Figure 3 shows the relation between the range of the secondary 

 undulations and various meteorological elements for some stations. 

 The thick full line ( ) shows the range (double amplitude) A (in cm.) 

 of the secondary undulations of longer periods (proper oscillations of 



bays), the thin full line ( ) the range (double amplitude) B (in cm.) 



of the secondary undulations of shorter periods (very long swells), the 

 thick broken lines (— — — — — ) the velocity of wind W (in m./sec.) 



at each station, the chain-line ( — • • — • •) the magnitude of the rate 



of change of the atmospheric pressure P, the chain-line ( — ) 



the pressure gradient G between the centre of the cyclone and each 



tidal station, and the dotted line ( ) the magnitude — — 



of the rate of change of the pressure gradient G. 



In the following we shall summarize the results obtained from the 

 present investigation : — 



(1) For stations situated along the Pacific coast of Honsyu, Sikoku, 



and Kyusyu conspicuous undulations of longer periods (proper 

 oscillations of bays) began six to twelve hours before the time 

 of nearest approach of the cyclonic centre to respective 

 stations, and conspicuous undulations of shorter periods (very 

 long swells or waves of one to three minutes periods) generallj?- 

 began somewhat later. 



(2) The conspicuous undulations of longer periods peculiar to the 



stations began about the time of arrival of the 755 mm. 

 isobar at the respective station. 



(3) The conspicuous secondary undulations under question seem to 



have no close relation with the velocity of wind, pressure 

 gradient, or the rate of variation of atmospheric pressure in 

 the vicinity of the tidal station. 



(4) It seems that the present conspicuous undulations were not 



produced hy the passage of lines of discontinuity and cold 

 fronts or by local showers. Besides, it seems difficult to 

 consider that the microbarographic oscillations in the vicinity 

 of the tidal stations became the prime cause of the undulations. 



(5) The secondary undulations in question seem to have some 



apparent relation with the time variation of the atmospheric 

 pressure gradient — the amplitude of the undulations is large 

 when the rate of change of the pressure gradient is large. 



(6) It seems likely that the present conspicuous undulations w^ere 



not produced directly by the meteorological changes in the 

 immediate vicinity of the tidal stations, but chiefl}' by the 

 waves which, were produced in some region near the centre 

 of the cyclone, perhaps by some microbarographic change, and 

 propagated toward the tidal stations. 



(7) As is naturally expected, the amplitude of the present secondary 



undulations has a close relation with the distance of the 

 c3'Xlonic centre from the tidal station. This relation is closer 

 in the case of the undulations of shorter periods than those of 

 longer periods. 



(8) There is a tendency for the secondary undulation to be large 



when the depression of the cyclonic centre is low. 



A detailed discussion will appear in a publication of the Central 

 Meteorological Observatory in the near future. 



90 



