REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 137 



high proportion of soda may be explained by supposing the sanidine to contain that 

 alkali ; 1 possibly also there may be amongst the microliths of the ground-mass little 

 crystals of plagioclase, the determination of which is impossible on account of their 

 small size and their confused arrangement. 



Let us now consider what are the stratigraphical relations between these phonolitic 

 masses and the surrounding basalt. According to Mr. Buchanan, no derangement of 

 the beds was found in any case at the contact of the two rocks. He was able to follow 

 the line of contact easily to the highest mass and procure specimens of it. The basalt 

 is much modified for some feet from the line of junction, the large crystals of augite and 

 olivine disappearing near the contact with the phonolite. The line of contact is 

 generally sharp, and many fragments of phonolite are seen enclosed in the immediately 

 bordering basalt, which is very fine grained. The grain grows gradually coarser, until 

 at a distance of 10 feet from the phonolite it reassumes the porphyritic structure which 

 this rock shows in other parts of the island. These two facts seem to show that the 

 phonolite rocks are the most ancient, and that the basalt has been poured out all 

 round them. There is no evidence, on the other hand, that the phonolite has been 

 erupted through the basalt mass. 



We shall now describe some rocks from " Foul House Bay," and as this name is 

 not on the chart we cannot follow geographical order in this case. They are coarser 

 grained than the other specimens from the island, dark coloured, with a blackish 

 tinge, and broken surfaces are shining and show a crystalline saccharoid texture. 

 Macroscopical greenish yellow granules of olivine, augite, and plagioclase are seen 

 in it. These rocks present obvious resemblances to certain peridotic diabases or 

 coarse-grained dolerites ; their microscopic characters also show the structure and com- 

 position of these lithological types. There is no distinct ground-mass, the crystals 

 being entangled. Sections of plagioclase show that this mineral is elongated following 

 the edge PjM, as is usual in the felspars of diabase and dolerite. This plagioclase 

 shows extinctions of 44°, and is thus probably anorthite. Olivine occurs in large 

 sections, rarely with crystallographic outlines, and is sometimes twinned, the two 

 individuals seeming to be united parallel to a pinacoid. This mineral is altered, 

 as is shown by the fissures being lined with an opaque black matter, and the sections 

 penetrated by delessite ; no serpentinisation, properly so called, is apparent. Delessite 

 is largely developed in other parts of the rock in question. Large, reddish, zonary 

 patches of augite fill the space between the other minerals. Magnetite or titaniferous 

 iron is very common. Besides delessite, some grains of calcite occur as products of 

 secondary formation. Another specimen, more decomposed, shows the same structure 

 and composition, except that olivine has almost entirely disappeared, its place being 



1 See Roth, Chem. Geol., vol. ii. p. 240. 



(PHTS. CHEM. CHALL. EXP. — PART VII. — 1889.) 18 



