HEAT — VOLCANOES. 701 



noes along great lines of mountains, and in linear series, as if over 

 profound fractures, are facts sustaining this view of their source. 



The necessities of .any single volcano, like Etna, or a single cluster 

 of them, like those of Hawaii, might be met by a separate lake 

 of fire beneath. But when an ocean like the Pacific is girted by 

 volcanoes, and also blotched all over with the results of volcanic 

 fires, such a hemisphere of volcanic action needs a vast sea of fire 

 beneath, if not a hemisphere of igneous action. A volcano is ever 

 discharging heat into the air by its lavas and gases, and must- have 

 some deep source below. 



Supposing the interior of the globe to be fluid, and this to be 

 the primary source of volcanic action, it does not follow that a 

 connection with the interior is retained by every active volcano. 

 After beginning on a fracture reaching through the crust, it may 

 have become cut off by cooling, so as afterwards to extend only to 

 a reservoir of liquid rock ; when several volcanoes have been 

 opened on a single profound fracture, they may afterwards have 

 become wholly disconnected from one another as well as from the 

 earth's interior. Known facts about volcanoes do not settle this 

 question, though favoring the idea of disconnection. Kilauea, on 

 the flanks of Mount Loa, is one of the largest volcanic craters on 

 the globe ; and yet eruptions occur at the summit of the same 

 mountain, 10,000 feet above the level of Kilauea, and so extensive 

 that the lavas flow off for 25 to 50 miles without any sign of sym- 

 pathy in the lower crater. If the two are ' connected, the siphon 

 has the liquid 10,000 feet higher in one leg than in the other. 



This connection is possible only on two suppositions:— (1) that it 

 is at such a depth that 10,000 feet is but a small fraction of the 

 whole length, and the additional pressure is more than counter- 

 balanced by the friction along the conduits ; or (2) that, if the 

 lavas rise in consequence of an inflating process, the difference of 

 length may not imply a corresponding difference of pressure. 



Even about Kilauea itself eruptions sometimes take place 

 through the upper walls of the crater to the surface (as at P, fig. 

 968), when the lavas are boiling freely in the bottom of the crater, 

 undisturbed by the ejection. 



While the linear arrangement of the volcanic mountains of a group is evi- 

 dence that they all originated in one grand breaking of the earth's crust, the 

 several volcanoes in a line may not stand over one prolonged fracture, but over 

 a series having a common direction, in the manner illustrated by the figures 

 on page 19. This was beyond question the mode of origin of the Hawaian 

 Islands. 



The islands of Oahu and Maui (see fig. 24, p. 31) consist each of two great 



