40 On the Constitution and Construction of the 



vent ; and, " as a bubble of air let into a barometer tube drives 

 the mercury into the Torricellian vacuum far above the baro- 

 metric height, aqueous vapour may raise a column of lava of a 

 height equal to its expansive force into the channels opening 

 into the craters."* We may thus have a continual alterna- 

 tion of columns of lava and steam rising in succession, the 

 consequence of which would be an alternate ejection of lava, 

 red-hot masses, and clouds of vapour, as well exemplified at 

 Stromboli. 



Now, to enable the steam to exert the prodigious force re- 

 quired, we must suppose both it and the lava on which it acts 

 to be lodged in a cavity, surrounded on all sides by solid resist- 

 ing walls, with no opening but the volcanic vent and the chinks 

 by which water is admitted. If lava were part of a central 

 fluid nucleus, water reaching it at any point, after being con- 

 verted into vapour, would glide along the under part of the 

 solid crust, and settle at the highest vaulted cavities, till ad- 

 ditions to its quantity or its temperature enabled it to open a 

 passage for itself, or for a portion of the lava, through the 

 crust. Water passing downwards from the ocean near the 

 shore, might thus create a volcano at the distance of 500 or 

 1000 miles inland, as readily as near the coast. But the fact 

 of all active volcanoes being near the sea, or large bodies of 

 water, is at variance with this supposition. On the contrary, 

 it lends support to the conelusion, first, that water is a ne- 

 cessary agent in volcanoes ; and, next, that the fluid matter 

 upon which it acts exists in isolated basins of various forms 

 and dimensions, confined within solid rocks. Thus, one basin 

 150 miles in length may exist under southern Italy, connecting 

 Vesuvius, the Lipari Isles, and Etna. There may be one 200 

 miles in length and breadth, under Iceland ; and a vast trough 

 4000 miles or more in length, but of comparatively small 

 breadth, may extend under the Andes. The sudden and si- 

 multaneous activity of three volcanoes in the Cordillera, far 

 distant from each other, which broke out from a state of re- 

 pose into violent eruption on the same day, favours the idea 

 of a subterranean connection between them ; and the existence 



^ Jischof s Paper, p. 38, 



