As plateaus usually display this edge on both sides, the border has been aptly compared to a double ledge or rim, between the two sides of which the table-land lies, often tolerably high above the sea level. If these rims, like mountains, are not contiguous to the plateau; if they are separated from it by a valley of greater or less width and depth, running parallel with the edge, they form what Humboldt has called natural circumvallations. Of such the Altai range, on the north side of the Asiatic central plateau, is an example. The hollow between the range and the plateau just mentioned is partly filled with inland seas. The Caucasus may, in like manner, be regarded as the circumvallation of the American plateau, separate from it by the Koor and the Aras (ancient Araxes) rims. Yet in the Caucasus another modification occurs—a partial linking of the plateau with the range at the west extremity, by the connecting chain of the Moschic Mountains. In like manner the Pyrenees, in their eastern half, form a circumvallation around the north side of the Castilian plateau, separated from it by the basin of the Ebro, and forming a perfect ring around Upper Castile and the elevated province of Biscay. In cases where a mountain chain rests upon a plateau, rising up in the very heart of it, its summits seem to be not high, although the basis, the true foot of the chain, may not be at the level of the plateau, but far lower, and such mountains may, therefore, be of great absolute height. The name superimposed mountains has been given to them. Such are the Kuenlun and the Thian-Shan ranges of Central Asia, the Guadarrama chain between Old and New Castile, and the Rocky Mountains in North America. These superimposed ranges often run near to and parallel with the rim or edge of the plateau, and seem to give it more completeness and breadth. The geologist employs the word “sutures” to designate such forms, because they serve to unite those parts of a plateau which are at different heights above the sea level. He regards the mountains as rising to fill enormous clefts which great convulsions have rent in the earth, and as passing up, while in their fluid state, to a height above the level of the plateau, and bridging over the abyss. In this way our mountains which rest on plateaus seem to have been formed, as indeed is indicated by their geological structure. The smaller plateaus display analogies kindred to those seen in the larger superimposed mountain ranges. The extinguished volcanic group of Auvergne rests upon the central plateau of southern France, which, according to Remond, has an average elevation of 1000 feet. The now silent volcanic group of the northern Rhine broke through the moderately elevated gray-wacke formation of that locality, and is, therefore, a superimposed range.