TIIE ERUPTIONS OF MAY, 1902, AT THE SOUFRIERE IN ST. VINCENT. 
319 
perfect mosaic. If regarded as true plutonic rocks they may be classified as gabbros, 
troctolites, anorthosites, &c. (or seeing that their felspar is anorthite, as allivalites, 
harrisites, &c.). They have evidently formed under plutonic conditions. Hornblende 
and olivine are frequent in them, hut these minerals are unstable at the levels to which 
the magma ascends before effusion, and are gradually dissolved. Professor Lacroix 
has described and figured* fused crystals of hornblende in nodules of this sort from 
St. Vincent, but we have not met with any in our specimens, where these minerals are 
always perfectly clear and sound. The felspars of these rocks resemble those of the 
early central portions of the phenocrysts of the lavas, though somewhat more rich in 
lime and containing more enclosures. The absence of outer zones of bytownite and 
labradorite shows that these felspars ceased growing at great depths. All the minerals, 
however, are like those met with, under certain conditions, in the lavas; hence there 
can be no doubt that these crystalline rocks are from the same magma, but they have 
formed at great depths and are more basic in character than the rocks erupted at the 
surface. 
A very interesting structural feature of these rocks is that their minerals are not in 
perfect contact with one another, but are separated by the thinnest possible films of 
brown glass (Plate 27, fig. 1). The vitreous material even passes into the cracks and 
cleavages of the felspars. It seems to enter the original cavities and enlarge them by 
corroding the surrounding mineral. This glass contains usually only small dusty 
grains of black magnetite. Near the edges of the specimens the finest veins unite to 
form thicker ones, which sometimes contain microlites of andesine and labradorite ; 
the glass is consequently similar to that of the hyalopilitic andesite matrix. This 
makes it probable that the glass veins are an injection of that portion of the magma 
which was liquid at the time of the eruption; it explains also the absence of zones 
surrounding the anorthite and intermediate in composition between it and the andesine 
of the glass. The very friable character of the nodules arises from the presence of 
these films of brittle glass separating the crystals. 
The subjoined analyses show the composition of two crystalline blocks collected in 
St. Vincent by Professor Lacroix. The second indicates the great abundance of 
anorthite in some of them ; compared with the first, it proves'a considerable range in 
composition in these blocks, though both of them are highly basic. The other three 
analyses are of anorthite, hypersthene, and hornblende, in a similar block from 
St. Kitts. Dr. Fels’s paper contains also particulars of the crystallographic and 
optical properties of these minerals. Analysis III. is interesting as proving that the 
felspar is not absolutely pure anorthite, but corresponds to Ab 12|-, An 8 7^. The 
felspar of Professor Lacroix’s rock must be rather more basic than this. The 
St. Kitts hypersthene contains 18 per cent, of iron oxides; a sample of hypersthene 
separated from the rocks of Martinique contained 27‘7 per cent.! of iron oxides ; if we 
* Op. cit., Plate 26, fig. 6, and p. 542. 
f Lacroix, A., op. cit., p. 506. 
