292 Baron Humboldt on Volcanoes. 



(1773-1822) almost without change in its elevation above the 

 level of the sea. Any diflPerence that appears may be looked 

 on as within the possible errors of measurement. 



Volcanoes whose summits reach far above the limits of per- 

 petual snow, like those of the Andes, present a variety of 

 peculiar features. The sudden melting of the snow in the 

 course of an eruption, not only occasions destructive floods, 

 torrents in which heaps of smoking ashes are floated away on 

 blocks of ice ; but the accumulation of ice and snow goes on 

 producing its influence uninterruptedly, and by filtration, into 

 the trachj'tic rocks, even whilst the volcano is jjerfectly qui- 

 escent. Caverns are thus gradually formed on the declivities 

 or at the foot of the burning mountain, and these become sub- 

 terraneous reservoirs of water, which communicate in various 

 ways, by narrow mouths, with the alpine rivulets of Quito. 

 The fishes of these alpine streams multiply greatly, particu- 

 larly in the gloom of the caverns ; and then, when the earth- 

 quakes come, which precede all eruptions of volcanoes in the 

 Andes, and the whole mass of the mountain is shaken, the 

 subterraneous caverns at once give way, and pom* out a deluge 

 of water, fishes, and tufaceous mud. This is the singular 

 phenomenon which the presence of the Pimelodes Cyclopum, 

 the Prenadilla of the inhabitants of the lofty plains of Quito, 

 attests. When, in the night between the 19th and 20th of 

 June 1698, the summit of Carguaii'azo, a burning mountain 

 18,000 feet high, crumbled together, so that no more than two 

 enormous rocky horns of the crater's edge remained, the 

 country, for nearly two square miles, was desolated with liquid 

 tufi", and argillaceous mud (lodazales), enclosing dead fishes. 

 In like manner the putrid fever of the mountain town, Ibarra, 

 to the north of Quito, which occurred seven years before, was 

 ascribed to an eruption of fish from the volcano Imbaburnu. 



Water and mud, which, in the volcanoes o:^ the Andes, do 

 not pour down from the crater itself, but from cavities in the 

 trachytic mass of th<j mountains, ought not, consequently, in 

 the strict sense of the phrase, to be reckoned among the num- 

 ber of proper volcanic phenomena. They are only mediately 

 connected with the activity of volcanoes, nearly in the same 

 measui-e as the irregular meteorological process, which, in my 



