284 
PROFESSOR T. G. BOXNEY OX A COLLECTIOX OF 
although we can often make a reasonable conjecture, we have no sure criteria for distin¬ 
guishing between specimens taken from a dyke and from a flow, if they happen to have 
solidified under approximately similar circumstances. Hence, seeing that while in the 
present extensive series from Socotra there are some which are certainly from intrusive 
masses and others which are almost as certainly from lava flows, there are several which 
I cannot separate, either lithologically or petrologically, I have thought it better to 
group them under one general title, and indicate in the description what seems the 
most probable history of the rock from which the specimen has been taken. The 
majority of these Socotra felstones are of a warm red or purplish colour, though some of 
a grey tint also occur. They vary from minutely crystalline (approaching on this side 
vein-granites) to very compact sub vitreous rocks which sometimes show a well marked 
fluidal structure. It will, perhaps, be more convenient to describe them as they occur 
geographically rather than to attempt to make a lithological arrangement. 
The first specimen [4049] is from a dyke in the Archaean series, not far from the 
coast of Gubbet Gollonsir, about 1^ mile east of Ras Bedu, at the western extremity 
of the northern coast of Socotra. It is a compact dull-coloured felsite, weathering a 
pale yellowish-brown, containing many little black crystals. Under the microscope 
the rock appears to have a glassy base, stained with ferrite so as to present a rather 
muddy aspect, in which are scattered many small and well-formed crystals, both of 
felspar and green hornblende; among the former orthoclase and a plagioclase, possibly 
albite, can be recognised ; the latter crystals are generally well formed but sometimes 
include portions of the ground mass. This, when examined with the two nicols, is 
to be seen full of minute microlitbs of felspar, but there appears to be some remains of 
a glassy base. There is no free quartz. The rock then appears to be intermediate 
between the hornblende andesites and the sanidine (or orthoclase) trachytes. From 
the same neighbourhood also comes a rather granular quartz-felsite containing some 
hornblende. Several specimens have been examined from an interesting series 
obtained on the plain near Kadhab village; unfortunately none of these occur in situ, 
but, as Professor Balfour informs me, they doubtless come from a part of the Haggler 
range lying to the south-east. [4206], a compact dark felsite with paler spots and 
wavy bands of a more crystalline material, containing small scattered* crystals of pink 
felspar, exhibits under the microscope a clear base (perhaps not wholly devitrified) 
studded with numberless minute granules and rods of opacite, indicating a fluidal 
structure. The spots and bands prove to be groups of spherulites crowded together 
with irregular interlocking edges.* There is some green hornblende, generally in 
clustered granules, possibly associated with minute tourmaline. The larger felspar 
crystals resemble orthoclase or sanidine, and there is a little free quartz. It is there¬ 
fore a rhyolite, but not one of the kind which, judging from this collection, is so com¬ 
mon in Socotra. Of this [4203, 4213, 4214] are types, “compact red felstones or 
* As in the figures of a devitrified glass: Daubrke, * Etudes Syntlietiques de Geologie Experimentale, 
l re Partie, pp. 170 and 171. 
