16 Transactions of the South African Philosophical Society. 



hundreds of miles from the mainland, and surrounded by water of 

 almost abysmal depth. Incidentally, also, I shall raise the question 

 of the origin of volcanoes, and shall adopt Dr. Sterry Hunt's theory 

 that lavas are re-melted portions of the already solidified crust and 

 even of the older sedimentaries, and that the force that gave rise to 

 the heat was crustal movements producing friction, thus reviving 

 the theory of Mr. Mallet, which, though universally held to be a 

 brilliant one, has been rejected on grounds which do not appear to be 

 sufficient. 



The question of oceanic islands is one of paramount importance 

 in the understanding of the form and figure of the globe, apart from 

 its purely geological interest. For it has been found in the measure- 

 ment of the force of gravity at various places by means of the 

 pendulum, that the value calculated by using any one of the estimates 

 of the true form of the globe is too great in oceanic islands and too 

 little on the coasts of the continents ; that is to say, the level of the 

 sea at places in mid-ocean is too low, and on the coasts too high, if 

 the curvature of the surface of the sea corresponded to its mathe- 

 matical value. The waters were said to heap up round the con- 

 tinents and sink in a basin-form in the middle. This was explained 

 by the attraction of the mass of the continents on the water. * 

 Fisher, f however, has shown that there is another possible explana- 

 tion, namely, that the average density of the rocks forming the 

 continent might be less than that forming oceanic islands, and, 

 indeed, the floor of the ocean as well. For supposing the theory of 

 Isostacy to be true, then the continents projecting so high above the 

 sea would have to have a very large base of material of low specific 

 gravity dipping into the rock-mass of the earth's crust in order to 

 float them, on the. principle of an iceberg floating in water. As the 

 average specific gravity of continental types of rock is 2*68, and that 

 of the basalts and andesites, of which oceanic islands were once 

 universally thought to consist of, is 2*96, the attraction of so great 

 a column of heavy material would affect the pendulum sufficiently 

 to let it swing the extra two or three times in the day, when placed 

 on an island in mid-ocean where the heavier rock was supposed to 

 extend downwards to the region of perpetual fusion. 



The presence of granite in islands so characteristically oceanic as 

 the Tristan d'Acunha group is subversive of all that we have so 

 confidently asserted in the past. Oceanic islands are purely volcanic, 

 say the text-books, and they then give a list of exceptions. First 

 and foremost are the Seychelles with a great variety of rock types 



* Suess, " Autlitz der Erde," vol. i. p. 3. 

 f "Physics of the Earth's Crusts." 



