C. Davison — Great Japanese Earthquake, 1891. 



25 



the proper proportion. The outer curve in each region corresponds 

 to five epicentres in a rectangle, and the inner to ten epicentres. 



In the central area of Mino and Owari, it will be noticed that the 

 curve of ten epicentres is forked to the south. The great fault-scarp 

 probably begins within the isolated curves in the north part of the 

 area, and follows the east branch of the fork ; but there can be little 

 doubt that a second fault traverses the branch proceeding towards 

 the south, though no corresponding scarp was observed upon the 

 surface. The curves in the regions marked D and F seem to point 

 to the existence of a pair of transverse faults, one on each side 

 of the Mino-Owari fault-system, and roughly parallel to it. The 

 earthquakes in the region A are probably connected with a strike- 

 fault, and those in the region B with another transverse fault. The 

 curves marked C and E are too small to offer any reliable guide as 

 to the origin of the earthquakes within them. 



In the following table are given the numbers of epicentres which 

 lie within each of these regions, for every year from 1885 preceding 

 the date of the great earthquake (October 28, 1891), and for every 

 month following it until the end of 1892.^ 



1 Occasionally an epicentre is defined in Professor Milne's catalogues as lying 

 between two consecutive rectangles. In these cases I have counted half with each, 

 and this accounts for the occurrence of fractional figures in the table. 



