332 VOLCANIC ERUPTIONS ON MARTINIQUE AND ST. VINCENT. 



sources. Gi'eat explosions have occurred, however, at each of the 

 localities referred to, which have thrown large quantities of dust and 

 mud to a height of several hundred feet, and sent out vast volumes of 

 steam to a height of many thousands of feet; but these phenomena 

 are seemingly the same, although marked by greater energy, as have 

 been observed in a large num))er of instances on both Martinique and 

 St. Vincent, where the hot dust, lapilli, stones, etc., ejected from the 

 summit crater in each instance, have accumulated to a great depth and 

 been invaded by surface water. While the subcraters mentioned 

 should not, in my opinion, be considered as true volcanoes, they sinui- 

 late many of the phenomena attending actual eruptions from deep 

 conduits. The columns of steam, heavily charged with dust and mud, 

 which arise from them have the convoluted or cauliiiower structure, 

 and at times expand at the top and take on nuishroom shapes in nuich 

 the same maimer as do the columns of steam similarly charged with 

 rock fragments that are blown into the air from a primary crater. 

 In each case the proximate cause is the same, namely, a steam explo- 

 sion. The solid material blown into the air in each instance is also of 

 the same natui'c for the reason that tbe hot dust and stones, to which 

 the superlicial explosions are due, were supplied by the eruption of the 

 summit crater. Since the o])served pheriomena are so similar, it may 

 be asked: What is the crucial test by which a true crater may ))e distin- 

 guished from a pseudocrater ? To formulate a definite answer to this 

 question is difficult. Perhaps the best reply that can be offered is that 

 the pscnidocraters are later in the time of their appearance than tiie 

 main eruptions which supplied the hot material necessary for their 

 production, and that they occur when the topographic conditions 

 previous to the eruptions favored the accunudation of a deep deposit 

 of hot debris. In addition, on ])oth Martinique and St. Vincent a com- 

 plete gradation in size and energy of the superticial steam explosions 

 has l)e(Mi observed, ranging from small ge3'ser-like spoutings (PI. I), 

 such as have occurred at hundreds of localities in valleys that were 

 deeply tilled with hot debris, and even on broad and comparatively 

 smooth surfaces covered with a thick sheet of similar material, up to 

 the markedly energetic explosions in the valleys of the Falaise and 

 Riviere Blanche. In many instances the smaller surface explosions 

 have been observed to follow heavy rains, and the same is true also of 

 the larger ex})losions referred to. The main ex})losions from the 

 summit craters have not only been vastly more energetic than anj^ 

 that have occurred at the sites of the ps(uidocraters, but the debris 

 blown out was incandescent, while the material thrown into tlie air by 

 the explosions in the hot accunudations in the ri\(M- valleys has not 

 been observed to be even red hot. The minor eruptions from the 

 summit crater, however, may l)e due to precisely the same immediate 

 cause as the eruptions of the pseudocraters — namely, the access of 



