A recent article1 on the drainage conditions in the region where the Salween, Mekong, and Yangtze are compressed within a space of 50 miles in their flow southward from the Tibetan Plateau is a bold, not to say a venturesome application of deductive geomorphology in an effort to give explanatory description of the river courses in a lofty mountain region. The trend of the main ranges and valleys, as well as the strike of the nearly vertical rock structures, on the Yiinnan-Tibet frontier is about north and south; but certain first-order tributary valleys, trending in part east or west, also turn to southward courses for part of their length; and, as it is assumed that such tributary valleys must originally have had wholly east or west transverse courses, an explanation for their southward turns is sought. The explanation proposed points out that the heavy rainfall comes with southwest winds and that the climatic control of vegetation on the higher slopes favors erosion on their windward aspects; it is therefore concluded that second-order tributaries, heading northward, would capture and divert the upstream parts of first-order tributaries farther north, and thus southward turns in the tributary drainage would be developed.