VI. 



PRESIDENT S AUDKESS. 



the Andesites is very common. They contain a Plagioclase 

 Felspar, Hornblende, Augite,and frequently a rhombic Pyroxene, 

 which from its dichroism is probably Ilypersthene, although it 

 shows none of the lamellar and acicular inclosures which aie 

 so characteristic of this mineral as found in larger masses and 

 which produce the peculiar metallic glance observable in it. 

 The dust ejected last August is exactly such as would result 

 from the, so to speak, internal explosion of a fluid mass con- 

 taining a few crystals of Felspar and of Pyroxenes, but the 

 sudden cooling of which has prcveiited general crystallization. 

 The quantity thus thrown out was prodigious and the depth to 

 which the country for many miles around was covered was very 

 great. At a distance of thirty miles it was constant employ- 

 ment for one man to keep the glass window of the compass-box 

 of a ship washed, so that the steersman might be able to keep 

 the course, and a specimen which I received through Mr. W. B. 

 Cooley, of Wolverhampton, is described as having been swept 

 up from the deck of a ship "500 miles out at sea from 

 Sumatra," and it is added that "the air was darkened and the sea 

 covered with floating pumice stone."* 



* Since this was in print I have had the opportunity of examining 

 and analysing another specimen of the ashes and one of the floating 

 pumice collected on the ship " Berhice," which was within forty miles 

 of Krakatoa at the time of the eruption, kindly sent to the Society by 

 Captain Ross, whose steam yacht carried the dredgitg parties at Oban 

 last year. The slight difference in composition between the ashes and 

 the pnmice is noticeable. Both samples were dried at about steam heat, 

 but the pumice had previously been washed, as the salt contained in it 

 made it somewhat sticky. 



Silica 



Alumina 



Peroxide of Iron . . 



Protoxide of Iron . . 



Lime 



Magnesia . . 



Oxide of Manganese 



Soda.. 



Potash 



Loss on Ignition . . 



