temperature of a given place on the sea-level, one can obtain with accuracy the difference in height between the two points. Two aneroid barometers were also carried, which were specially made for me--one registering heights to 20,000 feet, the other to 25,000 feet. Although I used these aneroids principally for differential heights along my route, as aneroids cannot always be relied upon for great accuracy, I found on checking these particular instruments with the boiling-point thermometers that they were always extremely accurate. This was, however, exceptional, and it would not do for any one to rely on aneroids alone for the exact measurement of mountain heights. There were in my outfit three artificial horizons--one with mercury, the others constructed with a plate glass. The latter had a special arrangement by which they could be levelled to a nicety. I found that for taking observations for latitude and longitude by the sun the mercury horizon was satisfactory, but when occultations had to be taken at night the plate-glass horizons were easier to work, and gave a more clearly defined reflection of stars and planets in such a bitterly cold climate as Tibet, where astronomical observations were always taken under great difficulty. The most useful instrument I carried on that expedition was a powerful telescope with astronomical eyepiece. Necessarily, I carried a great many compasses, which included prismatic, luminous, floating, and pocket compasses. Maximum and minimum thermometers were taken along to keep a record of the daily temperature, and I also took with me a box of drawing and painting materials, as well as all kinds of instruments for map-making, such as protractors, parallel rules, tape rules, section paper, note-books, etc. I had water-tight half-chronometer watches keeping Greenwich mean time, and three other watches. In order to work out on the spot my observations for latitude and longitude, I had with me such books as _Raper's Navigation_ and the _Nautical Almanac_ for the years 1897 and 1898, in which all the necessary tables for the computations were to be found.