704 WILLIAM HERBERT IIOBBS 



shelves, for mountain summits, and for intermontane valleys. 

 He gives, as the most noteworthy of aberrant districts, Hawaii, 

 Corsica, Sicily and Calabria, the Austrian Alps and the Carpathians, 

 large areas in Turkestan continued eastward to the Pamirs, Lake 

 Baikal, and localities within the valley of the Obi River in Siberia. 

 These, with the exception of the one last mentioned, from which 

 few seismic data are available, are notably the great earthquake 

 zones of the Eastern Hemisphere. 1 



hayford's figures analyzed with the idea of compensation 



eliminated 



Deflections corrected for topography to find attraction of hidden 

 masses. — We have thus seen that in many regions the evidence 

 favors the view that anomaly of gravity is connected with local 

 irregularity and not alone with general and orderly — systematic — 

 distribution of mass over a wide area. Hayford's gratuitous 

 assumption has been that the distribution of mass which produces 

 gravity anomaly (Ag.) is throughout regular and systematic; and 

 to this he has added the additional assumption of a failing earth, 

 which in the light of recent work does not appear to be well founded. 

 Pratt's hypothesis of compensation was designed to meet conditions 

 within an earth shell of supposed high plasticity, and Hayford has 

 contributed to this theory a determination of the supposed depth 

 of the compensation surface. His assumed proof has consisted in 

 testing by the method of least squares a number of hypothetical 

 systematic distributions of matter in the unknown region, with the 

 use of the measurements of deflection of the vertical and of gravity 

 within the area of the United States. His choice of the depth of 

 76 miles for complete compensation was based upon a reduction 

 of the sum of the squares of the residuals to about one-tenth of 

 that obtained upon the hypothesis of a rigid earth without impor- 

 tant local irregularities in distribution of matter. 



If it be true that local anomaly is to be ascribed largely to irregu- 

 lar local distribution of mass beneath and near the station, Hay- 

 ford's method is inapplicable and can only lead to erroneous results. 



1 Cf. de Montessus, "Les tremblements de Terre," Geographic Scismologiquc 

 (Paris, 1906), Map 1. 



