NORMATIVE DENSITIES AND ALTITUDES 397 



nents and ocean floors, as given in Table V beyond. This line cuts 

 through the little group of continental loci, those for Asia and North 

 America falling on it, and passes a little below the locus for the Atlantic 

 ridge (46) and above that for the Pacific (47). The locus for the earth's 

 crust lies a little below the line, being depressed, as it were, by the density 

 of the material of the ocean floors. 



The loci of the Atlantic floor are plotted twice. The locations on the 

 right represent the average depth, while those to the left represent the 

 average depth of the ridges, banks, or continental shelves on which stand 

 the islands from which come the analyzed rocks. The locus of the average 

 Atlantic falls far to the right of the calculated line, while that for the 

 ridges of the Pacific, if plotted, would fall far to the left of it. We are 

 therefore justified in thinking that the altitude which best corresponds to 

 the calculated density of the Pacific floor is that of its average depth, 

 while that for the Atlantic island rocks is the depth of the central ridge, 

 estimated to be about 1,000 fathoms. 



The locus (point A in the figure) of the density of the earth's crust 

 weighted for the areas of the continents and for those of the Pacific and 

 South Atlantic oceans (see page 390), and referred for altitude to the 

 average sphere level of the solid earth's crust, — 2,300 meters, as calcu- 

 lated by Wagner, 18 falls rather above the line, indicating that the density 

 is too low, because not sufficient weight has been given to the areas of very 

 heavy rocks of all the ocean floors. 



In figure 1, as well as in figure 2, the loci and the hyperbolas for the 

 values calculated as rocks with and without water are essentially alike. 

 Those for water-bearing rock fall above those for water-free rock because 

 of the lower densities, and the relative positions of some of the loci are 

 slightly different in the two, but either one would suffice to prove the 

 inverse relation between rock density and altitude. 



In figure 2 are shown the loci for 42 land areas, large and small, this 

 figure being an enlargement and extension of the upper left portion of 

 figure 1, the abscissal scale for altitudes being larger. Here, also, we see, 

 in more detail, the general drift, which shows clearly the inverse relation- 

 ship between the rock density and the average altitude. This is summed 

 up in the hyperbolas, 19 the solid line representing rocks calculated as 

 water-free and the dashed line when calculated as water-bearing. Many 



18 H. Wagner: Beitr. z. Geophysik, vol. ii, 1895, p. — . 

 H. R. Mill (Scot. Geogr. Mag., vol. 6, 1890, p. 185) calculates the "mean sphere 

 level" as 8,400 feet, equal 2,561 meters. 



10 The portions of these included here are so far from the vertices that they are prac- 

 tically straight lines. 



