328 E. 0. Hovey — Erujptions of 190^ of La Soufriere^ 



tropical showers percolating through the beds have come into 

 contact with the still highly heated interior, causing violent 

 outbursts of dust-laden steam. We saw one of these outbursts 

 from the Wallibou Yalley send up a column of such vapor fully 

 a mile in height. The action lasted for nearly an hour. The 

 secondary eruptions illustrated by figures 7 and 8 took place on 

 a clear, dry morning and must have been caused by the perco- 

 lating river waters. On May 30 we witnessed the throwing of 

 a dam across the stream and the formation of a temporary lake 

 by a heavy secondary outburst of dust-laden steam from the 

 lapilli-bed in the Wallibou Valley. This eruption is illustrated 

 in fig. 7. After the eruption ceased the little lake soon rose 

 to the top of the dam and quickly cut its way down to the 

 old level, sending a ^' mud-flow " down the gorge to the sea. 

 Such a lake in the valley of the Kabaka Dry Kiver cut its new 

 outlet through a narrow ridge of the old agglomerate constitut- 

 ing the wall of the canyon, forming as it did so a beautiful series 

 of channel-bowls, pot-holes and scratched corkscrew channels. 



When we first reached St. Vincent, the dust, especially that 

 covering: the Richmond estate, showed in marked manner the 

 wind-drift surface so familiar in the case of freshly-fallen snow, 

 and in many places these drifts were from three to four feet 

 deep (see fig. 6). There were several heavy rains between 

 May 24 and 29, so that the appearance of the surface was very 

 different on May 30 from what it was when I first saw it. Its 

 drifted character was not nearly as evident, and the beautiful 

 dendritic drainage, which was already in evidence on May 24, 

 had been greatly extended and intensified. Geological opera- 

 tions, which under ordinary conditions are performed so slowly 

 as to be imperceptible, were being carried forward rapidly 

 under our very eyes. One item of interest was the action of 

 the Wallibou River itself under the influence of the loose dust 

 and lapilli along its banks.^ Its waters became so overloaded 

 with sediment that they could only flow in pulsations, showing 

 that intervals of time were needed by the stream to gather 

 strength to force its way along with its load. On May 24 

 these waves or pulsations were from fifteen to forty seconds 

 apart. Such mud streams carried large bowlders down the 

 river bed to the sea. 



When the great cloud of ejecta rose from the Soufriere at 

 2 P. M., May 7, the portion which was traveling eastward 

 seemed suddenly to split, according to the accounts of eye- 

 witnesses, when it was some distance beyond the island, and 

 to send a part back to the land. This accounts for the fact 

 that unprotected windows in the eastern side of houses in the 



* First described by the author in the New York Times, June 29, 1902, 



