l^j^ F. OMORI. 



The sounds of the «econd type, which were sometiines like detona- 

 tions of tliuiider, were most frequent and distinct in the Neo-A^illey, 

 where, as I beUeve, their main origin really was. It is remarkable that 

 tremblings of the ground accompanying these sounds were invariably 

 very feeble, and often not to be felt at all, while severe sharp shocks 

 were usually not accompanied by distinctly audible sounds. This 

 peculiar phenomenon was ascertained likewise to have been observed 

 with the after-shocks of the Noto and Kagoshima earthquakes. The 

 following may be one of the possible explanations. 



Among ruimerous de[)ressi()ns of small pieces of ground produced 

 by the great eartlnpiake of 1<S1)1, there was one which took ])]ace on a 

 high inoiuitain flunk near the village of Higashi-Yokoyama, in the 

 Ono District, Mino, and which was cyHndrical in form, being about 

 9 feet in diameter and 10 feet in deptii. The top of the cavity was 

 just shewn on the surface slope as a circular aj)erture about 4 feet in 

 diameter. From examples like this, we may suppose that some 

 cavities have probably been formed undergrouiid which are not shewn 

 on the surface. Especially in the focal region there may exist big 

 vacaiil spaces, and the falling down into these of superincumbent 

 rock masses would give rise to sounds accom|)anied by small move- 

 ments of ground, as muffled s(junds are heard when we throw stones 

 into a deep well or IkjIc. On the other hand, shocks may originate 

 by fracturing of strata in distiicts near the focus which would cause 

 sh;u"p tremblings of the ground, but not be accompanied by loud 

 sounds. 



The slight rushing sounds often accompanying earth([uakes may, 

 as usually supposed, be caused by (piick tremors preceding the 

 principal earthquake motion. 



