The Rocks of Tristan cVAcunha. 15 



clearly stated, but they appear to be basaltic and andesitic rocks, 

 such as we find in the Drakensberg tuff necks. In the South 

 African Museum specimens there are fragments containing sphene. 



Middle Island is also composed of a tufaceous mass, and these 

 two islands, therefore, must be considered to be two great tuff necks, 

 such as are found along the Firth of Forth in Scotland, in the Eiffel 

 and Swabia in Germany, in Griqualand East and Griqualand West 

 in Cape Colony, and at Bingara in Australia ; but the extraordinary 

 fact of such a plug projecting out of the sea from deep water has 

 never been properly recognised. In the following pages we shall 

 have a good deal to say about such agglomerate necks. 



An examination of the present specimens can add to the above 

 description the occurrence of gneiss in Tristan d'Acunha, and the 

 inclusion of fragments of a sphene rock, possibly an andesite, in the 

 rocks from Nightingale Island, the latter thus showing some simil- 

 arity with the volcanic series of the Falkland Islands. Otherwise 

 our collection — though much more restricted in number of specimens 

 to the Challenger one — contains, with the exception of nepheline 

 rocks, very much the same types. There are, however, many points 

 exhibited in our specimens, such as the presence of sphene rocks, 

 which urgently call for a proper investigation of this interesting 

 group of volcanoes. 



The gneiss block was picked up by Mr. Hammond Tooke near the 

 settlement of Edinburgh. The finding of such a rock naturally excited 

 our interest, and there was expressed considerable doubt as to 

 whether the block had not been thrown out from a ship in ballast, 

 or had floated thither by icebergs from the far south in ancient 

 times, when possibly the island might have been lower, and the 

 ice floated much further north than it now does. In the Nightingale 

 rocks, however, there are fragments of a porphyritic rock imbedded in 

 the lavas ; these are minute fragments that have evidently been 

 derived from the shattering of a foreign igneous mass of an acid 

 type by explosions, such as occur in the throats of volcanoes, and 

 thus cannot have been drifted to the island. While the evidence is 

 not as good as one could wish for, and would not be admissible were 

 the islands more easy of access, or had a geologist been to the place 

 himself, nevertheless, it is reasonable to suppose that the granite 

 belongs to the islands, and has been actually spouted out of the throat 

 of the volcanoes in the same way as the granite and gneiss in the 

 Island of Ascension, a similar volcanic island in mid-ocean. 



A petrological description of the rocks follows at the end of the 

 paper. I propose first to discuss the bearing of these new finds 

 of rocks of a continental type on islands separated by so many 



