which travellers' routes do run relatively close together. Yet had we possessed no other material than the meridional routes, this uncertainty would have been greater still than it actually is. But fortunately for the northern half of the region in question we have the advantage of two latitudinal routes; in its middle portion also two, though it is true short ones, running from north-west to south-east; while in the south we are fortunate enough to have quite a »faggot» of latitudinal routes. Confining our attention in the meantime to the first mentioned, namely my route of 1896 and Wellby's of the same year, we have in them a fixed and incontrovertible basis for determining the direction in which the mountain-ranges run. These two routes in the part of Tibet with which we are at present dealing are situated between 35° and 36° N. lat., and follow throughout the whole of their extent vast latitudinal
valleys lying between equally vast main ranges, and the said ranges run as nearly as possible from west to east. The range, or rather system of ranges, north of my route is the Arka-tagh, and the system of ranges between my route and Wellby's is the Koko-schili, or at all events the westward continuation of the range which bears that Mongol name. The position and direction of these two mountain-systems being fixed, we might therefore reasonably consider it probable that the ranges which lie next to them on the south run on the whole parallel to them, or at the most deviate but slightly from the direction which they pursue. Nevertheless this does not appear to be altogether the case, for my route of 1900 does not cross any pass worth speaking about between Camp XLVIII and Camp LVI, any more than my 1901 route does between Camps XXXIII and XXXVIII; instead of that we travelled along the latitudinal valleys between the mountain-ranges, which in the region defined appear to have rather a north-west and south-east direction. In other words the mountain-ranges in this region seem to diverge towards the east, and they are moreover less distinct and less continuous than in the country to the north of them. Taking now another step farther south, we come to a vast, compact range, a true dominating main range of the first magnitude, namely that
IIIF Color HighRes
