ACCURACY OF THE FIELD MAP. 203
are large valleys from the left. These valleys from different directions, join in the sharp bend where the main valley turns to the north and N. E. On both sides there are wild steep rocks of very picturesque forms. Heaps of gravel, blocks and boulders fill the valley. The brook increases and is nearly perfectly free from ice, foaming in continuous rapids in its stony bed. The valley is very narrow, but there is still room for small cultivated fields and a few huts. Here are many well-built menis, some of original form. The valley turns more and more to the right, and finally we reach Tankse.
Here the first part of my journey in Tibet in 1906-1908, came to an end, and the great ring was closed. I had left Tankse on August 2Ist 1906, and now returned there on November 26th, 1907, the journey thus having taken one year, three months and seven days. Here we have, therefore, a means of controlling the exactitude of my field map, disregarding the astronomical observations and all other corrections from determined points, amongst which those of Ryder are the best. Colonel H. Byström has made such a graphic examination by joining all the 26 sheets which were constructed by Lieutenant C. J. O. Kjellström to the scale of i :200 000. Drawing my field maps I only made use of compass and watch and determined the distance by means of the length of the paces of my riding pony. Under such conditions as those in which this journey was made and which have been described above and in Vol. II and III of this work, one would expect a very large error. The continual winter storms, the cold, and the sometimes unfavourable ground, would very likely make the field map uncertain. My several crossings of the Manasarovar were, however, omitted, but the crossings of Lakes Lighten, Yeshil-köl, Pul-/so and Ngangtse-tso, as well as one days' journey on the Tsangpo and one on the Indus had to be included in the graphical calculation of Colonel Byström. Along the route of my fieldbooks, from Tankse and back to Tankse, he got a distance of 4,270 km. On the construction the length-error proved to be 109.5 km. or 2.56°I0 of 4,270, and the cross-error 1.52%i which may indeed be said to be a satisfactory result. Or, in other words, for a distance of 1 oo miles I would get, as an average, 102.56 miles. On the map in 15 sheets in 1:1,000,000, accompanying this work, Colonel Byström has tried to bring my field survey as much as possible in accordance with all other data existing.
IIIFカラー高解像度
