THE ERUPTIONS OF MAY, 1902, AT THE SOUFRIERE IN ST. VINCENT. 
313 
litic hypersthene andesites very similar to those above described. Probably they 
were washed out of the ash beds deposited by the eruptions of 1812 and 1718, or 
even at an earlier date. 
Blocks of Older Andesites. 
These were easily distinguished from the neAv bombs by their angular outlines (due 
to fracture), their less vesicular character, and the traces of weathering or fumarole 
action they often exhibit. Many of them were 2 feet in diameter, while some were 
5 feet long and 4 feet thick. Of these rocks the greater part were hyalopilitic 
hypersthene andesites with accessory olivine, and differ little from the new bombs 
of 1902, except in their more solid texture and the general absence of steam cavities. 
Their groundmass is more crystalline, as a rule, than that of the bombs, owing to 
their having cooled more slowly, but there is considerable variation in this respect, as 
some contain much brown glass, while others show very little of it. Many of them are 
very fresh, and must have been derived from the solidified magma of previous eruptions 
which occupied the passages leading to the base of the crater, or formed its floor. 
When the vitreous base is abundant, these rocks have a dark lustrous appearance ; 
when the groundmass is more crystalline they are more grey in colour. All carry 
phenocrysts of plagioclase, augite, and hypersthene ; olivine appears in most of them, 
and is often surrounded by hypersthene; hornblende is rare. The olivine may 
contain small dark brown octahedra of chromite or picotite. The hypersthene, when 
the groundmass is well crystallised, may have narrow borders of augite due to 
corrosion during the last stages of solidification. 
A few of these rocks contain as much olivine as hypersthene, and may be classified 
as hypersthene basalts (Plate 27, fig. 5). The great resemblance which these old 
andesites and old bombs present to the new products of the magma of 1902 proves that 
there has been little change in the nature of the materials emitted by the Soufriere 
for a very considerable period. 
An interesting structural variation presented by some of these andesites is flow- 
brecciation. In a few of them, although the rocks appear perfectly massive in the 
band specimens, the porphyritic crystals are broken to fragments (Plate 26, fig. 8). 
This is especially noticeable in the zoned felspars, as the handing does not then 
continue round the whole crystal, but stops abruptly at the broken edges. These 
rocks have a rather crystalline matrix, never a glassy one, and are not vesicular. 
They show no flow banding, but in the groundmass there are patches of different 
structure which look almost like enclosed fragments. It is clear that the brecciation 
took place before the rock was entirely solid, as later deposits have gathered on the 
broken faces of some of the larger felspars, forming thin continuous marginal zones. 
At the same time there can be no doubt that the matrix was already partly solid, as 
the brecciation affects it also. On Montagne Pelee^ brecciated rocks are common, and 
vol. ccvur.—A. 
* ‘La Montague Pelee,’ pp. 513, 514 (1904). 
2 S 
