44 H. I. JENSEN. 



perhaps forming among them, as the gases from burning 

 powder between the projectile and the breech of the gun, 

 might do it." 



3. Earthquakes resulting from volcanic evisceration 

 are not very frequent. Trough faults like the great Rift 

 Valley 1 may owe their origin to volcanic extravasation, 

 but even in areas thus undermined, the actual collapse is 

 most likely to take place when cooling proceeds most 

 rapidly, as at sunspot minima, and the cohesion of rock 

 strata thereby overcome. Closely allied are earthquakes 

 connected with the intrusion of sills and laccolites, as the 

 Port Resolution uplifts in island of Tanna, New Hebrides, 

 in 1878 and 1888. 2 



4. Earthquakes due to chemical degradation are rare, 

 and are mainly confined to districts in which there are 

 rocks of a soluble nature, such as gypsum, rocksalt or lime- 

 stone. Molten lava, says Milne ("Earthquakes," p. 283) 

 may by chemical action eat out great hollows in volcanic 

 regions. Earthquakes and landslips dependent on the first 

 kind of chemical degradation (e.<y., that which in 1840 

 caused Mt. Oernans in the Jura to fall) would take place 

 in wet periods ; those dependent on the corrosive properties 

 of molten lava would be most likely to occur at times 

 when other secondary causes are in full operation ; thus 

 the sudden expansion of heated vapours and gases conse- 

 quent on fall of barometric pressure at the earth's surface 

 would induce the corrosive magma to well out of its vent, 

 and on the return of normal atmospheric conditions the 

 spot is weakened and rendered liable to collapse. 



5. Changes of atmospheric pressure in themselves, and 

 changes of the gravitational influence of sun and moon, are 



1 See Prof. Gregory's work " The Great Eift Valley," Part III., Ch. xn., 

 pp. 214 - 236. 



3 See " Eeport of Aust. Assoc, for the Advancement of Science" for 1893 

 p. 209. 



