10 



s. Nakamura and K. Honda : 



Hakoné, Aug. 31, 1901. 



10 



:iu 



:ju 



40 



ùo 10" 



Fis. 5 a. 



10 



•,>o 



3U 



éO 



Hakoné, Au^^ 31, 1901. 



30 



40 



50 



i:ih 



1(1 20 



Fig. 5 b, 



30 



M 



50 



approximate values of periods, we get the numbers 12.98, 29.8, and 

 43.4, whence we deduce the periods 15.37, 6.65 and 4.64, minutes. 

 199.8-13 = 15.37. 



31= 6.43,-) 



30= 6. 65 J 



29= 6.88.J 



42= 4.75, 



43= 4.64,; 



44= 4.53. 



This example shows that the method is adapted for determining 

 longer periods. Of course the actual existence of these periods 

 must be tested in each case. This may be done by means of the 

 framework above mentioned, or by tlie method which will l:)e 

 explained presently. 



The determination of phases of several component oscillations 

 is a difficult task. For this purpose, the several components must 

 be separated from one anotber. To do this analytically will be too 



