The facts and photographs we collected were baffling. They did not correspond with the text books. Two volcanoes a hundred miles apart suddenly spouted death downward. Obviously they were connected along the island chain, with ocean to the east and ocean to the west. Telegraph cables were broken. Why? That which lay under the ocean was totally unknown, both events and topography. The biggest part of these volcanoes was submarine. Earthquakes at Pelée were relatively small but often continuous. Tidal waves were local and accompanied by downblasts of steam. The downblasts were at first supposed to be due to fallen avalanches from the upblasts. Then it appeared they were really sloping jets from concealed holes or cracks in the gulches, with inclined orifices amid the blocks of a cracked-up mountain. For at Pelée the blast that destroyed St. Pierre shot from the crater gulch in cascades of water and steam, while observers on high ground saw the horizon, or clear sky, over the crater. The speed of the blast was six miles in two minutes, or 180 miles per hour. This was different from the glow clouds in the later months, migrating slowly along cracks in the gulch bottom. Man’s perception of speed relative to himself has nothing to do with actual speeds. It may be argued that a miniature volcano erupts faster than a big volcanic system, but not if the whole terrestrial plexus of systems is taken into account. An eruption of Mauna Loa is a very slow affair, in comparison with the 10,000 underground squirtings of lava in cracks totally unperceived, except as tremors on seismograph. Pelée’s eruption was like turning on a hose. A structural valve or orifice, suddenly opened by underground heaving of the mountain block and letting out steam and mud, appears to be the only reasonable explanation of what happened. And the only agents possible were glowing stiff lava heating boiling water underground. Both of these were later identified. Grove Karl Gilbert of the U.S. Geological Survey, who had criticized favorably my manuscript on the Black Hills intrusive lavas, wrote me not to drop the enigma of Mount Pelée, because he found the published reports unsatisfying. In 1949, forty-seven years after the disaster, I published “Steam blast eruptions,” dealing with Pelée. In the interim I studied many volcanoes. Alexander Agassiz, who had been urging me to do a memoir on volcanoes, financed a trip to Vesuvius when it exploded and poured out lava in 1906. Ottajano northeast of Vesuvius was demolished by jets of gravel and stones; and Boscotrecase at the south was invaded by black streams of heavy, sprouting, bouldery slag. Here was a change of habit, from heaping up lavas for thirty-four years, to collapse, internal avalanching, and pure steam explosion accompanied by remnants of stirred lava flow.