the Rhine below Bingen, the Meuse in the Ardennes, or several of the Himalayan rivers in the gorges that they have cut through the youngest marginal ridges of the range.
Rapidly following the establishment of these two important classes of valleys came the recognition of the very antithesis of antecedent rivers in those streams which have grown by headward erosion along belts of weak structure, without relation to the initial trough lines. To these the term 'subsequent' has been applied. It is frequently in association with streams of this class that drainage areas are rearranged by the migration of divides, and that the upper waters of one river are captured by the headward growth of another. This is accomplished by a most beautiful process of inorganic natural selection, which leads to a survival of the fittest and thus brings about a most intimate adjustment of form to structure, whereby the more resistent rock masses come to constitute the divides, and the less resistent are cbosen for the excavation of valleys. Many workers have contributed to the solution of problems of this class; notably Heim, in his studies of the northern Alps (1876), and Löwl, who showed that, in folded mountain structures of great age, the original courses of streams might be greatly altered through the development of new lines of drainage (1882). A valuable summary of this subject is given by Philippson in his 'Studien über Wasserscheiden' (1886). The extraordinary depredations committed by the waxing Severn on the waning Thames have recently been set forth by Buckman. The turning of side branches from the slender trunk of the Meuse has been recognized in France. Many remarkable instances of stream captures have been found in the Appalachians, where the opportunity for the adjustment of streams to structures has been exceptionally good. Hayes and Campbell have, on the other hand, emphasized the importance of drainage modifications independent of the growth of subsequent streams on weak structures, but governed by a slight tilting of the region, whereby some streams are accelerated and their opponents are retarded. It should be noted that the proof of the adjustment or rearrangement of drainage marks a victory for the uniformitarian school that is even more significant than that gained in the case of the antecedent rivers; for in one case a growing mountain range is subdued by the concentrated discharge of a large drainage area; but in the other case, the mountain slowly melts away under the attacks of the weather alone on the headwater slopes of the growing valleys.
The reason why all these studies of land carving are of importance to the geographer is that they greatly enlarge the number of type forms that he may use in descriptions, and that they recognize the natural correlations among various forms which must otherwise be set forth in successive itemized statements. The brief terminology learned in