16 ART. 9. — S. IvUSAKABE. 



of the wave, as there exists more or less yielding in the rocks 

 through which the waves propagate, and also that, in view of 

 this inference, we do not see the necessity of assuming the path 

 of the tremors to be different from that of the principal shocks. 

 The present experiment relating to other modulus give, it seems to 

 me, still stronger foundations to support the above view. We must 

 not however forget that, it is unfortunately the common rule ra- 

 ther than the exception that a theory, however perfect it may 

 be, does not explain all the facts connected with it and also that 

 almost every phenomenon has more than a single cause, and this 

 is particularly true in the case of earthquakes. 



As the elastic constant varies during one cycle of bending 

 and all values at different phases of the cycle equally play their 

 parts in causing the vibratory motion, the apparent value of the 

 elastic constant during one complete vibration must be the mean 

 value of all the values at different phases. Now the mean elas- 

 ticity for one complete cycle being distinctly greater than what 

 is commonly adopted, the actual velocity of propagation for seis- 

 mic waves must be correspondingly greater than those given in 

 the above table, which are calculated by taking the square root 

 of the elasticity-density ratio. In the case, e.g., of a piece of 

 sandstone, the result of the experiment shows that the mean 

 value is 3'67 times greater than the constant term. Whence we 

 may infer that the actual velocity, in this case, would be probably 

 twice the value given in the table. 



Again, the velocity must necessarily diminish with an increase 

 of the amplitude of the wave, since the elasticity diminishes in 

 that case as explained above. From the example given there, 

 we may deduce the following to show how the velocity changes 

 with the amplitude. 



