1849.] LYELL ON THE STRUCTURE OF VOLCANOS. 223 



one mile in diameter surrounded by steep and lofty cliffs on every 

 side, save one, where the sea enters by a single passage nearly dry at 

 low water. In the interior of the small circular bay or crater there 



Fig. 9. 



Side view of the Island of St. Paul (N.E. side). Nine-pin i^ocks 

 two miles distant. Captain Blackwood, R.N. 



is a depth of thirty fathoms or 180 feet. The surface of the island 

 slopes away in every direction from the crest of rocks encircling this 

 crater. The highest peak is 820 feet above the level of the sea. If 

 we suppose considerable oscillations of level to occur by gradual move- 

 ments of upheaval and subsidence, the sea which has had power to 

 wear away part of the island and produce lofty and perpendicular 

 cliffs, would continue to keep open the single entrance, and as it deep- 

 ened it would also enter and undermine the walls of the crater, so as 

 to widen its area. Although by this means what is now the central and 

 higher portion of the island would be entirely destroyed, still high 

 interior cliffs would be produced, and a section of part of the volcano, 

 now submerged, would be laid open in the deep ravine excavated on 

 the eastern or lower side of the island. On every other side the rim 

 of the enlarged crater or caldera might remain unbroken. 



SOMMA. 



The evidence of Somma having been originally a submarine volcano, 

 has appeared more and more satisfactory in proportion as recent ob- 

 servations have been multiplied. MM. von Buch and Dufresnoy 

 affirm that the tuff which surrounds the mountain to the height of 

 1900 feet above the sea, contains marine shells analogous to those 

 which I found at the height of 2605 feet on the neighbouring volcanic 

 island of Ischia, all of which, except one, were of species now living in 

 the Mediterranean. As some of the component beds of lava preserve 

 throughout large spaces a uniform texture and are inclined at an angle 

 of 30°, it is inferred by MM. von Buch, De Beaumont, and Dufresnoy, 

 and probably with good reason, that they have now a much steeper 

 slope than they had originally. On such a slope, they observe, such 

 wide and compact sheets of lava could never have been formed. If, 

 instead of imagining the superimposed tuffs and lavas to have swelled 

 up like a great bubble according to the elevation-crater hypothesis, 

 we suppose that they gained their additional steepness when they 

 were traversed at successive periods by the dikes and veins with which 

 they are now reticulated, we may account for the high angle of their 

 dip, while at the same time the multitude of dikes, so far exceeding 

 those seen at any other neighbouring point of the Phlegrsean fields, 

 points to this spot as the grand focus of eruption in ancient as well 



VOL. VI. PART I. S 



