384 SEE— THE CAUSE OF EARTHQUAKES. [October 19, 



If we consider with attention how simple the motion of sub- 

 sidence or elevation, in a supposed fault, really is, we shall per- 

 ceive that such displacement cannot give the earth particle an ap- 

 preciable rotatory motion. For the two sides of a fault are con- 

 ceived as continuous and unbroken parts of the earth's crust, and 

 thus securely fixed at their backs, and moving only at their faces, 

 where the fracture exists. The motion is essentially like that of 

 a double cellar door opening or closing very slightly in the middle, 

 with rigid hinges at their backs. They can only go up or down 

 together, or one up and one down, with little or no horizontal 

 motion; and thus cannot produce a revolving tremor when opened 

 or shut. 



It is exactly so with the slipping of a supposed fault in the 

 earth's crust. It is mechanically inconceivable how the vertical sub- 

 sidence of a fault could give the earth particle a revolving motion in 

 the plane of the horizon. At most, such a slip could produce an 

 up-and-down oscillation, with the path of the ellipse described by 

 the particle very nearly in the vertical plane. 



Now the earthquake which destroyed San Francisco gave the 

 earth particle a large and conspicuous rotation almost parallel to 

 the horizon ; it was " a twister," and this rotatory character was 

 so marked as to attract instant attention. The conspicuous rota- 

 tion and the difficulty of explaining such a motion by the theory of 

 faults, as ordinarily stated, led the writer to question the validity 

 of that theory. As most of the violent earthquakes have both rota- 

 tion and vertical movement, it is evident that the difficulty felt in 

 explaining the San Francisco earthquake is very generally encoun- 

 tered in earthquake phenomena. 



It seems to have escaped the attention of seismologists that 

 rotatory earthquakes require explanation, and the theory of subsi- 

 dences of rocks and faults is incapable of furnishing it. How then 

 can motion of rotation be accounted for? 



To answer this question in the simplest way we may recall that 

 there are many impulses which can give a motion of rotation, either 

 by direct impact or by the " kick back," or recoil of reaction. If, 

 for example, we suppose an orifice to be forced through an under- 

 lying or overlying layer of rock so that lava escapes under great 



