REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 89 



most of them are grouped round a perfectly colourless prismatic crystal of plagioclase, 

 to which they are attached. Many crystals of magnetite and ilmenite are to be seen, 

 and plagioclase is more abundant than in the vitreous zone, although the crystalline 

 form is embryonic ; augite occurs in rosettes of little crystals. The shining black part 

 of the vein is perfectly compact, but the internal portion is slightly vesicular, the 

 pores being lined with a transparent coating of green secondary matter which also 

 penetrates the microscopic fissures of the rock. The whole mass of the dyke must have 

 been cooled rapidly. Olivine is scarcely to be found in this rock. 



All the rocks from Inaccessible dealt with so far conform more or less strictly to 

 the basaltic type. A rounded pebble picked up on the shore is a bronzite and biotite 

 andesite. This specimen shows that eruptive masses different in composition from those 

 of the coast must exist in the interior of the island. The appearance of the pebble shows 

 at once that it differs from the ordinary rocks such as those described above. It is 

 much lighter in colour, being whitish grey. The texture is fine-grained, the fracture 

 nearly plane, and no constituent minerals appear to the naked eye. Under the micro- 

 scope a colourless ground-mass is seen, formed chiefly of curved and twisted crystals of 

 plagioclase of indefinite outline, and all matted together. Mixed with these there are 

 some violet-coloured augite microliths, with irregular outlines, but evidently of the 

 same stage of consolidation. Some scales of biotite also appear. All these minerals 

 are of approximately uniform size, and have crystallised simultaneously. Small 

 yellowish crystals appear in the paste as isolated short prisms, with flattened summits, 

 and worn on the angles. Sometimes these occur as irregular grains with fractures, but 

 they are too minute to permit their forms to be definitely ascertained. These small 

 sections give straight extinction, and so far as they could be examined by convergent 

 light it has been proved that the plane of the optical axes is parallel to the brachy- 

 pinacoid. These crystals ought to be considered as bronzite, and the rock as a bronzite 

 andesite. 



C. — Rocks of Nightingale Island. 



Nightingale is the smallest island of the Tristan da Cunha group, lying towards 

 the south. It is surrounded by rocks, amongst which are two islets measuring 

 one-half by one-sixth of a mile. One of these, Middle Island, 150 feet high, with an 

 undulating summit, is situated in lat. 37° 25' 50" S., and long. 12° 29' 45" W. The 

 second islet, which also lies to the north of Nightingale, is Stoltenkoff Island, and has 

 a height of 325 feet. Nightingale Island is a mile long from east to west, and about 

 three-quarters of a mile broad. 1 A channel ten miles wide, and over 465 fathoms deep, 



1 For the natural history of tliia little group, see Thomson, The Atlantic, vol. i. p. 185 (with a map) ; Moseley, 

 Notes of a Naturalist on the Challenger, p. 126 ; Karr. Chall. Exp., vol. i., pp. 262 et seq. For its geology, see 

 Buchanan, Proc. Roy. Soc, vol. xxiv. pp. 614, 615. 



(PHYS. CHEM. CHALL. EXP. — PART VII. — 1889.) 12 



