on Earthquakes in Japan. 377 



Evidence that the rocks near to the edge of one of the bluffs 

 above Yokohama were subject to great stress during a recent 

 earthquake is obtained in the form of long cracks parallel 

 to the edge of the bluff. It is more difficult to account for 

 the damage done to buildings close to the foot of steep cliffs ; 

 but possibly an explanation may be afforded by reflexion of 

 vibrations causing increased activity in that region. 



Rotation of Bodies. 



After a severe earthquake it has often been observed that 

 many heavy bodies (like obelisks, gravestones, chimneys, &c.) 

 have more or less rotated. This phenomenon has been explained 

 in many ways. Some have supposed that they indicate a 

 vorticose motion in the ground; others that the phenomena are 

 due to reflected and direct shocks acting simultaneously on the 

 body which has been rotated. Mallet offers the explanation 

 that the rotation is due to the centre of friction of the base of 

 the body not coinciding with its geometrical centre. 



In certain cases no doubt these explanations, especially the 

 latter, may be correct. 



If vorticose motion is the cause of the rotation, we ought to 

 find^that all twisted bodies near the same place have been turned 

 in the same direction; and if we take Mallet's view, the 

 rotation ought to follow no definite law. 



What we actually find, after having examined a great num- 

 ber of cases, is that they do follow a law, but not that which 

 would be obtained on the supposition of vorticose motion. 



The law is that all similar bodies, such as gravestones, having 

 similar sides parallel, are in the same district rotated in the 

 same direction ; while another set, having their faces placed 

 at an angle to these, may be rotated in an opposite direction. 



This is accounted for if we suppose rotation to be due to a 

 direct shock. If the shock comes broadside on to the body, it 

 only tends to throw it over ; but if it comes obliquely to this 

 direction (but not exactly along a diagonal), it tends to tilt it 

 up on a corner. This shock may be resolved into two com- 

 ponents — one along the projection of the line joining the 

 centre of inertia and the corner, tending to tilt it up on that 

 corner, and the other at right angles to that direction tending 

 to whirl it round. This explanation is not only verified by 

 observations after an earthquake, but is easily verified by 

 experiment. It will be seen that, according to this explana- 

 tion, the direction of rotation may be used for the determina- 

 tion of the direction of vibration of a shock. One direction 

 produces a rotation corresponding to the hands of a watch, 

 whilst another gives an opposite rotation. It will also be 

 observed that a body tends to turn until it becomes broadside 

 on to the direction of shaking. 



