THE STRENGTH OF THE EARTH'S CRUST . 553 



as a slightly excessive density throughout the whole zone of com- 

 pensation; the Nashville basin as a slightly deficient density 

 through the same depth. The residuals should indicate an out- 

 standing excess and deficiency of mass respectively with the centers 

 at a depth near the middle of the zone of compensation. But 

 masses with centers at this depth and distance would have a very 

 diminished maximum effect upon the residuals of the deflections 

 of the vertical, and one largely modified by the effects of con- 

 tiguous regions. Heterogeneities of density nearer the surface 

 and not related to compensation would tend also to overshadow the 

 error involved in the hypothesis of local compensation. It would 

 appear then that the nature of the deflections is not very sensitive 

 for testing the relative probability of the hypotheses of local versus 

 regional compensation. The assistance of a computing office 

 for trying out several hypotheses would probably bring to light, 

 however, conclusions which would be more determinative. These 

 statements must be regarded, therefore, as forecasts not yet sub- 

 jected to the tests of computation. 



In view of the preceding discussion it would seem that the 

 deflection residuals of Solution H are chiefly of value for measuring 

 the heterogeneities of density not related to topography, nor to 

 the mantle of sedimentary rocks. This is especially true of the 

 Texas-Kansas region studied in detail, for there the region is one 

 of plains with an average elevation of about a thousand feet, and 

 the demands of local isostasy as postulated in Solution H would 

 call for a nearly uniform density under all this region. The out- 

 standing masses represent in large part, therefore, real and local 

 variations from a mean density of the continental crust. 



But if masses of excess or defect of density similar to those num- 

 bered '2 and 5 of Fig. 12 were widely extended, say to a radius of 

 500 rather than 100 km., they would tend much more strongly to 

 make for a local or intracontinental isostatic adjustment. They 

 would become then not outstanding masses but in large part com- 

 pensating masses. The outstanding masses represent the same 

 kind of variations, therefore, which if more broadly extended would 

 be in accord with an isostatic adjustment of topography to a differ- 

 ent level. They suggest that if the zone of compensation of the 

 continental crust be divided into three shells of 40 km. each in 



