152 AMSTERDAM ISLAND. 
of facts, we shall perhaps conclude in favour of the latter 
opinion. The expansive force of steam, under certain cir- 
cuinstances, is be3''ond calculation ; and steam is unquestion- 
ably formed in all ^'olcanic explosions, as the presence of a 
vast body of water seems to be indispensably necessary in 
these operations of nature : for tlie greater part of, if not all, 
real volcanoes are either islands, or are situated so near the 
sea as to ha^ e received a supply of water from it. It is well 
known that a torrent of looiling water rolled down the sides 
of mount Etna about the middle of last century, and the 
same circumstance has happened more than once in Vesuvius. 
It is also now understood that the earth contains a variety of 
substances, and those in large quantities, which, in contact 
with water, will cause ignition, and others that will support 
it without the presence of air. If then, by means of a rent 
or chasm in the bottom of the sea, a sufficient quantity of 
water be brought to act on a proportionate mass of materials 
capable of supporting ignition, the steam generated in con- 
sequence must heave up the superincumbent earth, and pro- 
bably, on some occasions, without disturbing the arrangement 
of its parts as they existed at the bottom of the sea previous 
to the eruption, except immediately in those places where 
the explosion burst out. Thus we see in many islands of vol- 
canic origin, as Teneriffe and Madeira, large portions of the 
surface where the action of fire is not in the least discernible. 
That new islands are produced from time to time, and pro- 
bably in this manner, history affords us various instances. 
One of the most considerable of the Lipari islands, called 
Vulcam, was created in the time of the Roman republic. 
Since the seventh century of the Christian aera three distinct 
7 
