国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

from 1904 to 1907. On the block pillars between which the wooden bridge is laid,
there was, in 1907, a very distinct water-mark 47 centimeters above the surface of
the spring-water pools. A Lama who had passed 6 years in the *gompa* asserted
that in 1903 the effluent had reached so far, a piece of information which was ob-
viously wrong. The bridge is old and the water-mark may date from 1846 and 1848
when the Stracheys found a strong current in the channel.

On August 24th and 25th, 1907, I surveyed the channel between the Mana-
sarovar and Rakas-tal, using exactly the same method as in 1901 through the desert
of Lop, although the distance there was more than eight times as long.¹ As the
difference of the altitude between the surfaces of the two lakes will change from
year to year and probably from one season to another — especially when there is no
outflow from the Manasarovar, I think it would be superfluous to record here all the
readings backwards and forwards from theodolite to staff. It will be sufficient to give
merely the result of the survey. I used Hildebrand's »Traveller's Theodolite» and
had a 2 m. staff, divided into meters, decimeters and centimeters, black and red.
As the ground, both in the bed of the channel and at its sides, is rather soft, con-
sisting of sand and fine gravel with some grass vegetation, the staff was put each
time on an iron plate to prevent it from sinking into the ground when turned
round 180°.

The distance between staff and theodolite was as a rule 50 m., measured with
a tape, sometimes more or less, depending on the configuration of the ground. There
are 104 stations in all and 208 readings. The result of all these readings showed that
the Rakas-tal then was 13.45 m., or 44.119 feet below the surface of the Manasarovar.

The measured line runs chiefly in the very bed of the channel, only sometimes
at its sides. The line, as measured with the tape, is 9.366 m. long. But this line
is not straight, especially at its beginning and end it meanders like the channel itself.
The neck between the lakes will here be about 9 k.m. broad or a little more than
5¹/₂ miles, so Ryder has made it too narrow, estimating its breadth at 3 miles. But
Ryder is very near the truth when he estimates the difference in height at 50 feet.
In 1909, 1910 and 1911 when the Manasarovar was about two meters higher
than in 1907, Ryder's 50 feet were probably perfectly correct.

In the channel, station N:o 4, just at the foot of Chiu-gompa, was the highest
level above the surface of the Manasarovar, namely 2.263 m. Thus the lake had to rise
2.263 m. before any water could flow over into the Rakas-tal. Ryder says, in 1904,
that »a rise of about 2 feet in the level of the lake would cause water to flow down
the channel», which proves that in 1904 the lake was considerably higher than in
1907. Remembering that in 1909 to 1911, an effluent really went out we get a
direct value of the considerable oscillations of the level even within a very few years.
Only between stations N:o 23 and 24 did I cross a point which was at the same level