286 THE LOP-NOR PROBLEM.
regarding the windings. In reality that route is 60o km. long. By adding up the separate sections of the southern route, we get 168o li, which, observing the same proportion, ought to be i 260 km., a distance which, if reckoned from Sa-tscheo, would reach all the way to Kerija. But in reality the difference is not so great, for the southern road via Ghas (Khas-nur), which starts at Sa-tscheo and ends at the mouth of the Tarim, measures onlÿ 700 km.
Still this method of calculation is misleading, for the distances between the principal stations are all given. .Beginning at Bayan-bulak, which seems to be situated in the vicinity of Sa-tscheo, we have the following stages: 200 + 150 + 730 + 300 = 1380 li to Ulan-tologhai, situated east of Khas-nur. From Ulan-tologhai westwards to the eastern shore of the Khas-nur, thence north to Khu-pu, and then on, after a westward bend to Nukitu-setsin, still north to the southern foot of the Nukitu-daban, and thence still farther west to Nukitu-schan-k'ou — gives altogether a distance of at least 300 li. These last-named points seem therefore to have been situated so close together that the traveller, who furnished the data of the itinerary and measured the distances, would appear to have preferred to take the entire route in the lump, instead of specifying the several short intervening sections.
From Sa-tscheo to Ulan-tologhai it is i 380 li or 400 km., and of this fully 300 li lie between Tsaghan-tschilaotu and Ulan-tologhai, corresponding to 87 km. The distance between Ulan-tologhai via Ghas to Nukitu-schan-k'ou is precisely the same; but if we measure 87 km. on the map, or say in round numbers 90 km. north from Ulan-tologhai via Ghas, we do not get farther than to the latitudinal valley between the Ak-ato-tagh and the Astin-tagh. If however we measure on the Wu-tschang map north from Ulan-tologhai a line equal to the distance between Tsaghan-tschilaotu and the place just named, we are brought to Nukitu-sekin, which, although it is counted to belong to the southern route, is placed on the map to the north of the northern route — a proof that the western parts of these two routes have been confounded together.
Kosloff is no doubt right in suspecting that the Nukitu-davan is identical with the Kurghan-davan, north of which pass we find also a spring (N.-sekin?) and a valley or »mountain mouth» (N.-schan-k'ou ?) opening to the north. Nor is the contrary proved by the fact that the actual distance between Ulan-tologhai and this valley-end amounts to 225 km., instead of the 87 km. calculated above. Evidently Nukitudavan, Nukitu-sekin, and Nukitu-schan-k'ou lie as near together as do Kurghan-davan, Kurghan-bulak, Kurghan-saj, and a possible Kurghan-sajning-aghis, or the Entrance of the Fortress Valley. On the other hand, Nukitu-qia-tarim seems to belong, as I have pointed out above, to the lowlands.
The topographical section in St. Petersburg appears to place greater reliance upon the Chinese map than it deserves, at least in respect of topographical matters. They have also marked on their great map of Asia the third and most northerly of the three routes we are discussing, as well as the roads that proceed north from Sa-tscheo, the lake Toli, and various other features; and from their map both routes and the lake just mentioned have been adopted by Sliders Hand-atlas. The interpretation however which the Russian map puts upon the northern route is certainly incorrect. On the Wu-tschang map the road in question runs north of Bulungir-
IIIFカラー高解像度
