338 voLCAisric eruptions on Martinique and st. vincent. 



rigid to acquire the typical shape frequentl}' to be seen in large 

 numbers of bombs about certain basaltic craters. 



The absence of characteristic bom))s on Martinique and St. V^incent 

 is in keeping with the composition of the lava thrown out. The fresh 

 lava is an andesite, having in a general way the composition of a 

 refractory brick, and unless very highly heated woukl not be plastic. 

 The dark color of the columns of steam rolling up from the craters 

 when in violent eruption, and the vast quantities of fragmentai 

 material showered on the adjacent land and sea is evidence that as 

 molten rock was forced up the volcanic conduits it became cooled and 

 stiffened before reaching the summits of the volcanoes and was shat- 

 tered by steam explosions, and the fragments blown into the air. Not 

 only are true volcanic bombs absent, but clots and splashes of plastic 

 or fluid rocks, such as are common about many volcanoes that have 

 erupted easih' fusible material, are also lacking. 



The fragments ejected were in many instances blown to a height of 

 many thousands of feet, the finer lapilli and dust being carried perhaps 

 5 or () miles high, and on falling were distri])uted in part through tlie 

 influence of the winds, in a general way in reference to size and weight. 

 The larger and heavier masses fell near the craters from which they 

 were projected, while much of the finer and lighter material was car- 

 ried great distances. Variations in the method of distribution were 

 caused by the direction of the hurricane-like blasts which swept down 

 from bothMont Pelee and La Souf riere during their mightier eruptions, 

 by the direction of the trade winds and upper-air currents, and by 

 tornado-like swirls in the greatly disturbed atmosphere. Immediately 

 following each great eruption, also, there was a strong indraft of cool 

 air from about the region covered with hot debris. The vastness of the 

 area on which the ejected material fell is indicated l)v the fall of 

 dust on Barbados, Trinidad, and on ships 275 miles southeast of St. 

 Vincent. 



Observations ri'|)()rtod by E. O. Honc}^ show that, contrary to earlier 

 accounts, written in part by myself, coarse material fell in St. Pierre. 

 The riddling of boiler plates one-fourth inch thick (PI. VI), in the 

 northern portion of the stricken city, by stone shot against them from 

 Mont Pelee, is evidence that the hurricanes of steam charged with hot 

 dust, which swept down from that volcano on May S or May 20, and 

 perhaps during later eruptions, were acconjpanied by a bombardment 

 of stones, no doubt hot, which were as di^adly as solid shot tired from 

 a cannon. 



Causes of dcat/i. —Respecting the gencM'al cause of death in St. 

 Pierre, the reports of various observers ditter more widely than in 

 connection with any other occurrence associated with the eruption of 

 Mont Pelee, unless it be in reference to the secondary craters referred 

 to above. Obviouslv manv deaths occuri-ed in St. Pierre from the 



