ST. VINCENT IN 1902, AND ON A VISIT TO MONTAGNE CELEB. 
285 
remains higher than B and C. This gap, D, leads down in the direction of Larikai 
and Morne Ronde. It was lowered considerably during the eruptions of the autumn 
of 1902, and this accounts for the greater deposits in the Larikai valley in the 
later eruptions. The Somma Ring (Plate 18) is seen to consist of beds of lava 
and tuff dipping outwards from the crater, conformably with the outer slope of 
the mountain. The whole of the interior of the crater is still quite bare and without 
any trace of returning vegetation. A few small patches of moss appear on and 
about the rim, and in somewhat greater abundance on the slopes outside. This is 
worth notice, for Mr. James Anderson, in 1784 (Part I., p. 461), found moss 
covering the inside of the crater and in great abundance on the cone at the confines 
of the grassy region and the barren, so that the conditions are presumably becoming 
similar to those which existed at that time, and the vegetation will eventually be 
as it was before the eruption of 1902. 
The Carib country was not visited*, but as far as could be judged from a distant 
view from near the top of the Soufriere, vegetation was returning in a manner 
similar to that in the Wallibu district, and Mr. Sands, who has since visited the 
district, assures me that this is the case. He states that there is water in the upper 
reaches of the Rabaka River, though it all sinks into the ground lower down, so that 
none reaches the sea. The old bed of the river has become blocked about a mile from 
the sea and a new course has been formed to the north of it. The restoration of 
the water supply to the district, by the repair of the old conduit, is under 
consideration. 
The difference between the character of the eruptions of the Soufriere and 
Montagne Pelee, mentioned in Part I., p. 533, appears to have continued since 1902, 
the outbursts from the former volcano being generally less frequent but more violent 
than those from the latter. 
The Return of Vegetation. 
Confining our attention for the present to the Wallibu district, it may be stated 
generally that the whole of the country north of Morne Garu was devastated in 
different degrees. The limit of the zone of devastation extended right up to the 
summit of Morne Garu, and the line followed the main ridge in a westerly direction 
almost to the sea, before reaching which it diverged so that the Richmond Planta¬ 
tion Works were included, but the south bank of the river was not materially 
injured (Plate 19). Within this area the bottoms of the valleys, which were covered 
by the incandescent avalanche, had their vegetation utterly destroyed. In other 
places, where the hot ash was only deposited in a thin layer, the roots were in many 
cases not killed outright and are now throwing up new shoots and leaves, though 
the large trees are almost universally killed, except in a few sheltered situations. 
* Arrangements had been made to return by what was then advertised as the last voyage of the Inter¬ 
colonial Service of the Royal Mail Company. 
