14 ART. 1. — S. KUSAKABE : FREQUENCY OF 



dependent on the geological nature of the region under considera- 

 tion^ hut it must be greatly affected by the geological distribution 

 of the rocks lying between the seismic focus and the said region. 



Ill other words, the seismic-wave-conductivity, if we may 

 be allowed to employ such a terra from some analogies in heat 

 and electricity, may be different for different rocks, so that in 

 one direction an earthquake may be propagated with comparatively 

 smaller fading than in another direction. 



Now, as was experimentally shown by the author himself,* 

 the amount of hysteresis generally decreases with an increase of 

 the elastic constants, while the latter increase with their age of 

 formation. Although nothing, as yet, can be said about any 

 numerical relation between hysteresis and the age of formation 

 of different rocks, yet in the rocks so far examined, a certain' 

 relation seems to exist. 



To illustrate this fact, the hysteresis curves for a few speci- 

 mens of rocks are shown in Figs. 2 and 3. If there were no 

 hysteresis, the curve would, of course, shrink to a single line, 

 so that the amount of hysteresis may be conventionally compared 

 by the area enclosed by the curve. It is, then, a matter of 

 fact that the area is enormously great for new rocks such as 

 sandstone and rhyolite, and gradually diminishes as rocks of 

 older ages are examined, until it becomes very small for the 

 oldest rocks of the Arcluean age. Thus it would not be a wild 

 conjecture to say that the amount of hysteresis gradually diminishes 

 from Cainozoic, through Mesozoic and Palieozoic, to Aicluean 

 rocks, in a definite, though not yet ascertained, ratio. 



An inference to be drawn from the above is that the seismic 



* This joiiriKil, Vol. XIX, Alt. 6, and Vol. XX., Ait. 9. The Pub. of the K. I. C. in 

 F. L. Nos. 14 and 17. Tokvo. 



