ON SElSMOLOGICAL INVESTIGATION. 79 



The methods and considerations which have led to these determina- 

 tions have been as follows : 



(1). Determination of Oi'igins by Comparisons bettoeen Time Intervals, 



Earthquakes from the same district will arrive at distant observing 

 stations at times the differences between Avhich will be constant. If, for 

 example, we have once determined the difference in time at which an 

 earthquake originating off the coast of Japan arrives at Batavia, Bombay, 

 Cape of Good Hope, Shide, &,c., whenever these differences are repeated 

 at four or more stations, without knowing anything about observations 

 in Japan, we can at once say where such an earthquake has originated. 

 It will be noted that our knowledge respecting the speed with which earth- 

 quake motion is transmitted enables us to give approximate values for the 

 time differences here considered. 



(2). Bij the Difference in the Times at ivhich the Maximum Motion 

 has been recorded at differerd Stations. 



In the present state of our knowledge all determinations of the position 

 of oi-igins from time intervals require the assumption that the velocity of 

 propagation of earthquake movement is constant. This condition is most 

 neai-ly fulfilled by the large waves of earthquakes. The methods by 

 which an earthquake origin may be determined from the differences 

 between the times at which it was recorded at distant stations are several. 

 The method of circles which is here employed has been selected chiefly on 

 account of its comparative simplicity in application. It is briefly as 

 follows : If the large waves of an earthquake reacli stations B, C, D, &c., 

 four, ten, twenty, cfec, minutes after reaching station A, then the centre 

 of a circle which passes through A and touches circles drawn round B, C, 

 D, &c., the radii of which are respectively 4 x 1°'6, 10 x 1°"6, 20 x 1°'6, kc, 

 will be the centre of the origin required. The constant l°-6 means 

 that the arcual velocity for large waves is taken at 1°"6 per minute, or 

 approximately 3 km. per second. In the British Association Report for 

 1899, p. 193, the speed there given was 2*5 km. per second, which appears 

 to be too low. The operation of drawing these circles is carried out 

 on a ' slate ' globe. For a complete solution observations are required 

 from at least four stations. With only three observations we are left to 

 choose between two possible centres, but as these may be widely separated 

 there is usually but little difficulty in selecting the one required. 



(3). By the Time Intervals between the Arrival of Preliminary Tremors 

 and Maximum Movement. 



From what has been said respecting preliminary tremors and large 

 waves it may be inferred that the interval in time between the appearance 

 of these two phases of earthquake motion at a given station has a relation 

 to the distance of that station from the origin. This relationship is shown 

 in fig. 1. An observer with this curve before him, although his time- 

 keeper may have failed, or although he may be so situated that it is 

 impossible to obtain accurate time, is immediately able to determine from 

 a well-defined seismogram the distance at which the motion it represents 

 originated. With this fact, the magnitude of his record, and a knowledge 

 of the physical configuration of districts from which earthquakes originate, 

 he is frequently able to locate an origin. With time records from several 



