there were noisy rumblings and steam was emitted; actual eruption is reported from April the twenty-sixth on, and yet the city was not evacuated. In St. Vincent local earthquakes have been on the increase for a year in the neighborhood of the volcano; people were actually frightened away from Windsor Forest on the northwest slope of Soufrière, as far back as May, 1901, by rumblings and quakings. The water of the lake has been seen to bubble and sulphurous coatings are described as being deposited on the rocks. So violent were the signals early in May, especially when the news of the Guerin disaster came from Martinique, that the leeward slopes of the Soufrière, at all times sparsely inhabited, were abandoned. Hence the small loss of life on that slope. On both islands, if the respective governments had maintained vulcanological stations with instruments, doubtless, there would have been perceived a gradually increasing series of signals of different sorts, tremors, sounds, sights, smells and temperatures. This record, if we had it now, would be invaluable. It cannot be reconstructed.
The question 'What actually happened in the eruptions?' involves no very great difficulty. The eruptions were small manifestations of a common type. Probably the Plinian eruption of Vesuvius was similar; the Central American volcanoes have quite the same history;. Krakatoa in 1883, Tarawera (New Zealand) in 1886, and Bandai San in Japan in 1888, all present phenomena identical in the main, with local variations. A slip of some sort liberated a steam column; the cause of the fracture or the source of the steam is one step too far back into theory to venture to treat it here. Release once started followed old vents, water-holes, and these vents were Soufrière and Pelée. The explosion that followed release of pressure tore away the walls of the fissure and its violence ground the material to powder. The material came from a depth where the rocks were hot, and it was heated further by friction. Those who saw the eruption of Pelée on the twentieth of May describe a black column of dust and rocks that looked like smoke with wonderful purling, interbillowing nodes that overrode each other like cauliflowers or 'brains'; this column shot up silently at first, followed by heavy detonations that finally became a continuous roar. The column was estimated by Lieutenant McCormick of the Potomac to be at least five miles high; he witnessed the spectacle from Fort de France. Lightning shot through the great billows in all directions in a network. When it reached its maximum height the column spread out like a flower on its stalk and the upper edges of the hard smoke-steam masses were lighted white by the rising sun. A perceptible cold wave was felt. A shower of gravel took place followed by fine dust which continued falling for an hour. This was the day of the funeral of Consul Prentis. A visit to St. Pierre later the same day showed that a terrific blast had
