66 PERCY QUENSEL • 



berg's information, boulders of this character were ver\- numerous in the talus 

 under a steep escarpment and can here hardl_\- have originated from higher eleva- 

 tions than about 400 m. 



In outward appearance this rock shows an aspect different to the compact 

 soda-trach}'te between i 000 and i 100 m. which on account of its dense texture 

 at first sight gives the impression of a quartzitic rock. The new specimen has 

 on the other hand a rough, loose-grained structure. Minute intergranular cavities 

 in the rock are seen under the microscope to be filled with limonitic matter, l^ut 

 allowing for this structural dissimilarity the rocks from the talus below Tierras 

 Blancas seem to be closely related to the soda-trachyte as described above (Fig. 21). 

 Taking into consideration the seemingly horizontal position of the la\a beds of 

 Masafuera, the conclusion must be that lavas of soda-trach\'tic composition have 

 been emitted at different times and that they have alternated with lavas of more 

 normal basaltic composition (see p. 71). 



Specimens of the rocks from between i 100 and i 400 m present again a 

 different aspect. Five samples from i 100, i 200, i 300 and i 420 m all have 

 very much the same appearance. The\' are all phyric lavas of an ash-grey colour. 

 Feldspar and, more rarely, olivine occur as phenocrysts in a groundmass of 

 basic plagioclase, augite and abundant magnetite. In a somewhat vesicular specimen 

 from I 100 m olivine is also present as sporadic grains in the groundmass. 



The groundmass has in most of the specimens a trachytic texture; the narrow 

 laths of feldspar circuiting the phenocr\'sts in a more or less well-defined 

 manner (Fig. 22). 



In my previous paper I named the rocks of this nature trachyandesite. 

 Hagerman says, referring to much the same specimens in Skottsberg's collection: 

 "Die Klassifizierung hiehergehoriger Gesteine ist etwas unsicher. Genijgende 

 Griinde sie als Trachyandesite zu bezeichnen, liegcn jedoch nicht vor" (13, p. 30). 

 Hagerman names the rock andesite in his jjaper. 



These rocks evidentl}' ha\e a rather wide distribution on Masafuera, probably 

 forming the whole complex of lava beds between I 100 and I 420 m. To certify 

 their petrographic position an analysis has now been made of a typical specimen 

 (Table I, Xo. 5). The calculated norm and Xiggli values definitely classify the 

 rock as a feldspar basalt, in many respects of similar composition as some of 

 the analysed basalts from Masatierra (Table I, Xo. i — 4). A difference of interest 

 is the still lower content of KgO; the Xiggli value /' is now 0.08 against O.15 — 0.20 

 in the basalts from Masatierra. 



These phyric feldspar basalts from Masafuera in general aspect also very 

 much resemble the aphyric feldspar basalts from between 400 — 600 m elevation 

 on Masatierra. Occasionalh' at lower levels they also on Masafuera may be aph\Tic 

 in texture (Fig. 23). They seem also in several respects to correspond to certain 

 phyric and aphyric basalts from the Hawaiian Islands, as described b\' W.ASIIINGTOX 

 (Petrology of the Hawaiian Islands. I. Kohala and Mauna Kea. Am. J. of Sc. V, 

 1923, p. 487). To this question I will return under a concluding heading dealing 

 with the chemical and petrological connections between the rock of the Juan 

 Fernandez Islands and other volcanic islands of the eastern Pacific. 



