. 
8 Alexis Perrey on Earthquakes. 
It cannot be too often repeated, that earthquakes are not of 
one single kind, identically the same. They are various both in 
causes and effects; I aim simply to bring out in relief the prin- 
been observed in the Pyrenees and the Andes. In the great 
valleys occupied by rivers, the mean direction, as calculated by 
rt, appears to be that of the course of the depression. I 
have shown this to be the fact with the basins of the Rhone and 
the Rhine, where the direction is nearly meridional, and the 
basin of the Danube, which has a transverse course, or from west 
east. 
In France, the departments most subject to earthquakes appear 
to be those about the mouths of the large rivers. The depart- 
Rhone forms a kind of node with that of the Sadne, is the only 
one which can compare with the kind just mentioned in number 
of earthquakes. 
Whatever may be the cause of the molecular vibration at any 
be spherical and concentric. How will it then be in a medium 
which is not homogeneous, or is of unequal density? This can- 
succeed one another through each point in the sphere of undula- 
tion and make successive shocks at the earth’s surface, the shocks 
directly over the centre or focus of the vibrations will be verti- 
eal: and the obliquity, or variation from verticality, will be 
greater the more remote the place of emergence at the surface is 
from the centre of vibration alluded to; or, the locality being 
fixed, the nearer this centre is to the surface. 
There can be no rotary shocks; the cases of apparent rotation 
ndicate the point from which it actually comes? I believe not. 
The difference in the rocks encoun! should produce deriva- 
tive and reflected undulations, as in the case of waves of sound. 
we have explained elsewhere. But does the direction of a shock 
