282 THE LOP-NOR PROBLEM.
called hills. On the other hand the mountain-range to which the Nukitu-daban belongs is not even represented on the map. But the height of absurdity is reached when mountains are put in the Ta Gobi, in that part of it which I crossed on my way from the Kerija-darja to the Tarim; for there does not exist there the slightest trace of a mountain any more than there does in the desert immediately north-east of Kerija. The fact of the signs for mountains being shown in these places must therefore be due either to the private belief of the cartographer, or to erroneous information given to him, or — and this is probably the most likely explanation — he intended by those signs to indicate the lofty sand-dunes which do exist there. Consequently from the occurrence of the conventional Chinese cartographical signs it is impossible, as von Richthofen points out, to conclude whether they mean mountain-chains capped with perpetual snow or whether they are merely low hills. The Chinaman in fact has so little appreciation of the relief of a country and of the plastic features of the earth that he does not make the slightest difference between a hill 5o m. high and a mountain 5000 m. high, taking both altitudes relatively. On the other hand, for horizontal topography he has a keen eye and a wonderfully developed power of apprehension. At the same time it is obvious, that a map may err in point of topography. In certain parts of the Wu-tschang map the errors are practically insignificant, because the materials from those regions are much more abundant than from the parts with regard to which the Chinese possessed only imperfect data. For instance, Khas-nur is placed too far to the north, and i 1/12° too far to the west, the consequence being that the southern road from Sa-tscheo is pushed too far to the north. As for the road from Khas-nur to Lop-nor, it is not possible to draw any conclusion whatever from the account given in the Si yii-sclzuei-tao-ki, for that carries us continuously west from the pass of Nukitu-daban, whereas Lop-nor lies north of that position. In fact the map gives a much nearer approximation to the truth than does the description in the accompanying text; for according to the former Khas-nor lies at any rate south-south-east from Lop-nor, although the distance is much too short, but the text puts it east of Lop-nor. • There can be little doubt that information dating from different periods, and derived from different authorities, has been collected and pieced together. The impression left upon the mind is that as far as Nukitu-daban the writer is describing the southern road, but from that point onwards he is describing the northern road. For he says, »after leaving Nukitu-daban you keep to the west and then come to Nukitu-setsin, and still going west you come next to Nukitu-schan-k'ou, and on beyond that until you touch successively the southern shore of Lop-nor, the Ike-ghaschon, and, still farther on, the Baghan-ghaschon, which is the southern bank of the Tarim river». Now this does not in the least agree with the map, for according to the latter Nukitu-sekin lies north of Nukitu-daban, and Ikeghaschon east (not west) of Nukitu-sekin. So long as we do not know the meaning of the word nukitu, which occurs, as we see, in four compound names distributed over a pretty wide area, and both amongst the mountains and on the plain, we are but fumbling in the dark in attempting to fix the position of the places indicated by these names. Himly translates the name Nukitu-qia-Tarim, which is used to mark the locality where the two roads divide south of Lop-nor, as meaning »abschüssiges Ackerland des Nuki» — tarim signifying »arable land», and qia »steep»,
IIIF Color HighRes
