of the bulb of the attached thermometer broken off, and floating on the mercury. Having quite expected this, I always checked the observations of the attached thermometer by another, but--how, it is not easy to say--the broken one invariably gave a correct temperature. Illustration--VIEW FROM AN ELEVATION OF 18,000 FEET OF THE EAST TOP OF KINCHINJHOW, AND OF TIBET, OVER THE RIDGE THAT CONNECTS IT WITH DONKIA. WILD SHEEP (_OVIS AMMON_) IN THE FOREGROUND. The Kinchinjhow spurs are not accessible to so great an elevation as those of Donkia, but they afford finer views over Tibet, across the ridge connecting Kinchinjow with Donkia. Broad summits here, as on the opposite side of the valley, are quite bare of snow at 18,000 feet, though where they project as sloping hog-backed spurs from the parent mountain, the snows of the latter roll down on them and form glacial caps, the reverse of glaciers in valleys, but which overflow, as it were, on all sides of the slopes, and are ribboned* [The convexity of the curves, however, seems to be upwards. Such reversed glaciers, ending abruptly on broad stony shoulders quite free of snow, should on no account be taken as indicating the lower limit of perpetual snow.] and crevassed. On the 18th of September I ascended the range which divides the Lachen from the Lachoong valley, to the Sebolah pass, a very sharp ridge of gneiss, striking north-west and dipping north-east, which runs south from Kinchinjhow to Chango-khang. A yak-track led across the Kinchinjhow glacier, along the bank of the lake, and thence westward up a very steep spur, on which was much glacial ice and snow, but few plants above 16,000 feet. At nearly 17,000 feet I passed two small lakes, on the banks of one of which I found bees, a May-fly (_Ephemera_) and gnat; the two latter bred on stones in the water, which (the day being fine) had a temperature of 53 degrees, while that of the large lake at the glacier, 1000 feet lower, was only 39 degrees.