508 



ISLAND OF ST. PAUL. 



[Ch. XXIX. 



on the east and west coast, and aie many of them nearly as naked and 

 barren of vegetation as when they first flowed. The tendency in vol- 

 canic vents to assume a linear arrangement, as seen in the volcanoes of 

 the Andes and Java on a grand scale, is exemplified by the cones and 

 craters of this small range in Palma. It has been conjectured that such 

 linearity in the direction of superficial outbreaks is connected with deep 

 fissures in the earth's crust communicating with a subjacent focus of sub- 

 terranean heat. 



By discussing at so much length the question whether the sea may or 

 may not have played an important part in enlarging the Caldera of 

 Palma, I have been desirous at least to show how many facts and obser- 

 vations are required to explain the structure and configuration of such 

 volcanic islands. It may be useful to cite, in illustration of the same sub- 

 ject, the present geographical condition of St. Paul's or Amsterdam 

 Island, in the Indian Ocean, midway between the Cape of Good Hope and 

 Austi'alia, 



Fie. 648. 



Map of the Island of St. Paul, in the Indian Ocean, lat. 38° 44' S., long. 77° 37' E., 

 surveyed by Capt. Blackwood, R. N., 1842. 



in this case the crater is only a mile in diameter and 180 feet deep, 

 and the sui-rounding clifts where loftiest about 800 feet high so that in 

 regard to size such a cone and crater are insignificant when compared 

 to the cone and Caldera of Palma or to such volcanic domes as Mounts 

 Loa and Kea in the Sandwich Islands. But the Island of St. Paul ex- 

 eirjplifies a class of insular volcanoes into which the ocean now entei's by 



