with perfect uniformity all round. Thus, one part of the surface will go down before another, and as this must usually occur by a cracking and sudden motion, the result will be an earthquake.
The seismic ribbons of which we have spoken are probably lines of weakness along which cracking habitually takes place. Along these lines there are enormous dislocations of the geological strata, and earthquakes are known to follow lines of faulting. The geologically recent elevation of the Pacific coast of South America is obviously, from this point of view, connected with the abundance of volcanoes and the frequency of earthquakes along the chain of the Andes.
One would think that the continual ejection of lava and ashes from an active volcano must leave a hollow under the mountain, and that some day the cavern would suddenly collapse. It has, however, been observed that volcanic eruptions and severe earthquakes are to some extent alternatives, so that it seems as though the volcanic vent served as a safety-valve for the liberation of the dangerous matter. But the theory of the collapsing cavern must not be wholly rejected, for some disastrous earthquakes affecting only very restricted areas, such as that of Casamicciola in Ischia, are hardly otherwise explicable. In this case Palmieri has attributed the formation of the cavern to evisceration under the town produced by hot mineral springs.
In the theories of which we have just spoken, the internal heat of the earth acts indirectly, by giving to gravitation an opportunity of coming into play. But as in volcanic eruptions enormous quantities of steam are usually emitted, it is probable that the pressure of steam is the force by which the lava and ashes are vomited forth, and that the steam is generated when water has got among hot internal rocks. From this point of view we can understand that an eruption will serve as a protection against earthquakes, and that the centers of disturbance will usually be submarine.
It may on the whole be safely concluded that a diversity of causes are operative, and that some earthquakes are due to one and others to other causes.
It would, however, be certainly wrong to look only to the interior of the earth for the causation of earthquakes, since the statistics of earthquakes clearly point to connections with processes external to the solid earth.
The laborious inquiries of M. Perrey show that there are more earthquakes at the time of full and change of moon than at other times, more when the moon is nearest to the earth and more when she is on the meridian than at the times and seasons when she is not in those positions relatively to the earth. The excess of earthquakes at these times is, however, not great, and an independent investigation of the Japanese earthquakes does not confirm Perrey's results. It is well, therefore, still to hold opinion in suspense on this point. If, however, Perrey's result should be confirmed, we must attribute it to
