rosy hues of

sunset.

The bad weather, which we generally had, was interspersed with lucid intervals of glorious blue skies, and of course I selected those lucid intervals for making my sketches. Generally speaking, what really happens is this: early in the morning, at sunrise, the weather is quite superb; the sky is unclouded azure, and the mountains are unbroken white. This lasts for some three hours, that is to say, till about 10 o'clock in the day, and that is what we used to call the bloom of the morning. Then, and then only, can you take your sketches. The air is extremely cold-biting cold-and after sketching for

G

a short time your fingers get perfectly numbed, and the only thing to do is to keep a supply of hot water close at hand, into which you can put your fingers, and so get a certain amount of warmth in them, which enables you to preserve their cunning for sketching. After 10 o'clock up come the clouds. You cannot tell how they form themselves. A little bit of vapour, no bigger than a man's hand, expands; fresh men's hands arise and clouds accumulate, till at last the whole atmosphere is clouded over. This lasts till about middle day. Then the clouds seem to turn into snow, and a certain amount of snow falls in the afternoon, which makes you very miserable in the evening, and you sit down to dinner with snow all around you, and your little tent also encrusted with snow. But towards midnight the clouds pass away, and stars come out, and it is a magnificent night. Then you have the sunrise as already described. Probably the sun when it rises will melt the thin snow during the bloom of the day. With that kind of weather you very seldom get a sunset view; but on that particular evening when I was at the Bhewsa Lake, the clouds somehow were lifted to display the setting sun, literally bathing Kanchanjanga in roseate light; and that happy moment, of course, I seized to make my sketch.

The Bhewsa Lake being comparatively low, we had to ascend from there once more towards the crest of the range, towards the point you will find named on the map as the Yakla Pass. On that ascent we found three very little lakes, for which I will adopt our President's designation of tarns. The first of these is overhung by some very weird, strange-looking rocks, piled one over the other, as it were by Titanic and Cyclopean hands; and in the back the view is not towards Tibet, but towards Darjiling. Leaving this tarn, No. I. (Fig. 5), we get to tarn No. II. (Fig.6). In this latter you will observe a very pointed hill tipped with This is the peak, well-known locally as Dopenti, which geographers will find marked on Stanford's map as Dobendi, another name for the same hill. Dopenti peak is the highest

snow.