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

as the cause which gives rise to the Raga-tsangpo. He observes that the Ki-chu,
Nyang-chu, Rong-chu and Shang flow against the Tsangpo itself.

Of the watershed Burrard says:¹ ›The water-parting between the Indian and
Tibet basins cannot be drawn with certainty: in places it is without doubt the Kailas
range, but the latter has been cut through from the north by feeders of the Brahma-
putra whose basins have not been determined. The Lhasa river, the Charta and
others drain through the north of the Kailas range, and pierce the Kailas range in
the same way as the Himalayan rivers pierce the Himalayan ranges.‹ The Chaktak-
tsangpo, however, pierces two ranges and rises on a third, which in this part is the
watershed between India and the plateau-land. The whole orographical arrangement can
hardly be compared with the Himalaya, as in the case of Chaktak-tsangpo the water-
parting is also on one of the most considerable ranges in the country north of the Tsangpo.

On his Chart XXX (Pl. XXIII) Burrard gives a good idea of the ›Himalayan
area drained by the Brahmaputra‹. Only between 84¹/₂° and 89° East long., is the
northern watershed incorrect. It is given from the materials brought home by Ryder.
But even from high stations it would be quite impossible to judge from how far
the northern tributaries come, as is clearly seen on the map. Chaktak-tsangpo, for
instance, is represented as rising too far south. The drainage area of Amchok-tso
is represented as making an apophysis to the north and as if the ultimate feeders
of the lake should reach further north than its neighbours. But in reality the area
of Amchok-tso, marked by Angden-la, is the southernmost point on the whole
watershed between 84¹/₂° and 89°. In some cases the watershed has been drawn
²/₃° too far south. On the whole section it is represented as situated south of the
30° lat., whereas it in reality crosses this parallel only at one place, Angden-la.

In the course of time this watershed has wandered up and down like a pen-
dulum. As nothing was known every explorer or geographer could accept, ad
libitum, whatever limit he personally thought to be the most likely one. In the
preceding chapters it has thus been seen how the watershed has sometimes been too
far north, sometimes too far south. It was my good luck to settle this feature, one
of the most important in Asia, although, from reasons easy to understand, my survey
could only be a reconnoitring, and the mathematically precise survey was to be left
to the future. To take only one example we find the northernmost feeder of the
Chaktak-tsangpo on the Ta-ch'ing map at 31¹/₂° North lat., on DUTREUIL DE RHINS'
Première transformation de la Carte Chinoise at 30° 55', on d'Anville's trans-
formation at 32° 5', on Saunders' map at 30° 27', on Ryder's map at 29° 42';
in reality it is situated on Sangmo-bertik-la, at 30° 15'. As I crossed the watershed
between 84¹/₂° and 89° only at five points, I could only sketch it roughly, and
between these five points there is, of course a great deal of work left to be done.