470 
DKS. T. ANDERSON AND J. S. FLETT ON THE ERUPTIONS OF THE 
frf)m our sigHt ])y the intervening ridges of Morne Ronde. Vast globidar Ijodies of 
fire were seen projected from tlie fiery furnace, and bursting fell back into it, or over 
it on tlie sui'rounding l^ushes, which were instantly set in flames. About four hours 
from the lava boiling over the crater it reached the sea, as we could observe from the 
reflection of the fire and the electric flashes attending it.” 
Now this looks very circumstantial, l)\it the fact remains that when we were at the 
mouth of the Larikai this summer we saw no lava in the bed of the .stream there. 
This torrent flo’ws down a tremendous gorge, cut in the old lavas and tuffs of this 
part of the mountain. On the shore below there had been a small alluvial fan, but, 
as already mentioned (p. 454), it had slid away and disappeared during the eruption 
of tins year. In the mouth of the gorge there was a low cliff, some 50 feet high or 
more, capped by a few feet of the new, hot sand. The cliff consisted of a material so 
exactly similar to the ash that overlay it that, if it had not been for the old burnt 
soil, it would have been hardly possible to find a line of demarcation between the two 
deposits. We were at once struck with the similarity, and came to the conclusion 
tliat here was the evidence of a dust avalanche eruption previous to that of this year. 
We could not land, as the bare cliff of volcanic sand was tumbling in frequent 
landslips into the sea, but we rowed quite close in, and could make out all the more 
important features of the deposit. It was almost unstratified, and had a few stones 
scattered through it, hut for the most })art was a yellow or brownish incoherent 
sand. It could not possibly have been an alluvial deposit, as it showed so little 
stratification, and such a torrent as the Larikai could not he siqDposed to lay down 
50 feet of sand in any part of the precipitous ravine. Further iq3 we could see 
where it had been rollino; down enormous boidders durino- the recent heavv rains. 
This was the last deposit in the Laril^ai, and it bore all the marks of a dust avalanche. 
It lay in the old eroded gorge behind a bend, whe’e a ])rojecting mass of rock had 
})rotected it from erosion. 
On the north-west corner of the Soufiiere lay the estate of Diivallie’s 
(De Volet’s), also known as Windsor Forest."^ This was at that time a sugar 
estate, hut is now devoted to grazing and cocoa. It is stated to have been 
“entirely covered with the matter thi'own out by the volcano; the sugar works 
totally covered and not discernible ; nine negroes killed, the rest escaped over the 
mountains and came to town much cut and bruised.” t 
According to Shephard it was “ entirely covered with sand and ashes.” This 
may have been the work of a mud lava, but it is more probable that from the 
position of the new cratei' that its emissions were directed principally towards the 
north side of the hill, and that the black cloud which descended there was even 
more destructive than in the })resent year, when most of the valleys in this quarter 
* P. Foster Huggins, ‘ An Account of the Eruptions of the St. Vincent Soufriere, 1902,’ p. 7. 
t Blue Book: ‘ Correspondence relating to the Volcanic Eruptions in St. Vincent and Martinique in 
May, 1902,’ p. 95. 
