46o JOSEPH BARRELL 



the mass is unsymmetrical about a vertical axis, observations must 

 be made along at least two lines at right angles to each other. 



A minimum number of observations will define an isolated 

 outstanding mass, but if several masses have their fields of force 

 notably overlapping, a larger number of observations becomes 

 necessary in order to differentiate their effects. An inspection of 

 Figs. 9 and lo shows that the shape of the curve of Fv between the 

 epicenter and a distance where it falls to one-tenth the maximum 

 value has more distinctive relation to the form of the immediately 

 adjacent mass than has the curve for Fh. But the available 

 geodetic data supply less information regarding the gradients of 

 the gravity anomalies than for the deflection residuals. The latter 

 are given along a certain belt of triangulation stations, whereas 

 the gravity stations are located at long distances apart. Further- 

 more, but few of the gravity stations coincide with deflection 

 stations. The present analysis will therefore rest upon the data 

 giving the curve for Fh. This curve is flat at the top, so that the 

 data will readily give the approximate value of the maximum 

 but will not determine closely its distance from the epicenter. The 

 value of d will, however, be determined ordinarily on two sides of 

 the epicenter by means of the deflection residuals and the mean 

 will give a more reliable figure than either alone. But according 

 to the form of the mass within those Hmits shown in Figs. 9 and 

 10 the value of d may range from 38° to 67^°. If the abnormal mass 

 is assumed to have a spherical form, its center will lie at an angle 

 of 55° below the maximum value of Fh and at a depth 1.4 the 

 distance to the epicenter. The error in locating the points of epi- 

 center and maximum Fh may cause the estimate of depth to be in 

 error 20 per cent and yet this figure will show definitely whether 

 the sphere Hes within the zone of compensation or in the centro- 

 sphere. If the mass, however, is in reahty a horizontally elongate 

 mass, the change in the distance EM from the epicenter to the point 

 of maximum Fh in two directions at right angles to the epicenter 

 will show that fact. A check on the form of the mass may be 

 obtained if the value of the deflection curve is known with fair 

 accuracy to a distance from the epicenter of three times the distance 

 of the maximum. Let the distance to the point of maximum value 



