Ch. XXIX.] OF YOLCAXIC ROCKS. 625 



a 





e 



Fig. 694 





& 





e 



d\ 







~> 



Section of the crater of Kilauea in the Sandwich Islands. (Dana.) 

 or, &. External boundaries of the chasm in the line of its shortest diameter 

 c, <?, /, d. Black ledge. g, h. Lake of lava. 



thickness, and nearly horizontal. Before this, we come to what is 

 called the " black ledge," c, e, and /, d, composed of similar stratified 

 materials. This ledge is three hnndred and forty-two feet in height 

 above the lake of lava, g, h, which it encircles. The chasm, a, 6, 

 and its walls hare probably been dne to a former sinking down of 

 the incumbent rocks undermined for a space by the fusion of their 

 foundations. The lower ledge, c, e, and /, d, may consist in part of 

 the mass which sank vertically, but part of it at least must be made 

 up of layers of lava, which have been seen to pour one after the other 

 over the "black ledge." If at any future period the heated fluid, 

 ascending from the volcanic focus to the bottom of the great chasm, 

 should augment in volume, and, before it can obtain relief, should 

 spread itself subterraneously, it may, melt still farther the subjacent 

 masses, and, causing a failure of support, may enlarge still more the 

 limits of the amphitheatre of Kilauea. There are distinct signs of 

 subsidences, from 100 to 200 feet perpendicular, which have occurred 

 in the neighborhood of Kilauea at various points, and they are each 

 bounded by vertical walls. If all of them were united, they would 

 constitute a sunken area equal to eight square miles, or twice the ex- 

 tent of Kilauea itself. Similar accidents are also likely to occur near 

 the summit of a dome like Mount Loa, for the hydrostatic pressure of 

 the lava, after it has risen to the edge or lip of the highest crater, a, 

 fig. 693, must be great and must create a tendency to lateral fis- 

 suring, in which case lava will be injected into every opening, and 

 may begin to undermine. If, then, some of the melted matter be 

 drawn off by escaping at a lower level, where the pressure would be 

 still greater, the whole top of the mountain, or a large part of it, 

 might fall in. 



Instances of such truncations, however caused, have occurred in 

 Java and in the Andes within the times of history, and to such events 

 we may perhaps refer a very common feature in the configuration of 

 volcanic mountains, — namely, that the present active cone of eruption 

 is surrounded by the ruins of a larger and older cone, usually present- 

 ing a crescent-shaped precipice towards the newer cone. In volcanoes 

 long since extinct, the erosive power of running water, or, in certain 

 cases, of the sea, may have greatly modified the shape of the " atrium," 

 or space between the older and newer cone, and the cavity may 

 thereby be prolonged downwards, and end in a ravine. In such 

 cases it may be impossible to determine how much of the missing- 

 rocks has been removed by explosion at the time when the original 

 40 



