1 86 G. F. BECKER ISOSTASY AND RADIOACTIVITY 



tiuu as well as the horizontal component. But Helmert's method is ap- 

 plicable at a comparatively small number of stations. Baron Roland 

 Eotvos's torsion balance** seems to afford a means of fixing the position 

 of locally attracting masses with some accuracy, since by it the radius of 

 curvature of the geoid is determinable; but this method is laborious in 

 the extreme, and many years must elapse before thorough surveys with 

 the torsion balance can be completed, even for a moderate number of 

 small areas. 



Meantime the only recourse is to general reasoning, trial hypotheses, 

 and experiments on the properties of matter. To me it seems clear that 

 gravity anomalies must be of two classes, which I shall take the liberty 

 of calling real anomalies and pseudo-anomalies, the latter being due to 

 irregular distributions of density and not affecting the real load per unit 

 area at the level of isostatic compensation, while the real anomalies rep- 

 resent real differences in load at that level. 



Areas of denudation and of deposition would seem to represent, at least 

 in part, what I have called real anomalies, for the removal of matter from 

 the outer surface of a column must reduce the pressure at its base, and 

 vice versa. But Messrs. Hayford and Bowie find it impossible to trace 

 any relation between the distribution of gravity anomalies and erosion or 

 sedimentation. This suggests that the effects of this actual transfer of 

 matter are masked by the effects of irregular distributions of density or 

 that the real anomalies are relatively small. 



As for the pseudo-anomalies, we know for a certainty that the distribu- 

 tion of densities in horizontal directions at the earth's surface is very 

 variable, while the exposures in deep wells or in deep cuts, such as the- 

 Grand Canyon of the Colorado, give no evidence that heterogeneity di- 

 minishes with depth. On the contrary, the existence and abundance of 

 dikes, sills, laccoliths, and batholiths make it highly improbable that 

 homogeneity, or even a gradual and regular increase of density, prevails 

 at any level above the deeper volcanic foci. Differences in density as- 

 cribable to mineral composition are not the only cause of these pseudo- 

 anomalies. It is well kno'WTi that the thermometric gradient varies 

 greatly with the locality and in a seemingly capricious manner. Of 

 course, this distribution of temperature also affects density. Voids, too, 

 give rise to differences in density, and these, whether as geodes or as 

 joints, may be sparsely scattered or closely grouped. With the assistance 

 of Mr. A. F. Melcher, I have recently shown that the volume of a column 

 or stratum of rock may incroaso tlirougli cvushing by an nnimint the 



** See Helmert's memoir on gravity and the mass distribution of tlie earth. Encyc. 

 der Math. Wiss,, vols. 6, 1, 7, article 23. 



