58 



As the maximum was recorded in Manila and in Tokyo at the same time the origin should be found on a 

 line all points in which are equidistant from these two stations. For similar reasons it should be on a lino 

 equidistant from Batavia and a point about 480 kms. from Irkutsk, on a line drawn from that place in 

 the direction of Guam. These two lines intersect in the region indicated. Guam and Say pan were, there- 

 fore, about 300 miles distant from the origin, and that this might well be the case is testified by the 

 destruction which took place. Had these islands been much nearer to the origin of a disturbance which 

 was so definitely recorded at stations all over the globe, it is likely that the destruction in Guam would have 

 been greatly intensified. 



The time of origin as deduced from the time of arrival of the maximum motion at Tokyo and Manila is 

 1.44, at Batavia 1.43, and at Irkutsk 1.40. The local times given for Guam are 1.35, 1.45 and 2.5. As a 

 close approximation to time of origin we shall adopt 1.44. 



In the following table velocities are expressed in degrees per minute. The eighth column indicates the 

 proportion of the wave path between the origin and the various stations which was suboceanic, while the 

 last column gives the amplitudes recorded at these stations : 



An inspection of the above table shows that the values for PI are such as might be expected, and also 

 that the velocity of propagation on an arcual path is not constant. The average velocity along a path 

 corresponding to a chord when the length exceeds 50 is more nearly constant. PS shows a fairly constant 

 arcual velocity to stations at less distances than 60 from the origin. To stations 60 to 90 distant the 

 rate of propagation increases to values that suggest a rate of transmission for P 2 . Beyond this distance 

 the rate decreases to values approximating to those obtained at the nearer stations. 



In the middle regions it may be noted that the increased rate is derived not only from records of Milne 

 pendulums, but from others adjusted to have different periods. It is likely, therefore, that the maxima 

 recorded by different types of instruments refer to the same phase of motion, and we are not dealing with 

 apparent maxima occasioned by coincidences of the period of the pendulum and that of its foundations. 



There does not appear to be any relationship between PS and the nature of the path, nor is anything 

 clearly shown with regard to decreasing amplitudes. (See Time Curve No. 49, Plate 1.) 



