1921.] Angenheister.—A Study of Pacific Earthquakes. 223 
VIII. The Upper Crust of the Earth according to Gravity- 
observations and Seismic behaviour. 
The results of gravity-observations suggest that at great depths the 
interior of the earth is built up according to the laws of hydrostatics. 
Differences from these laws are to be expected only in the upper crust. 
Thus at a distance of 120 km. below the surface according to Helmert, 
44-60 km. according to others, we have a surface in perfect hydrostatic 
equilibrium, so that in this thickness, which we may call the earth’s crust, 
the total effect of mass and stress over any area of, say, a few hundred 
square kilometres is the same. A good proof of this may be seen in the fact 
that gravity over great ocean-depths is the same as over low-lying lands of 
continents. The greater volume of continental blocks is compensated by 
the greater density under the ocean, where again the surplus of mass is 
compensated by the smaller density of water. According to Pratt the 
lower density of continents and mountains originates by certain masses 
expanding in the solidification of the earth’s crust, while other masses now 
in the bed of the ocean contracted and sank. H. Faye believes that on 
account of the overlying water the cooling under ocean takes place quicker 
than under the inferior conducting air over continents. In this way the 
earth’s crust under the ocean is thicker than under continents. We 
have seen that the seismic observations agree to a certain extent with 
this hypotheses. 
Firstly, disturbances of hydrostatic equilibrium, and thus stresses in 
the crust, are to be expected only to the lower boundary of the crust. More¬ 
over, stresses producing earthquakes are to be expected when neighbouring 
positive and negative gravity-disturbances suggest disturbance of a hydro¬ 
static equilibrium and resulting stresses in the crust. In the Pacific the most 
pronounced disturbance of this kind is near the Tonga Deep. Gravity above 
the Tonga Deep is 0*25 cm./sec. 2 too small, so that we must have a large 
superficial defect of mass. The westerly boundary of the Tonga Plateau, 
on the other hand, has a gravity surplus of 0-17 cm./sec. 2 , and thus a 
considerable surplus of surface mass. The elastic stress in the crust must 
therefore be directed from west to east— i.e., from the plateau to the deep. 
It will be remembered that in agreement with this the epicentres of Tongan 
earthquakes were all found on the west side of the deep. 
Secondly, below the boundary surface of the crust we may expect no 
difference of density, pressure, or elastic behaviour : i.e., these can only be 
expected in the crust itself. Consequently, only that part of the path of 
seismic waves which is on the crust can show the influence of different 
elasticity, &c. For great epicentral distances, therefore, we can expect 
only small differences of travelling-time of P and S with difference of azimuth 
under ocean or continent, as in this case only a small part of the seismic 
ray travels through the crust above the boundary surface. For small 
epicentral distances of P and S and for the surface waves of the main phase 
we may expect differences. We have seen that this is the case. 
Thirdly, denudation lightens the continent and increases the weight on 
the sea-floor. This in addition to the quicker cooling of the earth’s crust 
under ocean should give a thicker crust there than under continent. We 
have seen that the observations of surface waves give ample support to 
this hypothesis. 
Summary. 
1. For a strong Tongan earthquake of strength VIII the true tilt did 
not exceed 2 seconds of arc 6 minutes after the end of violent shaking. 
