THE GREAT EARTHQUAKE OF S. E JAPAN )49 



first and third sectors showed the pull towards the origin, while those 

 at stations in the second and the fourth showed the push from it. It 

 is also to be remarked that in all the other statioias except Titizima 

 the earth movenjents began by the push. Consequently thti seismic 

 action may have been applied in such a manner that the compression 

 along a deep stratum was directed towards the north and the west 

 while the counter action took place iu the upper side along the sectors 

 which showed the jAill. (PI. III.) 



PI. XVII. shov/s isochronals and isoseismals deduced from observations 

 made at different stations. Irregularities in these curves are due to 

 the variety of physical characters of the media through which tlie 

 seismic waves were propagated. The area in which the earthquake 

 was actually felt is about 1000 km. in diameter, rather smaller than 

 that of the great Japan Earthquake of 1707 or the Lisbon Earth- 

 quake of 1755. 



8. Observations at foreign stations. The present earthquake was 

 registered at seismological stations distributed in different parts of the 

 world. I have compiled a table showing the results of observations 

 at 48 such stations concerning the time arrivals of the lirst phase, 

 the second phase and the long waves'^ It gives a pretty good 

 " Laufzeitkurve," i. e. a curve showing the relation between the arcual 

 epicentral distance and the time of transmission of each phase. The 

 fairness of the curve may be due to the extraordinary magnitude of 

 the original disturbance and the accurate knowlege of the position of 

 origin. 



9- Statistics of the casualties and the distribution of seismis 

 intensity. Table II sliows the number of casualties in different 

 prefectures. 



I have prepard another table showing the numbers of cisualties 

 in the cities, towns, and villages most severely affected, and of the 

 houses which existid there before the catastrophe. Here this tabic is 

 omitted, but the general results are shown. The distribution in space 

 of the percentage of destroyed houses serves to indicate roughly the 

 distrilmtion of seisnuc intensity. These data are useful in determining 

 the isoseismal lines. 



4) Keports (in Japanese) of tlie Imp. Eartliq. Inv. Coinm., No. 100 A. p. 32. 



