ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JIAN FERNANDEZ ISLANDS 63 



basalt must be attributed to an iinasioii of /.eolitic matter. As there seems to 

 be a slight ditierence in refractive indices in the alteration product, this is prob- 

 ably a miNture of different zeolites or related material. It has, however, not 

 been possible to determine the definite nature of the zeolites. The fact that they 

 show no si<Tn of birefringence indicates that they must bclonj^ to the isotropic 

 i^roup or to those of ver}' near zero birefringence. 



It is of interest to note that Renard evidently has observed a very similar 

 circumstance in a lava How from Masatierra. when he says that sections of a 

 basic feldspar from a greyish very scoriaceous lava under the microscope are 

 seen to be "cracked and pervaded with zeolitic matter, which forms an irregular 

 network. This matter, which looks slightly gre\-, when seen in ordinary light, 

 remains obscured between crossed nicols." This description might as well refer 

 to the zeolitisation of the lava bed from Masafuera, especially as Renard in the 

 same lava mentions olivine, unifornil}- changed to iddingsite (Renard sa}-s hematite, 

 8, p. 176). 



It therefore seems probable that lava beds, in which the content of olivine has 

 been more or less changed to deuteric iddingsite, also have succumbed to a high 

 degree of zeolitisation. The fact that the large phenocrysts of olivine, onl\- 

 rimmed with iddingsite, show no signs of alteration, whereas the feldspars in the 

 same section may be completely altered to an isotropic substance of zeolitic 

 composition, must indicate that also the zeolitisation of the rock has taken place 

 at an earl\- stage and cannot be attributed to periods of normal weathering or 

 later thermal activit}-. 



In consequence ot tb.e changes in chemical composition which must have 

 taken place in connection with the formation of iddingsite and the zeolitisation 

 of the lava, it is no longer possible to establish the primary composition of the 

 rock in question. Though the whole aspect of the lava bed so obviously differs 

 from the normal olivine basalts on Masafuera, there is evidently not reason enough 

 with any degree of certitude to classify the rock as an olivine basanitoid. The 

 presence of alkaline rocks on ]\Iasat\iera must for the present, therefore, be 

 restricted to the occurrence of the soda-trachytic lava flows, now to be described. 



At an elevation between i ooo and i loo m there occur rocks of quite another 

 type. In outward appearance the\- are dense, aphanitic and aphyric in texture 

 and of a light }-eIlowish green colour. The dominant minerals are an acid plagio- 

 clase and. subordinate, orthoclase. 01i\"ine, a light green diopsidic pyroxene, and 

 ore minerals are sparingly found in small individuals between the feldspar laths, 

 which have a well-detined trachytoid orientation (Fig. 20). An analysis of this 

 rock was made for nn- tormer description and is reproduced in the annexed table 

 of the analyses of the Juan Fernandez rocks (Table I, Xo. 8). I named the rock 

 soda-trachyte. 



On my survey I had assumed that rocks of this composition only existed 

 on and formed the highest parts of the island, as at that time no higher elevations 

 were reached. New specimens in Skottsberg's later collection disprove my earlier 

 supposition in two essential respects. In the tirst place, rocks of much the same 



