Januabt 22, 1909] 



SCIENCE 



131 



little by little approaches the limit of the 

 resisting force there is a critical period 

 during which a relatively small additional 

 stress arising from some other source may 

 precipitate the catastrophe. A number 

 of possible sources for the additional 

 stress are known, the influence of several 

 has been fairly demonstrated in a statisti- 

 cal way, and it is on the whole probable 

 that a large majority of earthquakes owe 

 their precise dates to such contributory 

 causes. Many of the precipitating factors 

 are periodic in their character, and the 

 times of their maxima, or other favorable 

 phases, are known; so that, granting their 

 influence, they serve to restrict prediction 

 to certain epochs. They are not of 

 primary importance in forecasting, but 

 when the approximate date of a future 

 earthquake shall have been learned by 

 other means, they wiU serve to refine the 

 estimate of time. 



The principal known causes of periodic 

 variation of stress are bodily tides of the 

 earth; oceanic tides, which alternately 

 load and unload the sea bed near the 

 shore ; the winter load of snow on parts of 

 the land; annual and diurnal variations 

 of atmospheric pressure; diurnal varia- 

 tions of barometric gradient; and the 

 wandering of the earth's axis of rotation. 

 The relative importance of the several in- 

 fluences can not yet be indicated, but it is 

 known that their absolute importance is 

 not the same in all places. Three belong 

 to the coastal belts, two to the land; and 

 two belong to land and sea, but vary with 

 latitude. Their relative importance in 

 any particular locality may depend also 

 on the direction of the slowly growing 

 tectonic stress of the crust; for in order 

 to be effective the temporary or adventi- 

 tious stresses must be of such character as 

 to augment the tectonic stresses. Let me 

 illustrate this point. 



The tidal sway of an oceanic basin 



raises and lowers the surface very little 

 where the water is deep, but has a much 

 greater effect on the shoals bordering 

 coasts. The strip of sea bed following 

 the coast is subjected twice a day to the 

 addition of a heavy load of water, and in 

 the intervening hours is relieved of pres- 

 sure by the same amount. On the sea- 

 ward side of the strip there is a gradual 

 change in pressure, and on the landward 

 side, just at the water edge, an abrupt 

 change; and these pressure differences 

 cause strains and stresses in the crust be- 

 neath. The directions of the induced 

 strains lie in vertical planes at right 

 angles to the coast, and are competent to 

 increase or diminish tectonic stresses hav- 

 ing similar directions. On the coast of 

 Alaska near Mt. St. Elias the tectonic 

 changes in progress include an uplift of 

 mountains parallel to the coast, and the 

 main tectonic stresses may be assumed to 

 lie in vertical planes normal to the coast; 

 so that here the oceanic tides are competent 

 to precipitate earthquakes. But on the 

 California coast near San Francisco, 

 where the directions of the main tectonic 

 stresses are horizontal and are approxi- 

 mately parallel to the coast, the stresses 

 from oceanic tides may be ineffective. 

 On the other hand the stresses created in 

 the crust by the shifting of the axis of 

 rotation are probably better calculated to 

 augment tectonic stresses at San Francisco 

 than at Mt. St. Elias. 



Unfortunately the value to the fore- 

 caster of the periodic stresses is impaired 

 by the occurrence of other transient 

 stresses which are not periodic. The bar- 

 ometric gradients and extremes of pressure 

 connected with cyclonic storms are of this 

 class, and so are the pressure changes 

 arising when the sea is pushed against the 

 land or drawn from it by strong wind ; and 

 all these storm effects are at times much 

 greater than the rhythmic changes of cor- 



