219 



accumulates into elevated hills. At seven 

 leagues from Lisbon, near Colares, during the 

 terrible earthquake of the first of November, 

 1755, flames and a column of thick smoke were 

 seen to issue from the flanks of the rocks of 

 Alvidras, and, according to some witnesses, from 

 the bosom of the sea *. This smoke lasted se- 

 veral days, and it was the more abundant in 

 proportion as the subterraneous noise, which ac- 

 companied the shocks, was louder. 



Elastic fluids thrown into the atmosphere 

 may act locally on the barometer, not by their 

 mass, which is very small, compared to the mass 

 of the atmosphere ; but because, at the moment 

 of the great explosions, an ascending current is 

 probably formed, which diminishes the pressure 

 of the air. I am inclined to think, that in the 

 greater part of earthquakes nothing escapes 

 from the agitated earth ; and that, where ga- 

 zeous emanations and vapours take place, they 

 oftener accompany, or follow, than precede the 

 shocks. This last circumstance explains a fact, 

 which seems indubitable, I mean that mysterious 

 influence, in equinoctial America, of earthquakes 

 on the climate, and on the order of the dry and 

 rainy seasons. If the earth generally act on the 

 air only at the moment of the shocks, we can 

 conceive why it is so rare, that a sensible me 



* Phil. Trans., t. xlix, p. 414. 

 2 L 



