274 THE LOP-NOR PROBLEM.
likely that it alters a little every year. So far as one can see however, it did occupy the same position in 1876, and the rivers in the vicinity of Ajrilghan presented then on the whole the same appearance that they had in the year 1900. At all events in the years 1760-65, the Kara-koschun occupied the basin it now occupies, just as Kosloff says, basing his statement upon information supplied to him by the learned General Stubendorff, than whom no man is more skilled in Central Asian cartography.
Kosloff's statement, that two hundred years ago the Tarim flowed through the Schirge-tschapghan arm is taken from Pjevtsoff,* who writes thus : »According * Trail etc., p. 305-6.
to tradition the Jarkent-darja flowed 200 years ago farther to the north than its present lowermost course, and discharged into the little lake of Utschu-kul, which was connected with the Lop-nor by a sound. This tradition was confirmed by old Abdul Kerim, who said that in his grandfather's lifetime the river still flowed through the place named, but subsequently it changed its bed. The old bed of the Jarkentdarja, now called Schirge-tschapghan, can be distinctly traced even at the present time. Beside it there still remain in places stumps of trees, which once shaded its banks. Formerly these tree-stumps were very numerous, but the inhabitants of the neighbouring villages having consumed them as fuel, they are now far fewer in number.»
I have already shown, that during the last few years a considerable portion of the Tarim returned to the bed of the Schirge-tschapghan (Tokus-tarim), and that old tree-trunks which once accompanied that river extend into those parts of the Desert of Lop where I crossed it. Pjevtsoff's statements are always characterised by the greatest accuracy and conscientiousness, and he would not have written down the information of Abdul Kerim, unless he was convinced of its reliability. Moreover I am able to confirm it with maps and my own observations made on the spot. It is interesting to see how well Pjevtsoff's conclusions, based on questions he addressed to the natives, agree with those which I arrived at in 1896, and which are incorporated in the passage quoted above from the Zeitschrift. Pjevtsoff says, that his informant Abdul Kerim was i ro years old in 1890. Consequently he was born in 1780. Supposing his father to have been 4o when he was born, his birth would fall in the year 1740; and if his father was 4o when he was born, the grandfather's birth would take place in 1700. Now it was from his grandfather that Abdul Kerim derived his information. This ancestor seems also to have lived to an exceptionally great age; perhaps he was not born until 1720, in which case the two figures 40 must be altered to 30. The change of river-bed must have taken place in 1740 at the latest. Pjevtsoff says indeed, that it was in Abdul Kerim's »grandfather's lifetime», and since Abdul Kerim, who was born in 1780, derived his information from his grandfather, we are brought to the interval between 1700 (or 1720) and 179095 , when Abdul Kerim was old enough to discuss the matter with the old man. But if that were so, Abdul Kerim might very well have told Pjevtsoff that he himself, or at any rate his father, was alive when the river changed its bed. But as an actual fact it was only the grandfather who was alive when that event took place; consequently it must have occurred about the year 1740. My information was derived
IIIF Color HighRes
