REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 



77 



shore are hollowed out. On the Island of Tristan the gully lying behind the 

 settlement, in the centre of which the spring rises that supplies the village brook, 

 is formed in a similar way. It is banked by a vertical dyke, the thickness of which is 

 nearly 180 feet ; this injected rock has altered the encasing beds, which have become 

 schistose and break down readily. A large number of similar dykes can be seen in the 

 cliffs, but their thickness does not generally exceed one or two feet. The rocks of 

 the coast, presenting as they do good natural sections of the island, have enabled Mr. 

 Buchanan to establish at two points the existence of old vents, occupied now by 

 volcanic materials, which seemed to him products of subaerial eruption, slowly 

 deposited under water. This interpretation leads to the further admission, that 



certain parts of the Island of Tristan have, like several islands of the Atlantic, 

 been subjected to upheaval. 



In first describing the rocks that have been poured out as lavas, or projected as 

 incoherent volcanic materials, and now constitute the nearly horizontal beds, we must 

 point out, as one of the most important, a reddish yellow rock with large crystals of 

 augite. According to the observations of Mr. Buchanan, it has undergone profound 

 alteration under the inn uence of the dykes that traverse it. Some of the specimens of 

 it are almost completely disintegrated ; the augite crystals alone have resisted decom- 

 position, and they can be extracted with ease from the almost earthy mass that 

 encloses them. 



