ON THE EARTUQUAKE PHENOMENA OF JAPAN. 207 



along the eastern shores of Japan, it is seldom that they cross the moun- 

 tains to the north-western coast, which is singiilarlj free from earthquakes 



One inference to be drawn from the above observations is that the 

 disturbance, as felt upon the land, is to a great extent superficial, and on 

 reaching the mountains is either destroyed by reflections and refractions 

 or else absorbed by their mass. ' 



In the case of earthquakes where there were prominent vibrations 

 having a definite direction, a prolongation of lines parallel to these 

 directions through the observation stations has given intersections 

 corresponding with the locality in which we should seek for the origin of 

 the disturbance from the records of the post-cards. When prominent 

 movements in two or more directions had been recorded, the one which 

 had to be taken to represent the normal motion was indicated by the 

 approximate origin shown by the post-cards. 



These determinations were further checked by time observations 

 usually made in Yokohama, Tokio, and Chiba. Sometimes they were 

 also made m Saporo and Hakodate. Assuming these time observations 

 to have been correct, and the velocity of an earthquake to be constant it 

 is theoretically possible by several methods to determine the orio-in of a 

 disturbance. Although these methods fail in practice, chiefly o'Win? to 

 the fact, which I will speak of presently, that the assumed velocity is not 

 constant, the observations lead to very sure and practical deductions as 

 to the direction m which we are to look for the origin of a shock. 



II. Velocity of Propagation of an Earthquake Wave. 



The observations which l" have been enabled to make upon the 

 velocity with which earthquake motion is propagated are dependent upon 

 the accuracy of time observations made at the localities just mentioned 

 In iokio and Yokohama the observations were usually made automa- 

 tically, by means of an instrument (which I have described in Vol IV of 

 ^\^ '<l?.^°'^r^*',°''',?^*^®. Seismological Society of Japan,' under the name 

 of a Time Taker ), which gives a record of the time of a disturbance 

 without stopping or retarding the clock from which it was taken. At Chiba 

 and Kumagai the records were taken by telegraph operators by means of 

 watches which I provided for them. At the remaining stations the 

 observations were made similarly. All watches and clocks were from 

 time to time compared with a telegraphic signal sent daily from the 

 Meteorological Department in Tokio throughout Japan. The only ex- 

 ceptions wereHakodate and Saporo, where the observations were made 

 at observatories well provided with the necessary means of obtainincr 

 accurate local time. ° 



The conclusions which these observations lead me to draw are— 



1. Different earthquakes, although they may travel across the same 

 district, do so with different velocities, varying between several hundreds 

 and several thousands of feet per second. 



2. The same disturbance is propagated with a decreasing velocity 

 travelling very much more quickly across districts which are near to the 

 origm than across districts which are far removed. 



3. The greater the initial force producing a disturbance the .Greater 

 the velocity of propagation. ° 



As examples of the observations which have led to these deductions I 

 quote the lollowing : — 



1. The earthquake of October 25, 1881. 



The origin of this was about 41° N. lat. and 144° 15' E. long. From 



