14 REPORT OF THE CALIFORNIA EARTHQUAKE COMMISSION, 
dies down inversely as the square of the distance; the energy at P is, therefore, 
proportional to 
D +O a 
i} def Ba z= 2 tan atk 
0 IF o@+yt ey 
where D is the depth of the fault. The limits of the integral along the fault assume 
an infinite length for the fault; but the result is practically the same if the angle 
between two lines drawn from the ends of the fault to P is nearly 180°; that is, if the 


D 
distance of P from the fault is small compared with its length. The values of Ntan-1e 
have been plotted in fig. 4 in terms of y It will be noticed it is a function of — and is 
independent of the actual intensity of the disturbance at the fault-plane; but unlike 
Major Dutton’s energy curves, there is no point of inflection in our curve (this is due 







to the fact that the fault-plane is not confined to a considerable depth, but extends to 
the surface of the earth). We must therefore, to determine the depth of the fault, deter- 
mine the distance of the point where the force bears some definite proportion to the 
force in the immediate neighborhood of the fault-plane, for instance, where it is half as 
great. We find from the curve that the distance of this point is about 2.5 times the 
depth of the fault. We must, now, from our map of intensities determine the actual 
distance from the fault-plane of the points where the force had diminished to half its 
value at the fault-line. Professor Omori! has estimated that the acceleration on the 
made ground in San Francisco was somewhat less than 2,500 millimeters per second per 
second, and therefore the acceleration on rock at the fault-plane may be taken at about 
this value. According to Professor Omori’s scale, a force about half as great, or 1,200 
millimeters per second per second, corresponds to the degree VIII of the Rossi-Forel 
Scale, and this isoseismal occurs at a distance of about 20 km: from the fault-line opposite 
Point Arena; 20 km. is therefore 2.5 times the depth of the fault, which accordingly 
becomes 8 km. We must not attach too much importance to this result; the assump- 
tions and the data are all too inaccurate, but we can accept it as indicating that in the 
neighborhood of Point Arena the fault could hardly have extended to a greater depth 
than 20 km. and probably was not so deep. The distribution of isoseismals north of 


* Bull. Imperial Earthquake Investigation Commission, vol. 1, No. 1, p. 19. 
