REPORT ON THE PETROLOGY OF OCEANIC ISLANDS. 145 



are bounded by traces of the faces PyT. That this is the case, is ascertained by the 

 measurement of the angles and the direction of the cleavages, the latter exhibiting a 

 line of cleavage parallel to P, and another, but less pronounced system, parallel to the 

 prism. The angle of intersection of the trace of P and of the trace of the adjacent 

 face is about 100°; this face is therefore y. The other side of the section makes an 

 angle of about 64° with the trace of P, and is therefore T. The extinction for the 

 section in question is negative, and takes place at 27°. The felspar accordingly 

 approaches to bytownite. These observations have been made on a great number of 

 sections of the rock, and each time the angular values were approximately the 

 same. The symmetrical extinction of the sections showing albitic twins was about 

 40° ; this value is another proof of the exactness of our determinations. The plagio- 

 clase has almost always crystallised according to the albite law, sometimes associated 

 with that of the Carlsbad type. In some cases, also, this plagioclase is twinned 

 according to the Baveno law. Thus two crystals of plagioclase, both twinned according 

 to the albite law, can be observed grouped in such a way that the traces of M in the 

 two individuals make an angle of about 90°. The extinction of the albitic striae is 

 the same for the two crystals, being about 40°, from which we may conclude that 

 the section has been cut for both of the adjacent individuals in the zone P : k, and the 

 fact that the extinction of the albitic lamellae is the same in both confirms the sup- 

 position that they have a plane of this zone in common. The angular value of the 

 extinction seems also to indicate that the section is approximately perpendicular to the 

 edge P/M. The facts we have just mentioned thus prove the existence of the Baveno 

 twin in some crystals of this rock, and that of pericline has also been demonstrated. 

 Many of the plagioclastic sections show a zonary structure, especially those cut parallel to 

 M, on which are observed a series of concentric zones, the inner ones being disym- 

 metric hexagons, the outer cpiadrangular, representing traces of the faces Py. Thus the 

 internal hexagonal zones show supplementary traces of T. At the beginning of then* 

 growth the plagioclases crystallised with the faces of the prism, which became smaller in 

 proportion as the crystals formed, and finally disappeared when the last layers were 

 deposited on the nucleus. This fact may be generalised and applied to all the plagio- 

 clases in the rock, as the prismatic faces are wanting in the greater number of crystals, 

 or, if they are present, they play a very subordinate part. 



In this basalt the felspar sections often show alterations due to the action of the 

 magma; the angles are rounded off, the crystals often corroded, and penetrated 

 by the vitreous mass in which they are embedded. AVe cannot, however, explain, by 

 subsequent modifications, certain optical phenomena resembling the undulating extinc- 

 tion. At first sight one is tempted to ascribe these to the result of strain exerted on 

 crystals already formed. But they are explained by the manner in which the hemitropic 

 lamellae are entangled. When observed in polarised light, a good many sections are 



(PHYS. CHEM. CHALL. EST. PAIiT VII. 1S89.) 19 



