272 T, Teeada and T- Matuzawa 



Japan Sea side tlian on the Pacific coast. Theoretical discussions are 

 adduced in these papers. 



(c) Direction of first motion. The first motion of tlie eartli due 

 to tlie arrival of the shock has heen utilized from early days for the 

 estimation of the direction of the epicentre, and even of the focus. 

 Kawazoe^^^ showed hy his statistical investigation on a numher of 

 seismograms that the direction estimated by this riiethod falls within 

 a few degrees of that of the epicentre deterD)ined by the usual method 

 from the data of different stations, though sometimes a deviation amount- 

 ing to 10°, say, is met with. 



At a nieeting of the Tohyd Physico-Mathematical Society and 

 also of the E.I.C., in 1917, T. Shida^^^ read a paper on his investiga- 

 tion of the space distribution of the direction and sense of the initial 

 motion as an important indication of the ty]>e of disturbance at the 

 origin of the earthquake. Different exami)]eR of the case in which 

 two straight lines could be drawn, passing through the epicentre and 

 intersecting each other at an angle nearly equal to a riglit angle, 

 such that they divide the area of the map into four quadrants and 

 mark the boundaries of the regions with an alternate sense of the first 

 motion. Thus, while the stations lying in a certain quadrant show 

 a ' push ' from the epicentre, those situated in the adjacent quadrants 

 show a ' pull ' from the same point. His results have aroused great 

 interest among seismologists in this country. His step forward has 

 been, successfully followed by Saem. ISrakamura*^"^ K. Suda*^^ and 

 others^^\ The cases with much more complicated trends of the 

 ' nodal lines/ i.e. the lines marking the boundaries of two adjacent 

 regions with the opposite motion, have been brought forth and dis- 

 cussed with regard to the mode of disturbance at the origin, which 

 was considered as a combination of sim|)le or multiple 'sources'. 

 Theoretical considerations have been applied to different examples 

 with success, so that the distribution has now become a most im- 

 })ortant clue for the classification of the types of the original 

 disturbances. 



(1) J.M.S., 37 (1918), 53. 



(2) Not published. 



(3) J.M.S., 41 (1922), e. 1; 37 (1918), 390; 38 (1919), 395; 39 (1920), 134; 41 

 (1922), 139; 42 (1923), 1 ; Saito Hoonkwai Gaknzyutu Kenkyu H6kokii, No. 1 (1925). 



(4) U.t.S., 3 (1923), 14; 5 (1925), 36, 80. 



(5) T. Usiyaraa, J.M.S., 40 (1921), 135 ; 41 (1922), 4, 222 ; Kensinzih6, No. 1 (1925), 

 171. S. Aoki, J.M.S., 39 (1920), 192. K. Taguti, U.t.S., 4 (1924), 199, 



