Eruptions on Martinique and St Vincent 431 



degree to which the ascending steam is 

 debris-charged. If the energy is great 

 and the upward propelling force essen- 

 tially constant, it may well be inferred 

 that the column, as explained by Jag- 

 gar, will attain a great height before 

 the resistance it offers to the ascent of 

 fresh material causes an expansion at 

 the base. If, however, the steam driven 

 out at any stage in an eruption is 

 excessively loaded with debris, an ex- 

 pansion and overflow at the rim of a 

 crater might occur, no matter whether 

 the fall of previously discharged mate- 

 rial from aloft had begun or not. The 

 essential feature in a down-blast from a 

 crater seems to be that heavily debris- 

 charged steam behaves in many ways 

 like a fluid and will flow down steep 

 gradients and acquire great velocity 

 when the slope and other features of the 

 surface over which it progresses are fa- 

 vorable. The gradients on the slopes of 

 Mont Pelee and La Soufriere, within the 

 zone of destruction in each case, are 

 about 1 ,000 feet to a mile, and, as seems 

 evident, the finally accepted explanation 

 as to the controlling condition which 

 gave direction to the blasts which swept 

 them will include the principle just 

 stated. In this connection it is in- 

 structive to note certain observations 

 made by Messrs Anderson and Flett, 

 commissioners sent by the Royal Soci- 

 ety of London to stud)' the recent erup- 

 tions. On the evening of July 9 these 

 gentlemen were on a vessel near Car- 

 bet and witnessed an eruption of Mont 

 Pelee. 



"As the darkness deepened, a dull- 

 red reflection was seen in the trade-wind 

 cloud which covered the mountain sum- 

 mit. This became brighter and brighter 

 and soon we saw red-hot stones pro- 

 jected from the crater, bowling down 

 the mountain slopes and giving off glow- 

 ing sparks. Suddenly the cloud was 

 brightly illuminated, and the sailors 

 cried, ' The mountain bursts ! ' In an 

 incredibly short space of time a red-hot 



avalanche swept down to the sea. We 

 could not see the summit owing to the 

 intervening veil of cloud ; but the As- 

 sure and the lower parts of the moun- 

 tain were clear, and the glowing cata- 

 ract poured over them right down to 

 the shore of the bay. It was dull red, 

 with a billowy surface, reminding one 

 of a snow avalanche. In it there were 

 large stones, which stood out as streaks 

 of bright red/ tumbling down and emit- 

 ting showers of sparks. In a few min- 

 utes it was over. A low, angry growl 

 had burst from the mountain when this 

 avalanche was launched from the cra- 

 ter." The time occupied by the ava- 

 lanche to reach the sea was " possibly 

 a couple of minutes. It could not have 

 been much more. " 



' ' There is no doubt that the eruption 

 we witnessed was a counterpart of that 

 which destroyed St Pierre. . . . The 

 most peculiar feature of these eruptions 

 is the avalanche of incandescent sand 

 and- the great black cloud which accom- 

 panies it. The preliminary stages of 

 the eruption, which may occupy a few 

 days or only a few hours, consist of out- 

 bursts of steam, fine dust, and stones, 

 and the discharge of the crater lakes or 

 torrents of water or of mud. In them 

 there is nothing unusual, but as soon as 

 the throat of the crater is reached , a m ass 

 of incandescent lava rises and rolls over 

 the lip of the crater in the form of an 

 avalanche of red-hot dust. It is a lava 

 blown to pieces by the expansion of the 

 gases it contains. It rushes down the 

 slopes of the hill, carrying with it a 

 terrific blast, which mows down every- 

 thing in its path. The mixture of dust 

 and gases behaves in many ways like a 

 fluid. The exact chemical composition 

 of these gases remains unsettled. They 

 apparently consist principally of steam 

 and sulphurous acid." 



The account just quoted of a typical 

 down-blast from a volcano, seen under 

 favorable conditions by trained observ- 

 ers, is perhaps the best evidence on rec- 



