SOUFEIERE, AND ON A VISIT TO MONTAGNE PELEE, IN 1902. 
375 
that earlier cone in which the great Somma crater had been formed. It is possible 
that the great explosion which removed the summit of the old hill left a crater one 
lip of which was more than 1000 feet lower than the other, and on this lower lip the 
present craters have built up their later cone. Or it may be that, after that par¬ 
oxysm, the hill j^resented the appearance of a breached cone with a wide, open trough 
facing the south, and that the actual craters stand on the line of this fissure. 
The great valley which has already been described as lying on the south side of the 
Soufriere, between it and the Morne Garu, is deserving of special description, as it is 
here that the ejecta of the recent eruption have chiefly accumulated (see Plate 25, fig. 2). 
It runs across the island, and on the west is occupied by the Wallibu and Wallibu Dry 
Ptivers, on the east by the Pabaka Dry River, The dividing ridge is about 
2000 feet high, and is the lowest part of the axial chain of the island, so that it is 
quite jDOSsible to regard the Soufriere as a detached mountain not directly connected 
with the main ridge to the south. The rivers mentioned run in narrow ravines 
which have been incised in the floor of an open valley, so that it seems improbable 
that as they are at present they are the main factors in its formation. It seems 
rather that during some past epoch, when the jDhysical conditions were different from 
what they are now, this wide hollow was scooped out, and that subsequently the 
present rivers have cut depressions in its floor. 
It should be remembered that this is a structural dejDression sejjaratiiig the extinct 
and much eroded volcano of the Morne Garu from the still active and better preserved 
cone of the Soufriere. As such it has been a valley from the beginning, and is not 
entirely to be ascribed to erosion. This is shown by the dip of the rocks, which on 
both sides is towards the centre of the valley (see Plate 32, fig. I). It must have been 
from the earliest times one of the main drainage systems of the surface of St. 
Vincent. 
The streams flowing in this depression appear to have at some time or other reached 
a base-level, so that their cutting power was in abeyance and their valleys broadened 
and opened out laterally. This may have been during the subsidence, the climax of 
of which is marked by the 700 foot terrace. Subsequent elevation has enabled 
the streams to deepen their courses and to incise deep channels in the old valley 
bottom. 
Another factor which must have been of im23ortance in determining the formation 
of this trough is the absence of any hard lava streams in this part of the Soufriere—one 
of the most perplexing features in the structure of the mountain (see Plate 34, fig. 2). 
Lava flows are abundant on the north, the west, and the east sides of the volcano, but 
they are not found at the base of the hill on the south side, or in the part out of 
which these valleys have been eroded, though at higher elevations (above 1200 feet) 
they are fairly numerous and often seen in the stream sections. As these lavas always 
offer much greater resistance to the deepening of the river channels than do the tuffs 
and agglomerates, and generally stand out as cascades and waterfalls, it is easy to see 
