SOUFRIERE, AND ON A VISIT TO MONTAGNE PELEE, IN 1902. 
407 
yellowish in that quartei’, and this increased to a darker and more coppery colour, till at 4 p.m. it Iiegan 
to assume the appearance of a London fog. The western (leeward) side of the smoke-cloud remained 
vertical and fixed exactly in one position, and it could easily be seen to ascend. A good deal of ordinary 
cloud passed at this time, so the eastern edge was ill-defined, but we could still see the western edge 
distinctly. This pillar of smoke rose to 12° above the horizon (artificial). Its breadth was about 15° 
of arc. 
“ About 6 P.M. we began to see the Hashes of lightning play in the dense cloud. They were at the 
rate of about 3 per minute. They were not like ordinary lightning, but nnrch shorter, and seemed to 
travel slower. In addition to the “ lightning ” there were great flashes of a redder colour (the 
“ lightnings ” appeared a yellowish-white). The flashes were at intervals of about 5 minutes. The 
display was much more violent than that of Martinique on the 5th instant; the smoke-cloud was at least 
double the height, and I concluded that the eruption had been on a vastly greater scale. No detonations 
were heard, but an earthquake was felt at about 2.45 p.m. by nearly everyone I have sj^oken to on the 
Morne, but I did not notice it myself. Next day, so much cloud had come up and rain that we could 
not see the smoke-cloud; during this time the display of lightning and flashing diminished to, say, 
one-fifth of its intensity.” 
Further particulars supplied at a later date by Major Hoddeii and Ijy Mr. CtErald 
Dev^aux, of the Cul de Sac Factory, St. Lucia, enable us to amplify the aljove account 
ill some important respects. 
Detonations were distinctly though faintly heard on the 7th May coming from 
St. Vincent. They were loudest about 2 p.m., and lasted from about half-past one to 
five o’clock, and resembled distant thunder, or the sound of the discharge of big guns. 
There seems toj have been no considerable fall of ashes and no darkness, hut 
only a slight haze in the atmosphere. The amount of dust which fell was so slight 
as to form only a film, the thickness of which was too small to he measured. The 
ash was dry, and was best seen on the surfaces of the leaves, where, it formed a fine 
impalpable powder. 
Mr. Devaux adds, also, that “ at the time the detonations were most distinct, the 
sea (at 3 o’clock) receded from the beach five times in half an hour, about 25 feet 
each time.” 
Barbados. —As in the eruption of 1812, the island of Barbados, which lies almost 
ICO miles to windward of St. Vincent, received more of the dust emitted by tlie 
volcano than any of the neighbouring Windward Islands. In the month of May 
tliere appears to be a strong and })ersistent upper current of air, flowing in an east or 
north-easterly direction over the Caribljees, the return trade or ipjper anti-trade 
wind. The depth of the trade-wind stream at tliat time and place is not very 
accurately known : none of the mountains of the Windward Islands is sufficiently 
high to overtop it. It must be at least 10,000 feet deep, as tire steam pulls 
Avhich rose from Montague Pelee, in ordinary circumstances, were in\mriably carried 
bodily to leeward. But on the 11th July we were in Fort de Fi'ance, and saw there 
the tongue-shaped rolling steam cloud which accompanies the more important 
eruptions, and this rose to such a height that it was entirely in the upper anti-trade 
current, and floated away to the east or north-east. This is sufficient to show that in 
