304 HAYFORD— THE EARTH FROM [April 24. 



found at the surface only as the extreme pressure and temperature 

 conditions at great depths directly and indirectly produce differences. 



It should be kept clearly in mind that the geodetic evidence from 

 observations of the direction and intensity of gravity indicates 

 simply the present location of attracting masses, the present distri- 

 bution of density. It furnishes no direct evidence whatever as to 

 past distributions of density, or as to changes in density now in 

 progress. But an understanding of the present distribution of 

 density within the earth, especially near the surface, is so necessary 

 to a true understanding of the present state of stress and of viscous 

 flow in the earth that an understanding of the geodetic evidence is 

 fundamental to progress. 



Computations should be made in extension of those which have 

 been made by Darwin and Love. The new computations should, 

 however, deal with the actual irregular continents and mountains, 

 not with regular substitutes. The computations should also take 

 into account the bulk modulus of the materials composing the earth, 

 that is these materials should be assumed to be compressible. Such 

 computations will no doubt be both difficult and long. I believe 

 that even a moderately vigorous attack along this line will show con- 

 clusively that the earth does not behave as an elastic body under 

 the large loads superimposed upon it by the continents and moun- 

 tains. I believe that the computed stress-differences will be found 

 to be so large that the computation will be essentially a proof of 

 viscous yielding. 



Next make the contrasting assumption that the material compos- 

 ing the earth is competent to withstand but little shearing stress, 

 and that the pressure at any point is that due to gravitation acting 

 on the mass in the column extending from the point vertically to the 

 surface. Let it be assumed that isostatic compensation exists, is 

 uniformly distributed with respect to depth, and is complete at 

 depth 122 kilometers. Consider the actual topography and form a 

 mental picture as accurately as possible of the viscous flows which 

 would take place on the assumption that at each level the material 

 would flow horizontally from regions of greater pressure to regions 

 of less pressure along lines of maximum rate of change of pressure, 

 and that the time rate of such viscous flows would tend to be pro- 



