REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 59 



andesites ; but the pyroxenic mineral is bronzite. Plagioclastic microliths and little 

 reddish crystals of bronzite make up the ground-mass, and a good number of rather 

 large crystals of felspar also appear in it. At first sight these seem to be sanidine, 

 as they have the glassy appearance and the lines of fracture which one is accustomed 

 to consider as characteristic of this felspar ; but the homogeneity disappears with 

 polarised light, and the crystals are seen to be striated like plagioclase by the inter- 

 calation of a very large number of polysynthetic lamellae. Sometimes this felspar is 

 crystallised simultaneously according to the albite and Carlsbad laws. In certain cases 

 some individuals show a zonary structure. These extremely close striae recall similar 

 observations in sections of oligoclase and andesine, and this resemblance is confirmed 

 by the fact that the extinction in the felspar of this rock takes place at a very low 

 angle. 



The mineral identified as bronzite is always altered, and the decomposition shows 

 itself by the deep red tint which clothes the sections. Sometimes crystals cut 

 perpendicular to the prism, show an octagonal form like that of augite sections. This 

 form is, however, equally characteristic of bronzite, to which the optical properties in 

 parallel light plainly refer the crystals, but their small size and the alteration of the 

 mineral makes an examination by convergent light impracticable. These prismatic 

 sections always extinguish following the length, and never show pleochroism. The 

 alteration of this mineral not only changed the colour, but in some sections part of the 

 substance has been eliminated, and greenish matter deposited in the hollows. The red 

 colour produced by alteration makes these little prisms resemble certain olivines, but 

 the outlines of the sections and the elongated form of the prism do not confirm this 

 supposition. This bronzite is rarely found in sufficiently large crystals to induce micro- 

 porphyritic structure, but occasionally some are of such a size, and in this case they are 

 often deeply indented. A very pronounced fluidal structure appears round the larger 

 crystals of bronzite. The mineral may be traced from the large sections, on which its 

 determination is based, to extremely small microliths in the ground-mass. It is by 

 analogy also that the minute crystals of plagioclase in the ground-mass are related to 

 the larger individuals of the same species, the microliths being sometimes so minute 

 that the polysynthetic lamellae can hardly be discerned. Finally, we majr mention 

 amongst the constituent minerals of this andesite large and irregular sections of 

 magnetic iron, which usually appear as skeleton crystals. 



To andesite must be referred also the rock forming veins in the trachyte of the hill 

 known as " Crater of an old volcano." Darwin * thus describes the very numerous veins 

 in the earthy trachyte exposed on the sides of this mountain. The rock forming them 

 contains crystals of glassy felspar, some black microscopic grains, and small stains 

 of a dark tint. The ground-mass is very hard and compact, and the rock is more 



1 Darwin, Geol. Obs., pp. 44-45. 



