32 



Art. 9. -T. Terada : 



at the interval of periods 5(/-60! The curve reminds us ratlier 

 of the energy distribution of the kiininous spectra than the usual 

 probability curves. A secondary maximum is seen near the 

 period 90! The results for other years are of quite similar 

 character. 



Table III. 



For each season, the first column gives the number of trains and the 

 second the percentage of the total number. 



1912. 



X. 



The result must not be hastily interpreted as showing that 

 the absolute maximum of frequency is at the above mentioned 

 interval. If we could multiply the sensibility of the instruments 

 and turn at the same time the recording drum correspondingly 

 faster, the waves most conspicuous in our case might possibly be 

 replaced by others with a decidedly shorter periods. Besides, 



