St. Vincent^ and Mf. Pelee, Martinique. 345 



plateau also was sending skyward great colomns of steam, and 

 the whole formed a scene seldom witnessed, difficult to describe, 

 and never to be forgotten. The next daj we measured the 

 gorge and found that the Seche had deepened its channel at 

 least ten feet in the loosely compacted recent ash during the 

 hour which the flood lasted. 



In this instance it seems evident that there was close con- 

 nection between the heavy rain, the eruption, and the black 

 torrent. Two explanations present themselves for considera- 

 tion : {a) the crater may have thrown a mass of accumulated 

 rainwater and ashes bodily over into the head canyon of the 

 Seche; (J) the rain which fell into the crater may have been 

 the exciting cause of the eruption, but the mud-torrent may 

 have been due to the soaking of the heavy coat of ashes on the 

 steep outer slope of the old cone at the head of the Seche until 

 the resulting fluid mass slid off from the comparatively hard 

 surface beneath and poured down the gorge of the river. 

 There was plenty of water-soaked mud and ashes on the upper 

 part of the mountain to supply the avalanche and some of it 

 was on the ^qy^q of fluidity at the time of our visit, hence the 

 latter explanation seems the more reasonable. The little tribu- 

 tary of the Seche which empties into it on the southeast side 

 close to our point of observation, did not show any correspond- 

 ing torrent, because it does not head on the side of the great 

 cone. 



The mud-flows which have descended the Precheur River 

 canyon have had ample collecting ground in the '' Atrio del 

 caballo," to use a Vesuvian term, on the north and northeast 

 sides of the great crater, where the fine dust settles in vast 

 quantities ready, when sufficient water has been added to it, to 

 descend through a narrow gorge into the valley of the Precheur. 

 When I was walkino^ alonor the crater rim above the " Atrio" 

 June 20, my footsteps started small mud-flows down the outer 

 cone, so liquid was the mud at that time. The ordinary action 

 of the volcano is to deposit dust of impalpable fineness on the 

 inner face of the crater rim. When this deposit becomes thick, 

 it is ready to be swept off by a copious rain and carried through 

 the great southwestern gash, out of the crater and down the 

 gorge of the Piviere Blanche as a mud-torrent or flow. There 

 does not seem to the writer to be any need of locating '' mud 

 craters" at the heads of or along the line of the gorges which 

 have been the courses which these torrents of liquid mud have 

 followed to the sea. 



Where the tufl agglomerate of the old (outer) cone had been 

 freed from its coat of ashes, especially in its lower portion, 

 i. e., from 1000 to 2000 feet above tide, it showed a smooth, 

 somewhat fluted surface, the soft bowlders having been planed 



